101st Annual Conference Program

Mobile dental preventive programs aim to provide services to children who may not otherwise receive services due to geographic area or financial burdens. Using a collaborative model including a dentist, hygienists and dental hygiene students, children are able to receive preventive services including exams, cleanings, sealants, fluoride and proper education. By preventing cavities and other oral diseases in vulnerable populations, these programs help reduce the need for costly restorative or emergency treatments. In turn, this lowers the overall economic burden on families, schools, and the healthcare system while promoting long-term oral health.

Using data on 199,464 living organ donors during 1987-2024 in the US, this paper studies the role of social distance in the effects of gender on altruism. Leveraging the donor-recipient relationship, the analysis shows that women are more altruistic in the low social distance relationship where donors and recipients are socially closer. In particular, women donate to immediate and other family members such as children, spouses, and siblings more than to friends and anonymous recipients, relative to men. Even restricting the sample to friends and anonymous recipients, women donate to friends more than to anonymous recipients, relative to men. The paper then discusses implications for living organ donation.

The Special Smiles program is an oral health initiative designed to improve access to dental care for individuals with intellectual and developmental disabilities. Volunteers provide comprehensive oral health care, including free dental screenings and education to Special Olympics athletes and family members. By offering these services at no cost, the program helps to minimize the financial load associated with routine dental visits. Being able to educate athletes and their family members by offering practical strategies to support daily hygiene improves long-term oral healthcare.

Abstract Billions worldwide lack supply of basic drinking water, water from an improved source that does not take more than 30 minutes of collection time including queuing. Poor access to basic, clean water impedes national progress, resulting in a slower economy, worse health outcomes, and shorter lifespans. A complex set of factors affect the supply, including climate change, economic factors, population growth, poor policy coordination, and freshwater accessibility. Using panel data of countries with historical rural water deficits, our study evaluates the relative significance of factors determining the supply of safely managed drinking water and draws critical policy implications.

Equity returns exhibit persistent volatility and pronounced downside asymmetry that matters for how risk is measured and managed in real time. Standard stochastic volatility models with leverage typically attribute asymmetric return behavior solely to volatility–return feedback while imposing symmetric, parametric innovations. This paper develops a semiparametric Bayesian stochastic volatility framework that allows both leverage effects and asymmetric return innovations, relaxing the common restriction that all downside risk operates only through volatility dynamics. By rewriting the model in an unobserved-components form and approximating unknown disturbance distributions with Dirichlet process mixtures in both the measurement and state equations, the approach flexibly captures heavy tails and skewness without committing to restrictive functional forms. Empirical study using U.S. equity return evidence indicate that leverage and innovation asymmetry can coexist, are empirically distinguishable, and that excluding innovation asymmetry can affect predictive densities and inferred tail risk. Model comparison using WAIC favors specifications that account for distributional asymmetry, implying that policy-relevant assessments - such as stress-test loss projections, market-based risk indicators, and the calibration of countercyclical capital or margin tools - can be sensitive to whether downside risk is modeled as volatility transmission, innovation asymmetry, or both. More broadly, the results suggest that improving the distributional realism of volatility models can strengthen the informational content of risk metrics used in macro-financial surveillance and regulatory decision-making.

This research demonstrates how Artificial Intelligence could help investment professionals systematically allocate to international stock portfolios. Robust modeling approaches are needed to accurately forecast stock price dynamics. By expanding on Maranets (2024)1, this paper studies the benefit of using Recurrent Neural Networks (RNNs) – a machine-learning technique – to forecast daily returns of eleven G10 ETFs and construct long/short ETF portfolios using predicted return as an indicator. The research follows two paths: RNN model parameter tuning and ETF portfolio allocation using the best tuned models. In the first path, the entire sample from 2011 to 2022 is divided into three parts: initial training, validation, and testing. The selection of optimal RNN model parameters operating only on ETF historical price data without additional explanatory variables is performed on a composite ranking of three error metrics (MAPE, RMSE and MAE) over two iterations within the validation sample and following a greedy approach. In the second path, five daily rebalanced high-minus-low portfolios of G10 ETFs are constructed in the testing sample based on predicted return ranking from best-tuned RNN models in the first path, and portfolio performances are compared to a benchmark (iShares MSCI World ETF). This research builds on Maranets (2024) findings by introducing the epoch optimization as the third hyperparameter in addition to tuning of lookback window and LSTM units (the first and second hyperparameters). This paper studies the marginal improvement in traditional loss functions (MAPE, RMSE and MAE) of RNN LSTM per number of epochs in validation sample, detects the optimal number of epochs for training RNN return prediction models for G10 ETFs and reconstructs the Maranets (2024) framework for five high-minus-low portfolios of G10 ETFs with epoch optimization as the third hyperparameter. Per marginal improvement study, this research concludes that 30 to 40 epochs is the optimal number for training RNN LSTM return prediction models for G10 stock ETFs. Portfolios of Top 2/Bottom 2 and Top 5/Bottom 5 ETFs demonstrate the best volatility-adjusted returns and outperform the benchmark on volatility-adjusted basis with a low, positive correlation to the benchmark.

This study investigates the effects of governments’ containment and health policies and economic support measures on economic performance during COVID-19 in Taiwan, Japan, and South Korea. Using monthly OxCGRT indices and national Composite Coincident Indices (CCI) for 2020–2022, we apply unit root and cointegration tests, followed by bivariate VAR models on differenced series, Granger causality tests, and impulse response analysis. Results indicate heterogeneous relationships. In Taiwan, both policy indices Granger-cause CCI changes, with containment shocks producing positive and persistent effects on economic performance after 2–8 months. In Japan, bidirectional causality is found, and stricter containment measures exert temporary negative effects lasting around four months, with no significant role for economic support. South Korea shows similar negative containment effects (approximately five months) without reverse causality or detectable economic support impacts. The evidence suggests that Taiwan’s early implementation of relatively strict containment and health measures, without nationwide lockdowns, was associated with stronger economic resilience compared to the more prolonged stringent policies in Japan and South Korea. Economic support measures display complex effects that aggregate indices struggle to capture, possibly due to high saving rates and varying policy goals. These findings highlight the need for calibrated policy responses that balance public health objectives with economic stability and point to the value of disaggregated analyses for future pandemic preparedness.

This paper investigates nonlinearities in the inflation-growth-uncertainty relationship in Chinese provinces over the period 1992 to 2017 using nonlinear models and dynamic panel threshold models. We find that for the full sample period (1992–2017), inflation rates exceeding 9.7% are associated with a positive growth effect. Below this threshold, the correlation is insignificant. Since inflation rates above 9.7% were mainly observed in the early to mid-1990s, we restrict the sample to 1999–2017. In this period, the inflation threshold lowers to approximately 5.1%. Moreover, the relationship between inflation and growth shifts across the two regimes: below 5.1%, inflation is positively associated with growth, while above 5.1%, the effect turns negative and statistically insignificant. We further explore whether the effect of inflation on growth could be affected by uncertainty at the provincial level. For that purpose, we combine two recent uncertainty indices for the Chinese economy that are based on Chinese newspapers. We find that inflation only has a positive effect on growth for low levels of uncertainty. For high levels of uncertainty, the effect of inflation on growth turns negative and statistically insignificant.

This study investigates whether remittance inflows mitigate gender disparities in labor market outcomes across a panel of African economies. We apply a dynamic model to macro-level data on remittance receipts, gender-disaggregated labor force participation rates and wage employment, and covariates, including GDP per capita, dependency ratios, female secondary schooling attainment, and institutional quality indices. We employ the system GMM estimator to address endogeneity from potential reverse causality or unobserved confounders. Preliminary estimates reveal that remittances are associated with diminished aggregate participation in wage employment continent-wide and labor force engagement in North Africa, disproportionately affecting women, perhaps suggesting that such inflows substitute for market labor and facilitate withdrawal among those burdened by caregiving, rather than catalyzing formal sector entry. This suggests that such transfers alone fail to attenuate, and may exacerbate, gender gaps absent complementary interventions like subsidized childcare or gender-targeted vocational training.

Gender equality in labour markets remains a central development priority because access to productive employment, earnings opportunities, and economic empowerment are fundamental to individual and national welfare. Despite global progress, persistent gender disparities continue to shape employment opportunities, occupational mobility, wage outcomes, and access to productive resources. These inequalities limit women’s economic agency and constrain aggregate growth by preventing full utilisation of human capital. Using nationally representative data from the Annual Household Income and Expenditure Survey (AHIES) 2022–2024, this study examines the extent and determinants of gender gaps in labour market participation, employment quality, and earnings in Ghana. The analysis highlights structural, institutional, and socio economic barriers that reinforce gendered labour market outcomes, with pronounced disparities in informal employment, sectoral allocation, and income stability. The findings underscore the need for targeted policies that promote equitable labour market participation, enhance women’s economic empowerment, and support inclusive development.

Since price discrimination and selling below cost arise in the normal course of business and are usually legal for home firms, countering these practices by foreign firms provides a possible rationale for antidumping duties. If antidumping duties were to provide a systematic defense against predation by foreign firms, however, a strong ''fair-trade'' justification would remain. This is especially relevant in the current global trade situation where, e.g., European markets are flooded by Chinese producers heavily subsidized by their government. This paper adapts the classic entry-deterrence analysis of Dixit (1979) and Brander and Spencer (1981) to provide a simple treatment of predation, which is applicable with price leadership as well as quantity leadership. Although situations of cross-border predation appear to be quite rare, foreign firms may sometimes find themselves in leadership positions if they have to make shipments and/or set prices before their home rivals. This paper shows that, in the context of such an international leadership game, predation may occur without dumping and visa versa. Further, when dumping and predation do coexist, a sophisticated form of antidumping duty would prevent predation, but the simple antidumping duties that are generally observed in practice will often be insufficient. Consequently, the paper challenges the ''fair-trade'' view of antidumping policy as an antidote for predation.

Health is a central determinant of economic well-being, influencing individuals’ capacity to work, productivity, and long-term earnings. Poor health can interrupt labor market participation and deepen household vulnerability, particularly in settings where social protection is limited. Recent evidence, highlights the reinforcing cycle between poverty and adverse health outcomes, underscoring the importance of understanding how health shocks shape economic behavior. This study examines the impact of health shocks on labor supply decisions in Ghana using nationally representative data from the Annual Household Income and Expenditure Survey (AHIES) 2022–2024. The analysis explores effects on labor force participation, hours worked, and sectoral transitions, with attention to gender and spatial disparities. Results indicate that health shocks significantly reduce labor supply and increase income instability, especially among informal sector and rural workers. The findings highlight the need for integrated health insurance and social protection policies to strengthen household resilience.

Artificial intelligence is widely expected to become the defining general-purpose technology of the coming decades, with the potential to reshape productivity, the structure of production, and the distance between rich and poor economies. Whether Sub-Saharan Africa shares in those gains will depend far less on whether the region produces frontier AI than on whether its economies are equipped to absorb, deploy, and use AI-related technologies productively. This paper asks whether the data infrastructure on which such technologies depend has begun to raise income growth across forty-eight Sub-Saharan economies between 2007 and 2023. It builds an AI Preparedness Proxy Index that isolates the connectivity and data environment through which AI diffuses, and it keeps skilled labor and institutional quality as separate conditions rather than folding them into a single readiness score. The central result is sobering. Despite a decade of visible gains in connectivity, there is no evidence that improved preparedness has translated into faster growth, no sign that the payoff is larger where skilled labor is more abundant, and no acceleration in the years since modern AI entered general use. Trade openness is the only feature of these economies that is reliably associated with growth. Read in economic terms, the absence of a growth dividend is not reassuring but a warning. It indicates that most economies in the sample remain below the thresholds of skill, institutional quality, and productive use at which digital access begins to pay, and that on present trajectories they are poorly positioned to capture value from the AI transition. Closing that gap will require more than expanding connectivity. It calls for advanced skills, the absorptive capacity that turns access into productivity, and adoption anchored in export-oriented and other tradable activities where the returns to technology in the region are most visible.

When UAVs serve as aerial base stations in disaster response, they must allocate users, subchannels, and transmit power in a decentralized manner without explicit coordination. We extend the stochastic UAV allocation game formulation with bounded reward shaping drawn from three game-theoretic families: (1) auction mechanisms (Vickrey, posted-price) that induce efficient task allocation through pricing, (2) coordination games (Stag Hunt, Chicken) that balance collision avoidance against cooperative teaming, and (3) social dilemmas (Public Goods, Trust) that model shared costs and reciprocity. Our contributions are threefold: we prove that tabular Q-learning convergence is preserved under bounded shaping when rewards are clipped to [−M, M ]; we establish that incentive properties from the source games (truthfulness, equilibrium structure, free-rider dynamics) carry over to the shaped UAV setting; and we validate empirically that learned policies approach brute-force optimal joint rewards on small instances. Centralized tabular training falls within our theoretical guarantees, while independent learners show empirical convergence.

Costly peer reward mechanisms—under which all subjects in a group incur a cost to reward one another after observing their group members’ contribution decisions—have received considerable attention as an important mechanism of promoting cooperation in public goods games. However, existing studies typically assume a complete reward network, in which each subject is able to reward every other group member. In this paper, I relax this assumption by varying the density of the reward network, defined as the number of potential rewarders each subject has within the group, and examine whether the structure of peer reward networks affects the level of cooperation within the group. The experiment considers three reward network structures that differ in density: a complete reward network, in which each subject has three potential rewarders; a circle reward network, in which each subject has two potential rewarders; and a pair reward network, in which each subject has only one potential rewarder. In this experiment, subjects participate in a four-player public goods game. A total of 192 subjects voluntarily participated in the experiment. The experiment consists of four treatments: baseline, complete, circle, and pair. Each treatment includes four sessions, and each session consists of 15 periods. Each subject participates in only one treatment and one session. The experiment employs a partner-matching design; thus, group composition remains fixed throughout the session. Each period in the complete, circle, and pair treatments consists of two stages: a contribution stage followed by a peer reward stage. The baseline treatment includes only the contribution stage. According to the experimental results, the introduction of peer rewards significantly increases contributions to the public good, regardless of the density of the reward network. Furthermore, among the reward networks examined, individuals in a complete reward network receive higher reward amounts and face a higher probability of being rewarded, whereas individuals in a pair reward network experience greater reward intensity. With respect to efficiency, the results indicate that the implementation of peer reward mechanisms enhances efficiency across all types of reward networks.

This paper introduces a residual-based Shapley Population Variable Importance Measure (Residual SPVIM) framework for variable-wise specification testing in econometric models. The proposed procedure evaluates whether individual covariates remain systematically predictive of model residuals after fitting a working model, thereby identifying variable-specific functional-form misspecification. Building on recent advances in Shapley-based variable-importance inference, the method uses flexible machine-learning estimators to detect remaining residual structure while preserving the interpretability of correctly specified model components. Unlike classical omnibus specification tests, which provide only a global assessment of model adequacy, the Residual SPVIM framework localizes misspecification to the offending covariates. The procedure can detect omitted nonlinearities, interactions, and threshold effects while retaining the coefficient-based interpretation central to applied econometric analysis. Monte Carlo simulations demonstrate strong control of Type I error under correct specification and high power against both global and localized departures from the working model. More broadly, the paper shows how Shapley-based variable-importance methods, typically used for predictive interpretation, can be repurposed for formal econometric diagnostics. The proposed framework therefore provides a bridge between interpretable econometric modeling and modern machine-learning-based diagnostic tools.

Real-world cartels often mix closely monitored core members with loosely monitored peripheral participants—OPEC insiders face formal production audits while non-OPEC allies face far weaker verification; firms in high-enforcement jurisdictions collude alongside those in low-enforcement ones. Yet the theoretical literature typically treats the probability of detecting a member’s deviation as uniform across firms. We introduce firm-level heterogeneity in detection probabilities into a repeated quantity-setting game. Because outsiders are harder to monitor, they require a strictly larger minimum quota share than insiders to refrain from cheating—a loyalty rent that the cartel must pay to secure outsider compliance. When the monitoring gap is large enough, this rent turns positive: outsiders receive more than an equal share, compressing insider quotas and potentially rendering collusion infeasible. Endogenizing membership reveals a striking result: the formal alliance can collapse while tacit collusion survives under degraded monitoring, so institutional breakdown need not end cooperation. For policy, we uncover a welfare paradox: stronger internal monitoring unambiguously benefits the cartel, but harms consumers by enabling deeper output restriction—so policies that limit information exchange among cartel participants can destabilize existing collusion.

Lung cancer is the leading cause of cancer-related deaths worldwide. To combat this, the U.S. Preventive Services Task Force recommends annual screening for high-risk in- dividuals, based on findings from the National Lung Screening Trial (NLST). However, self-selection may reduce the effectiveness of screening in lowering mortality rates. This paper develops a dynamic discrete-choice model that incorporates both NLST partici- pation and lung cancer screening decisions to analyze the factors influencing screening behavior and how self-selection affects health outcomes and costs. Using data from NLST and the National Health Interview Survey (NHIS 2015), the model examines beliefs about lung cancer risk, survival rates, and the costs and benefits of screening, including uncer- tainties caused by false positives and false negatives. The findings show that individuals with lower lung cancer risk are more likely to undergo screening, driven by their unob- served preferences for proactive health investment, which are strongly correlated over time. While trial participants who undergo screening experience lower mortality rates, individuals in the general population who screen show higher mortality rates. Coun- terfactual analysis reveals that lung cancer survival benefits are limited, making 100% screening uptake unlikely. However, targeting underrepresented groups can reduce mor- tality by 27% at minimal cost, and annual lung cancer screening remains the most effective strategy for saving lives.

This paper examines whether crypto adoption weakens monetary-policy transmission through the banking system. Using an unbalanced annual country-level panel for 2020–2024, we combine a country-level crypto-adoption index with policy-rate, banking, macroeconomic, and institutional data. The primary dependent variable is the bank interest-rate spread, defined as the lending rate minus the deposit rate, which captures a key margin through which policy-rate changes are transmitted to borrowers and depositors. We estimate fixed-effects panel models with country and year effects and test whether the policy-rate/spread relationship varies with crypto adoption. The results indicate that crypto adoption weakens this relationship. In both contemporaneous and lagged crypto specifications, the interaction between the policy rate and crypto adoption is negative and statistically significant. Marginal-effects estimates show that the policy-rate effect on spreads is positive when crypto adoption is low, declines as adoption rises, and turns negative at high levels of adoption. Inflation results provide supporting evidence, with positive interactions between lagged policy rates and crypto adoption. Dynamic robustness checks reinforce the main finding: the lagged-crypto interaction remains significant after controlling for spread persistence, and the error-correction model shows that crypto adoption weakens the short-run response of spreads to policy-rate changes. The findings suggest that crypto adoption does not make monetary policy ineffective, but it may weaken an important bank-based channel of transmission.

This study examines the long- and short-term interest rate pass-through under low-interest-rate environments in the U.S. and Hong Kong from 2000 to 2023. We compare the pass-through mechanisms between policy rates, interbank rates, and retail lending rates in both economies. Our findings reveal that when policy rates approach zero lower bound, interest rate pass-through is hampered slightly in the United States, but much more in Hong Kong. Since Hong Kong dollar is pegged to the US dollar, monetary policy in Hong Kong follows that in the United States. The difference in the interest rate pass-thorough is probably due to different levels of competitiveness in the banking system.

This paper examines how firms form and update their inflation expectations and how these expectations influence price-setting behavior. Drawing on unique firm-level panel data from the Bangko Sentral ng Pilipinas (BSP) Business Expectations Survey from 2011 to 2024, the study offers novel micro-level insights from an emerging market perspective. Results show that firms’ expectations closely follow actual inflation but respond asymmetrically, i.e., overreacting to inflation surges while underreacting to disinflation. Domestic-oriented firms respond strongly in adjusting their prices when inflation shocks occur, while trade-oriented firms are more attentive to changes in external variables such as the exchange rate. While expectations play a significant role in price setting, adjustments appear to be downward-rigid. These findings underscore the role of information frictions and firm heterogeneity on inflation dynamics and highlight the importance of clear, targeted central bank communication for effective monetary policy transmission.

This paper investigates how uncertainty shocks originating in emerging market economies (EMEs) propagate internationally, and how trade integration and capital-account regimes shape the magnitude and persistence of these spillovers. While the literature has traditionally emphasized uncertainty shocks from advanced economies—particularly the United States—much less is known about the international macroeconomic consequences of uncertainty arising within EMEs themselves. Yet the cyclical properties of EMEs, characterised by recurrent stagflation, counter-cyclical trade balance, and capital-flow reversals, suggest that uncertainty shocks in these economies may generate markedly different transmission patterns. We combine empirical evidence from a two-step panel VAR with a two-country DSGE framework to study these mechanisms. Empirically, we identify domestic uncertainty shocks in China—a large EME with deep regional trade ties—and examine the spillover effects on eight Asian EMEs over 2010–2020 based on monthly data. The results show a complementarity pattern in the international transmission of uncertainty: rather than inducing capital flight toward neighbouring EMEs, a rise in Chinese uncertainty generates synchronized downturns across the region. Recipient economies experience output contractions, declining net exports, falling foreign reserves, and significant net capital outflows. These effects are systematically stronger among economies with higher bilateral trade integration, underscoring trade linkages as the dominant propagation channel. To rationalize these findings, we develop a two-country DSGE model featuring production networks through imported inputs, and credit markets subject to capital controls. The model replicates the empirical recessionary and stagflationary responses and highlights a negative trade-feedback loop: uncertainty in the source economy suppresses its import demand, which depresses the trading partner’s activity, generating further feedback to the source economy. Finally, we assess the role of financial liberalization. Capital inflow liberalization mitigates the real effects of uncertainty by dampening the rise in domestic lending rates and easing the contraction in imported inputs, while outflow liberalization produces competing forces that generate only muted net effects. Overall, our results highlight the trade network and capital-account regimes in shaping the international transmission of uncertainty within emerging markets.

This paper examines how households combine public inflation information with experience-based beliefs when forming inflation expectations, and whether the gain placed on public inflation signals varies with the inflation environment. We exploit a unique feature of the Bangko Sentral ng Pilipinas Consumer Expectations Survey that elicits expectations before and after respondents are informed of the latest official inflation rate, allowing us to directly identify expectation updating in a naturally occurring information environment. Using quarterly data from 2010 to 2024, we show that the gain households place on public inflation information increases with inflation persistence, implying that expectation updating is state dependent. This finding contrasts with the constant-gain implication of standard noisy-information models and provides empirical grounding for frameworks in which attention to public signals varies with macroeconomic conditions. Expectations formed after exposure to public inflation information also respond less strongly and less persistently to inflation shocks during high-persistence periods. The findings suggest that greater belief on common public inflation signals strengthens expectation anchoring and dampens the propagation of inflation shocks through expectations.

This paper analyzes the economic consequences of International Monetary Fund – supported programs across selected Asian economies, with particular emphasis on Pakistan as a recurrent program participant. It investigates whether these stabilization programs have fostered economic growth, macroeconomic stability, and strengthened institutional capacity in Pakistan relative to its South Asian peers. Pakistan remained reliant on repeated IMF stabilization programs and its regional peers Bangladesh and India successfully reduced and exited IMF engagement over the years. The analysis reports the policy question: What structural, fiscal and monetary policy choices enabled India and Bangladesh to exit IMF dependence and why Pakistan has not achieved similar outcomes? The paper adopts a comparative empirical approach under the timeframe of 2009–2024, considering the repeated IMF programs in Pakistan and sustained macroeconomic association in neighboring economies. The study used fixed-effects panel regression and difference-in-differences (DiD) estimation to asses the macroeconomic effects of IMF program participation, prioritizing on growth, inflation volatility, fiscal and external balance, export performance, and social spending. Propensity Score Matching (PSM) is applied to analyze counterfactual results for Pakistan. The analysis based on data from the IMF, World Bank, Asian Development Bank, and national monetary and fiscal authorities. The findings suggest that India and Bangladesh prioritized revenue mobilization, energy sector reforms, predictable exchange rate management, and export diversification inaddition by political feasibility. While in-contrast, Pakistan relied heavily on fiscal compression, price adjustments, international market instability, and structural constraints (tax expansion, energy sector inefficiencies). The paper concludes the institutional weaknesses and political economy constraints for Pakistan’s continued reliance on IMF programs. It also outlines the practical sequencing of reforms, revenue generating and sharing, export incentives, energy pricing and protection of social spending for Pakistan further adding to the policy debates on IMF programs designs for achieving lasting macroeconomic stability. Keywords: Political economy, IMF programs, Macroeconomic stability. JEL Classification Codes: E62, F33, O23

This study examines the impact of capital gains from real estate, gold, and equities on private final consumption expenditure (PFCE) in India, where household wealth is concentrated in tangible assets rather than equities as in advanced economies. A mixed-methods approach is used in the study. A Vector Error Correction Model (VECM) analyses quarterly data from 2011–12 to 2023–24 to study long-run and short-run relationships among PFCE, GDP, gold prices, housing prices, and the Nifty 50 index. A primary household survey complements this by assessing perceptions of changes in wealth, spending behaviour, realised and unrealised gains, and responses to interest-rate shifts. The impulse response analysis from the VECM shows that GDP and the NIFTY index exert strong positive effects on PFCE, with both generating significant initial increases in consumption that persist over the long run, reflecting the roles of economic growth and wealth creation in stimulating household expenditure. Gold prices also have a positive, though comparatively moderate and stable, impact on PFCE, suggesting that rising gold prices enhance perceived household wealth, thereby supporting consumption. In contrast, housing shocks have a persistent negative effect on PFCE, as increased housing-related expenditures or investments appear to reduce disposable income available for other consumption, potentially leading to crowding out. Survey findings show that moderate interest-rate changes have limited immediate effects, though lower rates encourage shifts toward equities, gold, and real estate. Realised gains show weak correlation with spending, indicating that unrealised wealth changes exert a stronger influence. Policymakers should therefore monitor asset-price trends as early signals of shifts in aggregate demand.

We study how access to local technology talent shapes banks' internal organization of financial technology (FinTech) activities. Using detailed job-posting data from China's largest recruitment platforms, we examine how city commercial banks allocate FinTech-related hiring across locations within their branch networks. Exploit-ing variation in banks' exposure to local IT agglomeration and instrumenting for local technology talent using the historical establishment of computer science programs, we show that banks allocate more FinTech hiring to non-headquarter cities with stronger IT ecosystems. This pattern is more pronounced for banks whose headquarters lack inherent IT advantages. Posting-level evidence indicates that FinTech hiring in IT-intensive cities is associated with higher skill requirements, consistent with labor market pooling, and is linked to a reallocation away from more easily automatable tasks. Together, the results highlight an organizational channel through which local technology ecosystems shape the diffusion of FinTech within traditional banking.

In a randomized control trial of customer-supplier firm pairs in New Zealand, we treat with information one firm in a pair and analyze its effects on expectations and actions of the directly treated firms (direct effect) and connected firms that did not directly receive information (spillover effect). We find strong spillover effects in expectations and actions. We show that the spillover effects on the connected firms’ expectations are driven by inter-firm communication, as opposed to observable actions. We embed inter-firm communication along the supply chain in a New Keynesian pricing problem and discuss the macroeconomic implications of input-output network-specific communication.

This paper investigates the macroeconomic consequences of biased expectations in banking, conceptualized as a form of bounded rationality. Using analyst-level earnings forecasts for U.S. banks between 2000 and 2019 and market reactions to stress-scenario announcements, I present empirical evidence challenging the Rational Expectations hypothesis and documenting systematic bias in bank projections. Bayesian model selection shows that both actual and forecasted earnings are best captured by dense models, yet persistent forecast errors suggest that bias stems from model complexity rather than omitted-variable bias. To assess aggregate effects, I embed adaptive learning into a New Keynesian framework with financial intermediaries, where agents employ ridge regression to mitigate multicollinearity. I introduce the notion of a Biased Perception Equilibrium, in which agents accept bias induced by regularization. This equilibrium yields muted bank responses and inflated franchise valuations under adverse shocks. Quantitative simulations reveal that post-2008 regulatory reforms heightened moral hazard and the banking equity premium, while biased beliefs tempered deleveraging. Furthermore, biased expectations impair the transmission of expansionary monetary policy by dampening the intertemporal substitution channel, reducing stimulus effectiveness by up to 12.5 percent under an ad hoc one percent maximum bias. These findings underscore how belief formation under model misspecification can distort financial dynamics and policy efficacy.

This paper provides a mathematical explanation for the flaws in the formulae used for computing the mean and variance of the returns of the portfolios as given in the Modern Portfolio Theory of Markowitz (1952) and also for the inefficacy of conducting regression analyses of the portfolio returns with respect to the market portfolio as advocated by Jensen (1969). Consequently, we get increasing mean price for increased logical sophistication of the active management strategy for portfolios indicating that it is possible to pick up the winners in an efficient market if accepting the additional risk that accompanies such strategies.

We study whether and how innovation intensity attracts foreign portfolio investment (FPI) using a panel of 60 countries from 1997 to 2021. Employing an instrumental variables strategy based on regional shift-share and global push instruments, we estimate the causal response of foreign debt and equity to patenting intensity. We investigate heterogeneity by development and institutions, as well as dynamics via local projections. Innovation increases net FPI, with larger effects for foreign equity than for debt, and the inflow of foreign equity rises with levels of development and with institutional quality. Moreover, our results suggest that a net foreign debt inflow is only supported at extremely high levels, with intermediate or low institutional quality and development showing no effect. FPI prefers to be allocated in countries with higher risk preference, as tested by legal origins, macroprudential policy levels, and financial openness. Our dynamic analysis suggests that foreign equity responses are immediate and persistent, whereas debt responses are modest and fade over time. Results are robust to alternative measures, instruments, and samples.

Over the past decade, green finance (GF) has emerged as a significant investment option for investors, including both green bonds (GB) and sustainable equities (SE). However, due to major uncertain events, economic policy uncertainty (EPU) and market volatility (VIX) with investor sentiment (IS) emerge as key drivers of market behavior, which ultimately affects the asset returns. Consequently, traditional equities (TE) and GF markets remain highly susceptible to uncertainties and IS. This study examines the role of EPU, VIX, and IS on TE, SE, and GBs using the daily data from 2015 to 2022. We take the MSCI World Index, S&P Dow Jones Sustainability Index, and the S&P Green Bond Index as proxies for TE, SE, and GBs, respectively. Moreover, IS is constructed using Thomson Reuters Market Psych Indices (TRMI). The study employs a robust methodological framework comprising the nonparametric causality-in-quantiles approach (Balcilar et al., 2016) and the wavelet analysis. Our findings from nonparametric causality-in-quantiles reveal the existence of causality in EPU/VIX/IS–TE, EPU/VIX/IS–SE, and EPU/VIX/IS–GB. Specifically, the findings indicate that (i) EPU predicts returns negatively at a few lower quantiles but positively at extreme upper quantiles for both TE and SE, whereas VIX exhibits a consistently negative influence across all quantiles for these asset classes; (ii) IS affects TE positively at lower quantiles but negatively at upper quantiles, while it exhibits positive predictability for SE mainly during bearish phases; and (iii) GBs provide more substantial hedging potential than both TE and SE, showing limited sensitivity to EPU and IS, and responding positively to VIX during upper quantiles. Furthermore, the time-frequency wavelet coherence analysis complements these findings. It shows that EPU, VIX, and IS display similar behavior toward TE and SE, with relationships strengthening significantly between late 2018 to 2021 amid major global events such as Brexit and the COVID-19 pandemic. However, the impact of these indicators remains weak for GBs, with VIX having the most pronounced impact. Overall, our results have important implications for policymakers and market participants when considering sustainable investments, particularly during periods of heightened uncertainty.

Financial market development is critical to corporate economic behavior. We take the establishment of the Shanghai Commercial Paper Exchange (SHCPE) in China as a quasi-natural experiment. Using data from Chinese A-share non-financial listed firms from 2011 to 2022, we employ a difference-in-differences (DID) model to investigate the impact of the construction of a unified commercial paper trading market on firm R&D investment and its corresponding mechanisms. The study finds that after the establishment of the SHCPE, firms with higher pre-reform commercial paper usage experienced a significantly larger increase in R&D investment than firms with lower usage. Mechanism analysis reveals that the establishment of the SHCPE allowed firms to better utilize commercial paper to improve their operational and solvency capabilities, which in turn increased R&D investment. Heterogeneity analysis indicates that the positive impact of the SHCPE's establishment on firm R&D investment is more pronounced in firms facing lower financing constraints, operating in more competitive industries, and benefiting from better regional business environments. Our research highlights the positive effect of establishing a unified commercial paper trading market on firm R&D investment, offering evidence on the economic consequences and mechanisms of targeted financial infrastructure reform in developing countries.

This paper investigates the evolution of industrial sophistication in Japan by constructing a novel establishment-level measure of "advancedness" that bridges international trade complexity with domestic production data. We develop a crosswalk mapping Harmonized System (HS6) product complexity indices to Japan Standard Industrial Classification (JSIC6) codes, achieving a match rate of 96.2% across unique industrial codes. Using this mapping, we define establishment advancedness as the sales-weighted average of the complexity of products manufactured by a firm. Our empirical analysis of Japanese firms from 2013 to 2022 documents three stylized facts. First, there is a distinct upward trend in the mean advancedness index for all establishments, rising from approximately 0.08 in 2013 to over 0.21 in 2022. Second, we observe significant heterogeneity based on establishment size; larger firms (300+ employees) consistently exhibit higher advancedness levels compared to smaller firms, though all size categories show positive growth trends. Third, we analyze the extensive margin of firm dynamics, specifically, the entry and exit of advanced product lines. Regression results indicate that an establishment's prior advancedness level is a significant predictor of its product portfolio adjustments, showing a negative correlation with the probability of adding new advanced products. These findings provide new insights into how firms accumulate capabilities and the micro-dynamics of industrial upgrading in advanced economies.

This paper examines the economic impact of intellectual property rights (IPRs) in the context of complementary vertical innovation by firms in upstream and downstream industries. I provide empirical evidence that industry downstreamness is associated with 1) a greater intensity of product versus process innovation, and 2) a greater relative importance of patents versus trade secrecy in firm appropriation strategies. I develop a novel model of endogenous growth to explore the implications of these findings. The model predicts that strengthening patent protection generates downstream-biased technical change in most cases. The corresponding reallocation of resources across industries creates distributional effects from patent policy that existing analysis do not account for. I examine the impact of these effects on economic growth and social welfare, and explore their implications for optimal patent policy.

Rural households in developing countries use labor migration as a strategy to insure themselves against exogenous shocks. With the introduction of mobile money (MM) technology, remittances from migrants are now available to these households at a cheaper and faster rate. Previous literature documents the direct impact of MM use to insure against shocks, but does not investigate the externalities it generates. I argue that rural households simultaneously use two insurance technologies: risk pooling within their own village, and labor migration. MM lowers the transaction cost of the migrant channel, inducing households to substitute away from the village pool toward external remittances. This substitution generates a negative externality, as households under-invest in the pool relative to the social optimum. Using a quasi-experimental design with three waves of panel data from rural Bangladesh, I estimate the causal effect of MM adoption on consumption smoothing after flood shocks. I find that MM users experience 1.1 percent greater consumption loss per standard deviation increase in flood intensity relative to non-users, rising to 2.8 percent relative to households in villages with no MM penetration. I find new evidence that the misallocation extends beyond adopters, as non-adopting households in MM villages suffer comparable consumption losses after flood shocks, due to pool thinning with no compensating remittance gain. The findings suggest that policies to increase MM use in rural areas should accompany incentives to participate in the village pool, to correct the misallocation.

This paper develops a multi-sector New Keynesian dynamic stochastic general equilibrium (DSGE) model with heterogeneous agents and two regions to examine the cross-region effects of spatially targeted LTV and property tax policies on housing markets. We endogenize the location choice of residence by allowing agents to choose the place of residence. By using Bayesian methods applied to Taiwanese quarterly data from 2010Q1 to 2019Q4 to estimate the model, we find that under a canonical Taylor rule, both spatially targeted LTV and property tax policies induce borrowers to migrate from the urban area to the rural area, causing a decrease in the urban housing prices and an increase in rural housing prices. Moreover, total urban housing consumption rises and rural housing consumption declines. Wage and labor hours increase (decrease) in the urban (rural) area. We also show that the effectiveness of a policy that lowers the urban LTV maximum limit to suppress urban housing prices depends on the accompanied interest rate rules and the households’ ability to move across regions. If the central bank implements a constant-interest-rate rule or an augmented Taylor rule which incorporates housing inflation, reducing the urban LTV ratio does not moderate urban housing prices or stir rural housing prices. If the fraction of urban borrowers residing in the urban area is constant, urban LTV shocks cause negligible effects on housing prices and housing consumption in both regions.

Coverage of pension is low in developing countries largely due the existence of significant informal employment that constrains governments to collect adequate tax revenues. As a result, pension systems in developing countries are usually contributory restricted to only public sector workers and private formal sector workers. The sizable proportion of informal employees grow into older adults without access to guaranteed pension income in their old age leading old age poverty and income insecurity as older people continue to rely in labor income even in their frailty. Introducing social pension to support older people in the informal sector could reduce old age poverty and increase consumption among older people. However, financing such pension scheme in developing countries will involve large fiscal costs. This paper develops a stochastic dynamic general equilibrium life-cycle model in a two-sector economy including formal and informal employment with labor or longevity uncertainties in the absence of complete markets to investigate welfare and fiscal impacts of introducing social pension scheme for informal workers. We analyze the most appropriate tax to finance social pension for informal workers. The model is calibrated to Ghana, a country with more than 85% of informal employment. Our quantitative analysis suggests that social pension to informal workers will require about 0.42 percentage points increase in consumption tax rate, 3.41 percentage points increase in labor income tax rate imposed on formal workers and 1.13 percentage points increase in capital income tax rate. The policy generates aggregate social welfare improvement due to its positive impact on capital accumulation and redistribution mechanism of transferring resources from the richer formal workers to informal workers when it is financed by consumption and capital income taxes. However, we find that when the social pension policy is financed by increasing labor income taxes of formal workers, informal workers experience welfare gains at the expense of formal workers in the long run as the negative impact of taxes outweighs the positive impact of improved capital accumulation. The findings of this study suggest that inclusion policies such as social pension scheme can be growth-enhancing for wider social welfare in developing countries.

We study whether social capital (SC)—defined as norms that foster cooperation and prioritize collective wellbeing—enhances the market valuation of ESG disclosure and whether this effect depends on sociopolitical institutions. While SC should increase stakeholders’ demand for sustainability, its ability to translate into market discipline may depend on whether individuals perceive themselves as legitimate agents of corporate accountability for collective good beyond self-interest. We argue that power distance (PD) culture and perceived responsibility across ESG domains jointly condition this translation mechanism. Using a cross-country sample of 1,220 firms from 33 countries and a difference-in-differences identification strategy around the COVID-19, we examine how SC interacts with PD culture in shaping ESG valuation. In high-PD societies, social capital does not reinforce the valuation of any ESG disclosures. Under a low-PD culture, social capital enhances the firm value of social and governance disclosures, but not of environmental disclosure. The absence of impact on environmental disclosure valuation echoes with the perception that government is the primary agent for solving climate issues, weakening decentralized market discipline. The results show that not all types of disclosures are valued by SC, firm value of disclosure is affected by stakeholders’ sense of direct agent and cultural hierarchy. These results provide policy insights into the conditions under which disclosure-based mechanisms are effective for advancing different dimensions of sustainability.

This paper studies how organizational human capital and network capital influence firms’ social outcomes. While prior research has emphasized formal governance mechanisms, incentive contracts, and demographic characteristics of leadership, relatively little is known about how the quality of human capital within organizations—and the networks through which this capital is connected—affects firms’ engagement with employees and other social stakeholders. Using a large panel of U.S. publicly listed firms, we examine changes in organizational human and network capital arising from shifts in key personnel and leadership composition that alter firms’ accumulated expertise and external connections. We construct measures capturing both the depth of organizational human capital, reflected in professional expertise and experience, and the breadth of network capital, reflected in external affiliations and inter-organizational connections. These measures allow us to analyze how changes in organizational capacity translate into observable social outcomes. Our empirical approach exploits plausibly exogenous variation in organizational human and network capital and applies a difference-in-differences design comparing affected firms to matched control firms. We focus on firm-level social outcomes, including employee-related practices, workplace safety, and social controversy incidence. The results indicate that increases in organizational human capital are associated with significant improvements in social outcomes, particularly in areas requiring monitoring, coordination, and compliance. Moreover, these effects are stronger when human capital is embedded within broader organizational networks, suggesting that network capital amplifies the effectiveness of expertise by facilitating information flow and reputational discipline. Additional analyses show that the effects are more pronounced in socially sensitive industries and during periods of heightened stakeholder scrutiny. The findings are robust to alternative specifications and controls for firm performance, investment activity, and firm size. This study contributes to the literature on organizational economics and corporate governance by highlighting organizational human and network capital as key, yet underexplored, drivers of firms’ social behavior. The results suggest that strengthening organizational capacity may be as important as formal rules in achieving meaningful improvements in corporate social outcomes.

Residential communities, schools, and businesses frequently coexist with nuclear reactors, raising a central policy and research question: How does proximity to nuclear energy infrastructure affect local housing markets, and under what conditions does perceived risk outweigh potential economic return? This question has gained renewed urgency as nuclear power is re-embedded in U.S. decarbonization strategies, including large-scale reactor restarts and planned deployment of small modular reactors (SMRs) and microreactors to support data-intensive industries. If current momentum continues, millions of Americans may reside within regulatory Emergency Planning Zones (EPZs) of nuclear facilities in coming decades, making the economic consequences of nuclear siting a first-order policy concern. This paper contributes to the literature by reframing nuclear proximity not solely as a source of downside risk, but as a joint risk–return tradeoff that can simultaneously depress housing values through safety concerns and support them through employment, infrastructure investment, and energy reliability. While prior studies document mixed and often context-specific effects, ranging from sharp post-accident price declines to negligible impacts in stable operating environments, there is no comprehensive, nationwide empirical assessment of U.S. real estate markets near nuclear facilities. Methodologically, the paper employs a spatial econometric framework combining hedonic pricing models with difference-in-differences and spatial fixed effects to identify the causal impact of nuclear proximity on residential property values. The analysis integrates geocoded housing transaction data with detailed information on reactor locations, operational status, and NRC-defined EPZ boundaries. To capture the role of perception and societal readiness, the empirical models are augmented with sentiment measures derived from media and policy discourse, allowing separation of economic fundamentals from shifts in risk perception. The dataset spans multiple decades of property transactions across the United States, covering communities near all operating nuclear reactors in 28 states. By systematically evaluating how housing markets respond across distance bands, regulatory zones, and sentiment regimes, the paper provides new evidence on when nuclear facilities function primarily as economic anchors versus perceived liabilities. These findings offer timely guidance for regulators, local governments, and investors as the United States considers large-scale nuclear expansion as part of the clean energy transition.

I document that legal frictions in labor markets influence household financial behavior. Using staggered state-court recognition of the Inevitable Disclosure Doctrine in the Unitied States from 1987 to 2018 as plausibly exogenous shocks to labor mobility in an event-time difference-in-differences setting, I show that restricting worker mobility reduces household participation in equity markets by about three percent of the mean. The effect operates through lower expected labor-earnings growth rather than increased labor-earnings risk, consistent with human-capital portfolio theory. Effects are concentrated among top-skill workers, especially for technicians/scientists, and above-median-wealth households. These results reveal a previously overlooked channel linking labor-market institutions to household portfolio choices, highlighting how legal limits on mobility can shape wealth accumulation and financial inclusion.

Climate change is projected to increase the frequency and intensity of short-duration, high-intensity precipitation events. In rural and marginal areas, where infrastructure, settlements, and economic activities are often concentrated in valley systems, this intensification is expected to lead to greater exposure to flash floods and spatially concentrated impacts. Model-based evidence suggests that these dynamics will place increasing pressure on already vulnerable territories. Climate risk models have advanced significantly, however they are rarely integrated with economic evaluation frameworks. Likewise, most economic assessments of adaptation remain detached from physically grounded impact scenarios and often fail to incorporate societal preferences. This paper addresses this gap by combining high-resolution hydroclimatic modelling with stated preference methods to estimate population-level welfare measures of flood impact reduction. The study focuses on north-eastern Italy, a climate-sensitive region where projected changes in convective precipitation are likely to exacerbate flooding over the coming decades. Flood exposure is simulated using convection-permitting climate projections and the Simplified Metastatistical Extreme Value approach, producing future scenarios for 2041–2050 and 2090–2099. These are translated into flooded surface estimates through a hydrological impact model. The economic valuation builds on a discrete choice experiment conducted in 2025 with a representative sample of 2,000 residents. The experimental design varies impact levels across five land use categories (residential, productive, agricultural, road, and touristic infrastructure) and includes a tax attribute, allowing for the estimation of marginal willingness to pay for avoided flooding. Individual willingness to pay is highest for residential protection, with touristic areas receiving the lowest valuation. These results are integrated with the physical projections to produce aggregate benefit indicators generalisable to the population of north-eastern Italy. These indicators define a monetary threshold against which adaptation costs can be evaluated to assess whether investments are socially justified. The study provides an operational framework for integrating economic valuation into climate risk assessment and strengthens the basis for benefit–cost analysis of flood adaptation under climate risk uncertainty.

Research on the impact of air pollution on labor supply, particularly in rural areas of developing countries, is considerably limited. This paper examines the impact of air pollution on hours worked in rural Bangladesh and evaluates the effectiveness of the Brick Kiln Control Act 2013 in reducing pollution. Utilizing panel household survey and satellite data on pollution, this study employs an instrumental variable approach that exploits the exogenous occurrence of thermal inversions. A 1 standard deviation increase in PM2.5 reduces weekly hours worked by 22.6% (4.8 hours per week) relative to the sample mean of 21.33 hours. The adverse effects are highly heterogeneous-older, male workers, residing in lower income households having susceptible members including children and the elderly experience the largest reductions in labor supply. Furthermore, estimates of regional and occupational heterogeneity reveal that traders working in Rajshahi and Rangpur experience the largest declines in labor supply. I find that time-use responses for domestic production, rest and avoidance-related behavioral responses are likely mechanisms driving this reduction in labor supply. I also provide quasi-experimental evidence on the effectiveness of the Brick Kiln Control Act. Relative to the pre-policy period, the Act reduced pollution by 24.86 µg/m³ (19.85%) by 2015, although this reduction declined to 13.87 µg/m³ (11.07%) by 2019. Importantly, I find that pollution reduction due to policy implementation leads to partial attenuation of the behavioral mechanisms, with individuals reducing their time spent in domestic production by 123 minutes, rest by 13 minutes and avoidance behavior by 10 minutes after policy implementation. This provides evidence that improved air quality allows workers to reallocate time away from non-market to market activities. Overall, these findings underscore the impact of environmental legislation on improving labor market outcomes in highly polluted, low-income settings.

I investigate whether power plant managers respond to the Clean Air Interstate Rule and Cross State Air Pollution Rule by switching fuel type, retiring polluting boilers, or installing new pollution control technology. These recent transboundary air pollution regulations have effectively reduced emissions in regulated states (Andaloussi & Isaken, 2022). Using pollution control technology data from the Energy Information Administration and boiler level data from the Public Utility Data Liberation project, I test if these reductions came from switching fuel type, boiler retirements, or new pollution control technologies. My estimation strategy assumes regulations act as exogenous cost shocks allowing me directly compare outcomes for regulated and unregulated plants. The results suggest that recent transboundary air pollution regulations effectively achieved emissions reductions by changing the composition of pollution control technology without causing boiler retirements. I find that regulated boilers are no less likely to retire compared to unregulated boilers. I find that plant managers change fuel type in response to SO2 regulation. I examine the different types of pollution control technologies installed separately for SO2 and NOx emissions. I find that plant managers are more likely to install dry type technologies in response to SO2 regulation and selective noncatalytic reduction technology in response to NOx regulation. These two categories of technologies can be classified as relatively cheap and less effective than their counterparts. The results show plant managers do not retire boilers but instead switch fuel type and install relatively cheap pollution control technology. These results contribute to an important conversation about the cost of compliance with air pollution regulation through observed responses rather than estimated costs.

This paper evaluates whether the national rollout of Liquified Petroleum Gas (LPG) connections under the Pradhan Mantri Ujjwala Yojana (PMUY) has contributed to reducing household energy poverty at the district level in India. Energy poverty remains a persistent development challenge in India, despite steady improvements in electricity access over the past decade (IEA,2023). While the PMUY was introduced in 2016 primarily to promote the adoption of cleaner cooking fuels among low-income households, its large-scale expansion of subsidised LPG access presents an important opportunity to examine whether such a programme may also influence broader dimensions of energy poverty. Using nationally representative consumption survey data from the National Sample Survey (NSS) for 2011-12 and 2023-2024, combined with administrative data on the number of LPG connections released under PMUY, the analysis exploits geographic variation in the intensity of programme rollout to identify its effects. A difference-in-differences framework, supported by an instrumental variable strategy based on political alignment of local legislators, is used to estimate the causal impact of LPG expansion on multiple measures of energy poverty. Quantile regressions further assess whether the effects are concentrated among districts with the highest deprivation. The study makes multiple contributions. First it contributes to the body of work focusing on determinants of energy poverty and its reduction in low-middle income countries (Khanna et al., 2019). Second, it leverages the latest survey round to calculate multidimensional energy poverty index and creates a decade long view of India’s energy transition. Third, the study adds new insight on how socio-economic channels of education and unemployment shape behavioural transitions towards modern energy use. Finally, the findings strengthen the case for national policies to solve energy poverty and support energy transitions which have spillover effects on health and economic growth (Banerjee et al., 2021; Zhang et al., 2019), directly aligning with UN SDG7. Main findings show a reduction in energy poverty, with stronger effects in rural areas. Education further amplifies these gains, highlighting that behavioural change and awareness are as important, especially in LMICs where coal and firewood are often free making socio-economic development critical for the full impact of policy.

Wind energy has expanded rapidly over the past two decades and accounts for a substantial share of new electricity generation capacity in the United States. A growing literature has documented both the economic benefits of wind energy—such as increased income, GDP, and home values—and its potential health costs, including concerns that low-frequency noise from wind turbines may lead to sleep disruption and increased suicide rates. This paper contributes to this literature by providing the first evidence on the impact of commercial wind turbine installations on local crime rates. Using Uniform Crime Reporting (UCR) data from 1995 to 2023 and a staggered difference-in-differences framework, we estimate the causal effects of wind energy development on a range of criminal offenses. We use the U.S. Census Bureau’s Government Finance Database to explore the mechanism through which wind farms affect local revenues and expenditures. We also explore other potential mechanisms, including changes in local population dynamics and migration patterns following wind turbine installation. Our findings indicate that county-level crime rates decrease about 3 years after the installation of wind turbines and local revenue and expenditures increase post-siting. These results highlight important heterogeneity in the social externalities of wind energy and provide valuable considerations for policymakers involved in the siting of renewable energy infrastructure.

Abstract: Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) are two advanced imaging modalities that offer their own advantages for diagnosing pathologies. The growing demand for these two modalities has amplified concerns about cost, resources, and accessibility within healthcare systems. Through a review of current literature, this presentation will compare both modalities for diagnostic outcomes, patient cost-effectiveness, as well as highlighting ways to ensure the most suitable modality is ordered for each patient. Furthermore, the presentation will emphasize the importance of imaging guidelines to enhance diagnostic efficiency while reducing unnecessary expenditures. Research/Policy Question: How can healthcare systems determine when MRI or CT is the most appropriate advanced imaging modality in order to maximize diagnostic accuracy, improve cost-effectiveness, and reduce unnecessary resource usage while maintaining high patient quality care? Contribution: Evaluating imaging modality guidelines and understanding the negative impacts for patients and healthcare facilities when the wrong imaging tests are ordered. Methods: Analyzing existing literature. A conceptual model will be utilized.

How do parents use information about one child to make decisions about subsequent fertility? Theoretical and empirical studies of the allocation of resources across children rely on assumptions about this relationship, yet limited evidence exists. We estimate how parents' fertility decisions respond to variation in children's initial endowments in West Africa. We also use exogenous variation in children's initial health from idiosyncratic variation in Harmattan-induced dust pollution. The empirical evidence for the existence of a tradeoff in quantity and quality in investments in children is enormous. Yet nearly all of it focuses on estimating the effect of a change in the quantity on subsequent investment in and outcomes of quality, with a near total absence of studies looking at how quality changes the quantity chosen and desired. Despite the absence of empirical findings of the effect of the health of a child born on subsequent fertility, researchers often need to make assumptions about this relationship as the basis for other studies. The most frequent assumption is that endogenous fertility is zero, which facilitates looking at the relationship between allocation of resources across siblings. Those that assume something other than no relationship do not agree. We estimate the effect of learning about quality on parents' decisions about future quantity of children. We estimate this in West Africa using Demographic and Health Surveys data, a region of the world with high fertility rates and extremely rare selective abortions. This location also provides exogenous variation in initial endowments in the form of dust pollution from the Harmattan, which reduces the health of babies who were in utero during high levels of exposure. We find that parents respond to relatively larger babies by decreasing subsequent fertility. This is true for male and female births, first and subsequent births, and high and low-education mothers. It does not appear to be driven by differential mortality. We estimate the effect on parents' decisions about the quantity of children in response to their children's quality, and we find that a reduction in quality increases the choice of quantity.

Primary Sclerosing Cholangitis is a rare, progressive liver disease where the bile ducts become inflamed and narrowed. Over time this restricts bile flow and damages liver tissue. With the progression of the disease, complications, cirrhosis, and liver failure can occur. As there is no approved medication that can prevent or treat the disease, it often progresses requiring a liver transplant, a heavy financial burden. Because PSC progression is unpredictable arising costs can be challenging for patients and families.

The Kingdom of Bahrain is a small island economy in the Arabian Gulf that gained independence from the United Kingdom in 1971. With a population of approximately 1.4 million, Bahrain is characterized by a distinctive demographic structure: an estimated 54 percent of residents are expatriates employed across key economic sectors, while the share of the population aged 65 years and older remains low at approximately 2.1 percent. Unlike several neighboring Gulf states, Bahrain has relatively limited oil reserves and began economic diversification at an earlier stage, positioning itself as a regional hub for banking, financial services, and tourism in the post–oil boom period. These structural features make Bahrain a compelling case for examining the interaction between economic structure, labor market composition, and public service provision in small, open economies. This presentation provides an overview of Bahrain’s economic development since independence, with particular emphasis on sectoral transformation and the role of expatriate labor. It further examines the organization and financing of the education and healthcare systems. Bahrain operates a two-tier healthcare system in which citizens receive universal healthcare free of charge, while non-citizens contribute partially to healthcare costs. As demographic dynamics, labor mobility, and fiscal pressures evolve, this dual structure raises important questions regarding equity, efficiency, and long-term sustainability of the health system. By situating Bahrain’s experience within a broader economic and policy context, this review contributes to discussions on post-oil economic transition, mixed healthcare financing models, and the governance challenges faced by economies with a high reliance on expatriate populations.

Accurately measuring provider quality of care is important to understand how to improve service delivery. However, current methods including clinical vignettes, checklists on best practices, patient surveys, and standardized patient visits do not capture more abstract features such as the friendliness of patient-provider interactions. This paper explores a novel method to measure provider quality using large language models (LLMs) to analyze transcripts from audio recordings of patient-provider interactions in Tanzania. Using data from a cluster-randomized trial evaluating the effectiveness of a youth-friendly loyalty program among drug shops, we used LLMs to analyze transcripts of standardized patients seeking care for HIV testing and contraception among providers known to discriminate against young women (N=139). We presented cleaned, translated transcripts to a customized version of OpenAI GPT-4.1 to develop a model to measure provider quality. After validating the model against known truths, we conducted exploratory analysis using a semi-structured conversation guide asking quantitative frequency-type questions (“how often were customers interrupted”), qualitative questions (“how well did shopkeepers respond to questions”) and theoretical questions (“what are the differences between friendly and non-friendly care”). Further, we used the model to develop a new quality of care scale, apply the scale to the dataset and rank interactions by quality. During model validation, the LLM demonstrated understanding of the nature of the data and accurately provided aggregate descriptive statistics. However, it struggled with questions asking it to recall facts from a specific transcript. During exploratory analysis, the LLM generated themes of provider quality of care that broadly align with a human reviewer. Based on a manual review of select transcripts by humans, the interactions were scored appropriately (an average of 12% discrepancy between the LLM and manual scoring) and produced accurate rank ordering. Preliminary analysis demonstrates a proof-of-concept of using LLMs to estimate provider quality. We expect improvements in accuracy with further refinement of the model and prompt engineering. This approach opens the possibility for scaling mixed-methods research at low cost and practical applications given the rise of audio recording in routine primary care settings, such as the rollout of “AI scribes” and other assisted note taking software.

Abstract The United States is the leading recipient of foreign direct investment (FDI) and portfolio investment. However, the rise in trade barriers since 2017 has impacted international capital flows. Tariffs, import quotas, and other trade restrictions influence investor perceptions, cost structures, market access, and capital allocation decisions. This paper examines the effect of trade barriers on capital inflows into the United States. Previous research suggests that increasing trade barriers can make U.S. assets less attractive to foreign investors. Higher barriers also reduce international trade volumes and productivity, which in turn lowers the expected returns on U.S. investments. Consequently, foreign investors may shift their capital to other more open or faster-growing economies, leading to declines in FDI and portfolio inflows. Using time-series regression analysis (2000–2024) and capital-flow data, we examine the effect of trade barriers on capital inflows into the United States. The findings highlight the broader economic consequences of protectionist policies on foreign capital and long-term growth prospects. Keywords: Foreign Direct Investment, Trade Barriers, International Capital Flows

This paper examines dynamic interactions among tariffs, monetary policy, and trade balances in the U.S. We adopt a factor augmented Vector Autoregression (FAVAR) model to U.S. macroeconomic, financial, and trade data during 1999Q1-2025Q2 to construct the factors that contribute to analyzing the impacts of tariffs and monetary policy on various key macroeconomic variables such as unemployment, inflation, and trade balances. The key questions we study are (1) What is the net impact of tariffs on the trade balances after accounting for monetary policy and exchange rate responses? (2) How the relationships evolve across the various tariff regimes, and (3) What sectors are most affected by the interaction of tariff and monetary policies. We will consider various tariff-intensive sectors, such as agriculture, automobiles and auto parts, electronics and machinery, textiles and apparel, high-tech and strategic industries, and raw materials. We will conduct a sub-sample analysis according to the trade liberalization era (pre-2018) and the initial protectionist turn (2018-2025), providing insightful implications for policymakers.

Japan has focused much of its monetary and fiscal policy on inflation targeting for over a decade, while balancing this objective with efforts to stimulate economic growth. This paper applies discrete wavelet analysis to Japanese quarterly macroeconomic data, and uses the decomposition to analyze the cycles across frequency ranges. The decomposed variables will be subsequently used to analyze fiscal and monetary policy relationships within a linearized wavelet-based model.

This study investigates whether tourism-led economic expansion alleviates or intensifies environmental pressure by adapting the Environmental Kuznets Curve (EKC) framework to capture both decoupling (inverted U) and renewed coupling (N-shaped) patterns. A composite Tourism Development Index is constructed via principal component analysis from international tourism receipts, arrivals, and expenditures, and its effects are examined on two distinct measures of environmental pressure: ecological footprint and per capita CO₂ emissions. The empirical analysis uses a balanced panel of 57 countries over 1995–2019, with economies classified as high-, middle-, or low-income according to the World Bank’s 2021 income thresholds to explore development-related heterogeneity. After accounting for cross-sectional dependence and cointegration, long-run elasticities are obtained using dynamic OLS, followed by instrumental variable quantile regression (IVQR) to trace how tourism, income, inequality, and trade openness influence different points of the conditional distribution of environmental stress. The estimates point to marked differences across both income groups and quantiles. For ecological footprint, tourism in high- and middle-income economies tends to follow an inverted U-shaped EKC pattern, whereas many low-income countries exhibit a cubic trajectory in which environmental pressure eventually rises again as tourism intensifies. For CO₂ emissions, such relinking patterns are prevalent across most quantiles and income categories, indicating that carbon-intensive aspects of tourism—especially long-distance transport and energy-demanding infrastructure—remain difficult to curb at later stages of development. Rising GDP is associated with higher ecological and carbon pressures in all groups, with the strongest elasticities observed in low-income economies. Income inequality and trade openness are also found to aggravate environmental degradation primarily in middle- and low-income countries, consistent with weaker regulatory capacity and slower diffusion of cleaner technologies.

Caring for individuals with Autism Spectrum Disorder (ASD), in severe cases (Level 3), involves specialized healthcare, special nutrition and education, as well as intensive support in daily activities. The greater the severity of the disorder, the more intense the monitoring and assistance must be, often requiring the constant presence of an adult caregiver, which typically implies the abandonment or reduction of work by one parent. This study analyses the additional living costs that families face in raising and caring for a child or young person with severe ASD in the state of São Paulo, Brazil. Specifically, we aim to: (a) estimate the difference in cost of living between families with a child or young person with ASD Level 3 and families without ASD members; (b) identify the expense components that most burden families with a member presenting severe ASD; and (c) assess the degree of household budget commitment associated with the presence of a family member with severe ASD. We conducted a quantitative analysis based on primary data collected from 115 families through an online questionnaire, comparing households with a member diagnosed with ASD Level 3 (ASD 3 families) against those with no ASD members (Typical families). The results, in 2022 values, indicated significantly higher expenses in families with a member with severe ASD, to the order of R$ 712.54 (US$ PPP 304.50) per capita per month compared to Typical families. Health expenses caused the largest budget difference between the two groups. When considering income lost by caregivers who abandon or reduce professional activities to meet the special needs of the child or young person with severe ASD, the difference reaches approximately R$ 2,067.00 (US$ PPP 883.33) per capita per month, in 2022 values, suggesting that costs can be considered catastrophic for these families. These findings corroborate trends described in international literature and can importantly support discussions of assistance and income transfer policies in Brazil aimed at families with children or young people with severe ASD, especially those in situations of greater socioeconomic vulnerability.

While policy makers from both sides of the aisle express the need for more affordable housing, the actual development of such housing frequently generates controversy at the local level. There remains a strong public perception that publicly subsidized rental housing negatively affects neighborhoods. However, research has found that affordable rental developments – particularly those financed through the federal Low-Income Housing Tax Credit (LIHTC) program – do not lead to declines in surrounding property values. In fact, LIHTC properties often have a positive, durable spillover effect on home values. Ironically, the positive spillover effects of LIHTC properties could have negative ramifications for local market-rate renters. The increased appeal of the community not only could lead to higher home values, but also to higher rents. On the other hand, the additional supply of rental units provided by the affordable housing development might moderate the effects of the higher land values. Our study focuses on both the spillover price and rent effects of LIHTC-financed affordable housing developments in Florida’s five major metropolitan areas. We augment the existing literature showing that LIHTC properties generally have neutral to positive effects on nearby home values. We cover largely untilled ground in assessing the effects that LIHTC developments have on nearby market-rate rents. Our study focuses on a state with some of the highest growth in rents and population throughout LIHTC’s 40-year history. We assess the effects of LIHTC developments placed in service between 1986 and 2023 on surrounding single-family and multi-family sale prices, as well as on the effective rents of units in nearby unsubsidized, professionally managed, multi-family properties. Our methodology incorporates both temporal and spatial difference-in-differences and controls for assorted unit, building, and neighborhood characteristics and trends. As in other parts of the country, we find that Florida’s LIHTC properties have a generally positive effect on area house prices. At the same time, they exert a downward pressure on rents in unsubsidized multi-family properties in the surrounding area. In short, LIHTC properties create affordable rental units for their tenants while making nearby market-rate units less costly, all while making the neighborhood more attractive.

This paper studies the option value of internal migration and migration patterns across heterogeneous households. I develop a quantitative spatial equilibrium model with mul- tiple locations, moving costs, idiosyncratic productivity component, life-cycle dynam- ics,and homeownership. The model is calibrated to 50 U.S. metropolitan areas, and I calculate how much additional consumption households would require to compensate for the loss of the migration option. I find that the average value of migration is 2.56% of consumption. Migration values are higher for younger and wealthier households. However, migration rates and migration values do not necessarily move together. Low- wealth households migrate more frequently, while high-wealth households obtain larger gains from migration. I show that this difference reflects heterogeneity in migration motives and in the size of gains conditional on moving. These findings suggest that migration rates alone cannot fully capture the economic value of migration.

Recent studies find that statistically there is zero economic impact from college football games for certain cities. We expand this analysis by looking at a larger group of counties and cities using sales tax data and find that there is almost no statistically significant economic impact in over 100 local geographic regions.

We explore recent student-athlete transfer data, focusing on the role of changes in NCAA regulations. We analyze nearly every transfer from an NCAA Division I men’s basketball program from the 2016 to 2023 seasons. Detailed results are provided by years of remaining eligibility and level of competition. A large increase in the number of transfers predated the beginning of Name-Image-Likeness payments. More players at all levels of competition have transferred as the regulations have eased, while players at the top level of competition were able on average to transfer to better (or less bad) teams. Further, fewer players at low levels of competition are transferring out of Division I altogether.

This study examines the effect of foreign teammates’ productivity on the productivity of domestic players using player-level data from the Korea Baseball Organization (KBO) from 2001 to 2025. To measure individual player productivity, this study uses Wins Created as the productivity indicator. The results show that, for pitchers, the productivity of both domestic and foreign teammates has a positive effect on domestic pitchers’ productivity. However, while the effect of domestic teammates tends to persist over a relatively longer period, the effect of foreign teammates appears to be short-lived. For hitters, the productivity of domestic teammates has a positive and persistent effect on domestic hitters’ productivity, whereas the productivity of foreign teammates has no statistically significant effect in the full sample. In addition, for regular hitters, the productivity of foreign teammates is found to have a negative effect. These findings suggest that the effect of foreign players in the KBO differs between pitchers and hitters. In particular, foreign pitchers are often used as starting pitchers and may therefore affect the pitching conditions and role allocation of domestic pitchers. In contrast, foreign hitters are more likely to be used as core offensive players occupying a separate roster role, rather than serving as direct competitors to domestic hitters.

Erin Matson is considered “the Michael Jordan” of NCAA Field Hockey. She received the Honda Sport Award in Field Hockey three times. She also starred on four NCAA title teams and three teams that went undefeated. When her career at the University of North Carolina ended, she made the unique transition to head coach of her college team. As coach, she led North Carolina to another NCAA title in her first year. This unique journey in college athletics leads one to wonder if Matson generated more value as a player or a coach. To address this issue, we will measure the marginal revenue product of a field hockey player. Field hockey is often referred to as a “non-revenue” sport where economic exploitation does not happen. Our analysis of Matson and other players will reveal that there exist field hockey players who are indeed exploited by their university employer.

In 2018 the National Football League adopted a rule penalizing players who lower the helmet to initiate contact, a reform that early descriptive evidence linked to a decline in concussions. We show that this decline was league-wide rather than a product of the rule itself. Exposure to the rule varied sharply across positions. For example, open-field tacklers faced near-universal enforcement, while quarterbacks, offensive linemen, and receivers faced almost none. We exploit this variation in a difference-in-differences design, estimating the effect of the 2018 rule on exposed positions relative to unexposed ones. Using over 200,000 player-week observations, we find no differential change in concussion rates among the positions the rule targeted - concussions fell equally across exposed and unexposed players. We also find no evidence of injury substitution toward other parts of the body; if anything, lower-extremity injuries declined among exposed positions as well. Because the decline appears even where the rule did not bind, it cannot be attributed to the rule's targeting of helmet-first contact and is more plausibly explained by contemporaneous league-wide changes in player safety. Crediting targeted contact rules with reductions that would have occurred regardless could potentially risk misdirecting future player-safety policies.

The recent inflation experience in the United States and the Federal Reserve’s actions to bring it under control created a heated debate on the causes of inflation. The debate centered on monetary easing during the pandemic, rising oil prices after the pandemic and the generationally high interest rates following the rate hikes by the Federal Reserve. Using vector autoregressive models, this paper investigates the extent to which each impacts inflation. Our findings reveal that the main driver of inflation in the U.S. are the movements in the price of oil. Unlike many expect, interest rates do not appear to have a statistically significant lowering effect on inflation. The second most influential driver of inflation is found to be the momentum coming from its past dynamics.

Our paper examines the determinants and short-term performance consequences of Generative Artificial Intelligence (GenAI) adoption in the U.S. banking sector following the exogenous technology shocks of ChatGPT, Claude, and Manus. We employ a dual-stage empirical design using a novel quarterly panel of 180 Bank Holding Companies (BHCs) from 2018 to 2025. First, a Spatial Linear Probability Model (LPM) with a time-varying, multi-asset network matrix reveals adoption is driven by strategic substitution rather than geographic “mimicry”. In other words, banks actively delay adoption when their structurally similar peers move first. Second, Stacked Difference-in-Differences (DiD) and Staggered Synthetic DiD (SDID) models demonstrate that early adoption carries a severe “Implementation Tax” that manifests through several channels. Smaller institutions suffer a large 42– 58% decline in market valuation (Tobin’s Q), while the largest institutions bear persistent penalties in profitability (ROE/ROA). We find that this Innovation Tax is partially offset by AI talent inflows, but only large banks successfully translate these human capital investments into measurable profitability gains (ROE). Our findings challenge the conventional notion of the “first-mover advantage” in technology adoption. We find that the market penalizes early GenAI adoption unless the firm possesses the scale to internalize the associated human capital costs. We contribute to the literature by providing novel causal evidence of an implementation penalty on adoption in banking, an “innovation tax”. This is particularly so, following the release of three unique GenAI models. Our results carry several policy implications.

Do Bribers and Bribed Add Financial Value to National Wealth? Kashi Nath Tiwari Is corruption wealth-producing or wealth-destroying? Does corruption have positive economic value for all the stakeholders (personal wealth of corruption-counterparties, national economies of corruption-counterparties, and the global economy in general)? There is heterogeneity in the world theater of corruption: age, gender, religion, geography (rural, urban), climate, wealth-level, stage of economic-development (agrarian or industrialized), size of the dwelling (small town where everyone knows each other v. large metropolitan area). What is corruption? Adam and Eve type of corruption (the serpent in The Garden of Eden tempting the two to commit the sin in eating the forbidden fruit). Lobbying? Constructing, bending, and utilizing the laws to subdue the rivals and competitors? African tribal-chiefs trading humans for their personal benefits? Increasing sales through price-garbling, colorful packaging, and manipulation of consumers? Public sector corruption versus private sector corruption? Domestic corruption versus transnational corruption. This paper examines the ease at which a company can purchase the favors from overseas officials to access their markets or to speed up the bureaucratic licensing process (to acquire the factors of production and to market the final products).

This paper explores and compares the factors that affect the economic success of Latino Musicians in terms of hourly wage during and post the Covid 19 era, using data from the 2021and 2023 American Community Survey, where musicians are defined as those who make their living from income derived from music-making: musicians and singers, music directors and composers, namely. The 2021 data reveal the importance of education and union membership on hourly wage for Latino musicians; however, English proficiency’s effect on hourly wage is minimal and there is no interaction effect between English proficiency and union on earnings. The same pattern applies to 2023. Quantile regression approach is adopted to examine the effects of English proficiency’s effects across the entire earnings distribution. The 2021 data shows that the relative importance of English proficiency is greater at the upper tier of the earnings distribution for Latino male musicians with a union membership: fluency in English is associated with a 3% increase in hourly wage for Latino male unionized musicians at the 50th hourly wage percentile, while such a premium is 26% for those at the 75th hourly wage percentile. This pattern does not apply to Latino female musicians. For Latino female musicians, educational attainment would contribute to their earnings, while marriage plays a negative role. However, the 2023 data show the relative importance of English proficiency is lower at the lower tier of the earnings distribution for Latino musicians with a union membership: fluency in English is associated with a 5.6% decrease in hourly wage for Latino unionized musicians at the 25th hourly wage percentile, while there are no interaction effects for median and upper tiers. Latino male unionized musicians at the 50th hourly wage percentile, fluency in English will increase their hourly wage by 36.6%.

Media reports have stated that the number of hit songs and albums performed by bands has declined over time. Possible reasons for this are organized into five theoretical areas: changing consumer tastes, technology allowing for solo projects, social media rewarding individual performers, financial incentives for solo artists, and record labels being less interested in bands. Within these theories, a portion of the blame has been pointed toward Spotify and their use of algorithms. Two data sets were constructed, the top 20 songs from Billboard’s annual Hot 100 chart for the years 1961 to 2025 and the top 20 albums from Billboard’s annual Top 200 chart for the same time period. Ordinary least squares was employed to determine that a significant decline in hits by bands had indeed occurred on both charts. A Mann-Whitney two-sample t-test was used to determine whether Spotify was to blame. The results indicate that Spotify is not to blame as the declining trend began years before Spotify came to the U.S. There is evidence that points toward changes in consumer tastes over time, as the decline in hit songs by bands correlates to an increase in the number of hits made using collaborations, particularly an increase in the number of hip-hop songs. These results raise a pedagogical question for popular music ensembles when considering these changing tastes by consumers.

In 1986, the first class of the Rock and Roll Hall of Fame was inducted. Today there are 425 individuals and groups in the Hall of Fame. It is argued in this study that the demographics of the inductees tell us much about the history of Rock and Roll. White males tend to be the industry leaders. And white males also tend to be ranked among the top performers. In contrast, women tend to be excluded. In addition, women have to do far more just to be considered among the best. The history of Rock and Roll is a history of gender discrimination. And the demographics of who is in the Hall of Fame tell that story.

South Korea ranks among the most gender-unequal countries in the OECD, yet its music industry exhibits a striking degree of gender parity rarely observed in other major music markets. Female artists, particularly girl groups, compete on equal commercial footing with their male counterparts in album sales, streaming, and stadium tours, a phenomenon with no clear parallel in the United States or other high-income economies. This paper argues that this paradox is not explained by domestic cultural progressivism but by the structure of government investment in K-pop as a national export industry. Beginning with the 1997 Asian financial crisis, the Korean government systematically treated cultural content as a strategic export sector, deploying direct subsidy, equity-based fund-of-funds structures, and a dedicated institutional mechanism to build K-pop into a globally competitive industry. We argue that an export-oriented industry funded primarily through public and quasi-public capital faces a different incentive structure than one driven purely by domestic private markets: maximizing global reach rather than short-term profit opens the door for female artists to thrive at a scale that domestic gender norms alone would not have produced. We document this relationship and discuss its implications for gender economics more broadly. Our findings suggest that labor market gender outcomes within an industry may reflect the structure of investment and ownership as much as prevailing cultural norms, and that industrial policy can generate gender equity as an unintended but meaningful byproduct.

The spread of COVID-19, which led to the implementation of numerous preventive measures, imposed substantial mental burdens on workers and fundamentally changed the work environment. Consequently, there were early concerns about the secondary effects of these preventive measures on mental health, including anxiety, loneliness, and depression. However, in Japan, the monthly suicide mortality rate did not significantly increase from January to June 2020 compared with the same month of the five-year average. In fact, the suicide rate among men significantly decreased. Based on this evidence, we examine whether those individuals with a history of suicidal ideation or suicide attempts experienced an improvement in mental health in early 2020 compared with 2019. Using a panel dataset, we performed a fixed-effects estimation and found that individuals with suicidal ideation or suicide attempts had a lower level of psychological stress in early 2020 compared with 2019. Moreover, our results show that their psychological stress was positively correlated with the average working hours of their local area during the early period of the COVID-19 pandemic. Average working hours at the prefectural level were associated with a greater reduction in distress among individuals with suicidal ideation compared with those without. While the increase in average working hours in 2020 reduced psychological stress for individuals without suicidal ideation, it had the opposite effect on those with suicidal ideation. We consider the possibility that their relative concern with others might influence their mental health. While many individuals experienced social and economic anxiety due to the uncertainty of the early period of the COVID-19 pandemic, our results suggest there might be some individuals whose mental stability recovered, consistent with the concept of interpersonal comparisons. Our results show that the mental health of individuals with suicidal ideation or suicide attempts in the previous year has largely improved during the early stage of the COVID-19 pandemic.

After the Covid-19 pandemic in 2020, the European Union adopted the European Next Generation EU Plan 2021-2026, supported by significant common funding. The development of the resulting Italian Recovery and Resilience Plan was much more for the country than the preparation of an emergency intervention (Gatti paper at the WEAI 99th Annual Conference, June 2024). It focused attention on issues that had not previously received systematic intervention. For women, the implementation of the Next Generation EU and the resulting Italian Recovery and Resilience Plan 2021-2026 meant the strengthening or new implementation of innovative policies for work and maternity. For women's health, the National Recovery and Resilience Plan 2021-2026 and the National Health Equity Program 2021-2027, also funded by European Union funds, have meant taking charge of community medicine, regional disparities in the provision of the National Health Service's Essential Levels of Care, and strengthening screening for the early detection of female cancers. For women's health, the objectives of the Italian Recovery and Resilience Plan and the National Health Equity Program are an acknowledgement of the structural weaknesses of the Italian National Health Service, even though it was created precisely to resolve these weaknesses. After having investigated the progress of the projects (Gatti paper at the WEAI 100 Annual Conference, June 2025), this paper, awaiting final data on the implementation of the Recovery and Resilience Plan in 2026, intends to concretely evaluate, comparing them with other European and international experiences, the characteristics of the tools implemented by the two Plans for improving women's health protection within the Italian National Health Service.

This paper examines how the COVID-19 pandemic reshaped healthcare recruiting across California's rural-urban divide. Using WageScape job-posting data from 2017 to 2022, we estimate difference-in-differences and event-study models at the ZIP-code-by-month-by-occupation level. We find that rural healthcare posting activity did not keep pace with urban posting growth after March 2020. Job-level duration and salary specifications indicate greater recruitment frictions in rural labor markets. Rural postings remained open longer, and posted wages rose modestly relative to urban areas. Deduplicated posting estimates further show weaker rural growth in unique vacancy creation. County-level evidence shows that rural employment lagged in emergency and hospital settings, while utilization and facility counts remained broadly stable. Clinic-specific results suggest some late-period recovery in rural clinic postings, but not enough to offset broader rural staffing constraints.

Abstract: Governments worldwide expanded social protection systems in response to the COVID‑19 pandemic. In India, the Pradhan Mantri Garib Kalyan Anna Yojana (PMGKAY) augmented the Public Distribution System by providing additional free foodgrains to more than 800 million beneficiaries. We use nationally representative longitudinal data from the Longitudinal Aging Study in India (LASI) to estimate the association between PMGKAY exposure and food insecurity using a difference‑in‑differences design with individual fixed effects. Because only one pre‑intervention observation is available, we use inverse probability weighting based on pre‑pandemic characteristics to ensure baseline comparability between eligible and comparison individuals. We find that PMGKAY exposure is associated with a 5.2 percentage points decline in moderate to severe food insecurity, equivalent to a 44 percent reduction relative to the baseline prevalence among eligible individuals. Estimated effects are concentrated in states with higher coverage and lower leakage in the Public Distribution System, where effective exposure to the program is plausibly greater. Consistent with an in‑kind transfer mechanism, reductions in food insecurity are not accompanied by increases in food expenditure. These findings suggest that large‑scale food transfers can play a protective role during aggregate shocks, but that their effectiveness depends critically on delivery capacity.

The collectible trading card industry is a $7B to $15B dollar industry with annual sales exceeding $3Billion. By comparison, the sports card industry is $13B to $15B. Yet, to date, almost no economic research exists to understand the nature of the collectible trading industry. The purpose of this study is to model the collectible trading industry and empirically test the robustness of our model. For empirical analysis we use the retail and secondary market prices of sealed packs from a trading card game called Magic the Gathering™. In each brand-new sealed pack, cards from the expansion of the game are randomly inserted. Thus, the actual contents in each pack are unknown. We find that inserted serialized cards significantly affect pact price, but more importantly the value of the highest priced card has the strongest effect on the value of packs sold on the secondary markets. Finally, there is a strong correlation between the retail price of packs and the price on the secondary market. This suggests that Hasbro™ who owns the Magic the Gathering™ brand pays attention to the consumer behavior in the secondary market when determining the retail pack price for each new expansion of the game.

For decades, political science has debated whether economic distress demobilizes voters or fuels electoral backlash. This study addresses this paradox by analyzing county-level turnout across the 2012, 2016, and 2020 US Presidential elections. Utilizing a two-way fixed effects model and integrating five macroeconomic indicators, we propose a unified framework grounded in the calculus of voting. We identify a two-speed realignment: intense populist narratives functioned as a shield that neutralized the traditional demobilizing penalty of acute unemployment, while a status threat engine drove participation surges in resourced, affluent counties rather than materially deprived ones. Conversely, chronic poverty acts as a permanent structural floor, leaving the most destitute trapped in a democratic dead zone despite intense narrative mobilization. Ultimately, these findings reveal a structural decoupling of aggregate participation from objective macroeconomic conditions, demonstrating that grievance-based narrative environments can fundamentally re-wire electoral behavior.

Generative AI agents based on large language models have evolved beyond conventional information-generation tools. By leveraging external tools and memory, they can autonomously plan and execute tasks while adaptively updating their behavior in response to environmental changes. The rapid development of multi-agent systems further accelerates the emergence of a society in which AI complements human decision-making. This study analyzes how the diffusion of generative AI agents affects economic activity, with particular emphasis on consumer behavior, market structure, and economic welfare. We focus on an “agentic economy,” defined as an economic environment in which assistant and service agents facilitate and mediate transactions and decision-making. A key feature of such an economy is that generative AI agents interact directly with consumers, enabling deeper inference of individual preferences and more efficient product recommendations. In current digital markets, consumers face a vast array of alternatives on online platforms, and, as rational inattention theory suggests, limited attention and information-processing capacity generate substantial search costs. By substituting for or complementing consumer search, generative AI agents can improve matching efficiency and potentially increase consumer surplus. We introduce the concept of “context” into the model, encompassing individual characteristics and consumption environments that fundamentally affect utility. Context is classified into publishable context, which is explicitly expressed in reviews or queries to agents, and unpublishable context, which is not directly disclosed. Although unpublishable context is not observable, generative AI agents can implicitly incorporate it through two mechanisms: (i) unpublishable context is non-explicitly reflected in the sequence of queries and responses between consumers and agents, and (ii) interactions with other users who share similar context are indirectly utilized through pre-training and model refinement. As a result, agent responses may reflect rich context even when consumers do not fully disclose it. These capabilities also reshape firm behavior by shifting competition away from pure scale expansion toward strategies that deepen relationships with heterogeneous consumers, encouraging innovation in product variety and quality. Using simulation-based analysis, this study quantitatively evaluates the effects of AI-driven search and recommendation mechanisms on prices, technology investment, consumer surplus, and economic welfare, shedding light on transformations in digital market competition.

In a seminal paper, Calvano et. al. (2020) demonstrated that reinforcement learning algorithmic agents engaged in price competition independently learned to collude on average, raising deep concerns about the future of competition in a variety of markets. However, the agents' collusion was relatively slow and haphazard. In this paper, we show that algorithmic agents based on deep neural networks also learn to collude in similar competitive environments. Importantly, collusion is several orders of magnitude faster and more than twice as large in magnitude. Moreover, agents learn to collude in an organized manner. We propose a new fundamentally new type of regulatory policy to address these issues that is explicitly based on intent as opposed to communication, which is absent in algorithmic competition. By using novel tools that have been developed to perform white box audits of neural networks, we show that our proposed regulation is easy to implement, effective at combatting collusion, and flexible even in non-stationary environments.

Mediation is a strategic tool in mitigating conflict in bargaining with incomplete information. We investigate the informed principal problem of mediator selection through two laboratory experiments. The theory of neutral optimum (Myerson, 1983) predicts that the informed principal’s optimal inscrutable choice is not the one that maximizes the ex-ante probability of peace in our environment due to th e conflicting interests of principal types. In uninformed settings, subjects overwhelmingly select the peace-maximizing mediator. When informed, however, subjects do not choose the neutral mediator more frequently than the peace-maximizing one. Instead, principals of different types diverge, each favoring compromises aligned with their own incentives. This divergence prevents the intertype compromise underlying the neutral optimum from emerging. We also document behavioral factors, such as projection bias, that contribute to these deviations. Our findings highlight the limits of neutral optimum theory in practice and show that while subjects recognize the need for inscrutable mediator selection, they struggle to coordinate on the theoretical compromise.

Using a calibrated model, this paper analyzes the optimal combination of taxation and auditing in a growing economy with public capital accumulation and tax evasion. We find that, under the benchmark calibration, over a wide range of values for the parameter capturing tolerance for tax evasion, the optimal tax rate remains stably around 40%, whereas the share of audit expenditure in government spending varies substantially. We then consider suboptimal policies that fix either the tax rate or the audit-expenditure share and allow only the other instrument to be chosen. To evaluate the degree of policy control provided by each instrument in isolation, we compute and assess the consumption-equivalent loss relative to the optimal combined policy. We find that the tax rate is a relatively effective policy instrument over a wide range of parameter values.

We propose a new model specification test for multiple-equation systems with cross-equation error and dynamic regressor--error dependences. Conventional tests often rely on exogeneity conditions strong enough to ensure consistency of the OLS estimator. These exogeneity conditions are violated when regressors and errors are dynamically dependent, rendering conventional model specification tests invalid. To address these limitations, we clarify the relationship among alternative exogeneity conditions, characterize the consistency of competing multiple-equation estimators, and propose a generalized Durbin estimator for multiple-equation systems with an intercept, cross-equation error and regressor--error dependences. We show that our estimator remains consistent under the weakest exogeneity condition. We then derive its asymptotic distribution and construct Wald tests. Our Monte Carlo experiments confirm that the bootstrap-based Wald test substantially improves finite-sample size control. An application of the bootstrap-based Wald test to the Fama--French multifactor models leaves the null hypothesis unrejected in cases where competing FGLS-based tests reject it.

This paper investigates the efficiency of Japanese stock market under wartime economic controls from 1930 to 1943. While Japan maintained active stock trading during the Second World War, state interventions---including capital controls, price regulations, and institutional restructuring---raised concerns about impaired market mechanisms. We employ an event study framework based on a CAPM-AR(p)-SV model, which accounts for serial correlation and stochastic volatility, to examine abnormal returns around major military and policy events. Our findings reveal persistent, cross-sectionally heterogeneous abnormal returns, particularly favoring zaibatsu-affiliated military firms, suggesting a deviation from semi-strong form efficiency. These patterns reflect structural asymmetries induced by wartime industrial policy, wherein market mechanisms amplified rather than neutralized distortions in capital allocation.

This paper investigates exchange rate pass-through (ERPT) in the Japanese cotton spinning industry from 1930 to 1939. A central challenge in the ERPT literature is the rigorous identification of the causal impact of exchange rate fluctuations on trade prices. To address this, we employ a Difference-in-Differences (DID) framework that exploits the unique monetary geography of the 1930s. Our identification strategy treats the de facto "Yen Block" (including the Kwantung Leased Territory and Manchuria) as the control group and non-Yen Block countries as the treatment group. We specifically focus on Japan’s exit from the gold standard in December 1931, an event that triggered a massive exchange rate shock in the non-Yen Block as Japan transitioned to a managed currency system. By comparing price adjustments across these two groups, we isolate the effect of currency depreciation from other confounding global macroeconomic shocks. Our identification-robust DID approach reveals significant Pricing-to-Market (PTM) behavior that standard ERPT estimations fail to capture. We find evidence of incomplete pass-through, suggesting that Japanese spinning firms strategically adjusted their markups to maintain market share in non-Yen Block regions following the rapid yen depreciation. This study contributes to the historical understanding of firm behavior in Japan’s leading export industry, which has been regarded as a major driver in mitigating the impact of the Great Depression on the Japanese economy. It also provides methodological insights into the estimation of ERPT during periods of extreme monetary volatility.

We propose Midastar, a novel extension of the threshold autoregression (TAR) to the Mixed Data Sampling (MIDAS) framework. In the regular Midastar, the target variable is observed less frequently than the threshold variable. In the reverse Midastar, the target variable is observed more frequently than the threshold variable. These models accurately capture threshold effects, whereas standard TAR with temporally aggregated data can point to spurious non-threshold effects. The parameters are estimated via profiling, and the no-threshold-effect hypothesis is tested via wild bootstrap. We establish the uniform consistency and asymptotic normality under much weaker conditions than in the literature. In particular, we overcome the challenges arising from the potential lack of stationarity. Our Monte Carlo simulations indicate that the proposed methods perform well in finite samples. We present two separate empirical applications to illustrate the practical values of the regular and reverse Midastar models. For the regular scenario, the target variable is monthly realized volatility measures of the crude oil market and the threshold variable is daily CBOE Volatility Index. For the reverse scenario, the target variable is monthly employment growth and the threshold variable is quarterly real GDP growth of the United States. Both studies indicate that the Midastar models are useful for modelling and predicting financial and macroeconomic indicators, whereas the single-frequency TAR model points to spurious non-threshold effects.

This study investigates the asymmetric and time-varying effects of global oil price volatility on core inflation dynamics in two advanced economies—the United States and the United Kingdom—using quarterly data from 1997 to 2024. Employing a combined Quantile-on-Quantile (QQ) regression framework, Wavelet Coherence, and Cross-Quantile analysis, the paper explores how fluctuations in Brent crude oil prices interact with monetary and fiscal policy variables and global economic policy uncertainty (GEPU) to shape inflation responses across different states and time horizons. The findings reveal that the United States displays greater resilience to oil price volatility, particularly under low-inflation regimes, owing to its diversified energy base, adaptive policy framework, and effective inflation targeting by the Federal Reserve. In contrast, the United Kingdom exhibits stronger and more persistent inflationary responses, especially under high-inflation conditions, reflecting its higher dependence on imported energy and the amplifying effects of exchange rate pass-through and fiscal spillovers. Wavelet decompositions highlight that oil price volatility exerts its influence mainly over medium to long horizons (8–16 quarters), indicating a gradual pass-through mechanism. Overall, the results underscore the role of structural and institutional differences in mediating the inflationary transmission of oil shocks. The policy implications emphasize the need for context-specific responses—maintaining energy diversification and policy flexibility in the United States, and enhancing energy security and fiscal–monetary coordination in the United Kingdom—to manage inflationary pressures in an era of global energy and policy uncertainty.

With the innovation of digital technology, this paper describes the process of how the traditional trading network evolves into the platform network and finally reaches the social commerce network from the perspective of network evolution. The digital economy has developed rapidly in the past decade, and the emergence of new digital technology has had an impact on the act of shopping. The e-commerce field continues to change and evolve, and social commerce has become the main form of online shopping as an emerging business model. Social commerce relies on social media platforms to deeply integrate social interaction with product transactions, breaking the limitations of traditional e-commerce, and opening up new consumer experiences and marketing strategies. Based on the theory of social division of labor, this paper applies the method of inframarginal analysis to explain the evolution process from traditional trade network to platform network and finally to social commerce network theoretically (Yang and Ng, 1993). Although there is a wealth of literature on social commerce in different areas, most of them are empirical studies on the impact of social commerce on a certain factor. The existing literature does not discuss the evolution of social commerce, and there is almost no theoretical model to support the relevant conclusions. The contribution of this paper is to introduce the inframarginal analysis from the perspective of the division of labor network, we classify and discuss the evolution of shopping structure, and study the conditions required for structure evolution theoretically. Three types of division-of-labor networks are established. We find corner solutions and conduct general comparative static analysis, then discuss the conditions and thresholds for network evolution. Results show that product transaction efficiency and information transaction efficiency jointly determine network evolution to the social commerce. Further analysis in this study includes thresholds and transaction efficiency, learning cost, and real per capita income.

One of the most common policies to prevent predatory pricing is to regulate loss leading prices through below-cost sales laws. Yet only a handful of research studies have been conducted on the competitive impact of loss leading and laws against it, both theoretically and empirically, yielding conflicting predictions. My research aims to empirically analyze the predatory aspects of loss leading pricing strategies and the effectiveness of below-cost sales (BCS) laws. Specifically, I study the retail gasoline market, leveraging state law changes in Wisconsin as a natural experiment. Analyzing station-week level price data over six years and employing a difference-in-differences approach, the research finds that all firms engage in loss leading pricing about 3.2% of the time when below-cost sales are permitted. However, the BCS law is not effective in preventing loss leading pricing at least in the first couple of years. Unlike the predatory theory predicts, loss leading pricing was primarily driven by small firms, not large firms, and although margins were squeezed for all firms, small firms' margins decreased more than large firms. Lastly, again contrary to conventional predatory pricing narratives, the absence of BCS laws did not significantly affect the total number of gas stations, regardless of the firm size. These findings do not support that loss leading pricing is predatory. This paper contributes to antitrust literature by providing insights into the impact of BCS laws on competition and challenging prevailing notions regarding loss leading pricing strategies.

Does Chinese development finance in Sub Saharan Africa build productive capability, and is that capability absorbed into measured productivity? We separate Chinese official finance into Fish, meaning infrastructure and physical capital, and Fishing, meaning technical assistance, education, health and agriculture, and we add a third object that the usual debate omits, the Fishing Environment, the trade and production structure that decides whether capability is used. The analysis uses harmonized public sources, AidData’s Global Chinese Development Finance Dataset 3.0, the World Development Indicators and Worldwide Governance Indicators, and the Penn World Table, for 45 economies over 2000 to 2021; the data and code are made available. The evidence is more cautious than the metaphor invites. A higher manufacturing-export share is robustly associated with higher total factor productivity, so the structure of the pond matters. We do not find a robust effect of capability finance on productivity: the Fishing coefficient is positive in most specifications but not significant across inverse hyperbolic sine, count and intensity measures, and it is imprecise under heteroskedasticity-based instruments. The mediation through trade structure that a strong reading of the thesis requires is not identified; the indirect effect is small, and a sensitivity parameter near zero would erase it. A manufacturing-export threshold appears in the point estimates, but its confidence interval is too wide to locate. We read the results as a conditional and unfinished answer, and we use them to identify the specific kinds of Chinese investment that would, on this evidence, raise the productivity payoff of the fishing channel.

This paper considers the endogenous formation of environmental R&D (ER&D) networks and examines the effects of carbon taxation under alternative regulatory timings: the commitment and non-commitment regimes. Under the commitment regime, the tax is set ex ante and is treated as credible, which induces coordinated ER&D investment and sustains the complete network as a stable and welfare-enhancing structure. Under the non-commitment regime, however, the flexibility in tax adjustment may weaken coordination, which results in a stable network structure that might not be socially desirable unless ER&D efficiency is low. Finally, comparisons between the stable networks under the alternative regulatory timings show that the commitment (non-commitment) regime yields higher welfare when ER&D efficiency is high (low). Our findings suggest useful policy implications for the efficient design of the carbon tax mechanism under the endogenous formation of ER&D organization, highlighting the need for supplementary measures to ensure socially desirable outcomes with stable network formation.

In the contemporary landscape of digital platforms and big data, information asymmetry has emerged as a pivotal issue in market competition. This phenomenon is particularly pronounced in the gaming industry, where platforms exploit their informational advantages to implement personalized pricing strategies, thereby significantly influencing the strategic decisions of independent game developers and altering market structures. This study employs the Hotelling model to construct a four-stage game-theoretic framework, examining how a platform designs listing fees and information fees to impact developers' decisions regarding market entry and pricing. In addition, we investigate the effects of personalized pricing on consumer surplus and market efficiency. The findings indicate that in a non-monopolized market, platforms can maximize their profits through personalized pricing, while independent developers adopt uniform pricing—given that at least one developer is permitted to pay a listing fee to be on the platform. Consequently, when personalized pricing is unrestricted, only the platform engages in such pricing behavior. A comparative analysis of different pricing strategies reveals that social welfare is maximized when all developers adopt uniform pricing. However, in the scenario where developers are precluded from joining the platform due to external constraints, consumer surplus under the platform's personalized pricing may surpass that under uniform pricing.

This paper develops a simple model of taste-based discrimination to suggest that a firm uses a compensating wage differential as a cover for discrimination. In the model, men and women work slightly different tasks, so there is a legitimate reason to pay a compensating wage differential. However, the value of the task difference is hard to establish, making it difficult for the workers to prove discrimination, and the third party instead observes the wage structure --wage components and the amounts-- which the firm controls. The firm then strategically misrepresents discrimination as a compensating wage differential by shifting the gender wage gap to the compensating wage differential component of the wage. This action makes it difficult for workers to win the discrimination lawsuit if they sue, and deters workers from suing in the first place, thus doubly protecting the firm from discrimination liability, and making the remedial function of equal pay legislations ineffectual. Empirically, this strategic action makes the detection of discrimination harder as the resulting wage structure resembles that of a compensating wage differential. Controlling for hours worked would underestimate the extent of discrimination as the gender difference in hours worked is a direct consequence of discrimination. We review some prominent discrimination cases to show how firms use compensating wage differentials as a cover for discrimination in practice.

This paper examines how gender and residential socioeconomic status shape hiring outcomes in Karachi’s information technology sector. Employers in Pakistan can openly state preferences regarding gender, residential location, and other characteristics, but the majority in the information technology sector choose not to do so. This creates an opportunity to examine whether discrimination persists when such biases are not explicitly stated. An analysis of explicitly gender-targeted job ads shows that men are preferred over women across most occupations, even in traditionally pink-collar roles. Results from a resume audit experiment, submitting 2,032 applications to 508 full-time job openings, show that men receive more callbacks for job interviews than women, even in the absence of explicit gender preferences in job ads. The study also indicates a significant premium favoring candidates from high-income areas, who receive 45% more callbacks than applicants from low-income neighborhoods. This advantage remains robust even after controlling for commuting distance. Qualitative interviews with human resource officials suggest that employers associate productivity with both gender and neighborhood socioeconomic status. Residential address acts as a proxy for class background and signals education, skills, and perceived “fit” in professional settings. These perceptions may reinforce stereotypes, disadvantaging women and candidates from low-income backgrounds.

The age at which individuals first marry is widely recognized as a key determinant of marital outcomes. While extensive economic and demographic research in Western contexts suggests that delaying marriage improves marital stability, less is known about how this relationship operates in rapidly transforming societies like China. Amid shifting family norms, rising divorce rates, and widening urban–rural disparities, China offers a unique setting to reassess the link between marriage timing and stability. This paper investigates how females' age at first marriage affects the risk of marital dissolution in China using nationally representative China Family Panel Studies and Cox proportional hazard models. Our results reveal a U-shaped pattern in the full sample: both early (<20) and later marriages (>31) are associated with significantly lower divorce risk relative to the 20–25 group. The inclusion of controls such as education and premarital or partner characteristics suggests that part of the initial association may reflect underlying partner characteristics. Stratified analyses uncover meaningful heterogeneity. Among rural individuals, the U-shaped relationship persists, aligning with theories of constrained matching and economic co-dependence. In contrast, urban individuals who marry earliest exhibit the highest marital stability, with the hazard of divorce increasing with marriage age, a result likely shaped by competitive urban marriage markets, “leftover women” stigma, and shifting expectations around career and family. By birth cohort, the divergence in marital stability by age at first marriage appears to diminish among individuals born after 1985, possibly reflecting shifting social norms and more individualized marital expectations in the post-reform era. Our findings contribute to the economics literature on household behavior and marital matching by illustrating how institutional context such as urbanization, value change, and education systems modulates the timing-stability link. The results refine Becker’s (1977) model of marital sorting by incorporating China’s unique demographic pressures and evolving cultural norms.

Objective: Low fertility has emerged as a global defining demographic challenge. Birth rates have declined across nearly every region of the world, producing an unprecedented convergence toward below-replacement fertility regardless of income level, culture, or development stage. These shifts raise concerns about population aging, labor force contraction, and long-run fiscal pressures. Even within the United States, where overall fertility has trended downward, there exists striking geographic heterogeneity. County-level general fertility rates range from above 65 births per 1,000 women in parts of Utah, South Dakota, and Alaska, to below 45 births per 1,000 in urban corridors of the Northeast and Pacific Coast. These wide, persistent differences raise a natural question: to what extent do these spatial patterns reflect causal effects of place, rather than compositional sorting or socioeconomic confounders? Housing costs, childcare availability, labor‐market structure, and local family norms are often cited as drivers of fertility. Yet whether “high-fertility places” truly cause higher fertility (or simply attract individuals predisposed to larger families) remains unclear. While descriptive variation in realized fertility is well documented, credible estimates of causal place effects on fertility are scarce. This study brings new causal evidence by leveraging the quasi-random assignment of U.S. military personnel to hundreds of installations across the country. Standard economic mechanisms centered on opportunity costs or the quantity–quality tradeoff increasingly fail to explain contemporary fertility patterns. An emerging view emphasizes the role of place: local social norms, community family structures, and local family environments may shape individuals’ fertility decisions. Our objective is to quantify how exogenous exposure to different geographic family environments affects subsequent fertility, marriage, and divorce among young service members. Methods: We construct a longitudinal dataset of enlisted personnel linking duty-station histories to annual fertility outcomes, dependent-status records, and marital transitions. New recruits complete initial training and are then assigned to installations in commuting zones nationwide according to manpower needs. We estimate causal place effects by regressing fertility outcomes on a full set of commuting-zone indicators, using inverse-propensity score weights to adjust for military branch and occupation. We estimate gender-specific place effects and examine marriage and divorce as related behavioral margins.

In this paper, I conduct research on whether there exists downward nominal wage rigidity and, if so, the reason why there is such downward nominal wage rigidity in the labor market of South Korea. To investigate whether nominal wage is rigid downwards, I utilize data from the Korea Labor and Income Panel Study (KLIPS). I find that, similarly with the findings of Park and Shin (2017), a significant proportion, 26.3%, of job stayers experience wage cuts. However, when I analyze the KLIPS using the histogram-location approach developed by Kahn (1997), I find that, in the labor market of South Korea, if the rates of wage changes less than zero belong to a bin, the proportion of job stayers included in the bin decreases by 46.4%. This implies that clearly there exists downward nominal wage rigidity in the labor market of South Korea. The downward nominal wage rigidity is just not strong enough to prevent all of the wage cuts. Since there is empirical evidence that nominal wage is rigid downwards, I examine the reason why there exists such downward nominal wage rigidity by implementing surveys targeting personnel directors and workers. The results of the surveys reveal that fairness (i.e., the need for acceptable reasons that can justify wage cuts), the preferences of firms and workers for increasing wage profiles, money illusion, labor unions, and the concerns about moral hazards, adverse selection, and turnover are the main factors that can explain downward nominal wage rigidity in the labor market of South Korea. This study is the first research that investigates downward nominal wage rigidity in the labor market of South Korea by applying the histogram-location approach developed by Kahn (1997) and conducting surveys targeting personnel directors and workers.

Driven by the profound impacts of AI and automation on global labor markets, this study examines the career trajectories of highly educated Millennials/Generation Y individuals in Taiwan (born in the 1980s). The analysis uses unique, long-term administrative data on labor records from 2014 to 2023. This data is linked to university entrance exam scores and family background characteristics, such as parental education and age. We precisely quantify how individual academic ability and socioeconomic background affect job-switching frequency and wage growth across various industries. Empirical results show that female workers undergo fewer job changes than male workers during their prime working years. However, when female workers transition jobs, the increase in their maximum insured salary is significantly smaller than that of male workers. This suggests that while women achieve greater job stability after switching positions, their mobility is associated with lower wage gains than men's. Additionally, workers with longer employment prior to changing jobs exhibit lower overall job mobility but experience higher wage growth after transitioning. Additionally, higher university entrance exam performance, measured by percentile ranks, is associated with reduced job-switching frequency and increased wage growth during transitions. This pattern suggests that initial academic aptitude at the time of college admission has a lasting impact on both job stability and earnings potential. With respect to family background, college-educated workers whose parents possess higher educational attainment achieve greater wage growth. However, parental education does not significantly influence job-switching frequency. Notably, having older parents is positively associated with both wage growth and the total number of job changes. Labor mobility patterns also differ substantially across sectors. Workers in industries that require advanced professional and technical skills, such as information and communication technology (ICT), receive the highest wage premiums when transitioning to other sectors. This reflects current technological and industrial trends. Collectively, these findings provide important insights into skilled labor dynamics. They have significant policy implications for the development of future labor regulations to address changing industrial and societal needs.

Minimum wages have been known to improve health and socioeconomic outcomes across many developed countries. However, the causal effect of minimum wages on crime is conflicting. Some studies shown the effect is negative, suggesting that the minimum wages improved crime rates, while other studies shown the contrary, suggesting the minimum wages increased crimes. There is little consensus on the effect of minimum wages on crime. This study contributes to this strand of the literature by examining the causal effect of minimum wages on crime in Japan. Using a difference-in-differences approach and 2002-2014 Crime Statistics in Japan, we showed that when prefectures increased the minimum wages following the Japanese 2008 policy, this reduced the crime rates, specifically property crime, of prefectures. Not significant effect was observed for violent crime. We also showed that the effect seems to be driven by break-and-enter for property crimes, suggesting that income effect from minimum wage policy may have blunted the profit-driven motivation for crime. Overall, this suggests that the minimum wages may be an effective policy tool to reducing the criminal behaviors in some countries.

Perspectives of Healthcare Human Resource Executives were collected through a nationwide survey to better understand current workforce challenges and strategies within the healthcare industry. The findings offer comprehensive insight into critical human resource practices that influence recruitment, hiring, and retention across multiple employee levels. The survey identified key areas experiencing the most significant employee shortages, highlighting roles that present ongoing recruitment challenges for healthcare organizations. Respondents also shared effective recruitment tools and strategies currently being used to attract qualified candidates, including both traditional and innovative approaches. In addition, the survey explored the types of incentives offered to different employee groups, such as clinical professionals, entry level management and executive positions, providing insight into how organizations tailor incentives to meet workforce needs. Human Resource Executives identified the top competencies used when hiring employees at various levels, emphasizing both technical skills and soft skills. Preferred customer service qualities were also discussed, illustrating how expectations differ based on employee role and level of patient interaction. The survey further examined the factors considered most important in employee selection, including experience, cultural fit, education, and interpersonal skills. Finally, the results addressed how organizations take into account employee retention and satisfaction, along with the tools used to support both. Respondents described strategies aimed at improving retention, such as professional development opportunities, competitive compensation, and supportive work environments. Tools used to enhance employee satisfaction were also highlighted, offering practical insights for strengthening workforce stability and engagement within healthcare organizations.

Assessing Trade Benefits of Regional Free Trade Agreements on Global Fish and Primary Agricultural Commodity Trade: 1989-2019. ABSTRACT This paper provides a comparative analysis of the impact of regional trade agreements (RTAs) on global trade in fish and primary agricultural commodities. We estimate trade creation and diversion effects using a Poisson pseudo-maximum likelihood (PPML) estimator with high-dimensional fixed effects and apply the split-panel jackknife (SPJ) bias-correction procedure of Weidner and Zylkin (2021) to obtain bias-corrected estimates. The results show that the EU, NAFTA, ASEAN, and SADC associations increased trade among members in both trade categories. The MERCOSUR agreement facilitated trade in fish among members, while the EAC increased only trade in primary agricultural commodities among association members. Our results reveal that the trade creation effects of RTAs are clearly dynamic. However, the export and import diversion effects show mixed trend patterns. The heterogeneity of RTA effects varies over time, across trading blocks, and shows notable asymmetric benefits within trading country pairs.

This paper investigates how communication between supervisors and subordinates affects salary outcomes in a fully remote work environment. The research question is whether remote communication—via text or online meetings—can substitute for in-person interaction in shaping promotions and salary growth, and whether supervisor ability moderates these effects. While prior studies highlight proximity bias and flexibility stigma in hybrid or on-site settings, little evidence exists on whether comparable mechanisms operate when all employees work remotely. The study contributes to the literature by leveraging a rare setting: a Japanese firm, operating fully remotely across Japan and 23 countries, for which complete administrative HR data and digital communication logs are available. These include monthly Slack messages, Zoom meeting counts, and meeting duration for each supervisor–subordinate pair. The dataset spans March 2023 to March 2024 and contains no measurement error from surveys or recall. To identify causal effects, the paper exploits routine supervisor rotations that occur every March and September. These rotations are shown to be exogenous to department characteristics and employee performance. Using this institutional feature, we implement a difference-in-differences-in-differences (DDD) model comparing employees who communicate with their newly assigned supervisor above versus below the median level (text or meeting time). The outcome is monthly salary, which directly maps to promotion levels in this firm. The results show that, on average, neither text communication nor meeting frequency with a new supervisor increases monthly salary or salary growth. However, supervisor heterogeneity fundamentally alters the effect: when the newly assigned supervisor is rated in the top 25% of all supervisors, subordinates who hold more meetings with that supervisor receive substantially higher salary outcomes. Meeting duration, rather than frequency, matters most. The largest effect appears when monthly meeting time is between four and eight hours—producing roughly 9–16% higher monthly salaries and 21–24% larger salary increases compared with those meeting less than four hours. In contrast, text communication exhibits no measurable effect. These findings indicate that high-quality, synchronous communication with highly capable supervisors enables knowledge transfer and accurate performance assessment in remote settings, mitigating concerns that remote work inherently disadvantages career progression.

This paper quantifies the substitutability between AI‑enabled medical robots and health‑care labor in hospitals. We ask whether AI robots act as substitutes or complements for different types of medical workers and use this to infer cross‑elasticities between AI capital and labor in the health‑care production function. We compile a novel dataset that merges NBER MPPES with an original, hand‑collected sample of roughly 30 news and trade‑press articles documenting hospital adoption of AI‑enabled robotic systems. Each article is coded for the adopting hospital, the type of robot, the year of installation, and the hospital’s location. We map hospitals to counties and build a county‑by‑year panel from the early 2000s through the last pre‑COVID year in NBER MPPES. This yields variation in local exposure to AI robots that we link to detailed measures of health‑care labor and output at the county–year level. Our empirical strategy is a stacked adoption difference‑in‑differences design that exploits staggered timing of AI robot adoption across counties. We estimate dynamic event‑study models of how employment, earnings, and procedure volumes for different categories of providers respond to robot adoption, conditioning on county and year fixed effects and flexible event‑time trends. Using these estimates, we recover cross‑elasticities of labor demand with respect to AI capital for surgeons, other physicians, and non‑physician staff. We also examine changes in procedure mix and intensity to understand whether AI robots shift tasks across worker types rather than simply displacing labor. The paper contributes to the economics of automation and health‑care technology by providing direct evidence on substitution versus complementarity between frontier AI capital and high‑skill medical labor. By quantifying these cross‑elasticities in a critical service sector, our results speak to the future of work for health professionals and to policy discussions on training, regulation, and the distributional impacts of AI in health care.

Over the past twenty-five years, advances in fintech, and the implementation and use of artificial intelligence, have ushered in an age of transformation in terms of access to finance, entrepreneurship, and labor market participation participation. However, research work in this area remains fragmented across disciplines, with limited integration between studies on digital financial inclusion, AI-driven economic empowerment, and gendered development outcomes. This paper bridges this gap by analyzing the evolution of the literature on fintech, digital credit, algorithmic decision-making, and AI-enabled financial systems. The paper further identifies persistent geographical, methodological, and conceptual imbalances, including limited macroeconomic analyses, uneven regional studies, and insufficient attention to algorithmic bias and AI governance. Ultimately, a structured review protocol - combined with bibliometric techniques - is applied to the body of the literature on the relationship between artificial intelligence (AI), fintech, and female economic outcomes in low- and middle-income countries (LMICs). This paper therefore contributes by providing the first integrated mapping of AI and fintech research, explicitly focused on female economic empowerment in LMICs, and by identifying the critical gaps that compromise cumulative knowledge and policy relevance.

forthcoming

Pregnancy and childbirth represent a major yet foreseeable shock to household finances. We study the financial impacts of this transition using a novel dataset linking parents in California birth records to their consumer credit reports. Our event study results show that mothers reduce credit card utilization after conception, with credit card balances falling by about 12% relative to the counterfactual. However, this self-insurance proves insufficient, and mothers experience persistent increases in financial distress after birth, with the incidence of any 30-day delinquency rising by as much as 20% relative to a pre-pregnancy baseline of about 5 percentage points. These impacts are unlikely to be explained by decreased credit access or intra-household specialization. Heterogeneity analysis reveals that the deterioration in financial outcomes is concentrated among privately insured mothers, while lower-income mothers on Medicaid do not show the same deterioration, consistent with buffering from expanded eligibility for safety net programs.

Public debt has continued to expand in many advanced economies, with Japan standing out for its exceptionally high debt level. Despite long-anticipated demographic pressures such as population aging and a declining birthrate, fiscal consolidation has proven politically difficult. One explanation for the persistence of fiscal deficits is public misperception or unawareness of fiscal conditions. Through an online survey experiment in Japan, this study examines whether (i) receiving information about the size of central government debt updates respondents' beliefs about debt and their preferences for central government spending and taxes, and (ii) receiving information about the average own-source revenue ratio of municipalities updates respondents' beliefs about that ratio and their preferences for municipal expenditure. The analysis finds that both treatments successfully update respondents' beliefs. However, neither treatment affects preferences for government expenditure at the national or local level. There is suggestive evidence that the debt information treatment increases support for raising tax revenue. The analysis also finds that trust in government is positively associated with support for higher spending and taxes. These results suggest that simply announcing debt levels or revenue structures is insufficient to build public support for fiscal consolidation.

Taiwan’s transition to democracy is an intriguing issue and the political evolution has also fostered a vibrant economic development worthy of further exploration. From 1895 to 1945 Taiwan was under the reign of Japan. After Japan was defeated by the US, Taiwan was reverted to China under the control of Kuomintang (KMT). To disguise its authoritarian control, KMT kept three elected institutions, the National Assembly, the Control Yuan and the Legislative Yuan. To maintain its claim to be Free China and represent the Whole China, KMT designated the representatives for the three institutions as the First Term and could serve until death or when KMT takes control of mainland China again, whichever comes first. Since KMT could never get back the mainland, the Free China claim serves as an excuse for KMT to oppress the opposition and Chiang, Kai-Shek, its leader, effectively acts like an emperor. Taiwanese have been fighting for their civil rights even under the colonial rule of Japan and even more so under the reign of KMT. However, KMT has deployed various gimmicks to suppress the call for democracy. For instance, the local dialect was despised and its use was forbidden publicly. Registered Birth Place (RGB) was recorded on the identification card, as a tool for segregation against native Taiwanese (TWN) and to confer privileges to the Mainlanders (MND), those following KMT to Taiwan from mainland China. Taiwanese suffer from political persecution and economic hardships during those years under martial law, no political parties could be formed and no press freedom. Anyone dared to voice discontent were at the risk of being jailed. In 1991 the First Term representatives came to an end. Thereafter, all representatives are elected in Taiwan. In 1996 the first time a president was elected directly by popular vote in Taiwan. The whole experience may look easy to outsiders. However, exploring the saga deeper would tell a different story.

Mangrove forests provide ecosystem services, including protecting against coastal erosion and storm-surge, yet their contribution to coastal property protection is undervalued in the market pricing of insurance premiums. This paper develops a bio-economic model to estimate the marginal contribution of each hectare of mangrove forest to homeowner storm and hurricane insurance using outputs from convolutional neural networks (CNN). In training the deep learning model, high-resolution Sentinel-2 satellite spatial data are extracted from 284 georeferenced field coordinates from NOAA Fisheries data in Florida. The CNN model’s estimated probability that an image chip contains dense mangrove is used as a proxy for mangrove density in a logistic regression to estimate the probability that a qualifying insurance claim occurs in specific locations. The baseline model shows that a unit increase in mangrove density reduces the log-odds of a qualifying claim by 32.7 and after controlling for location characteristics the log odds is -37.4. The predicted probabilities a qualifying claim are used in the expected insurance premium pricing rule. The results show that an average risk-neutral household living near a mangrove forest, with 80% insurance coverage, 20% baseline probability of qualifying loss and $50000 of insurance payout in the event of storm surge should have their insurance premium reduced by approximately $1,653 per annum for each additional hectare of nearby mangrove forest. The study shows that healthier mangrove forests serve as risk-mitigating natural capital, lowering the expected losses and reducing insurance premiums. It is recommended that such value-enhancing natural capital be incorporated into property valuation and insurance pricing.

This paper develops a structuralist macroeconomic model grounded in feminist economics’ critique of growth- and market-centered paradigms. The model explicitly links gendered labor structures to macroeconomic demand, distribution, and profitability, thereby illuminating mechanisms through which gender inequality is reproduced—or potentially transformed—at the structural level. This macroeconomic model addresses some fundamental feminist economics issues such as care-based investments in human capacity building, women’s labor force participation, women’s uptake of market work and household care burden. The paper engages with the fundamental distributive conflict and bargain between workers and capitalist class following Bhaduri and Marglin (1990) and further building on the demand and distribution frameworks of Braunstein, Van Staveren, and Tavani (2011), Braunstein and Tavani (2020), Heintz and Folbre (2022), the model incorporates care-burden majorly carried out by women as a primary channel for human capacity building and productivity growth where fertility choices of women play an important role both from the demand and supply side dynamics. Current ongoing discussions regarding declining fertility levels and low female labor force participation in various East-Asian and South-Asian country context provide a great background and context behind exploring the macroeconomic consequences of these gendered variables such as, fertility choices and care-burden borne primarily by women for human capacity building. The model incorporates investment in human capital as a key demand-side channel through which gender relations shape goods market equilibrium. On the supply side, producers’ equilibrium is modeled to capture how women’s fertility choices make a difference in terms of women’s relative labor force participation affecting productivity, markups, and profitability. Women’s relative participation in paid market work plays a central role in the model’s dynamics: increases in women’s paid employment alter capacity utilization and pricing behavior through productivity effects, while simultaneously influencing aggregate demand via human capital accumulation rooted in unpaid and care labor, reproduction of the labor force and productivity in market production sphere. es. In doing so, the paper contributes to feminist efforts to prioritizing care, justice, and collective well-being over growth and accumulation in an alternative macroeconomic framework.

Forthcoming

This paper studies how tightening access to household debt relief affects early-life human capital formation. I take advantage of the 2005 Bankruptcy Abuse Prevention and Consumer Protection Act (BAPCPA) for the USA as a quasi-natural experiment. The BAPCPA is a nationwide reform that substantially increased the cost and restrictions of personal bankruptcy. Although the reform altered households’ financial safety net, little is known about its implications for children’s early-life outcomes. I combine two complementary empirical designs. First, using U.S. natality microdata, I examine whether prenatal exposure to the post-BAPCPA financial environment affected birth outcomes such as birth weight, prematurity, and prenatal care utilization, leveraging variation in pre-reform local bankruptcy filing rates to identify differential exposure. Second, using longitudinal household data from the Panel Study of Income Dynamics and its Child Development Supplement, I study how children’s cognitive outcomes evolved before and after the reform, focusing on families with greater pre-reform financial vulnerability. The analysis emphasizes precise cohort timing around the October 2005 implementation of BAPCPA and employs difference-in-differences methods to assess the effects.

In a carbon-budget constrained economy, delayed adjustments following the bursting of a carbon bubble can complicate the macroeconomic response to emission shocks and generate asymmetric effects of conventional countercyclical monetary policy. While an increase in emissions may prompt a policy rate cut, such easing can unintentionally stimulate carbon-intensive activity and further raise emissions. Conversely, a rate hike in response to declining emissions may reinforce emission reductions. This asymmetry highlights the limits of monetary policy operating in isolation and underscores the need for coordination with enforceable carbon policies. This paper contributes to the macroeconomic analysis of climate change by embedding carbon-bubble dynamics into a standard Keynesian framework. The model captures not only income and interest-rate channels but also wealth and credit channels arising from the revaluation of carbon-intensive assets. I examine how carbon emission shocks affect output and inflation and assess the effectiveness of interest-rate targeting in addressing climate-related macroeconomic instability. The analysis yields two main results. First, a tightening carbon budget and the bursting of a carbon bubble can generate both inflationary and disinflationary forces: supply-side pressures may raise prices, while wealth losses, higher financing costs, and credit rationing can depress aggregate demand, producing stagnation accompanied by disinflation or deflation. Second, conventional monetary policy has limited effectiveness in addressing climate change because it does not differentiate between fossil-fuel and clean-energy sectors. Empirical evidence from a VAR model using U.S. data from 1970–2022 supports these predictions. Countercyclical monetary policy can either amplify or mitigate emission dynamics depending on the nature of the initial shock. Effective climate stabilization therefore requires an enforceable carbon policy that generates favorable emission shocks and complementary monetary measures—such as targeted support for low-carbon sectors—to facilitate a smooth transition toward clean energy.

One out of seven families with children experience food insecurity (USDA 2019), and each year, approximately 6 million households with children access a food bank to meet their nutritional needs (Feeding America 2014). While multiple government programs and non-profit organizations aim to alleviate hunger and improve nutritional intake, little is known about how the formal and informal safety nets interact. This project makes progress on this question by evaluating how demand for charitable services changes in response to an expansion in nutritional assistance. The free school meals program is the largest form of nutritional assistance for school-aged children. Since 2012, this program has dramatically expanded under the Community Eligibility Provision (CEP), which allows schools to offer free meals to all students, regardless of family income. As such, CEP offers additional nutritional resources to families who do not qualify for other forms of income assistance, such as SNAP, and those who may rely on aid from private and non-profit organizations. We combine information on the timing and implementation of CEP with data on household financial health from the Federal Reserve Bank of New York/Equifax Consumer Credit Panel (CCP), matched at the Census Tract level. For every individual in the CCP, we identify whether the school in their district participates in CEP. We restrict the CCP to individuals most likely to have a child in school, 25-54. We use a dose response difference-in-difference model to estimate the impact of CEP on bankruptcy, severe delinquency, amount of debt, access to credit, and number of accounts in severe delinquency (total, auto, mortgage, and revolving accounts) and mobility.

Building on last year’s “policy” presentation, now accounting and financial data has been procured for deeper analysis into past and current microeconomic situations for various colleges and universities in the U.S., as well as abroad. This presentation delves into widely used private industry operations management techniques, that U.S. colleges and universities implement, or need to implement, for continued success in the education market.  This presentation will then address the macroeconomic impacts that our U.S. colleges and universities have on U.S. GDP, as well as the GDP of other nations.  Analysis into the international market of post-secondary "tertiary" education services will be compared and contrasted with the U.S. educational institutions. Higher education institutions being “public”, “non-profit private”, or even “for-profit private”, as well as “R-1 research” versus “teaching focused”, and how each classification affects the quality of the education services provided, will be addressed. A big question that will be covered is how the U.S. became such a dominant international player in the higher education industry. An explanation of the evolution of the U.S. post-secondary environment, from the late 1600s, through the federal land grants of the late 1800s, to the present day. This presentation will explain relevant financial and managerial accounting practices, which provide higher education institutions with the necessary fiscal information to then implement microeconomic tools, and then allow education institutions to have continued financial sustainability. Marginal cost, marginal revenue, and demand elasticity calculations, as well as marginal productivity analysis will provide critical economic insight for educational institution management teams to experience continued success. Finally, this presentation will also address the international tariff environment on higher education service providers, as well past and current international policies on student travel visas.

County-level fiscal spillovers are crucial in determining the role of cross-county fiscal coordination in fulfilling national economic commitments. Using county-level economic and fiscal spending data from 2010 to 2018, we examine the spillover of fiscal spending across China's regions through output, trade, investment, and consumption linkages. Specifically, we find positive fiscal output, investment, and employment spillovers and a temporary positive consumption spillover that turns negative over time. Consistent with extant studies, we also find that the county-level fiscal spillover effect in China is comparable to that in the United States.

This study develops and estimates a structural equilibrium model to analyze the causal impact of the 1995 deregulation of university enrollment quotas in non-metropolitan areas on the demand for private education. Leveraging the estimated parameters, we conduct counterfactual simulations to evaluate the general equilibrium effects of a comprehensive deregulation of enrollment quotas. Our theoretical framework characterizes the interaction between universities and students as a two-stage game: in the first stage, universities determine their optimal enrollment capacities through Stackelberg competition; in the second stage, students engage in a rank-order tournament, endogenously determining their investment in private education given the capacity constraints. The model is estimated via the Simulated Method of Moments (SMM) using micro-data from the Education and Social Mobility Survey. The estimated model successfully replicates the salient empirical patterns following the 1995 policy shock. Specifically, while enrollment quotas in non-metropolitan areas expanded by approximately 40%, the persistent regulatory constraints in the Seoul Metropolitan Area (SMA) led to a 14.2 percentage point decline in private education participation in non-metropolitan regions. Our counterfactual analysis reveals that a nationwide deregulation—extending to SMA universities—would substantially mitigate the competition for college entrance, potentially reducing the national private education participation rate by 50% to the 20% range. These findings suggest that the liberalization of university enrollment quotas serves as a potent policy instrument for internalizing the negative externalities of overheated private education demand. Furthermore, our results underscore the necessity of incorporating university-student equilibrium and competitive tournament dynamics into the welfare analysis of educational policies.

This paper studies peer effects on household spending in tutoring programs (shadow education) in South Korea. Using the Korean Education Longitudinal Study (KELS), I exploit the transition from middle to high school, when students in many equalization areas were assigned to general academic high schools by district lottery and re-sorted into new schools and peer groups quasi-randomly. Peer intensity is measured as the leave-one-out mean of classmates’ tutoring expenditure. I estimate average treatment effects using propensity score matching (PSM) and augmented inverse probability weighting (AIPW), conditioning on rich pre-treatment covariates including baseline spending, achievement, and family background. Under conditional ignorability and overlap, the results reveal a threshold pattern: entering the top quartile of peer spending raises own tutoring expenditure by about $78–$91 (PPP 2007) per month, with the largest responses in Math and English. Smaller peer changes within the top half show weak or model-dependent effects, while downward moves lead to sharper and more immediate cuts. Placebo tests using income changes and value-added specifications point away from contemporaneous family income shocks. A back-of-the envelope mapping from the Top 25 percent transition implies that households raise tutoring spending by about 24 percent of a peer spending increase to maintain relative position. The findings document a strong, asymmetric peer mechanism consistent with a “rat race” dynamic in which relative position shapes household investment and contributes to overspending pressure in education.

This study aims to examine policy discourse surrounding the transition to an absolute grading system for the College Scholastic Ability Test (CSAT) using language network analysis and to provide implications for the future direction of university admissions. To this end, news articles were collected from two recent periods in which discussions on absolute grading for the CSAT were prominent: the 2017 period under the Moon Jae-in administration and December 2025 under the Lee Jae-myung administration. For each period, language network analysis was conducted to identify central keywords and to analyze degree centrality and betweenness centrality among key terms. The findings indicate that, in both 2017 and 2025, policy discussions were characterized by a top-down approach led by the presidential office, featuring terms such as “concern” and “criticism,” without presenting concrete goals or strategies. In addition, broad and ambiguous terms such as “improvement” and “adjustment” were frequently observed. Based on these results, this study suggests that future reforms of the CSAT should be grounded in clearly defined future competencies and desired human capital profiles, and should be developed through sufficient consultation with stakeholders in the field.

This study examines how a student’s ordinal rank in private tutoring expenditure within the classroom affects subsequent tutoring demand and academic achievement. Using unique data on South Korean fourth graders, I exploit variation in within-classroom rank conditional on absolute expenditure levels. Controlling for classroom fixed effects, I find that a higher percentile rank significantly increases short-term tutoring demand but has no significant effect on academic performance. The effects are concentrated among students with high tutoring expenditures, suggesting that competition is particularly intense among heavy users of private education and that incentives to maintain relative position strengthen with greater investment. I also find suggestive evidence that parental involvement increases with rank, pointing to a potential behavioral mechanism.

Supreme Court decisions affect market values not only of the direct parties to a case but also of firms exposed to similar legal issues. We identify such firms as those that file amicus curiae briefs supporting their preferred outcomes. Many of these firms report related litigation pending in lower courts and emphasize that the Court's decision will have important business implications. Firms are more likely to file if they are more sensitive to the legal change and face lower filing costs. We document that stock prices of amici firms respond significantly to Supreme Court decisions, and the magnitude of the price reaction is consistent with the selection effect in filing. Moreover, we find that a higher fraction of publicly traded amici supporting a given side is associated with a greater likelihood of a favorable ruling for that side. These findings have implications for judicial conflicts of interest.

Forthcoming

This paper analyzes heterogeneity in post-pandemic labor market adjustment across industries, contrasting healthcare and social assistance with leisure and hospitality. Using CPS monthly microdata (2015–2023), JOLTS separations, and LEHD Job-to-Job flows, we estimate sector-specific wage Phillips curves and link labor market tightness to wage growth and employer-to-employer mobility. Both sectors experienced significant increases in quit rates after 2020; however, the underlying drivers are markedly different. Leisure and hospitality workers showed substantial upward mobility across industry wage-premium quartiles and frequent exits to other sectors, consistent with improved outside options in a tight labor market. Conversely, healthcare and social assistance workers displayed limited outward mobility, higher rates of within-industry job downgrades, and separation elasticities that weakened further after COVID-19. These dynamics align with well-documented structural features of care industries—credential-based occupational closure, a disproportionately female and older workforce, and pandemic-induced deterioration in working conditions. Industry wage-premium rankings show no systematic upgrading for healthcare. Our findings suggest that wage compression disproportionately benefited mobile low-wage workers outside the care sector, while quits in healthcare reflect non-wage factors and deteriorating working conditions.

This paper studies housing structure depreciation and the heterogeneous effects of maintenance and renovation investments on preserving housing value. Using a flexible framework based on double machine learning that decouples housing structure and land, this study finds that housing structure values depreciate at an average rate of approximately 10 percent per year in the absence of any maintenance or renovation. It reveals substantial heterogeneity in the effectiveness of different types of investment. Routine maintenance plays a critical role in slowing depreciation, generating an average return of roughly 4 dollars in preserved housing structure value for every dollar spent. In contrast, renovation expenditures exhibit significantly lower returns, with an estimated return of only 0.6 dollars per dollar invested. The estimates are robust across a wide range of specifications and machine learning algorithms. Overall, the paper highlights the importance of maintenance behavior in housing markets and provides new evidence on the relative efficiency of maintenance investment in mitigating structural depreciation.

Research has shown that time-series stock market returns collected prior to the pandemic can predict COVID-19 retail firm bankruptcies by using a chaos-based model, which is a logistic regression. In that model, the Lyapunov exponent is the covariate. This study examines whether the model can be improved by adding a second covariate, the firm’s Debt to Asset ratio. The new model, and each of its covariates, is statistically significant at conventional levels.

We examine how Critical Audit Matter (CAM) industry norms and auditor–CEO social ties shape non‑professional investors’ assessments of financial reporting risk (FRR). Using an online experiment, we find that investors judge FRR to be lowest when the CAM disclosure aligns with the industry norm and no auditor–CEO social tie is present, compared with conditions involving a norm‑inconsistent CAM or a disclosed social tie. The presence of an auditor–CEO social tie also reduces investment interest, lowers perceived audit quality and auditor credibility, and increases perceived auditor liability for CAM‑related misstatements. In contrast, norm‑consistent CAMs lead to more favorable investment decisions, greater reliance on the audit report, and higher perceived management liability. These results highlight how CAM disclosures and auditor independence cues jointly influence investor judgment, offering insights for managers, auditors, and regulators as they consider the communication and governance implications of CAM reporting.

Corporate Location Decisions under Varying Global Corporate Laws Kashi Nath Tiwari KNT’s Academic Financial Research Business-laws and trade-laws are not identical across the globe. GATT, UNO, and other world organizations serving as watchdogs are either unable to implement the agreements or overlook the other way when the GATT rules are compromised. Authoritarian regimes across the globe are notorious with “do-not-care.” How do multinational corporations operate across the globe under varying rules and regulations? Capital control, currency control, restrictions of repatriation of profits, trade unions, tariffs, quotas, production-subsidies, production-taxes, consumption-subsidy, consumption-tax, nationalization, outright threats. Unwittingly, the corruption in the production sector and consumption sector is prevalent. Repatriation of profits (even if the funds are blocked) is easy by giving some money to enforcers. This paper studies the impact of heterogenous laws across the globe in doing business.

How is wealth inequality expressed in a supply-constrained housing market? Using household microdata from the Korean capital region, I document three cross-sectional patterns: owner-occupancy rises steeply with wealth but saturates at the top, ownership of non-primary-residence real estate is heavily concentrated there, and the poorest households are almost entirely excluded from such ownership. I build a lifecycle general-equilibrium model with collateral-constrained, indivisible housing and a fixed housing stock, and use it to ask which margins of the ownership distribution respond when inequality rises. The model implies an asymmetric restructuring: because owner-occupancy is already saturated at the top, higher inequality concentrates additional property holdings there rather than raising ownership, while at the bottom it pushes more households below the down-payment threshold and out of owneroccupancy. Equilibrium prices barely move, so the mechanism reorganizes who holds housing rather than changing its price level—a distribution puzzle rather than a price puzzle. The channel is general in form but conditional in strength, operating most strongly where down-payment constraints bind and housing supply is inelastic.

The extraordinary monetary policy interventions during the COVID-19 pandemic were notable not only for their unprecedented scale but also for the speed at which they were implemented. While expansionary monetary policy is intended to stabilize financial markets, the implications of rapid policy execution for banking system risk remain insufficiently understood. This paper examines whether the velocity of monetary policy adjustments amplifies financial fragility by affecting bank credit risk transmission. Using COVID-19 as an exogenous shock to identify monetary policy actions, we analyze Credit Default Swap (CDS) spreads for publicly traded banks in the United States and Canada over the period 2020–2022. We employ both panel-based methods, including Generalized Least Squares (GLS) and System Generalized Method of Moments (GMM), and time-series approaches, including Structural Vector Autoregressions (SVAR) and Local Projections, to capture the dynamic effects of policy speed on bank credit risk. Our results show that conventional monetary easing generally reduces bank credit risk; however, an excessively rapid pace of policy adjustment weakens these stabilizing effects and is associated with heightened volatility in CDS spreads, indicating potential policy overshooting and increased systemic fragility. These risk-amplifying effects are more pronounced for smaller banks and during periods of elevated financial stress. In contrast, the risk-mitigation effects of unconventional monetary policies, particularly quantitative easing, are ambiguous, emphasizing the complexity of crisis-driven interventions. This study contributes to the monetary policy and financial stability literature by highlighting the importance of policy execution speed, in addition to policy magnitude. The findings suggest that central banks should carefully calibrate the timing of interventions to balance swift crisis response with the preservation of banking system stability.

We assess the impact of living in a bank or protentional bank branch desert, or in an area with recent, local bank or credit union branch closures, on household access to credit. We use different distances to measure banking deserts and bank closures rates in urban, mixed and rural areas. For households, we look at credit access by 5 year age group. Overall, we find no economically significant effects on household access to credit We use a national 5% sample of US household credit records from 2017 to 2024, as well as FFIEC/S&P Global bank branch data and ACS tract characteristics. We employ a rich set of covariates and use propensity score matching and related models to control for unobserved confounders.

This paper examines how a widely accessible retail central bank digital currency (CBDC) interacts with bank stability, and what role bailout policy should play when households can move funds instantly into safe public money. A key concern is that a convenient, risk-free alternative to deposits could accelerate digital bank runs, even as it changes banks’ incentives to provide liquidity insurance. We build a Diamond–Dybvig-type model in which banks fund long-term investments with demandable deposits. Households face uncertain liquidity needs and choose whether to withdraw early or keep their deposits in the bank. Banks decide how much they will pay to early withdrawers, balancing the benefit of providing liquidity insurance against the cost of taking on more maturity mismatch and becoming more vulnerable to self-fulfilling runs. The government finances a public good through taxes and can either bail out banks in bad states or credibly commit to a no-bailout policy. We introduce a retail CBDC as a safe asset that depositors can hold alongside bank deposits. CBDC changes incentives in two ways: it gives depositors a safe outside option, so banks can provide less liquidity insurance with smaller early payouts, but it also creates an additional run margin if depositors expect trouble. We compare four environments—bailouts versus no bailouts, with and without a retail CBDC—and study run risk, banks’ maturity mismatch, crisis-related taxes, and depositor welfare. Three results stand out. First, when bailouts are expected, introducing CBDC can increase fragility: banks are more willing to promise high early withdrawals, and the availability of a safe asset makes it easier for funds to flee, raising expected fiscal costs. Second, under a credible no-bailout commitment, CBDC stabilizes: banks cut back on early payouts, maturity mismatches fall, and runs occur in fewer states, with lower required taxes. Third, when multiple equilibria exist, no-bailout regimes yield higher welfare, and the welfare gains are larger with CBDC. The results suggest that CBDC design choices and crisis policy are complements. A retail CBDC reduces fragility in our model when bailout expectations are credibly restrained; without that, it can increase incentives to run.

This paper studies how product bundling affects network participation, quality innovation, and welfare when network externalities arise in a primary monopolized market. We show that bundling breaks the symmetry between firms in the competitive market, uniquely positioning the dual-market firm as the high-quality supplier. The effects of bundling depend critically on the strength of network externalities. When network externalities are sufficiently weak, bundling expands network participation and improves overall welfare, stimulating quality investments in the network product and forcing the rival into a quality-enhancing survival defense. By contrast, when externalities are sufficiently strong, bundling severely restricts access to the network product, causing its output to contract. This drastic output contraction destroys firms' incentives to invest in quality, leading to a uniform decline in product qualities across the board. As a result, the severe access restriction under strong externalities lowers both consumer surplus and welfare. Bundling incentives depend critically on the firm's initial quality position. A low-quality firm always finds bundling profitable because it overturns the initial quality ranking, a scenario that explicitly triggers the aforementioned non-monotonic effects. By contrast, a high-quality firm bundles only over an intermediate range of network externalities, which unambiguously reduces consumer surplus and welfare.

The COVID-19 pandemic caused a sharp decline in international tourism and raised uncertainty about the extent of its recovery. Although Taiwan’s inbound visitor arrivals have gradually increased since border reopening, the recovery in visitor expenditure may not have followed the same path. This paper examines the recovery of average daily expenditure among inbound visitors from seven major source markets. Individual survey data from the Survey Research Data Archive for 2003–2019 are aggregated into quarterly series, and market-specific SARIMAX models are used to estimate conditional counterfactual expenditure paths for 2020–2024. Recovery is evaluated by comparing observed expenditure with the level expected under pre-pandemic expenditure relationships. The empirical results in the current study demonstrate that no single dynamic structure or set of explanatory variables performs best for all source markets. Average daily expenditure among visitors from China, Hong Kong and Macao, and Japan remained below the corresponding counterfactual levels. A similar shortfall was found for Korean visitors under the Benchmark model. Expenditure among Malaysian and US visitors was close to the counterfactual level, while expenditure among Singaporean visitors exceeded it. These findings indicate that visitor arrivals and expenditure intensity may recover at different rates. Source-market differences should therefore be considered in both the evaluation of tourism recovery and the design of tourism promotion policies.

Understanding the motivations behind antitrust laws is crucial given that courts continue to rely on legislative intent to guide their interpretation of these statutes. Accordingly, this paper revisits the origins of antitrust by examining why some states adopted antitrust statutes before the Sherman Act of 1890. We leverage large-scale digitized newspaper data and machine learning methods to measure antimonopoly sentiment during the 1880s and connect these patterns to states’ statutory choices. This analysis allows us to test competing explanations for the rise of antitrust, informing modern debates about the original aims of antitrust policy. Our findings provide the first quantitative evidence linking contemporaneous public sentiment to legislative outcomes in the pre-Sherman era. Newspaper coverage in the years leading up to early state antitrust statutes points largely toward a broad, popular concern with monopoly rather than a coordinated agrarian or small-business campaign. Regression results similarly show that newspaper attention to monopoly—rather than measures of agricultural, small-business, or monopoly-sector strength—best predicts early adoption. Together, these results suggest that the initial wave of antitrust legislation reflected diffuse public salience around monopoly rather than concentrated special-interest pressure, lending support to the “public interest” interpretation of antitrust’s origins.

This study investigates how digital platform capability (DPC) drives the adoption of circular economy practices (CEP) among Malaysian manufacturing SMEs, emphasizing the mechanisms and contextual conditions that shape this relationship. Drawing upon the resource-based view (RBV) and institutional theory, the research conceptualizes DPC as a strategic technological resource that fosters circular transformation through enhanced innovation and responsiveness to external pressures. Specifically, the study examines the mediating role of innovation orientation (IO)—reflecting an organization’s propensity to pursue and apply novel ideas—and the moderating effect of environmental regulatory pressure (ERP), which represents the external institutional forces encouraging sustainability-oriented change. An empirical analysis was conducted using data collected from dual respondents representing 237 Malaysian manufacturing SMEs, analyzed through structural equation modeling (SEM) to ensure reliability and robustness. The findings reveal that DPC exerts a significant positive effect on both IO and CEP, suggesting that digital platforms enhance a firm’s innovation capability and operational circularity. Furthermore, IO not only directly improves CEP but also mediates the relationship between DPC and CEP, confirming that innovation acts as a critical behavioral mechanism linking digital capability to sustainability outcomes. The moderating analysis indicates that ERP strengthens the positive association between IO and CEP, implying that firms operating under stricter environmental regulations derive greater sustainability benefits from innovation. However, ERP does not significantly moderate the DPC–CEP relationship, suggesting that digital capabilities yield consistent advantages regardless of regulatory intensity. This study advances the circular economy literature by positioning DPC as a technological enabler, IO as a behavioral conduit, and ERP as an institutional catalyst for circular transition. It contributes theoretically by integrating internal (RBV-based) and external (institutional) perspectives into a unified framework explaining digital-enabled sustainability. Practically, the findings offer strategic guidance for SMEs and policymakers by highlighting the need to strengthen digital and innovation capacities while aligning policy instruments to accelerate sustainable industrial transformation in emerging economies.

This study examines how local living subsidies affect the share of near-poor seniors in Taiwan using county- and city-level panel data from 2015 to 2024. The analysis considers gender differences and regional spillover effects. The results show strong persistence and clear spatial clustering in the distribution of near-poor elderly populations, suggesting that elderly economic vulnerability is structural rather than temporary. Higher per-capita living subsidies are associated with a larger share of near-poor seniors, with the strongest effect observed among women, indicating that female seniors are particularly responsive to public income support. Significant spatial dependence further implies that policy changes in one region can influence neighboring areas. Additional findings show that higher education levels are linked to lower elderly poverty risks, while higher housing prices reduce the share of near-poor seniors. Overall, the results suggest that local living subsidies play an important role in supporting aging in place and highlight the need to incorporate gender and spatial considerations into elderly welfare policy design.

This study examines the relationship among public childcare resources, female employment, and fertility in the context of Taiwan’s persistently low birth rate. Governments across East Asia have invested heavily in childcare infrastructure and family subsidies to mitigate declining fertility, yet policy outcomes remain uneven. Taiwan offers an ideal empirical setting, given its rapid expansion of public childcare programs since 2018 and the coexistence of high female labor participation with one of the world’s lowest fertility rates. Using county-level panel data for the period 2010–2024, this research applies a spatial Durbin model (SDM) to analyze how public childcare availability and female employment influence fertility. Public childcare resources are measured by the number of available childcare slots per 1,000 children aged 0–2, while female employment represents the labor market participation rate among women aged 20–39. Control variables include income level, housing costs, population density, educational attainment, and unemployment rate. The empirical analysis is expected to yield three key findings. First, public childcare availability is likely to have a positive direct effect on fertility, reflecting the reduction of time and financial burdens associated with childrearing. Second, higher female employment rates may exert a negative effect on fertility, underscoring the persistent work–family trade-off in a rigid labor market. Third, spatial spillover effects are anticipated, suggesting that the expansion of childcare facilities in neighboring areas can affect local fertility decisions through cross-district accessibility and information diffusion. By integrating spatial econometric methods with demographic policy analysis, this study provides quantitative evidence on how public childcare policies interact with labor market dynamics to shape fertility behavior. The findings are expected to offer valuable insights for regional policy coordination and the development of balanced family–work support systems in East Asian societies facing similar demographic challenges.

This study investigates the impact of publicly disclosed soil liquefaction potential maps on residential housing prices, employing a Difference-in-Differences approach integrated with a hedonic pricing model. The empirical analysis is based on housing transaction data from Kaohsiung, Taiwan, surrounding two major announcements: the 2016 release of the Preliminary-Level Soil Liquefaction Potential Map and the 2020 release of the Intermediate-Level Map. By constructing treatment and control groups according to spatial risk classifications and their reclassifications, this study evaluates how the disclosure of environmental risk affects market valuation. We found that the disclosure of high liquefaction risk information significantly depresses housing prices in affected zones, with price declines ranging from 2.8% to 3.7% following the announcements. Among various types of risk reclassifications, only properties reclassified into high-risk zones experienced a statistically significant decline in housing prices, suggesting that buyers respond most strongly when the perceived hazard escalates to the highest risk tier. Other risk transitions—such as moderate upgrades or downgrades—did not exhibit consistent or significant effects on price movements.

Title IV aid is capped by the Cost of Attendance (COA), which colleges publish and partially control through discretionary components. This paper tests whether four-year institutions use COA-setting as a strategic “headroom” lever to expand permissible aid and borrowing, and whether Pell Grants and federal loans substitute for institutional grants within the cap. We construct an institution-year panel merging IPEDS COA and component measures with IPEDS Student Financial Aid income-banded aid averages for full-time, first-time students, linked to Federal Student Aid aggregates on Title IV borrowing and Pell. We estimate dynamic panel models of log COA and non-tuition COA components as functions of institutional grants, Pell, and loan exposure, controlling for institutional resources, enrollment, selectivity, and state appropriations, with institution and year fixed effects. System GMM addresses persistence and simultaneous determination of price and aid. Results provide packaging-aware evidence on COA headroom and substitution among Pell, loans, and institutional grants across sectors and institutional resource levels.

Although college football has a long and rich cultural history in the U.S., relatively little attention has been paid to the lower divisions of the National Collegiate Athletic Association (NCAA). For many Division II football programs, stadium attendance is a major source of revenue. Consequently, administrators’ ability to set justifiable athletic department budgets and make informed ticket-pricing decisions depends on the accuracy of football attendance forecasts. Most existing attendance studies rely on short- run data. Schreyer and Ansari’s (2022) comprehensive review of sports attendance literature finds that only five percent of 195 studies used data from 20 or more seasons, while 80 percent used five years or less. To address this gap, this paper examines Division II college football attendance using a game-day data set spanning nineteen seasons (2001-2019), the longest period used in this context. In addition, this paper contributes to the literature by comparing alternative measures of outcome uncertainty (season wins, season win percentage, average points scored); employing instrumental variables in selected estimating equations to account for potential endogeneity between attendance and ticket prices; exploring how the effect of county population on attendance might change in sign or magnitude as population rises; considering of alternative measures of travel cost; and including independent variables typically excluded from previous studies. These variables include macroeconomic measures, multiple measures of weather conditions, proximity of the campus to a National Football League (NFL) stadium, the sum of points-per-game scored by each team, the presence of a marching band, and games involving a team identified as a historically-black college (HBC). The regression model takes general form in which attendance is the dependent variable, and economic characteristics, demographic variables, team performance measures and game attributes, are three sets of control variables. The random effects model is chosen because some of the variables of interest are time invariant (e.g. proximity to an NFL franchise and conference affiliation). The assumption is made that the school specific component of the error term is distributed independently of the explanatory variables.

As online course offerings become increasingly prevalent in institutions of higher learning, instructors are rapidly shifting from paper and pencil assessments to online assessments. Online assessments offer several key advantages, including reduced administrative costs, the ability to use a wider variety of multimedia resources, and faster analysis of results. To reduce the potential for academic dishonesty in online assessments, many instructors have adopted various proctoring solutions, though their effectiveness has not been studied in depth. This paper analyzes the role of human proctors used in online assessments, a preventative measure used in testing centers and classroom settings where students complete assessments online but under supervision. Moreover, we study the effect of self-reported test anxiety on exam scores, which can be heightened in the presence of a proctor. We conducted randomized controlled trials to estimate these effects in a large-enrollment macroeconomic principles course at a U.S. public university. Students took multiple open book online exams during the course; however, only one (determined randomly) was proctored. Our analysis also investigates the effect of proctoring and anxiety by gender and grade point average to further explore the impact that proctoring has on student performance.

The Korean Child Tax Credit (CTC), introduced in 2015, provides cash transfers to low-income households with children based on income and asset criteria. Using the Korea Welfare Panel Study from 2014 to 2024, we study the credit's effects on household consumption and debt. Because survey-based measures imperfectly replicate administrative eligibility records, we employ a Fuzzy Regression Discontinuity Design, constructing a combined running variable from the more binding of the income and asset thresholds. Reconstructed eligibility is significantly associated with actual receipt, producing a discontinuity of roughly 23 percentage points at the cutoff in the 2018 sample, driven primarily by the income criterion. However, the second-stage estimates reveal no statistically significant effects of CTC receipt on total living expenditures, food expenditures, private education expenditures, or debt reduction. The findings, robust to alternative income and asset definitions, suggest modest short-run behavioral responses among marginal recipients near the eligibility threshold.

We ask whether the size of the school and class a student attends in adolescence shapes labor market outcomes in adulthood. Linking the Korea Labor and Income Panel Study to administrative school records, we match individuals who entered the third year of middle school between 2005 and 2007 to the average class size, pupil--teacher ratio, and enrollment of their school, and follow them to 2022, when they are in their early thirties. Ordinary least squares estimates suggest that larger classes are associated with higher adult earnings, but this relationship is likely driven by the endogenous sorting of students across schools. Using an instrumental-variables strategy in the spirit of Hoxby(2000) that exploits idiosyncratic, trend-deviating fluctuations in the school-age population of a district, the positive association disappears and no robust effect of class or school size on earnings, employment, graduate-school enrollment, or completed schooling can be identified. The findings are consistent with a literature reporting weak effects of class-size reduction on test scores, and they raise questions about the case for continued investment in class-size reduction as Korea's youth population declines.

Drug free school zones (DFSZs) provide ground for increased penalties for drug crimes committed near schools or places where children gather. In this study, we utilize a novel data-set to estimate the causal effect of DFSZ enhancements on court sentencing outcomes in Nashville, TN. We exploit the variation of arrestees' location and use the instrumental variable two stage least squares estimator to study the causal effect of being charged with the DFSZ enhancement. Notably, we find that enhancement charges against eligible offenders are more likely to be applied to Black arrestees than White arrestees, leading to lasting impacts across races in sentencing outcomes. Our estimates suggest that being charged with the DFSZ enhancement increases the likelihood of an individual being found guilty by 11 percentage points (15\%) - driven by large effects for White defendants. The enhancement also increases sentence length for Black defendants by roughly 1.25 years. We present preliminary evidence of White defendants being offered better plea deals as the mechanism driving these results.

This study investigates whether preventive healthcare habits formed under parental influence during adolescence persist into adulthood. Using data from the National Longitudinal Study of Adolescent to Adult Health (Add Health), the analysis estimates the effect of receiving dental examinations in adolescence on adulthood dental exams using a selection-on-observables framework. In contrast to health behaviors analyzed in prior intergenerational transmission studies, dental exams are preventive, non-urgent, and largely independent of genetic or addictive influences. Results show a strong and persistent effect of adolescent dental exams on adult dental exam behavior that is robust across OLS, propensity score matching, and LASSO-based variable selection.

This paper investigates whether artificial intelligence adoption is associated with smaller gender wage gaps. Using data from the Current Population Survey matched with occupation-level AI exposure measures, I document that the gender wage gap is substantially smaller in high-AI-exposure occupations (6%) compared to low-exposure occupations (29%). To explore potential mechanisms, I develop and estimate a structural employer learning model where AI improves the precision of productivity signals. The estimated model implies that women face noisier productivity signals at baseline, and that AI would need to disproportionately improve female signal precision to rationalize the cross-sectional pattern. However, identification tests reveal important limitations: 62% of the cross-sectional Female × AI correlation is absorbed by pre-existing task objectivity measures, the within-occupation event study is null, and the industry-level IV estimate reverses sign, and age heterogeneity in the raw gap-AI gradient is inconsistent with the employer learning prediction. Robustness analysis across five independent AI exposure measures shows that the sign and magnitude of the Female × AI interaction varies substantially by measure. The contribution is therefore best understood as doc- umenting a robust cross-sectional correlation and proposing one theoretical mechanism among several, rather than establishing that AI causally reduces the gender wage gap.

Access to prescriptions opioid pain medications has become an important public health concern in the United States, yet researchers have not established clear causal evidence on how it affects fertility behavior and maternal-infant health. In this paper, I examine whether residing in counties with higher prescription opioid prescribing rates influences fertility, prenatal care utilization, maternal pregnancy behaviors, and infant health outcomes. County-level prescription opioid access may affect fertility through multiple biological, behavioral, and socioeconomic channels. Higher prescribing environments may worsen population health and economic instability while also altering healthcare utilization, contraceptive consistency, and pregnancy timing. Conditional on pregnancy, local prescription opioid access may further influence prenatal investments and infant health through changes in healthcare access, maternal behaviors, and local economic conditions. I combine restricted-use National Vital Statistics System (NVSS) natality microdata with county-level prescription opioid prescribing rates and state opioid policy data, including Prescription Drug Monitoring Program (PDMP) implementation and related opioid prescribing regulations. To identify causal effects, I exploit staggered adoption of opioid prescribing policies across states and variation in pre-policy opioid access across counties. Specifically, I examine whether fertility and maternal-infant health outcomes changed disproportionately in counties with higher levels of prescription opioid access following opioid prescribing reforms. I analyze fertility and maternal-infant health outcomes, including birth rates by maternal age, race, and education; prenatal care initiation and utilization; maternal pregnancy behaviors such as tobacco use and gestational weight gain; and infant outcomes including birthweight, low birthweight, preterm birth, cesarean delivery, and Apgar scores. I also examine heterogeneity across demographic and socioeconomic groups because economic instability and barriers to healthcare access may disproportionately affect vulnerable populations. This paper contributes to the literature in two ways. First, I jointly examine fertility behavior and downstream maternal-infant health outcomes within a unified framework. Second, I move beyond correlational evidence by utilizing quasi-experimental variation generated by opioid prescribing policies and differential county-level prescription opioid access. My findings will provide new evidence on how prescription opioid access shapes fertility and maternal-infant health outcomes in the United States.

A key issue facing policymakers in advanced economies is the dramatic aging of the population and the consequent increase in demand for long-term care. The long-term care sector in the United States is plagued by chronic labor shortages, and immigrant workers may be important in filling gaps in the health care workforce. However, while some empirical work has been done on immigrant labor supply in the context of nursing homes, little is known about home health care. In this study we start by using a state-level “shift-share” variable that captures variation in the foreign-born population at the state and year level to show that higher levels of immigration are associated with higher labor supply in occupations that provide the majority of patient-hours of home care (nursing aides, home health aides, and personal care aides), particularly for workers in home-based settings. We then estimate the effect of immigration levels on long-term care utilization using data from the Health and Retirement Study (HRS). The estimates suggest that higher immigration levels are associated with a significantly higher probability of receiving both paid (formal) and unpaid (informal) care, in some cases at the same time. Elderly individuals are significantly more likely to receive care because of high immigration levels if they have memory impairments, trouble getting in and out of bed, walking, or eating.

This study examines the determinants of tenure and promotion among accounting faculty members in U.S. universities. Using a hand-collected dataset of 669 tenure-track accounting faculty who earned Ph.D. degrees from U.S. universities, we investigate how employer prestige, research productivity, doctoral training, educational background, and individual characteristics jointly influence promotion from assistant professor to associate professor. We estimate both probit and logit models and report marginal effects to facilitate the economic interpretation of the estimated results. The results indicate that faculty members employed by highly ranked research institutions face significantly more demanding promotion standards than those employed by less research-intensive institutions. Promotion probabilities are also lower for faculty employed by private universities and for those who spent more years in full-time doctoral study. In contrast, research productivity and holding a bachelor's and/or master's degree in accounting are positively associated with promotion. After controlling for research productivity, employer characteristics, doctoral training, and other observable factors, male faculty members exhibit lower promotion probabilities, whereas faculty members who received their undergraduate education in China exhibit higher promotion probabilities. The research rankings of the doctoral institution and dissertation advisor, as well as faculty age, do not appear to have significant effects. Our findings are robust across alternative measures of research productivity, probit and logit specifications, and alternative measures of employer prestige. The study contributes to the literature on the accounting academic labor market by providing new evidence on the relative importance of employer prestige, research productivity, doctoral training, and individual characteristics in tenure and promotion decisions.

Interdependent Utility and Preferences Kashi Nath Tiwari KNT’s Academic Financial Research There are n-agents, n-countries, and n-factors of production, and n-goods: agents: a1, a2, a3, … an countries: c1, c2, c3, … cn factors of production: f1, f2, f3, … fn goods: g1, g2, g3, … gn In addition of the consumption-goods (cg), the utility of each agent is a function their own acctions as well as the actions taken by other agents: u1 = u1(cg1; a1, a2, a3, … an) u2 = u1(cg2; a2, a1, a3, … an) u3 = u3(cg3; a3, a1, a2, a3, … an) un = un(cgn; an, a1,a2, a3, … an-1) Both the short-run utility and the long-run utility are interdependent across the agency-domain. While your own actions are known to you, the actions of other agents are unknown. In a a game theoretic framework, this paper investigates the intertemporal utility maximizations within the framework of a stochastic optimal control theory.

Conventional theory suggests that monetary policy has stronger real effects when prices adjust slowly. We provide the first evidence on geographic variation in price stickiness and its consequences for monetary policy transmission. Using household expenditure data and retail scanner data that cover a large share of U.S. retail consumption from 2006 to 2019, we construct county-level measures of price stickiness. Variation in these measures is driven by differences in local consumption bundles across goods with differing inherent stickiness. We then examine how county-specific price stickiness affects the transmission of monetary policy to employment and inflation. Even when controlling for other factors known to be important to monetary policy transmission, local price stickiness has a significant impact on the responsiveness of employment and inflation to monetary shocks.

Empirical evidence shows that expansionary monetary policy can widen income inequality. Using a dataset with improved coverage of high-income households, this study shows that although the labor-income Gini declines following monetary easing, the total-income Gini increases due to a widening gap in financial income. This divergence reflects differences in household income sources and asset composition: top households earn a large share of income from risky financial assets that benefit from monetary easing, while bottom households rely primarily on labor income and hold interest-bearing assets whose returns fall when interest rates decline. To explain this pattern, I develop a modified two-agent New Keynesian framework that incorporates a financial accelerator mechanism. In this framework, risk-taking top households borrow deposits of bottom households as external finance through banks, so that monetary easing lowers the price of risk-taking—the lending spread—and redistributes financial income toward the top. As a result, the model can replicate the opposite impulse responses of the two Gini coefficients, consistent with the empirical findings. It highlights the importance of the financial income channel in understanding the redistributive effects of monetary policy.

This paper investigates the macroeconomic impact of the Eurosystem’s negative interest rate policy (NIRP). Using a medium-scale DSGE model with a banking sector, we capture two counteracting channels of lower (and possibly negative) interest rates: the expansionary bank lending and the contractionary net-worth channel. Our non-linear estimation on Euro area time series explicitly accounts for the time-varying effective lower bound. We find that NIRP had a positive effect on aggregate variables, suggesting that the bank lending channel outweighed the net-worth channel.

Abstract: This paper examines how monetary policy uncertainty shapes the transmission of monetary policy shocks through different macro-financial channels. While existing studies show that uncertainty can weaken the effects of monetary policy, less is known about which specific transmission mechanisms are altered when uncertainty about future monetary policy is high. Using monthly U.S. macro-financial data and high-frequency monetary policy surprises as external instruments, I estimate a smooth-transition proxy-SVAR in which lagged monetary policy uncertainty determines the state-dependent propagation of an identified monetary policy shock. The framework allows the impulse responses to differ between high- and low-uncertainty environments while maintaining a common contemporaneous shock identification. The analysis further decomposes Treasury yields into the expected short-rate path and term premium components, and private borrowing rates into the expected policy-rate path and a combined excess-premium component. The results show that high monetary policy uncertainty attenuates the initial pass-through of contractionary monetary policy shocks to Treasury yields, private borrowing rates, and real activity. In particular, the expected short-rate path responds less under high uncertainty, indicating a weaker policy-expectations channel. However, credit spreads, term premia, and combined excess premia display more persistent responses, especially at longer maturities and in private borrowing markets. These findings suggest that high monetary policy uncertainty does not simply shut down monetary policy transmission. Instead, it changes the composition of transmission by weakening the expectations channel while increasing the relative importance of the risk-premium and credit-cost channels. The paper contributes to the literature by opening the black box of state-dependent monetary policy transmission and identifying how uncertainty reshapes the relative strength of competing transmission mechanisms.

As one senior defense official observed, “the complexity we face exceeds our engineering grasp.” Defense acquisition programs continue to struggle with cost, performance, and schedule. One recently proposed solution is Artifiicial Intelligence (AI)-enhanced digital twins as a program management tool. By rapidly processing large datasets, these tools may improve management effectiveness by compensating for human limits. Used well, this framework might strengthen engineering and program management through more realistic scheduling, greater strategic flexibility, and improved resilience to unexpected challenges. However, this approach faces at least two major challenges. First, although digital twins offer a promising, high-confidence way to manage complexity in program execution, they are not foolproof and have led to spectacular failures. When the machines fail, organizations still need experienced and capable personnel who can intervene effectively. This creates a second concern: there are strong reasons to believe that heavy reliance on AI can erode the human capital needed for such intervention. However, well-designed AI-enhanced management tools may help mitigate that risk by supporting, rather than replacing, human expertise. Among other thigs, this indicates a need for more emphasis on human capital issues than is now evident.

The United States has increasingly relied on export controls as a central instrument of economic statecraft, particularly in its efforts to restrict diffusion of key technologies to firms in the People’s Republic of China (PRC). Despite the rapidly expanding use of export controls against the PRC, little empirical evidence exists on how these policies affect the financial performance of the foreign firms that become subject to them. This paper utilizes Moody’s Orbis database to construct the largest possible sample of both treated and control firms and applies modern difference-in-difference and event-study estimators to identify the causal effects of export-controls. Our results indicate that adding firms to the BIS Entity List causes significant declines in key measures of financial health such as revenue, employment, and assets but does not significantly impact indicators such as solvency ratio that may predict a firm’s survival. These findings suggest that U.S. export controls reduce the size of listed PRC firms but, on average, do not force them out of business.

forthcoming

Employee stock options are commonly valued using option-pricing frameworks that rely, directly or indirectly, on signals drawn from traded financial markets. Yet the informational content of listed-option market data is not uniform across contracts. This paper examines whether salient round-number strikes in listed equity option markets exhibit distinctive liquidity characteristics and considers the implications of those patterns for the valuation of employee stock options. The central argument is that visible market activity at salient strikes may overstate executable liquidity and thereby affect the interpretation of market-based valuation inputs. For forensic economists, the practical implication is that listed-option evidence can inform employee stock option valuation, but it should not be used mechanically without attention to liquidity quality, moneyness, and strike clustering.

Bangladesh has made some remarkable progress in alleviating poverty. However, there is a dearth of information on the dynamics of poverty in Bangladesh. Using the six rounds of the Mahabub Hossain Panel Data (MHPD), we investigate the entry into and exit out of poverty in Bangladesh. The MHPD dataset has been collecting data on rural Bangladeshi households since 1987-88, so it contains rich information on about four decades of income and other socioeconomic variables. Our preliminary analysis shows that most households faced at least one round of income poverty, whether using upper- or lower-poverty line of Bangladesh. We also measured the multidimensional poverty index (MPI) among those households, and there is much less transition into and out of MPI poverty. This indicates that rural households have better overall well-being and income poverty measurement does not capture all the improvements of overall well-being. We also run a Proportional Cox Hazard Model and find that households engaged in day labor are more likely to fall into poverty. Households dependent on farming are less likely to fall into poverty, and more likely to exit poverty. The findings give us a glimpse of what kind of households are experiencing transition in and out of poverty, and this can help craft policies to help households permanently exit poverty.

This paper compares stylized universal basic income (UBI) and earned income tax credit (EITC) reforms for aging Japan. Both policies are evaluated under the same fiscal discipline: a proportional tax on labor income is set so that tax revenue approximately covers transfer spending in each period. Using a stationary heterogeneous-agent model with calibrated labor supply, the analysis is anchored to grouped Japanese household income data (KSK 2023) and the official employment rate in 2023. We report how the two schedules differ in the tax rate required for fiscal balance, in labor participation and hours, in asset accumulation, in welfare, and in a stylized aging exercise. The study is not a fully calibrated life-cycle counterfactual for Japan; the earnings calibration improves fit to grouped income moments but still understates top-tail inequality relative to the survey data. At the baseline flat benefit level in the experiments, EITC requires a higher revenue-balancing tax rate than UBI (about 0.31 vs. 0.21); when the same tax rate is applied to both policies (about 0.26 at the midpoint), labor-force participation is higher under EITC, asset upper tails are thicker under EITC, and a larger share of households hold no liquid assets under EITC than under UBI. The exercise provides a disciplined Japan-focused comparison under a common fiscal closure rule.

This study develops an Overlapping Generations (OLG) framework to examine how public and private health expenditures interact and how their optimal mix is determined within an economy. Although both types of health spending have increased globally, countries differ substantially in their reliance on public versus private systems due to institutional structures, demographic pressures, and welfare arrangements. Existing research has mainly analyzed determinants of medical spending or focused solely on public provision, leaving limited understanding of how the two sectors jointly shape lifetime health capital and social welfare. To fill this gap, the study builds a dynamic general equilibrium model in which public and private health expenditures are endogenously chosen, and extends it by incorporating substitution elasticity, efficiency (quality), and congestion factors such as waiting time. Individuals live for three periods (childhood, young adulthood, and middle adulthood) and derive lifetime utility from consumption, savings, and health-related activities. Health capital is generated through a CES production function combining public and private healthcare inputs, allowing their relationship to vary from substitutes to complements. Public healthcare effectiveness operates through two channels: quality and congestion, both of which influence the marginal productivity of public spending. These channels shape the median voter’s preferred tax rate financing public healthcare, producing a political-economic equilibrium. This framework enables the model to assess how differences in efficiency, congestion, substitutability, and income distribution explain cross-country variation in public healthcare spending shares. Model results show that higher public-sector quality increases the marginal productivity of public expenditure and shifts voter preferences toward expanded public provision. Conversely, congestion and long waiting times reduce effective efficiency, raising reliance on private healthcare. Efficiency and congestion emerge as key determinants of the equilibrium public-spending share across countries, highlighting how institutional and economic environments generate distinct health-financing patterns. The study contributes by endogenously modeling the interaction between public and private health expenditures, allowing analysis of structural mechanisms often inaccessible to empirical approaches. By embedding qualitative factors such as efficiency in health-capital formation and congestion into a quantitative OLG framework, it clarifies when expanding the public sector enhances welfare and under what conditions the optimal public–private mix shifts.

A designer offers vertically-differentiated positions to agents in the absence of transfers. Agents have private outside options and may reject their offers ex-post. The designer has preferences over the quantity of agents who accept each position. We show that under a general condition on the distribution of outside options, an optimal mechanism for the designer offers all agents an identical lottery, and we characterize this mechanism. When our condition does not hold, the optimal mechanism may require screening agents by offering a menu of distinct lotteries. Our results follow from a decomposition of agents' participation probabilities in any feasible mechanism.

This paper develops a ``reversed Spence'' signaling model in which abandoning a university degree reveals entrepreneurial ability and commits the founder to high effort. The dropout threshold rises with degree prestige and likelihood of degree completion; investors respond with better terms. Especially at high-prestige universities and in fields where grades convey little information to the labor market, students value a credential guarantee, which universities provide through lenient grading standards, strengthening the dropout signal as a side effect. A bifurcation result shows that even a negligible dropout population, combined with prestige feedback, suffices to create distinct elite and credential university tiers.

Replication is a fundamental pillar of science and increasingly considered by policymakers when choosing and pursuing programs based on published research. We coordinated over one thousand researchers across 250 teams who performed 15,000 reanalyses (robustness checks) while simultaneously assessing the consistency and reproducibility of published research. Our sample consists of papers published in peer-reviewed journals in economics, political science and behavioural science from 2022 onwards. About 90% of teams reported the original papers to be computationally reproducible from the provided replication materials. Still, approximately 25% of teams identified some form of research inconsistency (colloquially, coding errors or discrepancies). Our teams of researchers found about 70% of reanalyses remained statistically significant and having the same sign as the original paper. We further classified original papers’ into: (1) their respective fields (JEL codes, etc.) and (2) their design-based research method (i.e. identification strategy). Results from randomized controlled trials appear more robust relative to instrumental variables when comparing t-curves of the original papers’ statistics to those of our teams of researchers’ reanalyses. In a similar analysis of t-curves, development economics and labour economics appear to be more robust than economic history. Our results by field and method roughly align with the priors of 250 researchers which were solicited prior to our analysis.

The Inequality Process (IP) (Angle, 1983-2025) is a competitive exclusion process that appears to maximize wealth production. It is a particle system; i.e., with no agents making decisions. Labor income statistics teem with patterns implied by the IP. The IP is a stochastic particle system isomorphic to that of the kinetic theory of gases up to the properties of particles and transition equations. Angle (1983) did not find the IP by tinkering with the kinetic theory’s particle system. This presentation shows that an IP phase diagram of particle wealth like that of the IP’s fit to labor income data implies ‘stylized facts’ (soft invariances) of stock price statistics. Quantitative finance (qfin) experts have written that there is no mathematical model "like one from physics" operating on stock prices. This presentation refutes that claim. The IP has been adopted as econophysics (Byrro Ribeiro, 2020:154ff). The IP resolves the “equity premium puzzle” (EPP) without necessarily negating its perceived alternative, the Capital Asset Pricing Model. Mazzoni (2018:172) describes the EPP as “probably the most prominent puzzle in modern economics”. A speculation: the IP is a step toward a general mathematical theory of markets maximizing wealth production by rewarding the producers of wealth. Perhaps a better name for the model than ‘Inequality Process’ would be the ‘Work to Wealth Process’. This presentation relies on IP phase-diagrams to make its points visually. In the 1980's and '90's, IP papers were "bounce back" rejected by most econ journals with comments such as "There's no market model here; it can't be right." This presentation is a long-delayed answer to that criticism. My goal is to convince others to use and improve on the IP leading to its discussion in textbooks and research on markets. Thanks to Prof Philip Shaw, the WEAI, and the NYC Workshop on Computational Econ and Complexity for providing a forum for IP findings.

I propose an estimation algorithm for Exponential Random Graph Models (ERGM), a popular statistical network model for estimating the structural parameters of strategic network formation in economics and finance. Existing methods often produce unreliable estimates of parameters for the triangle, a key network structure that captures the tendency of two individuals with friends in common to connect. Such unreliable estimates may lead to untrustworthy policy recommendations for networks with triangles. Through a variational mean-field approach, my algorithm addresses the two well-known difficulties when estimating the ERGM, the intractability of its normalizing constant and model degeneracy. In addition, I introduce ℓ2 regularization that ensures a unique solution to the mean-field approximation problem under suitable conditions. I provide a non-asymptotic optimization convergence rate analysis for my proposed algorithm under mild regularity conditions. Through Monte Carlo simulations, I demonstrate that my method achieves a perfect sign recovery rate for triangle parameters for small and mid-sized networks under perturbed initialization, compared to a 50% rate for existing algorithms. I provide the sensitivity analysis of estimates of ERGM parameters to hyperparameter choices, offering practical insights for implementation.

The Krusell and Smith (1998) algorithm used to solve heterogeneous-agent DSGE models with macroeconomic uncertainty involves an estimation step to obtain the endogenous aggregate laws of motion required for agents’ conditional expectations. The conventional approach to this estimation step requires the modeler to specify a parametric functional form of these aggregate laws of motion, which are typically taken to be log-linear in the aggregate states. In this paper, we propose an alternative nonparametric approach to this estimation step, which avoids this ex ante specification requirement and instead allows for flexible functional form to better capture any nonlinearities present in the true aggregate laws of motion. Using a simple Krusell-Smith economy with endogenous leisure, we show that our nonparametric approach dominates the conventional parametric approach along a number of important including fit and prediction error.

Consider a predictive regression with a financial asset, such as a stock market return, as the dependent variable, and a macroeconomic variable, such as consumption growth, as the independent variable or predictor. Such regressions in practice often fail to yield evidence of return predictability. Economic effects are small and coefficient estimates are statistically insignificant. In contrast, if the predictor variable is aggregated at a lower frequency, for example annual consumption growth rather than monthly consumption growth, evidence of return predictability is much stronger, both economically and statistically. The first contribution of the paper is a model to explain this result. I assume that the macroeconomic predictor is measured with error and is subject to adjustment dynamics. For example, aggregate consumption might respond slowly to a shock due to adjustment costs. I further assume that the macroeconomic predictor may contain a long memory or slow moving component. This can occur if the mean varies over time, there are structural breaks, or the series contains markov-switching properties, etc. Under these conditions, aggregating the predictor variable at a lower frequency is more likely to show evidence in favor of return predictability, relative to the original predictability regression. The second contribution of the paper is an instrumental variables estimator to recover the true parameter of the original predictability regression. The instrument is constructed using a lagged low frequency value of the predictor variable. I derive the necessary conditions for consistency of the estimator, and a method inference. I show that the estimator is robust to the presence of measurement error, adjustment dynamics, and long memory. Finally, I present empirical results using a few common macroeconomic predictor variables (consumption growth, employment growth, industrial production growth, and inflation) to predict the aggregate stock market. I show that the standard OLS return predictability regression fails to show evidence of return predictability. In contrast, the proposed instrumental variables estimator yields economically and statistically significant evidence of return predictability.

The purpose of this study is to evaluate graduate students' academic performance, whether or not they attend the same institution where they obtained their undergraduate degrees, based on the institutional data from former NTHU and non-NTHU undergraduates enrolled in NTHU graduate programs. Three main aspects will be discussed in this paper: academic performance, student satisfaction, and the awareness of and participation in international activities. Our empirical results first show that former NTHU undergraduates tend to receive higher GPAs than those of non-NTHU undergraduates, whereas those non-NTHU undergraduates display a relatively earnest attitude toward their graduate coursework and self-report more favorable learning outcomes. Second, the NTHU undergraduate alumni are more dissatisfied with various aspects of NTHU campus life than non-NTHU undergraduates. Lastly, former NTHU undergraduates demonstrated greater awareness of and participation in international activities than did graduate students from non-NTHU institutions, suggesting higher levels of institutional engagement. Our results may provide evidence supporting the hypothesis that former NTHU undergraduates continuing their graduate studies at NTHU with a sort of home field advantage over the non-NTHU undergraduates.

General Education (GE) in Taiwan was introduced to broaden students’ intellectual horizons and promote learning across disciplinary boundaries. However, the expansion of university autonomy has led to increasingly diverse GE models and growing ambiguity regarding GE’s educational purpose. As a result, concerns have emerged that students may approach GE as a source of ”easy credits,” prioritizing lenient grading and low workload over intellectual exploration. While prior studies document GE’s contribution to communication, critical thinking, and other generic competencies, few examine how students’ GE course-selection behaviors shape these outcomes. Using a rich administrative dataset from a research university in Taiwan, this study examines students’ GE enrollment patterns, course-taking strategies, and competency development among cohorts graduating between 2018 and 2024. The results indicate that students participate extensively in GE courses outside their home disciplinary domains, although substantial differences remain across disciplinary backgrounds. While course-taking strategies are associated with GE enrollment behavior, disciplinary background remains the primary determinant of how students navigate the GE curriculum. With respect to competency development, greater participation outside one's home disciplinary domain is positively associated with interdisciplinary competence and career adaptability. By contrast, neither course-selection strategies nor the breadth of cross-domain participation are systematically related to competency outcomes. These findings suggest that the effectiveness of GE may depend less on which courses students choose and more on the extent to which they engage with disciplines beyond their primary field of study.

This study investigates how instructional modality impacts academic achievement across racial, ethnic, socioeconomic, and gender groups in higher education, addressing concerns about equity following major post-pandemic shifts in course delivery. The research asks: To what extent do in-person, online, and hybrid/HyFlex modalities influence academic outcomes, and how do these effects vary for Black, Hispanic, Asian, first-generation, and female students? Unlike prior work, this study empirically assesses heterogeneous equity impacts using post-COVID institutional data. Drawing from nine undergraduate economics course sections at a public four-year Hispanic-Serving Institution (HSI), the analysis covers 273 students over three semesters, spanning the transition from in-person to emergency remote teaching, a fully online term, and a hybrid/HyFlex setting. The dataset includes course grades, ethnicity, gender, first-generation status, and class standing. Ordinary Least Squares (OLS) regression models, with interaction terms by modality and demographic factors, estimate differential effects. Findings reveal modest reductions in academic performance for fully online modalities, while hybrid/HyFlex formats significantly improve outcomes. Interaction estimates demonstrate that alternative learning environments substantially narrow achievement gaps for Black, Hispanic, first-generation, and female students. These results suggest that flexible instructional structures can mitigate longstanding inequities, providing greater accessibility, adaptability, and targeted support based on student background. The study offers quantitative evidence that hybrid/HyFlex and well-designed online models can enhance equity in higher education, providing more effective post-pandemic policy and instructional design.

State appropriations are a central funding source for public higher education in the United States, yet they fluctuate considerably with economic conditions. Because universities have limited ability to quickly adjust tuition or alternative revenue streams, changes in appropriations can directly affect enrollment capacity, program offerings, and long-run institutional decisions. This paper studies how fiscal rules shape the responsiveness of higher-education funding to economic shocks, and how these funding fluctuations in turn affect university outcomes. I focus on balanced budget requirements (BBRs), which vary significantly in strictness across U.S. states and constrain governments’ ability to smooth revenue shortfalls during downturns. Using institution-level panel data from 2002–2019, I exploit this cross-state variation by interacting a measure of BBR stringency with state unemployment rates to construct an instrument for higher-education appropriations. Stricter BBRs require states to close budget gaps immediately rather than borrowing or deferring adjustments, causing appropriations to fall more sharply when unemployment rises. This strategy isolates plausibly exogenous, institution-independent variation in university funding that is not driven by contemporaneous changes in student demand. The results show that fiscal institutions meaningfully amplify the cyclicality of appropriations. A one-percentage-point increase in unemployment leads strict-rule states to cut appropriations by substantially more than weak-rule states. These funding reductions have significant consequences for universities. A 1 percent decline in state appropriations reduces undergraduate enrollment by roughly 0.6 percent in both the current and following year, indicating limited short-run adjustment capacity. By contrast, there is little evidence of systematic tuition increases, suggesting that universities respond to fiscal stress primarily along the quantity margin rather than through higher prices. I also find compositional adjustments in program offerings, with institutions reallocating resources toward lower-cost fields during periods of fiscal contraction. Overall, the findings highlight how state fiscal rules transmit macroeconomic shocks into the higher-education sector. Strict BBRs magnify downturn-driven budget cuts, leading to reduced access and shifts in academic opportunities. These results underscore the importance of considering state budget institutions when evaluating the consequences of higher-education funding.

Climate change is one of the most serious challenges that the world currently faces, and there is a need to enforce stricter environmental regulations. Against this backdrop, the study examines the climate change effect on environmental stringency through global comparative analysis across G7 countries, China, India, and Russia, for the period 1990-2022. The method of moment quantile regression (MMQR) shows a positive and significant effect of climate change on the Environmental Stringency Index (ESI), with countries exhibiting higher baseline stringency responding more aggressively to rising temperatures. The economic growth, the level of technological development, trade openness, and inflation also strongly and significantly influence the environmental policies of a country. This study emphasizes the challenge of balancing economic growth with environmental protection, which gives useful guidance for shaping policies to tackle climate issues.

We examine how managers respond to physical climate risks through firms’ textual disclosures in China. We find that managers adopt a downward tone management (DTM) strategy in the Management Discussion and Analysis (MD&A) in response to physical climate risks, using overly pessimistic language that does not reflect firms’ actual performance and thereby misleads investors. The channel test reveals the underlying reason that managers deliberately use climate events as causal justifications for poor performance in their MD&A text, treating physical climate risks as a credible scapegoat to reduce investor scrutiny pressure. This attribution behavior is further supported by the heterogeneity tests, which show that the DTM effect is concentrated among firms with the strongest incentives to deflect blame, those that miss analyst forecasts, those with higher agency costs, and those with a history of financial fraud. Further analysis using the Paris Agreement as an exogenous shock shows that the DTM channel became stronger after 2015, when public awareness of climate risks increased and climate attribution became more credible to investors. This tone manipulation also occurs in the supply chain context, where major suppliers face extreme weather conditions, leading customer managers to conduct the DTM strategy. This study makes several key contributions to the climate finance literature. First, while prior research primarily focuses on the direct economic costs of climate change, we shift the focus toward managerial opportunism, revealing how physical climate risks trigger strategic impression management. Second, we extend the literature on tone management by providing rare empirical evidence for the incentives behind DTM. Integrating self-serving attribution theory, we demonstrate that managers exploit the external and uncontrollable nature of climate shocks as a credible scapegoat to deflect blame from internal strategic failures. This research offers important implications for regulators and investors in emerging markets, highlighting the need for increased scrutiny of qualitative disclosures to ensure they reflect genuine environmental risks rather than opportunistic managerial behavior.

This paper studies how climate risk, progressive taxation, and sovereign default interact in a small open economy. We develop a quantitative model in which climate-related disasters reduce output and raise income inequality, while the government can invest in climate adaptation to mitigate future damages. Adaptation is financed through distortionary income taxes and sovereign borrowing, both of which affect default incentives. Tax progressivity redistributes income toward lower-income households but alters the government’s fiscal capacity and borrowing costs, creating a trade-off between insurance, efficiency, and default risk. We show that greater tax progressivity can facilitate adaptation by stabilizing revenues and reducing inequality, but may also increase default risk when borrowing constraints tighten following severe climate shocks. Quantitatively, climate risk substantially amplifies sovereign spreads and default probabilities, especially in economies with limited fiscal space. Adaptation investment reduces long-run default risk but may raise short-run borrowing needs, making its effectiveness sensitive to the tax system. Our results highlight the importance of jointly designing climate adaptation and tax policies in climate-vulnerable economies facing sovereign risk.

Forthcoming

As a new policy instrument integrating governmental strategic intentions with market-oriented principle, Government-Guided Funds (GGFs) play a crucial role in correcting market failures and optimizing resource allocation. Based on a 2012-2022 prefecture-level city-industry panel dataset of China, this study investigates the impact of GGFs on intra-industry resource allocation efficiency and its underlying mechanisms. The findings show that GGF investments significantly improves intra-industry resource allocation efficiency, with an enhancement effect outperforming that of market-oriented private equity funds. This improvement is achieved mainly through narrowing corporate financing constraint gaps and stimulating intra-industry competition. Furthermore, GGF investment in upstream industries boosts the focal industry’s resource allocation efficiency via supply chain transmission, while a concentrated investment strategy strengthens GGFs’ guiding role. Heterogeneity analysis reveals that the GGFs’ policy efficacy is more pronounced in technology-intensive industries, industries with high external financing dependence, and non-advantageous industries. It is also stronger in regions with robust fiscal capacity, advanced financial development, and a high marketization level. These findings provide empirical evidence for GGFs’ investment orientation, and offer important policy implications for optimizing government-market synergy in fund operations and maximizing GGFs’ industry-policy guiding role.

Remittances have become one of the largest sources of external finance for developing countries, yet evidence on whether they translate into productive investment, and under what conditions, remains limited. This paper examines whether governance quality shapes the association between remittances and domestic investment in low- and middle-income economies. Using annual panel data for 104 countries from 2000 to 2023 and two-way fixed effects estimation, we show that remittances have a stronger association with domestic investment than FDI, foreign aid, or portfolio investment inflows. This association is particularly strong in low-income countries, where remittances may play a larger role in easing household-level financing constraints. We further show that the remittance-investment association tends to be stronger in countries with higher governance quality. Among the six governance dimensions we examine, this pattern is most pronounced for corruption control and, to a lesser extent, political stability. These results suggest that institutional reforms targeting these dimensions may help channel remittance inflows toward domestic investment.

This paper examines the trade effects of non-tariff measures (NTMs) in the context of the U.S.–China trade conflict, with a focus on their relationship to tariff policy. Using detailed product-level data on U.S. imports from China, the analysis estimates the ad valorem equivalents (AVEs) of NTMs to quantify their implicit trade costs in tariff-equivalent terms. The empirical strategy relies on a price-based approach, exploiting variation in unit values across products, and incorporates standard gravity-type controls and fixed effects to account for cross-country and product-level heterogeneity. The paper further compares the evolution of AVEs with observed tariff changes to assess the relative role of non-tariff and tariff measures in shaping trade costs. By translating regulatory measures into comparable tariff equivalents, this study provides a unified framework to evaluate the contribution of NTMs alongside conventional tariffs in trade policy.

Historically, many energy-related import (income) demand elasticities such as for oil, natural gas and also electricity tend to be close to unity. Over time, as countries develop and become more energy efficient and alternative energies become more widely available, the income elasticity tends to fall something less than one. The picture is very different for rare-earths and more broadly speaking for critical minerals (hereafter CMs). Many critical minerals have some properties of a source of energy (such as coal), but can more often be thought of an input, such as more traditional minerals such as gold, silver and copper, where there is some degree of substitutability among similar minerals. As a result, as new technologies develop, “new” CMs emerge as the leading, and often indispensable, input and demand and often imports rise rapidly. Ultimately, in the short run, one can observe import elasticities being far greater than unity, sometimes reaching a value of 5 or six. But, these high elasticities are often relatively shortlived, falling to unity, or even negative, as demand and scarcity drive up prices and alternative technologies and alternative CMs come online and in demand. As a results import demand elasticities often experience a “rise and fall” pattern in only over 5 or ten years. As CMs (there are more than 50 types, and even more sub-types) are now a ubiquitous part of the global economy, it is vital to understand the nature of these elasticities and evolutions of those responses over time. Economic models that assume, for example, a static, and often low elasticity of demand would lead to simulations and predictions that fall far from the mark. This paper estimates import demand elasticities for a handful of the most common CMs using detailed US and Japanese data and compare this to more traditional estimates of other energy demands (gas, oil, etc.)

This paper studies how immigrant-ancestry networks affect the prices, not just the volumes, of U.S. exports. The prevailing view treats these networks as a friction-reducing institution that expands trade volumes; we ask instead whether they raise the price an exporter realizes. We match shipment-level quantities from U.S. Customs and Border Protection (CBP) bill-of-lading manifests to cell-level unit values from the U.S. Census state–HS6 export statistics, and instrument county–destination ancestry stocks with the "leave-out push–pull" shift-share instruments of Burchardi et al. (2019). Stronger ancestry ties raise the export unit value while lowering the quantity shipped within a product–destination market. This price-up/quantity-down pattern is inconsistent with the demand-expansion channel that the network-trade literature emphasizes and points to a different role for the network: not a lubricant that broadens a market but a barrier that narrows it. A two-layer quantity decomposition shows that ancestry broadens a county's export scope—more exporting firms and more products—while concentrating each individual product market into fewer, larger sellers. To probe the price effect, we interact ancestry with destination- and product-specific import-demand elasticities: the premium loads on the within-product, across-destination variation in elasticity, a pattern a cost-side quality attribute cannot generate, which points to demand-side pricing rather than cost. The evidence is consistent with ancestry networks raising prices by concentrating market structure rather than by expanding demand, though we cannot rule out a role for higher product quality.

This study examines the rise of the pharmaceutical and medical devices sector in Costa Rica following the implementation of the Dominican Republic–Central America Free Trade Agreement (CAFTA-DR). It analyzes firm level data focusing on the factors that allowed the expansion of high-technology manufacturing, particularly in pharmaceuticals and medical devices industries, and the importance of trade liberalization, free trade zone (FTZ) incentives, skilled labor, and innovation. The study explores how these factors are associated with increases in exports and sales in high-tech industries. The analysis is based on firm-level cross-sectional data from the Business Environment and Enterprise Performance Survey, 2005, 2010 and 2023.

Fast charging stations are considered crucial to hastening electric vehicle adoption and re- ducing emissions, but due to a lack of data, there has been little research on competition or consumer behavior. I compile novel price and quantity data to study the effect of fast charging station openings on incumbent station charging quantities and prices. I implement a difference-in-differences event study model that leverages the timing of nearby station entry. I find that following the entry of another station from the same network nearby, charging decreases significantly at incumbent stations and one network responds by lowering prices. However, when a competitor enters nearby, charging and prices remain unchanged at incumbent stations. My results contrast with the retail gasoline market, where new station entry brings down prices at competing incumbents but not stations under the same ownership. These differences may be due to both charging station capacity constraints and product differentiation due to charging networks’ use of proprietary mobile applications and subscription plans. My results suggest policies aimed at reducing search frictions (such as price posting regulation) could increase the consumer benefits of current charging station subsidy programs.

Do carbon credit markets generate larger compliance responses as prices rise, or is there a price level at which the signal changes character? Oregon's Clean Fuels Program, in operation since 2016, provides an unusually rich setting to investigate this question. Credit prices in the program ranged from $47 to $165 per credit between January 2018 and May 2024, wide enough to ask whether the compliance response is linear or piecewise. There is a behavioral turning point at $150.45 per credit, and the market has not crossed it once since January 2021. The analysis uses publicly available Oregon CFP compliance data. The outcome variable is the volume-weighted average carbon intensity of ethanol, measured in gCO2e per megajoule, which serves as the identification vehicle because it is the only CFP pathway with sufficient compliance history to detect thresholds. The benchmark model is a dynamic autoregressive distributed lag specification with Newey-West HAC standard errors. The long-run credit price multiplier is -0.123, significant at the 1 percent level, meaning a 10 percent sustained increase in credit prices is associated with a 1.2 percent reduction in carbon intensity, with the full adjustment accumulating over 12 to 15 months. A Hansen endogenous threshold regression then identifies $150.45 per credit as the turning point. Below it, producer compliance is present but irregular. Above it, compliance is consistent and durable. The regime-specific long-run effects are -0.096 and -0.103, respectively. Instrumental variables estimation using the second lag of credit prices confirms causal identification, and recursive CUSUM diagnostics find no structural break across the sample. An infrastructure interaction model tests whether DC fast-charging port deployment conditions the credit price effect across regimes. It does. Both interaction terms are negative, meaning DCFC expansion amplifies compliance responses at any price level, but the amplification is significantly larger below the threshold than above it. The equality test rejects identical effects across regimes, F = 18.43, p < 0.001. Charging infrastructure and credit price recovery are not competing policy instruments. The return to infrastructure investment is largest when the market signal is weakest, which is precisely where Oregon sits today.

Concentrated demand growth is reshaping network industries worldwide. In electricity markets, localized load expansions interact with transmission constraints, potentially distorting price signals away from marginal generation costs. While rising demand increases average energy costs, binding transmission limits may generate spatial price differences that reflect congestion rather than underlying production costs. This paper estimates the causal effect of concentrated demand growth on spatial price distortions in a network-constrained market. Leveraging the rapid expansion of data centers in Virginia between 2015 and 2024 and a propensity score matching difference-in-differences design at the census-tract level, we show that transmission-induced inefficiencies increase wholesale electricity prices by \$2.49 per MWh, approximately 70 percent relative to pre-treatment levels. These distortions arise from the spatial misalignment between new load and generation capacity. Price effects increase with distance between data centers and nearby generation. Within census tracts where load growth and generation are co-located, inefficient pricing is lower when generation is from fossil fuel sources but higher when it is from renewable resources. Effects are larger during peak demand periods, in areas with more intensive load expansion, and in predominantly non-White communities. These results highlight how transmission constraints shape the distributional consequences of concentrated demand growth and inform load siting, generation placement, and transmission planning.

Artificial-intelligence-driven data centers are reversing two decades of flat U.S. electricity demand and have generated questions about how this growth will impact electricity prices. We quantify this effect using an hourly, unit-level least-cost dispatch model covering wholesale electricity markets in the continental United States. We find that existing data centers have already increased wholesale prices by 2 to 6% on average nationwide, with substantially larger effects in regions hosting major data center corridors. Extending the model through 2028, we show that if proposed construction proceeds under high-utilization scenarios, wholesale prices could rise dramatically (50%), while more moderate build-out yields smaller (20%) but still meaningful effects. Impacts vary due to utilization and build-out assumptions. Finally, we use the model to address several policy discussions including optimal data center siting decisions and renewable build-out uncertainty.

This paper studies how monetary policy affects firms through refinancing risk embedded in their debt maturity structure. Movements in the yield curve affect firms not only by changing current borrowing costs but also by altering expected future refinancing conditions. Firms with greater short-term debt exposure are more sensitive to these changes and adjust both financing and investment more strongly. We use euro-area loan-level data, matched to firm balance sheets, for 2018 to 2025. We exploit variation in firms’ predetermined rollover exposure and compare firms borrowing from the same bank at the same time. Firms with more debt coming due issue longer-maturity loans, showing active refinancing risk management. When the yield curve steepens, these firms issue debt with shorter average maturities. By decomposing the yield curve, we show that higher expected future short rates induce maturity extension, while higher term premia discourage it; the expectations channel dominates in our data. High-frequency monetary policy shocks confirm that firms respond primarily to changes in expected future rates, rather than to the current policy rate. Financing constraints limit firms’ ability to adjust their refinancing risk exposure. Firms with greater rollover exposure are small, highly leveraged, low-liquidity, and lack access to the bond market. These firms reduce borrowing and investment more when expected refinancing conditions deteriorate. Unconstrained firms that extend maturities face smaller investment declines. To interpret these findings, we develop a dynamic model in which firms choose debt maturity by trading off the cost of long-term borrowing against exposure to future refinancing conditions. The model highlights how expectations about future rates and term premia jointly shape maturity choices and amplify the effects of monetary policy. These findings imply that monetary policy affects real activity through a refinancing-risk channel and that the term structure plays a central role in shaping the transmission of monetary policy to firms.

The economic and social consequences of international migration have been studied extensively in destination countries, yet the costs borne by transit countries remain poorly understood. This paper estimates the causal effect of transit migration on local crime victimization in Mexico, the world’s most consequential transit country for northbound migration. I exploit spatial variation in Mexico’s commercial freight rail network, colloquially known as The Beast, as a instrument for predicted migrant flows, interacting the share of national rail routes passing through each municipality with annual counts of non-North American foreign arrivals. Using individual-level data from Mexico’s National Crime Victimization Survey (ENVIPE) across 2,462 municipalities from 2011 to 2019, I estimate a two-way fixed effects model that absorbs municipality and year-level confounders. The central finding is that transit migration increases local crime victimization. My preferred specification indicates that an additional 10,000 predicted transit migrants raises overall victimization by 0.24 percentage points, a 3 percent increase relative to the sample mean. The effect is concentrated in property crime, particularly vehicle parts theft, with smaller but statistically significant increases in kidnapping and rape. The absence of a broad increase in violent crime, combined with evidence that municipalities along The Beast experience demographic shifts toward more male, lower-educated, and economically marginalized populations, suggests that migrants themselves are the primary victims of the violent crimes observed, consistent with extensive qualitative evidence on predatory violence along transit corridors. Placebo tests that randomize both the timing of migrant crossings and the spatial assignment of rail exposure produce null effects, supporting a causal interpretation of the baseline results. Heterogeneity analysis further shows that, unlike settlement migration in the United States, transit migration flows away from established co-ethnic networks, amplifying migrant vulnerability by severing access to social capital and local knowledge. These findings suggest that transit countries bear measurable and previously underestimated social costs from international migration flows, with implications for the evaluation of U.S. immigration enforcement policies that increasingly redirect migrants toward longer and more dangerous overland routes through Mexico.

Limiting global warming to 1.5 degrees requires that cumulative carbon dioxide emissions remain within a finite remaining carbon budget. How this budget is allocated across countries raises questions of fairness and development. This paper evaluates whether equity-based carbon allocations are compatible with sustained economic growth in emerging and developing economies. We compute country-level fair shares of the remaining carbon budget under the equal-cumulative-per-capita (ECPC) principle. Using data for 162 countries between 1950 and 2023, we then estimate the historical relationship between income and per-capita CO2 emissions across income groups and use these elasticities to simulate cumulative emissions until 2050. Our results show that ECPC implies strongly negative remaining carbon budgets for most advanced economies, while lower-income countries retain positive but constrained allocations. Under historically observed income–emissions elasticities, many developing countries would exceed their fair shares when converging toward advanced-economy income levels. At the aggregate level, unused allocations offset only 17% of the combined carbon budget shortfall implied by countries exceeding their allocation and the negative fair shares arising from historical responsibilities. In a scenario in which we assume that the technology of advanced economies is transferred to all countries, the carbon budget coverage increases to 38%.

In 2002, the NCAA men’s basketball tournament adjusted the way teams are assigned to game sites. This meant moving away from a system in which a team’s travel was determined solely by which of four regions it was placed in, to a “pod system” intended to minimize travel, particularly for higher-seeded teams. Using data on tournament games played between 1985 and 2026, we examine whether the betting market correctly prices the relative travel burden faced by each team, and whether the pod system’s adoption changed the market’s ability to do so. This setting differs from most prior studies of travel and betting markets in that tournament games are played at neutral sites, so that both participating teams, rather than a single visiting team, must travel to the contest. We find that travel distance had no relationship to betting market accuracy prior to 2002, but that a significant relationship emerges in the years afterward. We further find that the direction of travel, and specifically eastward travel by the lower-seeded team, carries information beyond distance alone that the market fails to incorporate. We explore whether these findings translate into profitable betting strategies.

Since the Supreme Court overturned the Professional and Amateur Sports Protection Act in 2017, 39 states, Puerto Rico, and the District of Columbia have legalized betting on professional and amateur sports. As a result, betting on sports contests has exploded, rising from $5 billion (all in Nevada) in 2015 to $170 billion in 2025. There has also been a parallel increase in the social ills associated with gambling. College players have been suspended for betting on contests in which their teams participated, and professional players have faced criminal charges for actions that assured payoffs to gamblers. Studies have also shown that households in states that have legalized gambling have lower savings, higher credit card debt, and lower credit scores. All these factors should reduce the value of marriage and, according to the Becker model of marriage and divorce, cause more marriages to dissolve. I apply Callaway-Sant’Anna estimation to data from the American Household Survey to test whether states that have legalized sport gambling see a concomitant increase in divorce rates.

Forthcoming

This study examined the spatial and structural relationship between casinos and crime using the place-in-neighborhood (PIN) framework. It investigated how macro-level neighborhood conditions moderate the spatial effects of proximity to casinos. Using multilevel negative binomial models, we analyzed violent, non-violent, and other crime across 17,357 intersections nested within dissemination areas in Edmonton. At the micro-spatial scale, casino proximity does not independently correlate with higher crime counts. The effects for violent and other crimes are statistically insignificant, and the initial 0.5% increase for non-violent crime becomes insignificant once macro-level variables are included. At the macro level, concentrated disadvantage is strongly and consistently associated with higher crime across all categories, indicating that neighborhood structural conditions play a significant role in explaining crime variation. Cross-level interaction results indicate that the spatial influence of casinos is context-dependent. Concentrated disadvantage significantly increases the risks associated with casino proximity for violent and other crimes, but not for non-violent crime. These findings suggest that the criminogenic effect of casinos is not uniform and depends on neighborhood conditions. The results highlight the importance of considering structural context in crime prevention strategies.

Forthcoming

We study the determinants of US dollar demand across market participants and traded instruments using survey-based exchange rate and macroeconomic expectations. Leveraging granular foreign exchange trading data, we show that forward-looking expectations predict both currency returns and flows. Specifically, we show that the predictability of currency returns at long-horizons can be attributed to price pressure that originates from investors whose trading activity is aligned with survey expectations. To empirically establish the relevance of survey-based expectations for currency flows, we present three results: First, end-user investors increase their dollar holdings when they expect the US dollar to appreciate, whereas dealer banks supply dollar liquidity. Second, cross-sectionally, investors rebalance along the factor structure of currency risk into the US dollar following an expected dollar appreciation. Third, the predictive power of survey forecasts weakens when uncertainty or forecaster disagreement rises. Overall, our findings demonstrate that long-horizon expectations predict dollar demand across spot, swap, and forward currency markets. To rationalise these empirical findings, we develop an intermediation-based model of currency demand.

We study how non-bank financial institutions (NBFIs) generate systematic FX flows through the rebalancing of their currency hedges. When foreign portfolio values change, NBFIs adjust forward positions to maintain target hedge ratios. Using transaction-level EMIR data on Norwegian NBFIs, we show that flow–return pass-through is close to one for fully-hedged funds, strengthens monotonically with the hedge ratio, and is absent for unhedged funds. High-frequency event studies on USDNOK reveal that 1 billion NOK (100 million USD) in forward purchases moves the spot rate by approximately 0.2% within minutes, nearly tripling in high-volatility periods. Across G10, a panel 2SLS specification with currency and time fixed effects documents significant price impact for seven of nine non-USD pairs. NBFI hedging positions have tripled since 2016, suggesting that hedge-rebalancing flows have become an increasingly important component of FX order flow.

Driving under the influence is a common crime with substantial externalities. While prior research suggests the overall menu of sanctions produces deterrence, we focus here on one particular form of sanction enhancemen produced by "wash-out windows". In many states; after a period of 3-7 years, repeat offenders are forgiven of the prior convictions and treated as a first-time offender again. In studying the state of Washington I find indeed after 7 years, the sanctions of repeat offenders drop. Moreover, I find no evidence of sorting, suggesting there is not ex ante general deterrence. Being convicted at 6 years and 11 months raises the time spent in jail by roughly 35 days, an 170 percent increase over the expected sentence length for those convicted at 7 years and 1 month. While sanctions are more severe, there is not additional ex post reduction in recidivism, suggesting the extra 5 weeks in jail does not produce notable changes in reoffending either. This suggests there may be important non linearities in how sanctions affect crime, and additional punitive measures may not produce deterrence. We also find no evidence the additional sanctions reduce other measures of alcohol related crime or traffic accidents.

We study the impact of drug decriminalization on overdoses. Two papers have received substantial academic, media, and political attention regarding the impacts of decriminalization, reaching opposing conclusions despite using identical methods and near-identical data sources. To this end, we replicate and extend their approaches. We first document that slight differences in outcomes and pre-treatment window create different counterfactual weights, which led to different conclusions regarding the first-year impacts. We next document that with an extended time window, the differences in the methodologies between the two papers do not matter, and we find robust evidence that drug overdose, particularly after the second year, increased significantly by 0.8 per 100,000. Finally, we conduct additional analyses to investigate whether the arrival of the fentanyl crisis on the West Coast is driving the increase in overdoses.

“Point-in-time” counts of those experiencing unsheltered homelessness in the United States are at their highest levels in over a decade. In response, some cities have undertaken “encampment closures” in which individuals residing at an encampment site are made offers for shelter, but, regardless of take-up, must evacuate within a short period of time. Despite becoming a more widespread practice, little has been documented about the effects of these closures. I create a novel dataset of encampment locations and closures from 2017 to 2020 in Seattle, WA, construct a difference-in-differences approach with a matched comparison group designed to avoid spatial spillovers, and use administrative data (crime, emergency calls for service, complaints) as well as mobile-phone mobility data (foot traffic) to estimate their impact. Encampment closures do not prevent future encampments from being formed in the same location: about 46 percent of newer encampment closures intersect with the boundaries of at least one previously closed site, and the average time between these “repeats” is 121 days. While closures do seem to lead to a slight decrease in complaints in the short term, estimates for crime, 911 calls, and foot traffic in the areas surrounding encampments are relatively precise and cannot reject zero.

The SAFETY Act is a US law that attempts to create incentives encouraging anti-terrorism technologies by providing firms with third-party liability protections. Liability protections generally increase the extent these technologies are deployed. However, the possibility of a moral hazard creates ambiguity as to the effect liability protections have on the quality of anti terrorism technologies. I explore different incentives, in different scenarios, and in what cases firms have incentive to improve the quality of their anti-terrorism technologies.

Lead ammunition creates health risks to shooters and instructors at firing ranges because of the toxicity of lead and also creates environmental risks and cleanup costs. U.S. government agencies have explored using completely lead-free ammunition, called reduced-hazard training ammunition (RHTA), as an alternative. The authors of this report assess the benefits and costs of lead-free RHTA and lead training ammunition based on current market conditions and firing range maintenance practices, to inform decisions by law enforcement organizations that manage firing ranges or are considering investments in these facilities. The authors also provide strategies that law enforcement agencies can use to ensure reliable supplies and lower prices of training ammunition.

Hospital reorganization through specialty separation has emerged as a healthcare optimization strategy, yet its differential effects across specialties remain unclear. This natural experiment compared the impacts of hospital separation on emergency department (ED) quality between pediatric and obstetrics-gynecology (OB-Gyn) services. We analyzed 192,662 ED visits at Hsinchu Mackay Memorial Hospital (HMMH) and Hsinchu Municipal MacKay Children's Hospital (HMMCH), Taiwan, from November 2021 to October 2023, both before and after the establishment of a specialized children's hospital (HMMCH). Using difference-in-differences (DID) regression, pediatric (n=64,970) and OB-Gyn services (n=6,416) were compared against Medicine, Surgery, and Other specialties (MSO) as controls, evaluating waiting time, resident time, and observation time while adjusting for demographics, triage level, shift, and clinical characteristics. Post-separation, pediatric visits increased 74.3% (22,510 to 39,226) and OB-Gyn visits 41.0% (2,410 to 3,398). The hospital separation effect showed significant overall improvements: waiting time decreased by 0.09 hours (5.4 minutes, p<0.001), resident time decreased by 0.18 hours (10.8 minutes, p<0.001), and observation time decreased by 0.25 hours (15.0 minutes, p<0.001). However, the DID policy effect revealed striking heterogeneity between specialties. For waiting time, both showed similar relative increases versus controls, amounting to approximately 19 additional minutes, whereas for resident time, increases were attenuated. Critically, observation time effects diverged: pediatrics showed a significant relative increase, while OB-Gyn showed no significant change. Notably, OB-Gyn absolute observation time decreased 55% (0.84 to 0.38 hours), suggesting improved output-phase efficiency. These findings demonstrate that hospital separation produces specialty-specific effects: while both specialties experienced increased input-phase burden from volume growth, output-phase efficiency improved only for OB-Gyn. This heterogeneity implies that patient acuity patterns, resource requirements, and care complexity critically shape reorganization outcomes, warranting specialty-tailored planning in healthcare system transformations.

Goal: This study investigates the impact of Group Purchasing Organization (GPO) adoption on hospital supply cost efficiency in the United States. While GPOs are widely promoted as a cost containment strategy, empirical evidence on their causal effect, particularly at the point of initial adoption, remains limited. This research addresses that gap by evaluating whether hospitals adopting GPOs achieve greater improvements in supply cost efficiency compared to non-adopters. Methods: We used a nationally representative panel dataset from the American Hospital Association, covering general acute care hospitals from 2010 to 2019. The resulting study sample was an unbalanced panel of hospitals with 28,670 unique hospital-year observations. A generalized difference-in-differences event study design was applied to account for staggered adoption timing and treatment heterogeneity. Supply cost efficiency was measured using two ratios: (1) supply expenses as a share of total operating expenses and (2) supply expenses as a share of net patient revenue, capturing non-labor supply chain costs. Models controlled for hospital financial variables, organizational characteristics, and fixed effects. Principal Findings: GPO adoption was associated with statistically significant and sustained reductions in supply cost ratios. In the adoption year, the supply expense-to-total operating expense ratio decreased by 0.63%, with further declines to 0.73% two years post-adoption and 0.70% in the third year. Similar patterns were observed for the supply expense-to-net patient revenue ratio. No evidence of differential pre-trends was found, supporting the validity of the parallel trends assumption. Effects were more pronounced in system-affiliated and investor-owned hospitals. Practical Applications: The results suggest that GPO participation can deliver meaningful and enduring cost savings for hospitals, particularly for those with greater administrative and bargaining capacity. Policymakers and healthcare administrators may consider promoting GPO access, especially for under-resourced or rural hospitals, while also addressing concerns over procurement flexibility, vendor diversity, and potential impacts on care quality. Future research should incorporate quality-of-care measures and analyze differences across GPO types to inform optimal procurement strategies.

Offshore islands and other geographically isolated regions face persistent barriers to timely cardiovascular care, often requiring emergency providers to rely on costly, time- sensitive emergency transfers. This study examines how the introduction of a cardiac catheterization laboratory (Cath Lab) altered emergency cardiovascular care delivery in such a setting through three complementary empirical analyses spanning policy evalua- tion, patient-level treatment allocation, and aggregate operational dynamics assessment. Using 240 months of administrative records from 2004 to 2023 covering seven major diagnostic groups, the Difference-in-Differences Poisson framework identifies a substan- tial reduction in cardiac-related emergency transfers following the 2015 Cath Lab imple- mentation. Transfer volumes declined sharply after implementation and remained persis- tently lower in subsequent years, with reductions of approximately two-thirds in both the immediate and longer-term post-policy periods. Event-study estimates support the parallel- trends assumption, and the findings remain robust across alternative temporal aggregations. Patient-level Firth logistic estimates indicate that greater local treatment availability significantly reduced the likelihood of emergency transfer, while post-policy Poisson anal- yses show that expanded procedural activity and sustained cardiologist availability were associated with evolving referral patterns over time. These patterns are consistent with a process of local capacity expansion in which strengthened specialist resources reduced reliance on emergency transfers while preserving referral for clinically complex cases. The evidence indicates that targeted investment in specialized medical infrastructure can substantially alter emergency care delivery in geographically isolated health systems. The findings further suggest that infrastructure expansion is most effective when accom- panied by sustained specialist staffing and coordinated referral capacity, with practical im- plications for health system planning in remote and resource-constrained settings.

The integration of artificial intelligence (AI) into nursing homes has the potential to significantly enhance care quality, operational efficiency, and health outcomes for elderly residents in the United States. While AI applications have been increasingly adopted across various healthcare settings, empirical research examining their use, acceptance, and effectiveness in nursing homes remains limited. This study aims to conduct a comprehensive analysis of AI integration in U.S. nursing homes using a national survey framework to identify key drivers, barriers, and outcomes associated with AI adoption in long-term care facilities. The study employs a mixed-methods research design that combines quantitative and qualitative approaches. Quantitative data will be collected through a structured national survey targeting nursing homes and elderly residents, capturing perspectives on AI awareness, adoption willingness, anticipated applications, and perceived benefits and concerns. Survey measures will examine AI use across clinical care, chronic disease management, facility operations, and administrative decision-making. Statistical and econometric analyses will be applied to evaluate factors influencing AI adoption and to assess the relationship between AI utilization and resident health outcomes, including perceived well-being and care efficiency. This further analysis is expected to generate empirically grounded evidence on the current state and future trajectory of AI integration in U.S. nursing homes. By identifying critical determinants of successful adoption and evaluating the potential impacts of AI on care delivery and resident outcomes, the study aims to inform policymakers, healthcare administrators, and industry leaders. The findings will contribute to the development of evidence-based strategies to optimize the AI service ecosystem in nursing homes, addressing issues of adoption willingness, affordability, and sustainable implementation. Ultimately, this research seeks to support the responsible and effective use of AI technologies to improve long-term care for the aging population in the United States.

This paper examines how the 2025 U.S. “reciprocal tariffs” are priced in Taiwanese equity markets, focusing on export-oriented industries for which the United States is a primary destination. Using daily stock returns for Taiwan-listed firms in ten U.S.-oriented export industries, we implement an event study around the April 2, 2025 tariff announcement (with the first subsequent trading day response) and related policy updates. We estimate abnormal returns and cumulative abnormal returns (CARs) using a standard market-model framework. To explain heterogeneity in valuation effects, we run parsimonious cross-sectional regressions linking firm-level CARs to pre-event U.S. export exposure (US Exposure) and China export exposure (China Exposure), controlling for firm size, pre-event profitability, and industry fixed effects. The results indicate significantly negative market reactions around the tariff episode, with larger valuation losses among firms more dependent on the U.S. market. We further assess real-outcome consistency by examining post-event revenue momentum for a subset of firms using monthly revenue disclosures. The findings highlight export concentration as a key channel through which U.S. tariff policy shocks transmit to firm valuation in a small open economy, and underscore the importance of balanced export diversification and risk management under heightened trade-policy uncertainty.

China’s rapid emergence as the world’s leading producer and exporter of electric vehicles (EVs) has intensified trade tensions with the United States and the European Union, both of which have imposed tariffs in response to concerns over unfair competition. This paper examines whether these protectionist responses are justified by reviewing the literature on China’s EV industry, the sources of its competitiveness, and the economic arguments for and against trade intervention. The analysis finds that China’s rise was driven by a combination of extensive state support and market-based advantages. Government subsidies, industrial policies, infrastructure investment, and strategic supply-chain development played an important role in fostering the industry. At the same time, Chinese firms developed significant competitive advantages through economies of scale, vertical integration, technological innovation, and leadership in battery manufacturing and critical-mineral processing. The paper concludes that Western concerns regarding subsidies and market distortions are partially justified, but that the effectiveness of policy responses depends on their design. Targeted measures aimed at addressing specific distortions may be warranted, whereas broad and permanent trade barriers are unlikely to eliminate competitive advantages rooted in productivity, innovation, and industrial capabilities. Strengthening domestic competitiveness and supply-chain resilience may therefore offer a more effective long-term strategy than reliance on protectionism alone.

This study investigates whether the EU’s increasingly stringent carbon reduction policies were genuinely designed to address carbon leakage or if they inadvertently weakened domestic industrial competitiveness, leading to greater reliance on imports and ultimately contributing to the introduction of CBAM. As the EU imposed stricter environmental regulations, domestic producers faced higher carbon costs, making them less competitive compared to manufacturers in countries with weaker carbon policies. This study examines how such disparities have influenced trade flows, particularly in terms of trade diversion and trade creation, reshaping global trade dynamics beyond the stated environmental objectives. While most prior studies have assessed the trade effects of carbon policies using ex-ante models (e.g., CGE), this research employs an ex-post analysis to capture indirect effects. Using the Gravity Model, it analyzes how carbon reduction regulations have shaped international trade patterns and supply chain shifts. By observing variations in trade flows of CBAM-regulated products, the study provides empirical insights into whether these policies have contributed to trade restructuring and industrial protectionism. By adopting an ex-post approach, this study offers a new perspective on the broader economic implications of carbon reduction policies. Unlike previous research that primarily relies on theoretical simulations, it presents empirical evidence on trade adjustments triggered by environmental regulations. The findings contribute to a deeper understanding of how climate policies intersect with industrial competitiveness and global trade, shedding light on whether such measures serve dual purposes of environmental protection and economic self-interest.

Although most countries expand both fixed broadband and mobile cellular connectivity over time, the two technologies differ in their diffusion patterns and timing. This paper examines whether the relationship between internet connectivity and global value chain (GVC) participation varies across different types of connectivity and whether these associations persist over time. Information on GVC participation is drawn from the UNCTAD-Eora Global Value Chain Database. Internet connectivity types are measured using the fixed broadband and mobile cellular subscriptions, sourced from the International Telecommunication Union. The sample comprises an unbalanced panel of 930 observations covering 81 countries from 2002 to 2018. Of these countries, 45% are in Europe and 20% are in the America, followed by Asian (18%), Africa (13%), and Oceania (4%). High-income countries account for 65% of the sample, with the remaining 35% of middle- or low-income countries. To capture potential lagged relations, one-, three-, five-, and seven-year lags of internet connectivity variables are examined. In the fixed-effects model, both fixed broadband and mobile cellular subscriptions with a one-year lag are positively associated with GVC participation in the sample. However, persistence differs across income groups and connectivity types. In high-income countries, the association with fixed broadband disappears after one year, whereas mobile connectivity remains significant at three- and five-year lags, indicating greater persistence. In contrast, for low- and middle-income countries, significant associations are largely confined to the one-year lag, suggesting more limited persistence over time compared to high-income countries. To account for connection quality, fixed broadband bandwidth is considered separately. In high-income countries, higher fixed broadband bandwidth levels are positively associated with GVC participation only at a three-year period, with no persistence at shorter or longer lags. This pattern suggests that the association reflects existing infrastructure capacity in economies with already saturated connections. In contrast, for low- and middle-income countries, bandwidth is positively associated with GVC participation only at the one-year lag, with no evidence of persistence beyond this period. Considering both patterns together demonstrates that the role of broadband bandwidth in GVC participation varies across income groups, with more limited and short-term associations observed in low- and middle-income countries.

Undocumented immigrants are disproportationly likely to work in construction and they make up a disproportionate share in the construction labor force (Svajlenka, 2021). Little is know how US citizen workers of different racial groups respond to the removal of undocumented immigrants. Using American Community Survey (ACS) and the staggered county-by-county rollout of the deportation program Secure Communities (SC) between 2008 and 2013, the study analyzes the employment response of non-Hispanic Black/White US citizen workers to the removal of undocumented workers in the construction industry. Using staggered Callaway Sant’Anna difference-in-difference and controlling for county pre-treatment charactistics, the study does not find significant increase of the employment of either Black/White US citizen workers, despite a significant overall decrease in low-skilled construction labor supply. SC was found to lead to a significant decrease of Hispanic immigrant workers in construction, however, the effect seemed to be most prominent in the early adopter counties compared to the late adopter counties. The paper further suggests that US workers did not fully fulfill the job vacancies caused by the removal of undocumented immigrants in low-skilled construction.

This paper examines persistent racial disparities in drowning mortality in the United States, with a particular focus on American Indian and Alaska Native (AI/AN) populations. The central research question asks: To what extent have disparities in drowning mortality between AI/AN populations and other racial groups (White, Black, and Asian) persisted or changed over time, despite overall national declines in drowning rates? While prior research has established that AI/AN individuals experience the highest drowning mortality rates in the United States, less is known about how these disparities have evolved in recent decades and across geographic contexts. The paper contributes to the literature by providing a systematic, longitudinal comparison of drowning mortality rates across racial groups using nationally representative mortality data. It also extends existing work by examining state-level heterogeneity, particularly in Alaska and Hawaii, where environmental exposure and population composition may produce distinct patterns. Methodologically, the analysis employs a descriptive patterns-and-trends approach, calculating crude death rates (CDRs) for drowning (deaths per 100,000 population) by race, age, sex, and state. Mortality data are drawn from the United States restricted-use mortality files, while population denominators are obtained through the CDC WONDER system. Drowning deaths are identified using ICD-10 codes associated with unintentional and undetermined drowning. The study period spans 2003 to 2020, allowing for the examination of both long-term trends and more recent changes in disparity. The empirical strategy focuses on comparing CDRs across racial groups over time, with particular attention to divergence or convergence in rates. Subgroup analyses explore variation by age and sex to identify populations at elevated risk. Additionally, Alaska and Hawaii are treated as analytically distinct cases to assess how geographic and environmental factors intersect with racial disparities. Findings from this study will clarify whether national declines in drowning mortality have been equitably distributed and will identify populations and regions where disparities remain most pronounced. These results have implications for targeted public health interventions, resource allocation, and culturally responsive drowning prevention strategies in AI/AN communities.

The contemporary military establishment has developed historically into a central institution in modern nation-states, serving to protect private property and reinforce patriarchal hierarchy. Despite efforts to portray the military as gender-neutral, feminist researchers like Hudson et al. (2020)—building on Enloe—highlight how it is characterized by patriarchal practices. This paper draws on the feminist political economy lens to highlight militarization as a process rooted within patriarchy. Using existing literature, our conceptual framework proposes two hypotheses: (1) militarization lowers female share of employment, and (2) it restructures the gendered composition of employment within the care sector. Our results show that higher militarization is associated with a decline in female share of employment for feminized work, including specific types of care workers. These findings challenge the conventional norm of studying militarization within defense literature alone by showcasing the association militarization has with the civilian labor markets, especially the care sector and provide an initial empirical contribution to the ongoing discussion around militarization.

Artificial intelligence (AI) has become embedded in daily decision-making, information search, and the dissemination of scientific knowledge. Using incentivized large-scale economics experiments, I characterize trust in AI across three domains: AI as an advisor, AI as a source of information, and AI as an assistant for scientific writing. We find that relying on low-quality AI advisors without disclosures reduces individuals’ truth-detection rates below their inherent abilities, which are recovered once the AI’s true effectiveness is revealed. Using additional evidence from a large-scale cross-country global survey data, we discuss how trust in AI as an information source is strongly associated with trust in traditional information institutions (press, television, and news organizations), but not with generalized social trust or broad attitudes toward AI. In the third arm, we find that AI-generated abstracts hold potential to reach comparable levels of acceptability to original human-written abstracts with minimal revision, and that perceptions of AI authorship, rather than objective quality, drive much of the observed editing behavior. Overall, I deduce that trust in AI is a behavioral response shaped by disclosure, institutional environments, and perceptions of legitimacy, which collectively influence how the benefits and risks of AI are distributed across populations.

This paper examines the international variation in the prices of branded pharmaceuticals. We consider short-run profits, or quasi-rents, as representing each country’s contribution toward the global public good of therapeutic information embodied in new pharmaceuticals. We characterize globally optimal contributions through the Samuelson-Lindahl criteria and contrast the resulting outcomes with the Nash noncooperative equilibrium as developed in the Olson-Zeckhauser theory of international alliances. That theory predicts both the undersupply of public goods and the “exploitation” of large countries by small ones. We calculate national contributions to the global public good with data from a recent RAND report. Other countries’ contributions resulting from their prices are much lower than those in the United States, but their free riding is not complete. We also find that country size is a powerful determinant of contributions and that larger countries contribute disproportionately more. Finally, we suggest a cooperative policy approach that would move us closer to optimality with health benefits for all countries.

Abstract: Cheap Riders on the Public Goods Express Ted Bergstrom In 1965, Mancur Olson argued that there is ”exploitation of the great by the small” in the provision of global public goods. Where countries pursue their self- interest, but share public goods, such as defense, climate-control, or research, the outcome predicted by Nash equilibrium is one in which large, rich countries contribute a larger share of their national income to the public goods than smaller countries. In a 1966 article, Olson and Richard Zeckhauser illustrate this effect with the case of the NATO military alliance, where the largest, richest country, USA, spends more than twice as large a fraction of GNP on its military as do smaller NATO countries. Expenditures on global public goods fall far short of Pareto efficiency, while a Lindahl mechanism would result in Pareto efficient provision. But, as Zeck- hauser and Cuicui Chen observe, if countries in an alliance differ greatly in size, smaller countries will not willingly participate in a Lindahl mechanism, since they would prefer Nash equilibrium to Lindahl equilibrium. These authors suggest an alternative mechanism, which they call the ”Cheap-riding efficient equilibrium mechanism” which leads to a Pareto efficient outcome and which is preferred by all alliance members to Nash equilibrium. This paper presents a series of examples in which it compares Nash equilibrium to the cheap-riding equilibrium and explores other possibilities for international agreements that are superior to Nash equilibrium

This paper examines international coordination in pharmaceutical pricing and its implications for innovation incentives. Motivated by large cross-country differences in drug prices and renewed policy interest in burden-sharing arrangements, the study asks whether coordinated pricing policies can address free-riding while remaining stable among heterogeneous countries. I adopt a multi-country structural model of pharmaceutical markets linking country-specific price caps to firm pricing, consumer demand, and dynamic R&D investment. Countries choose pricing policies to maximize domestic welfare, generating cross-border spillovers through innovation. The model is calibrated using data on drug prices, market sizes, and branded sales shares across high-income and emerging economies. We evaluate counterfactual pricing coordination regimes, their stability, and their welfare implications.

In recent research, my colleagues and I explored the underlying economics of the global public good resulting from industry research and development. In this presentation, I describe the market setting in which innovative pharmaceuticals are introduced along with the underlying value of the therapeutic gains achieved. From this discussion, we draw important implications for the current debate over the prices charged for innovative pharmaceuticals

Urban spatial evolution directly impacts people’s lives. Since 2008, more people have lived in urban areas than in rural ones, highlighting the importance of understanding how cities develop. This paper investigates the trajectory of urban spatial development across 200 years in the United States and the impact of historical urban spatial structure on short- and long-term economic outcomes. Using a novel dataset constructed through the City Clustering Algorithm (CCA), this paper redefines historical urban boundaries, creating “synthetic cities” (Syncities) that more accurately reflect urban development from 1810 to 2010. Using instrumental variable regression, this paper identifies how the size and shape of cities influenced income performance in both the short and long terms. Specifically, we instrument for urban spatial structure using the predicted urban footprint. The findings highlight the long-term economic growth linked to urban expansion, although the short-term effects are not statistically significant. Cities with larger and more dispersed urban forms benefited economically in the long term. By introducing a novel approach to delineating historical cities and constructing a new spatial dataset for urban areas, this paper provides new insights into long-term urban development patterns and links the historical urban spatial structure to economic performance.

This study examines how digital literacy influences voter turnout in South Korea using National Assembly election data and the Digital Divide Survey across 17 provinces. Multilevel regression analysis shows that digital access has no significant overall effect, while digital capability strongly increases turnout and digital utilization reduces it by diverting attention to non-political activities. Income is the strongest individual predictor, and regional GDP also matters. Age-specific results reveal that access discourages turnout among younger voters but benefits older ones, and capability peaks among middle-aged groups. These findings highlight the importance of enhancing digital skills rather than focusing solely on digital access in promoting political participation.

This study investigates the role political regimes play in shaping the relationship between labor productivity and economic development, drawing on a unique historical dataset spanning 20 OECD countries from 1870 to 2011. Utilizing robust empirical methodologies, including fixed effects and Pooled Mean Group estimators, we identify distinct institutional mechanisms through which democracies outperform autocracies in converting productivity gains into sustainable, inclusive growth and macroeconomic stability. Our analysis makes two critical contributions: first, democracies significantly enhance the impact of productivity gains, leading to sustained growth and stability. Second, the interactive effects of labor productivity and education, while independently beneficial, yield diminishing returns under democratic governance, reflecting the redistributive pressures inherent in democracies. Contrary to assertions favoring autocratic efficiency, our findings underscore democracy’s substantial advantage in equitably distributing productivity gains and promoting stability. Thus, the research provides novel empirical evidence highlighting democracy's crucial role in maximizing the developmental impact of productivity advancements.

Where more locals staffed courts, police, and administration, Ukrainian communities collect more tax and govern more effectively today. Regions entered the Russian Empire with different traditions of self-rule, leaving variation in who held local office. Using newly transcribed 1897 census records, I measure the share of local officials and link it to modern communities. A 10-percentage-point increase in local representation is associated with 7.1% higher local tax revenue per capita in 2021. An IV strategy supports the results, which survive a battery of robustness checks. I argue that self-governance built durable local institutions that persisted across imperial and Soviet rule.

Does welfare retrenchment cause social unrest? I study this question by using parish-level expenditure on poor relief from post-Napoleonic England and applying generative AI to a corpus of 3.9 million newspaper column transcriptions to collect the category, date and location for each reported unrest incident. Reduced-form analysis leveraging a shift-share instrument finds a causal effect of welfare retrenchment on social unrest between 1819 and 1831, attributable to grievances arising from social and habitual comparisons. Simulations from a counterfactual analysis assuming blanket increases in welfare expenditure highlight the significance of social welfare provision in preventing large-scale unrest.

Forthcoming

Forthcoming

This paper estimates the dynamic trade effects of regional trade agreements (RTAs) using CEPII’s BACI bilateral trade data combined with the World Bank’s DTA 1.0 depth measures for 1995–2023. We implement Wooldridge (2025)’s extended two-way fixed effects (ETWFE) event-study for staggered treatment and compute all averages over cohorts that pass parallel-trends tests. Baseline results indicate sizable and persistent gains: RTAs raise bilateral trade by about 57% on average, with effects that build over time. Event-time estimates show the trade-creating impact rises from roughly 21% in the entry year, to about 85% by year 12, and exceeds 100% after year 19, suggesting long-horizon accumulation and durability. Cohort-specific ATTs reveal substantial cross-vintage heterogeneity: agreements implemented earlier deliver larger and more persistent trade gains than those adopted later. Calendar-year ATTs likewise indicate stronger trade creation in earlier periods. Heterogeneity by policy area and depth shows clear attenuation. Interaction estimates indicate that deeper coverage in WTO+ areas tends to reduce the marginal trade effect of an RTA, while additional WTOX depth does not consistently offset this attenuation.

This paper examines how inflation targeting offers an alternative to foreign exchange interventions for attracting debt inflows during economic crises in developing countries. Developing countries have traditionally relied on accumulating foreign exchange reserves to preserve economic stability and reduce exposure to external shocks. However, reserve accumulation carries a significant opportunity cost. We therefore investigate whether adopting inflation targeting can reduce the need for reserve accumulation while still sustaining or increasing debt inflows during crises. Drawing on capital flow data from 65 developing countries over the period 1990–2020, we estimate the differential effect of inflation targeting adoption while accounting for endogeneity and reserve accumulation. Our results indicate that reserve accumulation is associated with greater total debt inflows during crises in non-targeting countries. By contrast, targeting countries tend to experience rising debt inflows during crises even as reserves decline. Our results suggest that inflation targeting can substitute for reserve accumulation as a mechanism for attracting debt inflows during crisis episodes, particularly for private sector flows, without incurring the opportunity costs associated with holding reserves. We rationalize these results through a model of corporate intermediated borrowing that can explain the substitutability between reserves and inflation-targeting, as they both decrease financial intermediaries’ loan origination costs. The former by changing the leverage ratio of financial intermediaries; the latter by decreasing the expected rate of depreciation. Our findings highlight an additional benefit of inflation targeting for developing economies.

When a child dies, what happens to the schooling of the siblings who survive? The answer is not obvious: a death leaves fewer children to share parental resources, which may raise investment in each survivor, but it also disrupts the household, which may lower it. We model the net effect of a sibling death as a replacement-weighted quantity–quality reallocation gain minus a disruption cost, implying that the effect should be more positive where replacement was more constrained. Using the China Health and Retirement Longitudinal Study linked to a prefecture-year disaster panel, we will estimate the net effect of sibling-loss exposure on children’s educational attainment, and use heterogeneity in fertility constraints during China’s One-Child Policy era to test for the quantity–quality reallocation channel.

In this paper, we explore the determinants of sudden stops of international capital flows using interpretable machine learning models. That is, we focus on models that are not highly interpretable like ordinary least squares nor highly flexible like neural networks. Specifically, we use decision trees and related models. The benefit of this approach is that it can handle non-linearities such as threshold effects and interactions effects more readily than traditional econometrics methods. We focus our attention on sudden stops for emerging market economies since the consequences are more severe than for developed economies.

Although the stringency of conditionality is central to debates about the programs supported by the International Monetary Fund (IMF), most conventional measures of conditionality focus on the extent and scope of the conditions stipulated rather than the overall adjustment effort required in meeting them. In this sense, they provide only a partial guide. In this paper we attempt to complement existing approaches by quantifying the difference between the targets set in programs and the status quo prevailing at the time that they start. To do this, we construct a data set based on the Quantitative Performance Criteria (QPCs) included in IMF programs. Our data cover 981 QPCs, 160 arrangements and 55 countries over the period 2000-2025. We then devise and calculate a new measure of stringency, the Relative Adjustment Requirement (RAR). We find that while measures based on the number of conditions and their scope are closely correlated, our new measure of the stringency of conditionality is only weakly correlated with the other measures. We also discover that only a relatively small proportion of conditions (about 21 per cent) and only in a relatively small number of areas require adjustment. As many as a third of all IMF programs do not include any conditions that require adjustment. We then explore the implications of this for comparing IMF conditionality both across individual country programs and over time.

Natural resource management traditionally assumes stationary systems where historical relationships reliably predict future outcomes. This assumption is increasingly untenable as marine ecosystems experience non-stationary dynamics driven by ocean warming, acidification, and climate regime shifts. Delayed detection of such shifts forces managers to make harvest decisions based on outdated productivity assumptions, risking overharvesting in declining regimes or forgone economic opportunities in improving ones. While previous research has examined fishery management under various forms of uncertainty and the economics of regime shifts, these studies generally treat uncertainty sources in isolation, assume stationary productivity, or focus on deriving theoretically optimal policies rather than evaluating implementable rules under realistic monitoring and institutional constraints. This paper evaluates the performance of simple harvest control rules under realistic constraints on regime shift detection and management response, and examines how management choice affects the frequency and severity of economic disasters in resource-dependent industries. We use Prince William Sound pink salmon as a case study, leveraging a documented regime shift following ocean temperature changes in 1988–1989 that differentially altered productivity in the species' genetically distinct odd- and even-year broodlines. We compare three harvest control rules: maximum sustainable yield escapement, deterministic economic escapement assuming no regime changes, and stochastic economic escapement that anticipates potential regime shifts. Our analysis explicitly models time-lagged management responses simulating delays between when productivity regimes change and when managers incorporate new information into harvest decisions, arising from data collection requirements, institutional review processes, or reduced monitoring capacity. We link biological and economic performance to disaster declarations using federal criteria established under the Magnuson-Stevens Act as amended by the Fishery Resource Disasters Improvement Act. Results provide insights into fundamental tradeoffs between economic efficiency and robustness in non-stationary resource systems, with direct implications for the choice of management approaches under monitoring constraints and the design of disaster relief policy for resource-dependent communities.

Nonlinear equations of motion are central to many problems in resource economics and public health, such as fish reproduction and disease transmission. Because policy insights fundamentally hinge on these nonlinearities, standard linear approximations are inadequate. Furthermore, strategic interactions between jurisdictions compounds the computational burden of estimating these models. We develop a tractable Generalized Method of Moments (GMM) estimation framework for dynamic games that accommodates heterogeneous agents alongside fully nonlinear, interactive state dynamics. We provide sufficient conditions under which our approach is feasible and establish its consistency and asymptotic normality. Lastly, we apply our approach to a multi-country fishery game in East Asia and quantify strategic harvest responses to climate-driven shifts in ocean currents.

Traditional sectoral income measures such as Gross Domestic Product (GDP) and Net National Product (NNP) often omit natural capital depreciation, which can lead to systematic overstatements of economic performance in resource-dependent sectors. In response, the income and wealth accounting literature has emphasized wealth-based indicators as a more informative basis for assessing long-term economic performance. This study extends sectoral income and wealth accounting to a renewable-resource sector by adjusting GDP and NNP for changes in the estimated value of fishery assets and estimating sectoral wealth for the Peruvian marine extractive fishery over the period 1995–2022. The results show that conventional and green income measures, as well as total and per capita wealth, exhibit sustained upward trends over the study period. Physical indicators of the underlying natural asset -proxied by anchoveta biomass- are used to contextualize these monetary estimates under conditions of biological and environmental uncertainty. The findings illustrate how wealth-based measures can complement conventional income indicators in evaluating the economic performance of renewable-resource sectors and highlight the importance of integrating physical asset dynamics into sectoral income and wealth assessments.

Forthcoming

Forthcoming

This paper investigates the effect of hosting regular season Major League Baseball (MLB) games on emergency medical service (EMS) response times in San Francisco during 2024 and 2025. MLB games draw large crowds, causing traffic congestion, which may hinder EMS providers’ ability to reach patients quickly. Using a ring-based difference-in-differences identification strategy, I find that medical incidents that occurred within a quarter mile of Oracle Park experienced a 51.6\% delay if they occurred within an hour prior to the beginning of a San Francisco Giants game, and experienced a 28.4\% delay if they occurred within an hour after the end of the Giants game. Accounting for attendance heterogeneity, I find that, given an additional 1,000 spectators, medical incidents within a quarter mile Oracle Park experienced a 1.2\% delay if they occurred within an hour prior to the beginning of a Giants game, and experienced a 0.6\% delay if they occurred an hour after the conclusion of the game. These findings shed light on a previously under-examined disamenity of sporting events--EMS delays--which is particularly relevant in an era where large government subsidies are allocated to new professional sports stadiums.

Forthcoming

This study examines how competitive circumstances and physiological strain influence performance in collegiate soccer. Using a player-minute panel dataset from a female collegiate soccer team, we combine minute-level cardiovascular data with detailed match-event records to evaluate how match status and physiological conditions are associated with the probability of committing performance mistakes. The results indicate that teams holding a lead commit fewer mistakes than teams that are tied or trailing, with the largest effects concentrated in possession retention and missed shots. In contrast, trailing is not associated with a general increase in execution errors relative to a tied score, suggesting that adverse match circumstances do not systematically deteriorate performance quality. At the physiological level, elevated cardiovascular strain is associated with a higher probability of possession-related mistakes, while greater heart rate variability (HRV) is associated with improved execution quality. Both physiological relationships exhibit substantial temporal heterogeneity, with their marginal effects diminishing as matches progress. Overall, the findings suggest that performance in collegiate soccer is shaped by both external competitive incentives and internal physiological conditions, and that the influence of these factors evolves over the course of competition.

Unlike nonbanks, banks have an option to borrow liquidity from public entities – chiefly central banks – by pledging loans. The value of this option is hard to quantify because liquidity usage is endogenous and can undervalue the benefit of unused access. We address this challenge by using variation in the pledgeability of similar loans in the same area and period due to disparities between regional Federal Home Loan Banks’ collateral policies. We show that pledgeable loans are 10–25% more likely to be approved. Pledgeability is more valuable when loans cannot be originated-to-distribute (OTD), and reduces bank’s use of OTD. Pledgeability gives banks a competitive advantage over nonbanks, and increases banks’ risk tolerance. These results reveal a novel originate-to-pledge motive for lending that affects credit supply, risk-taking and market structure even outside crises.

Forthcoming

This paper studies the long-run effects of macroprudential policies on the resilience of the real sector to macroeconomic shocks. We use credit registry data from Turkey, exploiting the adoption of a new macroprudential policy regime starting in 2011 and the currency depreciation of 2018 as an empirical laboratory. Higher bank exposure to household-targeting macroprudential policies is associated with a reallocation of credit toward firms, especially in local currency and to higher-rated borrowers. This reallocation is stronger for exposed banks with higher shares of FX funding, consistent with a market-discipline mechanism. Macroprudential policies are associated with greater resilience to macroeconomic shocks: firms and provinces served by banks that engage in more prudent lending following the 2011 reforms are less affected after the 2018 depreciation, exhibiting smaller relative declines in investment, electricity consumption, export growth, and firm survival. Our results suggest that macroprudential policies do not merely ``lean against the wind'' in the short run, but can shift bank risk-taking in ways that may bolster real-sector resilience to macroeconomic shocks in the long run.

Platforms increasingly partner with banks to finance users, but partnership structures differ across platforms and over time. We model a capital-constrained, informed platform that partners with a capital-rich, uninformed bank, choosing between intermediation (on-lending bank funds) and information provision (communicating signals). The tension is that the platform internalises network externalities from on-platform trade but the bank does not, and the degree of this conflict is endogenous to the partnership choice. Our main result is a regime switch in borrower pool size: information provision is optimal for small pools, intermediation for large. This threshold increases with communication credibility and decreases with merchant investment costs and production risk. BNPL buyer financing is more complementary to information provision in merchant financing.

Forthcoming

To support credit-constrained small businesses, governments use loan guarantee programs that insure lenders against default risk. However, these programs face challenges in allocating appropriate loan sizes due to limited information about borrowers. I develop and estimate a two-stage screening model that combines two common tools in lending—a menu of contracts and soft information collection (e.g., interviews and site visits)—to mitigate borrower information asymmetry. Firms sort themselves by risk: low-risk borrowers accept more intensive soft information collection to obtain larger loans, while higher-risk borrowers choose smaller, fully guaranteed loans to avoid such scrutiny. In this framework, the intensity of information collection becomes part of the contract menu, encouraging risk-based self-selection and improving the agency’s ability to allocate loan sizes. Using detailed administrative data from South Korea's loan guarantee program, I estimate the model and evaluate welfare through counterfactual analysis. The menu's welfare gain more than doubles—from 3.9% to 8.7%—when combined with soft information collection. These results provide the first evidence that contract menus and soft information are complementary screening mechanisms, with implications for the design of lending markets beyond guarantee programs.

Existing literature demonstrates that peer age plays a crucial role in shaping student outcomes in elementary school. I build on this literature by highlighting that differences in developmental rates between boys and girls likely lead to varying theoretical effects of peer age based on gender. Using administrative data on public school students in North Carolina, I exploit cross-cohort, within-school variation in third-grade peer age to examine these effects. I find that both male and female peer age positively influence student test scores, but male peer age has an impact that is twice as large as female peer age. Further analysis reveals substantial differences in these effects based on students’ gender, age, and economic disadvantage. These findings underscore the importance of separately considering male and female peer age when assigning students to classrooms, providing practical strategies to enhance learning outcomes in elementary schools.

Nearly 40% of high school graduates do not enroll in college in the fall following their graduations. Despite not following the traditional path through college, many of these students attend college; one-quarter of college students are 25+ years old. Of these older students, 23% of male undergraduates over 25 are current or former military members. We seek to understand what influences these students to enroll in college. Specifically, we estimate the impact workplace peers' prior college enrollment has on a soldier’s own future college enrollment. We leverage random assignment of first term soldiers to peer groups in the US Army between 2005 and 2017 to estimate that a 10 p.p. increase in the fraction of peers with any prior college enrollment increases soldiers' likelihood of enrolling in college by 0.44 p.p within three years of their joining the peer group relative to a mean of 10.94 p.p. Stated in terms of the enrollment effect of Post-9/11 GI Bill education benefits, the magnitude of peers' impact is approximately equivalent to the impact of an increase of $12,600 in benefit generosity per enlisted soldier.

While interdisciplinary learning has become a cornerstone of modern higher education, limited research explores the temporal dynamics between academic performance and the timing of interdisciplinary engagement. This study investigates the relationship between students' academic performance and their interdisciplinary learning behavior in higher education. Using data from a technological university in Taiwan spanning the years 2015 to 2020 and analyzing a sample of 7,342 students, we employ ordinary least squares regression to examine whether students’ academic achievement in one semester predicts the extent of interdisciplinary engagement in the following semester. Interdisciplinary learning is measured by the proportion of cross-school, intra-school courses taken. The results suggest that students with higher academic performance are more likely to engage in interdisciplinary learning in subsequent semesters, particularly through cross-school courses. Furthermore, we find that students who engage in interdisciplinary learning at later stages of their studies tend to perform better academically than those who do so earlier. However, these patterns vary across academic disciplines, such as engineering, management, and the humanities, indicating that the impact of interdisciplinary learning is domain-specific. These findings highlight the importance of aligning interdisciplinary curriculum design not only with students’ academic trajectories but also with the characteristics of their respective fields, and suggest that flexible, performance-informed learning pathways may better support interdisciplinary development.

Does the weaponization of dollar finance undermine dollar dominance in global trade? Using a new panel of FX-neutral import invoicing shares from 2012 to 2023, I show that U.S. financial sanctions against Chinese entities induce a selective shift from dollar to RMB invoicing. The response is strongest among countries whose imports from China are concentrated in working-capital-intensive sectors, while countries with low working-capital exposure exhibit almost no RMB switching. This heterogeneity suggests that sanctions do not trigger broad de-dollarization uniformly; instead, they reshape invoicing choices along financially vulnerable supply-chain links. I develop a model in which dominant-currency invoicing offers normal-time liquidity and hedging benefits but exposes firms to sanctions-sensitive working-capital constraints. Exporters borrow to finance intermediate inputs, and banks provide cash-flow-based credit lines whose usable amount depends on the invoicing currency and the realized geopolitical credit state. As U.S. sanctions risk rises, dollar-linked credit lines become less reliable because of the threat of payment disruptions, asset freezes, compliance delays, and reduced dollar settlement capacity. Firms exposed to working-capital-intensive production therefore switch to RMB invoicing to stabilize borrowing capacity. In the model, working-capital exposure lowers the threshold at which a trade relationship abandons dollar invoicing. A quantitative calibration disciplines this threshold mechanism using the reduced-form RMB response. Removing working-capital heterogeneity eliminates the muted response among low-working-capital countries, showing that the working-capital channel is quantitatively central. The findings imply that sanctions can weaken dollar dominance not through universal de-dollarization, but through selective currency switching where dollar liquidity becomes a source of geopolitical funding risk.

This paper examines the drivers of exchange rate pass-through (ERPT) and price-setting dynamics in the Philippines using a novel panel of region-sector price microdata matched with measures of global uncertainty, global value chain (GVC) positioning, and dollarization. We explore how sectoral and external characteristics shape the inflationary effects of exchange rate movements. Results show that ERPT is heterogeneous and state-dependent. Price stickiness significantly reduces pass-through, particularly in downstream sectors more exposed to final consumption. Dollarization is positively associated with price flexibility and likely reflects an amplified transmission channel under external volatility. These findings highlight the importance of sectoral structures and foreign currency exposures in shaping inflation responses, which, in turn, have implications for monetary policy design in small open economies.

This paper investigates the effect of tariffs on labor market dynamics within OECD countries. I employ a Two-Stage Least Squares approach, using annual data of 32 countries over the period 2000-2017. Findings reveal that tariffs increase unemployment, whereas trade openness reduces overall unemployment. This study further explores how tariffs affect unemployment among high and low-skilled workers. Results indicate that tariffs increase unemployment among high-skilled workers, which supports the view that international trade plays a crucial role in shifting employment toward high-skilled industries. Results further reveal that tariffs also increase unemployment among low-skilled workers, suggesting a labor reallocation from agriculture to the service and industry sectors. Moreover, this paper provides new evidence that tariffs increase employment activities in the agricultural sector and reduce employment activities in the industry and service sectors, suggesting that the shift of employment from agriculture to the industry and particularly the service sector is attributable to globalization.

Unequal access to basic physical infrastructure has been a deterrent to the economic development of marginalized groups and women in many developing economies. Given the renewed policy impetus on basic physical infrastructures in India, we use two-way fixed effects to empirically analyze the dynamic caste- and gender-based differences in access to water, electricity, LPG, toilet, and perception of support from the government for clean fuel. In addition, we examine the impact of household infrastructure on women’s decision making and time use, by caste. Results show that marginalized caste households gained more in terms of one-time investment infrastructure – household toilet and LPG connection, highlighting the catch-up effect, but were worse off than non-marginalized groups in gains in dynamic household infrastructures – reliable electricity, firewood usage and subsidy for clean fuel. With better physical infrastructures, women from non-marginalized caste gained more autonomy than marginalized women underscoring the role of social structures or norms as a binding constraint.

This paper focuses on indirect labor market exposure to international trade shocks that occurs via competition for job matches among many heterogeneous types of workers and firms. We show that such competition causes the equilibrium earnings and employment incidence of trade shocks to diverge strongly from initial product market-based exposure. We construct a large-scale assignment model of the U.S. labor market that flexibly accommodates production complementarities and frictions from switching regions, industries, and even particular employers, and we estimate its parameters using LEHD job match records combined with firm-level imports and exports from the LFTTD. We feed the model a detailed labor demand shock derived from regression estimates of the firm-level employment impact of China's WTO entry using exogenous tariff gap variation via four channels, import and export competition and import and export access. Our model simulations reveal that labor market competition causes the shock's impact to spread to seemingly unaffected sectors and trickle down the skill ladder. Thus, despite disproportionate relative job losses at manufacturing firms, multinational firms, and high-paying firms, low-paid non-traded service workers and initially unemployed job-seekers account for a large share of earnings losses and particularly unemployment increases. We also find that even industries and trade categories whose firms enjoy net job growth due to improved import access (e.g. wholesale/retail multinationals) can feature earnings and employment losses for their workers.

Since the first Starbucks store unionized in Buffalo in 2021, over 700 locations have filed for elections, and broader organizing has surged. We examine whether this campaign sparked wider labour activism. Using an event study, we find that counties with a Starbucks election saw nearly five additional non-Starbucks elections on average – accounting for nearly 20% of the post-2021 surge. However, a staggered difference-in-differences design reveals no wage gains for restaurant workers, suggesting that union momentum has not yet shifted bargaining power. Employers may be waiting on the outcome of first contract negotiations before adjusting pay structures.

One common government intervention designed to improve public safety is life guard certification. Putatively, regulations requiring rigorous certification standards help reduce pool drownings. But such certification may be costly. This paper explores the pre-COVID 19 practice of some local governments providing subsidies or free lifeguard training. Exploiting the staggered implementation of the free/subsidized lifeguard training, the paper estimates the causal impacts of this type of local government intervention on drowning outcomes. The model estimates are derived from restricted-use CDC mortality measures for pool drownings and a unique census of local government-sponsored or subsidized lifeguard certification programs.

The paper analyses how Gen AI can be utilized in forecasting demand for a start – up telecommunications company planning on introducing innovative technology in 1980s. The original study is a Harvard Business School case on business forecasting techniques applied by the firm for predicting demand with the advent of the new technology. The case, specifically, examines the challenges faced by the telecommunications company applying for an FCC (Federal Communications Commission) cellular radio license in Cleveland in 1982. Generative AI model (ChatGPT 4o to be precise) is applied to forecast business demand with human interactions through prompt engineering. The findings suggest that AI generally improved the depth and structure of forecasting by adding several insights that were overlooked in independent analyses. The paper concludes that “prompt engineering” and “domain knowledge” of the human counterpart play an important role in human – AI interactive process for small business forecasting.

How groups decide on the distribution of losses is central to many economic and social interactions. We study this in a laboratory bargaining game where four participants with unequal endowments must unanimously agree on how to distribute a collective loss. Initial endowments are assigned either by chance or by individual performance in a real-effort task, and bargaining occurs with or without information about one's own position. We find that the source of inequality matters: when endowments are merit-based, high-endowment participants contribute less, while low-endowment participants contribute more. Detailed analysis shows that under merit those with the lowest endowment accept significantly higher losses for themselves and lower losses for those with the highest endowments than under luck. This suggests that legitimacy perceptions extend in particular to those most disadvantaged. By contrast, the availability of information about one's position has little effect on final agreements. These findings highlight how fairness views rooted in merit versus luck shape collective burden-sharing and can lead even the least-advantaged individuals to endorse meritocratic rather than egalitarian outcomes when inequalities are perceived as deserved.

In a provocative review of Joseph Stiglitz’s 2024 book The Road to Freedom, Vernon Smith challenges theoretical work on asymmetric or imperfect information: “…the standard and the modified theories are irrelevant because buyers and sellers in possession only of dispersed, private, decentralized, value information easily converge to competitive price-quantity allocations in experimental markets over time via learning in repeat transactions.” (Smith, “Review of Road to Freedom,” 2025) He went on to describe theories of asymmetric information as being “born falsified.” While it is true that lemon’s markets in lab experiments by run Holt and Sherman and others have been shown to generate low qualities and prices, the setup used was generally not as competitive as the Davis and Holt (“Equilibrium Cooperation in Three-Person, Choice-of-Partner Games,” 1994) experiment with two sellers per buyer, which generated a majority of high-quality deliveries in 10-round sequences with fixed groupings. This competitive treatment in this paper provides sellers with enough excess capacity to capture sales from a significant fraction of competitors. Smith’s comment about learning through repeat transactions is implemented with a second treatment that allows buyers to post public ex post ratings of specific sellers.

We investigate whether peer study time exerts an influence on individual academic performance within the context of Chinese middle schools. By leveraging nationally representative CEPS data, exploiting random class assignments, and using an instrumental variable approach, we identify a direct spillover pathway from classmates' time investment to a student's academic achievement. The findings show a statistically significant positive effect of peer study time on academic outcomes. The results are robust after accounting for endogeneity, and non-linear dynamics. Subgroup analysis reveals stronger peer effects among socioeconomically disadvantaged students, while mediation analysis uncovers a trade-off: peer-driven academic gains are partially offset by worsened psychological well-being. These results contribute to a more nuanced understanding of peer influence.

Early detection of Alzheimer’s disease (AD) is critical because its neurodegenerative effects are irreversible, and neuropathologic change as well as modifiable social–behavioral risk factors accumulate years before clinical diagnosis. Identifying higher-risk individuals earlier enables prevention, timely care, and more equitable resource allocation. We study prediction of cognitive performance from social determinants of health (SDOH) using the NIH NIA supported PREPARE Challenge Phase 2 dataset derived from the nationally representative Mex-Cog cohort from the 2003 and 2012 Mexican Health and Aging Study (MHAS). The target is a validated composite cognitive score in seven domains: orientation, immediate and delayed memory, attention, language, constructional praxis, and executive function, which derived from 2021 and 2016 MHAS. Through comparative evaluation of multiple methods, XGBoost was selected as the default prediction model due to its superior performance. Our results show that the proposed framework outperforms existing methods and achieves better predictive accuracy than the top results reported on the data challenge leaderboard. To better understand the relationship between input features and the composite cognitive score, we conducted a thorough post hoc analysis of the top contributing features, examining the mechanism by which these features are associated with cognitive scores. Our findings underscore the significance of flooring material as more than a mere housing characteristic – it serves as a powerful proxy for deeper structural determinants of health. Its predictive strength lies in its ability to reflect socioeconomic status (SES), environmental exposures, and access to healthcare, all of which shape long-term health outcomes.

We estimate a semiparametric smooth-coefficient stochastic frontier model using panel data from U.S. hospitals and medical laboratories (1990 to 2024). Specification tests reject the parametric model for acute care ($p = 0.01$) but fail to reject it for medical laboratories ($p = 0.80$). Nevertheless, the two estimators produce statistically distinct efficiency distributions in both sectors. Semiparametric estimates indicate \textcolor{red}{hospitals expand their service scale or scope using shared inputs, leading to a mild increase in overall efficiency (medians 1.05 and 1.04)}. \textcolor{red}{Additionally, capital utilization, labor flexibility, diversified asset utilization, and varied reimbursement sources of the account receivable} yields statistically significant managerial gains in hospitals, whereas we lack sufficient evidence for a similar effect in medical laboratories. Depending on the industry, reliable benchmarking may require flexible models beyond standard parametric time trends and interactions.

China's innovative city pilot (ICP) program, a flagship policy initiative launched in 2008, represents a strategic tool of smart authoritarianism aimed at fostering innovation. By creating localized experimental zones that blend market-like incentives with targeted state support, the program seeks to address longstanding autocratic inefficiencies such as resource misallocation favoring politically connected entities and government’s micro-meddling. Drawing on firm-level data from Chinese listed companies, we conduct difference-in-differences analysis and find that the policy significantly promotes R&D investment, with heterogeneous effects across firm types and institutional contexts. Specifically, the positive effects are more pronounced for private sector firms relative to state-owned enterprises, for private firms whose chairpersons or CEOs lack political capital or connections, and for firms operating in provinces with retarded marketization progress. However, the ICP program does not seem to boost innovative efficiency. These findings suggest that the ICP program functions as an adaptive mechanism within China's smart authoritarian framework: they mitigate institutional barriers for disadvantaged private actors by providing state-directed equalization, while mimicking aspects of free-market environments through selective openness and intervention.

Forthcoming

In the digital economy era, intelligent transformation has become a critical strategy for firms to maintain competitive advantages. This paper examines how China's National Comprehensive Big Data Pilot Zones (BDPZ), as a typical place-based policy instrument, affect the intelligent transformation of micro-enterprises. We further extend the analysis to the complementary National New-Generation Artificial Intelligence Innovation and Development Zones (AIIZ) and explore their interactive effects. Using a difference-in-differences framework with panel data of Chinese A-share listed firms from 2009 to 2023, we find that both policies significantly promote intelligent transformation, with stronger effects in dual-policy cities. Mechanism tests show that BDPZ works through improving data accessibility, quality, and security, while both policies also operate via stimulating AI patenting. Further analysis shows that the intelligent transformation induced by these policies is associated with significant gains in firm total factor productivity. Heterogeneity analysis shows that the positive impact is stronger for firms located in eastern regions, those operating in technology-intensive industries, non-SOEs, and enterprises with smaller size and shorter listing history. Overall, this study offers reliable empirical evidence concerning the micro-effects of data-oriented and AI-oriented regional policies, and provides targeted implications for policy-makers to build a more supportive ecosystem to advance industrial intelligence.

We examine how artificial intelligence (AI) adoption affects firm performance and the operational channels through which productivity gains may arise. Using annual firm-level data from Korea’s Survey of Business Activities, which directly measures AI use beginning in 2017, we estimate cohort-specific dynamic effects of AI adoption on operational efficiency (inventory and asset turnover), cost structure (operating expense ratios), profitability (operating margins), and total factor productivity (TFP) growth. To address selection into adoption, we combine propensity score matching with an event-time difference-in-differences design that compares adopters to observationally similar non-adopters. Estimated effects are heterogeneous across cohorts and outcomes. Inventory and asset turnover show mixed responses, while operating expense ratios tend to rise for early adopters, consistent with short-run implementation and reorganization costs. Improvements in operating margins are limited, whereas TFP growth shows suggestive medium-run gains, particularly among larger firms and those located in the Seoul Capital Area; small firms exhibit weaker and more variable effects, though post-2020 cohorts show encouraging signs of TFP improvements. By production structure, inventory turnover displays a J-curve pattern in both capital- and labor-intensive firms, with faster adjustment among capital-intensive firms, while TFP gains are more evident among labor-intensive firms. Overall, AI’s productivity impact appears to depend on complementary inputs and organizational capacity.

Accurately measuring the private benefits of political power is central to understanding the incentives that shape public-sector behavior, yet conventional data often miss informal and in-kind income associated with political privilege. This paper develops an empirical framework to detect income underreporting linked to political affiliation and applies it to Communist Party membership in China. Using household expenditure data from the China Household Income Project and Engel’s law, we construct imputed income measures that reveal substantial hidden income among party members. The gap between imputed and reported income ranges from 10 to 22 percent across survey years, indicating that standard estimates understate returns to party membership. Incorporating adjusted income produces estimated returns of up to 68 percent, two to four times larger than conventional estimates. An instrumental variables strategy confirms a sizable causal effect, highlighting important implications for inequality, rent extraction, and governance.

California’s Central Valley is known as one of the state’s higher-poverty regions, although economic conditions differ noticeably across counties and local communities. Some areas experience higher poverty rates and weaker labor market conditions than others. Using data from the U.S. Census Bureau, this study compares male and female labor force participation rates and examines whether gender differences in labor-market participation are larger in higher-poverty communities. Educational attainment and demographic characteristics are included as control variables, and GIS-based mapping is used to examine how these patterns vary across the region. Preliminary findings suggest that higher-poverty communities tend to exhibit wider gender gaps in labor force participation. The results may provide useful policy insight for addressing gender disparities in labor-market participation across economically vulnerable communities.

Small businesses are the main engine of economic growth and job creation in the US. Given the outsized role of small businesses in the economy, much attention has been given to the role that the government and the private sector can do to support them. Entrepreneurship training programs are a key policy instrument for supporting small businesses in many localities, yet credible evidence on their local economic impacts remains limited, largely due to data constraints. This paper evaluates the effectiveness of Small Business Development Centers (SBDCs) in California and Texas using a novel advisor-client matched panel dataset covering several thousand small businesses from several recent years. We examine who seeks technical assistance, how business outcomes evolve following engagement with SBDCs, and whether these programs generate measurable local economic effects, including job creation, investment, and new firm formation. A central contribution of the study is its focus on how local economic freedom shapes the returns to entrepreneurship training. By comparing SBDC activity and outcomes across cities and states with markedly different regulatory environments, we will analyze whether technical assistance is more effective in jurisdictions with fewer barriers to entry, lighter regulatory burdens, and greater flexibility for entrepreneurs. The contrast between California and Texas, and between metropolitan areas such as Los Angeles, San Francisco, Houston, and Dallas, provides a natural setting to study how institutional constraints interact with public support programs. Local economic freedom will be measured with Stansel’s Metropolitan Economic Freedom Index (MEFI). Are training programs and economic freedom substitutes or complements in the production of entrepreneurship? Our findings intend to shed light on whether entrepreneurship training complements local economic freedom by amplifying private initiative, or instead compensates for restrictive regulatory environments by helping firms navigate compliance costs and other negative incentives. The results have direct implications for debates over the appropriate role of government in fostering entrepreneurship, the efficient allocation of public resources, and the design of place-based economic development policies that enhance local economic freedom and small business dynamism.

Artificial intelligence (AI) is increasingly used in hiring and performance evaluation, yet little is known about how its presence shapes labor supply decisions before individuals apply for jobs. This paper studies how perceptions of AI-based hiring influence participation and other labor market outcomes, with a focus on gender differences. We propose that AI affects labor supply through two belief channels: perceived fairness of evaluation and perceived replaceability of one’s role. These perceptions may vary systematically across groups and induce anticipatory adjustments in career-related behavior. We investigate these mechanisms using an online experiment. In the first stage, participants complete incentivized real-effort tasks spanning routine, creative, and cognitive domains. In the second stage, participants report beliefs about their performance, AI familiarity and use, and attitudes toward AI-assisted evaluation, and then construct a hypothetical job profile by selectively disclosing personal attributes and task scores. This design allows us to separate performance from beliefs and to observe endogenous signaling choices in the presence of AI. The experiment addresses three questions: whether perceptions of fairness and AI-related anxiety differ by gender; whether AI exposure alters beliefs about job security and replaceability; and whether these beliefs translate into differences in effort, confidence, AI adoption, and self-presentation. By focusing on anticipatory behavior rather than realized hiring outcomes, the paper provides new evidence on how AI may reshape labor supply decisions and, ultimately, labor market representation.

Traditional microeconomics assumes that the marginal utility of consumption always diminishes, and that money functions merely as a budget constraint rather than a direct source of utility. These assumptions have shaped generations of theory and practice, yet they do not reflect how real people think, decide, and behave. In this paper, I present a new paradigm of Simultaneous Marginal Utility Maximization (SMUM), in which individuals derive utility not only from consumption, but also from money, time, and emotional outcomes. By introducing a multidimensional model that incorporates these four elements together, I demonstrate that the marginal utilities of money and time do not always diminish; instead, they often increase in critical situations such as scarcity, emergencies, decision deadlines, and emotionally intense moments. This insight challenges the foundations of neoclassical economics and opens a new path for understanding human behavior based on pure logic, pure rationality, and reality. The SMUM framework unifies economic and managerial decision-making within a simultaneous and dynamic structure, offering a more comprehensive explanation for consumption choices, saving decisions, time allocation, and strategic behavior. This work establishes the cornerstone of a broader paradigm that is called Life Economics and Management, and provides the analytical foundation for Latif’s School as an emerging direction in the future of economic thought.

While economic models often assume that work involves an exchange of time and effort for money, people can find moments of satisfaction and achievement and establish social relationships. To determine whether visualisable meaning influences labour supply, this study employs an economic experiment. In recent years, a growing body of literature has argued that workers do not only care about their wages but also about whether their jobs are meaningful. We investigate how minimal perceived meaning influences labor supply and experimented in Japan, following the experimental methodology of Ariely et al. (2008). The subjects were male and female who work in Osaka city, recruited via distributed flyers. The experiment was conducted on October 29th and 30th, 2024, with each subject in the same condition gathered in the same room. Subjects had their views blocked by partitions, preventing them from seeing what other subjects were doing. Subjects were randomly assigned to one of the two conditions, Meaningful (N =18) and Sisyphus (N = 20). In a laboratory environment, we manipulated the perceived meaningfulness of a repetitive task. In each of the two conditions, subjects received payment for completing a white puzzle with 40 pieces according to a declining-unit wage schedule. Before deciding whether to build each white puzzle, the subjects were told how much they had earned up to that point and how much they would earn for completing another puzzle. In the Meaningful condition, after the subject had completed each puzzle, the experimenter would put a sticky note on the partition in front of the subject and would give him or her new puzzle pieces. In the Sisyphus condition, the experimenter wouldn’t put any sticky notes after the subject completed each puzzle. This was the only difference between the two conditions. Unlike the results obtained by Ariely et al. (2008), our experimental results did not observe any effect of perceived meaningfulness on labor supply. Reasons for the different results include cultural background and the influence of other subjects' behavior.

Across three online experiments with 1,600 participants, we find that people work less when a portion of their earnings goes to various organizations, even when personal earnings are held constant. Having a positive opinion of the beneficiary only weakly diminishes this effect, and giving people the ability to choose the beneficiary only helps if participants have the option to opt out of earning money for a beneficiary. This gives evidence that people are averse to “working for others” and that they also value having choice in the matter. Our findings suggest that decreasing the salience of taxes and enhancing taxpayer autonomy may increase the efficiency of and support for redistributive systems.

Differences in technology and absorptive capacity between developed and developing countries are significant and account for total factor productivity differences that widen disparities in per capita income across countries. Countries that can absorb more innovations -- due to strong education systems, advanced infrastructure, more investments in R\&D or supportive policies--- will have higher relative productivity. This paper estimates the degree to which the impact of foreign direct investment on total factor productivity depends on the absorptive capacity (AC) of the recipient economy using a panel of \num{93} countries over the period 1996--2018 in a system of simultaneous equations. Given the plurality of absorptive capacity measures existing in the AC-related literature, the paper develops a novel approach to compute a multidimensional AC index to reflect the country's overall ability to assimilate and adopt foreign technology. The paper finds evidence that absorptive capacity exerts a positive role in enabling FDI to promote productivity growth

This paper studies whether local operating costs drive product market concentration in U.S. consumer-facing industries. Using a county × 4-digit NAICS panel from 2018 to 2023, we document that higher local rents are associated with greater concentration and a rightward shift in the establishment size distribution, consistent with a selection mechanism that raises the productivity threshold for survival. To interpret these findings, we develop a spatial oligopoly model with heterogeneous productivity, nested CES demand following Atkeson and Burstein (2008), and location-specific fixed operating costs that generate endogenously higher productivity cutoffs, fewer firms, and larger markups in high-rent locations. The model delivers an inherent welfare tradeoff: higher operating costs select for more productive firms but simultaneously reduce variety and raise markups through increased market power.

This paper shows that fiscal rules interact with financial market conditions to substantially amplify recession risks during fiscal adjustments. Using a panel of about 100 countries from 1990 to 2022, we document that consolidations undertaken during periods of elevated sovereign spreads impose four times larger recession costs when countries operate under stringent fiscal rule frameworks (7.9 percentage points) compared to more flexible arrangements (2.0 percentage points). The amplification mechanism reflects constraints on policy composition: stringent rules force governments to rely on spending cuts and front-loaded adjustments precisely when such policies are most costly. Budget balance rules generate the strongest effects, followed by debt rules, while expenditure rules show minimal amplification. Local projections reveal these costs persist for 3-4 years rather than dissipating quickly. We rationalize these patterns with a simple framework in which stringent rules prevent governments from substituting toward less costly consolidation instruments during stress. The amplification of recession risks under higher-quality fiscal rules reflects a tension between credibility and flexibility. Traditional quality indices reward credibility-enhancing features but do not assess state-contingent flexibility, particularly in the first-generation rules that dominate our sample. For emerging economies exposed to sudden stops and volatile capital flows, these results suggest fiscal rule design should incorporate well-designed escape clauses and greater flexibility in rule type selection.

This paper provides evidence of the impact of High Frequency Trading (HFT) on liquidity. I use data from the NASDAQ OMX that identifies the trades of 26 HFT firms on 120 randomly selected stocks listed on the NASDAQ. I find that HFT improves overall market liquidity. However, the liquidity improvements come at the expense of the non high frequency traders. Results indicate that trades in which the HFT supply liquidity to non HFTs have a significantly wider spreads. This impact is larger for smaller cap stocks. Additionally, price impacts are largest when HFT demand liquidity to non HFTs. whereas realized spreads are largest when HFT supply liquidity to non HFTs. These findings indicate that HFT improve market liquidity, but impose higher costs to non high frequency traders. These results were amplified during the crisis week of September 15th-19th of 2008.

Climate change has increased the frequency and severity of environmental shocks, influencing patterns of human migration. In rural, resource-dependent areas with incomplete credit markets, negative environmental shocks can simultaneously increase the utility of migration while decreasing its feasibility due to liquidity constraints. This “trapped populations” phenomenon is commonly invoked in the popular press, but causal evidence of its existence is limited. In this work, we utilize exogenous variation in non-wage income from the world’s longest-running universal cash transfer program to examine how liquidity constraints mediate adaptive migration responses to environmental shocks. We focus on Alaska, the fastest-warming US state, where rising temperatures accelerate environmental degradation. Rural Alaskan communities often live below the poverty line and depend on subsistence hunting for nutrition, disproportionately exposing them to environmental shocks. Alaska is also uniquely home to an annual universal cash transfer, the Alaska Permanent Fund Dividend (PFD), which varies over time and is plausibly exogenous to individual migration decisions. We embed a dynamic household consumption model within a discrete choice model of rural-urban migration to understand the influence of liquidity constraints on migration. The model demonstrates that cash transfers have the greatest effect on climate-driven migration in communities with a higher share of financially constrained potential migrants. To test this prediction, we harness a 17-year panel of bilateral migration flows between 344 Alaskan communities and exploit within-community variation in summer heat accumulation and temporal variation in PFD payments. Preliminary results show that negative shocks from heat exposure decrease outmigration from rural areas in the absence of the PFD. This effect, however, is mediated by cash transfers: an additional dollar of PFD payments dampens the negative effect of heat on migration, qualitatively reversing its effect at the average observed PFD payment. Consistent with binding liquidity constraints, this effect is only observed in communities in the lower half of the per-capita wage distribution. Our results provide causal, within-country evidence of a mechanism previously documented only in cross-country comparisons. By exploiting exogenous variation in cash transfer payments, we provide evidence that liquidity constraints limit climate adaptation in one of the most environmentally vulnerable areas on Earth.

Periods of resource scarcity often lead regulators to rely on command-and-control rationing rather than price-based mechanisms, despite longstanding insights showing such rules to be inefficient. Urban water management during drought is a prominent example: water agencies are frequently assigned uniform conservation mandates that ignore underlying heterogeneity in marginal costs and demand elasticities. At the same time, political and legal constraints often prevent water agencies from implementing full marginal-cost pricing to reflect scarcity. This paper quantifies the welfare losses created by the interaction of these two distortions, inefficient conservation mandates and inflexible prices, during California’s historic 2011–2017 drought. I assemble a novel panel dataset covering the full universe of roughly 400 California urban water suppliers, including detailed information on water consumption, prices, operating and capital expenditures, groundwater pumping, and hydrologic conditions. These suppliers together serve over 90% of the state’s 40 million residents. A 2015 policy shock required each supplier to meet a mandatory, percentage-based reduction in use relative to a pre-drought baseline, with targets varying substantially across suppliers and enforced with meaningful penalties. This heterogeneity generates plausibly exogenous variation useful for estimating key supply-side parameters. Using this variation, I estimate supplier-specific marginal costs via an instrumental variables strategy that leverages the divergence between observed and predicted, policy-exogenous conservation targets. The resulting 2SLS estimates reveal substantial heterogeneity: while the average marginal cost is about $253 per acre-foot—substantially higher than assumptions used in earlier work (e.g., Buck et al. 2016)—estimates range from below $70 in groundwater-abundant regions to more than $300 for suppliers reliant on imported surface water. I incorporate estimated parameters into a simulation model that reallocates drought-mandated conservation burdens across suppliers within hydrologic regions. Simulations show that allowing prices to flexibly reflect marginal cost would further reduce welfare losses, underscoring the inefficiencies created by binding price ceilings. Taken together, the results demonstrate how incorporating supplier-level cost heterogeneity can substantially improve drought policy design, even when prices cannot fully adjust to reflect full social costs. Such insights are crucial for the efficient design of future conservation mandates in a climate where scarcity conditions are intensifying.

Forthcoming

On Puget Sound shorelines in Washington state, USA, bulkheads and other hard armoring structures intended to prevent erosion and protect from coastal flooding often conflict with nearshore habitat conditions suitable for supporting conservation objectives. In this study we measure shoreline property owners' preferences over traditional and alternative shoreline management strategies, specifically hard armor, engineered soft shore, and natural shoreline approaches, along with on-site retreat (i.e., moving the home back from the shoreline). Using a Discrete Choice Experiment (DCE), we also measure the effects of financial incentives, specifically small grants and low-interest construction loans, on shoreline management decisions. We surveyed a random sample of 10,134 residential shoreline property owners in the Puget Sound region in the Spring 2025, achieving a 40.6% effective response rate (n = 3,822). The sample includes both armored and unarmored properties, allowing for an analysis of the relative effects of financial incentives on encouraging armor removal vs. armor avoidance. DCE data are used to fit multinomial, conditional, mixed, and latent class logit models from which we estimate willingness-to-pay for project types and features and the effects of financial incentives on property owner decisions. Shoreline property owners tended to be older (median age 71), higher income ($125,000 to $249,999 modal income bracket), and retired (69%). multinomial logit results indicate owners of natural shorelines are more price sensitive than owners of armored shorelines, the average shoreline property owner is unlikely to voluntarily choose on-site retreat under typical project prices, and grant offerings have small but significant influence on shoreline management choices across status quo shoreline conditions. These models can be used to inform incentive program design in the Puget Sound region.

Hydraulic fracturing (fracking) is often considered one of the most important innovations in the energy sector over the past five decades (Bartik et al., AEJ: Applied Economics, 2019). The combination of hydraulic fracturing and horizontal drilling has enabled a dramatic increase in shale gas production in recent years. While proponents highlight economic benefits—such as job creation and royalty payments to landowners—broader economic gains arise from cheaper natural gas and lower carbon emissions relative to other fuel sources. However, researchers across disciplines have also raised concerns about the environmental damages associated with fracking, including potential water contamination, air and noise pollution, truck traffic, and even minor earthquakes. Such adverse impacts have been widely documented in the economics literature (e.g., Olmstead et al., PNAS, 2015 on water contamination; Hill, JHE, 2018 on air pollution and infant health; Boslett et al., JEEM, 2015 on housing prices). Despite the growing body of work examining the impacts of fracking, the literature has been largely silent on fracking regulation. The regulatory landscape is further complicated by the absence of federal-level oversight, which has exempted fracking from key environmental laws and left regulation to state and local governments. As a result, communities exhibit substantial heterogeneity in their responses, ranging from symbolic resolutions to outright bans. In this study, we use panel data techniques to estimate how local regulations adopted by neighboring communities affect drilling activity by energy companies in municipalities without fracking regulations. We compiled a novel dataset, which combines information from Food & Water Watch (fracking bans in the US), the U.S. Census American Community Survey (municipality socioeconomic data), the National Center for Charitable Statistics (environmental group activity), and the Enverus database (drilling activity). Our results show that as the number of regulatory actions in neighboring communities increases, drilling activity within and around a municipality declines. This finding holds when extending the area of interest beyond municipal boundaries to 3-mile and 5-mile buffers. Additionally, we find that drilling activity increases just outside municipal borders where environmental groups have a strong presence, providing empirical evidence that societal pressure induces drilling companies to relocate operations outside city limits.

Investment in commercial-scale onshore wind farms in the United States is lower than needed to meet net zero carbon emissions by 2050. State-level policies could account for this relatively low investment. Since each state has their own tax laws, my contribution is analyzing the impact of these tax laws on wind farm investments. I use a synthetic control approach to examine the effects of the Texas Chapter 313 tax code on wind farm investments. Chapter 313 allows local governments in Texas to grant property tax abatements for new large capital-intensive investments. Wind farms fit within the definition of large capital-intensive investments and can therefore receive tax abatements through the Chapter. The data is from the United States Wind Turbine Database, the Database of State Incentives for Renewables & Efficiency, and the U.S. Wind Siting Regulation and Zoning Ordinances database, which include data for the 48 contiguous U.S. from the period 1982-2025. The results show that the Chapter 313 tax code had a significantly positive effect on wind farm investments in the state of Texas. The results can inform policymakers that state-level tax abatements for wind farms are an effective policy tool to significantly increase wind farm investments in the state.

Cement production generates roughly 8% of global CO₂ emissions, with nearly half arising from fossil fuel combustion in clinker production. In the short run, decarbonization relies on switching from coal to cleaner fuels such as natural gas. To promote lower-carbon construction materials, the recent Buy Clean Initiative introduces embodied-carbon standards in public procurement, reshaping eligibility for public contracts to prioritize low-emissions suppliers. The policy creates incentives for producers to adopt cleaner production methods to remain competitive in public contracting. This paper studies how green procurement policies affect firms’ fuel-switching decisions and market outcomes in a spatially differentiated industry. I construct a plant-level panel using the Portland Cement Association’s Plant Information Summary, including kiln technology, fuel use, and capacity. These data are merged with spatial data on natural gas pipelines and coal mines from the EIA, aggregate production and price data from the USGS, and fuel prices from the State Energy Data System. County-level cement demand is measured using BEA Input-Output tables, BLS employment data, and Census economic data. To study fuel-switching behavior, I exploit quasi-experimental variation in energy access costs driven by coal mine closures and proximity to natural gas pipelines. Preliminary results show that plants are more likely to switch to natural gas when pipeline access is closer, natural gas prices are lower, and kiln technologies are less energy efficient. I then quantify how policy effects vary across regions with heterogeneous fuel costs, infrastructure access, and market structure using a two-stage model of investment and procurement competition. In the first stage, plants decide whether to incur a fixed cost to switch fuels. In the second stage, suppliers compete in spatial procurement markets through second-score auctions, where buyers value bids and transportation costs. Procurement standards reshape competition across buyer types: conventional buyers remain price-oriented, while green buyers procure only from eligible low-emissions suppliers. Counterfactual simulations show how policy stringency affects fuel-switching incentives and market outcomes across regions. This paper contributes by developing a framework that links demand-side environmental policy to firms’ fuel-switching investment through procurement competition, and by introducing heterogeneous buyers to show how carbon-intensity thresholds reshape competition across markets.

Forthcoming

Forthcoming

This paper studies how managers matter for worker productivity by exploiting manager moves in Major League Baseball. I first show that manager effects are primarily match-specific using an AKM-type model. Then, staggered difference-in-differences results show positive effects of shared characteristics, including same position, both lefties, and same hometown. Effects of same position and both lefties are larger than same hometown. Effects are more pronounced for younger and less productive workers, indicating that mentoring is the primary mechanism while in-group bias is secondary. Finally, there is suggestive evidence that a manager's productivity as a worker is not a positive characteristic.

Forthcoming

We show that weak judicial enforcement and credit frictions jointly govern the transmission of corporate distress through trade-credit networks. Weak judicial enforcement increases the losses associated with defaulted trade credit, while credit frictions limit trade creditors’ ability to absorb those losses. We exploit variation in court congestion generated by a nationwide reform that reassigned pending bankruptcy cases across courts. Trade creditors are more likely to go bankrupt when their trade debtor’s case is handled by a congested court and when they face binding credit frictions. Bank relationships mitigate these frictions and insulate trade creditors from distress transmission. A counterfactual exercise shows that moving from the least to the most efficient courts would reduce bankruptcy propagation by 40% and increase aggregate sales by 2%.

This paper investigates feedback effects between bank health and zombie firms—financially distressed firms receiving subsidized credit. The literature focuses on how banks create zombies, overlooking zombies’ impact on bank health. Using Spanish firm-bank data (2005-2014), we document a vicious cycle: lower bank capital ratios are associated with higher zombie activity in served industries, while higher zombie prevalence is associated with reduced bank capital. We link this to a previously unexplored mechanism where banks respond appropriately to observable financial distress through higher provisioning, but overlook risks from relationship borrowers receiving subsidized rates. Our findings suggest that this feedback stems not from financial distress alone, but from the combination of distress with interest rate subsidies.

Forthcoming

Pretextual stops, sometimes called ``Terry stops,'' are a common policing strategy. They involve stopping a driver for minor infractions, such as tinted windows, to investigate other suspected criminal activity. For instance, a police officer may pull over a driver for a minor infraction and ask the driver, "Do you know why you were pulled over?'' The police officer asks this question, hoping the driver admits to a more severe crime. Although the U.S. Supreme Court ruled these traffic stops constitutional in 1996, there is an active debate about whether police officers should conduct these stops. While pretextual stops are commonly employed across police departments, such use has gained much attention and controversy. In 2024, California adopted a new legislation that restricted pretextual stops by increasing the cost of conducting them. Under the new bill, effective on January 1, 2024, when police officers conduct traffic stops, they must verbally explain the reason for pulling over the driver and record their reasoning on their body-worn video cameras. Failure to comply will result in the officer undergoing retraining after the first offense, followed by severe discipline for subsequent violations. In this paper, I leverage several administrative datasets and employ various identification strategies to study whether a statewide pretextual reform led to: (i) a reduction in “likely” pretextual stops, (ii) a compositional change in who is getting stops (by offense type and by race), (iii) any change in post-traffic stop outcome, and (iv) public safety as measured by reported crime, gun violence, and traffic accidents.

We provide estimates to inform the potential tradeoff between the crime reduction benefits and social costs associated with investments in policing. First, using stated-choice experiments in an original survey, we estimate willingness to pay for reductions in violent crime and police use of force. Our survey results indicate that citizens are willing to accept four additional violent crimes in return for one fewer force incident. Next, using data from Chicago and an identification strategy based on officers’ predetermined work schedules, we estimate causal effects of marginal officers on crime and use of force. Combining these estimates with our survey findings, we show that while marginal officers provide a crime reduction benefit which exceeds their salary costs, incorporating their impacts on social costs associated with use of force reverses conclusions about cost-effectiveness.

This paper documents the effects of employment protections for victims of domestic violence as established by California’s Senate Bill 400 implemented in January 2014 on the reporting of intimate partner violence. Results indicate that rates of intimate partner violence reports increased by 14 percent following the enactment of the law. This appears to be primarily driven by Black and Hispanic victims. Further, the increase is concentrated in tracts with higher unemployment rate and poverty rate as well as non-White and renter population, and those with below median income and education. Arrests for intimate partner violence also rose, driven by felony rather than misdemeanor cases, while 911 calls for domestic abuse remained unchanged – consistent with increased reporting rather than higher incidence. These findings suggest that strengthening job-retention rights for victims can lower the economic costs of disclosure, improve engagement with formal institutions, and reduce disparities in reporting across communities.

The FY 2024 NDAA mandated the development of an analytic framework for setting special and incentive (S&I) pays in the US military. This report documents analytic work in support of the development of such an analytic framework, including a review of the labor economics literature and the literature on the setting of military compensation, interviews with key subject matter experts on special and incentive pay policy, and analysis of quantitative data on S&I pay receipt among service members, as well as the framework itself and two example applications of the framework to S&I pays. We find that while many S&I pays have traditionally been seen as having the purpose of recognizing unusual levels of danger, hardship, or other adverse circumstances faced by service members, the use of recognition as a purpose is at odds with a major objective of the military compensation system, which is to support military force management objectives. We instead recast recognition as an incentive for members to take actions in support of force management objectives determined by DoD and the military services, following the labor economics theory of compensating wage differentials. This represents a significant shift in the conception of S&I pays within DoD and the services. The framework provides a guide for using the purpose of individual S&I pays, described in terms of the force management objectives they support, to set the eligibility requirements for and amounts of S&I pays.

The goal of Assignment Incentive Pay (AIP) is to incentivize service members to volunteer for positions that are difficult to fill or less desirable. We use structural modeling and quasi-experimental methods to evaluate the effect of AIP on the quantity and quality of members assigned to high-need areas. In our analysis, we focus on changes in two different AIP programs: the increase in AIP for assignments in Turkey in December 2022 and the April 2014 implementation of the Intercontinental Ballistic Missile Operations AIP program. We evaluate these changes using administrative Air Force Personnel Center data from FY 2011 through 2024.

Many labor markets are characterized by heterogeneous workers competing for jobs which differ in difficulty or required skills. Comparative advantage theories argue wage differences across jobs act as signals for the supply of skilled labor in a competitive market, facilitating the sorting of specialized workers to more complex jobs. In the United States' public sector, however, federal and state laws often constrain the amount wages are allowed to differ by job. For example, members of the military receive the same basic pay across all job types, given equal rank and years of service. I use NLSY97 data on military service members to calibrate an assignment model under different wage-setting policies. The model shows workers inefficiently sort into jobs when wages are held constant across job types---lowering overall output by 95 percent compared to when wages differ across jobs competitively. Output loss is mediated, but still present, at 56 percent when a simple sorting mechanism is introduced to allocate workers to jobs by their observed skill level. This mechanism mimics the military’s current policy to use standardized test scores and algorithmic-matching to help sort workers into jobs. The results in this paper show policies to centrally sort workers on observable characteristics may be less effective in increasing military output than policies that allow for wage differentiation by job type in the basic pay table. This finding is broadly applicable across public sector organizations and private firms with diverse workforces, numerous job types, and constrained wage-setting policies.

In credit markets, borrowers are more informed about the riskiness of a loan than lenders. Lenders are aware of this information asymmetry. They will not extend credit unless borrowers can credibly signal that they will not default. Online, there are viable markets for peer-to-peer, pseudonymous loans. The problem of asymmetric information is being mitigated, but how? When borrowers request a loan, they must specify the payment platform, e.g., Zelle, Venmo, etc. PayPal is one of the few such payment platforms that offers some degree of buyer protection. In this paper, we empirically test whether borrowers who request a loan with PayPal tagged as the repayment platform internalize the buyer protection by offering a lower interest rate (APR). Preliminary results show that PayPal tagged loan requests do indeed have a lower requested APR.

The Bureau of Indian Affairs holds approximately 55 million acres in trust within the United States. Using newly geocoded data linking individuals on trust lands to credit reports, we assess credit market differences compared to similarly creditworthy individuals elsewhere. Our matching estimators reveal trust land residents typically have smaller mortgage balances, higher credit card utilization, and greater debt from alternative financial institutions, increasing likelihood of extreme delinquencies. However, these land-specific effects are relatively small compared to person-specific factors. While policies encouraging lending on trust lands would help reduce capital access gaps, investments targeting individual financial capabilities would address these disparities more efficiently.

The CARES Act forbearance program, as one of the major pandemic-era policy interventions in the consumer credit markets, has provided short-term debt relief to millions of U.S. consumers. In this paper, we study the long-term impact of forbearance, particularly whether it could help distressed borrowers regain their footing post-forbearance. To overcome identification challenges, we use census block-group-level access to Internet Broadband as a novel instrument for the likelihood of forbearance takeup. Given the closure of bank branches, and call centers that were overwhelmed with the large volume of those applying, we find that fast internet access is an important determinant of receiving forbearance, We use this access-based variation in forbearance to identify causal effects on borrower performance and balance sheets for borrowers who were current on their mortgages at the onset of the pandemic, and then subsequently missed payments and entered forbearance. As designed, forbearance stabilized housing markets, even in the long-run: it persistently increases the likelihood of current mortgage status and reduces delinquencies through 2024. By contrast, we find no measurable effect on auto loan or bankcard delinquencies. Finally, the CARES ACT mandated that these missed payments should not directly impact credit scores. Consistent with this, we find a significant increase in credit access: new bankcard accounts and a concomitant decline in utilization rates.

This paper develops a static, multisectoral general equilibrium model of an input-output economy to study how defaultable credit impacts firms' debt structure and production decisions. The economy features N sectors with heterogeneous firms subject to idiosyncratic productivity shocks. Firms use a hybrid debt structure, issuing junior bonds to borrow capital and utilizing senior trade credit (a short-term debt with possibility of default) for intermediate input purchases. A key feature is that debt contracts are subject to limited enforcement (such as under limited liability), which grants firms the option to default on their obligations. In the event of default, a portion of the firm's output is lost as deadweight loss and the remaining output is transferred to creditors. The model divides the period into two stages for analyzing credit and production decisions relative to shock realization. We ask: How does default risk in the intermediate input market affect sectoral outcomes and firms' debt structure? Crucially, what is the magnitude of the welfare loss for the US economy caused by the sectoral misallocation of resources between final and intermediate outputs? We contribute by integrating the microeconomic mechanism of defaultable credit with an input-output structure, characterizing the equilibrium where firms optimally choose input levels while accounting for the endogenous default probability. Theoretically, a pecuniary externality creates suboptimal risk premiums, driving the overuse of senior trade credit and leading to sectoral output misallocation. This creates over-demand from upstream sectors, mainly producing intermediate inputs, forcing the economy to overuse resources in upstream sectors, leaving fewer resources for downstream sectors to produce final goods. This sectoral output misallocation, which can only be detected in a production network setup, appears as a reduction of aggregate TFP. The result is robust to varying model assumptions. In the second part of the paper, we quantitatively assess the welfare loss for the US economy by calibrating the model using BEA input-output tables, Compustat, the FISD dataset, and sectoral producer price indexes. Our results confirm the theoretical prediction that financial frictions lead to the significant overuse of trade credit.

We study two-player contests with delegation, in which each player begins with an initial probability of winning a prize and hires a delegate to exert effort on her behalf. Each delegate announces the period in which he will exert effort, and then chooses his effort level in the announced period. We show, among other things, that, if player 1 has a higher initial winning probability than player 2, then player 1 offers a less generous (or lower) contract to delegate 1, compared to the contract that player 2 offers to delegate 2, delegate 1 exerts effort before delegate 2 does, and delegate 1 exerts less effort than delegate 2. We also show that, despite both delegates having equal ability in the contest, delegate 1 earns a greater economic rent than delegate 2. Finally, we show that, as player 1's initial winning probability increases from one half, both players' contracts become lower less generous (or lower), the total effort level decreases, player 1's expected payoff increases, delegate 1's expected payoff may either increase or decrease, and both player 2's and delegate 2's expected payoffs decrease.

This paper studies collective decision-making under uncertainty when the initial vote yields only the narrowest possible majority. The majority of one is formally decisive yet may be too fragile to justify immediate implementation. We analyze a mechanism by which such an outcome may trigger further deliberation and the appointment of two additional decision-makers. The model applies to any committee that must choose between competing alternatives when delay is costly. We characterize the conditions under which committee members reopen deliberation, the conditions under which the added members vote informatively, and compare their error probability with that of a standard one-shot majority rule.

An increasing number of studies empirically examine the causes and consequences of corruption from various perspectives. Most of this literature relies on survey-based corruption indices or measures constructed by private risk-assessment firms. However, cross-country survey data on corruption suffer from important limitations, including comparability issues and the inherently hidden nature of corrupt activities. This study adopts the Multiple Indicators Multiple Causes (MIMIC) model to estimate the extent of corruption across countries, thereby offering an alternative approach to measuring corruption. A key advantage of the MIMIC framework is its ability to generate out-of-sample estimates through extrapolation, whereas survey-based measures typically provide corruption indices only for specific years. Moreover, by relying on observable indicators and structural relationships, MIMIC-based measures can mitigate biases arising from subjective perceptions. Using the QoG Standard Dataset (Teorell et al., 2018) from the Quality of Government Institute, we estimate latent corruption levels and find that our corruption estimates based on the 2014 cross-sectional data are highly and statistically significantly correlated with Transparency International’s Corruption Perceptions Index (CPI). These results provide supportive validation of our approach. Nevertheless, the findings should be regarded as preliminary. Future research will conduct additional robustness checks and systematically evaluate the relative strengths and limitations of alternative corruption measures.

Firms responded to the U.S.–China trade war by exploiting tariff avoidance channels, including small-package de minimis shipments and rerouting through third countries. Under U.S. law, shipments valued below $800 are exempt from tariffs and customs procedures, but are excluded from official trade statistics. Using quarterly country–product trade data from 2016 to 2024, the analysis captures both immediate and delayed adjustment margins. A Poisson Pseudo–Maximum Likelihood (PPML) difference-in-differences framework estimates how tariff exposure affected bilateral trade, with products classified as de minimis–eligible based on China’s Cross-Border E-Commerce export lists. U.S. imports from China declined sharply after tariff escalation, with de minimis–eligible products falling by 13.3 percent and non–de minimis goods declining by 25 percent, consistent with a reallocation of small-value goods into unreported de minimis channels. In contrast, non–de minimis goods were more likely rerouted through third countries, particularly those with preexisting trade ties to China, such as Mexico, South Korea, and Hong Kong. Tariff revenue rose primarily for non–de minimis Chinese goods, indicating that some firms continued importing directly and absorbed the duties rather than fully diverting shipments. These strategies contributed to significant blind spots in official trade statistics. A structural framework of export channel choice formalizes firm behavior across four discrete strategies: direct shipment, de minimis entry, rerouting, and exit. Comparative statics confirm that circumvention strategies diluted the effective incidence of tariffs and reshaped observed trade flows.

Trade finance plays a critical role in facilitating international trade by bridging payment timing gaps, easing working capital constraints, and mitigating information and contract enforcement asymmetries between trading partners. Despite its importance, empirical evidence on trade finance remains limited, particularly for emerging market economies. This study provides a comprehensive analysis of bank-intermediated trade finance in the Philippines using transaction-level data, covering export and import transactions for the period 2014-2025. The analysis adopts a two-part approach. First, it documents key stylized facts on the structure and dynamics of Philippine trade finance, examining instrument type, bank characteristics, firm sector and industry, currency denomination, counterparty country features, and maturity profiles. A complementary bank survey provides context to help interpret observed patterns. Using both stylized facts and survey results, a set of testable hypotheses is identified. Second, the study investigates how trade finance responds to counterparty characteristics, macroeconomic conditions, and global shocks. Using event studies, local projections, and panel regressions, it examines both the immediate and dynamic responses of trade finance to these drivers. This study fills important gaps in the literature and provides new evidence on the role of financial intermediation in supporting trade during both normal periods and episodes of global stress.

Using disaggregated data on Taiwan’s exports by industry and destination, this paper examines how export volumes and prices respond to exchange-rate movements and how these responses vary with product quality. A Taiwan-dollar appreciation reduces export volumes overall, but the elasticity is highly uneven. It is substantial for advanced and other non-advanced markets, yet effectively zero for China and Southeast Asia. Across specifications, export quality mitigates the decline in export volumes associated with appreciation, and a positive industry-level quality gradient is also observed in export prices. Crucially, this gradient emerges only where the scope for vertical differentiation is greater, namely in high-technology industries, industries with long quality ladders, and advanced-economy destinations, and disappears or reverses where such scope is limited. This pattern suggests that quality differentiation helps insulate exports from appreciation, largely through a compositional shift toward surviving high-quality varieties, particularly in markets where the differentiation literature predicts such reallocation to be strongest.

This article uses micro-level data from the Indian Human Development Survey to determine how the ownership of time-saving household appliances affects the health of children. The ownership of appliances such as a washing machine or refrigerator may be associated with a reduction in physical labor, a longer sedentary time, and a higher consumption of processed foods. In our analysis, we find that children aged 5-18 years who live in a household that owns a refrigerator experience an increase in the incidence of being overweight by 5.6 percentage points. The effect of owning a washing machine or refrigerator is greater, around 8 percentage points, for older children aged 12-18. When the sample is partitioned on the basis of gender, these outcomes are more pronounced among males. We instrument household ownership of time-saving appliances by average ownership rate among households with no children living in the same community as ownership of household durable goods is endogenous. With childhood obesity being a growing crisis in India, understanding some of the factors that might contribute to it will be a crucial step in fighting it.

Previous studies have documented that early childhood education (ECE) is crucial for children’s development. However, much of the existing evidence is based on research conducted in the United States and other Anglo/English-speaking countries. Consequently, far less is known about the effects of ECE attendance in non-Western societies with distinct demographic and cultural characteristics. The role of ECE in these contexts may differ from that in Western societies due to differences in family structure, childcare arrangements, credentialism, and social norms. Therefore, it is important to examine how ECE attendance affects children’s development in diverse cultural and social settings. Using a nationally representative longitudinal dataset of children in Taiwan, we examine the effects of ECE attendance on early cognitive, non-cognitive, and physical developmental outcomes. To address the potential endogeneity of ECE attendance, we employ a combination of propensity score matching and a difference-in-differences (DiD) approach. Specifically, we first use matching to construct a control group of children who did not attend ECE but share similar observed characteristics with the treatment group. We then apply a DiD model to the matched sample to estimate changes in children’s developmental outcomes after ECE attendance, relative to changes among children who did not attend ECE. Our results show that, compared with informal care settings, ECE attendance significantly improves children’s social competence and motor development. However, we do not find significant effects on cognitive or language development. These findings are robust to alternative matching methods and DiD estimators. Moreover, subgroup analyses reveal that the positive effects are more pronounced among children of less-educated mothers, suggesting that ECE may help compensate for disadvantaged family backgrounds. Our findings contribute new evidence to the literature on the impacts of early childhood education in East Asia, where empirical evidence remains limited. This study is particularly timely as many East Asian countries increasingly adopt ECE expansion policies as a core component of their pro-natalist strategies.

The rapid rise of financial technology (FinTech) has reshaped the global financial landscape by introducing new business models and intensifying competition within the banking sector. Central to this transformation are neobanks, which operate digitally through partnerships with licensed institutions, and challenger banks, which hold their own banking charters and compete directly with incumbents. This structural bifurcation has accelerated digital transformation among traditional banks, yet existing evidence on their market performance and risk characteristics remains mixed, with studies suggesting both competitive pressures and increasing systemic interconnectedness. This paper examines the dynamic connectedness between neobanks, challenger banks, and traditional banks using daily data from 2017 to 2025. We construct three global equal-weighted indices representing listed neobanks, licensed challenger banks, and large incumbent banks, and employ a time-varying parameter vector autoregression (TVP-VAR) connectedness framework to quantify total, directional, and net spillovers across these groups. The results reveal pronounced asymmetries in shock transmission linked to digital banking business models. Neobanks act as net shock transmitters during periods of elevated technology-sector sentiment, particularly during the 2020–2021 FinTech/SPAC boom and the 2022–2024 AI-driven revaluation. Challenger banks emerge as dominant transmitters during episodes of macro-financial stress, including the COVID-19 pandemic, the 2022 interest-rate shock, and the 2023 banking turmoil. Traditional banks are generally net receivers of shocks, except during the most severe stress episodes. Event-based robustness analyses confirm that connectedness is strongly regime-dependent and closely related to differences in licensing status, funding structures, and balance-sheet exposures. Overall, the findings provide new evidence on how digital banking models shape shock propagation and carry significant implications for financial stability in an increasingly digitalized banking system.

We study how an aging population and a sustained decline in fertility affect bank liquidity creation and whether and how it impacts banking market structure and business models. For identification, we exploit the variation in fertility rates and population age that is observed across the U.S. states. We show that consistent with the life cycle hypothesis, an aging population is negatively associated with banks’ liquidity creation. The aging environment is also associated with reduced competition. Further investigation shows that an aging population positively predicts bank non-interest income growth: Banks that are located in a state with an aging population are more likely systematically to increase the diversification of their activities. The downward pressure on their profits as a result of aging demographics leads to a reduction in traditional interest income and maturity transformation activities and an incentive to increase the share of revenue that is generated by non-interest income activities.

After more than three decades of intensive development of multifactor models, no asset pricing studies in finance empirically test the extent to which more advanced models can explain bank stock returns. A natural question is: Are bank stocks anomalous to more advanced multifactor models for nonfinancial firms? That is, are banks so different from nonfinancial firms that they cannot be effectively priced by traditional models? This paper contributes to the asset pricing literature by investigating whether nonfinancial multifactor models can explain bank stock returns. Using standard Fama and MacBeth (1973) cross-sectional regression analyses, we comparatively study the CAPM plus five well-known multifactor models that have evolved over time. Also, we include the recently proposed Mean Variance Pricing (MVP) model. Using size- and beta-sorted portfolios and out-of-sample cross-sectional regression tests, we find that market factor loadings (i.e., beta estimates) in the CAPM are significantly priced in the cross section of average bank stock returns but contrary to theory have negative signs. Particularly in the case of size-sorted portfolios, as various factors are added to multifactor models, their ability to price bank stocks markedly improves. Also, relative to other models, we find that the MVP model does a superior job of pricing bank stocks. In size-sorted bank stock portfolios, it achieves almost perfect out-of-sample goodness of fit (viz., R-squared of 99 percent) and has highly significant factor loadings associated with zeta risk (viz., t-statistic exceeding 4). Tests based on other test asset portfolios corroborate our findings. Also, graphical analyses of out-of-sample average predicted (fitted) excess returns and actual excess returns confirm that average mispricing errors decline as factors are added to multifactor models. However, mispricing errors are normally smaller for the two-factor MVP model compared to other models, especially in the case of size-sorted portfolios. In view of these findings, we conclude that, as factors are added to multifactor models, their ability to price bank stocks improves. Even so, the parsimonious two-factor MVP model outperforms other multifactor models. We infer that bank stocks are less anomalous to asset pricing factors used in nonfinancial studies than previously believed.

Using a large dataset of firm-bank and ownership information for 23 European countries from 2008 to 2015, we explore how firm and bank heterogeneity shape the effects of bank switching on firm performance. We show that adding banks who are knowledgeable about economy increases firm investment and trade credit, while leaving loan volumes unchanged and raising interest expenses. Adding banks specialized in the firm’s industry leads to higher output growth and productivity. These effects are economically meaningful and persist among financially constrained firms. Overall, our findings highlight the role of banks as providers of information and complementary services, operating through channels beyond traditional investment and financing effects.

This study investigates the potential impact of open-source contributions on national economic growth, with a particular focus on emerging technology sectors including Artificial Intelligence (AI), Unmanned Aerial Vehicles (UAV), Blockchain, Quantum Computing, and RISC-V. Addressing the limitations of traditional economic growth models in capturing technological progress, the research incorporates GitHub contribution data into an extended Solow Growth Model as a proxy for innovation input and technological advancement, utilizing a cross-country panel dataset from 2012 to 2021. The empirical analysis demonstrates that overall open-source activity exerts a significant positive influence on GDP growth, validating the role of open-source innovation as a modern driver of economic performance. Among the examined sectors, contributions related to RISC-V, AI, Blockchain, and Quantum Computing show significant positive impacts on GDP growth, highlighting the importance of open-source ecosystems in fostering technology diffusion and industrial upgrading. Conversely, contributions in the UAV sector exhibit no significant impact, likely reflecting the industry’s current focus on proprietary military and commercial platforms where open-source contributions are less directly linked to broader economic outcomes. Additionally, the analysis confirms that labor input and capital formation remain consistent drivers of economic growth. Overall, the findings support the integration of open-source innovation into contemporary economic growth frameworks and provide empirical evidence to guide policymakers in enhancing national innovation capacity and digital competitiveness.

How is productivity growth linked to technology diffusion in individual industries and digitalization? This paper examines the implications of digital transformation on productivity performance in the manufacturing and service-producing sectors, accounting for the technological divergence between frontier and laggard industries. We employ empirical analysis based on industry-level data from the UK over the sample period 1995-2020. To accommodate certain idiosyncrasies and heterogeneity of such industries, we estimate a measure of productivity by controlling for capacity utilization, time varying markup and accounting for externalities between industries. We find that economically significant productivity slowdown is associated with the industry-level divergence. While the productivity growth depends negatively on the speed of divergence, the speed of convergence between the frontier industries and the rest depends positively on digital innovation. Additionally, digitalization has a greater growth effect for industries that are farther from the technological frontier. We also show evidence of convergence in industrial clubs to support the “ripple effect” of digital spillovers. This paper contributes nuanced perspectives on policy formulation and organisational structure reforms for tackling the productivity puzzle.

Public Capital and Labor Productivity in Mexico: A Vector Error Correction Model, 1960-2022. Abstract* This paper investigates the important question of whether there is a long run relationship between the private (public) capital stock and economic output (labor productivity) in Mexico over the 1960-2022 period. Several important findings are uncovered. First, it is established that the underlying private production function exhibits constant returns to scale and that the included regressors are non-stationary in level form (controlling for the presence of a single break). Second, the non-stationary, I(1), variables included in the underlying production function have a stable and long-term (cointegrating) relationship even in the presence of a regime (shift and trend) break. Next, the VECM methodology determined that some of the included variables can be treated as weakly exogenous, viz., the of stock public capital and the labor force. Fourth, both FMOLS and DOLS estimation of the underlying production (labor productivity) function for Mexico is undertaken and the results provide robust evidence in favor of the argument that both the private and public stocks of capital increase economic output and labor productivity over the period under review. The estimates also reveal the lackluster performance of Mexico’s labor productivity over the period in question. The estimators utlized in this study are extremely consistent even in the presence of both endogeneity and serial correlation of any order. From a policy standpoint, these results suggest that across-the-board cuts in public spending to meet targeted fiscal deficits as a proportion of GDP should be made carefully (paying attention to their composition) so as not to undermine public capital (infrastructure) spending which, in turn, may negatively affect long run growth and labor productivity. *Rough draft. Please do not quote without permission. Prepared for presentation at the WEAI Meetings in Denver, CO., June 30th, 2026. Keywords Dynamic Ordinary Least Squares (DOLS), Fully Modified Ordinary Least Squares (FMOLS), economic output, Gregory-Hansen cointegration single-break test, Impulse response function (IRF), public capital stock, Johansen Cointegration test, labor productivity, KPSS no unit root test, single-break (Zivot-Andrews) unit root test,, and vector error correction model (VECM). JEL C22, 040 and 054

In Japan, where the population is rapidly ageing, older adults’ labour force participation has become crucial. However, employers in Japan consider older adults to be lacking adequate skills, possibly compromising their employability; yet, little research has been conducted on how older adults’ skills, particularly digital skills, affect their employment. Inspired by previous research, this study examined the effects of digital skills on employment and wages among older adults (those aged 65 and above). The KHPS/JHPS was employed for analysis, as it is one of the few datasets that clearly defines digital skills. Descriptive statistics revealed that few older adults possess digital skills, and that many are employed in blue-collar jobs. A treatment effect analysis with dual robustness was conducted to clarify the causal relationship among digital skills, employment, and wages. Because the study’s treatment effect analysis could not incorporate unobserved heterogeneity, a fixed effects analysis was performed thereafter. Both analyses yielded similar results regarding the effect of digital skills on white-collar employment. Possessing multiple digital skills significantly increased older adults’ probability of obtaining white-collar jobs by 5–6%, suggesting that improving older adults’ digital skills could broaden their employment options and promote their employment.

This paper examines how gender composition influences leadership selection and team dynamics in collaborative settings. Using a field experiment conducted in a three-day design-thinking program for first-year undergraduates at a U.S. university, participants were randomly assigned within cohorts to teams that varied in gender composition. After an initial period of interaction, each team selected a leader from among its members. The study combines administrative data, peer and self-evaluations, survey measures of individual experience and team dynamics, and objective team performance evaluations by external judges. The research design allows for causal comparisons across team environments and leadership configurations. The analysis contributes to understanding how diversity and informal leadership interact to shape collaboration and performance in team-based work.

The Affordable Care Act’s Medicaid expansion provision allows states to expand coverage to able-bodied adults earning up to 138 percent of the Federal Poverty Level. Initially, only twenty-four states chose to expand Medicaid, with another fifteen states expanding coverage since 2014. Hospitals in rural areas and other proponents of Medicaid expansion in non-expansion states have advocated for expansion, arguing that decreased health care coverage in rural areas makes it more difficult for providers to profitably serve these communities, potentially leading to more hospital closures. Using a staggered difference-in-differences approach on American Community Survey microdata from 2010 to 2021, this paper examines how Medicaid expansion impacts rural healthcare workers’ decision to remain in or migrate out of rural communities. Results suggest that Medicaid expansion increases the likelihood that rural health workers will migrate to urban areas, primarily driven by health care specific occupations in the hospital sector.

Homeownership has long been considered a major component of the American Dream and a desirable policy goal. While a substantial literature examines the benefits of homeownership and the factors underlying homeownership gaps, far less is known about whether economic shocks—particularly positive shocks such as economic booms—generate spillover effects on individuals’ housing tenure decisions. We address this gap by examining how the economic shock induced by the fracking boom affected local homeownership rates. To mitigate concerns about endogeneity, we employ an instrumental variables (IV) approach. Our preliminary results reveal a statistically significant and positive relationship between oil and gas employment and homeownership, indicating that the fracking boom increased the likelihood of homeownership. Further analysis shows that this effect persists beyond the boom period, suggesting that the homeownership effect from the boom is not temporary. We also conduct subsample analyses to explore heterogeneity in homeownership responses across demographic groups. Given the substantial economic gains associated with the fracking boom and the potential benefits of homeownership, our findings highlight the important role of positive economic shocks in shaping housing tenure decisions and carry broader implications for housing and regional development policy.

Households can improve investment outcomes by investing in financial literacy or by seeking professional financial advice, yet relatively little is known about how these two mechanisms interact to shape household portfolio risk. While a large literature studies financial literacy and financial advice separately, fewer studies examine their joint role in determining both stock market participation and the composition of household portfolios, particularly the share of risky assets held in equities. We develop a theoretical model in which households choose whether to invest in financial literacy and whether to seek financial advice under uncertainty about advisor quality and market conditions. In the model, financial literacy directly affects households’ portfolio risk-taking by improving their ability to evaluate risky assets, and indirectly affects outcomes by increasing the likelihood of selecting high-quality advisors. Financial advice, in turn, improves portfolio outcomes only when households possess sufficient financial literacy to assess and appropriately implement advisory recommendations (Calcagno and Monticone, 2015; Ludwig et al., 2023). As a result, as financial literacy advances, it may act as complement for financial advice before functioning as substitutes in improving portfolio efficiency and shaping households’ exposure to financial risk. We test the model’s predictions using data from the 1998–2022 Survey of Consumer Finances. Employing logistic and Tobit regression models, we examine both the extensive margin of stock market participation and the intensive margin of household portfolio risk, measured by the share of equities in total financial assets. Our empirical results indicate that the use of financial advisors is positively associated with stock ownership and higher equity shares, particularly among households with lower levels of financial literacy, consistent with a substitutability relationship in household risk-taking. These findings remain robust across income groups and alternative measures of risk preferences. Overall, the results highlight the importance of jointly considering financial literacy and financial advice when evaluating household portfolio behavior. In line with the Federal Reserve’s emphasis on expanding public financial education, our findings carry important implications for financial education policy, the regulation and design of advisory services, and the promotion of efficient household risk-taking.

We study how generative artificial intelligence (GenAI) affects creative goods markets using data from a large stock images marketplace. In December 2022, the platform announced it would allow artists to sell GenAI-produced images, subject to two conditions: all GenAI images must be labeled, and GenAI would be prohibited in certain markets. We exploit this policy variation using a difference-in-differences design. We estimate a 136% increase in image production and 47% increase in active artists, accompanied by a 15% decline in non-GenAI production and 29% decline in non-GenAI active artists. On the demand side, total sales increase by 82%, but non-GenAI sales fall by 28% and non-GenAI sales rates decline by 5%. We use image embeddings to develop novel measures of image quality, characteristics, and market composition. We find that quality increases by 10% overall and 2% for non-GenAI content, suggesting low-quality non-GenAI artists exit. Our product characteristic measures indicate that GenAI production is heavily concentrated in characteristic space, potentially limiting the value of these images to the market. Non-GenAI production responds to GenAI by differentiating into niches. Our results suggest that GenAI can lead to market expansion and quality improvements, but substantially crowds out non-GenAI production—a finding with direct implications for copyright policy and the longer-run sustainability of original creative work.

While a growing literature studies how firms respond to geopolitical shocks, much less is known about how governments reallocate economic resources across firms with different exposure to rival economies through procurement and fiscal support. In the big-power rivalry through trade war and tech war between the U.S. and China, the U.S. has adopted a de-risking strategy to reduce its economic interdependence with China and control or ban high-tech exports to China. China, on the contrary, has pursued a counter-decoupling strategy to resist decoupling. Against this backdrop, the US government has employed a punitive strategy that reduces support to firms with high exposure to the rival economy, China. By conducting DID analysis before and after 2018, we find that U.S. listed companies with supply chain ties to China, particularly with Chinese customers in various tiers, are penalized by a lower frequency and value of federal government contracts granted. An examination of the ongoing procurement contracts suggests the existence of withdrawal and funding cuts for pre-existing procurement contracts post-2018. The largely indistinguishable differences in the negative effects of input-output linkages to China on securing public procurement between defense and civilian contracts, one-bidder and multiple-bidders scenarios and sole-source and multiple-sources scenarios point to the uniform applications of the tightened screening policies. In contrast, motivated by economic and social stabilization and the maintenance of technological links with the US and the west, the Chinese government has adopted a compensatory and supportive strategy that increases procurement contracting from those domestic companies with a high exposure to the US, especially to US suppliers. At the same time, the government reduces fiscal subsidy to those public contractor firms. This reveals the intention of the Chinese government to sustain Chinese companies with high exposure to the adverse shocks from the rival economy and to stockpile and hoard certain products whose critical components have or are expected to be restricted or banned to Chinese companies. Notably, the Chinese government shifts from providing general subsidies to companies with US supplier links to increase procurement from those companies to support them to continue to produce such products, a targeted subsidy policy.

AI has become a central source of factual information. Despite the deployment of a common global model, trust in AI-provided information may differ substantially across countries. We combine evidence from a large-scale incentivized cross-country experiment with data from three global surveys (World Values Survey, Reuters, and Ipsos) to document cross-country differences and examine their sources. We show that trust in AI-provided information varies markedly across countries and demographic groups. Trust in AI as an information source is strongly associated with trust in traditional information institutions (press, television, and news organizations), but not with generalized social trust or broad attitudes toward AI. These findings imply that trust in AI is shaped by pre-existing information-institution environments, such that the same model can generate different patterns of adoption, reliance, and information risk across countries, with important implications for the design of policies governing AI adoption.

AI is rapidly reshaping financial forecasting, yet the persistence of algorithmic averse behavior and the pitfalls of concentrated trust remain poorly understood. This paper demonstrates that heterogeneous trust in AI-generated signals-not mean acceptance-drives cross-sectional dispersion and polarizes aggregate outcomes, creating a stability externality even when forecasts are accurate. A structural framework shows that individuals combine subjective priors with noisy AI signals, where the trust parameter balances idiosyncratic versus correlated error components. Crucially, the same trust weight that renders individual behavior Bayesian-optimal amplifies aggregate volatility through correlated model errors, generating a wedge between private and social optima. Empirical evidence from a controlled survey experiment with undergraduate economics students reveals stark heterogeneity: risk-averse agents exhibit significantly higher AI trust, while overconfidence acts as a rigid cognitive barrier to belief revision, suppressing updating. Social interaction injects substantial noise, rendering group-level trust measures uninformative compared to individual-level structural estimates. These findings refute the notion of uniform algorithmic appreciation, showing instead that individually rational reliance on AI generates selection bias, dynamic polarization, and systemic fragility. The analysis establishes that widespread adoption of accurate AI can destabilize markets precisely because heterogeneous trust prevents correlated errors from diversifying away, highlighting an urgent need for governance mechanisms that account for user-level behavioral differences rather than average acceptance.

In this paper, we identify level-k reasoning in repeated games that operates at the level of a supergame strategy, rather than at the level of individual rounds. First, we develop a model of level-k reasoning that incorporates choices over strategies as well as beliefs about strategies chosen by others. Then, using data from the Indefinitely Repeated Prisoner's Dilemma that includes elicited strategies and beliefs about strategies, we classify a substantial fraction of subjects as level-1 or level-2. Moreover, we show that when level-k reasoning operates at the level of a strategy, cognitive ability and experience both predict higher level reasoning.

Why can a population exposed to the same information appear stable while remaining locally fragile? This paper shows that neither networks, heterogeneous priors, nor bounded-rationality frictions in aggregation are required for local consensus fragility. In a delayed social-learning model with costly, incomplete, and asynchronous reassessment, mean-only interaction alone generates a sharp instability threshold: below it, directional imbalances are damped; above it, they are amplified. The model therefore identifies a structured reservoir of unresolved adjustment through which a calm average may coexist with a centred benchmark that is close to local instability.

Airport congestion imposes annual costs exceeding $8 billion on U.S. airlines and $17 billion on passengers, motivating billions in capacity investments worldwide (Ball et al., 2010). Yet a critical gap persists between projected and realized benefits: expansion projects rely on simulation models with strong assumptions about demand and behavioral responses, mirroring patterns where ex ante projections overstate net benefits across major public infrastructure investments (Fowlie et al., 2018; Duranton and Turner, 2011). This paper provides the first causal, ex post evaluation of a major U.S. hub runway expansion. Phase One of the O'Hare Modernization Program, completed in 2008, was part of an $8 billion reconfiguration transforming intersecting runways into a parallel system. I employ difference-in-differences with exact flight-level matching using 2003 – 2010 data and matched flights from comparable airports as controls. Three key findings emerge. First, operational improvements were asymmetric: arrival delays fell 4.5 minutes (25 percent) while departure delays showed no significant change. This reflects both the nature of the investment: Phase One introduced a new dedicated arrival runway while the departure-side improvement was merely an extension of an existing runway, and the harder capacity constraint faced by arriving flights (Brueckner, 2009). Second, the expansion produced substantial welfare gains: the arrival improvement corresponds to roughly $200 million in annual passenger time savings, and fare analysis reveals average reductions of $9 per arrival passenger. Third, welfare gains were unevenly distributed, dominant carriers retained partial pricing power, limiting consumer pass-through and raising questions about returns from publicly subsidized infrastructure. These findings contribute to a growing literature on realized returns to public infrastructure investment. Targeted investments addressing binding constraints can yield substantial returns, but complementary instruments, including congestion pricing and air traffic management reforms, are necessary to sustain efficiency gains and ensure benefits are broadly shared.

High temperatures have been found to worsen mental health, but evidence from lowand middle-income countries remains scarce, where adaptive capacity is limited. Using nationwide weekly administrative data covering the entire population and a survey series representative of urban Mexico, we estimate effects of heat on self-reported well-being, emergency department visits for mental illness, and suicide using a unified empirical design. Comparing hotter and cooler weeks across years within the same locations and calendar week, we find that one extra day above 30'C (86'F) increases suicides by 5% and emergency department visits for mental illness by 2.4%, and lowers reported mental well-being. The effect sizes exceed comparable U.S. estimates but are attenuated in municipalities served by mental health workers, suggesting protection through access to care. Gender patterns differ from U.S. studies: women experience larger emergency department visit increases, while suicide effects are concentrated among men. Our results indicate that heat imposes mental health burdens from reduced well-being to clinical demand and mortality, highlighting mental health as a key component of climate damages and a priority for adaptation investments in health systems.

Forthcoming

This article estimates short-run employment effects of California’s sectoral minimum wage for fast-food restaurants. On 1 April 2024, covered limited-service chains faced a $20 hourly floor while the statewide minimum wage was $16 (up from $15.50 in 2023). Using establishment-level mobility data and a difference-in-differences design restricted to the counties without local wage ordinances in CA, we proxy on-site staffing with weekly long-duration device visits (>4 hours). Baseline estimates indicate an average 8% decline in on-site staffing intensity at treated outlets relative to comparable full-service and retail establishments. An event-study shows no pre-trends and a gradual, persistent post-announcement decline, consistent with anticipatory staffing adjustments. Placebo estimates for exempt ‘enclosed’ venues are null, and effects are similar across urban and rural locations. The results imply that a sectoral wage floor induces meaningful reductions in on-site staffing intensity at continuing establishments, with adjustments beginning before implementation rather than through immediate large-scale workforce changes.

On April 1, 2024, California implemented a $20 hourly wage floor for workers in large chains in fast-food restaurants and snack and non-alcoholic beverage bars. The new standard, which corresponds to 69 percent of the state’s median full-time wage, surpasses all prior benchmarks in minimum wage policies and research. We use DiD and DDD event study methods and granular data that permit distinguishing covered from uncovered restaurant establishments. Our granular pay data come from Glassdoor job postings and Square payroll data, granular mobility-based employment data from Advan, and granular prices scraped from over 2,000 restaurants in California and control states. We find that the policy increased average weekly wages for covered fast food workers by about 11 percent and did not reduce employment. Compared to controls, prices increased by 1.5 percent, equivalent to 6 cents for a $4 item. Employers passed about 50 percent of the higher wage costs to consumers as higher prices, consistent with a monopsony model.

This paper provides a simple user's guide to synthetic control methods and makes the case that synthetic control methods are almost always superior to traditional parametric difference-in-difference estimators which almost always violate the parallel trends assumption. The paper provides several examples in one-shot, staggered adoption, panel settings. We provide some examples of difference-in-difference results versus synthetic control results including the effect of California's $20 minimum wage on prices, the effect of Medicaid expansion on mortality and the effect of Brexit on GDP.

We study the employment effects of California’s fast-food minimum wage, which raised the wage floor to $20 per hour for large fast-food chains beginning in April 2024. A key challenge in evaluating this policy is that statutory coverage depends on firm size and cannot be cleanly addressed using commonly used aggregate datasets that rely on industry classifications. Using establishment-level data that allow us to identify establishments subject to the law’s coverage rule, we estimate the impact of the policy on employment at covered restaurants, as well as on other margins of adjustment. Our research highlights the importance of statute-consistent treatment measurement in evaluations of sector-specific minimum wage policies.

We study how deposit rate shocks transmit across banking markets through digital social ties. Depositors’ inattention implies that households react to outside rate changes only when social networks make these changes salient, inducing local banks to raise their own rates. Using merger-driven shocks to local deposit rates and county-level social connectedness, we show that small banks increase rates in response to shocks occurring in socially linked but geographically distant counties. Spillovers are economically meaningful, persistent, and stronger in competitive markets and in counties with more financially sophisticated households. The findings suggest that digital social ties stimulate depositor search and reduce the spatial segmentation of deposit markets.

Forthcoming

We study the incidence and macroeconomic consequences of wholesale funding crises using a novel narrative chronology and newly constructed data on banks' liability structures for 147 countries from 1800 to 2024. Wholesale crises have occurred with unprecedented frequency over the past three decades and have become more common than retail bank runs, especially in advanced economies. We show that wholesale crises are predictable using observable fundamentals such as non-deposit funding growth and bank capital ratios. The macroeconomic fallout from wholesale funding crises is severe when accompanied by contractions in non-deposit funding, and can exceed the costs of retail deposit runs, consistent with wholesale creditors possessing superior information about bank solvency. While deposit insurance reduces the likelihood of retail bank runs, it increases the probability of wholesale funding crises, indicating that run risk is reallocated from insured retail depositors to uninsured wholesale creditors.

We estimate zip-level “housing cost sensitivities” (HCS)—the percent change in local home prices or rents per 1% increase in national home prices caused by Ben-David, Towbin, and Weber (2024) mortgage-rate shocks. Using Zillow home-value and rent indices for over 13,000 zip codes, we find wide heterogeneity. When mortgage-rate shocks raise national home prices by 1%, local home-price HCS span from 0.3 to 2.0% (IQR), while rent HCS range from –1.5% to +1.35%. We find the HCS are strongly regressive: lower-income zips exhibit substantially larger responses for both prices and rents. Zip-level home-price HCS correlates closely with the city-level inverse supply elasticity of Guren et al. (2021), yet meaningful unexplained variation remains. Beyond supply constraints, we find that demand amplification (natural amenities), land-cost pass-through, and housing-market liquidity all shape HCS. Our results quantify where mortgage-rate movements most strongly transmit into local housing costs and clarify the mechanisms underlying that heterogeneity.

This paper studies how an exogenous expansion in access to homeownership shaped the geography of economic opportunity in the U.S. in the 20th century. We examine the rise of mail-order kit housing in the United States between 1910 and 1930, an innovation that lowered construction costs, standardized designs, and tied housing to railroads. Using newly digitized data on kit-housing firms’ mill locations and full-count census records, we exploit spatial variation in transportation costs as a source of exogenous variation in town-level homeownership access. We estimate the contemporaneous effects of increased homeownership on residential tenure, educational attainment, and housing values, and examine long-run impacts on intergenerational educational and occupational mobility. To complement the town-level analysis, we link novel person-level purchase records from a leading kit-housing firm to census data to study the characteristics of adopters and validate our identification strategy.

We study lender market power in United States equity securities lending and its implications for market efficiency, measured by adjusted spreads, market illiquidity, and transaction costs. Using lending data, we construct security level measures of lender market power and loan scarcity, where scarcity is an equilibrium outcome that reflects both short selling demand and the willingness of lenders to release inventory. High lender market power constrains short sellers and lowers the elasticity of loan supply, and these supply side shocks are reflected in higher lending fees and worse trading outcomes in the underlying shares. Both unconditionally and after conditioning on proxies for short selling pressure, higher lender market power is associated with wider adjusted spreads, greater illiquidity, and higher lending fees. The findings indicate that lender market power impairs market efficiency and motivate greater transparency and oversight of lender concentration and fee schedules in securities lending.

I analyze a multi-period call market with informed traders, liquidity traders, and a market maker. Informed traders receive heterogeneous, noisy estimates of the fundamental value of the traded asset. Informed traders specialize in knowledge acquisition, resulting in an information structure previously unstudied in the literature. I investigate the relationship between knowledge specialization, trading strategies, and the informational content of market prices. I discuss what my results imply regarding the informational efficiency of relevant real-world financial markets.

This paper studies whether oil-market narratives have macroeconomic effects, and whether those effects differ depending on whether narratives convey directional information or uncertainty. I construct new monthly oil narrative sentiment and uncertainty indices from Wall Street Journal and New York Times articles from 1953 to 2022 and classify oil-related coverage into five narrative types. Using the typespecific narrative decomposition as external instruments in a Proxy-SVAR, I identify two structurally distinct narrative shocks. A bullish narrative sentiment shock raises the real oil price persistently, increases world industrial production, raises world oil production, and gradually passes through to consumer prices, consistent with a demand-side information channel. By contrast, an elevated narrative uncertainty shock reduces world oil production and world industrial production, while producing only a muted response of the real oil price and consumer prices, consistent with a supply-side real options channel. When both shocks hit simultaneously, the combined effect produces a stagflationary pattern in which oil prices and inflation rise while production and activity respond weakly. These findings show that oil-market narratives contain macroeconomically relevant variation beyond physical oil shocks and that separating directional tone from ambiguity is essential for understanding how oil news transmits to the broader economy.

Prediction markets have rapidly expanded beyond political forecasting into financial domains, yet whether their prices conform to established pricing theory remains an open question. This paper benchmarks prediction-market prices on Kalshi and Polymarket against Black–Scholes binary-option probabilities across four asset classes: cryptocurrencies (Bitcoin, Ethereum), major equity indexes (S&P 500, NASDAQ-100), gold, and eggs. Using a unified within-contract Δlogit specification, I find that all eight asset–exchange specifications maintain a long-run equilibrium with their no-arbitrage benchmarks, confirming that prediction-market prices are systematically anchored to financial theory rather than driven by sentiment alone. Markets nonetheless under-react to benchmark signals, and efficiency varies sharply across asset classes: cryptocurrency markets track their Black–Scholes benchmarks most closely (β_MZ ≈ 0.72 to 0.85), followed by gold (≈ 0.37), while long-dated equity-index and egg contracts are the most sluggish (≈ 0.04 to 0.10). Every long-dated market converges toward efficiency as settlement approaches. A threshold error-correction analysis shows that equity-index markets correct large mispricings roughly three to five times faster than small ones, and the SETAR optimal-threshold estimator confirms significant asymmetry for six of eight assets. A new triangular analysis that prices Kalshi S&P 500 bucket contracts against an actual listed-option condor spread shows that the listed-option spread, not the Black–Scholes formula, is the binding benchmark: Kalshi cointegrates with the condor but under-reacts to it on the trade date and closes the gap only over about three days, leaving a narrow but real convergence trade on roughly one S&P trading day in eight. These findings suggest prediction markets serve a legitimate price-discovery function whose efficiency is shaped by time horizon, derivatives-market depth, and contract design.

We study a monopolist’s choice between screening consumers through price mechanisms (two-part tariffs) and quantity mechanisms (menus of bundles). Our main result shows that the firm strictly prefers quantity mechanisms, except when the type distribution has an increasing hazard rate, in which case the two screening mechanisms are equivalent. While quantity mechanisms maximize profits, price mechanisms always deliver both greater consumer surplus and higher total welfare. The divergence between the two mechanisms can be stark. Under a decreasing hazard rate, the optimal quantity mechanism produces partial or full separation across types, while the optimal price mechanism collapses to a single two-part tariff. Yet, under certain conditions, this simple two-part tariff can replicate the outcomes of a fully separating quantity mechanism. These findings suggest a novel policy lever for regulating market power: mandating firms to employ two-part tariffs in place of more complex nonlinear pricing schemes.

We study price competition where consumers are uniformly distributed along a circle representing product varieties. Firms pay a fixed cost to enter the market, then produce a product variety represented by their location on the circle at a private per unit cost. The demand for a firm’s good is determined by the prices firms set simultaneously. In this setting we show that under a standard regularity condition a unique equilibrium exists: firms enter if their costs are below a threshold, then set prices based on own costs, expected rival costs, and the number of entrants. We show that the threshold is inversely related to entry costs and the number of potential entrants, and average prices may fall as the threshold rises. Finally, we provide closed form expressions for various welfare measures in the market that help identify the impact of private information. Interestingly, cost variance positively affects both consumer and producer surplus. While we provide sufficient conditions for positive entry restrictions to be welfare improving, we also show that for many common families of cost distributions the sufficient conditions do not hold and in fact is optimal from both consumer and producer perspective to set zero entry restrictions.

Auction researchers have made remarkable progress in developing and applying structural econometric models to interpret real world data. A structural approach allows for estimation of latent (unobserved) characteristics that are critical to mechanism design, policy work, and evaluating the efficacy of an auction setting. One ubiquitous challenge of course is whether the right assumptions have been maintained in the model. If the model is wrong, the estimates are biased and therefore not helpful. Typically a researcher tries to hedge this concern by justifying why assumptions are maintained, often appealing to the nature of the object or structure of the market, but these claims typically are not supported formally. We fill this gap by developing an econometric test of something fundamental: The informational paradigm that a research maintains. We provide a test that directly uses observed bid data meaning it can be used immediately to help guide (or support) the environment a researcher maintains when modeling a real-world setting. Our approach facilitates a data-driven way of helping guide assumptions. We show that our test performs well in simulated data, extends to a wide array of more complicated settings, and then apply it to a number of real-world data sets finding results that often complement justifications suggested by previous researchers.

This study develops a monopolistic-competition framework of multi-product firms (MPFs) in which branding shapes consumer preferences and the balance between within-firm and market-level competition. We examine the effects of trade integration on firm behavior and market outcomes under homogeneous and heterogeneous firms. In the homogeneous-firm benchmark, trade generates relatively symmetric expansion in MPFs’ R&D for product innovation and brand appeal, while intensified competition gradually discourages entry, yielding an inverse-U relationship between market size and the mass of firms. By contrast, under firm heterogeneity, trade induces asymmetric responses: large brands leverage their underlying advantages to expand, whereas small brands progressively contract under intensified competition. As a result, the widening gap between aggregate market output and average firm supply allows the mass of firms to continue increasing as markets expand. More broadly, the analysis illustrates how firm heterogeneity reshapes the trade–competition relationship, generating a dual competitive structure in which dominant firms expand brand families while continuing to compete alongside a persistent fringe of small firms.

This paper examines the trade effects of EU-imposed sanctions using a staggered DID gravity framework. While existing studies primarily rely on conventional gravity models or standard DID estimators, this study explicitly accounts for the staggered timing of sanction episodes and heterogeneous treatment effects across country pairs. Using sanction data from the Global Sanctions DataBase (GSDB) and bilateral trade data from the UN Commodity Trade Statistics Database, we find that the average effect of EU sanctions is limited when all sanction episodes are pooled together. However, the results reveal substantial heterogeneity across sanction types. Trade sanctions consistently reduce trade flows under both conventional TWFE and extended TWFE specifications, whereas travel, financial, and military and arms sanctions do not show robust average effects. Dynamic event-study estimates further indicate that trade sanctions generate a persistent decline in trade flows after implementation. These findings suggest that the trade effects of sanctions are concentrated in instruments that directly restrict commercial exchange and underscore the importance of distinguishing among sanction types when evaluating the economic consequences of sanctions.

This paper studies the role of supply chain reallocations in the transmission of trade shocks. I introduce a global sourcing framework in a multi-country, multi-sector general equilibrium environment, with both variable and fixed costs of sourcing inputs internationally. The endogenous responses of supply chains reflect producers' sourcing location choices and entry decisions. I then calibrate the model to the World Input-Output Database and quantify the trade and welfare consequences of two hypothetical trade policy changes: the US-China trade conflict and the effects of preferential trade agreements. The quantitative analysis reveals two novel global value chain (GVC) transmission channels. First, allowing endogenous supply chain reallocations amplifies the trade and welfare consequences of shocks to variable trade costs. Second, changes in fixed sourcing costs are essential in welfare evaluation and could generate a 2-3 times larger impact than similar changes in variable trade costs.

We provide first causal evidence of the effect of foreign Technical Barriers to Trade (TBTs) on the local environment in a developed country. We combine data on TBT adoption by US’ export destinations, bilateral US exports, and local industrial composition in a shift-share design to estimate the effect of exposure to foreign TBTs on regional pollution emissions in the US. We estimate that a 1 standard deviation (s.d.) increase in TBT treatment leads to 0.14, 0.05, and 0.13 s.d. increase in SO2, PM10, and VOC emissions growth, respectively. These effects are stronger when the foreign TBTs are unharmonized with US standards, mandate considerable production adjustments, and are subject to observable compliance. Unlike TBTs, exposure to foreign tariffs reduces local emissions growth by directly restricting trade flows. Our findings highlight the environmental consequences of non-traditional trade policy instruments for policymakers seeking to balance openness, standardization, and sustainability.

G7 countries accounted for approximately 53% of the world's total petroleum consumption; thus, any shocks in the oil market pose considerable economic implications. To this end, we investigate the response of G7 real effective exchange rates (REER) to oil supply, demand, and risk shocks, considering various economic circumstances. Given the high fluctuations of oil markets over our sample period (from February 28, 2013, to March 14, 2023), we apply the cross-quantilogram approach to measure the fat-tailed linkage between oil shocks and REER. Our results demonstrate that G7 REERs are highly exposed to oil demand shocks while being least exposed to oil supply shocks. The REER of Canada encounters a depreciation in response to the supply and risk shocks, while that of the UK experiences a depreciation to supply and demand shocks in the long memory. Additionally, we find that REERs of Italy, Japan, the US, and Germany are less exposed to oil market shocks. Our findings are robust considering the rolling-window-based cross-quantilogram that accounts for bearish, normal, and bullish market conditions.

Batteries have emerged as a potential solution to the intermittency of renewable generation. However, batteries are distinct from traditional generation sources due to their role as arbitrageurs; this makes analyzing their behavior challenging, as their willingness to charge and discharge is driven by their beliefs over the future value of stored energy. To understand the strategic nature of battery behavior, we look to the Texas electricity market, which has seen a rapid expansion in battery capacity from just over 3 GW in 2022 to over 12 GW by the second half of 2025. First, using novel, high-frequency data from the Electric Reliability Council of Texas spanning 2023–2025, we recover the dynamic marginal costs that rationalize the real-time bidding behavior of battery operators. To accomplish this, we extend the seminal supply-function framework of Hortaçsu and Puller (2008) to incorporate the dynamic decision-making of grid-scale battery storage. We find that recovered marginal costs track closely with afternoon peak prices, consistent with battery operators arbitraging price spreads between low and high demand periods. At their core, these estimates of marginal costs allow us to uncover the strategic nature of operator behavior. Our approach also offers regulatory value: because market-power mitigation in electricity markets traditionally relies on cost-based tests, our method provides a transparent, data-driven way to infer the underlying costs that batteries reveal through their bids. Next, we use our model of optimal storage behavior to construct two benchmarks: a perfectly competitive benchmark, and a strategic operator benchmark. We then ask: "When batteries fail to arbitrage efficiently, how much are these deviations driven by strategic incentives versus uncertainty or unsophistication?" We use the gaps between our benchmarks and realized outcomes to separately identify these two mechanisms. We use our approach to explore heterogeneity across firms. We document behavioral differences that vary by firm size and location. Lastly, we explore the role that independent optimizers –firms that manage a portfolio of separately-owned batteries– play in shaping competition between firms.

The energy transition from a carbon-intensive world to a low-carbon one depends to a large extent on the availability of certain critical minerals like lithium, nickel, cobalt, and graphite. These minerals are essential for the manufacture of clean energy technologies such as lithium-ion batteries for electric vehicles. During the 1980s, advances in such clean energy technologies were concentrated in Japan and the U.S., while China was a relatively minor player. Today, China is the leader in clean energy technologies, controlling the vast majority of critical mineral supply chains required for the energy transition. How did China transform its challenges into opportunities for developing clean energy and critical mineral industries? Drawing on historical state records and official planning documents dating back to the 1980s and data on state funding and innovation, I find that state industrial policies have been crucial behind China’s success story. I compare these data with those from Japan and the U.S. to show China ‘catching up’ relative to its competitors. I also use a novel dataset of mining-related acquisition deals from 2010 to 2023 and analyze the nature of mergers and acquisitions in the lithium, nickel, and cobalt industries. The findings reveal systematic tactics used by Chinese corporations for securing mineral supply chains, including premium bidding, majority stake acquisitions, loan forgiveness, and investments linked to the Belt and Road Initiative. Using industry annual reports of selected mining corporations, I document the extent of state funding that has enabled these corporations to advance these tactics to dominate the mineral supply chains beyond China’s domestic border. The paper concludes with policy implications for advancing the clean energy transition and underscores the pivotal role of the state in this process.

Government ownership has been extensively studied in financial economics, particularly concerning privatizations and government interventions. Many studies examine the effects of government ownership on firm performance and find that partial state ownership can enhance performance by improving sales, profits, firm value, and productivity (Boubakri et al., 2005, 2008; Dyck & Zingales, 2004; Gupta, 2005). However, studies also highlight the drawbacks of government ownership. Recent research indicates that government ownership can suppress CEO compensation and incentives, thereby narrowing the pay gap between CEOs and employees (e.g., Hui & Fang, 2023; Mengistae & Xu, 2004). While a reduced pay gap may lower agency costs by curbing excessive executive compensation (Jensen & Murphy, 1990), it could also have unintended consequences. In competitive labor markets, lower compensation may demotivate executives or increase CEO turnover, which is often associated with subsequent poor performance (e.g., Fee & Hadlock, 2004). Thus, reductions in CEO pay under government control could lead to leadership instability and deter long-term strategic investments. This study aims to explore how government control influences the CEO-employee pay gap, specifically whether this effect is mainly driven by lower CEO compensation rather than higher employee wages. Additionally, we investigate how pay compression affects CEO turnover and its impact on firm performance, agency costs, and operational efficiency. Using a dataset of 6,769 firm-year observations from 30 countries spanning from 2002 to 2019, we analyze the mechanisms by which suppressed CEO compensation affects firm decision-making under government control. Our initial analyses reveal that firms under government control exhibit a significantly narrower CEO-employee pay gap—over 3,000% lower—primarily due to reductions in CEO compensation rather than increases in employee wages. This pay compression correlates with increased managerial instability, as CEOs in government-controlled firms face a 45.8% higher likelihood of turnover. Therefore, we anticipate that the interaction between government control and the standardized pay ratio is significantly associated with lower Tobin’s Q, reduced capital expenditures relative to assets, greater investment inefficiency, higher agency costs, and lower return on assets.

The fact that CEOs (and top management executives) display heterogeneous characteristics with implications for corporate policy choices and firm outcomes traces back to the seminal work of Hambrick and Mason (1984), which launched upper echelons theory. The role of CEO tenure was highlighted by Hambrick and Fukutomi (1991), who theorized that firm performance would display a hump (inverted U) shape related to CEO tenure. Various studies have also examined the effect of CEO age on corporate policies and outcomes. For example, Serfling (2014) finds that risk-taking behavior decreases with age. Belenzon et al (2019) find that as a CEO grows older, firm investment, growth, and profitability decline, but probability of survival increases. However, CEO tenure and age are inherently correlated (older CEOs tend to have longer tenures) making it difficult to disentangle the effects of firm-specific experience (tenure) from life-cycle effects (age). In this paper, we attempt to distinguish between the CEO tenure and age effects and study the roles they play in shaping management style in the domains of investment, financial, and strategic policy choices as well as firm performance. We study a cross-section of CEOs of U.S. companies in place in 2023 and classify them by their birth years as either belonging to the Baby Boomer or Gen X cohorts. Within each cohort we create three tenure buckets (less than five years; between five and ten years; more than ten years). To capture tenure effects, we focus on comparing CEOs with low tenures (less than five years) to those with high tenures (more than ten years), as these categories reflect distinct contrasts in firm-specific leadership experience. We calculate a total of twelve style variables, which are essentially accounting/financial and performance ratios. For comparison purposes, we use propensity score matching to create samples matched by CEO gender, firm size, and industry sector. CEO birth cohorts display some interesting distinguishing characteristics, but our key finding is that firm-specific CEO tenure remains the primary driver in determining management style.

The rising prominence of Environmental, Social, and Governance (ESG) factors in investment decisions necessitate a deeper understanding of the mechanisms driving corporate sustainability performance. This study addresses a critical gap by investigating the impact of Common Institutional Ownership (CIO)—where a single institutional investor holds shares in multiple competing firms—on the ESG performance of Taiwanese listed companies over the period 2016 to 2024. The study employs robust panel data regression techniques to analyze the relationship between CIO and firms’ overall ESG scores, as well as their individual Environmental (E), Social (S), and Governance (G) pillar scores. The analysis yields two significant and nuanced findings with substantial implications for investors, regulators, and corporate boards. First, the study finds a positive and statistically significant association between an increase in CIO and a higher overall ESG score. This effect is primarily driven by substantial improvements in the Environmental and Social pillars. This suggests that shared institutional owners may actively influence portfolio firms toward greater sustainability efforts and stakeholder consideration. Crucially, this positive influence is counterbalanced by a negative association between CIO and the Governance (G) pillar. This finding signals a potential governance blind spot, suggesting that institutional investors may overlook or deprioritize firm-specific, traditional governance issues when their focus is primarily on cross-firm ESG factors relevant to systemic risk mitigation. Second, the research establishes that the relationship between CIO and corporate ESG performance is heterogeneously moderated by specific firm structures. Key moderators include family operation status, the presence of CEO duality, and board structure. This result underscores that the mechanism through which CIO influences corporate behavior is context-dependent. For example, the effect of CIO may be weaker in strongly family-controlled firms due to entrenched power structures, or stronger where CEO duality is present, offering institutional investors a clearer target for concentrated pressure. Therefore, investors and regulators must adopt a tailored and structure-aware approach to leverage the positive effects of CIO while mitigating its potential adverse impact on robust corporate governance.

Quantitative courses such as econometrics are central to developing students’ capacity for empirical research and data-driven reasoning. However, many students enter these courses with substantial gaps in prerequisite knowledge -- particularly in probability, statistics, and basic quantitative skills -- which hinder their ability to interpret models, understand assumptions, and apply econometric methods. This paper proposes a redesign of econometrics instruction that combines diagnostic assessment, structured remediation, and the pedagogically guided integration of generative artificial intelligence (GenAI). The redesign begins with a diagnostic pre-test that measures students’ baseline proficiency in core statistical concepts and identifies individual learning deficits. The pre-test is administered prior to two midterms and a final exam in econometrics-related courses. Based on diagnostic results, students receive targeted remediation supported by AI tools such as ChatGPT, Gemini, and Copilot, which provide personalized explanations, step-by-step derivations, and simulated practice problems. Instruction is reorganized around a Problem-Based Learning (PBL) framework, introducing econometric techniques through concrete empirical research questions. This approach shifts emphasis from abstract formulas to identification strategies, interpretation, and causal reasoning. To evaluate the effectiveness of AI-supported learning, the project integrates multiple data sources, including post-tests, administrative academic records, learning-platform activity logs, and a detailed AI-usage survey. Learning outcomes are analyzed using econometric models that account for heterogeneity in student motivation, prior preparation, and AI-usage patterns. The empirical strategy incorporates background controls, lag structures to mitigate reverse causality, and instrumental variable approaches to address endogeneity in AI usage. Preliminary evidence from the first implementation -- based on data from approximately 30–80 undergraduate and graduate students across different course levels -- reveals substantial heterogeneity in baseline statistical knowledge, with large deficits in probability and model interpretation. Using linked pre- and post-test data, administrative records, platform logs, and AI-usage surveys, we find that students who actively engage in AI-supported remediation exhibit larger gains in conceptual understanding than observationally similar peers. These gains are concentrated among students who use AI to clarify reasoning and verify derivations rather than obtain direct answers. Overall, the results suggest that diagnostically guided and pedagogically structured AI integration can enhance learning in econometrics, especially for students with weaker initial preparation.

Abstract The educational value of generative artificial intelligence (AI) remains contested. While AI is often promoted as a transformative learning technology, existing evidence regarding its effectiveness remains mixed. A key unresolved question is whether AI primarily enhances learning outcomes directly or influences learning through changes in students’ behavioral responses to educational technology. This study investigates the causal effects of a lowintensity AI-assisted instructional intervention among elementary school students using a randomized controlled trial (RCT). Building on the concept of learning friction, we conceptualize AI as a technology that lowers operational and cognitive barriers to learning rather than directly transmitting knowledge. To address treatment non-compliance and identify causal effects among actual users, we employ instrumental variable (IV) estimation and estimate Local Average Treatment Effects (LATE) on three dimensions of student outcomes: cognitive performance, learning motivation, and adaptability to new technologies. The results provide little evidence that low-intensity AI exposure improves cognitive achievement or learning motivation. In contrast, AI exposure generates a statistically significant and substantively meaningful increase in the Adaptability Index, corresponding to approximately 0.4–0.5 standard deviations. These findings suggest that the primary educational value of low-intensity AI interventions may not lie in immediate academic gains, but rather in strengthening students’ ability to engage with, adjust to, and learn from emerging technologies. This study contributes to the growing literature on AI in education by introducing a learning-friction perspective and demonstrating that even modest AI interventions can shape important non-cognitive outcomes. The findings further suggest that adaptability may represent an overlooked channel through which AI influences long-term learning and human capital development.

Graded examination scripts often contain much information that could inform student learning, yet this information may never reach students. This is either because institutional regulations do not allow access to exam scripts (as in the case of final exams in the UK) or because students tend to not retrieve scripts when they already know their grades. What would students learn if they retrieved exam scripts that would otherwise remain uncollected? This question is investigated with a four-year field experiment with first year university students. In this experiment students receive their midterm exam grades but only receive detailed feedback by retrieving their script. A randomly selected group of students receives reminders to retrieve their script and suggestions on how to use it for revision for the final exam. We find that the effects of retrieving exam scripts vary with ability. Students with the highest university entry scores benefit substantially from retrieving their scripts while the benefits to students with lower entry scores are at best small and not statistically significant. These results can be explained by two key features of detailed feedback: that at least some prior study is necessary for it to be minimally effective and that it becomes more effective with better academic preparation. Because higher ability students tend to have better academic preparation, they benefit more from it.

Skills such as curiosity and critical thinking empower students to explore ideas and adapt to change, yet fostering these skills in low-capacity schools remains challenging. Moreover, though examples of successful pedagogical innovations in developing countries exist, it is not always clear why others may fail. This study offers evidence on both fronts to address a central policy question: should pedagogical reforms build internal teacher capacity or rely on external agents? We evaluate two delivery models for curiosity-based science education through a randomized trial spanning 150 government middle schools and 7,185 students in rural Uttar Pradesh, India. We compare teacher training—where science teachers receive intensive workshops and ongoing support to integrate inquiry-driven pedagogy—against external instruction, where specialized facilitators deliver the curriculum directly to students. Both interventions use identical content aligned with the state syllabus and grounded in cognitive science research on curiosity. The design emulates a realistic policy choice between investing in teacher capacity, which could offer long-term benefits, or outsourcing to external instructors, who may be less resistant to changes in the curriculum. We measure impacts on curiosity, critical thinking, growth mindset, and academic performance. We additionally track classroom practices through classroom observations as well as student and teacher self-reports. We situate this comparison within the realities of public education systems characterized by workload constraints, hierarchical structures, and limited instructional support, providing evidence for scalable approaches to strengthening pedagogy in low-capacity settings.

Motivated by the observation of declining foreign direct investment (FDI) inflows into China and increasing FDI exit (foreign divestment) in recent years, we explore whether and how the US-China trade and tech war and geopolitical tension have cast a shadow over the business environment in China and prompted global multinational enterprises (MNEs) to reduce FDI activities in China. Exploiting the shock generated by the US-China competition (trade war and tech war), which is largely exogenous to global MNEs, we investigate various channel categories that contribute to MNEs’ readjustment in terms of FDI entry and foreign divestiture. We analyse several widely conjectured channel categories, which include the direct effects of tariff hikes in raising production costs and the indirect effects encompassing heightened political risk, global value chain (GVC) or global supply chain (GSC) restructuring, the “race-to-the-bottom” “involutionary competition” in the Chinese market, legal and regulatory uncertainty, and anti-foreign sentiment. In addition, we also consider some longer-term structural changes such as rising costs of factors of production, the phasing out of preferential tax advantage for foreign companies, tightened environmental regulation, and China’s economic slowdown and weak consumption growth. We gauge the firm-level exposure to US-China competition risk and its various channel categories through two avenues. One is to conduct textual analysis of corporate earnings conference call transcripts, and the other is to construct proxy metrics of various risk channels from actual data. The textual data-based risk measurement allows us to understand global MNEs’ perceptions of and concerns over China-specific business environment risks and identify which channel categories generate most pronounced concerns for an individual MNE. The actual data-based risk measures help us compare MNEs’ articulated risk exposure and perception with the intensity of the risk they are expected to face that is inferred from the proxy measures, which will complement the text data-based measures or provide an alternative measurement. We further conduct firm-level and country-industry-level analysis to investigate how global MNEs readjust FDI entry and exit related to China amid the US-China competition.

Although self-defense is the most commonly cited reason for gun purchases, there is little causal evidence that gun demand responds to changes in crime rates. We study the effect of crime on gun demand in three ways. First, we examine large, persistent changes in crime stemming from 2009 police hiring grants in U.S. cities. We use grant receipt as an instrument to estimate the effect of crime on firearm demand, employing three different measures of local gun demand. Across all measures, we find no evidence that crime affects firearm demand. Second, we analyze the effects of massive, short-run spikes in local crime. Specifically, we identify crime shocks over two-week intervals for Massachusetts towns. Using detailed data on gun sales and concealed-carry weapon (CCW) permits, we find that gun purchases increase following a local crime spike. However, these spikes are extremely rare and the effects are modest, with an implied elasticity of gun sales with respect to crime ranging from 0.1 to 0.26. CCW permit applications do not respond, possibly because obtaining a permit requires additional steps. Third, we consider how firearm demand responds to shocks that increase fears of future gun control. County-level responses in demand are primarily related to political partisanship and are uncorrelated with local crime.

This paper conceptualizes the deterrent effect of capital punishment as estimating a “demand for murder” where price is the likelihood of being apprehended and punished (specifically, executed) for committing a murder, and quantity the number of murders. The objective is not to estimate an accurate and universally applicable measure of the deterrent effect of capital punishment, but rather to provide an instructional example showing: i) relevance of economic models and techniques for examining real world issues; ii) how different assumptions and/or data can provide different results; iii) assumptions which underly the basic supply and demand model; and iv) the distinction between small and big critical thinking as presented by Siegfried and Colander (2022). The introduction considers issues related to teaching economics. Section 2 reviews selected literature which present differing views of the use of capital punishment and it’s deterrent effect. Section 3 empirically estimates demand curves for murder using time-series and then cross-section data; and demonstrates the difficulty of trying to estimate the deterrent effect of execution. Section 4 performs an even simpler analysis of time series and cross section data, deriving an even wider range of estimates. The conclusion summarizes results and examines the appropriateness of this example for presenting fundamental concepts in introductory economics courses.

This paper aims to study the CNY/USD exchange rate deviation from its central parity since the exchange regime reform of August 2015 in China. Relying on the self-exciting threshold autoregressive (SETAR) model, it attempts to identify the band of fluctuation beyond which the People’s bank of China (PBOC, Chinese central bank) starts to intervene in the foreign exchange market to stop further fluctuations. The results shows that the single identified threshold is lower than the announced official one, implying an earlier one-sided intervention in foreign exchange markets.

The 2023 banking crisis, precipitated by the collapse of Silicon Valley Bank, exposed a systemic failure in interest rate risk management following the Federal Reserve’s rapid monetary tightening. This paper investigates the divergence in equity performance between traditional regional banks (proxied by the KRE ETF) and diversified financial institutions (proxied by the XLF ETF) during the 2022–2023 cycle to isolate the market value of duration hedging and balance sheet agility. Utilizing a multiple time-series regression framework that controls for both long-term duration shocks (10-Year Treasury Yield) and short-term funding costs (2-Year Treasury Yield), the study demonstrates that unhedged regional banks fell victim to a macroeconomic "double-edged sword," suffering severe, simultaneous penalties from unrealized portfolio losses (AOCI) and rapid spikes in deposit-funding costs. Conversely, diversified institutions that actively neutralized mismatches exhibited a quantifiable "Agility Premium," mitigating the penalty of interest rate shocks by 60% to 70% and effectively decoupling their equity returns from catastrophic volatility. Furthermore, a regime-switching analysis reveals that this structural divergence was severely amplified by investor psychology. The market transitioned violently from salience bias in 2022—fundamentally ignoring accumulating unrealized losses—to herd behavior and contagion in 2023, resulting in a systemic breakdown of rational asset pricing. Ultimately, these findings challenge the traditional view of hedging as a mere regulatory compliance cost, demonstrating instead that effective duration management serves as a primary driver of alpha and a critical defense against both mathematical interest-rate risk and irrational market panic.

The background for IPO success research is based on the underpricing puzzle (rooted in asymmetric information between informed and uninformed investors) and long-term IPO underperformance (compared to market benchmarks). Our purpose is to determine if hedge fund performance can explain long-run IPO success. This fills in two research gaps. First, it extends the recent short-run research that finds hedge funds are significantly asso-ciated with IPO success. Second, it addresses scarcity in long-run IPO success research by testing periods longer than five years. We use the same success measure and a similar re-gression model as short-run research with regression variables adapted to each long-run period. This measure has a different interpretation when applied to long-run tests as mis-pricing is no longer the major consideration but is replaced by long-run IPO price perfor-mance. We offer three new findings. First, long-run IPO success is poor when hedge funds perform well. Second, in contrast to hedge funds results, IPO success is greater when mar-kets perform well. Third, the Vuong test shows that hedge fund performance explains long-run IPO success better than market performance. We conclude that the long-run findings are similar to short-run findings except strong IPO success now coincides with strong market performance.

We conduct a virtual marketplace experiment to examine whether framing a welfare transfer associated with a negative externality affects market participants' decisions. Participants — 155 undergraduates at Santa Clara University — are randomly assigned as sellers, buyers, or bystanders in a modified version of the Bartling et al. (2015, 2019) task, in which trading a product generates an external cost to bystanders. Across sessions, we randomly assign a framing statement that describes the welfare transfer as market participants profiting at bystanders' expense, or no framing statement (control). We find that the "taking" frame substantially reduces both sellers' ask prices and buyers' willingness to purchase relative to the control. These results suggest that framing welfare transfers in explicit terms meaningfully constrains the behavior of market participants, even in an incentivized setting.

Accurate perception of social relationships is central to human social cognition, yet little is known about how well individuals perceive incoming social ties within their natural social environments. We collected data from 3,077 adolescents across 117 classrooms and analyzed complete social networks comprising four directed relational layers—friends, best friends, enemies, and worst enemies—alongside adolescents’ predictions of who would nominate them in each layer. Across all layers, perceptual accuracy was strikingly low. Only 0.27% of students correctly assessed all of their incoming ties. Errors were overwhelmingly driven by underestimation: adolescents systematically failed to recognize how many peers named them, both positively and negatively. Misperception was structured rather than random. Popularity (in-degree) was strongly associated with underestimation, whereas sociability (out-degree) was associated with overestimation. These structural patterns were consistent across relational layers and error types and showed no systematic improvement with age. Together, these findings reveal robust structural asymmetries in social cognition, suggesting that young people perceive far less of their social worlds than there is.

Affirmative action policies shape access to higher education in roughly one-quarter of countries worldwide. However, considerable debate persists over whether such policies should promote equality of opportunity through skill development or equality of outcomes through preferential selection. These approaches correspond to developmental (ex-ante) policies that foster human capital formation before competition and preferential (ex-post) policies that modify selection rules at the assignment stage. While both aim to improve representation, they differ in timing, mechanism, and potential behavioral consequences. This paper provides causal evidence on how the timing and magnitude of affirmative action affect performance and representation. Using a controlled laboratory experiment that incorporates a rank-order tournament, I compare developmental interventions that enhance capabilities before competition with preferential policies that grant disadvantaged participants priority access to winning slots after performance is realized. The results show that timing matters. Developmental policies generate substantial performance gains, particularly among disadvantaged participants, while achieving representation outcomes comparable to quotas. Preferential policies do not significantly affect performance relative to baseline conditions. These findings highlight the importance of distinguishing between policies that build skills and those that reallocate opportunities, and underscore the role of behavioral responses in the design of affirmative action policy.

Despite educational and career advancements among women, substantial gender differences persist in labor market outcomes, particularly in leadership roles. Giving advice and making decisions are integral components of effective leadership. While the majority of Economics literature on advice focuses on how receiving advice improves performance, there is limited research examining advice-giving behavior itself. In this study, we use an incentivized experiment to investigate the plausible channels that can explain the gender gap in advice-giving. In this study we study how task-associated-stereotypes impact individuals' decisions to give advice and their assertiveness in doing so. The experimental design incorporates different task types to understand how stereotypes influence advice-giving decisions while also examining the role of self-evaluation as a potential barrier. Additionally, the study tests whether performance feedback can revise self-evaluation and subsequently affect advice-giving behavior. Preliminary results reveal gender differences in male-stereotyped tasks. Specifically, while performance does not differ by gender, females are less likely to give advice on male-stereotyped tasks and demonstrate lower assertiveness due to lower self-evaluation. Feedback helps females revise their performance beliefs and increases their likelihood of giving advice. I also find no gender differences in female-stereotyped or gender-neutral tasks.

This article investigates the impact of political lobbying on designing revenue neutral pollution taxes within a heterogeneous energy industry. Integrated tax-subsidy policies, where high polluting subsectors are taxed and the revenues are used to subsidize low polluting subsectors, are deemed more politically palatable. We develop a multi-sector model where firms are ranked according to their pollution intensity. Firms maximize profit and are taxed by a regulator to internalize their pollution damages. Only high emitting firms create a lobby group that influences their pollution tax. The regulator chooses pollution taxes within each subsector to maximize their own welfare which is equal to social welfare plus a weighted value for bribes from the lobbying group. Our theoretical results show that when lobbying influence is introduced, the high emitting pollution tax declines relative to Pigouvian benchmarks and taxes on other subsectors are affected. In the binding revenue neutral case, the taxes in the other sectors are higher while subsidies in the cleanest subsectors are lower. In the non-binding revenue neutral case, the magnitude of all taxes and subsidies decline because the value of marginal product of labor increases in the high polluting lobbying sector relative to all other sectors. Social welfare is lower in the presence of lobbying. The model highlights how political influence reallocates tax burdens across energy sectors, even in settings where welfare is nearly flat over large regions of tax space.

I develop a novel spatial environmental growth model in which comparative advantage endogenously shifts across regions. Research investment can be directed toward clean and dirty sectors, as well as regions with high concentrations of either. Environmental regulation affects polluting production across regions and redirects research investment, shifting the comparative advantage of dirty sectors toward regions with low regulatory exposure. Cross-regional and cross-sectoral labor reallocation interacts with shifting comparative advantage, determining local employment and pollution patterns. I apply the model to evaluate local labor market consequences of the Clean Air Act 1990 Amendment, a place-based regulation with differential compliance costs across regions. Quantitative results show that conventional estimates that ignore general equilibrium effects substantially overstate the Act's negative impacts on polluting jobs, especially over the longer run.

The objective of this article is to analyze the welfare effects of the European Union (EU) methane regulations, specifically the “You Collect, We Buy” program and the maximum methane intensity standard, to reduce methane emissions from oil producers across varying flaring costs. We develop a sequential game where a planner chooses a methane price to incentivize methane storage as well as a methane-intensity stan- dard while a monopolist chooses energy output and methane storage. We find that the optimal methane price is equal to the marginal damages from pollution plus the marginal social value of producing the good weighted by the elasticity of the proportion of stored methane due to an increase in energy production. The methane price tries to internalize two externalities: the effect of pollution damages and the reduction in monopoly output. The optimal methane intensity standard is affected by the optimal methane price. The relationship between the two policies need to be considered to maximize welfare. Firms with low marginal flaring costs should be faced with a more stringent maximum methane intensity standard and a high methane price to reduce leakages and allow for more methane capture. Welfare is higher when the maximum methane intensity standard is binding compared to the non-binding scenario. As the marginal cost of flaring increases, the welfare difference in the two scenarios decline because a higher flaring cost acts as an another incentive to reduce methane emissions.

Forthcoming

We study how parental divorce affects children’s human capital formation, with particular attention to the heterogeneity in these effects. Using linked administrative data from Texas, we follow over six million children born between 1986 and 1997 from birth through schooling, combining birth records, marital records, and criminal justice records with rich administrative education data. To estimate the ATT of divorce, our main empirical design is a stacked event-study, which uses students whose parents divorce after the event window as controls. On average, math and reading scores fall by 0.02–0.03 standard deviations around divorce; reading scores largely recover within four years and math scores partially recover. However, average effects on behavioral outcomes are large and persistent: four years after divorce, school absences are 10 percent higher and disciplinary incidents are 34 percent higher. To characterize the heterogeneity in the effects of divorce, we develop a new variance-decomposition method that recovers the variance of treatment effects. This method relies on assumptions analogous to those in generalized difference-in-differences settings but extended to the variance of outcomes. We find substantial heterogeneity in the effects of divorce on reading, discipline, and attendance—approximately one-third of children benefit from divorce on these dimensions. In contrast, for math achievement, we cannot reject that effects are uniformly negative. This heterogeneity is only partially due to observable characteristics: boys, white children, and children from non-poor families experience the largest adverse effects of divorce. Our findings imply that divorce is broadly harmful across key dimensions of child development, but that some students fare better after their parents' divorce, at least on certain dimensions. Methodologically, the variance-based approach offers a general tool for measuring treatment-effect heterogeneity in panel settings where only average event-study effects are typically reported.

What happens when financial support of the elderly falls explicitly on their adult children? This paper studies California’s Old Age Assistance (OAA) program, which included “responsible relative” requirements that rendered elderly individuals ineligible for public support if their adult children had sufficient income. However, private transfers from those “responsible” children were compulsory. Using linked census records from 1900 to 1950, I assess the requirement’s impact on OAA participation, retirement decisions and intergenerational coresidence among potential recipients as well as the labor supply and migration decisions of their children. To do so, I exploit the requirement's repeal in 1949 and quasirandom variation in who had and who was a “responsible” relative, determined by the combination of income, gender and the number of one's own dependents. To estimate effects of the requirement on potential recipients, I adopt a differences-in-differences strategy using parents with only daughters (who were exempt from responsibility if married) as a control group randomly less likely to be exposed. Because an additional child increased the income necessary to be "responsible", I use twin births as exogenous declines in the likelihood of an adult child being compelled to support and estimate a difference-in-difference between those with parents old enough to qualify for OAA versus those without. This project contributes to the literature on adult children as a source of elderly support. It differs from most research in this literature by studying a setting where government support fills gaps in private support, not vice versa, and is the first to assess the impacts of mandating filial support. The findings speak to contemporary policy debates around population aging and the family as a potential substitute to public assistance. This research also contributes to previous work on the intergenerational consequences of expanding access to elderly benefits. While Fetter et al (2025) and Costa (1998) find that the Social Security Act reduced reliance on children, this paper is first to evaluate a legal change explicitly shifting support for those elderly most in need away from their children. Lastly, this project assesses the effects of a means-test that reaches beyond the household on welfare uptake.

There is a large body of literature in economics on the topic of motherhood penalty—labor market interruptions and the resulting financial implications due to childbearing and rearing. According to the American Community Survey 2023 data, there are about 5.1 percent of children under the age of 18 years who have some form of physical and/or mental disability. To my knowledge, the impact of a child’s special needs on the motherhood penalty, however, remains unexplored. This paper fills this gap. I use an ordered logistic regression model to conduct the empirical analysis. The data come from Public Use Microdata Samples (PUMS) published by the US Census Bureau. I use 1 percent PUMS data for 2023. Initial results of this study indicate that women, in general, are more likely to work fewer hours than men. Women with children, however, are more likely to work longer hours than men with children. This difference between fathers and mothers is further accentuated when the child has a disability. That is, mothers are more likely to work longer hours than fathers, and mothers of disabled children work even longer hours than fathers of disabled children.

Using a Synthetic Control Method, we analyze what happened to economic institutions in Moldova after citizens of the country received the right to travel to European Union countries with a simplified visa. This policy constituted an exogenous change in the ability to emigrate away from Moldova. After this policy change, we find economic institutions improved in Moldova, and these effects are robust to sensitivity analysis. We also provide evidence of a mechanism: diaspora voters were most likely to vote against illiberal parties. Our results suggest emigration improves economic institutions and that immigration restrictions in immigrant-receiving countries negatively influence institutions and economic development in immigrant-sending countries.

This paper investigates the relationships among alcohol prohibition, institutions – proxied for by measures of economic freedom – and Baumolian Entrepreneurship across United States counties using prohibition data from Sechrist (2012), Metropolitan Economic Freedom data from Stansel (2023), and Net Productive Entrepreneurship data from Callais, et. al (2026). We aim to connect prohibition, modern institutions and entrepreneurial outcomes by testing whether early institutions of prohibition work through current institutions to affect modern entrepreneurship. We investigate this channel using prohibition as an instrument for institutions in 2SLS entrepreneurship regressions, and test over-identifying restrictions to assess instrument validity. If a strong and valid instrument, we will explore the likelihood that modern institutions serve as a conduit through which prohibition might have lasting effects on entrepreneurship.

A substantial body of empirical and theoretical research has examined the impact of globalization and regional institutions on economic growth. Entrepreneurship, particularly in its nascent stage, is widely recognized as a key driver of economic development. This study integrates these strands of literature by exploring how globalization and institutional quality interacts to influence nascent entrepreneurship. Using multi-level regression analysis and data from 20 emerging economies, we investigate the moderating role of institutions on the relationship between globalization (economic, social, and political) and nascent entrepreneurship. Institutional variables include state ownership of assets, protection of private property rights, fiscal health, capital mobility controls, tariffs, administrative requirements, regulatory burden, impartial public administration, and credit market regulations. Our findings suggest that political and informational dimensions of globalization are positively associated with nascent entrepreneurship, while social globalization exhibits a negative correlation. Importantly, institutions can mitigate adverse effects of globalization.

Forthcoming

Forthcoming

We examine the interaction between money market mutual fund (MMMF) flows, bank deposits, and bank lending using quarterly U.S. bank-level data from 2010–2023. We find that MMMFs and bank deposits are economically meaningful substitutes: a 10% increase in MMMF flows is associated with a 0.76% decline in bank deposits, driven by Government and Treasury funds, while tax-exempt and prime funds act as complements. Deposit substitution is associated with lower commercial real estate, residential, and consumer lending, particularly among smaller banks. Results are robust to IV specifications and county-level tests, with the floating-NAV reform providing quasi-experimental support for the funding-channel mechanism.

We present a dynamic reputation model of the labor market, where workers differ in their innate propensity to commit crimes and employers can gather information about workers’ criminal history from a publicly available record. Working while young enhances human capital, which increases future labor market opportunities. The criminal record provides valuable information with regard to a worker’s future propensity to commit crimes, and also deters crimes. On the other hand, it also makes it more difficult for young workers who have committed crimes to accumulate human capital, and can thereby enhance recidivism. We show that the optimal policy balances the two effects, and that under plausible assumptions leads to sealing of records of young offenders, particularly when the social cost of these crimes are not too high. Our model also suggests that other arguments advanced in the literature in favor of the sealing of juvenile records, namely that juvenile records are less predictive of recidivism, and that juveniles have less self-control, are less salient.

This paper examines the effect of replacing arrests with criminal citations on defendant outcomes and subsequent criminal charges. Using web-scraped Maryland court records, I implement a difference-in-discontinuities design that exploits a January 1, 2013 policy shock encouraging citations for low-level offenses. The policy increased the issuance of citations by 47.7% among citation-eligible defendants and 39.3% among citation-eligible first-time defendants. Reduced-form estimates show sizable decreases in pretrial detention for these groups, falling by 29.2% and 40.4%, respectively. Citation-eligible defendants are also 4.8% more likely to have their cases dismissed. However, reduced pretrial detention among eligible defendants is offset by increased pretrial detention among defendants charged with more serious offenses that are ineligible for citation, leaving aggregate pretrial pretrial detention unchanged. Finally, I find little evidence that the policy increased failure-to-appears or re-offending.

An estimated quarter of the U.S. workforce needs an occupational license in order to work and all states partially restrict people with a criminal record from attaining licenses. However, a majority of states have passed reforms to these laws since 2015. In this study, we develop a theoretical model for the effect of collateral consequences for licensing and test this model using novel legislative data on reforms and data from the Current Population Survey and the American Community Survey from 2010 to 2024. Our results suggest collateral consequence reforms significantly increased employment and increased earnings for Black and Hispanic workers.

DACA (Deferred Action for Childhood Arrivals) provides temporary work authorization and deportation protection to individuals meeting strict age, residency, education, and criminal history criteria. Although not its primary focus, DACA’s requirements shift young people’s incentives, and this project investigates whether those shifts prompted changes in student misbehavior. Specifically, leveraging restricted-access, student-level data from Texas, I use a Difference-in-Differences approach to assess DACA’s impact on juvenile crime and school discipline. I compare likely undocumented students to those who are not by combining two unique administrative flags—showcasing granularity not seen in previous literature. Relative to likely undocumented students’ averages, DACA decreased in-school and out-of-school suspensions per student per year by 10.9% and 16.5%, respectively. Transfers to alternative education programs, a proxy for juvenile crime, also fell by 13.7% per student per year due to DACA. All findings are significant at the 1% level and robust to sample and specification variations. These results have important implications for disciplinary policy, demonstrating how students’ incentives can be shifted to deter juvenile crime and misbehavior.

Survey-based discrete choice experiments are widely used to estimate willingness to pay (WTP) for non-wage amenities. Recent innovations in elicitation methods, such as Bayesian Adaptive Choice Experiments (BACE), can generate precise, individualized WTP estimates, making possible a more stringent field test of survey-based discrete choice methods against person-level realized choices. In this paper, we estimate WTP across two key amenities in the U.S. Army: (i) the compensation recruits require to accept longer initial service obligations, and (ii) recruits’ willingness to pay for a first duty station of their choosing. We implement two complementary approaches. First, we conduct a randomized controlled trial (RCT) that randomizes enlistment bonuses across these amenities, allowing us to estimate WTP for station of choice and for extending recruits’ contractually obligated service from 3 years to 4, 5, and 6 years. Second, we administer a BACE survey to a sample of potential recruits—nearly half of whom subsequently enlist—to estimate the same parameters. We compare estimates across the RCT and survey-based approaches, examining both internal consistency within the survey and external consistency with real-world enlistment choices. Together, these analyses provide a field test of survey-based valuation methods and suggest that targeted changes to Army enlistment incentive structures could yield substantial cost savings.

This report for the 14th Quadrennial Review of Military Compensation (QRMC) focuses on allowances: the Basic Allowance for Housing (BAH), Basic Allowance for Subsistence (BAS), and Cost-of-Living Allowances (COLAs). A companion report focuses exclusively on BAH, responding to congressional questions posed about it in the Fiscal Year 2023 National Defense Authorization Act. We found that BAH is, on average, higher than what civilians of comparable income spend on rent and utilities. However, BAH is also volatile: in any given year, more than half of military housing areas (MHAs) have a greater than 10 percentage point spread across the year-over-year BAH changes for different paygrades in that MHA. This volatility is the result of the Department of Defense estimating the cost of six different housing profiles in each of the about 300 MHAs every year. We recommend three possible courses of action to reform BAH, each of which use other government-generated data and would make BAH more predictable. We also found that the statutory definition of BAS is unclear and has led to it drifting upward over time relative to the price of food in a way that Congress probably did not intend, and we recommend tying its level to the US Department of Agriculture’s estimate of the cost of a liberal food plan for an adult man. Finally, COLAs—both in the contiguous United States and overseas— are in part driven by the Living Pattern Survey, which is infrequent, is not verified, and can lead to counterintuitive outcomes. Commissaries and exchanges keep records of the volume of sales to active duty servicemembers, which could directly verify how much access to on-base savings affects the local cost of living for servicemembers.

This report for the 14th Quadrennial Review of Military Compensation (QRMC) responds to questions related to the Basic Allowance for Housing (BAH) posed by Congress in the 2023 National Defense Authorization Act (NDAA). BAH is a form of compensation designed to help members rent adequate housing near their duty station, and the amount depends upon their rank, their military housing area (MHA), and whether they have dependents. We found that BAH is, on average, higher than what civilians of comparable income spend on rent and utilities. The only exception is the BAH paid to members in the W1 paygrade without dependents. However, BAH is also volatile: in any given year, more than half of MHAs have a greater than 10 percentage point spread across the year-over-year BAH changes for different paygrades in that MHA. This may contribute to some members’ dissatisfaction with BAH, as may comparison to on-base privatized housing standards, which are higher than BAH standards for some paygrades. We also found that over the long term, BAH responds to changes in housing markets very well, but its built-in lag of about 6 to 18 months means that it does not keep up with rapid housing cost inflation.

The Office of the Under Secretary of War for Personnel & Readiness contracted with CNA to answer the following compensation policy questions: Do current Basic Allowance for Housing (BAH) adequacy standards (for housing size, type of home, etc.) and sampling methodologies result in BAH rates commensurate with actual servicemember choices? If not, why? CNA developed and administered the 2025 BAH Adequacy Survey to determine the extent to which BAH helps servicemembers secure their basic housing needs. The vast majority (84 percent) of BAH recipients are not living at their BAH standard and nearly half of BAH recipients overspend their BAH by 5 percent or more. Incongruencies between perceptions of BAH adequacy and reported costs indicate that BAH recipients’ understanding of BAH does not align with policy; they consider BAH inadequate even when it covers their spending. Based on our study, we provide recommendations aimed at improving servicemember housing experiences and future iterations of the BAH Adequacy Survey.

Monetary policy may have an information effect when information asymmetry exists between the central bank and the private sector. That is, the private sector believes that an interest rate decision reveals the central bank’s private information about the state of the economy, and uses the interest rate as a signal to update beliefs about the state of the economy. However, the literature has yet to agree on whether the information effect of monetary policy is quantitatively important. In this paper, we first show that the reduced-form estimation of the Fed information effect is biased due to the correlation between monetary policy surprises and the unobserved shocks. We then develop a New Keynesian model with a two-way learning process: the central bank learns from lagged aggregate inflation and output, and firms learn from individual marginal costs and the interest rate. Our calibrated model shows that the information effect reduces the output gaps caused by demand shocks and noise shocks but may lead to a temporary rise in inflation after a contractionary monetary policy shock.

This paper examines the spillover effects of US unconventional monetary policies (UMP) on international capital inflows to 33 emerging market economies (EMEs) from 2003Q1 to 2019Q4, focusing on quantitative easing (QE) and its normalization phases, i.e., tapering and quantitative tightening (QT). While there is already a substantial literature on the earlier phases of UMP, little research to date has examined the later QT phase, and only a few studies have asked whether the effects of easing and tightening are symmetric. Using a dynamic panel fixed-effects model with changes in the Wu-Xia shadow federal funds rate interacted with phase-specific dummies, we find pronounced differences in the magnitudes of the effects across policy phases. The spillover effects during QE were large, whereas the normalization phase as a whole was relatively muted. Despite the financial market stress associated with the 2013 taper tantrum, we find no evidence of sustained quantity effects on EME capital inflows during the actual tapering implementation period. QT, by contrast, had statistically and economically significant effects, specifically on portfolio and FDI inflows. Lastly, our analysis highlights several important determinants of EME capital inflows, particularly global risk aversion, domestic political stability, and expected investment returns driven by EME macroeconomic fundamentals.

How should monetary policy react to an increase in import tariffs? This study examines the appropriate monetary policy response to tariff increases and argues that unconventional monetary policy, particularly forward guidance, is essential for an optimal response. Whereas previous studies emphasize mere monetary easing, this is insufficient. The analysis extends a small open economy model with nominal rigidities, following the framework of Monacelli (2025, mimeo), by incorporating a fiscal authority that finances expenditures through distortionary income taxation and debt issuance, governed by a class of Bohn rules. The government is assumed to be benevolent, operating as a social planner that maximizes household utility. We employ a numerical analysis. Specifically, a numerical exercise considers a 20% increase in import tariffs, which raises consumer price index (CPI) inflation because CPI includes import price inflation. Higher CPI inflation reduces the real burden of government debt, prompting the fiscal authority to cut taxes to stabilize its fiscal position. This tax cut increases the natural rate of output and exerts downward pressure on the output gap. Although monetary policy attempts to stimulate output by reducing the nominal interest rate, the zero lower-bound constraint prevents further easing, generating deflation and a persistent negative output gap. If the social planner follows policy under commitment, the implied dynamics require the nominal interest rate to remain at zero for an extended period. This outcome establishes the necessity for forward guidance as an unconventional monetary policy to address an increase in import tariffs. We extend this model to a two-country framework to capture the repercussions of an increase in import tariffs in one country. These results closely mirror those of the small open economy model. The tariff-imposing country bears primary welfare losses, whereas the partner country—which does not impose tariffs—remains comparatively unaffected. Therefore, import tariffs operate not as a “beggar‐thy‐neighbor” policy but as a “beggar-thyself” policy. Policy under commitment leads to a sustained zero nominal interest rate, reinforcing the role of forward guidance as an essential unconventional monetary policy, even in a two-country setting.

I construct a measure of the stance of monetary policy as the spread between the 2 year Treasury Note rate and the estimated Clarida, Gali & Gertler (2000) target rate. This spread captures deviations of market priced expected FOMC behavior over a policy sensitive duration and the historical benchmark expected policy path. Shocks to the spread are unexpected changes to the expected policy path and are monetary policy shocks that manifest in advance of any actual change of the funds rate. Granger Causality tests over 5 and 10 year rolling windows suggest that the spread leads the funds rate in all non zero lower bound episodes. Factor Vec- tor Autoregression analysis reveals economically and 90% statistically significant responses to spread shocks on impact and in short windows thereafter in the duration of unemployment spells, residential construction spending, building permits, and money and banking aggregates. A rich set of variables that indicate monetary policy operating with a lead. Exploiting the immediate response in money aggre- gates via bank deposits, I estimate a medium scale New Keynesian model with a money demand channel, where deposits are utilized as an observable in place of the funds rate. The model-based counterfactual funds rate in a robust to the number of instruments the FOMC utilizes and reveals the stance of monetary policy across single instrument, quantitative easing or tightening historical episodes. The peak rate of 5.5% in late 2024 is equivalent to the peak rate of 2.5% in the middle of 2019 in terms of the tightness of monetary conditions. The FOMC achieved 12 percentage points of greater monetary stimulus in the COVID recession than in the Great Recession.

We examine the exposure of leading Chinese firms to stock price crash risk amid the US-China trade war and rising geopolitical hostility and economic policy uncertainty. We find that Chinese multinationals with ex ante outward investments in the US through cross-border M&A or green field projects before the onset of the trade war exhibit lower stock price crash risk during the period of heightened policy uncertainty and hostility. This finding remains robust to a battery of additional checks, including multifaceted parallel trends tests and entropy balancing. In addition, cross-sectional heterogeneity analyses unveil that high-tech companies are the primary driver of the main findings. Since the outbreak of the trade conflict, Chinese high-tech firms with established outward investments in the US demonstrate stronger innovation resilience under tightening intellectual property transfer restrictions. Our empirical evidence suggests that established ties through ex ante outward foreign investments act as conduits for positive strategic and technological externalities, effectively sheltering firms from the adverse effects of trade and geo-economic policy turmoil.

This paper develops a two-country model with heterogeneous firms à la Melitz (2003) to understand how openness to trade and horizontal FDI interacts with search frictions in the labor market to shape aggregate unemployment. The model intends to shed light on two interrelated questions: first, how does aggregate unemployment vary with labor market frictions under different degrees of openness to trade and FDI?; and second, how does investment liberalization affect unemployment for different levels of labor market frictions? The results show that lower costs of FDI can widen or narrow the gap in unemployment rates across countries by attenuating the reallocations of workers to the comparative-advantage sectors induced by labor market frictions. The paper also highlights how labor market frictions interact with trade and FDI costs, to shape the decision firms make between exports and FDI.

We empirically examine how market competitiveness affects Chinese firms’ greenwashing behavior based on information of publicly listed Chinese firms between 2011 and 2020 acquired from multiple sources. We adopt a peer-relative estimate to quantify firms’ greenwashing behavior. To effectively mitigate the endogeneity issue, we compute two industry-level instrumental variables—import penetration shocks, and foreign tariff shocks—that are highly related to market competitiveness but orthogonal to firm-level greenwashing. It is found that a higher level of market concentration (or less competition) significantly reduces greenwashing. Further mechanism analysis reveals that firms’ ownership structure plays an important role. The negative influence of market concentration is mainly reflected among state-owned firms, whose greenwashing behavior responds negatively to Chinese government’s ESG (environmental, social, governance) mandating disclosure policy. When the overall greenwashing score is disaggregated into the three ESG dimensions, the reduction among state-owned firms is primarily driven by a decline in governance (G) greenwashing. In contrast, relative to private firms, state-owned firms display an upward trend in greenwashing related to environmental (E) performance.

The Earned Income Tax Credit (EITC) and minimum wage are two of the most impactful policies for low-income households, with many states now having policies that go well beyond the federal policy. While the labor market and spillover effects of each policy have been extensively investigated, whether and how these policies may enhance or mitigate each other has been largely overlooked, with the most recent empirical evidence for labor outcomes being 15 years old. This research presents our first attempts to revisit and update -- in terms of data and empirical methods – this sparse evidence, and it makes three contributions. First, it demonstrates the enormous growth in these programs since 2010 and clarifies the different ways in which the two policies may interact. Second, we revisit and update the two studies that provide empirical evidence on the joint effect on labor outcomes, Leigh (2010) and Neumark & Wascher (2011), and find very different results when using more current data. Our third contribution adapts the distributional difference-in-difference estimator from Cengiz et al (2019) to provide new evidence on this interaction. Our preliminary results suggest that 1) increased minimum wages lead to negligible (dis)employment effects, especially for the very-low-income households with children most affected by EITC, and 2) longstanding concerns that the EITC may lead to lower wages, which a minimum wage could help mitigate, have empirical support. While still preliminary, our findings all suggest that this possible interaction is likely growing in importance and requires additional investigation.

In April 2022, the State of Texas began transporting migrants to unsuspecting cities across the United States as part of Operation Lone Star. The program emerged amid increasing migrant arrivals at the US-Mexico border and longstanding public debate over the relationship between immigration and crime. Since then, thousands of migrants have been bused to sanctuary cities nationwide. Using novel busing data from the Texas Division of Emergency Management and incident-level crime data from the FBI’s National Incident-Based Reporting System from 2021 to 2024, we leverage the exogenous ``buslift'' of migrants in a difference-in-differences framework to estimate the causal effect of migrant inflows on crime in destination cities. Relative to non-recipient cities, the arrival of migrants has no distinguishable effects on overall crime or crimes committed by Hispanic offenders. The precisely estimated null effects allow us to rule out economically meaningful increases in crime in destination cities with high confidence. Using a triple difference approach, we find reductions in Hispanic crimes relative to offenses committed by non-Hispanic offenders. Overall, we find no evidence that migrant arrivals increased crime, with estimates suggesting relative declines in Hispanic offending within recipient cities.

Regulations, information constraints, and administrative burdens are key drivers of provider behavior; we use recent state-level policies targeting gabapentin as a natu- ral experiment to study these mechanisms in prescribing decisions. Gabapentin, an increasingly misused medication with rapidly expanding off-label use, has been sub- ject to two distinct regulatory interventions: Drug-of-Concern (DOC) designation, which expands PDMP reporting without constraining prescribing, and Schedule V classification, which imposes binding prescribing and documentation require- ments. Using a novel dataset linking MarketScan claims (2016–2023) with CDC multiple-cause-of-death records, and a stacked difference-in-differences framework, we document four main findings. First, DOC designation has no detectable effect on gabapentin prescribing. Second, Schedule V classification leads to a sustained decline in prescribing, reducing treatment rates by 0.2–0.3 percentage points (from a 3–4% baseline) within eight quarters, driven by reduced continuation among established users. Third, we find little evidence of harmful substitution: opioid prescribing does not increase, co-prescribing with opioids declines, and pregabalin substitution is modest and delayed. Fourth, Schedule V classification is associated with a meaningful reduction in gabapentin-involved mortality, including opioid-co- involved overdose deaths, with no corresponding change following DOC designation. Mechanism tests exploiting variation in PDMP mandates and patient histories show that neither salience nor information effects generate measurable changes in pre- scribing; instead, the behavioral response is concentrated in settings where Schedule V classification raises administrative hassle costs. Collectively, the results demonstrate that only binding regulatory frictions meaningfully reduce gabapentin use and associated harms, while monitoring-only policies are largely ineffective.

Undocumented students in the United States are barred from accessing federal financial aid. Some states also actively prohibit enrollment at public universities, while others have implemented policies to make higher education more accessible. Yet, data constraints make estimating the effects of these policies difficult. In this paper, we study the effect of the California DREAM Act, which granted undocumented students access to Cal Grants, the state’s primary form of financial aid, through the California Dream Act Application (CADAA). Equipped with the universe of CADAA applications, we overcome the need to proxy for immigration status as is usually done in the literature and are thus able to produce better estimates of the effects of aid for this population. Using a regression discontinuity design (RDD), we first present reduced form estimates of the intent-to-treat (ITT) effect—i.e., the effect of scoring above annual Cal Grant offer cutoffs on various outcomes that proxy for educational persistence. While we observe minimal noncompliance, we also implement a fuzzy RDD to estimate the local average treatment effect (LATE), where a student’s score relative to the cutoff instruments for whether they were actually offered an award. Ultimately, we find evidence that scoring above the offer threshold affects awardees’ postsecondary outcomes.

This paper explores the impact of a dramatic change in migration and trade policy - closing the U.S.-Mexico border to non-essential travel. Using variation in the population of U.S. border counties relative to border-adjacent Mexican counties, we explore the impact of closing the border during the COVID-19 pandemic on employment outcomes in the U.S. We find evidence suggesting that border closures adversely affected businesses on the U.S.-side which were more heavily dependent on Mexican consumers who were prohibited from traveling to the U.S. during the COVID pandemic. Moreover, the negative impacts are strongest for important industries which were not deemed “essential,” and which were thus left more vulnerable to the policy change. These results are robust to controlling for various measures of exposure to the COVID pandemic, so that we can better isolate the impacts of the border closures and rule out that the estimated impacts are driven by the pandemic itself. As such, the paper contributes to the wider migration and trade literatures by quantifying the negative impacts of restricting cross-border flows on U.S. businesses and workers.

This paper examines the labor market impacts of the State of Texas’ Operation Lone Star migrant busing program, focusing on the arrival of asylum seekers in the Denver metropolitan area between 2022 and 2024. Unlike prior migration shocks studied in the economics literature such as refugee inflows driven by natural disasters or voluntary labor migration, Operation Lone Star represents a deliberate, state-orchestrated redistribution of migrants targeting sanctuary cities. We argue that the economic effects of the program may operate through a mix of demand-side effects and informal labor market expansion which raises distinct questions about labor market integration, particularly given asylum seekers’ constrained access to formal employment due to delayed work authorization. Empirically, the paper employs synthetic controls and difference-in-differences methods to estimate the causal effects of migrant arrivals on employment, establishment counts, and wages across industries in the Denver MSA using Quarterly Census of Employment and Wages (QCEW) data. Preliminary results suggest limited aggregate labor market disruption, with heterogeneous effects across sectors. These patterns are consistent with a framework in which asylum seekers contribute to local economic activity through consumption and informal work while facing barriers to formal labor market participation. The paper contributes to the literature by providing one of the first quantitative assessments of Project Lone Star and by integrating political economy theory with causal labor market analysis to better understand how contemporary migration governance shapes economic outcomes in receiving cities.

We study the effects of a major immigration reform on alleged and substantiated maltreatment of Hispanic children using administrative data from child protective services agencies. Secure Communities (SC) ties federal immigration enforcement to local law enforcement. We exploit the staggered rollout of SC across counties to estimate a dynamic treatment effect model. We find that SC implementation increased the number of Hispanic children found to be victims of child maltreatment as well as the likelihood that maltreatment allegations for Hispanic children are substantiated. Results are consistent with stricter immigration enforcement increasing maltreatment among Hispanic children while reducing reporting rates.

Sleep deprivation imposes significant public health and economic burdens. While researchers studying events like daylight saving time transitions have quantified the impacts of population-wide sleep shifts, less is known about the consequences of acute, voluntary, and recreation-driven sleep loss. This paper investigates this gap by studying the 2002 FIFA World Cup, hosted in South Korea and Japan. The extreme time difference meant that US-based fans sacrificed significant sleep to watch live matches. We track fatal accidents in areas with large German populations on days when the German national soccer team played early morning games. Germany has by far the largest number of Americans who trace their ancestry to a foreign country, and they made it to the final of this World Cup. Areas with greater than 30% German heritage experienced increases in fatal car accidents of 40% relative to control areas after German games. The effects are dose-dependent and rise as the share of the German population increases. Our results are larger for crucial tournament games and non-alcohol-related incidents, consistent with sleep-deprived driving. Effects are driven by male drivers, mirroring World Cup viewer demographics. Placebo tests using the 2006 World Cup, with no games played during normal U.S. sleeping hours, confirm that sleep disruption, not the sporting event itself, drives our findings.

Travelers routinely face a trade-off between minimizing travel time and reducing exposure to traffic risk, yet this speed–safety trade-off has rarely been quantified using revealed route choices at scale. This paper asks: How much travel time are individuals willing to give up to reduce traffic accident risk, and what does this imply for the value of safety in everyday transportation decisions? Quantifying this trade-off is central to transportation policy, infrastructure investment, and the economic valuation of safety improvements. We develop a revealed-preference framework that compares alternative routes connecting the same origin–destination pairs, focusing on the trade-off between travel time and expected accident risk. For each origin–destination pair, we construct the fastest feasible route and a safer alternative route, and measure differences in predicted travel time and exposure to traffic fatality risk along the route over a recent period (2020-2022). To ensure behavioral relevance, origin–destination pairs are sampled using distance-based weighting to reflect observed travel patterns. We estimate travelers' trade-off between travel time and safety using a Poisson pseudo–maximum likelihood (PPML) gravity-style specification, in which route choice intensity depends on travel time and accident risk, while controlling for origin–destination fixed effects. This specification accommodates zero flows, heteroskedasticity, and non-linearities common in large-scale transportation data. Identification comes from within–origin–destination variation in travel time and risk exposure across alternative routes, holding trip purpose and geography fixed. Estimated coefficients are used to recover the marginal rate of substitution between travel time and accident risk. We translate these trade-offs into a value of statistical life by combining estimated willingness-to-pay for risk reductions with standard relationships between fatality risk and expected mortality. This approach bridges transportation route choice behavior with safety valuation methods traditionally derived from labor markets or stated-preference surveys. The paper contributes in three ways. First, it provides a scalable revealed-preference method for estimating speed–safety trade-offs using real-world route alternatives. Second, it offers new evidence on the valuation of safety in routine travel decisions rather than occupational settings. Third, it delivers policy-relevant insights for navigation systems, roadway design, speed management, and the welfare evaluation of transportation policies that affect travel time and accident risk.

The mass rapid transit (MRT) or metro system, known for its low pollution, high punctuality, and frequent service, plays a crucial role in green transportation. However, in practice, metro systems heavily rely on electric energy for passenger transport and station operations, leading to carbon emissions and environmental challenges. Therefore, metro operators need to gather information on electricity utilization efficiency and carbon emission efficiency to effectively implement "energy-saving and carbon-reduction" initiatives. Unfortunately, these issues have been largely overlooked in existing literature. This paper develops two models—a two-stage mixed network Directional Distance Function (DDF) model and a traditional DDF model—to measure environmental, energy, and carbon efficiencies. The study evaluates these efficiencies using time series data from the Taipei MRT (TRTC) from 2005 to 2022. The empirical analysis shows that, using the mixed network DDF model, the average environmental efficiency is 0.9916. The mean electricity efficiency values are 0.9997, 0.99966, and 0.9997 for the entire system, production, and service stages, respectively, indicating that these efficiencies are close to the frontier level. The average carbon emission efficiency values are 0.9599, 0.9700, and 0.9504 for the entire system, production, and service stages, respectively. This indicates that carbon emission efficiency has not yet reached the efficiency frontier, primarily due to lower efficiency in the station system. To improve environmental efficiency, operators should prioritize enhancing carbon emission efficiency during service processes, which could lead to an average carbon reduction of approximately 4.96% per quarter. Furthermore, the traditional DDF model yields efficiency values of 0.9895, 0.9860, and 0.9830 for environmental, electricity, and carbon emissions, respectively. These findings indicate that these efficiency values still fall short of the efficiency frontier. The empirical analysis demonstrates that the proposed mixed network DDF model provides operators with valuable insights into overall environmental efficiency, production efficiency, and service performance. It also offers energy and carbon emission efficiency data for medium- and high-capacity systems and stations, which can assist operators in implementing effective "energy-saving and carbon-reduction" measures.

How does an increase in public sector salaries affect labor market dynamics in settings where public sector jobs are allocated through merit-based exams? While such salary increases are often promoted as tools for attracting, retaining, and motivating high-quality employees to deliver effective services, their broader labor market implications remain underexplored. Existing literature has focused largely on improvements in worker selection and performance (e.g., De Ree et al., 2018; Finan et al., 2017), but not on how wage increases affect job search and employment decisions. Leveraging a natural experiment in Bangladesh, where public sector salaries nearly doubled in December 2015, we estimate the causal impact of the salary reform on college graduate employment using a difference-in-differences strategy. Our design exploits a strict upper age limit of 30 years for government job eligibility, comparing labor market outcomes of college graduates just below and above this threshold before and after the reform. Using multiple rounds of Bangladesh Labor Force Survey (LFS) data, we document a 14 percentage points increase in unemployment among college graduates due to the salary hike. We argue that this is due to the postponement of labor market participation in the private sector to prepare for highly competitive merit-based public exams, which became even more competitive after the salary surge. We find that the effect on unemployment is higher for recent and female college graduates. The effect of the salary increase on unemployment also persists in the long run: Four years after the salary increase, college graduates eligible to apply for government jobs during the policy change but now ineligible demonstrate a three-percentage-point rise in the probability of unemployment. More saliently, this group of college graduates is 6.8 percentage points less likely to participate in the labor market, implying a large share of these college graduates become discouraged jobseekers. The related social costs are significant, as individuals in this age group of college graduates are less likely to be married and are more likely to live in financial difficulties.

This paper examines the role of social networks in the diffusion of labor strikes in the United States. Using new data from the Cornell ILR Labor Action Tracker (2021–2024) and Facebook-based county connectedness measures, we document that strikes spread primarily through social networks rather than spatial, industrial, or political linkages. A 1% increase in network exposure is associated with a 2.3% contemporaneous and 0.8% lagged increase in strike activity. We show that both informational and behavioral channels drive diffusion. Policy and administrative environments, such as right-to-work laws and public sector notice requirements, shape the timing and persistence of these network effects.

We provide a comprehensive analysis of the effects of minimum wages on blacks, and on the relative impacts on blacks vs. whites. We study not only teenagers – the focus of much of the minimum wage-employment literature – but also other low-skill groups. We focus on employment, which has been the prime concern with the minimum wage research literature. We find evidence that job loss effects from higher minimum wages are more evident for blacks – and more so for black men. In contrast, they are not very detectable for whites. Moreover, the effects of minimum wages are often large enough to generate adverse effects on earnings (and relative earnings) of blacks. Given strong residential segregation by race in the United States, the race difference in the effects of minimum wages implies that any adverse impacts fall on areas with a high black population share. We explore additional evidence on whether minimum wage effects are also more adverse in black areas, regardless of individual race. We find weak evidence of this heterogeneity, although it does accentuate the concentration of the adverse effects of minimum wages in areas where the black population is concentrated.

The risk premium is the additional return above uncovered interest rate parity risk-averse foreign investors require to hold domestic assets. This premium is well-known to vary over time, although an economic explanation of this variation has largely proved elusive. In this paper, we propose a disaggregated measures of economic policy uncertainty as alternative economic explanations for the risk premium. In particular, we find that monetary policy uncertainty in both Japan and the US contribute significantly to movements in the risk premium on the Japanese yen since the mid-1990s

Despite being a cornerstone of many international monetary and macroeconomic models many studies have found that Uncovered Interest Rate Parity (UIP) often doesn’t hold. There is still a lack of consensus, however, about the reasons for such failures. We suggest that limits to arbitrage is a likely reason that has not been studied sufficiently. We test several implications of imperfect capital mobility and find support for this hypothesis. We investigate deviations from UIP for nine major Asian currencies using daily data from 2009–2022 at 3-month, 12-month, and 5-year horizons against two base currencies: the U.S. dollar and the Japanese yen. Departures from UIP have often been assumed to be due to efficiently priced risk premia or by limits to arbitrage and capital mobility. However, we find that deviations from UIP are significantly larger when measured against the yen than the dollar. If the efficient risk premia explanation was correct the yen was less risky than the dollar. That seems to be unlikely to be correct. On the other hand, the yen market is much smaller than the dollar market which would be consistent with lower capital mobility and greater limits to arbitrage against the yen than the dollar. While most UIP studies focus on average deviations based on regression analysis. We also study the patterns of deviations and find that for many of the currencies there are numerous cases of prolonged deviations from the average UIP biases in the same direction and that these shorter-term patterns are often significant. Such patterns are inconsistent with many of the explanations based on factors such as current account deficits and debt to GDP ratios which tend to change only fairly slowly. This implies that many of the deviations are due in part to shorter-term factors which could usefully be studied.

Current eminent domain literature has predominantly focused on land use ex-ante takings. This paper extends the discourse by formalizing Michelman’s (1967) concept of “demoralization costs” associated with ex-post takings through a structured economic model, building on Blume, Rubinfeld, and Shapiro (1984). Using both Cournot and Stackelberg game-theoretic frameworks, the study investigates the social implications of undercompensation and strategic household resistance for higher compensation. The analysis reveals that in government-led and unilateral scenarios, the government tends to undercompensate, striking a cost-benefit balance. Conversely, household-led games demonstrate that stronger resistance compels the government to increase compensation to mitigate opposition. These findings inform a proposed compensation policy complemented by a threshold rule, designed to restore social optimality by curbing excessive resistance and balancing governmental authority with household dissent. By addressing the ex-post consequences of undercompensation, this study expands eminent domain literature beyond its traditional focus on ex-ante land use decisions. It also strengthens the justification for fair compensation, the means approach, by showing that it not only resolves land assembly problems but also reduces social waste by minimizing resistance and fostering societal progress toward the rule of law and civil governance.

This paper analyzes the association of experiencing a nearby homicide on individual voter turnout in the 2016 U.S. general election in Houston, Texas. We assemble a novel individual-level dataset by merging geocoded murder incident records with the state voter file. Our empirical strategy exploits the as-if random timing of murders around election day in a difference-in-differences design: individuals living within one mile of a murder in the 180 days before the election serve as the treatment group, and those liv- ing near murders that occurred in the 180 days after the election form a control group. We find that pre-election local homicides substantially depress turnout: treated indi- viduals are about 7.6 percentage points less likely to vote in 2016 relative to the control group. Younger voters (millennials) and politically unaffiliated (non-partisan) voters exhibit especially large turnout declines when exposed to local violence, with millen- nials experiencing a 12 percentage point reduction and non-partisans experiencing an 8 percentage point reduction. By contrast, we find that all racial groups experience similar demobilization effects. Men and women also show comparable turnout declines. We discuss mechanisms for the depressive effect of community violence on voting, and consider the implications for democratic representation and public safety.

Card and Dahl (2011) demonstrates that upset losses in the NFL cause an increase in intimate partner violence. We replicate these results, finding coefficients consistent with those in the original paper. We then conduct a series of expansions and robustness checks, finding that the original paper's results are robust to the inclusion of additional years and teams, as well as adjusting the methodology used to approach favorite teams and police agency inclusion.

This paper will analyze the institutions of youth, gender, and work in Lebanon, Palestine, and Jordan from 2000 to 2020. In this analysis, the paper will present results that show how the institutions of family background, “expected” work, and education influence labor markets over time. During a time with turmoil in international financial markets and regional conflicts, these institutions prove to be remarkably resilient in their ability to determine labor market outcomes including employment, wages, and wage-gaps. Using labor market surveys, ILOSTAT, and School to Work transition surveys, this paper finds a strong connection between father’s education and labor market outcomes of sons, but not daughters. Additionally, regional dummies and other family background controls persist in their importance throughout the period, especially in Lebanon. Overall, despite differential political structures, Jordanian and Palestinian labor markets still appear to be remarkably similar in relation to gender, youth, and work.

This paper examines the role of mobile financial services in shaping women’s participation in overseas labor markets and explores the barriers and opportunities for adopting digital financial technologies. We have surveyed around 2,500 randomly selected low-skilled international migrant workers in Bangladesh, divided into five subsamples of 500 individuals each: (i) female return migrants not planning to migrate again; (ii) female return migrants planning further migration; (iii) women intending to migrate for the first time; (iv) male return migrants not planning to migrate again; and (v) male return migrants planning further migration. Preliminary findings indicate that compared with women not planning to migrate again, women intending further migration tend to be more financially and digitally literate, more empowered, and more likely to use mobile money services for remittance. The gender divide along these characteristics is much greater. Relative to male migrants, female migrants display significantly lower levels of financial, digital, and digital financial literacy. Results also reveal a gender gap in digital financial services adoption: men are significantly more likely than women to use mobile money services for international transfers. The findings hold policy implications for supporting women’s financial inclusion and international labor market participation in developing countries.

This study investigates gender differences in job burnout in the United Arab Emirates (UAE), with a particular focus on why women may be more likely to experience burnout in a rapidly transforming labor market. Although men continue to represent the majority of the workforce in the UAE, female labor force participation has been steadily increasing, supported by national development initiatives and broader social reforms. This shift provides an important setting to explore how occupational stress and burnout manifest differently across genders within the UAE’s unique socio-cultural and demographic environment. Building on a substantial body of literature suggesting that women often report higher levels of emotional and physical exhaustion, this study evaluates whether commonly identified drivers of burnout—such as workload, job characteristics, and family responsibilities—are sufficient to explain observed gender disparities. In addition, the study explores a further dimension: the role of gender role expectations and perceptions of empowerment in shaping women’s experiences of work-related stress. We propose that the pressure associated with balancing professional responsibilities alongside family and societal expectations may intensify burnout among women. To analyze these dynamics, the study utilizes original survey data collected from a diverse group of workers in the UAE, including both nationals and expatriates from a wide range of cultural, ethnic, and religious backgrounds. This diversity offers an important perspective for understanding how gendered experiences of burnout vary across different groups within a highly multicultural workforce. The survey includes detailed questions on work–life balance, perceived autonomy, emotional labor, personal views on gender roles and empowerment, and employment conditions. By examining these issues, this research aims to contribute both to the academic literature and to policy discussions by identifying the key mechanisms behind gender differences in burnout. The findings are expected to provide insights that can inform workplace policies and gender equity initiatives tailored to the distinctive characteristics of labor markets in the Gulf region, thereby improving our understanding of how gender, culture, and social norms shape occupational well-being in rapidly modernizing economies.

Forthcoming

We study how individualized, proxy-based environmental information shapes beliefs, defensive investments, and demand for precise diagnostics in the context of waterborne lead exposure. Leveraging the mandated disclosure of water service line inventories under the 2021 Lead and Copper Rule Revision, we conduct a pre-registered field experiment in four large U.S. cities that discloses to treated households whether their building’s service line is recorded as lead. Disclosure induces sharp belief updating. Among homes recorded as having lead service lines (LSLs), perceived LSL probability rises by 21 percentage points relative to non-lead homes; meanwhile, belief confidence increases by 0.2 points (on a 1-5 scale) across all treated respondents. Willingness to pay for a pitcher-style lead filter falls by $8 on average among the non-lead respondents, while the effect on the lead respondents is ambiguous. At the same time, willingness to pay for a professional tap-water lead inspection increases by $9 among lead respondents and decreases by $7 among the non-lead respondents. We develop a stylized model in which households’ optimal decisions depend on both the mean and variance of perceived tap-water contamination, which rationalizes these patterns. Back-of-the-envelope calculations suggest that disclosure of water service line material type alone could generate approximately $17 million in social value across our four study cities, even before accounting for any health gains from subsequent mitigation practices.

In recent decades, several utility sectors, including electricity, nuclear power, and natural gas have undergone substantial consolidation, driven by expectations that economies of scale would enhance efficiency and service reliability. Despite facing similar infrastructure challenges and regulatory environments, the U.S. drinking water sector remains largely decentralized with more than 50,000 community drinking water systems, over 80% of which serve fewer than 3,300 people. The small scale of most systems limits opportunities to realize economies of scale and can reduce operational efficiency and complicate regulatory compliance, as reflected in the consistently higher rates of Safe Drinking Water Act violations among smaller systems (EPA, 2024). As a result, policymakers have increasingly promoted system consolidation as a strategy to address these challenges, with several states implementing legislation and programs to mandate or encourage consolidation. Despite these efforts, there exists limited empirical evidence on whether water system consolidation achieves its intended goals. This study examines the impact of consolidation on system expenses, customer rates, and water quality, with preliminary results suggesting improvements in regulatory compliance but limited evidence of changes in expenses or rates. Combining data on consolidations from California’s Consolidation Data Dashboard with system-level expense and rate information from the CA Electronic Annual Report (eAR) and records of Safe Drinking Water Act (SDWA) violations from the EPA, I employ a staggered rollout difference-in-differences design to estimate the impact of consolidation on per-unit system expenses, customer water bills, and SDWA violations. Preliminary results suggest a statistically significant decrease in SDWA violations following consolidation, suggesting improvements in regulatory compliance and water quality. However, preliminary results show no statistically significant impact on total, operating, or capital expenses, nor on monthly water bills for the typical consumer. These findings offer an initial step toward understanding how consolidation impacts the performance of drinking water systems and its potential as a strategy for improving efficiency and regulatory compliance. This study also contributes to the broader economic literature on public service provision and scale efficiency in regulated utilities and informs ongoing policy discussions about how to ensure safe and affordable drinking water.

As climate change worsens, adaptation becomes increasingly necessary — yet today's adaptation may constrain tomorrow's options. We study these questions in agriculture, where irrigation is a key adaptive tool. Using parcel-level groundwater pumping data across California's Central Valley, we estimate how pumping responds to temperature shocks and aquifer depth, model how these factors will shape groundwater resources through the end of the century, and quantify the welfare consequences of climate change inclusive of adaptation. Our dataset combines annual parcel-level pumping with monthly groundwater depth and daily weather data. Using a climate econometrics framework with unit and year fixed effects, we find that each day above 38°C increases annual groundwater consumption by 15.9 AF (1.6% of the mean) and expenditures by $775 (1.8% of the mean). A 10% increase in aquifer depth reduces consumption by 67 AF (6.6% of the mean), implying farmers value avoiding deeper aquifers at $2,014 annually. We use these estimates to parametrize a dynamic model of adaptive irrigation and groundwater depletion, simulated forward using standard climate projections. Under RCP4.5 (RCP8.5), the aquifer is 68% (162%) deeper by end-of-century than it would be absent climate change. A "natural scarcity" effect — whereby deeper aquifers raise costs and reduce extraction — mitigates roughly 71% (62%) of climate-driven depletion. Hotter temperatures and aquifer depletion reduce farmer welfare by $1.1 billion ($3.5 billion) per year by end-of-century. This represents 4% (14%) of current aggregate crop revenues in the San Joaquin Valley. Finally, we use our simulation model to solve for the socially optimal path of groundwater pumping, where pumping taxes internalize the open-access externality. This lets us assess whether climate change exacerbates the long-run stock externality. We contribute to literatures on climate adaptation and water resource dynamics, by showing direct evidence that groundwater pumping is a valuable but self-limiting form of adaptation.

This paper investigates the impact of the Association of Boxing Commissions and Combative Sports’ 2017 changes to the 10-9 judging criteria of MMA fights, intended “to evolve” judging to focus on the result of action (i.e., damage) when selecting round winners. Utilizing detailed round-by-round fighter performance statistics from the UFC and 25 years of Nevada scorecards, findings suggest that judges increased their scoring weight for knockdowns and choke attempts, which damage the opponent, and decreased the weight given to certain fight outcomes more commonly associated with positional changes or a lack of damage – sweeps, non-slam takedowns, and missed power strikes to the head and body. Previously documented biases towards heavier favorites and the fighter who won the most recent round remain supported in the data.

Understanding performance under pressure is critical to strategic management and behavioral economics. Existing literature often relies on sports with direct opponent interaction, such as tennis or soccer. In those settings, an individual’s performance is linked to the opponent’s defensive capabilities, making it difficult to disentangle pure psychological pressure from physical opposition. This paper isolates the impact of competitive pressure by analyzing professional golf match play, where we focus on putting, a setting defined by parallel competition without direct interference. We examine how performance fluctuates based on two key variables: the competitive state (leading versus trailing) and the uncertainty surrounding an opponent’s potential result. By utilizing golf as a natural experiment, we remove the confounding variable of direct defense, offering a cleaner model for business competition where firms must execute internal strategies while facing external competitive benchmarks. The findings shed light on how expectations and outcome uncertainty drive performance when the only variable is the player’s response to pressure.

Both economists and the public at large believe that entry by rival sports leagues results in skyrocketing salaries. The examples we commonly use, however - from the American League in 1901 to the USFL in 1983 – pertain to team sports. The sport that has seen the most recent rival entry, professional golf, follows a very different business model. Rather than pay pre-determined salaries, players on the PGA Tour participate in a series of rank-order tournaments, where the total prize money is determined by the tournament sponsors. We apply difference-in-differences estimation to data from 2016 through 2025 to determine the impact of the entry of LIV Golf on the prize money offered on the PGA Tour.

I build on previous research on risk taking decisions in this paper, in which I compare outcomes in elite professional golf tournaments conducted across different states of prices for wayward shots. The US Open is one of the most penal tournaments in which golfers compete, and two recent championships, the 2019 and 2021 versions, were conducted at golf courses that also host annual PGA Tour events. In several dimensions of play, the outcome of golfers’ shots exhibit evidence of more risk averse decision making on their part. Their behavior is consistent with economic theory, which posits that harsher penalties for wayward risk taking will cause agents to adopt more conservative strategies.

This study provides an assessment of the BRICS countries’ ambition to restructure the global political-economic-social architectures (PESAs) in this modern era of fractures and/or fragility in Western-led global alliances and institutions. Current data showed that the BRICS countries have higher gross domestic products than the G7 countries; and the BRICS aspires to restructure the global PESAs with the ultimate aim of overtaking the G7 countries. This paper asserts that the assessment of the governance quality performance in the G7 and the BRICS countries is one of the empirical tools with which research scholars can measure whether the BRICS’ ambition to restructure the global PESAs will succeed in this modern era. Given the BRICS’ desire to shift the global power once held by the G7 countries since the formation in 1975, this study is guided by some pertinent research questions predicated on governance quality performance. Do the BRICS countries have stronger and more powerful PESAs than the G7 countries? Can the BRICS countries restructure the existing global PESAs in this modern era? To answer these questions, we use the Worldwide Governance Indicators to construct three indices of governance quality performance: Political Governance Quality Performance (PGQP) Index, Economic Governance Quality Performance (EGQP) Index, and Social Governance Quality Performance (SGQP) Index. Utilizing the difference-in-means statistical method for analysis, the estimated results indicate that the G7 countries have stronger PGQP, FEGQP, and SGQP than the BRICS countries. To further comprehend the governance quality performance, we provide the country-specific bar charts to display the governance quality performance pathways for each country. All the G7 countries operate in the functionally effective governance (FEG) performance pathway regarding PGQP, EGQP, and SGQP while the BRICS countries, except South African, operate in the functionally ineffective governance (FIG) performance pathway. Among the BRICS, South Africa has its PGQP, EGQP, and SGQP in the FEG pathway while the United Arab Emirates and Saudi Arabia show similar pattern. Importantly, the estimated results reveal the BRICS as the incohesive bloc with respect to governance quality performance; therefore, they face challenges in restructuring the global PESAs given the growing concerns of China-centric driven globalization.

How Western donors respond to China’s expanding development finance remains contested, with competing hypotheses and limited systematic evidence. This study estimates the effect of Chinese aid on bilateral official development assistance (ODA) provided by donors in the OECD Development Assistance Committee (OECD‑DAC). Using a Poisson pseudo‑maximum likelihood estimator on a four‑dimensional panel covering 31 donors, 130 recipients, and 13 sectors from 2001 to 2019, the analysis exploits within‑recipient‑sector‑year variation in Chinese aid shocks and incorporates an extensive set of multi‑way fixed effects to address endogeneity concerns. While no average competitive response is detected across all donors, we find consistent evidence that Japan systematically increased its ODA commitments in reaction to Chinese engagement, amounting to an estimated US$ 5.4 billion, or 2.5% of Japan’s total ODA commitments in our sample during the study period. Japan’s competitive responses are concentrated in geographically proximate and more democratic recipients, consistent with its geopolitical and normative priorities. No comparable response is detected for other major OECD‑DAC donors, including the United States, Germany, France, and the United Kingdom. Taken together, the results show that Japan’s behaviour illustrates how a traditional donor can strategically deploy ODA as part of a broader foreign policy and industrial strategy, but the scale of its response remains modest. Combined with the muted reactions of other donors, this suggests that the OECD‑DAC system is more resilient to China’s emergence as a major donor than often assumed.

This study explores how global geopolitical risk (GPR) influences commodity price dynamics through a multi-method econometric lens. Focusing on the prices of crude oil, gasoline, and precious metals, we investigate whether commodity markets respond asymmetrically to increases versus decreases in geopolitical tensions, how these effects unfold over time, and whether their influence persists across different historical periods. To do so, we apply a nonlinear autoregressive distributed lag model to identify short- and long-run asymmetries, a vector autoregression framework with impulse response functions to trace the dynamic transmission of GPR shocks, and time-varying Granger causality tests using both rolling and recursive windows to capture shifts in the predictive power of GPR over time. Our results reveal that positive GPR shocks exert a stronger and more persistent influence on commodity prices than negative shocks, particularly for oil. The dynamic and time-varying analyses further show that GPRs impact intensifies during major geopolitical crises but is episodic and commodity-specific. These findings underscore the importance of modeling geopolitical risk as an asymmetric and evolving force in global commodity markets, with implications for energy policy, investment strategy, and market risk management.

We examine bank supervision as an information-production process using confidential supervisory ratings for U.S. bank holding companies from 2005 to 2023. We construct a benchmark model using prior ratings, observable financial characteristics, and market data, and analyze deviations from this benchmark as a measure of supervisory innovation. The benchmark explains 70–80 percent of rating variation, indicating that supervision is predominantly systematic and grounded in observable fundamentals. We show that ratings deviate more from the benchmark when examiners have greater access to private information. We also document persistent examiner-level heterogeneity in stringency and procyclical rating patterns that are independent of information access. These supervisory innovations have real effects on bank behavior: unexpected stringency reduces asset growth and lending while increasing capitalization; unexpected leniency is associated with higher future risk-taking.

Forthcoming

This study examines the impact of supervisory factors on de novo banks’ initial charter choice. We find that banks from states with higher supervisory fees, those lacking readily accessible state supervisory cost information, and those located farther from their state regulators are more likely to opt for a national charter. These results suggest that supervisory cost, transparency, and proximity significantly influence new banks’ initial charter decisions. Our findings offer valuable insights into strategies for fostering de novo bank formation and have important implications for the design and effectiveness of banking supervision.

Why do some U.S. counties generate more new businesses than others, even after accounting for differences in income, education, and population? This paper argues that long-run civic institutional continuity, not initial historical conditions and not contemporaneous religious activity, is a measurable predictor of modern entrepreneurial intensity at the county level. The paper introduces a Church Persistence Index linking county-level congregation counts from the Association of Religion Data Archives (ARDA) across four time points: 1890, 1990, 2010, and 2020. Each year's church density is median-standardized, winsorized, and weighted by its average Spearman rank correlation with the other years, so that counties with genuinely consistent institutional presence receive higher scores than those with a single historical peak. The index is standardized to mean zero and unit standard deviation across approximately 3,000 U.S. counties covering 82–95% of the continental United States. The primary outcome is new establishment births per capita in 2022 from the Census Bureau's Business Dynamics Statistics, a measure that captures entrepreneurial intensity at the local level and avoids the compositional problems of net entry. OLS regressions include controls for 1989 median household income, 1990 educational attainment, 2010 log population, and state fixed effects, with HC3 robust standard errors. A one-standard-deviation increase in persistence is associated with 5.75% more new establishment births per capita. The paper makes three contributions. First, it demonstrates that institutional continuity, not initial conditions, drives the result: 1890 church density alone is statistically indistinguishable from zero once long-run persistence is controlled, directly challenging the deep-roots literature's emphasis on historical starting points. Second, mechanism tests show that persistence strongly predicts modern educational attainment, and education predicts firm births, but only 9% of the total persistence effect is mediated through formal schooling, the remainder likely reflects trust, social networks, and civic coordination capacity. Third, the persistence index framework is generalizable to any repeatedly measured, spatially granular civic institution. Results are robust to alternative index constructions, inclusion of contemporaneous church density, drop-one-control exercises, and substitution of historical for modern population controls.

During the studio system in Hollywood, the largest and most successful studios were vertically integrated through production, distribution, and exhibition. In 1948, vertically integrated studios were forced to divest from their theaters following United States v. Paramount Pictures. This paper examines the effects of vertical disintegration on product characteristics in the U.S. film industry. Using difference-in-differences models, I compare vertically integrated studios to non-integrated studios before and after the Paramount Case. I also analyze short-term and medium-term stock market responses surrounding the decision. The analysis uses film-level and firm-level data from the American film industry spanning the 1930s through the 1950s, including information on film budgets, genres, star participation, and studio stock prices. I find that vertical disintegration affected film characteristics, particularly genre selection and the use of star talent. I also find that the post-case distribution of film budgets exhibited fewer extremely high-budget productions, a result that differs from prior expectations in the literature. These findings provide suggestive evidence that changes in product characteristics are connected to risk management by studios. Overall, the paper contributes to the literature on vertical integration and antitrust policy by showing that vertical disintegration can alter product characteristics, offering a relevant lesson with policy implications for current antitrust debates.

Overconfidence bias is among the most studied phenomena in behavioral finance, linked to excessive trading and capital misallocation. Existing literature largely treats it as a fixed in- dividual trait, paying little attention to whether the visual environment in which trading occurs can itself intensify the bias. This paper asks whether color framing in retail trading platforms amplifies overconfidence in novice traders and pushes position sizing away from statistically optimal decisions. The growth of retail fintech platforms has lowered barriers to market participation, draw- ing in users with little prior trading experience. College-age investors are of particular rele- vance given their representation among new retail market entrants and their exposure to gam- ified digital environments, suggesting potential susceptibility to interface-induced behavioral distortions. The study employs an incentivized laboratory experiment in which color framing is the sole treatment variable. Participants are randomly assigned to either a high-contrast red and green interface, replicating a palette common in retail trading platforms, or a neutral interface, with all other task elements held identical. We hypothesize that high-contrast color framing will push position sizing above the Kelly-optimal level, producing overbetting consistent with overconfidence amplification. Overconfidence is operationalized as mean deviation from the Kelly fraction, disaggregated to distinguish systematic overbetting from random noise. If color framing shifts position sizing in the predicted direction, the implication is that a single design choice may function as a policy variable capable of disadvantaging less experi- enced market participants.

Arts education is widely considered a vital force in student development, yet it remains highly vulnerable to district budget cuts. This study leverages a natural experiment in Massachusetts, where high school graduation requirements are mandated at the district rather than the state level, to quantify the impact of arts requirements on academic and behavioral outcomes. Analyzing 222 public school districts, we utilize ordinary least squares (OLS) and a Post-Double-Selection (PDS) Lasso framework to control for high-dimensional characteristics. We find that while arts mandates successfully increase student arts participation by 6 percentage points, they have no statistically significant effect on the global sample. However, critical tradeoffs emerge when stratifying by district types. Small, Title I-eligible districts experience a 5 percentage point increase in chronic absenteeism under an arts requirement, indicating resource constraints and program crowding. Conversely, large, affluent districts see a minor but significant 0.8 percentage point reduction in four-year graduation rates. Ultimately, curriculum requirements can create structural tradeoffs with scheduling flexibility and funding that inadvertently hinder student success.

We examine the effects of the state price density (SPD) and absolute risk aversion, as implied by crude oil options and futures markets, on domestic oil drilling activity. We measure the SPD and risk aversion across the distribution of returns, allowing us to separately identify the cost of hedging declining oil prices from the cost of hedging rising oil prices. Our results suggest that an increase in the cost of hedging oil prices results in a decline in domestic oil production. The results suggests that it is not just uncertainty that matters for producers, but also the risk premium associated with the uncertainty, or equivalently the cost of hedging uncertainty. We find some evidence of asymmetry in the response of drilling to the changes in the SPD and risk aversion. The response of rigs is stronger to changes in SPD or ARA in the region of declining oil returns, compared to the region of increasing oil returns. This implies that producers respond more strongly when there is a shock to the cost of hedging lower oil prices, relative to a shock to the cost of hedging higher oil prices. The results are consistent with the role of the U.S. as a major oil producer. Our results are economically and statistically significant. For example, in our baseline specification, a one standard deviation increase in the SPD in the region of declining returns, results in a 2.120% cumulative decline in drilling rigs over a 13 week period. To highlight the asymmetry, we find that a one standard deviation increase in the SPD in the region of increasing returns results in a 1.623% cumulative decline in drilling rigs over the same timeframe. When we use absolute risk aversion instead of the SPD, the the cumulative impacts are 1.724% in the region of declining returns and 1.389% in the region of increasing returns.

The Generalized Axiom of Revealed Preference (GARP) is necessary and sufficient for a set of consumer choices to be rationalized by a consistent utility function. However, when a data set fails to satisfy GARP the violations may not be statistically significant. We propose a new nonparametric test to evaluate if a violation of GARP is statistically significant. The new nonparametric test is evaluated using simulated data from a utility function that satisfies GARP, but with measurement error added to the prices, quantities, and expenditure. As the amount of measurement error increases in the simulated data the number of violations of revealed preference increases and the new test procedure performs well with one, five, and ten percent measurement error but is less precise in detecting the true unitality function when measurement error is high a twenty percent. Using financial monetary data, the test fails to detect that the data are consistent with GARP.

Over the past decade, global macroeconomic performance has been increasingly shaped by two major sources of uncertainty: geopolitical risk and global supply chain pressures. As countries become more integrated through global production networks, the combination of heightened geopolitical tensions and growing interdependence has amplified concerns about the macroeconomic consequences of these shocks. Using panel local projections with sign-restricted identification for 35 countries from 1998Q1 to 2024Q1, we find that while both shocks are contractionary, supply chain disruptions are roughly twice as damaging to GDP and three times as inflationary as geopolitical shocks. Geopolitical shocks primarily induce demand-driven slowdowns, whereas supply chain shocks trigger persistent supply-side disruptions. The effects vary across income groups: advanced economies display stronger and more sustained inflationary responses, while emerging markets suffer deeper and longer-lasting output losses. Advanced economies currencies appreciate at medium horizons whereas emerging markets exchange rates depreciate persistently. These differences reflect structural factors such as trade exposure, policy credibility, and financial constraints. Our findings highlight the need for tailored stabilization strategies that account for both the source of uncertainty and country-specific vulnerabilities.

This paper is the first to construct a green bond-augmented monetary aggregate, enabling us to examine how central banks can advance a climate-aligned agenda without compromising their primary mandate of price stability and economic growth. Early evidence by Jawadi et al. (2024) suggests that the Federal Reserve may already be responding to climate-related developments, despite its public and explicit rejection of any climate-policing role. Our contribution offers both a rigorous theoretical foundation and robust empirical evidence showing that information embedded in climate-driven green finance markets can (i) improve the measurement of monetary policy without requiring an explicit climate mandate, and (ii) enhance the effectiveness of those core objectives themselves. Our theoretical and empirical results indicate that the green bond market serves as a conduit for transmitting climate-related information into the conduct of monetary policy, while monetary policy actions in turn exert measurable effects on the green bond market. Moreover, our findings suggest that disregarding the informational content of the green bond market introduces a systematic measurement distortion that is correlated with the policy stance, a welfare distortion that is likely to intensify as the market share of green bonds continues to grow.

This paper examines how oil news shocks, measured as unexpected changes in oil futures prices around OPEC announcements, affect U.S. consumer sentiment and spending. The central question is whether expectation-driven oil market news, rather than realized supply disruptions, transmits to households through changes in inflation expectations, perceived purchasing power, and consumption behavior. To address this question, I estimate a monthly proxy-VAR model that uses high-frequency oil futures price surprises as an external instrument to identify oil news shocks. The baseline analysis uses U.S. data from 1978:1 to 2019:12 and is extended through 2024 to assess post-pandemic dynamics. The results show that oil news shocks lead to a delayed decline in both the Index of Consumer Sentiment (ICS) and real Personal Consumption Expenditures (PCE), with negative effects emerging approximately one year after the shock. A decomposition of sentiment indicates that these responses are driven by persistent declines in expectations about personal finances, business conditions, and buying conditions, alongside an immediate increase in inflation expectations. Disaggregated consumption responses reveal heterogeneous effects across categories: durable goods spending rises briefly before turning negative, while nondurables, services, and energy-related spending decline immediately and persistently. The paper contributes by linking oil news shocks to detailed measures of household expectations and consumption behavior, and by identifying the expectation channels, particularly inflation expectations and perceived purchasing power, through which these shocks propagate. Overall, the findings highlight the importance of expectation-driven oil market disturbances for understanding consumer behavior and macroeconomic dynamics.

Portfolio diversification is as important to debt management as it is to asset management. In this paper, we focus on diversification of sovereign debt issuance through greater reliance on inflation-indexed bonds for an emerging economy, Colombia. Using an arbitrage-free dynamic term structure model of fixed-coupon and inflation-indexed bond prices, we account for inflation and liquidity risk premia and calculate the forward-looking net benefit of issuing inflation-indexed bonds over nominal bonds. Our results suggest that the Colombian government could lower its funding costs by as much as 0.69 percent through increased issuance of inflation-indexed debt.

This paper examines the transmission of monetary policy to corporate debt markets through the debt refinancing channel. We show that firms with greater refinancing risk earn higher subsequent bond returns on average. However, this positive risk premium disappears during periods of elevated interest rates and monetary tightening. When firms approaching maturity walls are forced to refinance maturing debt at substantially higher borrowing costs, the resulting rollover losses increase financial distress and depress subsequent bond returns. These adverse effects are particularly pronounced among firms with low cash holdings, high leverage, and speculative-grade credit ratings. We further demonstrate that this monetary policy transmission mechanism is distinct from the liquidity-shock channel emphasized in prior studies of rollover risk and that its effects extend to the primary bond market. Overall, we contribute to the literature by demonstrating that refinancing risk constitutes a state-dependent transmission mechanism of monetary policy.

Theory predicts that yield shocks trigger large-scale investor rebalancing --- flights from risky to safe assets. We show that in the European sovereign bond market, these flights are impeded: following stress events, trading volume falls by 30\% over the following week rather than increasing. The mechanism is a collapse in liquidity supply. Bid-ask spreads rise by 56\% and market depth falls sharply, preventing investors from rebalancing. We identify the supply channel using the fact that on the MTS platform, the same bond trades on both a domestic and a pan-European market. Since domestic market makers hold disproportionately more of their own sovereign's bonds, stress events reduce their risk-bearing capacity more severely. Consistent with this, we find that bid-ask spreads widen more, depth falls more, and trading volume drops more on the domestic platform. These results show that the concentrated exposure of bank-dealers to sovereign risk can freeze a central secondary market precisely when rebalancing is most needed.

We study the role that cross-border firm-to-firm credit plays in financing exporters. Exploiting the exogenous shock of US tariffs on Chinese goods in 2018–2019, we examine the response of Colombian firms - bystanders not targeted by trade policy - to redirected US demand. Using credit registry information for cross-border and domestic non-financial firm financing, we find that almost 40 percent of the total credit sourced by exporters came from cross-border firm-to-firm credit at end-2019, which represented 80 percent of their cross-border credit. In contrast to traditional trade credit, which is typically short-term, firm-to-firm credit has an average maturity of almost 2 years, and has characteristics resembling bank lending. Our findings highlight an overlooked financial channel underpinning the international trade network.

We study how households adjust consumption following the introduction of a loan- to-value (LTV) restriction. Using population-wide electronic expenditure records and administrative data, we find that renters increase consumption after the policy, the opposite of the intended saving response. As home purchases become more difficult, prospective buyers delay both homeownership and the savings needed for a down pay- ment. This effect is robust to renter-only comparisons and is strongest in regions where the LTV restriction is more likely to bind. At the same time, the policy operates as in- tended among renters who transition into homeownership under the constraint, as these homebuyers reduce consumption before buying and keep it persistently lower thereafter. Our findings show that borrower-based macroprudential policies can reshape intertem- poral behavior in unintended ways, with important implications for homeownership, aggregate demand, and the transmission of housing regulation to the real economy.

While a large literature has examined how bank credit shocks affect firms or households, it has not accounted for the fact that such shocks may simultaneously impact both. In this paper, we overcome this limitation and disentangle the real impact of a credit market disruption into the effect of firm-side credit shocks, individual-side credit shocks, and their interaction. To this end, we construct a novel dataset linking Norwegian employees to their employers and their respective bank relationships. We show that individuals’ labor income and consumption decline by 1–2% when only they or only their employer face a credit shock, compared to the benchmark where neither do. However, when individuals and their employer simultaneously face a credit shock, labor income and consumption decline by nearly 6%, revealing a strong amplification effect. This amplification arises because personal credit constraints hinder individuals’ consumption smoothing and job search when confronted with wage cuts or layoffs triggered by their employer’s credit constraints. Our findings suggest that this mechanism also shapes the aggregate transmission of credit shocks.

This paper examines why legal central bank independence (CBI) indices that measure de jure CBI may fail to capture monetary autonomy as exercised in practice. Taiwan presents a theoretically important puzzle: despite relatively low scores on widely used legal CBI measures, particularly the Cukierman-Webb-Neyapti legal index, the Central Bank of the Republic of China (Taiwan) has maintained substantial policy autonomy, monetary credibility, and institutional continuity from the 1990s through the 2020s. Using Taiwan as a theory-building qualitative case study, the paper argues that de facto CBI is produced through institutional power distributions and rules-in-use. Institutional power distributions refer to how Formal Power, Informal Power, and Traditional Power are allocated within monetary governance. Formal Power is codified authority; Informal Power is relational, reputational, informational, and strategic authority; and Traditional Power is historically embedded authority. Rules-in-use explain how these powers are interpreted, applied, negotiated, and exercised in practice. Drawing on archival documents, policy records, secondary literature, and elite interviews, the study shows how Taiwan’s de facto autonomy has been sustained despite relatively low legal-index scores. The paper contributes to the CBI literature by developing a broader institutional conception of de facto CBI. It argues that legal CBI scores may understate actual monetary autonomy when professional authority, information asymmetry, institutional memory, seniority, political acumen, and rules-in-use become institutionally consequential in practice.

Behavioral finance has stressed the role that sentiment sometimes plays in financial behavior. While available measures of sentiment offer a combination of the types of pure sentiment considered in behavioral finance with other factors which influence actors’ expectations. It is worthwhile to consider what role these indices play in influencing various types of financial behavior. Most of the applications to date in the area of finance have been to the behavior of financial markets. A few studies have investigated the effects of sentiment on specific types of credit in the US but none of which we are aware have investigated aggregate credit growth and its main components in a unified framework. That is the purpose of this paper. Controlling for standard economic factors we investigate the effects of a number of often used measures of sentiment on US credit growth across five credit categories: total private credit, residential mortgage credit, commercial and industrial loans, corporate debt, and consumer credit. Our sentiment indices include survey-based, text-based, and market-based measures. We also consider an index of economic policy uncertainty. The results show that sentiment effects are largely credit-type specific. Lending standards, as measured by the senior loan officer opinion survey, are the most consistent supply-side driver, significant across four of the five credit categories. News sentiment has the broadest demand-side reach, significant across four categories. Small business optimism explains the most variation in residential mortgage credit, an unexpected finding that points to a household-business balance sheet channel. No single index dominates across all categories. These findings suggest that the choice of sentiment measures should be matched to the specific credit market of interest rather than relying on a single index.

We study how foreign exchange reserve holdings affect firms’ cost of international borrowing. Using transaction-level bond issuance data from 94 countries, we document that higher reserve buffers are associated with lower international borrowing spreads. The effect operates through reduced currency volatility and lower crisis risk, which improve investors’ assessment of currency-related repayment risk. The impact is stronger in countries with weaker macro fundamentals, lower sovereign currency credit quality, and more active foreign exchange intervention, as well as among financially weaker firms and during periods of financial stress. The results highlight the role of reserve policy in shaping firms’ access to international capital markets.

This paper investigates the optimal design of deposit insurance coverage limits. Specifically, we address the policy question of whether a cap on deposit insurance payouts—whereby the insurance authority pays failing bank depositors less than the bank’s promised short-term return—improves social welfare. We construct a modified Diamond-Dybvig framework characterized by a continuum of banks, aggregate uncertainty regarding bank failures, and costly asset liquidation. Distinct from recent literature, our model incorporates a privately funded, ex-ante deposit insurance scheme, sequential servicing, and endogenous bank contracts that react to regulatory constraints. We characterize the global optimum for a social planner and demonstrate that this allocation can be decentralized via a bank contract, although the system remains susceptible to a bank panic Nash equilibrium. Our primary contribution lies in the analysis of coverage caps. By solving a constrained optimization problem where the insurance payout is allowed to be less than or equal to the promised consumption, we prove that the optimal policy requires full coverage. Consequently, imposing a cap on deposit insurance is suboptimal in this environment. Finally, we extend the model to include heterogeneous endowments. We demonstrate that our main result is robust to this extension, showing that the optimality of full coverage persists even when depositors differ in their initial wealth.

This paper investigates whether global Basel regulatory standards meaningfully improve bank stability, or whether national institutional quality ultimately determines regulatory outcomes. While Basel I to III were designed to harmonize capital regulation and strengthen systemic resilience, empirical evidence indicates substantial cross-country variation in enforcement and effectiveness. This research asks: Do Basel Accords reduce non-performing loans (NPLs), and is their impact contingent on institutional strength? The paper contributes to the literature by explicitly modeling the interaction between global regulation and national governance, bridging two strands of research that have largely developed separately. Prior Basel studies focus on capital adequacy and risk weighted reforms, while institutional finance research highlights legal origins and governance capacity as determinants of financial performance. Yet no empirical study has directly tested whether Basel’s effectiveness depends on institutional quality, particularly in the post Basel III era. To address this gap, I construct a cross-country panel dataset covering 2000–2023 that merges Basel implementation measures from BIS reports with institutional quality indices from the World Governance Indicators, along with NPL ratios from different datasets. The empirical strategy employs country and year fixed-effects models to estimate within-country effects of regulatory adoption, supplemented with staggered DID estimators to account for variation in Basel implementation timing across jurisdictions. The core specification includes an interaction term between Basel adoption and institutional quality to test whether regulatory effectiveness is conditional on governance strength. The expected geographic scope includes both advanced and emerging economies, enabling comparison between countries with strong institutional environments and those characterized by weaker enforcement capacity. Preliminary expectations suggest that Basel adoption is associated with improvements in asset quality primarily in countries with strong institutional environments, while effects may be negligible where governance is weak. The findings aim to inform policy discussions on the viability of global regulatory harmonization and the importance of institutional strength for achieving bank stability. The results may have implications for Basel IV and for emerging weak institution economies seeking to signal credibility through regulatory adoption.

Whether banks with higher ESG ratings exhibit lower systemic risk is still an open question that demands further investigation. The existing results are mixed and mainly dependent on how the banking risk is measured, whether the source of the risk is systematic or idiosyncratic, and which sources of the ESG rating data are used. This paper advances the existing literature by first examining the impact of the joint and separate effects of the ESG ratings on a battery of risk indices represented by the cost of capital, Merton’s (1974) distance-to-default and expected default frequency, and levered and unlevered equity betas. Second, and more specifically, ESG rating data is separately obtained from two different, and most frequently used, ESG rating agencies, the London Stock Exchange Group (LSEG), formerly known as Refinitiv, and the MSCI, formerly known as Morgan Stanley Capital International. Applying instrumental variable approaches to a panel data of 242 US banks from 2016 to 2023 and accounting for potential endogeneity arising from reverse causality, omitted variables, and measurement errors, the results provide partial support for the argument that banks with higher ESG ratings exhibit lower cost of capital, while the direction and significance of the relationship appear to vary considerably both by the sources of the data (LSEG vs. MSCI) and across the different formulations of the risk indices. Among the components of the ESG scores, banks with higher environmental ratings tend to experience lower cost of capital and lower betas under both rating specifications. Using the quadratic version of the ESG scores, the results provide strong support for the existence of nonlinearities and idiosyncratic risk effects highlighting the presence of the “over-investment behavior” by banks found by previous studies.

Using loan-level regulatory data, this paper finds that banks reduce risk-taking in their loan portfolio after bank-level uncertainty rises. Banks also increase interest rates and require collateral in loan contracts after positive uncertainty shocks, especially among banks with less regulatory capital that are also subject to mark-to-market accounting regulations. In response, public firms dependent on bank credit reduce investment, employment and hoard liquidity. Private firms turn to shorter duration and expensive credit, like credit cards and trade credit, increasing illiquidity. Local employment also declines after bank uncertainty increases. Taken together, uncertainty shocks in the banking system significantly affect the economy.

This paper studies the environmental consequences of airport-level Zero Emission Vehicle (ZEV) regulations in the United States. We examine how these policies affect pass-through dynamics, airline pricing, service quality, network responses, and consumer welfare.We exploit ZEV adoption across airports and implement reduced-form analysis to estimate the effect of ZEV mandates on charging fees, airline operating costs, ticket prices, delay time, and route adjustment. We find that ZEV adoption increases airport charging fees, and airlines partially pass through these costs to passengers. Our results quantify the distribution of benefits and costs across consumers, airlines, and local communities, and inform the optimal design of environmental policy in networked transportation markets.

We argue that mergers can function as a form of market access, enabling acquired products to “export” into new markets by leveraging the acquirer’s distribution infrastructure. We develop a model in which post-merger expansion depends on two key product characteristics: similarity to the acquirer’s portfolio (which governs cannibalization) and latent demand. Using mergers in the U.S. spirits industry from 2010–2019, we find that a one standard deviation increase in similarity is associated with approximately 30% less post-merger store coverage, while a one standard deviation increase in latent demand is associated with approximately 27% more coverage. Effects are amplified for foreign acquisitions. Our findings suggest that variety gains from mergers are concentrated in low-similarity, high-demand products.

Understanding volatility connectedness of G7 housing markets is essential for assessing global financial stability. This paper investigated dynamic risk transmission among G7 housing markets from 1970 to 2025 using a dynamic conditional correlation generalized autoregressive conditional heteroskedasticity framework. Total connectedness index surged amid major crises. Net pairwise directional connectedness revealed a structural asymmetry. Canada and United Kingdom acted as net total directional connectedness (NET) transmitters, whereas United States market, characterized by its substantial depth, and endogenously absorbed external shocks as a volatility sink. Contagion tests confirmed that market synchronization during periods of distress reflects fundamental interdependence. Panel analysis identified heterogeneous drivers of volatility connectedness. While funding costs served as primary catalysts for markets with distinct directional pathways. Real GDP growth and valuation soundness emerged as critical determinants for markets with NET ambiguous group. Connecting empirical volatility dynamics with regulatory efficacy revealed inherent constraints of static macroprudential regimes. This approach ensures that policy responses are sensitive to both fundamental macro-financial drivers and valuation risks, thereby strengthening the structural resilience of global housing market against vulnerabilities arising from financial liberalization.

This paper studies how Social Security transforms a common payroll-tax reform into different effective labor wedges across older workers, and whether those wedge differ- ences change retirement behavior. Standard policy analysis often treats payroll-tax changes as uniform incentive shifts, but for seniors the relevant object is not the statutory tax rate itself. It is the effective wedge on work after the benefit formula, claiming rules, and the Retirement Earnings Test are taken into account. Exploiting the 2011–2012 U.S. payroll-tax holiday with monthly SIPP data on workers ages 62–70, we find that retire- ment responses vary systematically with baseline tax exposure, claiming status, and the strength of marginal benefit accrual. Already-claimed workers respond most, consistent with facing a near-pure tax. A dynamic retirement model calibrated to pre-reform mo- ments shows that the Retirement Earnings Test is a major barrier to pre-FRA work, and that benefit linkage sustains late-career employment. The contribution is to show that, for seniors, the incidence of payroll-tax reform depends on Social Security’s institutional mapping from the statutory tax rate to the distribution of effective wedges, not on the statutory rate alone.

We study how transportation access shapes women’s labor market outcomes in Korea, moving beyond large metros to mid-sized cities and suburban/rural areas. First, we document stylized facts linking lower access to longer commutes and lower labor force participation (LFP), disaggregating by marital/children status and by mode (subway, bus, car). Second, we construct mode-specific accessibility indices and gender-specific accessibility that weight job opportunities by female employment shares and work-hour flexibility. Third, leveraging exogenous network expansions (subway/bus/road interchanges), we estimate causal effects of improved availability on women’s LFP, hours, earnings, and wages using staggered event-study/DID designs. We decompose total impacts into (a) composition via in-/out-migration, (b) behavioral responses of incumbents (including mode switching), and (c) demand-side changes using ancillary establishment and vacancy data. Results will clarify where (rural/suburban/urban) access gains are largest and for whom they matter most, informing cost-effective transit and land-use policies that alleviate commuting burdens for mothers.

Across both domestic and international contexts, national and sub-national governments have crafted tax policies to attract or retain residents, motivated at least in part by the belief that individuals and families move between jurisdictions to minimize their tax burden. One common example of such policies in the United States are state-level elderly income tax benefits, or tax policies that reduce the tax burden on elderly individuals as a function of their age or type of income. This research uses confidential administrative tax records from 1999 through 2022 and a range of empirical approaches to estimate how interstate elderly migration responds to these policies. First, we use state-to-state migration flows to estimate the impact of changes in elderly tax benefits on inter-state migration, looking at the near-universe of tax changes over the first two decades of the 21st century. We find that while elderly filers are marginally more responsive to overall after-tax rates than their younger counterparts, there is no evidence they respond to elderly-specific tax incentives. Next, we study several highly visible changes to elderly tax preferences that occurred at different times in different states, using a difference-in-difference approach to estimate both the migration response to and cost of these policies. These analyses suggest at most modest migration responses that fall well short of compensating for the first order impact of these tax changes on revenues. Our preliminary findings thus suggest that elderly-specific tax incentives do not serve as an effective means to attract or retain elderly residents, and rather operate as a transfer to and within the elderly population.

The gender gap in higher education has reversed over the past several decades, with women now comprising a substantial majority of college graduates. This paper investigates the implications of this shift for marriage market dynamics by examining how the declining ratio of college-educated men to women affects marriage rates among college educated women. To estimate the causal impact, we implement an instrumental variable (IV) strategy that exploits variation in state-level funding cuts to public colleges—an exogenous shock that influenced gender-specific college enrollment patterns but is plausibly unrelated to marriage decisions. In the first stage, we use data from the Integrated Postsecondary Education Data System (IPEDS) to predict changes in the female-to-male college enrollment and completion ratios driven by these funding reductions. In the second stage, we link these predicted values to marriage outcomes using data from the American Community Survey (ACS) and the Current Population Survey (CPS). The findings indicate that a relative decline in the availability of similarly educated men significantly reduces marriage rates among college-educated women, highlighting how shifts in educational attainment can constrain partner availability and alter patterns of union formation.

As early as 2012, individual states implemented streamlined billing and healthcare routines that facilitated providing immediate postpartum long-acting reversible contraception, or LARC devices, to new mothers. Providing immediate contraceptive options post-birth reduces barriers to additional doctor visits for busy new mothers. In mid-2016, the Department of Health and Human Services Centers for Medicare and Medicaid Services released an Information Bulletin detailing payment and policy rules that all states could use to similarly streamline billing and payment issues that had hampered their ability to provide LARCs immediately postpartum. In the following years, dozens of states adopted these policies and increased the dispensing of LARC devices postpartum. Using data on the years of state-level Medicaid policy changes and data from the CDC on Medicaid-covered births, I examine the impact of increased availability of LARC devices for postpartum women affects the total number of higher-order births and the interval time between births for existing mothers post the policy change.

We estimate the effect of a major expansion of in-person college education in California prisons on the disciplinary records, participation in rehabilitative programs, and employment of incarcerated students. In 2014, California allowed public colleges and universities to receive funding for courses offered inside of prisons. Since then, college availability has expanded from one prison to all but one prison in the California system. Using person-level administrative data, we estimate a fuzzy-triple-difference design that leverages differences in the timing and extent of college expansion across prisons, as well as differences in inmates’ propensity to attend college courses on the basis of their characteristics upon incarceration. This is a new approach in the literature that more plausibly addresses unobserved differences between participants and non-participants, as well as between prisons with large and small programs, than do previous studies. We find that college attendance substantially reduces disciplinary incidents, and past enrollment in college courses increases participation in other rehabilitative programs. However, we find no evidence for a reduction in disciplinary incidents while enrolled in college, nor do we find evidence that college attendance affects the likelihood of holding a high-value job while incarcerated.

This paper examines the effects of the 2019 universalization of Mexico’s Social Pension Program (PAM), one of the country’s most expansive and politically salient social programs. The reform simultaneously increased the cash transfer and extended eligibility to all individuals aged 65 and over, regardless of income or contributory pension status. Using nationally representative data from the ENIGH and a triple-differences (DDD) identification strategy, we estimate the causal effect of the universalization on poverty and labor market outcomes. Our empirical approach exploits variation across time (pre- and post-reform), age (eligible vs. ineligible), and pension scheme status (non-contributory vs. contributory), allowing us to separate the effects of expanded eligibility from those of increased benefit levels. We find strong increases in take-up rates and no significant change in overall poverty rates, suggesting that many new beneficiaries were not economically vulnerable. However, we document a surprising increase in extreme poverty, concentrated among low-income elderly who responded to the reform by exiting the labor force. This reduction in labor supply, driven by a significant drop in employment among individuals in the bottom income quartile, suggests that the pension acted as a substitute for labor income rather than a supplement. Taken together, the results highlight the trade-offs inherent in universal pension programs: while broader access reduces administrative exclusion, extending transfers to economically secure individuals may dilute redistributive impacts and generate behavioral responses that offset potential welfare gains.

In 2005, the Vietnamese government implemented a free health insurance program for children under the age of six. This paper investigates the long-term impact of this program on children's educational attainment. Using nationally representative data and leveraging differences in exposure across cohorts, I find that the program increases the likelihood of completing primary and lower secondary education. I also find that the effects are particularly pronounced for children in rural areas. These findings suggest that the free health insurance program for children can serve as an effective tool not only for improving access to healthcare but also for fostering human capital accumulation.

I analyze the impact of one of the largest waves of land redistribution in Mexico on agricultural productivity. Leveraging newly digitized agricultural census data in a difference-in-differences design, I compare outcomes between municipalities with higher and lower rates of land reform from 1930 to 1950 to estimate the effects on agricultural production. I find that an increase in \textit{ejido} land significantly reduces the value of agricultural capital and expenditures in the medium-run. Though the magnitudes of these effects indicate a moderate impact, consistent underinvestment in the agricultural sector likely reduced productivity over time. These findings complement previous studies by offering evidence on the mechanisms underlying the negative long-run impacts of land reform on development indicators in Mexico.

Abstract: In an open economy New Keynesian model, a higher import tariff induces a central bank with a dual (narrow) mandate to reduce the relative weight on the domestic inflation rate in the target rule that underpins optimal discretionary monetary policy. Owing to this, the gain from better stabilization of the output gap is traded one-for-one for higher inflation variability, leaving overall welfare unchanged. Under a broader mandate that includes real exchange rate stability, a central bank’s losses actually decrease as the import tariff increases. The gains are modest and derive from the central bank showing greater concern in the expanded target rule for the output gap at the expense of inflation and the real exchange rate. The output-inflation variability trade-off improves under a broad mandate.

This paper investigates the cross-country heterogeneity of the New Keynesian Phillips Curve (NKPC) across European Monetary Union (EMU). Utilizing a comprehensive dataset of Eurozone economies, we employ a dual econometric approach—combining panel data analysis and individual country-level time series estimation—to identify structural variations in inflation dynamics. Our analysis focuses on two primary dimensions of heterogeneity: the slope of the Phillips Curve (the sensitivity of inflation to economic slack) and the mechanisms of inflation expectation formation (the weight of forward-looking versus backward-looking behavior). A key contribution of this study is the rigorous treatment of model and parameter uncertainty. Beyond standard aggregate models, we draw on recent methodological contributions from regional and intranational studies to examine cross-country variations in expectation formation. Preliminary results reveal previously undocumented layers of heterogeneity within the Eurozone’s inflation dynamics.

In August 2020, the Federal Reserve replaced Flexible Inflation Targeting (FIT) with Flexible Average Inflation Targeting (FAIT), introducing make-up strategies that allow inflation to temporarily exceed the 2% target. Using a synthetic control approach, we estimate that FAIT raised CPI inflation by about 1 percentage point and core CPI inflation by 0.5 percentage points, suggesting a moderate impact net of food and energy and a largely temporary effect. Short- to medium-term inflation expectations increased by approximately 0.8 percentage points, while long-term expectations remained anchored. The effects of FAIT on economic activity were, if anything, minimal. Our results are robust across multiple specifications, including alternative price indices, synthetic control estimators, control groups and adjustments for global supply chain pressures, economic activity, fiscal policy, commodity prices, interest rates and monetary aggregates. The differing macroeconomic outcomes under FAIT versus a counterfactual FIT characterized by moderate inflationary effects, negligible real effects and anchored long-term expectations, are consistent with the hypothesis of a steeper-than-expected post-pandemic Phillips curve in the New Keynesian model.

The Opportunity Zone program was designed to encourage investment in distressed communities across the United States. Early research found no evidence of impacts of the program on employment, earnings, or poverty of zone residents, but some evidence of positive effects on employment among businesses in zones. Using the latest survey-based as well as administrative data, we adopt a longer-run and more comprehensive perspective on the labor market impacts of OZs. We find that OZ designation increases job creation among businesses within zones (“workplace employment”). However, newly created jobs in zones are offset nearly one-to-one by declines in nearby low-income tracts. While we detect gains in OZ resident employment over the longer run, the increase comes from jobs with workplaces outside of OZs that, in light of the changing demographic composition of zones, are likely held by new as opposed to existing residents. Overall, our results suggest that OZs have limited benefits for existing residents of targeted areas and are associated mainly with spatial reallocation of jobs and residential location. Finally, the gains in workplace employment occur in urban but not rural areas.

This paper examines how the economic effects of population aging vary across space. Using U.S. county data from 2000 to 2020, we estimate the impact of changes in the share of adults aged 60 and older on GDP per capita growth. We instrument for aging using lagged age structure from historic censuses to address endogeneity. We find that a 10 percent increase in the 60+ population share reduces GDP per capita by roughly 4–5 percent, but this average masks substantial spatial heterogeneity. Aging has little measurable effect in large, economically integrated labor markets, while generating sizable growth declines in smaller and more isolated regions. Sectoral analysis reveals contractions in agriculture, construction, and manufacturing alongside expansions in health care and professional services; however, only economically integrated counties appear able to offset sectoral contraction and avoid aggregate losses. These results suggest that the local consequences of demographic aging depend not only on industrial structure but also on regional integration and labor market connectivity.

Entrepreneurship serves as an engine for economic growth and is a source of employment opportunities in both developed and developing countries. However, the relationship between entrepreneurship and unemployment remains ambiguous. This study explores this relationship by estimating the causal effect of total early-stage entrepreneurial activities (TEA) on the unemployment rate. We estimate this effect across 29 countries in Europe from 2001 to 2022 using several identification strategies. Our preferred model finds that a one percentage point increase in TEA leads to a decrease in the unemployment rate by 0.427 percentage points. This indicates that an increase in entrepreneurial activity significantly reduces the unemployment rate. Additionally, our study examines the effects of four startup policy initiatives put in place by the European Union (EU) over our sample period, using a Difference-in-Differences (DID) approach. We find that the DID estimates reveal no statistically significant effect for any policies observed, due to the timing and effectiveness of the policy interventions. Finally, our study leverages the theoretical Okun’s law framework to create an alternative measure of entrepreneurship, "firm growth". Our analysis reveals that an increase in firm growth is associated with a decrease in the unemployment rate. This reinforces the robustness of our main findings.

The demographic transition, a combined effect on modern society's demographics through lower fertility and mortality, has shaped an unprecedented demographic structure in most advanced countries. Under this new structure, mature cohorts (e.g., 40 to 59 years old) are more populous than their younger counterparts (20 to 39 years old). In the economy, mature cohorts tend to save, while younger cohorts tend to borrow. This change in the composition of savers and borrowers creates a saving glut, pushing down interest rates and driving up asset prices, particularly housing prices in metropolitan areas. This housing boom is often cited as a potential culprit of extremely low fertility in the lowest-low fertility countries, particularly East Asian countries, and also poses a threat to financial stability because most home buyers borrow to finance their purchases of new homes. Against this backdrop, we propose a new policy measure to address the potentially harmful price surge in the metropolitan housing market. As the price surge is driven by excess demand for properties, not their use, a bond indexed to the housing prices of a particular area will quench speculative demand for properties. As the government collects property taxes, it can reliably issue bonds backed by the future tax revenue from the properties. Since housing supply cannot respond instantaneously to a sudden surge in demand, which would otherwise lead to speculative bubbles, timely bond supply will reduce price bubbles in the housing market. Potential home buyers can buy a bond for the neighborhood where they plan to buy a home, rather than the physical property. To show the effectiveness of this policy measure, we incorporate this bond in an overlapping generations model with a housing market. The price surge under the demographic transition becomes more moderate with the housing price-indexed bond. The welfare analysis finds that the younger generation's welfare improves under the new bond, as they have a measure to protect themselves against a housing price surge, and housing prices rise more moderately.

This paper studies how origin-based social networks shape migration decisions by influencing the resilience of informal institutions. Using panel data on Chinese households between 1986 and 2008 and exploiting exogenous variation in rural migration induced by agricultural price shocks, I show that temporary rural-urban migration significantly weakens cross-household risk sharing within villages, but only in villages without active kinship networks. A theoretical framework distinguishing between limited commitment and information frictions, together with empirical tests using household transfer patterns, shows that this deterioration does not arise from a reduced ability to enforce informal transfer obligations on migrant households, but instead primarily reflects information problems that limit communities’ ability to insure migrants whose incomes are harder to observe. These findings highlight the role of kinship networks in facilitating information flows across geographic distances and identify an important channel through which strong origin networks can encourage migration: by reducing the potential loss of informal insurance associated with migration. Consistent with this mechanism, as China relaxed administrative restrictions on temporary migration, villages with active kinship networks experienced significantly larger increases in migration, with the strongest effects among households facing greater income risk and therefore most averse to losing the benefits of informal insurance.

In a decentralized setting, the state is comprised of both the central and local governments, each with its own distinct and sometimes conflicting preferences. To what extent do local government adhere to central government policies? And what explains the variation in policy compliance among local authorities? We argue that local governments challenge central directives when they conflict with local interests. They are more likely to conform to central policies when they perceive high uncertainties from the central government, while delaying compliance to prioritize local interests in times of low uncertainties. We test our theory by investigating the local variations in the implementation of land marketization policies in China using a unique dataset we compiled from millions of land transaction records, firm data, and administrative data from 2007 to 2013. We find that heightened local compliance following President Xi's assumption of power in 2012 and the presence of the Central Inspection Team in selective localities in 2013, both indicative of high uncertainties. Conversely, there exists a large variation in local compliance following the central directive prohibiting non-market approaches to transfer non-agricultural land, which is characterized by low uncertainties. Additionally, we find that higher impact of State-Owned Enterprises on local economies causes the degree of delayed local compliance.

In a post land reform environment, violence and redistribution are substitutes for an opportunistic incumbent attempting industrialization. We setup a theoretical model to explore this relationship for individual and group level violence across two types of policy driven post land-reform industrialization (PLRI)— small and large-scale, in the Indian state of West Ben- gal. We find strong evidence for our theoretical prediction of an inverse relationship between land redistribution and both types of violence during the period of small scale industrialization after an industrial policy was announced in 1994-95 by the incumbent. This relationship however breaks down for individual level violence during attempts at large-scale industrialization between 2006-2011, even though both types of violence increase in this phase.

This paper examines three landmark eminent domain cases—Poletown Neighborhood Council v. Detroit, County of Wayne v. Hathcock, and Kelo v. New London—often cited as evidence of inconsistent judicial outcomes. Using a comparative case analysis, it introduces Contextual Cost–Benefit Analysis (CCBA), Using a comparative case analysis, it introduces Contextual Cost–Benefit Analysis (CCBA), a framework that interprets takings decisions as pragmatic evaluations in which costs and benefits are weighted by their contextual salience. The analysis shows that courts tend to uphold takings when projected benefits credibly outweigh associated costs, and to reject them when the tradeoff appears unfavorable. This satisficing logic, grounded in Simonian bounded rationality and Northian institutional constraints, resembles the heuristic structure of the Hand Rule in judicial reasoning rather than the marginal calculus of standard economic models. Beyond clarifying how courts balance doctrinal structure with policy reasoning, CCBA demonstrates the limits of optimization-based economic reasoning in explaining legal decision-making in practice.

Objective: This study examines disparities in pain treatment by analyzing prescription opioid use among adults born in the United States (US) and first-generation immigrants. We evaluate whether differences in healthcare access contribute to lower opioid receipt among immigrants, whether the disparity changes with improved access, and how length of residence in the US shapes these patterns. The goal is to inform equitable pain management policies and clinical practice. Methods: Using nationally representative data from 2005 to 2020, we estimate probit models to assess the association between immigrant status and the probability of receiving a prescription opioid. The analysis proceeds in stages: estimating overall differences in opioid prescribing; testing whether immigrants have lower healthcare access; evaluating whether access explains the disparity; examining interactions between immigrant status and access; and assessing differences by years lived in the United States. We compute average marginal effects to quantify differences in opioid receipt across groups and access scenarios. Results: Immigrants are substantially less likely than US-born adults to receive prescription opioids. Although immigrants have lower rates of health insurance coverage and usual source of care, adjusting for these factors explains a small share of the disparity. The difference in opioid receipt increases, rather than decreases, at higher levels of access, and the largest gaps are observed among immigrants with long-term residence who have access to healthcare. These findings indicate that equal access does not lead to equal treatment. Conclusions: Policies that focus solely on expanding healthcare coverage or connecting patients to regular care are unlikely to eliminate disparities in pain treatment. Reducing inequities in opioid prescribing will require approaches that address clinical decision-making, provider communication, and cultural and linguistic barriers that persist once patients enter the healthcare system. Ensuring equitable pain management will require integrating these considerations into clinical guidelines, training, and quality improvement initiatives.

The nationwide expansion of Medicare’s Home Health Value-Based Purchasing (HHVBP) model links agency reimbursement to performance on hospitalization rates and functional improvement measures. While the program aims to enhance care quality and reduce Medicare spending, its performance benchmarks may systematically disadvantage agencies serving cognitively impaired and medically complex patients, for whom functional gains are more difficult to achieve. This paper examines whether HHVBP generates differential penalties for agencies facing greater patient complexity. I combine publicly available HHVBP performance files with Home Health Compare data collected between January 2018 and March 2025 to construct a baseline agency-level complexity index. In the absence of diagnosis-level data, complexity is proxied by the dispersion of functional improvement outcomes across mobility and self-care domains. Agencies in the top quartile of this distribution are classified as high complexity. Using event study and difference-in-differences designs with agency, state, and time fixed effects, I estimate the dynamic impact of HHVBP on performance outcomes. The results show that high-complexity agencies experience disproportionately larger declines in rewarded performance measures following the nationwide implementation. The findings are consistent with incentive misalignment in value-based payment design and raise concerns about unintended consequences for access and equity among medically complex aging populations.

This paper examines early career outcomes of the Class of 2017 from the top 57 doctoral programs in economics. Focusing on those with initial tenure-track (TT) placements, I find that women are 12.6% less likely than men to receive tenure 8 years after graduation. Although this gap is slightly lower than the 15.8% observed from the Class of 2008 (Chen et al., 2022), the gender difference in tenure and promotion (T&P) remains significant in the profession. The preliminary results also show that the female ratio in the doctoral program has a positive effect on T&P, as does having a publication or a revise-and-resubmit (R&R) in a top 50 journal during graduate studies. Receiving teaching awards as a graduate student also improves T&P prospects in the U.S., especially for international students. Finally, a closer investigation indicates that the gender gap mainly stems from TT positions outside the U.S. Consistent with recent evidence that the gender disparity has diminished in initial job placement, including TT positions (Chen, 2025), I find no gender gap in the T&P among U.S. faculty, regardless of citizenship. This finding highlights encouraging progress in the profession. However, it calls for continued institutional support for junior female faculty globally and for greater recognition of the persistent challenges women face in advancing their careers beyond the hiring stage.

This paper examines whether state-mandated mental health education in K–12 schools improves adolescent mental health and reduces disparities across racial, ethnic, and socioeconomic groups. Adolescents from marginalized communities face elevated risks of depression, suicidal ideation, and related mental health challenges due to structural racism, cultural stressors, and economic hardship. At the same time, they are less likely to receive timely mental health care because of stigma, limited access to providers, and financial and institutional barriers. School-based mental health education mandates represent a potentially scalable and equity-enhancing policy intervention, designed to improve mental health literacy, normalize help-seeking, and promote early identification of mental health concerns among youth. We exploit policy variation from three early-adopting states—New York and Virginia, which enacted mandates in 2018, and Florida, which followed in 2019—to estimate the causal effects of these mandates on adolescent mental health outcomes. Using a quasi-experimental difference-in-differences design, we compare changes in mental health outcomes for youth in mandate states before and after policy implementation to contemporaneous trends in demographically and economically similar comparison states without such requirements. This approach allows us to isolate the impact of mandated instruction from broader national trends in adolescent mental health. Our empirical analysis draws on two large, nationally representative datasets: the National Survey on Drug Use and Health (1990–2023) and the National Survey of Children’s Health (2016–2023). We focus on adolescents ages 12–17 and examine outcomes including depressive symptoms, suicidal ideation, and mental health help-seeking behaviors. We also assess heterogeneous effects by race/ethnicity, family income, and parental education to evaluate whether mandates narrow or widen existing disparities. This study contributes to the growing literature on social policy and youth mental health by providing rigorous, policy-relevant evidence on the effectiveness of school-based mental health education. By focusing explicitly on distributional impacts, the findings will inform debates over whether education-based interventions can serve as effective tools for reducing mental health inequality and promoting adolescent well-being at scale in the United States.

We re-visit the nexus between environmental regulation and FDI inflow in an efficiency-oriented framework. Specifically, we examine the influencing role of environmental regulation on environmentally-oriented FDI inflow efficiency—a country’s ability to attract FDI to the largest extent possible while proportionally reducing emissions. We use a panel dataset of 40 countries for the period 2002-2020 and adopt Directional Distance Function within the framework of Data Envelopment Analysis to estimate FDI inflow efficiency - relative performance of the countries in attracting FDI compared to the best performers in the group. A two-way fixed effects panel regression model is then used to estimate the impact of environmental regulation on FDI inflow efficiency. Our results show insignificant impact of environmental regulation on FDI inflow efficiency. However, economic freedom consisting of size of the government, legal system and property rights, market regulation, significantly improves host countries’ FDI inflow efficiency. Size of host country’s market also plays an important role on FDI attractiveness. Our results are robust to endogeneity due to both reverse causality and omitted variables. Based on our empirical findings, we suggest that an economy can attract FDI not only by lowering environmental standard and acting as pollution haven, but also by strengthening the government’s role and enhancing the economic freedom that investors like to enjoy in the host economies.

The European Union’s Carbon Border Adjustment Mechanism (CBAM), scheduled for full implementation in 2026, is intended to reduce carbon leakage and strengthen global climate mitigation incentives through carbon-related border adjustments. As a long-run climate policy, its effects depend on international trade linkages, evolving climate policies, forward-looking investment responses, structural transformation, and climate–damage feedback. Despite its dynamic nature, the long-run global effects of CBAM remain insufficiently quantified. This paper develops a dynamic spatial integrated assessment model (IAM) that integrates forward-looking investment, sector-level bilateral trade, endogenous climate dynamics, and climate–economy feedback within a unified framework to evaluate the long-run effects of CBAM. The model features 13 global regions and multiple production sectors, allowing for endogenous capital accumulation, structural transformation, and policy-induced abatement. CBAM is modeled as a border tariff based on carbon intensity and abatement rates, with its effects shaped by cross-regional interactions and policy environments. The framework allows comparisons across alternative assumptions regarding capital adjustment, CBAM coverage, and future climate policy paths. The results reveal several key long-run patterns. First, under baseline scenarios, CBAM reduces cumulative global emissions and shifts the global temperature trajectory downward, although the effects remain modest. Second, these effects become substantially stronger when capital adjusts endogenously: relative to a static benchmark with fixed capital paths, CBAM generates larger emissions reductions and more pronounced welfare effects as investment responds to policy incentives. Third, the impacts of CBAM increase significantly as more countries and sectors adopt similar mechanisms, reflecting stronger general equilibrium and expectation-driven effects. The analysis also highlights substantial regional heterogeneity. Developed economies generally experience persistent welfare gains and structural shifts toward non-agricultural sectors, while developing economies face persistent welfare losses and greater pressure on agricultural employment. Most global emissions reductions occur in developing regions, implying an uneven distribution of mitigation burdens. Overall, the paper provides a quantitative framework for evaluating the long-run environmental, distributional, and macroeconomic consequences of border carbon policies in an interconnected global economy.

Forthcoming

Digitalization and the rapid diffusion of artificial intelligence (AI) are transforming energy systems worldwide, yet their net climate impact remains unclear. Does new technologies and digitalization increase carbon emissions through increased electricity demand or enable decarbonization through efficiency gains? Using a global panel of 184 countries from 2000-2021, we examine how digitalization affects power-sector CO2 emissions per capita. We construct a digitalization index and employ fixed effects models, heterogeneity analyses, and a difference-in-differences design around the 2015 Paris Agreement. Results reveal three main patterns. First, higher digitalization is associated with increased power-sector emissions. Second, this effect is substantially stronger in fossil-intensive electricity systems. Third, in highly digitalized countries with cleaner electricity systems, digitalization instead contributes to lower power-sector emissions, a finding reinforced by post-Paris Agreement trends. The findings suggest that digitalization and by extension AI adoption is not inherently green. Its climate impact depends on crossing both a digital readiness threshold and a decarbonization threshold. This highlights the importance of complementary energy policies in shaping whether digital transformation accelerates or hinders climate mitigation.

Forthcoming

Despite well-documented returns to female athletic participation in health, income, and labor market outcomes, women remain significantly underrepresented in college athletics relative to their enrollment shares. We analyze the institutional and economic determinants of female participation using an unbalanced panel of approximately 1,700 US college and university athletic programs over 2005-2021, drawn from the Equity in Athletics Disclosure Act (EADA). Our data span all five major competitive classifications - NCAA Divisions I, II, and III, the NAIA, and the NJCAA -providing the first comprehensive documentation of female athletic participation across the full spectrum of US intercollegiate athletics. Female participation grew from approximately 200,000 athletes in 2005 to nearly 300,000 in 2021, with the female participation rate rising from roughly 5 to 7 percent of female enrollment across all classifications. Panel models with school, classification, and year fixed effects show that female enrollment is the strongest predictor of women's teams and participants, consistent with the Title IX proportionality mechanism. State employment is positively associated with female participation, while state income and population are negatively associated, consistent with a labor market returns channel. These findings establish a pre-NIL and pre-House v. NCAA baseline for evaluating the gender equity consequences of the new commercial environment in college athletics.

The Covid-19 pandemic substantially disrupted the college experience, leading many students to defer admission or take leaves of absence during the 2020–2021 academic year. Yet not all dimensions of college life were equally affected. Place-based aspects of the college experience, such as athletics and campus social life, were more directly disrupted by pandemic restrictions than academic credit accumulation itself. This paper examines whether enrollment at colleges and universities with a stronger athletic focus responded differently to the pandemic than enrollment at otherwise similar institutions. We measure institutional sports focus using multiple sources, including win-loss records, Directors’ Cup points, athletic department expenditures, and Niche.com athletics ratings. We also account for other place-based dimensions of campus life, including social life. Finally, we examine whether pandemic-related enrollment disruptions at sports-centric institutions affected subsequent athletic performance, testing whether changes in enrollment translated into changes in on-field outcomes.

College sports is being rebuilt in real time. NIL deals, donor driven collectives, and the transfer portal have created a functioning athlete labor market long before the NCAA has figured out how to regulate one. At the same time, revenue sharing proposals, escalating media rights contracts, and the rapid normalization of sports betting are reshaping the financial incentives that drive conferences and institutions. This paper maps how these forces interact and why they are pushing collegiate athletics toward a fully market based model faster than policymakers anticipated. We show how NIL structures increasingly operate as de facto compensation systems, how the transfer portal amplifies competitive mobility, and how media rights realignment concentrates resources in ways that widen the gap between conferences. We also examine how legalized sports betting introduces new integrity, data rights, and compliance pressures that cut across all of these domains. The result is a landscape defined less by amateurism than by market coordination problems. We close by outlining the regulatory choices that will determine whether this emerging system stabilizes or fractures further.

Forthcoming

This paper examines how banks adjust their loan pricing when firms have a higher share of index fund investors as shareholders. Using syndicated loan data, we find that loan spreads increase with passive ownership and provide evidence consistent with higher loan spreads reflecting increased risk due to reduced shareholder oversight. Supporting this interpretation, we find stronger effects among firms in which shareholder oversight is more important in disciplining management. However, the increase in spreads is not fully accounted for by changes in firm risk. Further analysis points towards banks increasing their monitoring efforts in response to changes in shareholder composition.

We study how shortfalls in banks’ capital and liquidity ratios affect their supply of credit to U.S. corporations. Using micro (bank)-level survey data on changes in lending standards and terms, we find that shortfalls in bank capital (relative to their internal targets) translate into significantly tighter lending standards and price terms – especially during crisis times. Capital shortfalls also hinder banks’ provision of credit. However, liquidity shortfalls do not appear to have a consistent effect on banks’ credit supply.

Forthcoming

Forthcoming

Across the developed world, private credit has emerged as the fastest growing alternative source of financing, growing from less than 10% to nearly one-third of non investment grade corporate debt in the United States. Yet the exact nature of borrowing firms and private credit's impact on firm dynamics remains unknown, leaving critical questions about its role within the structure of the financial system and financial stability unanswered. This paper addresses both using a novel database, merging the largest private credit deal data with U.S. Census microdata. Borrowing firms are typically mature, larger than the median U.S. firm, and concentrated in intangible intensive industries. Using a synthetic control identification strategy, we document a fundamental tradeoff: private credit raises firm employment and innovation while simultaneously increasing firm exit likelihood. These effects are driven by a relaxation of financing constraints rather than substitution away from other financing sources, with elevated exit risk reflecting tighter creditor control. The mechanism enabling this relaxation is industry specialization, whereby concentrating in a focused set of sectors allows them to extend credit where other financing sources cannot.

We study how private equity (PE) sponsors influence loan contracting through their relationships with bank’s legal counsel (“lender law firm”) in leveraged buyouts (LBO). We show that stronger PE–Lender Law Firm relationships are associated with fewer loan covenants. This result is robust when we instrument for the relationship using geographic distance between PE sponsors and law firms. Loans involving relationship lender law firms also feature higher default rates and higher interest spread. Lead banks that use a sponsor’s relationship law firm are more likely to receive future deal mandate from the sponsor. Overall, we highlight PE sponsors’ ties to lender law firms as an important channel of PE bargaining power in buyout financing. Our findings also raise concerns that lead banks, competing for deal mandate and operating under an originate-to-distribute model, may have weakened incentives to insist on stringent negotiation.

We document that life insurers have expanded their provision of private credit over the past decade, totaling $849 billion, or 14%, on life insurers' balance sheets in 2024. A substantial part of the growth stems from private credit extension to financial borrowers and to privately placed asset-backed securities. We document that private equity-owned (PE-owned) life insurers drive these trends. We also provide evidence the growth of these investments accounts for 61% of PE-owned insurers' annuity market share increase and that PE-owned insurers have more access to these investment through affiliated issuers. The results are concentrated in the indexed annuity market for which the new forms of private credit investments facilitate a better maturity and cashflow match.

How do housing markets process salient local information such as crime? This paper provides the first joint estimates of attention elasticity and memory decay in housing price capitalization, offering micro-foundations for behavioral information processing in asset markets. Using 310,000 residential transactions in Cook County, Illinois (2018--2023), I exploit triennial property tax reassessment cycles as exogenous shocks to buyer attention and employ repeat-sales and spatial designs to isolate cognitive from compositional dynamics. To address endogenous crime exposure, I apply the bunching-based control function of Caetano, Caetano, and Nielsen (2024), which exploits mass points at zero crime to correct for spatial sorting. I examine seven crime types across four distance bands (100m-1000m) and four temporal windows (3-12 months prior to sale). Three core findings emerge. First, crime capitalization is strongly attention-dependent: price discounts are 30% larger during reassessment years, when buyers are prompted to evaluate property values. This attention elasticity (0.31) validates Sims' (2003) rational inattention framework-price adjustment depends on when agents attend, not solely on fundamentals. Second, within-property repeat-sales estimates reveal a 31% decline in price sensitivity to crime within nine months, implying behavioral memory decay rather than permanent learning. Decomposing total cross-sectional decay (55%), I show that 57% reflects genuine forgetting and 43% reflects compositional selection-lower-quality properties sell when crime is salient, while higher-quality ones wait. Third, a spatial test reveals pronounced overweighting: price effects decay 97% from 100m to 1000m, far exceeding the 14% decay in actual crime correlation, evidencing localized salience rather than rational risk assessment. Together, these results reconcile Linden and Rockoff's (2008) identification challenge-distinguishing risk-based from salience-driven capitalization-and bridge rational inattention with salience theory (Bordalo, Gennaioli, and Shleifer, 2012). The findings imply that housing markets exhibit bounded information efficiency, where attention timing and salience mediate how fundamentals capitalize into asset prices. Beyond housing, the framework formalizes the behavioral lifecycle of local “micro-narratives”, offering a cognitive basis for narrative economics (Shiller, 2019). Policy implications follow: reducing the visibility of crime may mitigate its economic impact as effectively as reducing its incidence, underscoring the importance of information environments in spatial market behavior.

Inclusionary Zoning requires developers to set aside a share of new units at affordable rates to low/moderate income tenants. Nearly 900 cities have adopted an inclusionary ordinance. In California, more than a third of localities have adopted an ordinance, with the number of associated affordable units rivaling the largest federal housing policies. Yet economic theory tells us these policies act as a tax on construction, and will reduce the overall supply of housing, potentially reducing affordability in the long run. Quantifying this trade off has been difficult due to heterogeneous policy, data quality issues, and the endogeneity of local inclusionary zoning adoption. I estimate the effect of inclusionary zoning using novel administrative data covering all cities and counties in California. The data covers more than 500 localities over up to a 30 year period. It measures more than 20 parameters of a city’s inclusionary ordinance, which allows me to construct an aggregate measure of stringency. I perform a difference-in-difference estimation around a recent change in California policy. I leverage a statewide change in preempting state law for identification, performing a difference-in-difference around an exogenous reactivation of local inclusionary ordinances. I find the typical inclusionary zoning ordinance reduces annual new residential construction by 5%. Parallel trends leading up to policy reactivation reassuringly suggest the necessary identifying assumption holds. Accounting for the stringency of policies turns out to be critical. Heterogeneous effects by policy stringency may explain divergent results in the prior literature. I also examine the interaction between inclusionary zoning ordinances and the Low Income Housing Tax Credit (LIHTC), the largest federal affordable housing policy. Developers may use the same low income units they built to comply with local inclusionary zoning policy to qualify for LIHTC. Indeed, I estimate that adoption of a typical inclusionary ordinance increases LIHTC funding by more than 30%. This softens inclusionary ordinances’ reduction in local housing supply. Note that because there is a fixed pool of credits to be allocated, this is occurring at the expense of development in other jurisdictions.

As the short-term rental market has grown, some cities seek to regulate it to boost access to affordable housing. We analyze the effects on the rental market of the most severe short- term rental policy to date: New York City’s de facto short-term rental ban. The efficacy of such policies depends on a city’s policies like rent controls that make long-term renting unattractive. Property owners may explore other avenues, such as gray markets, before converting units to long-term rentals. Using a synthetic difference-in-differences approach and Zillow's Observed Rent Index, we find no statistically significant evidence that rents fell because of the ban. At best, rents fell by less than two percent. Additionally, we find some evidence that rental listings fell rather than rose following the ban. Together, these results suggest owners may be exploring other options besides converting units to long-term rentals.

The increasing intensity of natural disasters underscores the importance of current mitigating strategies. However, how do existing strategies affect recovery trajectories remain less discovered. This study investigates the economic impacts of evacuation and displacement in Typhoon Morakot which caused more than 700 deaths in Taiwan in 2009. Applying the instrumental variable method, this study identifies the effects of evacuation in the short run and displacement in the long run. Our findings on the short-term and long-term impacts of evacuation are diverse. After one year of the strike, evacuation has no effect on transitory post-strike income. In the long run, the displacement effect is negative, accounting for nearly one-third of victims; average income. Some splitting sample analyses are also performed. Based on the results this study address suggestions about recovery strategies.

This study examines whether stronger gender equality attitudes are associated with more objective and impartial views toward non-heterosexual groups (hereafter referred to as LGBT+). Over recent decades, Taiwan has promoted substantive gender equality through anti-discrimination policies and expanded educational resources. While these efforts aim to foster respect and equal treatment, it remains an open question whether the diffusion of gender equality norms reduces traditional stereotypes toward LGBT+ populations. Using data from the ``Gender Group’’ modules of the Taiwan Social Change Survey Project (TSCS) for 2012 and 2022, this study conducts quantitative regression analyses based on survey items capturing gender equality and gender diversity attitudes. The results show that support for shared economic responsibilities, joint caregiving, parental leave sharing, and women’s autonomy in work and family choices is significantly associated with lower levels of bias toward gender diversity. These beliefs mitigate common stereotypes, such as negative views regarding homosexuals’ private lives, marriage, gender expression, and parenting. The analysis further indicates that older individuals, men, supporters of the People First Party and the Kuomintang, and those with frequent religious participation exhibit higher levels of bias. In contrast, higher educational attainment is strongly associated with greater recognition and acceptance of gender diversity. Overall, the findings suggest that progress in gender equality is accompanied by broader social acceptance of gender diversity, though substantial heterogeneity persists across demographic, political, and religious groups. From a policy perspective, sustained implementation of gender equality legislation -- such as the Gender Equality Education Act and the Act of Gender Equality in Employment -- can promote inclusion not only by improving women’s status but also through norm diffusion. Strengthening gender equality and diversity education across all schooling levels, together with targeted public outreach, is essential to reduce persistent stereotypes and achieve more comprehensive social inclusion.

This paper explores how culturally inherited beliefs about gender equity may influence men’s relative representation in pink-collar, nursing, and social-work occupations. We find that increases in gender equity are associated with lower relative representation of men in pink-collar, nursing, and social-work occupations, a result we dub the pink paradox. Our preferred estimates suggest that a one standard deviation increase in gender equity is associated with a 3.8 percentage point (14.6%) decrease in the relative probability a man has a pink collar occupation, a 1.7 percentage point (50%) decrease in the relative probability he works in nursing, and a 0.11 percentage point (22%) decrease in the relative probability he works in social work. While surprising, we find these results on multiple, independent samples and across numerous alternative specifications. Our findings suggest that men from more gender equitable backgrounds avoid occupations with higher female shares and highlight potential unintended consequences of increasing gender equity for occupational sex segregation.

Niederle and Vesterlund (2007) argue that one reason women earn, on average, less than men is that women tend to avoid competition. Jurajda and Münich (2011) and Ors et al. (2013) study the performance of university candidates in competitive entrance exams and find that women are less likely than men to succeed. Blau and Kahn (2017) find that psychological attributes have a small-to-moderate impact on the gender pay gap in the US. Other papers find that women perform similarly to men in competitive environments (Lavy, 2013). In particular, several papers found that high-ability women are as competitive as men (Atkinson et al., 2003; Paserman, 2022). We use a unique dataset to examine the likelihood that women take a competitive test and their performance in it. Our dataset contains information on all students who graduated from Israeli universities from 1995 to 2021. For each student, the database contains a rich set of socio-demographic information, as well as high school and university GPAs, and post-graduation salaries. Notably, the dataset includes the students' Psychometric Entrance Test (PET) scores. The PET is the Israeli equivalent of the US SAT, and it carries considerable weight in university admissions decisions, particularly for prestigious programs. The PET exams are well-suited for testing the competitiveness hypothesis, as grades are awarded based on relative performance among peers, and for examining whether women underperform in highly competitive environments. We find that, first, conditional on high school GPAs, women achieve lower PET scores. Second, we find that, conditional on the first PET score, women are less likely to retake the PET. Third, we find that both differences are minor among high-ability women. To rule out alternative explanations, we examine men and women who had similar incentives to retake the PET and graduated from the same programs. To determine whether our results are driven by women having lower motivation to attend prestigious programs, we examine students' university GPAs. Finally, we control for the possibility that women avoid taking a second PET because they do not want to delay their studies. Our results hold after controlling for these possible explanations.

The recent spread of online sports betting in the United States has led to a rapidly expanding literature which estimates impacts of this access on a wide variety of outcomes. We show, however, that much of the variation in access to online sports betting is confounded with access to online casino gaming. Additionally, a majority of industry revenues across important subsets of treated states come from casino gaming rather than sports betting. We then replicate one early published study estimating the impacts of access to mobile sports betting on self-reported mental health, and illustrate the degree to which their results are robust to the exclusion of states with legalized casino gaming. Given the data used in other studies, as well as when impacts are most pronounced relative to legalization and COVID-induced sports shutdowns in 2020, this observation calls into question whether the estimated impacts should properly be attributed to legalized sports betting.

In Murphy v. National Collegiate Athletic Association (2018), the Supreme Court struck down the federal law that barred states from legalizing sports betting. Supporters argue that LSB generates entertainment value while filling state tax coffers, boosts consumer surplus, and harmonious gambling enhances well-being. Critics counter that it exposes millions of new bettors to an activity prone to overoptimism about financial returns and self-control problems. Emerging research links LSB to reduced savings and investment, lower credit scores, higher debt, and more bankruptcies, increased leisurely activities with especially strong effects among young men and lower-income groups. There is also a rise in child maltreatment, particularly in rural areas, and unexpected losses by local sports teams increase intimate partner violence in LSB states relative more than non-LSB states. Mixed findings have emerged regarding mental health and other self-reported outcome. In this paper, we examine the effects of legalized sports gambling on public health out- comes using a staggered difference-in-differences framework and a comprehensive dataset derived from private health insurance and Medicaid claims. Our results indicate that legal- izing sports gambling is associated with a small increase in new diagnoses of gambling ad- diction and even larger increases when we include a broader set of mental health disorders and utilization of treatment. Among individuals with a prior diagnosis of gambling addiction (our “at risk” sample), legalization is linked to worsening mental health outcomes. Further- more, we document spillover effects: in households with a member at risk, legalization of sports gambling correlates with an increase in mental health diagnoses and utilization of treatment.

Internal investigations into allegations of workplace misconduct are undertaken on a confidential basis. But confidentiality cannot be assured. Investigated employees may be revealed by the investigation, including by word of mouth and by disclosure requirements to future potential employers. Based on an experiment fielded on a large nationally representative sample, this study provides the first evidence of direct employment harm to an employee investigated for workplace misconduct or sexual harassment. Subjects express considerable opposition to a callback of applicants who were investigated for workplace misconduct, even when the investigation did not find misconduct. The findings add to the evidence documenting that any association with stigma harms employment prospects.

Economic models often assume goods are either complements or substitutes. We challenge this by incorporating transaction costs into a theoretical framework that predicts utility-based complements may appear as substitutes when transaction costs are high. We empirically test our framework using retailer-level alcohol spirit sales in Washington State and their visibility, potential familiarity, and proximity to marijuana dispensaries as proxies for transaction costs in a staggered difference-in-differences design. Results reveal expected heterogeneity across our transaction cost measures, supporting our theoretical framework and highlighting their importance in identifying the relationship between goods.

This study investigates the causal effects of SNAP participation on household expenditures, with a central focus on non-food related expenditures. We examine total non-food spending as well as six subcategories of non-food expenditures using the 2000-2011 waves of the Consumer Expenditure Survey (CE). Selection into SNAP is addressed by employing an IV approach that exploits variation in state-level policies and program administration to instrument for SNAP participation, conditional on household and state-level characteristics. Respondent misreporting is addressed by adopting an approach based on parametric methods for misclassified binary dependent variables that produces consistent estimates when using instrumental variables. The analyses reveal that, after adjusting for misreporting, SNAP participation among low-income households increases expenditure on non-food by 27 percent, while increasing expenditure on food by nearly 41 percent per year. The results indicate that SNAP participation significantly increases most of the non-food subcategories. The results are generally similar but measured less precisely when separately examining households with and without children. Overall, the findings of this study indicate that SNAP is serving its intended purpose of increasing household expenditures on food, while also allowing households to allocate some of their income to various non-food spending categories.

Cultural norms fundamentally shape how individuals perceive and engage with formal institutions, affecting decisions central to economic well-being. Among these norms, family ties, which emphasize parent–child relationships and strong familial obligations, are one of the most primitive. This paper examines how culturally transmitted family ties affect participation in U.S. safety net programs. Stronger family ties may reduce reliance on public assistance by providing informal insurance or increasing stigma associated with program use, yet they may also increase needs by limiting labor mobility and expanding household size. Identifying their causal effect is challenging because family ties are closely intertwined with socioeconomic characteristics that also influence program participation. To address this, we focus on U.S.-born individuals with foreign-born parents, holding constant institutional and macroeconomic environments while exploiting plausibly exogenous variation in norms transmitted through parental country of origin. We link country‑level measures of family ties from the World Values Survey to individual data from the 1994–2025 Current Population Survey Annual Social and Economic Supplement. We estimate the effect of inherited family-tie strength on participation in a broad set of safety‑net programs, including TANF, SNAP, SSI, EITC, WIC, Medicaid, and school lunch, and classify these programs by benefit type (cash versus in kind). Our empirical strategy leverages variation in family-tie strength across and within parental national origins while controlling for predetermined individual characteristics, parental immigration attributes, and environmental covariates. We also conduct extensive heterogeneity analyses and robustness checks. Our results show no statistically significant relationship between family ties and overall safety‑net participation. However, analyses by program type reveal important differences in the effects of family ties across programs. Individuals with stronger family ties are more likely to participate in cash or near‑cash programs but are less likely to participate in in‑kind programs. Evidence on underlying mechanisms indicates that greater intra family resource sharing, rather than labor market disadvantages or differences in household composition, is the primary channel explaining these patterns. These findings underscore the importance of designing policies that account for cultural norms and informal resource sharing incentives.

Why does regional integration in Africa remain so low, and could the variation in border design during the colonial era explain it? We address these questions by developing a Border Artificiality Index (BAI) for 39 Sub-Saharan African countries that captures two theoretically distinct dimensions of border artificiality: a spatial dimension capturing geometric irregularity and misalignment with natural geographic features, and a social dimension that captures human and institutional fragmentation effects through ethnic partitioning and the disruption of precolonial political structures. Preliminary evidence suggests that these dimensions predict different economic outcomes and operate through distinct channels. The social dimension is more closely associated with weaker formal trade integration, which is consistent with its theoretical role in eroding the cross-border trust and institutional coherence that trade relationships require. The spatial dimension, which captures the physical distortion of market geographies and misorientation of infrastructure away from regional neighbors, shows stronger predictive power for structural and developmental constraints, which is consistent with its role in raising the fixed costs of regional connectivity. Validation exercises confirm strong face validity, a clear two-factor structure, and substantial correspondence with earlier measures of border artificiality. By making colonial legacies visible and measurable, we argue that colonial border artificiality is not a single inherited condition but a multidimensional legacy with distinct economic consequences. The social and spatial dimensions of that legacy bind through different mechanisms and call for different policy responses. Improving regional integration, therefore, requires interventions that target the specific channel through which inherited borders constrain economic coordination, rather than generic trade liberalization alone.

Prior literature has pointed out that Mexico’s liberalization of trade and foreign investment beginning in the 1980s affected the distribution of wages between production and non-production workers. The vertically specialized maquiladora sector attracted significant attention in this regard because of its association with outsourcing, foreign direct investment, and global value chains. This study utilizes multiple Mexican government data sources that shed light on different aspects of the relationship between wage inequality and global supply chains. The main findings are twofold. Trade and maquila activity are associated with (1) higher relative wages for non-production workers (as compared with production workers) and (2) lower relative employment of non-production workers. The first finding at first seems to support previous studies which argue that trade liberalization increased the skill premium in Mexico by raising demand for skilled labor (that literature typically considers non-production workers to be skilled). However, the second finding suggests that trade and maquila activity may have actually reduced the demand for skilled labor. These results highlight the potential importance of other mechanisms for explaining linkages between trade and wage inequality.

This paper investigates the impact of resource booms on export value, concentration, composition, and sophistication in resource-rich developing economies. Using a shift-share instrument that leverages heterogeneous exposure to Chinese demand after China’s 2001 WTO accession and the ex-ante composition of export baskets, I examine the causal effects on export baskets and sectoral employment of Brazilian local economies. The findings reveal increased export values and concentration in more exposed regions, with a shift from resource-based manufactures to primary products and declining export sophistication. Despite wage growth in primary and service sectors, primary employment remained stable while manufacturing jobs contracted, resembling a Dutch disease pattern. These results underscore the trade-offs of resource booms, where short-term gains in export value and sectoral wages may be offset by long-term development challenges. Given Brazil’s similarities to other commodity exporters, these findings may indicate similar trends emerging across developing economies.

This paper estimates bias in the Consumer Price Index (CPI) for elderly individuals in the United States using Engel curve methodology and household-level data from the Panel Study of Income Dynamics (1999–2023, unbalanced panel of 142,819 household observations). The analysis focuses on how well the CPI reflects changes in the cost of living for older Americans, with particular attention to differences by gender and marital status. Results show that from 1999–2011, the CPI overstated the increase in the cost of living for both elderly and non-elderly households, consistent with what are observed in the past literature. In contrast, from 2011–2023, the CPI understated cost increases, with the bias generally larger for non-elderly households, a new result that has not been discovered in the past. Within the elderly group from 2011-2023, the CPI understated the annual rate of increase in the cost of living for married individuals and male individuals. These findings highlight the variation in CPI bias across demographic groups and time periods, with implications for inflation-indexed programs such as Social Security. By identifying these discrepancies, the study contributes to a better understanding of how inflation affects different segments of the population and informs policy efforts aimed at improving the accuracy of cost-of-living adjustments.

An individual who is boundedly rational is more likely to sub-optimize in his lifecycle plan because he is able to foresee only up to a degree. These individuals may have planning horizons shorter than a full length and their preferences are known to be characterized by time-inconsistency. Some argue that the consumption profile of the boundedly rational agents may solve the lifecycle consumption puzzle by dissolving the discrepancy between the economic theory and the data, even in general equilibrium. This research extends this idea and tries to answer whether the prediction of the bounded rationality model is robust to the addition of leisure to the household preference. By exploring the consumption-leisure trade-off among the boundedly rational agents, this research provides a theoretic background to optimal leisure choice (thus labor supply) and retirement over a lifecycle for households who have short planning horizons, as well as who face uninsurable risk. Individuals may not make adequate decisions regarding when to retire partly because their ability to foresee into the future is limited. In this study, I contribute to the literature by examining a model that gives rise to an insight into endogenous retirement decision for the boundedly rational individuals under the environment of uninsurable risk. One may imagine that increasing retirement age will increase saving and total wealth, but decrease leisure time, and vice versa. Therefore, the trade-off between the utility gain from early retirement and the reduction of retirement benefit plays a vital role in determining the optimal retirement age. This paper looks for the necessary condition for retirement specifically when the household has a productivity shock, high or low, which may reveal only later in life, but not anticipated in advance. Because of the precautionary saving motive under the risk, retirement age for boundedly rational households can arrive sooner than the case without risk, but later than the one with full rationality under risk. This paper specifically aims to find the conditions of such outcome.

Abstract There are many challenges to achieving a circular economy. A main reason is that once a household product is malfunctioning, consumers will choose to dispose of the product and replace it with a newly manufactured product as opposed to repairing the product for further reuse. This creates many environmental challenges such as extensive raw material extraction, more industrial emissions in the production process, more greenhouse gas emissions, more landfill use, etc. There are several reasons why consumers choose to throw away their malfunctioning household product and replace it with a newly manufactured product. This paper, based on a consumer survey of 282 survey respondents, examines the reasons consumers choose the replacement method as opposed to the repair and reuse method. These reasons are ranked according to the most important to the least important reasons that consumers gave for not repairing and reusing their own household product. The results of this survey will allow policy makers to focus in on those policies that will be more effective at dealing with the problems of a throwaway society.

Does Quality of Governance Matter in the Resource–Environment Nexus? Evidence from Resource-Intensive and Non-Resource-Intensive West African Countries Abstract The current study examines the environmental effects of natural resource extraction and governance quality across 13 West African countries for the period 1986–2024. Despite the growing literature on resource–environment dynamics, only a few studies have explored the comparative effects between resource-intensive countries (RIC) and nonresource- intensive countries (NRIC), and none have accounted for the moderating role of governance quality within the West African context. For this reason, this paper focuses on the comparative analysis of West Africa’s RIC and NRIC while incorporating quality of governance (QOG) as both a direct determinant and as an institutional moderator of the natural resource–environment relationship. Unlike most resource–environment studies, this study utilizes the load capacity factor as a comprehensive environmental quality indicator. Using panel-corrected standard error (PCSE) and feasible generalized least squares (FGLS) techniques within the load capacity curve (LCC) framework, the study finds that the EKC hypothesis holds for the full sample, while the LCC hypothesis is validated only for NRIC. Natural resource rents improve ecological sustainability across both samples, but this effect is conditioned on governance quality. While in resource-intensive countries, better governance paradoxically intensifies ecological pressure through the formalisation of extraction, in non-resource-intensive countries, governance directly enhances ecological sustainability. Renewable energy consistently improves environmental quality, while industrialisation degrades it regardless of resource group. The short-run causality analysis reveals a bidirectional relationship between economic growth and ecological sustainability, and confirms that renewable energy Granger-causes economic growth. The study recommends differentiated environmental governance frameworks, sector-specific enforcement in resource-intensive economies, and expanded renewable energy investment across the sub-region. Keywords: Load capacity factor, Load capacity curve (LCC), Environmental sustainability, Natural resources, Quality of governance JEL Classification: Q01, Q32, Q56, C33

It is hoped that the provision of technology through the Clean Development Mechanism stimulates innovation and encourage developing countries to decontaminate and decarbonise. However, in countries with immature property rights and political systems, there is a risk that imitation of developed countries' technology and tax evasion involving bribery occur, and that the pollution haven hypothesis arises as a result of environmental policies. This study presents the mechanism of the pollution haven hypothesis by analysing the impact of environmental taxes, environmental tax evasion, and intellectual property rights (IPR) on economic growth, pollution, market share, corruption, and imitation in the North-South model with imitation. Economic growth and pollution reduction through environmental taxation in South imply the Porter hypothesis, but environmental tax evasion impedes this hypothesis. Environmental tax cuts in South with low IPR levels and tax evasion in South with high IPR levels lead to an increase in imitation rates, in addition to pollution havens and leakage from North to South, through labour migration between the manufacturing sector and the R&D sector. Depending on the IPR level in South, it is possible to avoid this pollution haven hypothesis. The deterioration of the quality of the political system by reducing fines for corrupt officials leads to the political pollution haven hypothesis, regardless of the level of IPR, but it can be avoided by improving the quality of the political system. This study suggests that the pollution haven hypothesis can be avoided by appropriate institutional design for patent rights and environmental regulations.

This study challenges mainstream theories of public spending by exploring the neglected link between financial development and social protection. We test two competing hypotheses: first, the shield hypothesis, which posits that financial development enhances state fiscal capacity and efficiency, thereby increasing social protection expenditures; and second, the substitution hypothesis, which suggests that financial development improves private risk management, reducing the need for public spending. Analyzing panel data from 61 countries (1996–2021) and using an Instrumental Variables Fixed Effects (IV/FE) framework to address endogeneity, the study finds that financial institutions and financial institutions depth provide robust support for the shield hypothesis. Furthermore, while the dependency ratio and unemployment increase social protection expenditures, exports reduce them. These findings validate the notion that financial institutions development bolsters, rather than diminishes, the state’s redistributive functions, offering pivotal implications for contemporary welfare state frameworks and development policy.

Over the past two decades, losses from natural hazard events have increased substantially, in part due to climate change. These losses have increasingly manifested as disruptions in homeowners insurance markets, with many high-risk areas experiencing reduced availability and affordability of coverage. While property-level risk characteristics are a central determinant of insurers’ portfolio management decisions, a growing literature highlights the role of policyholder credit scores in underwriting and premium setting, alongside the importance of regulatory environments in shaping insurers’ market participation. Yet even after accounting for risk exposure, credit risk, and regulation, substantial unexplained variation in insurance availability persists within narrowly defined risk regions. This paper proposes a framework in which insurers’ adjustment strategies—raising premiums versus non-renewing policies—depend on the elasticity of demand faced within a given risk pool. We argue that policyholder credit scores provide a parsimonious and insurer-relevant proxy for this elasticity, capturing households’ liquidity constraints, debt capacity, and sensitivity to premium increases. We develop a stylized model in which insurers optimally choose between price and quantity adjustments for high-risk policies based on underlying loss risk, regulatory constraints, and the distribution of demand elasticities proxied by credit scores. We then test the model’s predictions using ZIP-code–level data from the Federal Insurance Office (FIO), CoreLogic hazard metrics, and the Federal Reserve Bank of New York / Equifax Consumer Credit Panel. Using coarsened exact matching to compare ZIP codes with similar expected losses, claims experience, and loss ratios, we examine whether credit scores predict differential changes in policy non-renewals and premiums over time. The results indicate that, conditional on risk and regulation, insurers are more likely to rely on non-renewals in lower-credit-score areas, while adjusting primarily through premium increases in higher-credit-score areas. These findings suggest that disruptions in homeowners insurance markets reflect selection on demand elasticity rather than risk alone, with important implications for equity and regional resilience to climate-related hazards.

Wildfire risk has intensified across the western United States, generating substantial economic losses. From 2000 to 2017, large wildfires burned more than twice as much land annually as compared to the period 1984 to 1999. These changes in fire regimes pose new challenges for private forest managers, who must decide how to adjust management practices such as harvest timing, planting density, and fuel treatments in response to changing risk. While theoretical forestry models predict that wildfire risk can alter optimal management decisions, there is little empirical evidence directly documenting how landowners have responded to increasing wildfire risk. This study provides the first large-scale empirical evidence on how wildfire risk affects management decisions throughout the forest rotation on private land. We combine plot-level data from the U.S. Forest Service's Forest Inventory and Analysis (FIA) program with geospatial wildfire perimeters from the Monitoring Trends in Burn Severity (MTBS) dataset. Using confidential FIA plot coordinates, we assemble a panel of forest plots across Oregon, Washington, and California from 2001 to 2023, observing management decisions and stand characteristics over time. For each plot, we construct time-varying measures of wildfire exposure that capture the size, timing, and proximity of nearby fires. We then examine how planting density, the likelihood of intermediate fuel treatments, and the probability of harvest respond to nearby wildfire events. The results reveal consistent evidence of adaptive management responses to changing wildfire risk. Plots exposed to nearby wildfires exhibit longer rotation ages, lower planting densities, and a higher likelihood of intermediate fuel treatments. These responses align with predictions from natural resource economic theory, in which increasing fire risk induces landowners to adjust management to reduce expected losses. A simulation shows how these management responses have interacted to lower wood volumes and fuels in private western forests in the years since 2001. This study demonstrates that landowner adaptation to fire risk is likely to play an important role in shaping future forest conditions and regional fire regimes.

This paper examines how various urban agglomeration effects adjust natural disasters' short-run economic impacts. Using wind strength to proxy disaster intensity and electricity consumption to measure economic activity, I show that natural disasters reduce aggregate economic activity by 3 percent at the metropolitan statistical area (MSA). Lower job diversity amplifies losses, while centralized road networks and greater internal goods sourcing mitigate them. Leveraging granular industry specialization profiles, I find that increases in specialization in non-tradable local industries magnify the adverse economic effects of natural disasters.

Wildfires can threaten the provision of safe drinking water by altering source water characteristics or by directly damaging drinking water infrastructure. Wildfire-induced changes to source waters can disrupt normal drinking water treatment practices and may challenge systems with limited operational capacity or flexibility. Recent work shows that wildfires are associated with increases in the concentrations of certain drinking water contaminants as well as the frequency of Safe Drinking Water Act violations. Previous studies find the effects of wildfire on surface water sources can last as long as fifteen years post-wildfire in some locations. However, previous studies have focused on a small subset of contaminants and geographic areas over a shorter time horizon that excludes years of exceptional wildfire frequency after 2016. We study the impacts of wildfires on drinking water quality from 2000-2022 by linking a rich set of information on wildfire burn locations and severity to precise drinking water system intake locations and contaminant concentrations in all public drinking water systems in the Western U.S. Using geospatial techniques to characterize the extent of wildfire impacts to intake catchment areas, we demonstrate how the presence of an upstream wildfire affects drinking water quality delivered by downstream public water systems at different distances and time periods post-fire. Our analysis focuses on two types of outcomes related to drinking water quality: likelihood of contaminant detection and contaminant concentration. We include numerous classes of drinking water contaminants such as disinfectant byproducts, nutrients, volatile organic chemicals, and metals. We test for heterogeneous impacts across public water system size and source, watershed characteristics, and states. In all models, we include a set of controls for weather, temporal, and state variation. Our research builds on a growing literature aiming to better understand the impacts of climate change on water resources.

In US data, the unemployment rate is highly persistent over the business cycle, but standard calibrated search models predict little persistence after shocks. I show that incorporating duration dependence in job finding and separation rates significantly increases the persistence of the unemployment rate. To do so, I embed two-state duration dependence into an otherwise-standard Diamond-Mortensen-Pissarides search model. Intuitively, after recessions, the composition of the labor force shifts toward the long-term unemployed and the short-term employed, decreasing the aggregate job finding rate, increasing the aggregate job separation rate, and prolonging the effects of shocks. In my empirical calibration, duration dependence increases the persistence of the unemployment rate by 90\%, measured in half-lives. The model is flexible enough to match the tenure profiles of job-finding and separation rates for infinitely many combinations of pure duration dependence and permanent heterogeneity, and the extent to which duration dependence increases persistence depends on the share of observed duration dependence that is driven by pure duration dependence rather than permanent heterogeneity.

Can targeting firm debt relief improve stabilization policy in a financial crisis? I study the stabilization effects of firm-specific debt relief in a financial crisis. I construct a DSGE model of heterogeneous firms, calibrated to U.S. data, including the unconditional size distribution of firms. Firms face persistent idiosyncratic risk and financial frictions that give rise to an endogenous distribution over capital, debt, and productivity, while leading to capital misallocation and life-cycle effects. I model financial frictions by assuming collateralized borrowing. A shock to firms’ access to credit exacerbates misallocation, leading to a crisis similar to that of the Great Recession. Importantly, as the effects of the shock are not proportional across the distribution, the model is solved non-linearly and equilibrium decision rules are being resolved for all firms at each date. This disproportional misallocation creates a role for targeted policy: the government can allocate resources to financially constrained firms via debt relief, allowing firms to invest more. I show that policy targeting firms with the highest level of excess return to investment in a crisis reduces the drop in output by over 27 percent. To consider policy targets with historical precedent, I study firm size and age as alternative targets. The model is well-suited for this, as it produces an age-size distribution of firms matching U.S. data; this distribution is untargeted in the calibration. To my knowledge, this work is the first of its kind to be able to produce such size, and age-size distributions matching U.S. data. Though most firms constrained by financial frictions are small firms, I find policy targeting medium size firms outperforms other size targets, while policy targeting young firms outperforms other age targets. Medium size firms have a larger efficient size than small firms; they continue to grow more, creating a more persistent effect of debt relief policy.

This paper studies a modification to Social Security that distributes a universal lump-sum transfer to individuals early in life in exchange for an actuarially fair reduction in retirement benefits. The transfer is welfare-enhancing, particularly for credit-constrained individuals. Using survey data, we first document persistent disparities in credit access by income and race. Next we evaluate the transfer policy in a quantitative overlapping-generations life-cycle model with incomplete markets and race and education heterogeneity. We find that a lump-sum transfer of $40,000 at age 25 generates positive welfare gains for all race-education groups, equivalent to a 1.0–2.6 percent increase in lifetime consumption. Welfare gains are driven by improved consumption smoothing and reduced borrowing costs, and they decline sharply if the transfer is introduced later in life. We show that interactions with means-tested programs, differential mortality, and bequest motives matter for the magnitude, but not the sign, of welfare effects. Overall, a universal program that transfers a portion of an individual's Social Security wealth earlier in life improves welfare without increasing government outlays.

The uncovered interest parity (UIP) condition predicts that the higher-interest-rate currency should depreciate so as to equalize expected returns across currencies. Yet despite its central role in international macro models, this condition is often rejected in the empirical data. This paper argues that UIP coefficient estimates are inherently regime-dependent, a latent feature that becomes visible once one studies their term structure. Using historical Eurocurrency/LIBOR deposit rates at the 1-, 3-, 6-, and 12-month maturities, I document three empirical facts. First, rolling-window UIP coefficient estimates are negatively related to the volatility of interest rate differentials, and this monotonic relationship evolves gradually over time rather than shifting discretely around the 2008 Global Financial Crisis. Second, both the sign and the magnitude of UIP coefficients vary across maturities: under normal times, short-maturity currency trades earn larger premia in high-volatility states, whereas near the zero lower bound the premium shifts toward longer-maturity deposits. Third, these maturity patterns are difficult to reconcile with a pure ``delayed-overshooting'' explanation based solely on sluggish exchange rate responses to monetary shocks, but they are also hard to explain using a standard finance-based ``risk premium'' argument alone. Instead, the term-structure evidence from UIP regressions points to the coexistence of both mechanisms, with time-varying risk premia\textemdash shaped in part by the signs of yield curve slopes\textemdash playing a central role in the UIP puzzle.

Climate change is increasingly framed as a source of systemic financial risk. Abrupt repricing of carbon-intensive assets, extreme climate events, or sudden transition policies may trigger episodes of macro-financial stress often described as “green swans.” Existing debates on climate policy and inflation, however, largely emphasize production costs, relative price changes, or supply-side disruptions. This paper highlights a distinct mechanism. Climate-related financial crises operate as fiscal shocks: climate-Minsky risk reduces the expected present value of future primary surpluses, and the price level adjusts to restore the government’s intertemporal budget constraint. I develop a simple model in which the government finances expenditures by issuing short-term nominal debt and the central bank conducts monetary policy through interest-rate targeting. Fiscal policy is assumed to be non-Ricardian, so future primary surpluses do not adjust endogenously to stabilize the real value of outstanding debt. The model is featured with the dependence of fiscal outcomes on climate-related regime risk. In normal times, the government runs a baseline real primary surplus. During a climate-crisis regime, fiscal capacity deteriorates due to increased public expenditures and revenue losses associated with financial sector support, disaster relief, transition subsidies, and erosion of carbon-intensive tax bases. The economy switches probabilistically between normal and climate-crisis regimes following a Markov process, capturing the low-frequency and persistent nature of climate-related fiscal stress. Holding nominal debt predetermined, increases in the probability, severity, or persistence of climate crises raise the equilibrium price level on impact by weakening the expected fiscal backing of government debt. In this paper, I also clarify why monetary tightening alone may be insufficient to control inflation in the presence of climate-driven fiscal fragility. Higher interest rates raise debt-service costs and can further weaken fiscal backing, reinforcing price-level pressure. Climate-related financial instability thus transforms climate risk into fiscal risk, with inflation emerging as the adjustment mechanism under the fiscal theory of price level.

We study how U.S. aluminum tariffs propagate into beer prices through competitive pricing interactions. Independent craft brewers face spot aluminum costs and raise prices first, by approximately 1-1.5 percent, around the imposition of tariffs in March 2025. Macro-owned craft brands, which compete with independent craft but share parent firms and procurement with macro brands, increase next, before their parent firms' input cost increases have fully materialized. Macro brand prices adjust last. The staggered pattern within multi-brand firms is consistent with strategic complementarity: firms raise prices on brands whose rivals have already increased prices. Strategic pricing interactions amplify the pass-through of input tariffs beyond direct cost transmission.

How do governments manage debt dynamics in a fiscally dominant regime? We build a simple model that highlights the ’maturity-structure’ channel under fiscal dominance. Using a novel central government security level dataset between 1999 and 2022 for India, we identify periods of fiscal dominance using a Regime Switching model. We find that the probability of issuing a long-term security is approximately 22 percentage points higher under a fiscally dominant regime compared to other regimes. Our results suggest that maturity extension appears to be a distinctive feature of fiscally dominant policy regimes and contributes to the interest burden of debt.

We analyze the interaction between the Taiwan central bank’s profits and its policies. To earn large and consistent profits, the Taiwan central bank significantly expanded its balance sheet and relied on inexpensive short-term domestic funding to invest in longer-term foreign debt securities. In doing so, the central bank engineered a massive duration and currency mismatch on its balance sheet to capture term and currency risk premiums. We also argue that these large profits could not have been realized without a low-rate policy combined with heavy regulations on domestic financial institutions. In particular, the hedging requirement imposed on Taiwanese financial institutions reduces their post-hedging returns on foreign investments. Consequently, this regulation amounts to financial repression by restricting the overseas investment opportunities of Taiwan’s financial intermediaries and effectively trapping their assets within the domestic market. In other words, the profits earned by the central bank are, in effect, an implicit tax levied on domestic depositors. Besides, we consider two counterfactual scenarios in which the Taiwan central bank cannot conduct its current strategy. We find that when the central bank fully hedges its currency risks, these excess returns decline by 50%. Moreover, when the central bank is required to close its duration mismatch, its excess returns drop by 74%. These two exercises raise the question of whether the large excess returns could be sustained in the absence of financial regulations.

Forecasting time series is a common problem, in finance and other fields. Neural networks and other machine learning methods often produce models with high accuracy, but they provide little insight into the forecasting process. Here we provide a methodology for a human-in-the-loop forecasting process, where the forecasting regimes are automatically assigned. This will allow the forecaster to provide insight into the type of situation that is being forecast, as well as whether the accuracy is expected by be high or low. This methodology is applied in this paper to UK inflation forecasting. In this particular application the method shows promise for determining when the forecast is extrapolating and greater error is expected (such as during the Covid-19 pandemic).

Index numbers such as Paasche, Laspeyres, and Divisia indices are often used to form aggregates of goods with similar characteristics. When the number of goods are aggregated to form three or more composite goods which are used to evaluate consumer choices, the aggregates together may fail to pass the Generalized Axiom of Revealed Preference (GARP) which is necessary and sufficient for a set of consumer choices to be rationalized by a consistent utility function. We develop a new nonparametric test to evaluate if the aggregates together may be consistent with GARP. The new nonparametric test is evaluated using aggregated simulated data generated from a utility function that satisfies GARP. We add five, ten, and twenty percent measurement error to the price and quantity index numbers from the data that are consistent with GARP and test if they aggregates satisfy GARP when the data have measurement error. We find that the Divisia indices perform better than the Paasche and Laspeyres indices but all indices perform less well when the measurement error is high at twenty percent. Using consumer data for durable goods, nondurable goods and services, only the Divisia indices are consistent with rational consumer behavior.

The rise of the digital economy has the potential to reshape international financial markets. How stable are stablecoins? Is there interdependence among their returns and volatilities? Do return and volatility spillovers exist between the stablecoin market and other financial markets? To address these questions, we employ a multivariate GARCH-in-mean framework to examine the relationship between volatility and returns of leading stablecoins, as well as spillovers within the cryptocurrency market and between crypto and traditional financial markets. We find that although major stablecoins exhibit deviations from their $1 peg, and each stablecoin has its distinct volatility dynamics. Our results provide strong evidence of shock and volatility transmission among leading stablecoins, as well as significant volatility spillovers between the stablecoin market, Bitcoin, and the U.S. stock market. These findings highlight the central role of stablecoins in both cryptocurrency and broader financial markets.

This paper examines whether large language models (LLMs) can extract structured, forward-looking signals from institutional and corporate text that improve real-time detection of US business cycle turning points. A multi-source corpus spanning 1980 to the present is constructed from Federal Reserve statements, minutes, transcripts, and Beige Book reports; International Monetary Fund and European Central Bank communications; Bank for International Settlements Quarterly Reviews; and corporate earnings call transcripts. Each document is stamped with its public release date and enters the analysis only on or after that date, imposing a strict real-time information constraint. An LLM scores every document along several economically interpretable dimensions (recession risk, growth outlook, financial stress, labor market concern, forward guidance, and corporate distress) using a fixed zero-shot prompt with anchored rubrics to limit score drift across documents and decades. The resulting series are aggregated into monthly composite and source-specific signals and evaluated in two distinct roles: as predictors within probit and Markov-switching forecasting models, and as direct probabilistic recession forecasts elicited from the LLM itself. Performance is benchmarked against the Chauvet-Piger Markov-switching model, the term spread probit, the Sahm rule, and dictionary-based sentiment applied to the same text, with all models estimated recursively on an expanding window. Forecast accuracy is assessed at three-, six-, and twelve-month horizons across six recession episodes using quadratic probability scores, AUROC, calibration diagrams, and Diebold-Mariano tests, with bootstrap confidence intervals and episode-level case studies addressing the small number of recessions. The central threat of LLM knowledge contamination, scoring historical documents from post-hoc knowledge rather than contemporaneous textual content, is confronted directly through a systematic information-stripping test, a validation centered on the 2020 COVID recession whose onset left no prior institutional signal, and an explicit conservative-bound interpretation of measured gains. The paper contributes a scalable, reproducible framework for converting institutional text into real-time macroeconomic indicators and offers the first systematic comparison of LLM-derived signals against the leading business cycle dating benchmark. It further tests whether these signals augment, rather than merely replicate, the Markov-switching benchmark and whether their value differs across Federal Reserve, international, and corporate sources.

In this study sponsored by the Office of the Assistant Secretary of the Navy for Manpower & Reserve Affairs, we examined patterns in Selected Reserve (SELRES) recruiting, retention, and inventory over time and across different occupations. We used CNA’s comprehensive individual-level administrative data to construct metrics describing inventory health and to estimate the responsiveness of SELRES retention and affiliation to selected policy levers. Authorized endstrength has been relatively stable since 2016. SELRES inventory peaked in the 2019–2020 timeframe before dropping to a low in 2022 due to a combination of factors, including lower retention, affiliation, and non– prior service recruiting after COVID-19. Many of these patterns have since reversed, and SELRES inventory has returned to levels more in balance with billets authorized. Although overall inventory levels have recovered from the 2022 lows, we found that manning health varied across occupational communities. To examine manning health, we created a metric of overall community health using measures of inventory, fill, and fit across different communities. We found that some large enlisted communities appeared undermanned relative to requirements while others had a substantial mismatch between billet requirements and the skills or occupational specialties of those filling the billets. Finally, we used a difference-in-differences approach to estimate how policy levers, such as bonuses and advancement opportunity, affect retention and affiliation. Our analysis on reenlistment bonuses found that increases in these bonuses in recent years improved reenlistment rates for early and mid-career reservists but not for late-career reservists. We also found that affiliation bonuses were associated positively with affiliation probabilities for sailors who left the active component within their first 6 years of service. Finally, we found a positive relationship between advancement opportunity and retention rates. Our paper produced several policy recommendations, including to reallocate reenlistment bonus resources from senior to junior personnel, condition affiliation and prior service bonuses on zone, reallocate reenlistment bonuses from well-manned communities to undermanned communities, and to increase advancement opportunities for communities we identified as having low inventory health and advancement opportunity.

The United States federal government awards billions of dollars in contracts annually to local firms for them to provide goods and services. In addition to acquiring the needed goods and services for the government, these contracts help support local economies and keep people employed. A reduction in government spending reduces employment and wages in these economies, and may also have other impacts. The reduction in spending may also affect the military’s ability to recruit and retain members. Fewer jobs which were supported by federal dollars make joining the military more attractive to individuals most impacted by spending reductions. This same dynamic may make it more attractive to remain in the military instead of separating when given the chance. Using administrative Air Force personnel data, we leverage the institutional details of the Budget Control Act of 2011 and transaction-level government spending data to estimate the impact of local reductions in spending on military recruitment and retention.

forthcoming

Marine Corps Recruiting Command (MCRC) depends on the success of each of its recruiting stations (RS) to meet its contract mission. Monitoring RS performance is critical to predict shortfalls and adjust accordingly. Our research questions asks, "How can MCRC forecast the likelihood of RS success?" To address this question, CNA to develop a tool to forecast the likelihood of RS success, as defined as meeting mission with at least 63% Alpha and 95% Tier 1 contracts. A new tool is especially important because the recruiting environment has changed. Specifically, interview yields from contacts and contract yields from new working applicants have fallen due to lower propensity rates and the introduction of Military Health System Genesis. We use logistic least absolute shrinkage and selection operator (LASSO) regressions and extreme gradient boosting (XGBoost) to predict RS success based on previous RS performance, RS personnel, economic conditions, and characteristics of the local population. To evaluate performance, we train and test our model on separate samples. Our training sample spans from 2020 through 2024, while our testing sample uses 2025 outcomes. We find the LASSO model outperforms XGBoos and past heuristics. The predictions remain useful for up to four months in advance. To facilitate the use of the tool, we produced a dashboard that visualizes our tool to help MCRC leadership quickly identify recruiting stations at the greatest risk for failure.

Over the past decade, global macroeconomic performance has been increasingly shaped by two major sources of uncertainty: geopolitical risk and global supply chain pressures. As countries become more integrated through global production networks, the combination of heightened geopolitical tensions and growing interdependence has amplified concerns about the macroeconomic consequences of these shocks. Using panel local projections with sign-restricted identification for 35 countries from 1998Q1 to 2024Q1, we find that while both shocks are contractionary, supply chain disruptions are roughly twice as damaging to GDP and three times as inflationary as geopolitical shocks. Geopolitical shocks primarily induce demand-driven slowdowns, whereas supply chain shocks trigger persistent supply-side disruptions. The effects vary across income groups: advanced economies display stronger and more sustained inflationary responses, while emerging markets suffer deeper and longer-lasting output losses. Advanced economies currencies appreciate at medium horizons whereas emerging markets exchange rates depreciate persistently. These differences reflect structural factors such as trade exposure, policy credibility, and financial constraints. Our findings highlight the need for tailored stabilization strategies that account for both the source of uncertainty and country-specific vulnerabilities.

This paper examines the relationship between economic narratives surround-ing geopolitical events and their actual economic impacts. By employing a large-language model on a large set of newspaper articles, we identify whether the narrative of a geopolitical risk (GPR) shock is seen as acting on the supply or on the demand side. For the classification of the narrative, we equip the large-language model with a questionnaire to analyze the event’s economic characterization. A Vector Autoregression model allows us to validate our GPR narrative indices by assessing whether the narratives align with economic realities. By identifying the nature of the geopolitical risk shocks in real time, central banks can more easily gauge the inherent risk to inflation and thus make better informed decisions.

This study uses bank-level data from December 1999 to January 2026 to examine how bank lending is affected by national elections in Taiwan. Our results yield three primary insights. First, we find little evidence of a broad expansion in total lending prior to elections, although state-owned banks increase total lending in the immediate pre-election quarter. Moreover, lending behavior varies across loan categories, with firm lending increasing earlier in the pre-election period and household-related lending increasing primarily in the 3 months immediately preceding elections. Second, we obtain no robust evidence of partisan differences, suggesting that electoral incentives may be more influential than are partisan preferences. Third, we obtain some evidence that the growth rate of preferential housing loans, primarily provided by state-owned banks, increases prior to elections, particularly during the 6 months preceding elections. The share of preferential housing loans also increases most clearly in the immediate pre-election quarter. We also find some evidence of a decrease in loan quality following elections and inaugurations after the introduction of the Preferential Housing Loans for the Youth program. These results suggest that politically influenced lending operates through targeted credit allocation in household-related and policy-oriented housing lending.

This study examines exchange rate pass-through (ERPT) to U.S. import prices over two decades, analyzing data by country-of-origin and industry. We find significant heterogeneity in ERPT, with higher pass-through for commodities priced in dollars and key U.S. trading partners like Canada and Mexico. While short-term exchange rate fluctuations have limited effects, foreign production and transportation costs play a more substantial role. Over time, both costs and exchange rates exert greater influence on import prices. The dominance of low pass-through industries in U.S. trade explains the muted aggregate effects reported in prior studies. However, advanced panel estimation techniques suggest traditional time series analyses miss key features of ERPT. Our findings underscore the need to account for sectoral variations in import price models to more accurately capture the impact of foreign economic shocks on U.S. import prices.

This paper studies whether commuting constraints can generate gender wage gaps through spatial sorting, even in the absence of discrimination or individual-level wage--commuting compensation. A worker workplace choice model is developed in which individuals choose workplaces based on commuting costs, allowing commuting sensitivity to differ by gender. When women are more sensitive to commuting costs in settings where higher-paying jobs require longer commutes, differences in access to high-wage workplaces generate gender wage gaps. Empirically, the paper estimates gender-specific gravity equations using Korean Census commuting flows and regional wage data. Because the structural model implies a conditional mean representation of commuting flows in terms of workplace choice probabilities, the gravity equations are estimated using Poisson pseudo–maximum likelihood. The estimates imply a baseline gender wage gap of 1.88 in three-month average earnings, driven by gender differences in access to high-wage workplaces due to commuting constraints, and counterfactual exercises show that equalizing commuting sensitivity across genders would not only eliminate this gap but reverse its sign.

Work-from-home (WFH) arrangements are increasingly common in post-pandemic labor markets. This paper investigates whether WFH reduces gender gaps in labor supply, focusing on periods of heightened seasonal caregiving demands, such as summer school breaks. Using data from the Current Population Survey (CPS), I compare women in WFH-possible jobs who work remotely to those who work on-site in the same types of jobs. Employing a Difference-in-Differences (DiD) approach, I estimate how remote work affects labor force participation, earnings, working hours, and absence from work. The results show that WFH substantially mitigates the seasonal decline in women’s labor supply by providing greater flexibility in work location and timing. These findings suggest that WFH can be an effective tool for reducing gender gaps in labor force participation, helping women maintain employment and working hours during periods of increased household responsibilities.

Women comprise the majority of minimum wage workers, making gender the third most important predictor of being affected by a minimum wage policy. Women are also more likely to prefer flexibility in their working hours, exhibiting more variation along their extensive (employment) and intensive (hours) margins. But, the minimum wage literature has had very little to say about gender or women and has heavily favored estimates along the extensive rather than the intensive margin. In this study, gendered effects of minimum wages are explored to see what role increasing the minimum wage may have in closing (or maybe expanding) any gender gaps in labor market outcomes. Seven large minimum increases of one dollar or more are used as the treatments, happening across three different Canadian provinces, and occurring at different times in the low inflation period of the late 2010s to early 2020s. The policy effects were mostly positive for wages and negative for hours while more mixed for employment. Most often, whomever had the larger wage increase, also had the larger hour decrease, with larger effects for women than for men. Gender gaps are also shown to be closing in wages and in employment, but may be expanding in hours.

Modern information and communication technologies have opened up new possibilities in the way people live and work. This study examines the impact of a new mandate in Colombia that increased residential download speed, exploring its effect on the likelihood of employment for women of prime working age. We use seven years of data from Colombia’s National Quality of Life survey (ENCV), a nationally representative data for Colombia spanning from 2012 to 2019 and employ event study analysis. To complement these findings, we use four treatment effect estimation methods assuming unconfoundedness. Our results indicate that women in households with internet connection and in regions that experienced the highest increase in download speed are more likely to work and less likely to use most of their time doing housework. Our results from event studies show that after the mandate, women who experienced an increase in internet quality are more likely to do market work by four percentage points. This effect is sizeable given women's low working participation rate in our data (52 percent). Pre-mandate estimates confirm that the increase in working women is not explained by a difference in trends between women experiencing increased internet quality and their counterparts. Moreover, our results on working status are robust to restricting the violations of parallel trends to be no larger than the maximal pre-treatment violation. Also, a placebo exercise comparing women in regions that do not experience a significant increase in internet speed shows no evidence of an effect. Last, we provide several results consistent with the hypothesis that programs targeting free women's time may increase their probability of working. This timely study aligns with the UN’s 2030 Agenda and the emphasis on information and communication technologies (ICTs) as enablers of the Sustainable Development Goals (SDGs). Understanding the effects of internet access on employment is crucial globally, as governments strive to provide reliable, affordable internet for all. Our findings, drawn from a nationally representative sample, offer insights relevant to countries with similar challenges in gender disparities in labor participation and internet connectivity.

Despite substantial advances in cancer detection and treatment, a significant share of patients in the United States refuse recommended cancer-directed therapy, potentially contributing to preventable mortality and persistent disparities in cancer outcomes. This study examines the socioeconomic and demographic determinants of cancer treatment refusal using a large population-based dataset. We use the Incidence – SEER Research Plus Limited-Field Data, 21 Registries (November 2024 Submission), covering cancer diagnoses from 2000 to 2022 and representing approximately 47.9% of the U.S. population. Treatment refusal is identified using SEER variables explicitly documenting patient refusal of recommended cancer-directed surgery and radiation therapy, distinguishing refusal from cases where treatment was not recommended or contraindicated. We evaluate refusal patterns across cancer type, stage at diagnosis, and treatment modality, and assess associations with patient characteristics including age, sex, race/ethnicity, and marital status, along with county-level socioeconomic indicators such as income, insurance status, and rurality. We estimate multivariable logistic regression models to identify predictors of refusal and to quantify adjusted odds across patient subgroups. Preliminary analyses suggest that treatment refusal is not randomly distributed but is concentrated among specific demographic and socioeconomic groups. The findings will provide evidence on which populations are most vulnerable to refusing cancer treatment and highlight potential policy and clinical interventions such as targeted patient navigation, improved shared decision-making, and community-based outreach to reduce refusal rates and improve equity in cancer care.

Emerging evidence suggests that take-up of the social safety net amongst system-involved individuals (those who are directly or indirectly impacted by criminal justice system) is limited, and that barriers to take-up spill over to system impacted populations more broadly. This study asks: (a) to what extent do administrative burdens fall disproportionately on system impacted individuals, and (b) which program design features most strongly shape their willingness and ability to enroll? Using pilot data from a survey of low income adults (n = 866), descriptive analyses assess differences in SNAP application rates and experiences or expectations of administrative burden between system impacted and non impacted respondents. A conjoint experiment evaluated which program design features causally increase application interest, focusing on barriers (work requirements, drug testing, eligibility for individuals with drug related felony convictions), features that may reduce learning and compliance costs (virtual appointments, text message reminders), and features aimed at reducing psychological costs (caseworkers with lived experience of public benefits or system impact). Contrary to common rhetoric, system‑impacted individuals reported higher SNAP application rates than their non-system‑impacted counterparts, controlling for myriad covariates. For the experience or expectation of administrative burdens, only on the dimension of understanding how to apply or recertify were system-impacted individuals’ burdens disparately greater. However, categorization of a participant’s state of residence’s eligibility policy for individuals with drug‑related felony histories was a significant factor on several dimensions of burden. Living in a state where individuals are eligible for SNAP with previous drug-related felonies without barriers or modifications was associated with significantly lower experienced of perceived difficulty of understanding eligibility and how to apply or recertify. The conjoint experiment revealed relatively weak average preferences across program attributes, with one exception, namely that respondents preferred programs with work requirements. Utilizing Bayesian Autoregression Trees (BART) to assess heterogeneous effects, counter to common rhetoric, system impact did not meaningfully shape preferences for work requirements or drug testing. System-impact status was a notably important source of heterogeneity in preference for caseworkers who were system-impacted, specifically preferring those without this lived experienced, counter to expectations.

Background: The US has state variation in edentulism, commonly known as total tooth loss, in older adults. This paper identifies key social determinants of health (SDOH) associated with state differences in older adult edentulism that can be addressed with partnerships between dental providers and policy makers. Methods: Data was collected from three publicly available data sets including, County Health Rankings, provided by the University of Wisconsin Population Health Institute, the U.S. Centers for Disease Control and Prevention (CDC), and the Centers for Medicare and Medicaid Services (CMS). Results come from multiple regression analysis of panel data for all states and the District of Columbia for even years 2012-2020. Results: A variety of SDOH, health behaviors, and demographics are associated with older adult edentulism with the largest effects for education and smoking. We have other variables of interest and overall, our model explains over 80 percent of the variation. Conclusion: Policy making in states should emphasize both education and smoking cessation programs for individuals. Future research may also consider the relationship between state regulated dental insurance and dental practice location. States may also consider policies to encourage better dental professional training and incentives to address smoking cessation for patients.

The Affordable Care Act (ACA), passed in 2010, increased access to healthcare as more Americans obtained health insurance through this health reform. With this increased access, the healthcare industry has been put under increased pressure to address the gap between supply and demand for healthcare staff. To meet the new demand, there has been a notable increase in employment of medical practitioners with advanced training, namely Physician Assistants (PAs) and Nurse Practitioners (NPs) (cite). The extant literature finds that PAs have the lowest malpractice rates. However, this literature does not condition upon scope of care. When we control for state regulatory structure, we find that PAs who provide care as semi-independent or independent workers have a malpractice adverse action rate that is statistically indifferent from that of MDs. These results provide evidence consistent with the conclusion that lower malpractice rates among PAs exist because of a regulatory shield in most states that necessitates a liability-shifting supervising PA.

This study examines how economic growth responds to improvements in business conditions over time. Using dynamic panel models and panel vector autoregressions (pVARs) for 82 countries over the period 2003–2022, we investigate the timing, persistence, and transmission of growth responses to changes in the business environment, as measured by the Economist Intelligence Unit’s (EIU) Business Environment Ranking (BER). The estimated orthogonalized impulse response functions suggest that economic growth temporarily accelerates following an exogenous shock in the business environment, with positive effects gradually unfolding for up to three years. Extending the pVAR model to include investment and trade openness reveals that changes in the business environment influence growth directly and indirectly through gross fixed capital formation and external sector channels, with investment emerging as the more persistent transmission mechanism. Overall, improvements in the business environment are associated with transitional growth gains, and the timing and magnitude of the effects differ systematically across countries. Advanced economies experience larger and more persistent growth gains than emerging markets and developing economies. Countries with better initial business environments, greater financial openness, and higher levels of human capital also exhibit stronger cumulative responses, with the peak effects attained at longer horizons.

This paper documents the contribution of human capital to the acceleration of economic convergence over the last decades. We show that absolute convergence in human capital began around 1970, preceding recent evidence of absolute GDP per capita convergence since the 2000s. We estimate that the β-convergence coefficient in GDP per capita, conditional on the initial level of human capital, is twice as high as the unconditional β-coefficient, highlighting a crucial role of human capital in closing the gap between poor and rich economies. This significant contribution of human capital to income convergence is robust to alternative measures of income growth, broader definitions of human capital (including those accounting for education quality), fixed effects, and the potential role of institutions in the convergence process.

Identifying how a specific pair of macroeconomic outcomes interrelate over time can make a valuable contribution to devising effective macroeconomic policy packages for development. I study the relationship between export performance and the real exchange rate, undertaking a systematic test of four theories of the co-evolution of these two important indicators. The object of this study is Thailand over the period 1970- 2020, a uniquely applicable environment for examining the theories at hand. I conduct a comparative test of the four theories through a series of structural vector autoregression (SVAR) estimations, supporting these feature exercises with preliminary empirics which utilize some innovative techniques and a local projection which delivers a set of reference results. I find that Thailand’s export growth over the period has been concentrated in export booms dominated by technical manufacturing and coinciding with rapid real exchange rate appreciations. Comparative SVAR testing yields evidence that Thailand has experienced dynamics reflecting endogenous investment booms and more fragile support for theories of external financing and commodity cycles. I find evidence of competitiveness dynamics in single-equation models but these are not identified within SVAR estimations.

The Somali community of Eastleigh, Nairobi has become a key part of the wider Nairobi and Kenyan economy. Observers have stated that up to 40% of Nairobi’s GDP is driven from its Somali diaspora community. As a team, we wanted to look at some of the drivers of the success of businesspeople in Eastleigh, the business hub of Nairobi’s Somali community. Our team of Somali and Swahili speaking researchers conducted 253 interviews with a selection of business owners in the Eastleigh area. Setting up a multinomial logit regression based on five different metrics of success, we saw that there was a highly significant correlation between being Somalia born, rather than being a Kenyan Somali internal immigrant from Northeastern Kenya. There is also a significant correlation between being a successful Somali born businessman in Eastleigh and having close family connections in Western nations, tentatively giving evidence of the acknowledged finance role of diaspora relations among the Somali community. These results have interesting applications to questions of diaspora analysis, cross-border financial activity, and community cohesiveness in the face of state collapse.

We examine changes in mental healthcare utilization caused by the 2022 Supreme Court case, Dobbs v Jackson Women’s Health Organization, and the subsequent differences in state-level abortion laws. Prior evidence documents declines in self-reported mental health in states that became more restrictive, but relies on survey measures that cannot distinguish changes in actual care-seeking. We study whether Dobbs caused measurable shifts in mental healthcare utilization using administrative claims from 200,000 continuously enrolled, privately insured adults drawn from the 2020--2023 Merative MarketScan Commercial Claims and Encounters database. Employing a difference-in-differences design on a balanced daily panel, we compare changes in outpatient psychotherapy and psychotropic medication utilization following the Dobbs leak (May 2, 2022) among adults residing in states that enacted more restrictive abortion laws relative to those in states that enacted explicit protections. Event-study estimates confirm parallel pre-period trends and indicate that effects are concentrated in the two months following the leak, consistent with an acute stress response. Our findings reveal a difference in effect across the two arms of mental healthcare. On the medication side, individuals seemed to increase utilization in response to their self-reported stress. Those in restrictive states exhibited a significant increase in antidepressant initiation relative to those in protective states (xtreg: +0.0031, p < 0.10 ; IRR: 1.06, p<0.05), representing a 4.4% increase over the pre-period mean. They also saw a significant increase in the probability of a dose increase (xtreg: +0.0050, p < 0.01 ; IRR: 1.09, p<0.05). On the psychotherapy side, we find the opposite pattern. Individuals in restrictive states were significantly less likely to begin or re-enter a therapy episode (xtreg: -0.0210, p < 0.01 ; IRR: 0.92, p<0.01) and significantly more likely to disengage from ongoing therapy (xtreg: +0.0100, p< 0.05). Together, these results are consistent with a treatment modality substitution in which policy-induced stress increased pharmaceutical care-seeking while simultaneously suppressing psychotherapeutic engagement. This may reflect differential access barriers, cost, or privacy concerns associated with ongoing therapy in restrictive states. These findings suggest that the full mental health burden of abortion restrictions may be understated by utilization counts alone.

Research / Policy Question: This paper asks whether large-scale public health interventions that reduce infectious disease exposure in early life generate improvements in longevity. Specifically, it examines the long-run effects of early-life and childhood exposure to the U.S. malaria eradication campaign of the 1920s on later-life mortality. The policy relevance stems from ongoing debates about the costs and long-term benefits of malaria eradication and new vaccine deployment in malaria-endemic countries today. Contribution: The paper provides the first population-level evidence on how early-life reductions in malaria risk affected survival into older ages in the United States. Although prior research shows that malaria eradication improved midlife socioeconomic outcomes, no study has traced its consequences for old-age mortality. Methods and Models: The empirical strategy exploits cross-county variation in baseline malaria risk, measured by the ecological Malaria Risk Index, and cross-cohort differences in age at exposure to the 1920 eradication campaign. We also estimate state-level impacts on cumulative mortality rates using a difference-in-differences framework with birth-state and birth-cohort fixed effects. Data, Time Frame, and Geographic Scope: The main analysis uses individuals born in rural counties between 1900–1940 in all U.S. regions except the Northeast, linked to deaths occurring 1975–2005 as observed in the Death Master Files (DMF) of the Social Security Administration data extracted from the Censoc Project. Results: We find that in-utero and early-life exposure to malaria eradication increases longevity by about one month per one–standard deviation increase in the birth-county malaria risk index. The effects exhibit dynamic complementarities, with larger gains among individuals who were also exposed to the hookworm eradication campaign a decade earlier and among children in states with stricter child labor policies. Analysis of census data indicates that improvements in education, reductions in disability, higher income, and greater wealth are plausible channels through which early-life exposure translated into longer lives. Evidence from World War II enlistment records further shows nontrivial increases in cognitive ability and height among affected individuals.

Crime imposes substantial social and economic costs in the United States, yet little evidence documents how early-life health interventions shape later criminal involvement. This gap is likely due to scarcity of data linking criminals to their early childhood experiences. Building on evidence that early-life medical care improves socially desirable adult outcomes, this paper examines how early health conditions and medical interventions impact later criminal behavior to uncover cost-effective strategies for improving public safety. We study this question using newly linked administrative data combining Texas birth registries with statewide criminal histories. Our empirical strategy leveraging the very low birth weight (VLBW) cutoff at 1,500 grams that determines the eligibility for free intensive neonatal care. The identifying assumption is that infants born at 1,490 grams are essentially identical to those born at 1,510 grams, except for the additional medical attention lower-weight infants receive. Our regression discontinuity design reveals that medical treatments received by VLBW infants reduce their likelihood of committing crimes in adulthood. We find similar declines across Black and White individuals; however, because baseline conviction rates are higher for Black individuals, VLBW treatment reduces the racial gap in convictions by 83%. These findings suggest that differences in criminal behavior can be partially explained by malleable health differences present at birth. Our natural experiment setting offers several advantages for exploring these questions. First, low birth weight infants face elevated risks for physical, cognitive, and behavioral problems, many associated with later criminal activity. Second, by focusing on infants with nearly identical birth weights around the threshold, our design minimizes confounding from family background and socioeconomic factors. Previous research using this strategy has shown that VLBW interventions improve intermediate health outcomes and academic achievement, both of which may affect incentives to engage in criminal activity. Third, our analyses use administrative data from Texas, the second most populous state and the state with the highest number of incarcerated residents. Texas’s large, diverse population (13.6% Black, 39% Hispanic) allows us to investigate heterogeneity in early health intervention impacts across racial and ethnic groups with precision.

The paper explores the dynamics of inter-fuel substitution, focusing on the impact of biofuels in the aviation sector and their potential in decarbonizing aviation in California. Specifically, the research investigates two key objectives: (1) understanding the historical interplay of substitution or complementary relationships among general transportation fuels after the mid-2010’s introduction of biofuels, and (2) evaluating the extent to which biofuel adoption drives inter-fuel substitution in particularly aviation sector. Ordinary Least Squares (OLS) and Seemingly Unrelated Regression (SUR) methods are used to analyze the data from U.S. Energy Information Administration (EIA) and U.S. Environmental Protection Agency (EPA) for 1980-2022 to provide a comprehensive view of biofuel integration. Results reveal that while biofuels have not significantly changed traditional substitution and complementarity relationships in the general transportation sector, they serve as a viable alternative to kerosene-based jet fuel in aviation. A 1% increase in jet fuel price leads to 1.043% increase of SAF volumes consumed, demonstrating that SAF has become a valid substitute for a traditional kerosine-based jet fuel. The effect of price changes of SAF on jet fuel volumes still remains limited because of small shares of biofuels in aviation. The paper demonstrates biofuels' growing potential to reduce dependency on fossil fuels and their important contribution to achieving long-term sustainability in aviation.

I examine how targeted federal grants affect state and local spending on public transit. The analysis uses comprehensive U.S. expenditure data from 2000--2019 and exploits an exogenous shock from the 2009 American Recovery and Reinvestment Act (ARRA). ARRA funds were apportioned to Urbanized Areas through preexisting formula programs, independent of potential changes in transit investment. Using ARRA apportionments as an instrument, I find that each \$1 of exogenous federal grants generates a \$0.20 annual increase in capital transit spending from all sources. This average effect reflects two distinct phases between 2009--2019: an initial rise in federally funded expenditures with no displacement of state or local spending (the flypaper effect), followed by substantial crowding-in of state funding to the same localities. I develop a conceptual model of local officials’ investment decisions that distinguishes between guaranteed and flexible funding from each source, the latter requiring costly negotiation. In this framework, the initial flypaper effect arises from the stickiness of guaranteed funds, while the subsequent state crowding-in results from increased negotiation for flexible funding. This negotiation can be driven by two mechanisms: (i) higher returns to additional investment or (ii) lower costs of negotiation. Empirical evidence on the nature of additional spending rejects the first mechanism, while cross-state variation in institutional settings supports the second. Taken together, these results suggest that federal grants empowered local officials to secure additional flexible state funding by reducing the cost of negotiation, leading to a disproportionate and persistent increase in total public transit spending.

The US Department of Transportation (DOT) adopted a new policy of mandatory service guarantees in 2024 – customers are entitled to a refund of the ticket price if their flight is cancelled, significantly delayed or significantly changes. Using DOT on-time performance data we examine how carriers responded to this new DOT policy. By separating based on carrier types, we find some modest changes in flight delays based on carrier types with both low-cost carriers and ultra low-cost carriers exhibiting noticeable service quality improvements.

This paper documents a two-decade decline in U.S. consumers’ use of cash and the consequences for policy makers responsible for currency management (bills and coins) using nationally representative data from the Federal Reserve Bank of Atlanta’s Survey and Diary of Consumer Payment Choice (S/DCPC). The S/DCPC data provide a comprehensive picture of U.S. consumer payment preferences, adoption, and use of various payment instruments from an annual recall-based survey (SCPC, 2008-2024) and a daily payment diary kept by consumers over a three-day periods during October 1-31 each year (DCPC, 2012-2024). The S/DCPC data show a long-run decline in consumer cash use from about one-third of payments in 2008 to about 14 percent in 2024. Like consumers in other countries, U.S. consumers’ use of cash before and during COVID-19 dropped significantly but has remained relatively stable since. Cash use declined partly because payments increased over time in places where cash is not accepted (e.g., online) but also decreased even where cash still is accepted widely (e.g., in-person payments), suggesting a secular change in consumer preferences. Indeed, the S/DCPC data also reveal similar declines in U.S. consumer’s self-reported unconditional preferences for cash payments. Most payments made with cash have remained relatively low value over time. We extend the literature by estimating conventional multinomial Probit/Logit models of payment choice that control for individual consumer characteristics, and payment value and location, economic conditions, and changes in survey methodology over an unprecedented length of time. Although the data and results suggest cash is not “dead”, or even “dying” any time soon, the trend is unequivocally negative. Banks, policy makers, and cash industry participants who are considering investments and retooling in the U.S. cash infrastructure need reliable forecasts of future cash use. We investigate the relevant factors, develop long-run forecasts, and discuss the uncertainty.

The GENIUS Act establishes a regulatory framework for stablecoins and could expand pathways for their use for payments. Little research exists on the use of cryptocurrency as a payment method. In this paper, we examine how using cryptocurrency for payment has evolved in recent years, identify which consumers’ characteristics are associated with using cryptocurrency for payment, and assess how these characteristics have changed. We use data from the 2022 and 2025 Survey of Household Economics and Decisionmaking (SHED), in which the share of U.S. consumers using cryptocurrency for payment has been very small and declined slightly from 2.7% in 2022 to 2.4% in 2025. From a logit model using a pooled sample across both years, we find a statistically significant negative coefficient for the 2025-year dummy. From logit models using two separate 2022 and 2025 samples, we find seven characteristics are statistically significantly (at 5% level) positively associated with the probability of using cryptocurrency for payment for both years, and while three characteristics (age 18-34, self-employed, and using alternative financial services (AFS)) have a larger average marginal effect for 2025 than for 2022, four characteristics (age 35-54, male, credit score being “fair/good”, risk tolerance being the highest 9-10) have a smaller average marginal effect for 2025 than for 2022. Additional five characteristics (Black, Hispanic, other minority race, credit score being “very poor/poor,” and “don’t know credit score”) are statistically significantly positively associated for 2022 but not or only marginally associated for 2025. Our results indicate that the decline in the use of cryptocurrency for payment spans many consumer groups. Our results also highlight that the positive association between the use of cryptocurrency for payment and using AFS may be due to these consumers’ insufficient access to more traditional digital payment methods, warranting continued monitoring and further research.

I explore consumers’ sensitivity to ATM surcharges using a uniquely detailed, high-frequency panel data set. Taking advantage of daily variation in withdrawals across thousands of ATMs, I am able to directly estimate the price elasticity of demand for ATM cash withdrawals, accounting for potential endogeneity using rich fixed effects as well as structural instrumental variable strategy. The results indicate consumers are considerably less sensitive to ATM surcharges than suggested by previous studies that used aggregate data or relied on strong structural assumptions. I also find substantial heterogeneity in cash demand and surcharge sensitivity across different demographic groups, location types, as well as competitive environments. The results suggest clear ways in which ATM operators could increase profits by adjusting their prices.

The social cost of carbon dioxide (SC-CO2) quantifies the monetized damages from a marginal unit of CO2 emissions and is a central input to climate policy. Existing global estimates account for damages including temperature-related mortality, agriculture, energy, and sea-level rise, but omit damages from fire PM2.5 mortality---a pathway through which warming amplifies wildfire activity and associated smoke exposure. We estimate a global SC-CO2 incorporating fire PM2.5 mortality as a fifth damage sector in the GIVE integrated assessment model. To parameterize the damage function, we pool fire PM2.5 concentration data from multiple global fire model simulations and estimate country-level relationships between fire PM2.5 exposure and global mean temperature (GMT) change using a fixed-effects regression. We find that incorporating fire PM2.5 mortality increases the global SC-CO2 by 23%, from $208 to $255 per ton CO2 at a 2% discount rate. FUND regions contributing the largest fire PM2.5 damage components are the Former Soviet Union, United States, and Western Europe, reflecting large estimated increases in fire PM2.5 exposure per degree of warming and large exposed populations.

How should 2035 Nationally Determined Contributions (NDCs) be judged when countries reach the same income levels under different technological and structural conditions? This paper proposes a stochastic, effort-sharing benchmark for evaluating NDC ambition relative to historically observed low-carbon development paths. We classify 102 economies into seven development paths and estimate path-specific emissions–income elasticity surfaces using a functional-coefficient panel model with greenhouse-gas and income data spanning 1900–2022. We define a Latecomer Advantage Index as the cumulative elasticity gap between followers and earlier developers over common income ranges, and use it to construct two counterfactual benchmark paths: a same-path predecessor path, which serves as an equity reference, and a deep-decoupling frontier path, which serves as an aspirational reference. Forecast uncertainty to 2035 is incorporated through functional principal components, residual bootstrap, and Monte Carlo income paths. We find that a latecomer advantage is widespread at the entry stage of industrialization: followers generally increase emissions less than earlier developers at the same income level. However, this advantage is not permanent. It often remains when followers are compared with same-path predecessors, but turns into a follower gap at higher income levels when the benchmark is the deep-decoupling frontier. The per-capita advantage also does not automatically translate into lower total emissions, especially where population grows rapidly. Applied to available 2035 NDCs, the framework shows that the same pledge can appear ambitious relative to comparable predecessors while remaining insufficient relative to best-practice or temperature-based benchmarks—China's roughly flat pledge, for instance, falls below the aspirational frontier yet is rated "highly insufficient" on temperature grounds. The paper therefore complements 1.5°C-consistency assessments by diagnosing the relative ambition of NDCs against historically realized development paths.

Economists’ disagreements about the economics of climate are due, in part, to disagreements about the cost of decarbonization. Integrated Assessment Models (IAMs) assume that abating greenhouse gas (GHG) emissions reduces the output available for consumption or non-climate investments, while Keynesian models assume that climate investments can increase output through an investment multiplier. Abatement costs are usually estimated from engineering studies, but in principle those costs should be observable in macroeconomic data. This paper seeks to estimate abatement costs using a panel data of 177 countries from 1990 and 2022. Countries have achieved a wide range of GHG emissions control rates since 1990. I estimate the statistical relationship between GDP per capita and the GHG intensity of output across several specifications. My preferred two-way fixed-effects results find that a 1% reduction in GHG intensity of output is associated with a 0.49% increase in output. International data provides mixed evidence for rising abatement costs. These results are correlational rather than causal, but the paper provides some recommendations for future research using international data to inform climate economics.

Motivation: Fire-prone states must substantially increase prescribed burning to reverse historic fire suppression, restore ecology, and ultimately lessen the frequency and intensity of wildfires. Despite the advantages of prescribed fires, this tool has been underutilized. Air quality regulations create a perverse incentive to reduce the fuel backlog via wildfires rather than via prescribed fire because smoke from prescribed fires could lead to violations of the National Ambient Air Quality Standards (NAAQS). Research Questions: This project quantifies the degree to which clean air rules lessened hazardous fuel reduction in the United States and the implications of these practices for air quality. First, we ask: have federal air quality regulations reduced the prevalence of prescribed burning in the US? Second, to the extent regulations did lessen fuel treatment, we measure the subsequent implications of air quality regulation for wildfires and smoke exposure. Data and Methods: We assemble a grid cell-year panel from 2007-2023 restricted to states in the western continental US. We match to each proximate air quality and NAAQS compliance status from the US EPA and prescribed burning on National Forest land from the US Forest Service. Wildfire data comes from Monitoring Trends in Burn Severity (MTBS). We test for a decrease in prescribed burning after particulate matter standards became stricter in locations that are near or above the new compliance threshold relative to locations that are just below the compliance threshold using panel data with cell and year fixed effects. Contributions: This project provides novel empirical evidence of effects of NAAQS on fuel management. The economics of wildfire management and suppression is of increasing interest. Within this work, the roles of air quality regulation has been widely speculated but never established beyond anecdotal evidence. Moreover, the analysis may inform policymakers on the degree to which current policy presents a conflict between hazardous fuel reduction and incentives to abate air pollution.

Access to clean cooking fuels is essential for improving health, equity, and environmental sustainability globally. Despite ongoing policy efforts, a large number of households, particularly in developing countries, continue to rely on polluting cooking fuels, contributing to adverse health outcomes, environmental degradation, and economic hardship. This study uses panel data covering 102 developing countries from 2000 to 2022 to empirically examine the effects of clean cooking access on human wellbeing. Employing fixed effects and instrumental variable approaches, we address endogeneity concerns and identify causal impacts. We also investigate rural-urban disparities, and find that urban populations experience larger improvements in health and nutrition per percentage point increase in clean cooking access. Focusing on sub-Saharan Africa, we simulate multiple scenarios of increased clean cooking access and estimate substantial gains: Raising rural access from 10% to 25% could avert approximately 84,000 infant deaths and prevent 1.68 million cases of child stunting. Achieving universal access in urban areas—an increase from 35% to 100%—could avert 377,000 infant deaths and yield over 11 million disability-adjusted life years, reflecting substantial reductions in household air pollution-related disease burden. These findings underscore the need for urgent, inclusive, and context specific policy interventions to close the clean cooking gap to advance global health and development goals. While rural households remain a priority, the scale of potential benefits in urban areas highlights the importance of a dual-focus approach in clean energy planning

Significant decarbonization demands a system that can keep pace with shifting policy priorities, technological updates, and the economic realities shaping emissions. This paper develops a structured overview of important objectives in contemporary carbon-market design by clarifying the typologies of carbon permits and carbon credits, examining the interactions between regulators, exchanges, and corporate actors, and reviewing the Global Carbon market model, which showcases a best-practice issuance pathway. By applying our modified Vickrey price auction (VPA) algorithm, we show how integrating auction design with verification and rollout processes can improve permit allocation while reducing common compliance obstacles. Our framework outlines how trans-parent permit allocation enables tradable credit, how cycles can be optimized to foster market confidence, and how interoperable registries can improve transparency and scalability.

The invasion of lionfish (Pterois miles) in the Mediterranean Sea presents a growing ecological challenge with increasingly significant economic implications. This article analyzes the economic dimensions of the invasion, focusing on its impacts on small-scale fisheries, tourism-dependent coastal economies, and associated livelihood systems. By preying on commercially valuable species, lionfish exert direct pressure on fisheries operating in economically fragile regions, contributing to reduced productivity, income instability, and potential food security concerns. Drawing on empirical observations and regional case studies, the analysis highlights sector-specific economic effects. In fisheries, lionfish alter catch composition, increase bycatch, and raise operational and safety costs due to their venomous spines. In the tourism sector, economic impacts are largely perception-driven, requiring adaptive responses by dive operators, while the species’ visual appeal simultaneously creates niche tourism opportunities. The article further examines economic adaptation strategies centered on the valorization of lionfish as a marketable product. Case studies from Cyprus, Greece, and Turkey illustrate diverse approaches, including structured removal programs, gastronomy-led market development, and public engagement initiatives. A comparative analysis with the rapid commercialization of the invasive blue crab in Tunisia demonstrates the potential for invasive biomass to transition from economic liability to asset under supportive institutional and market conditions. The findings suggest that valorization can partially mitigate the economic impacts of the lionfish invasion while aligning economic incentives with ecological management objectives. However, successful scaling requires coordinated investment in infrastructure, safety protocols, supply chains, and consumer awareness, as well as continued ecological monitoring to ensure long-term effectiveness.

Introduction The Conservation Reserve Program (CRP) currently enrolls approximately 16 million acres of cropland. The program specifically aims to cost-effectively reduce erosion and, for this reason, targets cropland that is marginal, in terms of erodibility and productivity. This research seeks to measure how and to what extent land enrolling in the CRP differs from land not participating in the program. Data To examine the productivity and land characteristics of land in CRP, we link spatially explicit program data to publicly available soil data, as well as yield from remote sensing and from the Risk Management Agency (RMA). We link our panel of parcel data to operators and landowners to examine differences within and between farming operations and landholdings. For the yield analysis, we observe almost four million panel fields in the Midwest and Central and Northern Great Plains over the period 2013 through 2024. For all other comparisons, we observe over 16 million fields during this period. Methods The primary approach compares land that is eventually offered to or enrolled in CRP to land within the same tract or county that does not participate, using multidimensional fixed effects. We also extend this comparison to within land sharing a common primary operator and compare conditional distributions, using deviations from county or tract averages. We conduct this analysis for multiple characteristics of interest, including productivity, land classification, and erodibility. Results Preliminary results indicate that land enrolling in CRP in the following year is between five and six percent lower in estimated productivity, on average, compared to land not subsequently enrolled within the same county, using the remote sensing measure of yield. The difference in productivity is smaller in magnitude within owner. Land enrolled in the following year also tends to have lower predicted productivity and higher erodibility. Conclusions There exist longstanding policy debates around the kinds of land which CRP enrolls and what kind of farmland it might be removing from the rental market and the productive sector. These debates are directly tied to program decisions, including payment rates and targeting methodologies.

Introduction. Identifying the long-run economic and environmental effects of land reclamation is empirically difficult because few large-scale historical cases exist. This study examines the consequences of coastal polders constructed under Bangladesh’s Coastal Embankment Project (CEP) during the 1960s and 1970s, which protected approximately 1.2 million hectares from saline intrusion. While polders are often viewed as tools for agricultural modernization and climate resilience, little evidence exists on their long-run impacts. This paper examines whether polders facilitated structural transformation or instead generated a “polder trap,” in which settlement expanded into marginal land without sustained productivity gains. Data and Methods. I construct a new subdistrict-level (upazila) panel dataset linking geocoded polder boundaries to agricultural censuses (1975–2019), population censuses (1941–2011), satellite-derived surface water data (1984–2021), nightlight intensity, and tidal river data. The dataset includes 463 upazilas nationwide, including 138 within polder-protected coastal zones. Agricultural outcomes include cropping intensity, HYV adoption shares, irrigation use, and land use patterns. Environmental outcomes include water occurrence and persistent waterlogging measures. Because polders were non-randomly placed in low-lying flood-prone regions, I develop a hydro-geomorphological instrumental variable based on the interaction between low elevation and proximity to tidal rivers. This interaction captures engineering suitability for embankment construction while separately controlling for elevation and river proximity. I estimate two-stage least squares models and complement them with spatial regression discontinuity and difference-in-differences designs exploiting cyclone shocks and post-1990 adaptation projects. Results and Conclusions. Preliminary results indicate that polders generated substantial early population growth relative to comparable non-polder areas. However, instrumental-variable estimates show weaker adoption of high-yield rice varieties and no sustained gains in agricultural productivity within polder regions. Satellite evidence further indicates greater long-run surface water persistence consistent with drainage congestion and waterlogging. These findings suggest that while polders attracted settlement, they did not generate sustained structural transformation. Instead, large-scale land reclamation in environmentally fragile regions may have anchored labor and settlement in marginal land without long-run productivity growth. The results provide policy-relevant lessons for climate adaptation planning in low-lying delta economies.

Corn production relies heavily on chemical fertilizers. Due to concerns over yield penalties from unpredictable weather and inconsistent application guidance, farmers often use a “build and maintain” approach for phosphorus (P) and potassium (K). This practice may reduce profits due to P and K losses and lead to water contamination through P runoff. This study has three objectives: (1) to examine whether farmers are willing to accept payment to adopt a P and K management practice that may reduce P and K application rates, P runoff, and yields but maximize profits; (2) to investigate how changes in P and K prices affect farmers’ adoption preferences; (3) to explore the impact of socio-economic factors on those preferences. We use a choice experiment (CE) to accomplish these objectives. The CE presents hypothetical scenarios with various combinations of five attributes and their levels with the adoption of profit-maximizing P and K rates. Additionally, an information treatment is included to provide hypothetical price changes for P and K. Respondents are asked to state their preferences among two alternative options and a status quo on each choice card. A mixed-mode survey was conducted following the methods as outlined by Dillman et al. (2014). A total of 328 complete surveys were received from a random sample of 2,500 corn farmers across 12 Midwest states. We employ mixed logit models and latent class models to investigate farmers’ heterogeneous preferences. We find that, on average, farmers’ marginal willingness to accept for reducing P, K, and yield per acre amounts to $0.49 per pound, $1.38 per pound, and $2.40 per bushel, respectively. However, their preferences are heterogeneous. Our findings suggest that some farmers are willing to forgo a portion of compensation due to the reduction in P costs. In general, farmers have no preferences for reducing P runoff, but certain farmers may value the environmental benefits associated with decreased P runoff. Our findings indicate that price changes in P and K do not influence farmers’ preferences for different attributes. Moreover, younger farmers and those with higher education attainments are more likely to adopt the proposed practice.

Agriculture is highly vulnerable to climate change. Recent research has highlighted producer responses through adjustments in yields, crop choices, and harvest timing, but less is known about land-use responses. This paper examines how farmers adapt their land-use decisions in response to changing climate conditions. We ask three main research questions: 1) How do weather events influence land conservation and abandonment? 2) Are producers using CRP as a climate adaptation strategy? and 3) Are conserved and abandoned lands substitutes? To answer these questions, we compile data for CRP contracts, cropland abandonment from Xie et al. (2024), the Cropland Data Layer, and PRISM weather records for the contiguous US from 2008-2023. We estimate the short-run effects of weather events on both land conservation and land abandonment using a panel model with unit and time fixed effects. We study long-run responses using a long-difference model. Identification comes from deviations in extreme degree days and precipitation from the location-specific means. Recent concerns with these approaches center on the fact that locations may systematically experience different levels of warming under climate change. Specifically, global warming induces mechanical trends in extreme temperature exposure that are correlated with a location’s baseline temperature (Jones et al. 2025). To address this, we use a recentering procedure that creates a distribution of counterfactual shocks, averages across these counterfactuals, and yields a single exogenous treatment shock (Borusyak and Hull 2023). Preliminary results find that producers respond to extreme heat by reducing cropland cultivation and modestly increasing land abandonment, indicating that abandonment serves as a limited climate adaptation strategy. In contrast, CRP enrollment declines following extreme weather events in both the short and long run. This suggests that the CRP is not being used as a climate adaptation strategy. Long-run results show that responses in cultivation and abandonment flatten over time, implying limited scope for adaptation through these channels. Finally, CRP and abandoned lands differ systematically in historical land use, crop types, and climate characteristics. These differences suggest that CRP targets landscapes distinct from those naturally prone to abandonment, with important implications for land-use change, environmental additionality, and climate adaptation.

I document that immigrants are more likely than natives to increase their labor force participation following surprise monetary contractions. Specifically, I find that after a 100 basis point monetary policy surprise, immigrants’ labor force participation increases by 20 to 30 basis points more than that of natives. Financial frictions affecting immigrants and an “added worker effect” are two key mechanisms that account for immigrants’ participation response. My analysis is based on panel regressions using monthly individual-level data from the Survey of Income and Program Participation (SIPP) over the period 1996 to 2019. I control for individual fixed effects, state-by-month fixed effects, and time-varying demographic characteristics, as well as for differential effects of changes in aggregate conditions using standard macroeconomic indicators. My findings contribute to our understating of the heterogeneous impact of monetary policy across demographic groups, while also highlighting that there are non-trivial effects of monetary policy on labor supply.

In this paper, I examine the impact of Fed monetary policy on global stock market returns and explore the role of financial development in transmitting these effects. To achieve this, I develop a two-country Dynamic Stochastic General Equilibrium (DSGE) model that accounts for heterogeneity in financial development across economies. The framework incorporates key real-world financial market features, including nominal price rigidity, monetary policy shocks, risk premium shocks, and uncovered interest parity shocks, within an incomplete asset market structure.

We study the racial incidence of monetary policy in a two-asset HANK model calibrated to the U.S. economy. Black and White households share identical preferences and financial technology but differ through three empirically disciplined channels: a residual earnings gap from CPS microdata, race-specific job-finding and separation rates from matched CPS flows, and race-specific entry wealth from the Survey of Consumer Finances. These differences generate persistent gaps in liquid wealth, illiquid wealth, hand-to-mouth status, and borrowing exposure. Monetary policy therefore transmits unevenly across racial groups through debt service, precautionary saving, portfolio rebalancing, and cyclical labor-market transitions. We show that a contractionary monetary shock disproportionately burdens Black households, and a counterfactual decomposition reveals that intergenerational wealth differences and differential separation-rate cyclicality are the dominant channels. The framework provides a tractable way to quantify the racial dimension of monetary transmission that standard representative-agent and race-neutral heterogeneous-agent models cannot capture.

This paper introduces the dilution rate of government debt, the ratio of new short-term debt to unmatured long-term debt, to add state-contingent maturity management to the fiscal theory of the price level. Unlike existing linearized long-term debt approaches requiring constant-duration structures, the dilution rate permits dynamic maturity policy. A dilution rule expands conventional determinacy regions and introduces novel dynamics. Surprise maturity lengthening produces inflationary effects resembling a tax cut. When monetary policy aggressively implements conventional open market operations -- quantitative tightening with rising rates, easing with falling rates -- credible escalatory maturity lengthening suppresses inflation within the government budget constraint analogously to how the Taylor principle suppresses inflation within the Fisher equation. This theory may explain why 2023--2025 U.S. inflation remained well below its post-COVID peak despite continued deficits.

This paper investigates how access to generative artificial intelligence (AI), specifically ChatGPT, affects student attention, cognitive effort, and learning outcomes. The central research question is whether AI assistance acts as a complement to learning or as a substitute for attention, potentially undermining long-term knowledge retention. The paper contributes to the emerging literature on AI in education by introducing a formal attention-based mechanism through which AI alters learning behavior, and by providing causal experimental evidence on how AI precision shapes effort allocation and performance. To guide empirical analysis, the paper develops a two-period signal detection model in which students choose whether to exert costly attention effort in the presence of an AI signal of varying precision. The model predicts that higher AI accuracy reduces early effort by lowering the marginal returns to attention, thereby improving short-term task performance but weakening subsequent performance when AI is removed. The empirical strategy relies on a controlled laboratory experiment conducted with undergraduate students at the University of Tennessee. Participants were randomly assigned to one of three groups: (i) no AI access, (ii) AI assistance with 100% accuracy, and (iii) AI assistance with 50% accuracy. The experiment simulates a classroom environment with instructional modules, AI-assisted assessments, and a final quiz without AI access. Key outcomes include task performance, quiz performance, and time spent on learning activities as a proxy for attention. The final analytical sample consists of 58 participants, with additional pilot data used for robustness. Results show that access to fully accurate AI significantly increases performance on AI-assisted tasks but is associated with lower performance on the subsequent quiz, consistent with a substitution away from attention toward AI reliance. Exposure to unreliable AI does not improve performance and, in some cases, reduces engagement with instructional material. Although some estimates lack statistical significance due to limited sample size, the overall pattern supports a trade-off between cognitive effort and AI use.

The rapid expansion of legalized sports betting by U.S. states after a 2018 U.S. Supreme Court ruling created the conditions to evaluate the effects of the legalization of sports betting on college student outcomes. This paper has two specific aims: 1) To test whether the legalization of sports betting affects college student retention and graduation rates; and 2) To determine whether state-level differences in the design of sports betting policies differentially effect college student outcomes. We use publicly available data on college student outcomes and a difference-in-differences research design to estimate the effect of changes in state-level sports betting policy on college student and student-athlete outcomes. The analyses in this study will help provide policymakers with causal evidence as to whether the legalization of sports betting has had negative consequences on the academic outcomes of college students.

Labor markets generally price in any shocks to non-pecuniary amenities. Theoretically, a negative shift in amenities will cause workers on the margin to find new jobs if not compensated for this disutility. This paper combines unique administrative from the U.S. Army combined with a natural experiment. From 2003-2006, the military increased its presence in the Iraq, and needed additional soldiers. To meet goals, the Army conditionally randomly assigned several units to extend deployment lengths for about three months on average without offering additional compensation.. We find a 15-percentage point decrease in likelihood of reenlistment at the end of the first term for extended soldiers controlling for unit, year, and other soldier characteristics. These findings show that when firms do not allow for compensating wage differentials workers will separate when faced with shocks to job quality.

Algorithmic risk assessment (ARA) tools are increasingly used across the criminal justice system to inform decisions at nearly every stage—from pretrial release to sentencing, parole, and post-confinement supervision. As of 2023, more than 42 states required the use of ARA tools at some point in the judicial process. Some states mandate their use during the pretrial stage, while others apply them to sentencing, parole, or post-confinement supervision decisions. Although prior research has examined the effects of ARA tools themselves, our study is the first to investigate the effects of ARA laws that mandate or regulate their use. We address several research questions, focusing on whether ARA laws influence reoffending, probation outcomes, sentence length, and parole decisions. We also examine whether the effects of these laws vary across demographic groups, with particular attention to age, criminal history, and race. To do so, we merge newly compiled data on the content, scope, and effective dates of state ARA laws with restricted-access individual-level administrative data from the U.S. National Corrections Reporting Program. These data include more than 22 million prison-term records and over 4 million post-confinement records from 2000 to 2020, covering the universe of state prisoners. Our identification strategy leverages variation in the timing of ARA law adoption across states using staggered-treatment difference-in-differences estimators. We initially focus on ARA laws governing parole decisions, examining whether these laws affect the likelihood of parole and subsequent rates of reoffending.

We study how short-term mandatory overtime affects worker wellness and productivity in a high-burnout public-sector setting. In 2022-23, the Chicago Police Department imposed cyclical, pre-scheduled cancellations of officers’ regular days off. Using daily officer-level records, we estimate stacked difference-in-differences models that leverage policy-induced short-run variation in workload. We find little evidence that the policy meaningfully affects attendance or wellness: we rule out attendance declines above 1.6 percent, injury increases above 1.7 percent, and sick-leave increases above 2.7 percent. Effects on productivity and enforcement outcomes are generally modest. We rule out declines in calls for service response greater than 2 percent, with congruent but less precise estimates for rarer

In this paper, I leverage exogenous variation in the timing of food assistance benefit disbursal to estimate a causal intent-to-treat effect of food scarcity on incidence of injury. It is well-documented that Supplemental Nutrition Assistance Program (SNAP) recipients spend the majority of their benefits in the first two weeks after disbursal, leading to heightened food scarcity at the end of the benefit month. Prior research has shown that this pattern, known as the SNAP cycle, can negatively affect cognitive performance and health, but a link between food scarcity and safety has not yet been established. Using 10 years of workers’ compensation claims data from Oregon, I find that individuals who are likely SNAP recipients experience an approximate 5 percent increase in occupational injuries at the end of the benefit month relative to the beginning. My preliminary findings suggest that this pattern is primarily driven by cognitive, rather than physiological, impairments. Ongoing extensions of this work consider heterogeneity across states and occupations, as well as the moderating effects of paycheck cycles, staggered disbursal, and increases in benefit amounts.

To operate the UI system, state workforce agencies incur roughly $3 billion in administrative costs each year. Yet models of optimal unemployment insurance (UI) implicitly assume no administrative costs, i.e., one dollar of additional tax revenue translates to one dollar of additional benefit payments. This paper examines the impact of UI these administrative costs on optimal UI policy. First, I add administrative costs to the standard Baily-Chetty framework. Next, using accounting data from state workforce agencies from 2002-2019, I use two separate approaches (IV and accounting approach) to estimate the administrative costs associated with an additional dollar of UI benefit expenditure. Each dollar of benefit payments generates administrative costs of around five cents. Lastly, I evaluate the optimal policy and welfare implications. Administrative costs reduce the optimal UI replacement rate from 28% to 21%, with larger reductions in alternative specifications.

Workers displaced by trade, automation, or other structural shocks face a form of labor-market risk that is neither ordinary unemployment nor disability. The loss is persistent, because displaced workers often lose the market value of task-, firm-, industry-, or occupation-specific human capital, but it is not necessarily permanent, because workers can rebuild human capital through reemployment and on-the-job investment. This paper develops a theory of human capital insurance (HCI) for this intermediate risk. In a Ben-Porath model with displacement risk, workers choose between risk-free financial saving and risky human capital investment. Because no asset pays off when displacement destroys human capital, workers self-insure through precautionary saving and underinvest in human capital. The first-best HCI policy is the missing Arrow security for risky human capital: it pays the displaced worker the fall in continuation wealth caused by the human capital reset and finances these claims with actuarially fair premia. I then restrict the planner to a real-world wage insurance instrument modeled on Reemployment Trade Adjustment Assistance (RTAA), which replaces a fraction of the post-displacement wage loss for a limited duration. I show that wage insurance can replicate first-best HCI when the duration is allowed to depend on age and human capital type, and I quantify the second-best gains of simpler scalar policies. Empirically, I use Trade Adjustment Assistance participant records and a Bowlus-Robinson price-deflation strategy to construct a human capital loss proxy. A proof-of-concept estimate for non-RTAA participants implies a residual human capital loss of about 44 percent over the first three post-displacement quarters. Preliminary indirect-inference results imply that RTAA-style wage insurance raises mid-life ex-ante human capital by roughly 25 percent and generates a certainty-equivalent welfare gain of about 5 percent; the best scalar wage insurance policy in the current calibration raises mid-life ex-ante human capital by roughly 62 percent and generates a welfare gain of about 10 percent.

This paper uses innovative new micro transactions data on consumption and income from a payment diary (Gilyard and Schuh, 2025) to estimate daily consumption Euler equations for the first time. Point estimates for total consumption and income closely replicate the literature, verifying data quality, although annual estimates are imprecise. As usual, the life-cycle permanent-income hypothesis (LC-PIH) model is rejected due to excess sensitivity when estimated with the simplest specifications, data aggregated across agents, and consumers constrained by liquidity and wealth; novel self-reported data on unexpected consumption cannot explain excess sensitivity. However, disaggregated daily consumption and income reveal new insights. Daily consumption smoothing is more evident for: 1) consumers who receive discrete income more often (weekly); and 2) non-bill expenditures that minimize the distinction between continual consumption and discrete expenditures (Aguiar and Hurst, 2005). Convenience samples of “Big” transactions data likely reflect non-trivial selection effects.

Macroeconomic and household finance research often carefully accounts for income shocks but completely neglects expense shocks. We use a survey linked to credit bureau records to estimate expense shocks' frequency, size, and financial implications. Expense shocks are more common and larger than income shocks. Very large expense shocks are several times more common than very large income drops. Researchers can use our results and publicly available data to account for expense shocks. When we do so in an otherwise standard model, we find: (1) many costly decisions to increase liquidity such as asset liquidations or loan defaults are driven not by strategic considerations but by expense shocks, (2) smoother consumption implies less insurance, not more, (3) expense shocks explain most precautionary wealth, and (4) expense shocks mostly determine the marginal propensity to consume. Hence, expense shocks matter, and they matter more than income shocks in many important contexts.

A tension exists in the fiscal stimulus literature: household-level studies estimate substantial marginal propensities to consume (MPCs) out of temporary government transfers, yet aggregate time-series data from the 2001 and 2008 U.S. tax rebate episodes show little discernible impact on total consumer spending. This apparent conflict has cast doubt on whether existing heterogeneous agent New Keynesian (HANK) models calibrated to micro MPC data can reliably predict aggregate outcomes. We bring new daily retail spending data to the March 2021 Economic Impact Payments—which were roughly twice the size of the 2008 rebates and distributed predominantly within the first week—and show a sharp increase in daily aggregate spending whose timing tracks the stimulus disbursement, well above that which could be explained by the relatively slower vaccine rollout and economic reopening. We support this interpretation with two further tests: the spending response is similar among households vaccinated before March and those vaccinated afterward, and the increase in March is largely concentrated among households below the income eligibility threshold for the payments, even though spending growth from the broader reopening is similar across income groups over a longer period. We then revisit monthly Census retail sales data spanning all three stimulus episodes, constructing counterfactuals based on intertemporal MPCs implying a monthly retail sales MPC of approximately 0.2. We compare these counterfactuals to those predicted by a Kalman smoother-based interpolation. For 2001 and 2008, both the actual data and the no-stimulus counterfactual remain within the forecast error bands—with the exception of the month of September 11, 2001—indicating that these stimulus payments were too small and too slowly disbursed to produce a statistically distinguishable aggregate signal. For 2021, actual spending lies well above the error bands, while the counterfactual that subtracts the MPC-implied stimulus effect fits comfortably within them. Our findings suggest that the micro-macro MPC disconnect is an artifact of statistical power, not a fundamental inconsistency, and that calibrating HANK models to micro MPC estimates remains well-founded.

This study examines how young workers value work arrangements, with a focus on remote-work flexibility, access to mentoring, and the presence of employer monitoring. We field a discrete choice experiment with 317 undergraduate students, varying job characteristics across hypothetical job offers to identify individuals’ tradeoffs among these three attributes. We use a mixed logit model to quantify the degree to which respondents value each job attribute, as well as estimating how the value of working from home (WFH) changes when mentoring opportunities or monitoring practices are introduced. Descriptive statistics highlight substantial heterogeneity in preferences for WFH, which persist after controlling for observable characteristics such as gender or field of study. On average, respondents view fully remote work negatively, but value flexibility in the form of arrangements that permit one or two days of WFH per week. Mentoring emerges as a consistently valued attribute, and its availability reduces the value respondents place on the ability to WFH, suggesting that early-career workers perceive a tension between remote work and effective mentoring. In contrast, the presence of computer-based monitoring has little systematic association with job desirability.

This paper examines the importance of remedial credits on baccalaureate degree attainment. Using an administrative dataset in the state of Kansas, the paper shows that appropriate placement into remedial credits is essential for maximizing degree attainment for students who otherwise would experience a relatively higher barrier to degree attainment. Students who need remedial credits and receive a remedial course perform comparably with students who did not need a remedial class and did not take one. Conversely, students who are over-placed into remedial coursework (those who do not appear to need remedial math but are placed into a remedial course) experience a significant degree penalty. The impact of remedial credits on student degree outcomes is compounded by the presence of complex pathways that students take between their first institution and their final institution. This paper quantifies the prevalence of complex pathways by institution and uses an innovative instrumental variables approach to decompose the effect of student-institution match and the effect of institution quality on degree outcomes.

Title IX exemptions allow educational institutions, that are associated with religious organizations to bypass specific anti-discriminatory rules if compliance conflicts with the organization’s religious beliefs. This exemption can be requested after a Title IX complaint has been filed, but some universities preemptively filed for exemptions. Over a 100 private, religiously-affiliated colleges have received Title IX exemptions first beginning in 2014, yet the effects of these exemptions are understudied. Using a coarsened exact matching (CEM) framework and data from the Integrated Postsecondary Education Data System (IPEDS), we find that receiving a Title IX exemption has no discernible effect on tuition or enrollment, slightly increases the percentage of students who are women, and decreases the number of external gifts the school receives.

We investigate the impact of virtual care adoption on primary care accessibility, utilisation, and patient welfare using administrative claims data from British Columbia, Canada, spanning 2018 to 2022. In this universal single-payer setting, where point-of-service costs are zero, time acts as the shadow price of access. We contribute to the literature by isolating the distance-mitigating effects of virtual care from endogenous provider sorting and quantifying welfare implications through an opportunity cost framework. Methodologically, we employ difference-in-differences specifications to estimate utilisation responses and non-parametric comparisons of travel distance distributions to analyse spatial frictions. We further estimate the lower bound of Compensating Variation and the upper bound of Equivalent Variation to assess the economic value of time savings generated by virtual modalities. We document four primary findings. First, virtual care adoption is associated with increased visit frequency for established patients (intensive margin) but is uncorrelated with patient participation (extensive margin), which contracted during the pandemic. Second, virtual visits exhibit first-order stochastic dominance over in-person visits regarding travel distance, even within fixed patient-provider pairs, indicating that the technology mitigates spatial frictions. Third, utilisation responses display significant heterogeneity: females, working-age adults, and residents of rural areas with strong metropolitan ties experience the largest relative increases in visits. Finally, welfare analysis reveals that while the savings from substituting in-person visits with virtual ones are relatively homogeneous across genders, the welfare losses from a hypothetical removal of virtual care would be highly concentrated among non-metropolitan residents and females. These results suggest that virtual care primarily benefits those facing the highest physical access barriers by relaxing non-price rationing.

How does working more or fewer hours affect workers' health? An existing literature finds that retirement is usually a net positive for health. But could simply cutting back to part-time work deliver comparable health benefits – or perhaps even greater health benefits, if there is value in keeping active? We measure effects of hours worked per week on a variety of health outcomes in a population of older US adults from the Health and Retirement Survey (HRS), allowing for nonlinear effects. We use an econometric approach which exploits the fact that there is bunching at 0 and 40 hours of work to perform model selection. This bunching is helpful for model selection because it creates discontinuities in all moments of the outcome variable, and because, under the assumption that treatment effects are continuous, those discontinuities are solely attributable to selection bias. Expanding on Caetano's (2015) test for regression models, we construct a diagnostic for selection on observables models based on whether the model can simultaneously exactly account for the discontinuities at both bunching points in the first three moments of all 11 health outcomes – a total of 66 discontinuities to be exactly explained. Many models unambiguously fail this test, but models which control for detailed health history and demographics pass. The blind spot of the test is if there are confounders which are continuous in their first three moments at both bunching points. We argue this is very unlikely. For instance, a joint test for continuity in the first three moments at both bunching points rejects for 87 out of 88 observed variables correlated with hours worked. We also develop a sensitivity analysis to assess whether a failure to reject due only to sampling error could mask substantial bias. Our main finding is that most health outcomes are maximized at 20-30 hours of work per week. Effects of working that many hours (relative to zero hours worked) range between 0 and 0.1 SD, depending on the outcome. Our results suggest that part-time work on average delivers at least as many health benefits as retirement.

We examine the long-term impact of a large-scale women’s empowerment policy on psychological well-being, using state-level amendments to inheritance laws in India as a natural experiment. Exploiting variation in the timing and geography of the Hindu Succession Act Amendment within a difference-in-differences framework, we find that women exposed to the reform exhibit significantly better mental health outcomes in later life. Specifically, the reform reduced the likelihood of depressive symptoms by 5.9–6.4 percentage points and improved a composite mental health index by 0.08σ to 0.093σ. Falsification tests show no effects among individuals from the same state-cohort groups belonging to communities to which the reform did not apply. We demonstrate improvements in multiple socio-economic indicators as mechanisms for the downstream effects on mental health: treated households are 4.2 percentage points more likely to own land, and treated women report a 0.11σ increase in autonomy. These results highlight the long-term benefits of women’s empowerment policies in improving their mental well-being.

Sufficient sleep is critical to human health and cognitive function, yet identifying exogenous sources of sleep disruption remains methodologically challenging. We exploit the timing of the full moon as a plausibly exogenous source of ambient nighttime light to estimate its effects on sleep and fatal car accidents. Using self-reported sleep data from the Behavioral Risk Factor Surveillance System (BRFSS) and fatal accident data from the Fatality Analysis Reporting System (FARS) spanning 1975–2023, we find that the days immediately following the full moon are associated with significant reductions in sleep and increases in fatal car accidents. Effects are strongest in sunny states during non-winter months, consistent with an ambient moonlight mechanism, and are concentrated in the pre-smartphone era (pre-2011), suggesting that simulated light from devices may have displaced the moon's historical role in disrupting sleep. In our preferred specification looking at sunny states in non-winter months before 2011, we find a 2.2% increase in both accidents and fatalities following the full moon. Placebo simulations confirm that it is unlikely that our results could be an artifact of the underlying variation in the data. These findings contribute to the growing economics of sleep literature by documenting a novel, historically meaningful source of sleep disruption and its downstream consequences for road safety.

While the labor market and health effects of immigration enforcement measures are well-documented in the economics literature, virtually no studies explore the role of these interventions in determining immigrant workers’ commuting patterns, despite the question’s policy relevance. Examining the connection between such programs and foreign-born workers’ transportation mobility would not only inform us of immigrant communities’ perceptions of enforcement but would also carry urban planning, environmental, and labor market implications. My paper exploits the implementation of employment verification mandates, like E-Verify, and police-based programs, such as 287(g) agreements, to create a quasi-experimental framework to analyze how immigrants’ behavior responds to these policy tools. Using American Community Survey data, I find that E-Verify and jail agreements increase workers’ likelihood of using public transportation, with the former also reducing their travel times. Conversely, task force agreements prompt immigrants to substitute away from using public transportation and drive to work instead. My results are highly statistically significant and are particularly pronounced for likely undocumented immigrants.

This paper examines the effects of degree-restrictive free college programs on postsecondary credential attainment. While prior work on broad-access programs, such as Tennessee Promise, focuses on increasing overall enrollment and degree completion, less is known about how restricting eligible fields of study shapes educational outcomes. I study three state programs, Work Ready Kentucky, Arkansas Future, and Build Dakota, all of which subsidize tuition only for selected fields. Using institution-level data on credentials conferred, I find that the introduction of these programs is associated with a significant decline in associate’s degrees awarded by two-year institutions in participating states, alongside a substantial increase in certificate attainment and overall graduation rates. These results suggest that degree restrictions shift students toward shorter, more targeted credentials rather than increasing traditional degree completion, highlighting an important tradeoff in the design of free college programs with implications for workforce development and the composition of postsecondary attainment.

When a platform enters the markets it creates, the entry often impacts users of the platform, including both downstream rivals and non-rivals. This paper investigates this issue in the app economy by studying the effects of platform entry on third-party apps. We leverage natural language processing and unsupervised machine learning to cluster apps using pairwise cosine similarity, which provides a measure of horizontal differentiation within app categories. We find positive effects of rivalry from first-party apps. Both entry by Apple and Google into a cluster stimulates downloads of third-party apps, though the magnitudes differ. We also find that apps that more closely resemble the most popular apps have fewer downloads, whereas greater similarity to ordinary competitors enhances downloads. These findings yield important implications for developers in choosing market segments and designing effective differentiation strategies.

I study how third-party firms respond when a platform owner enters its own marketplace, analyzing Apple's entries into App Store submarkets from 2016-2021. Using text embeddings to define markets and a staggered difference-indifferences design, I find that Apple's entry deters new competitors and shifts incumbents' monetization strategies, but effects vary widely: many markets show no meaningful response, while others move in opposing directions across a host of monetization and quality outcomes. Responses depend on how Apple enters and apps' competitive proximity to Apple. This heterogeneity suggests targeted oversight rather than categorical restrictions on platform-owner entry.

This paper studies how commission rates affect app entry across platforms. We exploit a sharp increase in Android game commission rates—from 30% to 50%—in China in 2014 to estimate the impact on both Android and iOS app stores in a difference-in-differences framework. We document large declines in the number and quality of new games. Game entry falls by 47% on Android and by 30% on iOS. The industry-wide share of high-quality games decreases by 12 percentage points. These findings demonstrate platform complementarity on the supply side and show that higher ad-valorem fees discourage quality provision.

Ensuring adequate competition in mobile telecommunication is a basic policy goal addressed by regulators such as the U.S. Federal Communications Commission (FCC) when designing auction rules or approving secondary transactions. Ensuring adequate competition is also the basic question when anti-trust authorities review mergers, such as the "4-to-3" merger in 2020 between Sprint and T-Mobile in the U.S. A concern is that too much concentration can lead to a small number of firms exerting market power to decrease social welfare. However, in telecommunication, increased concentration may also lead to efficiency gains due to various scale effects. For example, Hazlett and Crandall (2024a) presents evidence that the Sprint/T-Mobile merger benefited consumers even though it led to fewer firms competing in the U.S. mobile market. This raises several basic questions. Does more concentration always benefit consumers and if not what effects determine the "optimal" number of competing firms in a market? How does this depend on the market size, the amount of bandwidth available or the scale effects that are at play? In this work, we present a game theoretic framework to give insights into these questions. Our model is based on a framework for Cournot competition with congestion effects that has been used to study competition in wireless markets and other setting where users are sensitive to congestion effects. Our model captures several different scale effects. We account for the investment incentives of a service provider as a function of their bandwidth holdings, where a provider with more spectrum may see a larger return from investing in infrastructure. We consider a congestion model that can capture "multiplexing gains" achieved by serving more users with proportionally more bandwidth. We also consider how operational cost may scale with an increasing number of users. And finally, we look at how imperfect collusion can affect the competition. We use this model to compare the economic welfare obtained with different number of competing firms and different distributions of bandwidth. Our results indicate that the optimal number of firms may vary depending on these effects, suggesting that policymakers should carefully account for them when evaluating market competition.

This study estimates the impact of long-term exposure to air pollution on high school academic performance using a panel dataset of Tehran high schools between 2014 and 2023. We match school-level standardized exam scores in eight subjects with school-specific pollution exposure measures for PM2.5, PM10, and NO₂ constructed from the city’s monitoring network using spatial interpolation, along with detailed school-level institutional and geographic characteristics. Our empirical strategies include fixed-effect models with alternative specifications to address heterogeneity, nonlinearity, and endogeneity concerns. The results show that cumulative exposure to air pollution significantly reduces performance across most subjects, with the largest declines observed in subjects relying more heavily on memory and verbal processing, particularly Sociology and Grammar in the Humanities track, and Biology in the Sciences track. Moreover, schools located closer to major transport-related pollution sources, especially bus terminals and gasoline stations, experience substantially larger adverse effects, highlighting the role of localized pollution hotspots. The findings remain robust across alternative specifications, including district fixed effects, instrumental-variable estimation, and propensity score matching. These results suggest that sustained improvements in urban air quality and targeted interventions around high-exposure schools are essential not only for public health, but also for protecting educational equity and long-term human capital development in highly polluted urban environments.

Air pollution warnings are designed to protect public health by reducing outdoor exposure to hazardous particulate matter (PM), yet they may generate unintended behavioral spillovers. This study examines whether PM warnings inadvertently increase influenza transmission by encouraging individuals to spend more time indoors, where viral spread is more likely. Using daily province-level panel data from South Korea for 2015–2019, I investigate the impact of official PM10 and PM2.5 warnings on influenza incidence. I combine a standard binary measure of warnings with an exposure-weighted treatment that accounts for the population share within each province covered by city-level alerts. The empirical strategy relies on fixed-effects panel models with province fixed effects and flexible year–season–weekday fixed effects, supplemented by Fourier terms to capture seasonal flu patterns. Weather conditions, ambient PM concentrations, local socioeconomic characteristics, and vaccination rates are included as controls. Event-study and stacked difference-in-differences approaches are used to examine dynamic responses. Across specifications, PM warnings lead to a 10–15 percent increase in influenza incidence, with effects strongest among children and teenagers. Event-study estimates show no pre-trends and reveal a delayed surge in flu cases peaking 25–40 days after warnings—consistent with secondary transmission dynamics rather than direct pollution effects. Exposure-weighted estimates indicate substantially larger effects once measurement error from spatial aggregation is addressed. Search behavior data from Naver DataLab show significant increases in queries for indoor activities, masks, and air purifiers, and decreases in outdoor-activity searches, confirming warning-induced shifts toward indoor congregation. Effects are stronger on business days, during daytime hours when emergency text alerts are issued, and after the 2017 rollout of Emergency Disaster Texts (EDTs). The increase is concentrated in outpatient cases, indicating primarily mild-to-moderate illness. Robustness tests—including placebo timing, non-overlapping event windows, alternative pollution controls, and an examination of next-day pollution levels—support a causal interpretation. Overall, the findings reveal a previously overlooked behavioral externality: while PM warnings reduce outdoor pollution exposure, they can unintentionally heighten viral transmission risk by increasing indoor crowding. These results highlight the complexity of information-based environmental policies and underscore the need to incorporate infectious-disease considerations into warning-system design.

Obtaining a better understanding of the physician decision-making process is crucial, as it has significant consequences for patient health and healthcare costs. One factor that is likely to affect physician decision-making but has been largely overlooked is environmental quality, particularly air quality. It is well documented that exposure to air pollution can adversely affect cognitive functioning and alter risk attitudes, thereby influencing decision-making. This study adds new evidence to this line of research by examining how air pollution affects physicians’ decisions regarding cesarean deliveries (C-section). Our analysis is based on a combination of high-resolution air quality data with administrative birth records in Taiwan from 2004 to 2022. We exclude elective and pre-scheduled C-section, and medically necessary cases as these deliveries are unlikely to be affected by air quality. Leveraging variation in air quality within physicians over time, we employ a fixed-effects model to investigate how air pollution exposure on the day of delivery affects the likelihood that a physician chooses to perform a C-section. Our findings provide strong evidence that physicians’ decision to perform a C-section is significantly influenced by PM2.5 levels. A 1 μg/m³ increase in same-day PM2.5 leads to more than 3 additional C-section cases per 100,000 births. This represents a 0.27% increase in the probability of receiving a C-section relative to the average C-section rate of 1.3%. These results are robust to different model specifications and supported by a placebo test that demonstrates that air pollution does not affect birth-related emergency indicators, such as Apgar scores or birth weight . Additionally, we observe that these effects are heterogeneous, with stronger impacts observed among male physicians, deliveries in weekdays, and physicians with more delivery experience in a year. PM2.5 increases their C-section rates more significantly, possibly reflecting experience-induced risk aversion under poor air quality. Moreover, no significant effects are observed among physicians working in private hospitals compared with those in public hospitals. These results contribute to the growing body of evidence linking air pollution to decision-making. We highlight important implications for reducing medical burdens and provide new insights into how environmental risks may influence clinical judgment.

This paper analyzes the causal effect of air pollution on the number of traffic accidents in California from 2012 to 2019. The empirical strategy leverages exogenous daily variation in wind direction as well as the occurrence and strength of thermal inversions to account for changes in the daily average concentration of PM2.5. The estimation technique relies on the control function approach to address endogeneity and Poisson Pseudo-MLE. This analysis integrates county-by-day data from the U.S. Environmental Protection Agency, the National Centers for Environmental Information, the North American Regional Reanalysis, and the California Highway Patrol. I find that, on average, each additional μg⁄m3 of PM2.5 increases the number of accidents resulting in serious, minor and possible but non-visible injuries by 1%, 0.6% and 0.49%, respectively. Similarly, collisions resulting only in property-damage increase by 0.65%. I find no effect on fatal accidents.

Conventional VAR practice attributes unexpected inflation to monetary or other structural shocks and treats fiscal policy as neutral; fiscal variables are often omitted. In these settings, shocks to primary surpluses are not allowed to be an independent source of inflation. This paper removes that implicit restriction. We distinguish “funded” from “unfunded” fiscal shocks in post-war U.S. data. “Funded” shocks are offset by future primary surpluses so that the present value of current and future primary surpluses is unchanged; “unfunded” shocks are not offset. Because the real market value of government debt can be interpreted as the market’s assessment of the present value of expected future primary surpluses, the two cases imply distinct restrictions—for example, a funded deficit shock should raise the current real market value of debt one-for-one with the deficit, whereas no such restriction holds for unfunded shocks. Within a parsimonious structural VAR, we exploit budget-identity-based restrictions to identify funded versus unfunded shocks and quantify their effects on economic activity, achieving identification without relying on large structural models. Using U.S. data from 1959 to 2025, we find that unfunded fiscal shocks, on impact, raise inflation, boost real GDP growth, lower nominal return on government debt, and increase short- and long-term interest rates; funded deficit shocks leave inflation, real GDP, and nominal debt returns near zero, suggesting that funded shocks are close to exhibiting Ricardian neutrality.

The traditional bank-sovereign nexus has been a major financial stability concern, particularly following the eurozone crisis and the post-Covid rise in sovereign debt. However, non-bank financial institutions (NBFIs) have grown significantly to take center stage in international financial markets, particularly as key intermediaries in sovereign bond markets. This paper examines the transmission of risks among banks, sovereigns and NBFIs using bank-level data for European banks and aggregate data worldwide. We find that banks’ exposure to their domestic sovereign has recently become less important in explaining the risk co-movement between banks and sovereigns. Instead, banks’ exposure to domestic NBFIs has gained importance as an explanatory factor for the same co-movement. This is driven primarily by exposures to riskier sovereigns, either domestic or cross-border. We also find evidence that the importance of NBFI holdings of sovereign debt as drivers of the correlation between NBFI risk with sovereign risk has risen. Together, our results indicate that the traditional sovereign-bank nexus has morphed into a broader sovereign-bank-NBFI nexus.

In a small-open-economy New Keynesian model, we explore the implications of both rational, fully informed agents and boundedly rational agents, coupled with central bank prediction errors, on inflation dynamics. Central bank prediction errors serve as monetary shocks, leading to suboptimal short-term inflation and output gap dynamics. Regardless of how agents process information and form their expectations, inflation persistence is linked to the persistence of the monetary shocks. However, differences emerge in inflation volatility and sensitivity to domestic slack due to disagreements between rational and boundedly rational agents regarding macroeconomic persistence. This affects inflation forecastability in small open economies. We support these findings with evidence from a set of emerging market economies, analyzing the role of judgment in forming expectations with survey-based forecasts from Consensus Economics(R) and the International Monetary Fund. Finally, we introduce the Divine Coincidence Curve, which illustrates conditions under which monetary policy can simultaneously reduce output gap and inflation volatility. The presence of boundedly rational agents alters this relationship, influencing policy effectiveness and its transmission mechanism.

This paper examines how foreign investment in the UK and its components such as foreign direct investment (FDI) and foreign portfolio investment (FPI) affect the domestic macroeconomic conditions. We estimate a Dynamic Factor Model to extract comovements from 31 indicators to obtain three composite measures of the UK macroeconomic conditions, which correspond to real economic activity, monetary policy, and aggregate uncertainty. We find that shocks to FPI and equity-type foreign investments have the most significant positive impact on UK domestic economic activity and reducing aggregate risk, while shocks to debt-type foreign investments reduce domestic interest rates. As for the effect of structural shocks to the UK domestic markets, we find that an unexpected domestic monetary tightening has significant negative impact on FPI and debt-type foreign investments, while shocks to UK domestic output and risk conditions do not have significant impact on foreign investments.

Ownership stability also has a significant positive effect on the ESG performance of manufacturing firms. Large shareholders with long-term stable shareholdings are more inclined to support corporate sustainable development strategies and form a long-term governance consensus, thereby ensuring the continuity and stability of corporate strategic goals. They also have sufficient motivation and capability to monitor management’s ESG performance, providing sustained resource investment and oversight assurance for ESG projects. The mechanism tests confirm that ownership stability enhances ESG performance through two mediating mechanisms of information asymmetry and resource allocation efficiency. Specifically, large shareholders with long-term stable holdings have strong incentives to continuously monitor management, curb selective disclosure, and improve the quality of corporate information, thereby enhancing ESG performance. Moreover, a stable ownership structure sends positive signals to the outside world, alleviates firms’ financing constraints, improves resource allocation efficiency, and thus provides sustained resource support for ESG practices. Further analysis shows that the positive influence of ownership stability on ESG performance is especially pronounced in firms characterized by lower managerial overconfidence, weaker industry competition, and higher analyst coverage. This result suggests that the governance value of ownership stability is most salient in contexts where internal decision-making is conservative, external competitive pressure is muted, and external information scrutiny is robust.

Understanding whether countries are converging in their sustainable development trajectory is crucial for assessing global coordination efforts and the effectiveness of international sustainability frameworks. We analyze convergence in Economic, Environmental, Social, and Governance (EESG) performance for 56 countries from 2000 to 2022 using σ, absolute β, conditional β, stochastic, and club convergence tests. Results show significant but incomplete convergence, with global half-lives of 8.66 years under absolute β and 7.22 years under conditional β convergence, alongside notable regional variation. Stochastic convergence is present globally but absent in some regions, and club convergence reveals four distinct groups with divergent trajectories. These findings highlight that global sustainable development progresses at multiple speeds, shaped by institutional and macroeconomic factors, and underscore the need for context-specific, differentiated policy support to maintain convergence momentum and prevent widening disparities.

This study aims to explore the relationship between green finance, economic globalization, and ESG performance, utilizing annual data from 113 countries spanning from 2000 to 2020. The Westerlund cointegration test and the pooled mean group (PMG) estimation method are used to determine the long-term equilibrium relationship between the variables. The empirical results reveal green finance and economic globalization significantly promote ESG performance in the long term. However, in the short term, economic globalization has a significant inhibitory effect on ESG performance, while the impact of green finance is insignificant. Moreover, green finance significantly promotes ESG performance in the long term for low-pollution countries, but not for high-pollution countries, while economic globalization promotes ESG performance in the long term for both low-pollution countries and high-pollution countries. These findings remind policy-makers should take into consideration the long-term and short-term effects of green finance and economic globalization, when formulating policies aimed at enhancing ESG performance.

This paper investigates the validity of firm-level markup estimation techniques, particularly the cost accounting (CA) approach—a straightforward, transparent, and data-thrifty alternative that sees minimal modern application. While the production function (PF) approach has quickly become a workhorse method, several biases challenge its validity. The trend in U.S. markups is qualitatively and quantitatively robust to CA, suggesting that it is not driven by biases. I develop a validation test utilizing Dorfman-Steiner's (1954) advertising equation. CA tends to outperform PF, dramatically so in some cross sections. CA’s data-thrifty nature enables use when alternatives are infeasible; I illustrate this with several applications.

This study re-examines intergenerational income elasticity (IGE) in Korea by situating it within the structural transformation triggered by the 1997–98 Asian Financial Crisis. IGE, defined as the elasticity of children’s lifetime income with respect to parental income, is a key indicator of equality of opportunity. Earlier estimates based on the Korean Labor and Income Panel Study (KLIPS) yielded relatively low IGEs (around 0.2), suggesting high mobility. However, these estimates were produced during a period of extraordinary income volatility, delayed labor-market entry, and altered household formation, all of which compressed observed earnings profiles and biased mobility assessments downward. The Crisis destabilized education as the dominant parental investment channel by eroding its expected returns and amplifying uncertainty. As bankruptcies and unemployment proliferated, families increasingly relied on alternative mechanisms—particularly real estate transfers and inheritances—to secure offspring’s economic positions. These shifts imply that intergenerational persistence may have changed not in magnitude alone, but in its underlying transmission channels. To document this evolution, we compare pre-crisis father–son pairs observed in 1998 with post-crisis pairs formed in 2023 using retrospective parental characteristics. Treating the Crisis as a quasi-structural shock allows us to trace how human-capital-based transmission weakened while asset-based persistence gained salience. Methodologically, we employ a Two-Sample Instrumental Variables (TSIV) strategy that links KLIPS parental attributes to income distributions from the Household Income and Expenditure Survey. Unlike conventional estimators that correct life-cycle bias by scaling coefficients using an age-specific adjustment factor (θₛ), TSIV exploits the full age range of sons to recover parental permanent income directly. This approach better accommodates life-cycle heterogeneity and mitigates both classical and non-classical measurement error without relying on ex post attenuation corrections. Consequently, TSIV provides a conceptually cleaner benchmark for long-run IGE estimation. Our findings show that conventional multi-year OLS averages understate income persistence. TSIV-corrected IGEs are markedly higher—0.393 before the Crisis and 0.279 afterward—indicating that Korean mobility exceeds U.S. levels but has risen since the Crisis. Yet the expanding role of asset-based transfers suggests that intergenerational advantage has not diminished; rather, its mechanisms are shifting from labor income to wealth inheritance.

Forthcoming

We use data on the proliferation of cable programming at a local level to evaluate the effects of widely televised baseball - such as Braves and Cubs games through the superstations TBS and WGNA - on minor league baseball attendance. Our approach exploits heterogeneity in the rates of cable expansion at the designated market area level and changes to minor league baseball attendance in those areas from 1974 through 2010 in a staggered difference-in-differences framework, with continuous treatment. Recent work has suggested that television served as an “obvious” substitute for minor league baseball (Szymanski, 2025) in the 1950s. However, shortly after the proliferation of cable television and associated superstations in the late-1970s, minor league attendance increased dramatically. We hypothesize that as cable subscriptions reached new areas with minor league teams, these broadcasts served as complements to local minor league attendance, introducing the sport to a wider number of households that then sought out local options to view. While we are in the preliminary stages of analysis, a result implying such an effect suggests that the substitution relationships between television and attendance may not be stable over time.

Workplace injuries generate substantial costs for both workers and employers. Injured workers lose income and incur medical expenses, while employers lose access to valuable human capital. These costs highlight the importance of understanding the determinants of successful return from workplace injury. We examine return-to-play outcomes following one of the most common workplace injuries in professional sports: ulnar collateral ligament (UCL) damage among Major League Baseball pitchers. Using data on 433 MLB pitchers who underwent Tommy John surgery between 2000 and 2024, we analyze three dimensions of post-surgery recovery: the probability of returning to MLB, the duration of rehabilitation, and the timing of return using Cox proportional hazards models. We find that relief pitchers are significantly less likely to return to MLB than middle relievers, and starting pitchers who successfully return complete rehabilitation process more quickly and exhibit higher rates of return over time. Longer MLB careers prior to surgery reduce the likelihood and rate of returning, while greater pre-surgery workload is associated with a higher probability and faster rate of return, suggesting that more established pitchers may possess greater human capital and stronger incentives to complete rehabilitation successfully.

Forthcoming

The first major wave of occupational licensing and regulation in the United States occurred during the Progressive Era around the turn of the 20th Century, concurrent with rapid urbanization, technological innovation, and the emergence of entirely new occupations in fields ranging from healthcare to engineering. In this paper, we reassess the effects of occupational licensing, certification, and registration laws during this time period using a variety of novel data. First, we use recently-compiled data on the history of state and federal licensing policies, which significantly improves upon the accuracy of previously-available datasets. Second, we use data from full-count Census microdata. In addition to covering 100% of the U.S. population and providing fine geographic detail, we leverage the original write-in job titles recorded by Census enumerators to identify workers in licensed occupations. This reduces measurement error in workers' occupational affiliations and allows us to identify dozens of occupations that are not separately identifiable in the Census Bureau's public classification. Finally, we leverage new data on the qualifications needed to obtain a license for a subset of occupations including barbers and beauticians to study the response of employment to changes in mandatory training hours.

Near 25% of U.S. physicians are foreign-trained. Alongside federal immigration policy, states show increasing efforts to facilitate their integration. Yet we know little about whether such policies attract and retain foreign physicians, or about their implications for public health. A key empirical challenge is the lack of policy variation that is both widespread and plausibly exogenous. I study a major wave of state medical licensing reforms for foreign medical graduates (FMGs) from 1960 through the 1970s, when more than 30 state medical boards removed citizenship requirements for licensure. Using a newly digitized individual-level dataset covering all physicians in the US, and a staggered difference-in-differences design based on the quasi-random timing of removing citizenship requirement, I show that these reforms: (1) substantially increase the flow of new licenses to FMGs, raising total licenses granted by nearly 15%; (2) did not translate into a significant increase in the total stock of licensed FMGs; (3) the null stock effects were largely explained by reduced inflows of FMGs migrating from other states and by license renewals/updates among existing in-state FMGs; (4) the effects on licensing flows predominantly occurs after the easing of federal constraints on the transition to permanent residence; and (5) I find no evidence of worsening elderly or infant mortality. Together, these results suggest that lowering licensing barriers is not a panacea for physician shortage: while it resulted in no detectable adverse impacts on public health, it primarily reallocates, rather than expands, the physician workforce and only boosts license flows when federal work authorization is widely accessible.

Welfare analysis allows policymakers to weigh the various benefits and costs of active labor market programs on equal terms. However, these methods require strong assumptions for policies without an extensive evidence base. In this paper, I propose an approach for evaluating policy proposals using a minimum effect standard. Instead of asking whether a policy meets a certain welfare threshold, I invert welfare analysis functions to identify the minimum effect that policymakers should expect from a program to meet their welfare improvement goals, given expected costs. To do so, I identify a set of outcomes commonly reported in causal policy analyses of workforce training programs that are sufficient to determine that a program will meet welfare goals with a given level of confidence. I apply this approach to the context of subsidized employment programs. A large literature estimates the causal effects of workforce development programs on a variety of outcomes (e.g., see Card, Kluve, and Weber 2018) or uses these estimates to perform welfare analysis (Hendren and Sprung-Keyser 2020). This paper provides an intuitive way to use this research to inform policy decisions without requiring high-quality causal estimates on a broad set of outcomes. I also contribute to literature on optimal policy choice with uncertain benefits and costs (Manski 2021). This literature focuses on maximizing welfare by defining an optimal choice function across a defined policy space. I simplify this approach by focusing on one policy at a time but still considering how uncertainty in policy outcomes affects policy choices. Card, D., Kluve, J., & Weber, A. (2018). What works? A meta analysis of recent active labor market program evaluations. Journal of the European Economic Association, 16(3), 894-931. Finkelstein, A., & Hendren, N. (2020). Welfare Analysis Meets Causal Inference. Journal of Economic Perspectives, 34(4), 146–167. https://doi.org/10.1257/jep.34.4.146 Hendren, N., & Sprung-Keyser, B. (2020). A unified welfare analysis of government policies. The Quarterly journal of economics, 135(3), 1209-1318.

There is growing interest in understanding the effects of employer consolidations on the labor market. This paper examines whether U.S. college and university mergers affect faculty employment and wages. Using a difference-in-differences research design, I exploit local differences in exposure to college and university mergers occurring between 2000 and 2022. I find that while faculty employment increases at merging institutions, non-merging institutions located in markets exposed to mergers experience a decline in employment in the seven years following a merger. I do not find evidence of a wage effect at both merging and non-merging institutions. The employment reductions in non-merging institutions are stronger in markets where mergers increase local labor market concentration. Additionally, they appear to be largely driven by decreases in student enrollment. I observe corresponding increases in student enrollment at merging institutions, suggesting that they take advantage of their stronger position to capture a larger share of the product market, thereby reducing overall labor demand at non-merging institutions.

Income and wealth inequality have risen sharply within advanced economies, raising concerns about their social and health consequences. While the link between inequality and adverse health outcomes remains debated, the phenomenon of “deaths of despair”—drug overdoses, alcohol-related diseases, and suicide—has drawn attention in the United States. This study extends the analysis globally by examining whether labor market deregulation, specifically employment protection legislation (EPL) reforms that weaken workers’ bargaining power, contributes to deaths of despair. Using OECD indicators of employment protection and WHO mortality data for 26 advanced economies from 1998 to 2019, we estimate the dynamic effects of major EPL reforms on age-standardized death rates. Applying Jordà’s (2005) local projection method, we trace the temporal response of mortality to deregulation. Results show a significant rise in deaths from drug use disorders and suicide within five years of major EPL reforms. Further analysis indicates that governmental social transfers, unemployment benefit replacement rates, public health coverage, avoidable mortality, and poverty levels mediate these effects. These findings highlight the health risks of labor market deregulation and underscore the need for policies that strengthen social safety nets to mitigate deaths of despair.

The primary purpose of unemployment insurance (UI) programs is to afford beneficiaries the ability to smooth consumption over employment-related income shocks. However, little is known about whether UI prevents extreme material hardship. In this paper, we evaluate the effect of UI on one extreme outcome: homelessness. Using New York State and New York City administrative data, we estimate the impact of UI eligibility on homelessness using a regression discontinuity design (RDD) that exploits cutoffs based on workers' earnings in the past year. We find that being barely eligible for UI reduces homelessness within a year by 1.1 percentage points (a 19 percent decline). These effects persist for up to three years, well past the expiration of benefits. Not accounting for how UI prevents extreme distress undervalues the benefits of UI.

Using the United Nations' (UN) Principles for Responsible Management Education (PRME) policy as a quasi-natural experiment, we examine the effects of the Top Management Team (TMT) possessing Master of Business Administration (MBA) degrees on the corporate environmental performance in China. We find that the PRME-aligned MBA educational experiences of TMT have a significant positive impact on corporate environmental outcomes. Analysis of mechanisms suggests that TMT executives with PRME-aligned MBA qualifications face increased pressure to improve their corporate reputation and enhance the quality of information disclosure, thereby fostering corporate sustainability. Heterogeneity analysis suggests that the beneficial effects are more significant in high-profit firms, non-state-owned enterprises, and organizations managed by people with international experience.

Standard Navy Diver (ND) career paths are currently predicated on functional diversification, with NDs taking training and assignments in multiple mission area pillars over the course of their career. Some enterprise stakeholders believe that this approach has degraded ND core skills, constraining diver capability and leadership competency, and reports have found capability gaps in diving and salvage operations planning and coordination and inadequate supervision leading to diving mishaps in undersea medical support and salvage operations. The rating faces additional issues with respect to manning levels, training and qualifications, and executive agency. US Fleet Forces Command (USFF) N43 (Director, Fleet Maintenance) asked CNA to conduct a comprehensive study of the ND rating structure to determine the feasibility and advisability of mission area specialization. Using multiple methods, including SME discussions, focus groups, and inventory projection modeling, this study addressed the benefits and drawbacks of increased skill specialization and ways to address manning issues, training and qualification issues, and the executive agency question.

This memo outlines the methodology and findings of our analysis concerning insurance rates inside and outside Military Housing Areas (MHAs) for the continental United States (CONUS). Our analysis aimed to answer two primary questions: 1. Are there differences in insurance premiums paid within and outside MHAs? 2. To what extent does natural disaster risk affect insurance premiums in key US defense communities? From analyzing data on paid homeowners’ insurance premiums, we find that average paid premiums vary significantly between MHAs and non-MHAs, with some of that variation attributed to natural disaster risk. Along with this, our analysis found that zip codes connected with military installations are more likely to be in areas with higher risks of natural disasters, according to the distribution of risk categories in MHAs versus the risk distribution in the entire CONUS.

This paper studies the causal effects of physical fitness test failure and mandatory probation at the U.S. Naval Academy from 2002-2019 using a regression discontinuity design that exploits a sharp passing time cutoff in the 1.5 mile run event. Physical probation produces meaningful short run fitness improvements that persist for up to three semesters before marginal fitness test failers and passers converge to a common performance level. Fitness test failure also carries significant non-fitness consequences. Students who barely fail are less likely to graduate and perform worse on academic GPA trajectory measures, a crossover effect not detected by GPA level specifications. Both fitness and non-fitness effects are heterogeneous by seniority and season. Fall failures produce strong fitness improvement with minimal non-fitness consequences, while Spring failures produce minimal fitness improvement but carry significant retention and academic costs concentrated among students approaching the seniority threshold beyond which separation could incur financial or military service obligation.

Objective: Traumatic Brain Injury is a “signature injury” from the Global War on Terrorism. There is a paucity of research regarding its association with return to duty (RTD) in the combat theater. This exploratory study examines the association of documented TBI and other polytrauma, as well as blast injury, with return to duty in a cohort of surviving Veterans. Methods: Using available data from the DoD and VA Infrastructure for Clinical Intelligence Theater Medical Store Encounter database, we examined medical encounters for a cohort of surviving Veterans. We used ICD codes to identify TBI, Spinal Cord Injury (SCI), Burns, and blast injuries. We used Generalized Estimating Equations with a binomial family and a logit link to examine the association of these injuries with RTD, compared with other injuries and non-injury events in the combat theater between 2003 and 2015. Results: We examined 360,030 medical encounters for 66,215 surviving Veterans. Each Veteran had an average of 5.4 encounters. Generally, RTD declined with TBI severity, ranging from mild (OR=0.46; 95% CI 0.44, 0.49) to penetrating (0.07; 95% CI 0.04, 0.14), along with SCI (OR=0.14; 95% CI 0.08, 0.21) and burns (OR=0.61; 95% CI 0.48, 0.77). Blast as a mechanism of injury (OR= 0.62; 95% CI 0.58, 0.65), along with non-battle injuries (OR=0.43; 95% CI 0.42, 0.44) and battle injuries (OR=0.28; 95% CI 0.27, 0.30), relative to non-injury encounters, was also associated with reduced RTD. Finally, relative to 2002, RTD generally increased after 2006, with 2005 (OR=0.53; 95% CI 0.31, 0.92) the worst year, and 2013 (OR=3.80; 95% CI 2.22, 6.50) the best. Conclusions: This exploratory study of combat theater medical records shows that relative to other conditions, increasing TBI severity generally decreased RTD, as did other polytrauma conditions of SCI and burns. Blast mechanism of injury, as well as non-battle and battle injuries, also decreased RTD relative to non-injury medical events. The worst year for RTD was 2005, and the best was 2013 during the 2002-2015 study period. To fully understand the lifetime burden of combat TBI and polytrauma, it is important to start with the initial outcomes in theater.

This paper investigates the employment effects of initiating transactions between South Korean manufacturing SMEs and large business group firms (BGFs) over the period of 2015 to 2022. Using firm-to-firm transaction data and a panel difference-in-differences framework, we find that SMEs beginning to transact with BGFs experience a 3.5%p increase in employment growth. This effect strengthens with higher transaction intensity and persists over time. Further analysis reveals that the observed growth is closely associated with expansion of transaction networks and a higher probability of exporting. Employment gains are primarily driven by hiring surges, among full-time workers, underscoring the quality of job creation. Taken together, these findings suggest that rather than restricting inter-firm transactions out of concern for potential negative effects, SME policies should identify and promote the specific channels through which such transactions enhance SME growth.

This paper examines how artificial intelligence (AI) development is related to firm productivity and business dynamism in Japan. Using firm-level panel data from the Basic Survey of Japanese Business Structure and Activities matched with the Institute of Intellectual Property (IIP) patent database, we use AI-related patent applications as a proxy for AI development and analyze its association with productivity, firm organization, and industry-level reallocation. The empirical analysis combines fixed-effects estimation, event-study methods, and inverse probability weighting to address potential endogeneity. We find that firms engaging in AI development tend to exhibit higher productivity in the medium to long run, while displaying distinctive dynamics around the timing of AI development. In particular, productivity temporarily declines around the onset of AI development and subsequently improves. This pattern is consistent with adjustment costs associated with the adoption of new general-purpose technologies, but it is also compatible with endogenous selection into AI development, and therefore, the analysis does not make strong causal claims. Importantly, the association between AI development and productivity is heterogeneous across firms. The positive relationship is more pronounced among firms with higher initial productivity levels and larger firm size, while it is weaker among smaller and lower-productivity firms, highlighting the role of complementary assets such as skilled labor and organizational capabilities. We further show that AI development is associated with changes in firm organization and labor composition. While total employment does not decline significantly, the share of high-skilled workers increases, and the number of subsidiaries tends to decrease. At the industry level, greater AI development is associated with stronger reallocation toward more productive firms, linking AI development to business dynamism in the Japanese economy.

This paper examines how industrial robot imports affect trade upgrading and production outcomes in Chinese manufacturing firms. We construct an import-based firm-level proxy for robot adoption by combining the Annual Survey of Industrial Firms with China Customs Trade Statistics for 2000–2012. We first estimate firm fixed-effects models with industry-year fixed effects and then address endogeneity using inverse-probability-weighted difference-in-differences and two-stage least squares. The results show that robot imports are associated with higher firm output, value added, labor productivity, and employment. More importantly, robot-importing firms export more and export higher-quality products, while importing less in value terms but importing higher-quality intermediate inputs. These findings suggest that robot adoption is linked not only to efficiency gains within firms but also to a two-sided process of trade upgrading, whereby firms reorganize input sourcing toward better intermediates and translate improved production capability into higher-quality exports.

When consumers are borrowing constrained and mortgages are costly to refinance, the “tilt” induced by higher trend inflation front loading the profile of real payments of fixed amortization mortgages has real effects on consumer’s life cycle housing and con sumption. To quantify the welfare costs of anticipated inflation, we introduce nominal mortgage contracts into a calibrated life-cycle housing tenure choice model. We find that even moderate levels of anticipated inflation generate meaningful welfare costs by distorting life-cycle housing and consumption decisions. When house prices are held fixed, a one-percentage-point increase in inflation reduces welfare by about 0.08 percent of lifetime consumption. The welfare cost is substantially larger for college-educated and higher-income households, who are more likely to be homeowners. In general equilibrium, inflation lowers housing demand and reduces real house prices, partially mitigating the tightening of borrowing constraints. As a result, the welfare cost of inflation falls as the housing supply elasticity declines and is roughly a third as large with a fixed supply of housing. We find that the Volcker disinflation increased steady state welfare by nearly 0.05 %, and that households in cities with low housing supply elasticity saw roughly half the welfare gains of cities with the most elastic supply.

This paper examines changes in the distribution of home values since 1930 using US Census data. Inequality in housing prices and in rents fell from 1930 to 1970 driven by improvements in basic structural characteristics like indoor plumbing. Homeownership also rose during this period. Inequality has risen since 1970 though structural characteristics play a more limited role. Instead geographic decompositions show that inequality between neighborhoods has consistently been two to three times larger than inequality between cities, showing that much of housing inequality reflects differences in hyper-local housing values within the same labor market. We also plan to measure housing appreciation at the neighborhood level since 1940 to understand how housing appreciation has varied across neighborhoods with differing housing values, incomes, and other characteristics.

Changes in mortgage rates impact homeowners’ consumption patterns. Individ- ual small landlords also own a significant share of US rentals, and mortgage rate changes may influence rent-setting behavior. Do small landlords pass lower borrow- ing costs on to their tenants? To isolate this channel, we track landlord-property pairs over time and use rate reductions from refinancing and automatic ARM re- sets. Landlords pass through between 10% and 30% of savings, reducing asking and closing rents on new listings. This reveals an overlooked mechanism through which easing monetary policy reaches renters, but also shows that resets to lower rates mainly increase landlords’ profit margins.

This paper revisits the macro-housing VAR framework of Fratantoni and Schuh (2003) using updated data to include the housing boom-bust around the Great Financial Crisis. Replicating the original model and extending the sample through 2019, we find that the main qualitative dynamics are preserved, although responses exhibit reduced persistence consistent with changes in macroeconomic volatility. We then update the model specification to incorporate subsequent improvements in VAR macro modeling. The new system better captures short-run business cycle dynamics, yielding impulse responses that are less persistent, more tightly estimated, and more economically interpretable, even under bootstrap inference. We expand the VAR model to include housing and financial variables motivated by recent research developments, which likely offers even more insight from regional heterogeneity in a heterogeneous-agent VAR (HAVAR). Homeownership and rental prices have little impact on the macro dynamics but lending conditions play a meaningful role in housing market outcomes. The paper concludes with estimation of a structural VAR model to help clarify the underlying transmission mechanisms in the VAR.

The rapid diffusion of artificial intelligence (AI) is reshaping production, innovation, and growth, but its implications for the household cost of living remain poorly understood. This paper studies how large-scale AI deployment affects consumer expenses through the electricity sector. We develop a macroeconomic framework in which AI data centers generate a surge in electricity demand, requiring costly expansion of generation, transmission, and distribution capacity. Because electric utilities operate under regulated cost-recovery pricing, these infrastructure investments are passed through to end users as higher electricity rates. As a result, households face rising utility bills even when AI improves productivity and lowers the prices of goods and services. Embedding this mechanism in a New Keynesian DSGE model, we show that AI-driven energy inflation raises headline CPI, induces tighter monetary policy, and reduces real household purchasing power.

This paper develops a pseudo-representative agent New Keynesian (PRANK) model of consumption's response to a household energy expenditure shock. A calibrated version of our PRANK model, which allows labor income volatility to be higher during recessions, predicts that non-energy consumption expenditures respond more negatively to energy expenditure shocks that hit during a recession versus an expansion. Estimates of a nonlinear model using pre-pandemic U.S. data show that (i) an increase in household energy expenditures reduces non-energy consumption, (ii) the response to a household energy expenditures shock is stronger when it impacts the economy during a recession rather than an expansion, and (iii) asymmetry over the business cycle is driven primarily by purchases of vehicles and furnishings.

It is commonly argued that emerging technologies are critical to national security due to extensive military and civilian applications. How does this shape the orientation of high growth, early-stage firms? This paper proposes two new measures to empirically measure this relationship using a unique dataset of venture-backed firms incorporated between 2018 and 2025. First, a text based measure to identify early-stage firms developing technologies deemed by governments to be critical to their national security. Second, a measure of “defense-exposure” that captures the degree of textual alignment between solicitations for R&D through the SBIR/STTR program by the US government, particularly the Department of Defense, and descriptions of high-growth, early- stage, firms. Through the measure, I document three preliminary descriptive patterns. First, defense exposure is nearly exclusively concentrated among firms developing critical technologies. Second, most defense exposure is low-intensity, consistent with broad dual-use relevance rather than defense specialization. Third, among firms developing critical technologies, exposure to defense demand is increasing relative to exposure to civilian government demand.

This paper studies whether import liberalization changes firms’ technological direction rela- tive to the global frontier. Using China’s WTO accession, the paper maps product-level tariff reductions into IPC technology classes and combines this exposure with Chinese firm-patent records and text-based similarity to frontier patents. Technology classes exposed to larger tar- iff cuts show short-horizon increases in utility-model local and best-match similarity, but little centroid movement. Invention-side outcomes do not show the same short-run gain; at longer horizons, granted-invention best-match similarity declines. Rank estimates point to relative repositioning among incumbent utility-model patentees, while a firm-IPC risk-set analysis finds little broad entry into new technology classes but changes in entrants’ frontier positions con- ditional on entry. The results suggest that import liberalization changes where firms innovate, with limited broad movement toward the global frontier.

Mass layoffs are typically viewed as detrimental to regional economies. Previous works focused on negative economic consequences such as increased unemployment, declining wages, and reduced consumer spending. However, mass layoffs may act as bridges between otherwise disconnected innovation hubs by facilitating the flow of information and ideas as displaced workers move to new firms. Despite these potential, positive spillover effects have received little attention which could offer different perspectives to understanding mass layoffs. This study fills this gap and examines the short- and long-term effects of mass layoffs on innovation of firms receiving displaced workers. I use novel micro-level data that merge LinkedIn employment histories with patent records. This enables detailed tracking of individual worker mobility, firm characteristics, and patenting activity before and after layoffs. Descriptive statistics show that two-thirds of displaced workers move to small firms (< 50 employees), while the rest join larger firms. To estimate the impact of this labor reallocation on innovation, I employ a difference-in-differences design, comparing firms that absorbed displaced workers to similar firms that received no such workers in unaffected metropolitan statistical areas. The results show that small firms absorbing displaced workers experience a short-term increase in innovation outcomes. Over the long term, these innovative gains remain positive up to seven years. In contrast, large firms show no sustained impact on innovation. I further analyze channels and identify three mechanisms: (1) knowledge transfer from displaced workers, (2) formation of new teams through their networks, and (3) increased diversity in human capital in receiving firms. These channels appear to boost the innovative capacity of host organizations despite the overall disruptive nature of mass layoffs. These findings highlight the direction for policy interventions such as targeted support or hiring incentives for small firms to transform labor market disruptions into opportunities for innovation growth.

Green technologies that are equally clean can differ substantially in their cost structures and welfare implications, yet little is known about whether market forces direct innovation toward the socially optimal technological path when multiple green options coexist. This paper studies innovation path choices in China’s electric vehicle (EV) battery industry, asking whether market driven innovation aligns with the social planner’s preferred path and how industrial policies can correct potential distortions. The paper contributes by quantifying distortions in firms’ technology path choices between two emerging clean technologies and by evaluating how alternative policy instruments shape the direction of green innovation. I focus on two dominant battery technologies, Lithium-Iron-Phosphate (LFP) and Nickel-Cobalt-Manganese (NCM), which differ sharply in innovation difficulty and production costs. I develop and estimate a finite-horizon dynamic structural model in which battery suppliers choose how much to innovate along each technological path by improving energy-density frontiers. They then bargain over battery prices with downstream EV manufacturers, who compete in prices for consumers choosing among differentiated vehicles. The framework compares firms' profit-maximizing innovation decisions to those of a social planner, who maximizes total surplus, encompassing consumer surplus, producer surplus, and environmental benefits. The model is estimated using a comprehensive panel dataset covering China’s passenger vehicle market from 2017 to 2023, combining vehicle-level sales, detailed vehicle characteristics, supplier relationships, battery input prices, and policy variables. I estimate consumer demand using a nested-logit model, recover battery marginal costs through a bilateral bargaining framework, and identify innovation cost parameters for leading battery suppliers. Counterfactual analyses reveal substantial distortions in innovation direction. Relative to the social optimum, the market undertakes only one-fourth as much innovation in LFP, a technology with high sunk innovation costs but low marginal production costs, and twice as much in NCM, which has the opposite cost structure. This divergence is driven primarily by vertical separation and downstream competition, not uninternalized spillovers or environmental benefits. Policy simulations indicate that R&D subsidies are more effective than performance-based consumer subsidies in aligning innovation with the socially preferred path, highlighting industrial policy’s role in shaping both the level and direction of green innovation.

There is abundant empirical evidence linking a lack of health insurance to reduced healthcare access and poorer health outcomes in the United States. While the Affordable Care Act (ACA) aimed to close this gap for the broader population, the Medicare program has historically served a similar role for those aging into eligibility. This project analyzes the impact of both Medicare eligibility and the ACA on insurance status, inpatient treatment, and mortality using data from the Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS). To mitigate selection bias, the sample is restricted to patients admitted through the emergency department with "non-deferrable" conditions, building on the regression discontinuity (RD) design of Card, Dobkin, and Maestas (2009). The impact of Medicare is estimated using an RD design at the age-65 threshold, while the ACA's impact is evaluated using a difference-in-differences (DD) framework comparing 55–64-year-olds to a control group of 65–74-year-olds around the 2014 implementation. Changes in disparities are further evaluated using triple-difference analyses comparing disadvantaged cohorts to baseline groups. The results indicate that both policies successfully generated absolute health benefits: turning 65 and the post-2014 ACA landscape both resulted in significant overall reductions in mortality. However, the policies diverged in their impact on inequality. Upon reaching Medicare eligibility at age 65, disadvantaged groups experienced a distinct reduction in mortality disparities compared to the baseline population. In contrast, while the ACA benefited all patients generally, it yielded no evidence of a similar reduction in health disparities.

This paper examines the effect of the Affordable Care Act’s (ACA) Medicaid expansion on job security among low-income adults in the United States. While a growing literature studies the labor market consequences of public health insurance, relatively little is known about how Medicaid expansion affected job stability and mobility for individuals unlikely to have employer-sponsored insurance (ESI). Focusing on marginalized, low-income populations unconditional on health status, this study extends prior work that emphasizes job lock and pre-existing conditions. Using longitudinal data from the Panel Study of Income Dynamics (PSID) spanning 2009–2019, I implement a difference-in-differences research design with individual fixed effects, exploiting cross-state variation in Medicaid expansion while excluding early and late expanding states. Job security is measured using indicators of employment stability, including tenure-based outcomes and labor mobility measures. Linear probability models are used for binary outcomes, while count-based specifications capture changes in the duration of employment. This paper contributes to the literature by shifting the focus from health-condition-driven labor supply responses to job security among low-income individuals for whom employer-sponsored insurance and traditional job lock mechanisms are less salient. By leveraging rich historical panel data, this study improves target population identification relative to cross-sectional analyses, strengthening internal validity. The findings provide new evidence on the labor market implications of Medicaid expansion for economically vulnerable workers and inform ongoing policy debates on the non-health effects of public health insurance expansions.

Growing demand for home care services, driven by an aging population and a shift towards home and community-based care, has worsened worker shortages exacerbated by low wages. To address this, some states have introduced wage pass-through policies aimed at increasing compensation for direct care workers like home health aides. However, lacking systematic data on home care providers and their workers, the impact of these policies remain unclear. For instance, it is unknown whether home care providers increased worker compensation following these policies, particularly given that states have limited resources to enforce compliance. Moreover, to the extent that states set minimum wages but do not fully fund the increments with additional Medicaid reimbursement, home care providers may decrease supply of aide services to minimize costs, potentially leading to adverse effects on aide employment. Using Medicare Cost Report data spanning 2004 to 2019 on approximately 5,000 free-standing home health agencies across 36 states (with 6 treatment states), we use a stacked difference-in-differences approach to analyze the effects of wage pass-through policies on home health aide compensation and service supply. Our findings show that wage pass-through policies led to a 15% increase in wages ($4,051 [SE=1,968]) and benefits ($755 [SE=269]) per aide per year in the three years following implementation. We observe no significant changes in agencies' employment of aides or supply of aide services. We also find no evidence that the share of directly employed (versus contract) aides increased, suggesting that these policies had minimal effects on aide labor shortages. Finally, we do not observe significant impacts of wage pass-through policies on agencies’ decisions to enter or exit the home care market. Together, these results suggest that wage pass-through policies have increased aide compensation with minimal unintended consequences.

The Census Bureau is amid an agency-wide effort to innovate the enterprise approach to the data lifecycle with improvements in data collection, processing, imputation, and the use of administrative records. This period of transformation has created a prime opportunity to begin implementing the modernization of the Survey of Income and Program Participation (SIPP) that is part of SIPP’s continuous improvement program. SIPP, one of Census’ flagship surveys, is a nationally representative survey that measures monthly changes in employment, income, household dynamics, and government program participation. In this paper, we discuss the challenges that SIPP faces in the current survey operations landscape, from declining response rates to rising collection costs, and the opportunities for modernization. Our priorities in the new SIPP design are to reduce respondent burden, move to multimode data collection, and increase the use of administrative records. With these priorities in mind, we outline the roadmap for SIPP’s modernization and the steps The Census Bureau is taking to improve upon the SIPP program.

Since approximately 2020, rising housing costs in the U.S. have increased the salience of housing affordability for policymakers, researchers, and the public. One increasingly popular indicator of housing affordability is the share of the population facing a ‘housing cost burden,’ i.e. paying over 30% of their income for their rent or mortgage. The Survey of Income and Program Participation (SIPP) includes detailed data on income, earnings, housing, and utility costs over a moderately long time period (2013 through 2023) alongside measures of program participation and household well-being. The availability of these data provides a compelling opportunity to estimate and analyze the extent of housing cost burdens and how they have changed over time. In this paper, I present a new set of SIPP summary estimates of national and subnational housing cost burden prevalence. I summarize these estimates by demographic characteristics, economic characteristics (e.g. earnings, income, employment), household well-being characteristics (e.g. food security and reported non-payment of rent or mortgage), and housing tenure. These estimates provide new context for discussions of housing affordability, especially regarding variation in rent burden dynamics across the income distribution. SIPP data also allow me to investigate the relationship between the social safety net and self-reported difficulties paying for housing. Preliminary findings provide some evidence of limited change over time in the prevalence of housing cost burdens, but meaningful variation in the time trends across the income distribution as more high income households became cost burdened over time. Similar shifts also occur across demographic categories and measures of well-being. Beyond these new population statistics, I present models predicting either individual rent burden or the likelihood of non-payment of rent or mortgage conditioning primarily on economic variables while controlling for other demographic factors. I also fit models of state-level variation in rent burden trends for states with sufficient sample size, to showcase geographic differences in these trends over time. Taken together, my results provide a novel snapshot of the evolution of housing affordability in recent years and showcase variability in the regional and demographic context of these estimates.

Traditional measures such as poverty, income, and net worth do not tell the full story of how long a household could sustain itself if earnings stopped. This paper develops a liquidity measurement framework using the 2017–2023 Survey of Income and Program Participation to quantify short-run economic preparedness. The framework centers on interpretable metrics: months of income replacement (liquid assets expressed as months of household income), fixed-dollar emergency savings thresholds (e.g., $500, $1,000, $2,000), and obligation-based coverage measures, such as the number of months of rent or mortgage payments a household could cover from liquid resources. These measures allow liquidity to be evaluated both relative to income and relative to recurring commitments. SIPP’s detailed wealth and asset information, combined with income, program participation, and demographics in the same microdata, supports consistent household-level indicators and subgroup comparisons over time. The framework addresses questions such as: What share of households would exhaust liquid resources within one or three months of income loss? How common are limited buffers among higher-income households? And how does liquidity coverage of housing costs vary across income bands and demographic groups? Together, these indicators provide a transparent and reproducible way to measure short-run economic preparedness using public-use SIPP data in a way that can inform SIPP Modernization.

The Survey of Income and Program Participation (SIPP) is a longitudinal nationally representative survey with a particular focus on income, employment, and government program participation. For many years hot deck has been the default method of imputing missing data in the SIPP. Hot deck is an attractive imputation method due to its simplicity in implementation across different variable types. Due to this simplicity, it is a procedure that end-users are able to understand, leading to increased trust of the SIPP. However, hot deck is not without its limitations. These limitations include the need for stratifier variables to be categorical and the potential for small donor cells. Such problems are even more pronounced in the current environment where survey non-response are increasing. As part of the greater SIPP Modernization effort, we investigate the extent to which advanced imputations methods enhance the quality and accuracy of for imputed item-level nonresponse, relative to hot deck. Using data from the 2022 SIPP, we compare hot deck against four model-based imputation methods: non-iterative Linear and Logistic regressions, Multiple Imputation by Chained Equations (MICE), and Extreme Gradient Boosted Trees (XGBoost). These methods were applied across four sets of SIPP variables varying in type (i.e. continuous, categorical, and binary), and in non-response rates. We randomly split the observed, non-hot decked, ground-truth observations into two components: 80% as the training sample and 20% as the test sample. The training sample was used to develop the imputation models, which were then applied to and evaluated on the test sample. The models are evaluated using various metrics according to the type of outcome variable: accuracy, sensitivity, specificity, and F1 score estimates for categorical variables; and mean error, correlation, and root mean squared error estimates for continuous variables. The processing time for these models were also assessed to understand the implications from a production standpoint. Our results indicate that MICE and XGBoost significantly improved the quality of the imputations and reduced processing time relative to hot deck imputation. We discuss the implications of our work for the SIPP, the SIPP Modernization efforts, and the applications for other surveys.

I study who gains and who loses from exclusive contracts in the video streaming market, where studios grant streaming services exclusive rights to distribute their content. Streaming services can use these contracts to differentiate their content offerings and soften competition, while studios may leverage them to negotiate higher license fees from streaming services. I develop and estimate a structural model that incorporates bargaining between streaming services and studios, streaming services setting subscription prices, and consumer demand for subscriptions and titles. I find that streaming services that lack in-house content (like Hulu) gain from exclusive contracts, while those with extensive in-house content (like Netflix) see minimal or even negative effects. For studios, exclusive contracts benefit small studios with weak bargaining power but harm large ones with strong bargaining power. While exclusive contracts harm consumers due to reduced title distribution and higher subscription prices, they may benefit consumers in the long run by stimulating content production and streaming service entry.

I study the welfare implications of exclusive dealing in U.S. retail. Using a novel dataset, I document widespread use of contracts that restrict local entry by rival stores. Stores with exclusives face fewer competitors and higher prices, suggesting anticompetitive effects, yet many grocers in under-served neighborhoods use them, implying potential entry benefits in lowdemand markets. I estimate a structural model of household store choice and a landlord–retailer entry game to capture equilibrium pricing and entry. Exclusive dealing benefits landlords and large retailers. A ban raises welfare for some households but increases the number of households living in food deserts

This paper studies the competitive effects of vertical integration and foreclosure by BYD, the leading Chinese manufacturer in the successive oligopoly markets of power batteries and electric vehicles (EVs). We find that integration generates pro-competitive effects by eliminating double marginalization, while foreclosure produces anti-competitive effects by raising rivals' costs. The magnitude of foreclosure's impact is significantly moderated by several factors, and most critically by learning-by-doing. With learning-by-doing, foreclosure creates a total welfare loss of RMB 1.12 billion; without it, the loss shrinks to RMB 0.46 billion. We also find that foreclosure reduces BYD's profits; this is reversed if foreclosure targets only the closest EV substitutes. Our results highlight the fact that dynamic learning and market conditions are vital for assessing vertical strategies in high-tech industries.

In 2010, Live Nation—the largest promoter in the live music industry—merged with Ticketmaster, the leading primary ticketing agency. Following the merger, numerous complaints emerged alleging that independent venues were pressured to contract with Ticketmaster to access Live Nation-promoted concerts. We develop a methodology to quantify the extent of foreclosure on non-Ticketmaster venues without requiring data on stakeholder profit allocations. Using a comprehensive dataset on artists, concerts, and venues, we estimate a model of promoters’ venue choices together with an event’s revenue-generating process, allowing for restricted choice sets. We find that foreclosure accounts for roughly 8% of the merger’s combined impact on Live Nation’s shift toward Ticketmaster-contracted venues, with the remainder reflecting a genuine post-merger preference for Ticketmaster venues. Foreclosure is far from uniform: it concentrates among events featuring top artists—a scarce input for which venues have no ready substitute—where Live Nation’s ability to foreclose is greatest and where it accounts for about a quarter of the post-merger shift. These findings reveal a novel channel for anti-competitive harm from vertical integration, whereby the affected parties— independent venues—do not compete directly with Live Nation’s promotion services or Ticketmaster’s ticketing operations.

This paper investigates how crude oil prices and their volatility influence exchange rate dynamics in six African economies—Algeria, Nigeria, Sierra Leone, South Africa, Uganda, and Zambia—using both linear and nonlinear time-series models. We estimate symmetric autoregressive distributed lag (ARDL) and nonlinear ARDL (NARDL) models to assess whether oil-related variables affect real effective exchange rates (REER) symmetrically or asymmetrically in the short and long run. Our empirical strategy proceeds in two stages. We first estimate ARDL models under the assumption that changes in oil prices and volatility have symmetric effects on exchange rates, identifying their separate short- and long-run influences. We then relax this assumption by decomposing oil price and volatility changes into positive and negative components within the NARDL framework. This allows formal testing for asymmetries using Wald statistics and examines whether exchange rates respond differently to increases versus decreases in oil prices and volatility. The results yield three key findings. First, crude oil prices and volatility have stronger effects on exchange rates in the short run than in the long run across most countries. Second, introducing asymmetry significantly improves explanatory power: the NARDL models reveal more cointegrating relationships and uncover additional significant effects not captured under symmetry. Crude oil prices show pronounced short-run asymmetric effects, while volatility exhibits more persistent, long-run asymmetric influences. Third, the magnitude and direction of impacts differ systematically between oil exporters (Algeria, Nigeria) and oil importers (Sierra Leone, South Africa, Uganda, Zambia), reflecting differences in trade structures, commodity export profiles, and macroeconomic adjustment mechanisms. These findings have important policy implications. Since oil market shocks primarily affect African exchange rates through short-run, asymmetric channels, policymakers should employ timely stabilization tools—such as reserve management, hedging strategies, and flexible exchange-rate interventions—to mitigate immediate volatility. The persistence of volatility effects also underscores the need for long-term economic diversification and reforms that strengthen macroeconomic resilience. Overall, the results demonstrate that accounting for asymmetry is essential for understanding how global oil market fluctuations are transmitted to exchange rates in African economies and for designing effective policy responses.

This paper examines the impact of increases in sovereign borrowing on firms' ability to access international capital markets during different credit market conditions in the financial center (U.S.) and in the periphery. Conditions in the financial center are captured by the U.S. broker-dealer leverage, as a proxy for booms and busts in global liquidity. Conditions in the periphery are captured by different levels of sovereign credit risk and sovereign debt. Using a structural VAR approach, we identify shocks to government issuance by exploiting the fact that current macroeconomic conditions do not respond to issuance of longer-term debt. This is because longer-term financing is not used immediately as it typically finances long-term investment projects such as infrastructure. We find that, in emerging countries, government issuance crowds out firms' access to international capital markets during periods of low global liquidity and high sovereign risk. By contrast, in advanced countries, government issuance increases, i.e. crowds in, firm issuance during periods of high sovereign debt and low sovereign risk. These results imply that, depending on market conditions, government borrowing has the ability to crowd in or crowd out firm borrowing.

Official Development Assistance (ODA) to developing countries is highly volatile, with coefficient of variation three times higher than domestic revenues. This paper examines the macroeconomic effects of aid volatility using a panel VAR with external instruments for eleven aid-dependent countries over 1995–2019. We instrument contemporaneous aid shocks with lagged donor commitments, exploiting the temporal structure of aid delivery for identification. A one percentage point unexpected aid inflow generates 3.122 percent real appreciation, 0.491 percentage point GDP contraction, and 2.616 percentage point deflation on impact. Effects persist for approximately two years. Counter-intuitively, extremely aid-dependent countries (> 9.5% of GDP) experience substantially smaller effects than moderately dependent countries (5–9.5%), with appreciation of only 1.224 versus 3.786 percent. This adaptation effect suggests that high-dependence countries have developed institutional mechanisms to manage volatility through decades of experience. These finding indicate that aid delivery mechanisms matter for macroeconomic stability. Multi-year commitments and better donor coordination could reduce volatility costs. Moderately aid-dependent countries may benefit most from institutional capacity building, as they face substantial exposure without the adaptive mechanisms of extremely high-dependence countries.

Rising temperatures and their health consequences have become one of the most prominent global concerns in recent decades. A growing body of research has documented that exposure to extreme heat is linked to increased mortality. Although the effect of rising temperatures on mortality is relatively well established, its impact on morbidity—particularly the associated monetary costs—remains less well understood. Morbidity is not only more prevalent than mortality but also generates substantial healthcare expenditures. A better understanding of the temperature–morbidity relationship is therefore crucial for informing resource allocation, guiding public health strategies, and protecting vulnerable populations. This study contributes to the literature by estimating the causal effect of temperature on emergency department (ED) visits and their associated medical expenditures. By matching high-resolution weather data with comprehensive medical claims data from 2000 to 2023 in Taiwan at fine spatial and temporal scales, we estimate a high-dimensional fixed effects model that accounts for nonlinearity, location-specific time-invariant unobservables, and seasonality in ED utilization. Consistent with existing studies, our findings indicate that rising temperatures significantly increase ED visits. On extremely hot days with temperatures exceeding 30°C, ED visit rates rise by approximately 2.5 per 100,000 persons relative to mild days with temperatures between 22–24°C. Compared with the average ED visit rate of 79.1 per 100,000 persons, this represents an increase of about 3%. Using detailed visit-level medical billing data, we further estimate that medical expenditures increase by NT$24,412 (approximately US$813) on days with temperatures above 30°C compared to days with temperatures between 22–24°C, corresponding to a 19% increase relative to the average daily medical expenditures of NT$125,833 (US$4,194). Moreover, our heterogeneity analysis by income quintile shows that the temperature effect is largest among the bottom income quintile, indicating that high temperatures disproportionately affect lower-income populations. Taken together, these results contribute to the growing body of research linking climate change to healthcare burdens and underscore important implications for strengthening emergency response capacity and promoting environmental justice in a warming climate.

This paper examines whether litigation-based environmental enforcement reduces corporate pollution and whether its effectiveness depends on institutional independence. We study China’s procuratorate-led environmental public interest litigation (EPIL) reform using firm-level pollution data from the National Tax Survey Database. We find that EPIL significantly reduces firms’ pollution-emission intensity on average, but its effect is limited for major taxpayers and high-revenue firms under localized procuratorial governance. A subsequent procuratorial de-localization reform, which reduced local government influence over procuratorates, substantially strengthens EPIL’s pollution-abatement effect among these locally important firms. Mechanism evidence shows that de-localization improves the judicial handling of environmental disputes, strengthens administrative enforcement, raises penalty intensity, and encourages pollution-control investment and green innovation. These findings suggest that litigation-based environmental governance can induce corporate pollution abatement, but its effectiveness depends critically on whether litigation initiators are insulated from local government interference.

Do firms internalize expected liability costs when choosing contractual partners? This paper studies how joint and several liability rules shape market structure by altering firms’ contracting decisions. I examine the hazardous waste industry, where waste generators contract with disposal facilities and may be held jointly liable under the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA, or “Superfund”) for environmental damages caused by their disposal partners. Economic theory predicts that when liability is shared between contracting partners, generators have incentives to contract with “clean” disposal firms, but that these incentives weaken when expected joint liability diminishes. To test this prediction, I exploit a natural experiment created by the Supreme Court’s 2009 Burlington Northern & Santa Fe Railway Co. v. United States decision, which resolved a long-standing circuit split over “arranger liability” and sharply narrowed the scope of joint liability in some federal circuits, but not others. Using comprehensive administrative data from the EPA’s RCRA Biennial Reports from 2003–2017—covering over nine million hazardous waste transactions—I construct market shares for disposal firms, defined as the share of hazardous waste contracts originating from a given industry–state pair. I measure firm-level environmental risk using a Bayesian learning framework in which generators update beliefs about a disposal firm’s accident risk based on its history of regulatory violations. I implement a triple-differences research design that compares changes in market shares of dirty versus clean disposal firms before and after the ruling, across affected and unaffected circuits. The results show that weakening joint liability led to a substantial reallocation of contracts toward higher-risk disposal firms. In the preferred specification, the relative market share of dirty disposal firms increased by 28.7 percent in treated markets after Burlington Northern, with the largest gains accruing to the dirtiest firms. These findings provide causal evidence that tort reform affects market outcomes through partner choice, rather than solely through firms’ own precautionary decisions. Broadly, the results highlight how liability regimes shape incentives for environmental quality in supply chains and suggest that weakening joint liability can increase the likelihood of environmental harm by shifting activity toward riskier firms.

This article examines how public climate information is absorbed within the trading day and how that absorption reshapes capital flows. It exploits the institutional design of CDP’s annual score release, which occurs through a centralized, embargoed window that synchronizes disclosure timing across firms. This structure mitigates confounding from staggered announcements and generates quasi-experimental variation in the arrival of climate-related information. Using high-frequency data on U.S. large-cap equities, the article aligns intraday abnormal net flows and liquidity measures to the disclosure day. Trading is partitioned into short intervals (1-minute and 20-minute blocks) to trace where within the session climate information is impounded, moving beyond daily returns to trading-clock resolution. Identification relies on two complementary placebo tests. A random-score design permutes firms’ CDP scores while preserving the true disclosure day, maintaining timing but breaking the link between firm characteristics and flows. A random-date design assigns firms’ true scores to pseudo-event dates, preserving information content while severing the synchronized timing channel. Together, these tests isolate the joint role of credible information and coordinated arrival. Three findings emerge. First, capital reallocation is rapid but concentrated late in the day. Abnormal buying pressure associated with higher climate scores is negligible at the open, near zero through midday, and becomes economically and statistically significant in the final trading block before the close. This pattern is consistent with institutional execution strategies that target the closing auction to minimize tracking error. Second, the intraday profile is robust to liquidity and volatility controls but disappears under both placebo designs, indicating that synchronized disclosure is essential. Third, effects show little persistence in multi-day windows, consistent with semi-strong efficiency. The article contributes new high-frequency evidence to the climate-finance literature. By linking embargoed CDP release timing to minute-level order flow and liquidity, it shows that climate information is primarily impounded within the trading day, especially near the close, rather than through prolonged post-event drift that daily studies may miss.

Are emerging markets' central banks credibly committed to inflation targeting? This paper applies the "Optimal Policy Perturbation" (OPP) framework of Barnichon and Mesters (2023) to evaluate monetary policy credibility across emerging economies. Combining impulse responses from a Bayesian VAR with professional forecasts from the Economist Intelligence Unit, I exploit the dual interpretation of the OPP statistic. First, under the assumption that observed policy rates are locally optimal given policymakers' information sets, I adopt a revealed-preference approach and recover time-varying implicit policy weights by minimizing deviations of the OPP statistic from zero. Second, under fixed benchmark weights, I use the OPP statistic to assess whether policy rates could have been adjusted to further reduce the policymaker's loss, and to quantify the direction and magnitude of such adjustments. The evidence indicates that inflation-target regimes in emerging markets have become increasingly credible since the early 2000s, with stronger anchoring of inflation expectations and reduced inflation persistence. While central banks face sharper trade-offs during global crises, monetary policy behavior typically reverts toward inflation stabilization, highlighting the flexibility and resilience of modern inflation-target frameworks.

This paper examines the impact of political advertising on household consumption behavior in the United States. Using a regression discontinuity design that exploits variation in advertisement exposure across media market borders, we show that exposure to political advertisements significantly affects household consumption. Weekly county-level nondurable spending increases by 0.24% for every additional one standard deviation of positive economic-themed advertisements. The effect is heterogeneous by county-level demographics, with a stronger consumption response from counties with a higher proportion of older and low-income residents. We also provide evidence on the underlying mechanisms, showing that these effects operate through changes in household sentiment and the economic information conveyed in the advertisements.

Food Engel curves for U.S. households have been drifting upward for the past twenty years, in contrast to their prior tendency to drift downward. Prior interpretations of Engel curve drift as reflecting unmeasured real income gains imply that this reversal means that real incomes have been growing much more slowly in recent years than in the latter part of the twentieth century. Evidence from Mexican and UK data suggest that the reversal is not necessarily confined to the United States only. This raises questions about why consumer behavior has changed and whether CPI bias has changed dramatically.

This paper analyzes employment and structural transformation in 21 Sub-Saharan African economies (1990-2018) using the Economic Transformation Database. A Divisia decomposition of employment-to-population ratios shows that while agriculture’s role is declining, most labor is shifting into services rather than manufacturing. Productivity gains are concentrated in agriculture and capital-intensive sectors, whereas job creation has been strongest in low-productivity services and construction. Econometric results highlight heterogeneous effects of trade, FDI, and education across sectors. We conclude that Africa’s transformation remains incomplete and fragile, requiring active industrial policy and demand-led strategies to align job creation with productivity growth.

We estimate the effect of basketball expenditures on the probability of making the NCAA men’s basketball tournament. In contrast to prior studies that focus on total athletic expenditures, we examine the relationship between program success and more granular spending categories within men’s basketball. Using publicly available Financial Reporting System (FRS) data from 2017–2023 (excluding 2020), we analyze spending across sub-categories, focusing on coaching and facilities, to identify whether these investments are associated with NCAA tournament appearances and seeding. We find that an additional $1 million paid in coaching salaries increases the probability of making the NCAA men’s basketball tournament by approximately 10 to 20 percentage points and improves a team’s expected tournament seeding by roughly 1.5 to 2 seeds. Conversely, an additional $1 million in facilities-related debt is associated with a lower probability of tournament participation and worse expected seeding. These results suggest that spending on facilities acts as a substitute for coaching investments and may have a detrimental effect on program success.

This study analyzes the extent to which technical skills and psychological resilience influence annual win rates and global ATP rankings among male professional tennis players. The objective is to move beyond performance assessments that rely almost exclusively on technical attributes. Using player-level panel data from 2020 to 2024, I merge ATP official ratings—Serve, Return, and Under Pressure—with match-level statistics from public databases to form a consistent dataset. To capture resilience more precisely, I construct a new measure, New Under-Pressure Performance, incorporating comeback win rate as an observable indicator of a player’s ability to recover under adverse conditions. The empirical analysis applies pooled OLS and fixed-effects panel regressions to identify the main determinants of competitive outcomes. The results show that serve and return skills remain the principal technical drivers of win rates and ATP rankings, consistent with earlier work in the sports economics literature. Psychological resilience also has a statistically significant and economically relevant association with win rates. The newly proposed under-pressure measure performs better than the ATP’s original index in explaining variation in win rates, suggesting that the capacity to respond effectively in difficult match situations captures a distinct, measurable dimension of player performance. The study contributes to the literature on behavioral and psychological determinants of athletic outcomes by demonstrating how resilience can be inferred from observed match data without relying on survey-based or subjective assessments. The findings underscore the value of integrating psychological indicators into player evaluation, coaching decisions, and performance modeling. The empirical framework can be extended to other racket sports, such as table tennis and badminton, to examine how technical and psychological factors interact across different competitive environments.

August 29, 2016, Colin Kaepernick took a knee during the San Francisco 49ers’ first preseason game. The protest quickly became league wide, and spread to player protest instances in the WNBA, the WNSL, NBA, college football and other professional sports. The NFL player protests ended after NFL owners and the Players Association on an anthem policy in 2018. This study estimates how players’ anthem protests effected NFL game day attendance. This study makes three contributions to the emerging literature on the effects of athlete protests, and NFL players’ anthem protests in particular. First, it introduces a variable to measure the effects of the BLM movement on attendance, comparing it with the effect of the share of Trump voters. Of the two previous studies only (Sperling and Vandegrift, 2022), only the latter measured the effects of the share Trump voters, and it didn’t include a measure for BLM. Because fans supporting the BLM (and Kaepernick) potentially increased ticket sales, omitting their effect may overstate the negative effect of player protests on attendance. Second, this study uses tickets sales as the dependent variable, enabling for the estimation of the effect of the players’ anthem protests on gate revenue. Conversely the two previous studies Brown and Sheridan (2020) and Sperling and Vandegrift (2022) used TV viewership as the dependent variable. In distinction to TV viewers fans opposed to the protests have the option of protesting in-person. Third, our measure of tickets sold is team-specific, whereas the two previous studies relied on TV ratings aggregated to large regional markets, which may yield less precise estimates because it can mask heterogeneity across of teams, local economic conditions, and demographic composition. With these data we use the clubs fixed-effects identification design, which controls for unobserved team heterogeneity to estimate the effect of player protests on attendance. The results show that fans protesting the player protests by boycotting the game reduced game day attendance. Further we estimate how this reduction was offset somewhat by MAGA sympathetic fans some choosing to protest in-person, and BLM sympathetic fans attending to support the player protests.

Healthcare spending in the United States remains persistently high relative to other high-income countries, with national health expenditures reaching $4.9 trillion in 2023, raising concerns about cost containment. A central question in U.S. healthcare policy is how to reduce spending without compromising access or quality. One potentially important margin for cost containment is the reallocation of primary care delivery towards midlevel providers, particularly nurse practitioners (NPs) and physician assistants (PAs), who are reimbursed at lower rates and may exhibit different practice patterns than physicians. Whether this reallocation affects aggregate healthcare spending depends critically on the regulatory environment governing midlevel practice, which varies substantially across states and over time. This paper examines the causal relationship between state-level scope-of-practice (SOP) reforms for midlevel providers and healthcare spending, leveraging a staggered adoption framework over the period 2012–2021. The empirical strategy exploits variation in the timing of midlevel practitioner SOP expansion and uses a staggered difference-in-differences approach to examine labor-market dynamics in the healthcare industry. Preliminary analysis draws from publicly available data, including workforce supply measures from the Health Resources and Services Administration Area Health Resources File and spending from the CMS Geographic Variation Public Use File which provides standardized Medicare fee-for-service per capita spending at the state level, separately for Part A and Part B. Preliminary evidence illustrates how midlevel SOP reforms affect NP and PA workforce supply and, through that channel, healthcare spending. Ongoing work extends the analysis using commercial claims data covering a large share of the employer-sponsored insurance population, allowing for a decomposition of spending into price and quantity components. This research contributes to the literature on occupational licensing and healthcare labor markets by clarifying how regulatory changes affect both provider supply and healthcare spending.

Shortages of healthcare professionals, particularly in mental health, remain an ongoing policy dilemma. Licensing requirements may be one potential factor contributing to this shortage. In 2022, Illinois eliminated the exam requirement for master’s level social workers (MSWs). Using data on the number of MSWs from several states, we estimate the effect of this policy change on the number of MSWs in Illinois using the synthetic control method. We estimate that the elimination of the exam requirement increased the number of MSWs in Illinois by more than 2,600 licensees-- a 50% increase. We also provide evidence that our estimates are both economically and statistically significant.

Physical therapists are critical evaluators for movement, pain management, and recovery of patients. Yet, the tasks and duties a physical therapist may perform varies across states and have not been recorded, even by the leading professional associations. This study contacted each state regulatory board to develop the first ever dataset on physical therapy scope of practice differences across states. Variations of physical therapy practice include the ability to refer patients, file for reimbursement, diagnose patients, develop treatment plans, and the length of time for which they can provide recurring care for a patient. As these differences in job tasks affect practitioner autonomy, we develop a physical therapy health provider shortage area map to identify if areas with reduced autonomy are more likely to have lapses in available patient care.

This study examines how changes in chiropractic scope-of-practice (SOP) regulations shape health labor market outcomes. I construct a state-level index of SOP restrictiveness for 2011 and 2024, and use this index to define a proxy treatment indicator identifying states with a reduction in regulatory restrictiveness. Because no natural policy experiment divides states into treated and control groups, I employ a pseudo-difference-in-differences strategy that leverages binary measure of regulatory change. States that had a reduction in SOP restrictiveness are classified as treated. I then estimate the effect of these regulatory changes on chiropractor establishments per 100,000 population and employees per 100,000 population at the county level with standard errors clustered at the county level. This approach contributes methodologically by demonstrating how to construct credible comparison groups when formal treatment assignment is absent and the results offer insights into how regulations may affect health-sector labor market. The results show an increase in chiropractor establishments per 100,000 population and employees per 100,000 population for states with a reduction in SOP restrictiveness.

In this study, we examine the impact of Medicaid dental coverage expansions on both the utilization of dental services and the supply of dental providers. While most states require Medicaid to cover emergency dental services for adults, coverage for other dental services varies across states. We first compile state-level data on adult dental coverage from 2012 to 2024. Exploiting the gradual roll out of Medicaid dental coverage expansions, we estimate their impact on dental service use. We begin by using BRFSS data, and our variable of interest is the timing of respondents’ most recent visit to a dentist. Using a Difference-in-Differences model, we estimate the effects of Medicaid dental coverage expansions on the use of dental care. We find that Medicaid dental expansion increased dental visits among low-income individuals by 4%, suggesting that the policy improved dental health among Medicaid recipients. Using data from the Area Health Resource Files, we then examine the effects of Medicaid dental coverage expansions on the number of dentists, especially those in private practice. We find that Medicaid dental expansion had little impact on the number of dentists within a state. The estimates are precise enough to rule out effects exceeding 10%. We make contributions to the literature by bridging the demand and supply sides of dental services.

This paper examines how intangible capital is related to firms' market power in product and labor markets using an unbalanced panel of Japanese listed firms from 1980 to 2024. We estimate firm-level markups and markdowns following the production-function approaches employed by the standard literature and then relate these measures to two types of intangible capital: R&D capital and organizational capital. The empirical results show that organizational capital is positively associated with markups and negatively associated with markdowns, while the role of R&D capital is relatively limited and less robust. These findings suggest that organizational capital may strengthen firms' product-market position while also encouraging rent sharing with workers. The stagnant accumulation of organizational capital may therefore help explain why Japan experienced weak price and wage growth since the 2000s.

Since China’s accession to the WTO, the impact of increased competition from Chinese imports (the “China shock”) on employment and productivity in many developed countries has become a major concern for policy makers. The share of manufacturing workers in the total number of employees has been declining, and Japan is no exception. The paper empirically examines the impact of the increase in imports from China on employment using questionnaire information of the Census of Manufactures and the Economic Census for Business Activity as well as the Trade Statistics of Japan and the National Freight Flows Survey (Logistics Census). The main results are twofold. First, imports of intermediate products from China have a positive impact on employment at Japanese firms. Second, however, imports of capital products from China might have a negative effect on employment growth. Thus, reducing trade barriers in intermediate products, participating in global value chains, and supporting inter- and intra-industry labor mobility for specific workers, regions, and industries that are negatively affected by capital goods are key to employment growth in Japan.

This paper examines how outward foreign direct investment (FDI) is associated with domestic plant dynamics and employment adjustment among Korean manufacturing firms over 2018--2023. We utilize firm-plant panel data for manufacturing firms, constructed by linking the Survey of Business Activities and the Mining and Manufacturing Survey, to compare multinational enterprises (MNEs) with non-MNEs. The results show that although plant birth is also observed among MNEs, the increase in domestic plant death is more pronounced relative to non-MNEs. MNEs also exhibit lower net employment growth and higher excess job reallocation, suggesting that outward FDI is associated with domestic employment adjustment and plant-level restructuring. These patterns are more pronounced among MNEs with overseas plants in emerging economies and among small and medium-sized enterprises (SMEs), and the robustness analyses provide similar evidence. Overall, the findings indicate that Korean outward FDI is associated with domestic plant reorganization.

While prescription-drug monitoring programs (PDMPs) have been found e↵ective in reducing prescriptions and misuse of medical opioids, the evidence on their impacts on other outcomes, particularly those related to child welfare, is mixed. Combining a di↵erence-in-di↵erences design with a heterogeneity-robust estimation approach, we estimate the e↵ects of PDMPs on how, and how much, parents spend time with their children. We find that PDMPs increase time spent on both active and passive childcare, with much of the e↵ect for active care driven by increases in relatively engaging forms of childcare per se, education care, and medical care. We find much larger negative e↵ects for parents with disabilities or mobility issues, and argue that this type of heterogeneity can help reconcile some of the apparently conflicting results from the literature on PDMPs and child outcomes. We further show that the positive e↵ects of PDMPs are amplified in states with legal marijuana laws. Our findings highlight the role of policy to address the needs of sensitive populations and mitigate the potential downsides of substance-related policy interventions.

This paper quantifies how Conditional Pocket Money (CPM)––allowances contingent on study effort––shapes adolescent skill formation. Using a dynamic structural model that separates selection into CPM from its causal effects, I show that eliminating CPM would reduce cognitive test scores by 4.43% (0.122 SD) by late adolescence, with gains accruing to children with both low and high self-regulation. Families who do not use this parenting strategy do so because they cannot find a contract that satisfies the child’s participation constraint within their budget and time constraints. I evaluate two policies targeting these households: an unconditional cash transfer and an after-school program. The cash transfer relaxes budget constraints, enabling CPM adoption and generating a 23.79% increase in cognitive skill levels. The after-school program directly substitutes for parental monitoring, producing larger gains of 27.94% at equivalent cost.

The globalized food system has made trade a crucial factor in achieving food security. The Jameel Index for Food Trade and Vulnerability (Jameel Index) was developed to guide policy and investment decisions that can support sustainable international development and food security through food trade. It combines a range of meta-indicators, including dependency on food and animal feed imports and the concentration risk of trading with only a small number of partners. A major finding was that nations with a high level of food security span a wide range of food import vulnerability whilst nations with poor food security are predominately low in food trade vulnerability due to the lack of adequate domestic production and lack of financial ability to import food. The Jameel Index framework analyzes future Food Trade and Vulnerability through models and scenarios of population and economic growth and climate change up to 2050. An analysis using results from the IFPRI Food Modeling Framework: The IFPRI IMPACT-GLOBE Macro Economic modeling linkage and a new bespoke Bi-Lateral Trade model developed by IFPRI for the Jameel Index found that low and lower middle-income counties will increase their food security but at the cost of increase food trade vulnerability. Climate Change by 2050 will have some significant impacts for certain regions but overall, the impacts on food security and trade vulnerability are minor compared to increase food demand from population and economic growth. This paper will present the Jameel Index projections to 2050 followed by a moderated panel of food trade and development experts to discuss the policy implications of the results and future analyses.

Abstract Statement of the research: Rural finance plays a crucial role in promoting sustainable agricultural development in China. This paper systematically examines the impact of rural finance on agricultural green total factor productivity and its underlying mechanisms. Research model and data: Using provincial panel data from 2007 to 2019, this study employs the meta-frontier undesirable super-efficiency slack-based measure (SBM) model and the Global Malmquist–Luenberger (GML) index to measure agricultural green total factor productivity. Empirical Results: The empirical results show that rural finance has a significant positive effect on agricultural green total factor productivity. Furthermore, rural finance indirectly promotes enhances agricultural green total factor productivity by influencing the scale of agricultural operations, industrial integration, and technological innovation. The promotion effect is found to be stronger in the eastern and central regions as well as in non-grain-producing areas. Contribution made by the paper: Based on the empirical findings, this study offers the following policy implications. First, the government should strengthen support for rural finance, with particular attention to economically underdeveloped western regions. Expanding subsidies for rural financial products and encouraging institutional innovation—such as cross-regional integration of financial resources and greater participation of private capital—can help alleviate financing constraints and improve the efficiency of rural financial services. Second, the rural financial service market should be improved by establishing robust credit evaluation systems and financing information platforms to reduce information asymmetry. Financial products should be tailored to local conditions and designed to support the adoption of green and low-carbon agricultural technologies. Third, financing channels for rural households should be broadened by promoting products such as rural housing mortgages, forest rights mortgage loans, and loans secured by land contract and management rights, while strengthening risk management mechanisms to enhance the resilience of rural financial services.

Based on Industry–University Cooperation (IUC) performance data covering more than 1,000 universities and research institutions in Japan published by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), this study constructs a balanced panel of 139 universities with complete observations from 2018 to 2022. To estimate cost efficiency, we employ multiple Stochastic Frontier Analysis (SFA) specifications, including Battese and Coelli (1995) and Greene’s true fixed effects model (2005). Lagged explanatory variables are incorporated to mitigate potential reverse causality, and a series of robustness checks are conducted. Beyond estimating efficiency scores, the study further analyzes the influencing factors of cost efficiency. The empirical results show that the average cost efficiency ranges from 0.70 to 0.82, indicating substantial room for improvement. National and public universities consistently exhibit higher efficiency, while private and medical universities tend to perform less efficiently. Efficiency notably improved in 2020. Among internal and external factors, the number of specialized IUC practitioners significantly improves efficiency, whereas researcher numbers shows weak or inconsistent effects. Organizational structures and invention ownership systems do not exhibit statistically significant impacts. These findings suggest that opportunities for direct interaction between practitioners and researchers may contribute more to improving efficiency than formal organizational arrangements. As a key policy tool for economic development, IUC receives substantial public investment. This study provides updated empirical evidence from Japan and addresses an important gap in the relevant field.

This paper studies how venture capital (VC) reshapes startups’ innovation choices by insuring against default risk, and explores the macroeconomic implications of this mechanism. I develop a dynamic general equilibrium model in which startups choose betweenconservative(low-risk, low-return)andaggressive(high-risk, high-return)inno- vation while endogenously selecting their financing mode. Debt financing features state uncontingent repayments and exposes startups to default. By contrast, VC financing is a state-contingent dynamic contract with one-sided limited commitment from startup. Evidence from a new dataset linking VC deals, balance sheets, and patents supports the model predictions: VC-backed startups begin with higher leverage, show greater post-financing profit dispersion, and generate more high-quality patents. Calibrated to financing and innovation data, eliminating VC reduces the aggregate output by 5 percent and the mass of large firms by 11 percent, despite only 0.2 percent of startups ever receiving VC.

This paper examines how ideological orientation shapes citizens’ evaluations of and commitment to democracy in Latin America. Using data from the Latinobarómetro survey (1995–2020), we construct a country-year dataset covering more than two decades and multiple countries. For each country and year, we estimate median values of key variables to capture consistent, cross-national patterns in democratic attitudes. Preliminary results based on panel regressions using country-level medians reveal a clear nonlinear relationship between ideology and democratic attitudes. Satisfaction with democracy follows a U-shaped curve: individuals at both ideological extremes report greater satisfaction than moderates, though right-leaning respondents are consistently more satisfied than those on the left. Support for democracy also displays a U-shaped pattern, with higher support among the ideological extremes and lower support near the center. These findings suggest that political moderation does not necessarily correspond to stronger democratic attitudes. A working hypothesis emerging from these results is that ideological commitment, on either end of the spectrum, reinforces both satisfaction with democratic performance and attachment to democracy as a system, while centrists tend to express more conditional or ambivalent support.

This paper asks what is the welfare value of Medicare Part D, and how much of that value reflects financial protection versus a mechanical transfer from beneficiaries to the public budget? Despite extensive evidence on Part D’s effects on prescription-drug spending and utilization, less is known about how these changes translate into a money-metric measure of consumer welfare that can be compared directly to program costs. The paper’s contribution is to provide a transparent welfare evaluation of Part D using an optimization-based framework that decomposes the program’s value into (i) a transfer component driven by reduced out-of-pocket payments and (ii) a “pure insurance” component capturing the welfare gains from lowering out-of-pocket risk in high-need states. We implement this framework using the Medical Expenditure Panel Survey (MEPS) covering 2000–2012 for the United States. Outcomes include annual total prescription-drug spending, out-of-pocket prescription-drug payments, and payer-specific drug spending (public versus private). Insurance variables are used to identify Medicare beneficiaries and Part D coverage. To estimate causal impacts, we use a difference-in-differences design comparing changes before versus after Part D implementation between groups differentially exposed to the policy, controlling for demographic characteristics and fixed effects and incorporating MEPS survey weights. The welfare calculation maps reduced-form estimates into an interpretable willingness-to-pay for Part D and reports a decomposition into transfer and insurance components using a standard consumption-based proxy for resources and conventional risk-aversion values. In our baseline estimates, Part D lowers out-of-pocket prescription-drug spending by $1,200 per beneficiary-year and increases public prescription-drug spending by $2,100, while total prescription-drug spending rises by $500, implying substantial crowd-out of private payments. The resulting willingness-to-pay for Part D is $1,000 per beneficiary-year, with 45% attributed to pure insurance value beyond the mechanical transfer. Comparing willingness-to-pay to the implied fiscal cost yields a net welfare effect of −$1,100 per beneficiary-year. These results provide a unified welfare interpretation of Part D’s spending and payer-mix effects and inform policy discussions about the tradeoff between program costs, crowd-out, and beneficiaries’ valuation of prescription-drug insurance.

This paper examines how insurance benefit design at the Medicaid-CHIP eligibility boundary may affect mental health treatment patterns among children. With one in six U.S. children experiencing mental health conditions yet only half receiving treatment, understanding how public insurance transitions influence healthcare utilization is an important policy question. I employ a sharp regression discontinuity design using the National Survey of Children's Health (NSCH) from 2016 to 2023, with an analytical sample of 55,584 children with mental health conditions. The running variable is constructed as the difference between household Federal Poverty Level and state-specific Medicaid income eligibility limits derived from Kaiser Family Foundation data, using data-driven optimal bandwidth selection and local linear polynomial estimation. The main finding reveals that children just above the Medicaid eligibility threshold show a 20-30 percentage point higher probability of mental health medication use. This effect is concentrated among children with active anxiety conditions, in states without continuous eligibility provisions, and during 2016-2019. Notably, no discontinuities are found for treatment receipt or access barriers, suggesting possible cost-driven substitution toward medication—families facing higher out-of-pocket costs under CHIP may substitute toward relatively less expensive medication over therapy. The findings suggest potential policy implications for benefit harmonization between Medicaid and CHIP regarding mental health cost-sharing differentials and provide supportive evidence for continuous eligibility policies now federally mandated as of 2024.

Financial incentives designed to encourage smoking cessation during pregnancy have emerged as a promising intervention, yet their potential to reduce health inequalities remains insufficiently examined. While maternal smoking substantially elevates risks of miscarriage, fetal death, prematurity, and low birthweight, smoking rates among pregnant women across Europe remain concerning, ranging from 10% in the UK to 22% in France, with pronounced socioeconomic disparities. Although previous research confirms that financial incentives improve abstinence rates and birth outcomes, their differential effects across income groups have received limited attention. The theoretical framework suggests competing mechanisms. The diminishing marginal utility implies that financial incentives should be more effective among low-income women. However, this population also experiences higher addiction levels that could potentially diminish incentive effectiveness. Understanding these dynamics is essential for determining whether financial incentives can genuinely serve as equity-enhancing policy tools. This paper examines how financial incentives for smoking cessation during pregnancy affect socioeconomic disparities in both maternal smoking behavior and birth outcomes, while accounting for differences in addiction levels. Drawing on data from the Financial Incentives for Smoking Cessation in Pregnancy (FISCP) randomized controlled trial conducted in France, we build upon initial findings demonstrating that financial incentives more than doubled continuous abstinence rates and reduced low birthweight incidence by six percentage points. This analysis reveals complex patterns in how pregnant women across income levels respond to these incentives. Financial incentives increase smoking abstinence across all socioeconomic groups. High-income women achieve higher continuous abstinence rates, while low-income women who do not achieve complete cessation nonetheless significantly reduce their cigarette consumption. The effectiveness of financial incentives diminishes among more heavily addicted women, yet addiction differences do not fully account for the differential impacts observed across income groups. Most importantly, the health benefits concentrate among low-income women, with financial incentives reducing low birthweight risk exclusively in this population. This pro-poor effect occurs because low-income women face elevated baseline tobacco exposure, rendering even partial cessation clinically significant for birth outcomes. These findings demonstrate that financial incentives can function as equity-enhancing policy instruments, effectively reducing maternal smoking and adverse birth outcomes among the most vulnerable populations.

We empirically examine the impact of the Electronic Case File (ECF) system on judicial efficiency and reform using a Differences-in-Differences (DiD) approach. Taiwan introduced the ECF system in criminal cases in 2015. Existing literature has not examined whether efficiency gaps between case types vary before and after policy implementation, limiting causal assessment. We use the Kinmen District Court of Taiwan as a case study, and analyze efficiency variation between treatment and control groups, considering equipment upgrades and policy stability. The results show that the ECF system significantly improves judicial efficiency and remains robust during the COVID-19 pandemic. We also note potential indirect benefits for the business environment. The main contribution is to apply a DiD framework to examine the impact of the ECF system on judicial reform and judicial efficiency.

This study examines the dynamic relationships among violent crime, property crime, and crime-related search activity using a panel vector autoregression (VAR) framework applied to monthly data from Chicago, Los Angeles, and San Antonio, 2011–2021. Crime-related search behavior is measured through city-specific Google Trends data for both Google Web Search and Google News Search, operationalized as metro-level proxies for public attention rather than as indicators of a shared national information environment. Results indicate substantial within-category persistence, with autoregressive effects dominating system dynamics for both crime series. We find no statistically robust evidence of predictive content for crime outcomes in either search measure. Both crime categories, however, significantly precede increases in search activity, a pattern consistent with public attention responding to crime conditions rather than shaping them. Impulse response and variance decomposition analyses confirm that cross-category spillovers are negligible and that crime dynamics are largely self-contained within categories. An important measurement qualification bears directly on interpretation: the two search series differ substantially in empirical stability. Google Web Search exhibits broadly distributed, month-to-month variation across all three cities. Google News Search, by contrast, is near-zero in the overwhelming majority of months in San Antonio, with only occasional positive values, suggesting low search volume, threshold effects, or platform-specific censoring in that market. This cross-city heterogeneity in proxy quality implies that null findings for the News series should not be interpreted as evidence against attention-based mechanisms; they may instead reflect limited measurable variation. Web Search appears to provide a more stable and informationally consistent signal in this dataset. Persistence is identified as the dominant structural feature of short-term crime dynamics. Implications for crime forecasting and for the appropriate use of platform-specific search proxies are discussed.

Why did honor-based femicide disappear among Mizrahi Jews in Israel while persisting among Palestinian citizens, under the same state and legal system, across the same historical period? Drawing on an original dataset of femicide cases in Israel since 1948, assembled from archival court records, administrative documents, and press coverage, this paper argues that the divergence cannot be explained by cultural assimilation. The disappearance of honor killing among Mizrahi Jews was produced by the Israeli state's melting pot policy, which stigmatized Mizrahi culture as backward, Oriental, and Arab. The practice became dishonorable not through the moral recognition that women's lives are equal in worth, but because it was associated with Arabness at the moment when Arabness was being made into a stigma incompatible with Jewish national identity. The paper tests and complicates Kwame Anthony Appiah's argument that honor killing ends when it becomes dishonorable. It confirms his prediction while revealing a mechanism he does not theorize: coercive stigmatization by a dominant state, in which a practice becomes dishonorable through ethnic othering rather than community-internal reform. The same discursive act that ended honor killing among Mizrahi Jews entrenched it among Palestinian citizens by racializing the practice as an Arab cultural problem, thereby justifying state non-intervention and producing a near-fivefold gap in the clearance rates for Arab versus Jewish women's murders that persists to the present. The paper contributes to scholarship on honor norms, the racialization of gender-based violence, and femicide in Israel and the broader region.

In recent years, the role of primary commodity price shocks on business cycles in emerging market economies has attracted increasing attention, reflecting their high dependence on commodity production and exports. By contrast, the relationship between commodity prices, external borrowings, and domestic policies is relatively unexplored. This paper examines the effects of commodity price fluctuations on external borrowings in emerging market economies and investigates optimal policies, focusing on the role of financial frictions. A small-open-economy model with commodity goods (exportable goods) and importable goods sectors is developed, incorporating debt dollarization and a nonlinear collateral constraint on external borrowing. The model is calibrated to commodity-exporting economies and numerical simulations are conducted to analyze the responses of aggregate variables to external commodity price shocks. Further, the study examines optimal policies for financial stabilization. The simulation results demonstrate that a negative shock on the commodity price can lower the collateral prices and raise the probability of a financial crisis or a sudden stop of external borrowing. This is due to the fact that the economy's external borrowing level and the consumption ratio between commodity and importable goods may deviate from the socially optimal level, particularly when the collateral constraint is binding. The implied optimal policies involve subsidies to correct these inefficiencies. The findings could provide policymakers with useful guidance in designing domestic policies.

We investigate the informational content and forecasting potential of disaggregated GDP revisions. Using dynamic linear regression models, we show that revisions to consumption growth predict future consumption growth, improving forecast accuracy by 2.6%. As a consequence, consumption growth revisions also have explanatory power for GDP growth, improving forecasting accuracy by 4.4% relative to an autoregressive benchmark model. This is despite the fact that GDP growth revisions typically do little to improve forecast performance. The information contained in consumption growth revisions likely goes beyond information reflected in benchmark forecasting surveys.

This paper analyzes the properties of forecast error in the Survey of Professional Forecasters in relation to economic policy uncertainty. By examining the quarterly forecast error of key macroeconomic variables alongside the economic policy uncertainty measure from 1985 to 2020, I demonstrate that most real activity variables show significant negative responses to economic policy uncertainty. However, forecasts related to these variables are substantially sluggish, resulting in long-lasting forecast errors. In other words, the findings indicate that slow response of forecasters is attributed to forecast error in the SPF through both static and dynamic frameworks.

Concern about climate change is increasingly cited as a reason to have fewer children, on the logic that a smaller population emits less. We argue that this reasoning overlooks a political-economy feedback: when environmental preferences are transmitted across generations, green fertility--the fertility of environmentally minded households--helps determine the composition of the future electorate and hence the stringency of climate policy. We develop a political-economy model in which individuals differ in their support for climate policy, preferences are imperfectly transmitted across generations, and the environmentalist share of the population determines policy stringency. Green fertility affects cumulative emissions through two opposing channels: a positive scale effect, since a larger population emits more, and a negative composition effect, since a larger environmentalist share supports stronger policy. We first study a one-country model of emissions regulation and show that, near a political threshold, lower environmentalist fertility can increase cumulative emissions by weakening future climate policy. We then extend the model to directed technical change, where the environmentalist share in a frontier country determines subsidies for green research and innovations diffuse globally with a lag. In this setting, higher green fertility in a frontier country keeps the scale effect localized while the effect of policy-induced innovation spreads globally. Simulations calibrated to the U.S. electricity sector show that the effects are nonlinear: small changes in green fertility have modest effects away from innovation thresholds, but can sharply reduce cumulative emissions when they shift the economy from a brown innovation path to a green innovation path. The results suggest that environmentally motivated fertility decisions may have long-run political and technological consequences that complicate simple individual carbon-accounting calculations.

This article is an empirical analysis of how firms strategically use lobbying by joining business associations to influence energy regulations and market structure. We use a cross-country dataset by the World Bank Enterprise Surveys containing firm-level measures of lobbying activity, regulatory exposure and firm-characteristics. We find that lobbying has a positive effect on the probability of facing energy performance standards. Energy performance standards and lobbying reduce competition facing a firm and increase monopoly status – thus providing empirical causal evidence in the Regulatory Capture literature. Stigler (1971) defined Regulatory Capture as a phenomenon where policies are sometimes influenced (captured) by the industry they are expected to regulate, thus reducing welfare. In the political-economy literature, firms lobby against regulation stringency (Stigler, 1971; Peltzman, 1976; Bo, 2006) to reduce their cost (Fisman, 2001, Faccio, 2004). However, the Industrial Organization (IO) literature argues that may firms lobby for stricter regulations to erect barriers to entry. Salop and Scheffman’s (1983, 1986) theoretical work shows firms use regulatory instruments as “cost-raising strategies” for rivals. Empirical research supporting this mechanism are correlational and/or specific to a region or industry (Maloney and McCormick, 1982; Ryan, 2012) and lack firm-level causal estimates. We contribute to the existing literature by bridging this gap. Our identification strategy entails: (1) establishing a positive, statistically significant causal effect of firm-level lobbying through business associations on regulatory burden using Instrumental Variables; (2) temporal evidence at the sector-level to show past regulations do not affect current group lobbying activity, to address reverse causality; and (3), showing past regulations affect current market-structure outcomes. A large set of firm-level characteristics are included to minimize omitted variable bias. We use “presence of an external advisory board” as an instrument for joining business associations that lobby. We find that joining a business association leads to a 64.2 % increase in the probability of facing an energy performance standard. Energy performance standards lead to a 3.4% increase in the probability of becoming a monopoly. Our findings are relevant for WEAI colleagues interested in IO and the political-economy of energy regulations.

Despite the ubiquity of overlapping land management regulations, the cost-effective design of payment for ecosystem services (PES) in the presence of overlapping command-and-control regulation remains poorly understood. We study this question using Brazil's Reflorestar program, a large-scale PES for forest restoration that operates under the Forest Code, which requires landowners to maintain native vegetation on a share of their land. We analyze enrollment patterns, estimate willingness to accept (WTA) using plausibly exogenous spatial variation in offered payments, and estimate additionality using staggered difference-in-differences. We find positive reforestation impacts with magnitudes and costs comparable to avoided deforestation initiatives. We also examine strategies to increase cost-effectiveness. Our results suggest that, in the presence of overlapping command-and-control regulation, the cost-effectiveness of restoration PES is substantially improved by targeting private rural properties that are already compliant with the Forest Code or are in municipalities with greater enforcement capacity.

This paper examines how domestic stock exchange competition shapes international stock return commonality. Using a novel hand-collected dataset on stock exchange competition and ownership covering 78 countries from 1960 to 2023, we show that greater domestic exchange competition is associated with significantly lower return synchronization with both the S&P 500 and the World ex-U.S. index. Economically, an additional domestic exchange reduces commonality with the S&P 500 by approximately 2.3 percentage points and with the World ex-U.S. index by 7–9 percentage points, controlling for GDP per capita, exchange size, and global crisis periods. These effects are strongest in high-income countries, consistent with stronger domestic arbitrage and more efficient local price discovery. In contrast, partial U.S. ownership of domestic exchanges increases return commonality with global benchmarks, particularly the S&P 500, highlighting the role of exchange ownership and institutional linkages in international financial integration. Our findings demonstrate that market infrastructure—specifically exchange competition and ownership—constitutes a key but previously overlooked determinant of global return co-movement.

Theoretical models predict that the choice of invoice currency in inter-firm trade is determined by negotiation, reflecting the relative bargaining power between an exporter and an importer. On the other hand, in intra-firm trade, the choice of invoice currency is determined by the centralization of exchange rate risk management at the group's head office. To disentangle two different mechanisms in currency invoicing, we conducted a questionnaire survey on Japanese unlisted manufacturers. Our unique questionnaire items include information on the relative size of trade partners compared to Japanese SMEs and on the invoice currency used for export or import with trade partners. By using destination-trade partner datasets, we found the following results: first, in the intra-firm trade, a firm with more equity share of own foreign subsidiary tends to choose trade partner’s currency and US dollar so that the head office assumes exchange rate risk, which is consistent with the currency risk sharing hypothesis; second, in the inter-firm trades, the larger the relative size of trade partner compared to Japanese SMEs more tend to choose foreign currencies, which is consistent with the bargaining hypothesis in currency invoicing.

This paper provides some preliminary documentation of a new database of the prices of services for Western Europe for the period 1870 to 1914. The data are drawn from a series of travel guides published by the German publishing house Baedeker over this period. We use the data to revisit some of the questions about deviations from the law of one price and real exchange rate fluctuationsfirst explored by Crucini, Telmer and Zachariadis (2005). We report some preliminary tests of deviations from the law of one price for four expenditure categories for four large European cities, and also constrict a simple cost of tourism index and compare it to historical estimates of consumer price indexes.

This study examines whether the effect of ESG label adoption on fund-level sus tainability performance varies with geographic revenueexposure. Usingastaggered difference-in-differences design withmonthlypaneldataon19,892open-endfunds fromMorningstarbetween2020and2025. WeestimatetheimpactofESGself-labeling onESGriskscores andtest for heterogeneity across thirteen global regions. Two findings emerge. First, ESG label adoption reduces fund-level ESG risk by 0.05–0.07 points on average, suggesting that labels correspond to measurable im provements in portfolio sustainability. Second, this effect varies substantially by ge ography. Funds with greater pre-treatment exposure to Africa experience attenu ated ESG improvements, implying that highly Africa-exposed funds show minimal ESGriskreductionuponlabeling. SimilarpatternsemergeforEuropeEmergingand Japan although they are less significant, while UK-exposed funds exhibit amplified improvements. Thefindingssupportaninstitutionalinterpretation: ESGlabelinginducesstronger responses where regulatory oversight is stringent and weaker responses where dis closure frameworks are underdeveloped. This pattern raises concerns about unin tended consequences of sustainable finance frameworks. If ESG rating methodolo gies systematically favor investment in jurisdictions with strong disclosure infras tructure, they may redirect capital away from emerging markets where sustainable investment is most needed

We document that foreign economic policy uncertainty (EPUF ) has significant incremental predictive power for excess U.S. stock returns in the presence of domestic EPU, both in aggregate and for returns of portfolios constructed on firm characteristics, for 6 to 12-months-ahead horizons. We find that EPUF shocks primarily transmit to equity prices through cash flow news rather than the discount rate news channel. We examine whether responses of select macro-financial variables to an adverse EPUF shock are consistent with this transmission mechanism. Corporate investment outlays, payouts, and aggregate credit demand decline in response to such a shock.

Employment has shifted away from routine tasks amenable to automation and into non-routine work. This study proposes that a more granular task approach advances our understanding of how changing task demand affects demand for specific skills. Applying factor analysis to the near universe of relevant O*NET task attributes, the analysis estimates a comprehensive set of granular tasks. Merging the task data with U.S. individual level data from the 1980 decennial census and the 2019 ACS, this study tracks the reallocation of labor shares across tasks overall, and by education and gender. The results identify specific tasks driving employment polarization, present a new channel for the de-skilling of non-college educated male employment, and uncover novel gender differences associated with the shift away from routine work and into non-routine work. First, with respect to high-wage employment polarization, employment flows into decision-making tasks explain approximately 40% of high-wage employment growth and flows into technology and information related tasks explain an additional 50%. While employment flows into interpersonal tasks contribute little to employment growth in the upper portion of the wage-skill distribution. Second, it is common to associate low and middle skill jobs with routine or non-routine physical work. However, the granular tasks reveal that abstract work is not only prevalent, but also intensive at all skill levels. The change in the incidence of non-routine abstract work for non-college educated individuals has been understudied in existing literature, yet this is an important channel impacting human capital, skill requirements, and wages. Finally, employment shifts for women display a different pattern relative to those for men. Both college and non-college educated women began the period with sizeable employment shares concentrated over a few tasks. Over the next 40 years, both groups expanded employment shares into a wide range of non-routine abstract tasks, particularly so for college educated women. While female employment was becoming more broad based across non-routine abstract tasks, both college and non-college men's employment became more concentrated in a few tasks. College educated men shifted away from mid-wage tasks, particularly away from interpersonal tasks, and into higher wage information and technology tasks.

I develop a novel online experiment with 387 subjects on Prolific and simple utility model to examine how workers respond to reported work meaning. Workers value different aspects of work and are willing to give up other incentives for their best work match. Roughly 31% of workers do not value meaning in their work, 27% do not value work pay, 30% value both, and 11% cannot be characterized. Overall, 57% of workers are willing to sacrifice up to 14% of possible pay in their pursuit of meaningful work. For workers who value work meaning, I estimate positive impacts of meaningfulness on the quantity and quality of output, although these effects occur only for a task with prosocial framing. Finally, I validate a light-touch treatment designed to increase worker awareness of the value they place on meaning, again finding effects only for workers who seek meaning in their work. Importantly, raising worker awareness of work meaning has no impact on the types of incentives they seek out. These results offer insights into how such interventions and preferences for work meaning might be leveraged in organizational settings.

Using novel data from the Survey of Household Economics and Decisionmaking (SHED) on self-assessed changes in job amenities among individuals who changed jobs between 2020 and 2024, we estimate the effect of labor market tightness on changes in multiple dimensions of job quality. We find evidence that higher levels of job vacancies per person, as measured at the state level by the Job Openings and Labor Turnover Survey (JOLTS), and at the CBSA level by Lightcast job postings, led to a greater likelihood that workers changed jobs and saw improvements in pay and benefits, interest in the work, better opportunities for advancement, and overall job quality in their new jobs. Specifically, we find that a 10 percent increase in vacancies per thousand residents is associated with a 7 percent increase in the likelihood of changing jobs and an 11 percent increase in the likelihood of moving to an overall better job. We find increases of comparable magnitude for most other attributes we can measure, with few differences by workers' gender, but slightly more by their education. Our results are robust to multiple measures of labor market tightness across data sources, geographic definitions of the labor market, and instrumental variable strategies. Our results suggest that labor market tightness not only improves worker pay, in line with earlier studies, but also several job amenities more broadly. Consequently, the benefits to workers of a tight labor market may be underestimated when based on pay alone.

This paper shows that employed workers in the U.S. have unbiased beliefs about the average wage of outside offers they could potentially receive in the next four months, irrespective of their gender, education, experience, wage, and job search status. Using each worker's belief and actual wage offers as two noisy measures of the worker's productivity, I find that over 80% of the wage variation across employed workers is due to heterogeneity in worker productivity, and there is no significant sorting between worker productivity and the job-specific wage rate.

Generative Artificial Intelligence (AI) promises to expand instrumental rationality in public administration. However, its true impact on public policy decision-making is still highly debated. In this systematic review, conducted according to PRISMA 2020 guidelines, we examine whether AI tools reduce or exacerbate cognitive biases in public service contexts. We synthesize 27 empirical studies from Web of Science and Scopus up to January 2026. Our review reveals a paradoxical "jagged technological frontier": AI greatly improves analytical performance in structured, probabilistic tasks. Yet, it often hinders critical thinking and synthesis in open-ended policy design. To address this paradox and reduce the risk of 'algorithmic bureaucracy,' we propose the Symbiotic Governance Framework. This normative model presents eight main principles for human-AI collaboration, including task-appropriate modality, mandatory human involvement, and contestability. Our findings show that preventing algorithmic bureaucracy requires more than just adopting new technology. Rigorous, task-calibrated human-in-the-loop oversight is essential.

This paper identifies a significant correlation between demographic, socioeconomic, and environmental factors and academic success in Brazil’s National High School Exam (ENEM). High writing performance is positively associated with being female, identifying as white, and having access to a computer at home. Conversely, lower grades are strongly associated with being male, being over 18, and attending schools with high failure rates. A critical finding of this research is the measurable impact of climate on cognitive performance. Adverse meteorological parameters—specifically minimum temperature and wind speed—act as physical constraints that hinder the concentration and decision-making required for proficient writing. Regional disparities further emphasize this, with students in Espírito Santo performing best and those in Rio de Janeiro performing worst. We conclude that "rationality" and cognitive skill development are not purely internal processes but are heavily mediated by the external environment. Structural inequalities and climate risks function as systemic barriers that limit a student's ability to make rational decisions and demonstrate proficiency. Therefore, achieving educational equity requires addressing not only pedagogical quality but also the environmental and structural constraints that disproportionately affect human capital development.

This study examines whether adults in Brazil's Federal District — a population with a strong concentration of public-sector workers and politically engaged citizens — revise their policy preferences when confronted with counter-ideological evidence. Using a survey experiment (N = 217), we cluster respondents into three ideological profiles (Left, Centre, Right) via K-means on 14 binary items and then model belief revision across eight policy domains with a mixed-effects logistic regression with respondent random intercepts. To address potential circularity between the clustering items and the belief-revision outcomes, we implement a leave-one-out (LOO) clustering procedure as the primary specification: for each outcome item, cluster labels are estimated from the remaining 13 items. LOO cluster labels agree with the full-sample solution in 86.1% of cases, and the headline interaction odds ratios (OR = 0.154 and 0.208 for Left and Centre vs. Right) are virtually unchanged from the baseline (0.174 and 0.184), confirming robustness to the circularity concern. The interaction between ideological congruence and cluster membership is the central result: respondents are substantially less likely to revise positions that are congruent with their ideological cluster. This pattern persists across alternative moderator specifications (vote choice, continuous ideological score) and is inconsistent with a pure demand-characteristics explanation. Implications for evidence-based policymaking are discussed.

What did the COVID-19 pandemic ultimately reveal about the capacity of modern governments to design policy under crisis, govern through prolonged uncertainty, and convert emergency intervention into durable institutional resilience? This paper advances the central finding that COVID functioned as a global governance stress test, exposing a persistent gap between crisis mobilization and long-term institutional resilience. Rather than treating the pandemic as a singular public health or economic shock, the analysis positions COVID as a test of federal capacity, institutional coordination, and policy learning across the lifecycle of crisis, governance, and recovery. The paper consolidates three sequential studies conducted over three years to offer a unified interpretation of the pandemic as a governance phenomenon. Its overarching aim is to define COVID analytically as a stress test of state capacity and to assess whether extraordinary crisis responses translated into sustained governance resilience during recovery. Methodologically, the study employs a retrospective, multi-stage governance analysis that synthesizes three approaches. First, it analyzes the United States government’s 4.1 trillion dollar emergency response using public policy and constitutional frameworks to evaluate crisis decision-making under uncertainty. Second, it examines governance during prolonged disruption, focusing on how political institutions managed extended shutdowns, social externalities, and declining public trust. Third, it introduces and applies a Governance Resilience Index to evaluate post-COVID recovery outcomes across five major United States industries between 2023 and 2025, emphasizing institutional continuity, regulatory adaptability, equity, and governance quality rather than spending levels. Across all three phases, the analysis finds that emergency policymaking was effective at stabilization but weak at institutional transformation. While crisis interventions prevented systemic collapse, they largely failed to generate durable governance capacity for long-term resilience. Recovery outcomes were uneven, revealing a gap between rapid policy mobilization and sustained institutional learning. Taken together, these findings reinforce the conclusion that COVID should be understood as a governance stress test that exposed limits of emergency authority and fragile policy continuity. The paper concludes by proposing a framework for crisis governance that links policy design, institutional capacity, and recovery performance, offering guidance for shocks from climate risk, geopolitical instability, and technological disruption.

The emergence of cryptocurrencies has ushered in a new era of the modern economy. While much of the existing literature focuses on the economic and financial implications of cryptocurrencies, few studies have examined the potential risks they pose to modern public management, particularly in the realm of governance. This research seeks to address this gap by exploring the issue through five dimensions: economy, equity and inclusion, efficiency and effectiveness, accountability, and public–private partnership. Accordingly, we propose a pentagonal conceptual framework to illustrate the multifaceted relationship between cryptocurrency and public governance. First, from an economic perspective, cryptocurrencies embody the principle of decentralization, operating independently of national governments and central banks. Their relatively low cost of issuance and circulation aligns with core economic principles, reflecting market-driven efficiency. Second, in terms of equity and inclusion, cryptocurrencies may enhance financial accessibility. Traditional fiat money systems, regulated by central banks, often favor upper social classes. By contrast, cryptocurrencies offer a more impartial and borderless medium of exchange, potentially reducing barriers linked to class, nationality, or geography. Third, regarding efficiency and effectiveness, cryptocurrencies facilitate peer-to-peer transactions that transcend national borders, thereby promoting global commerce and trade. Operating in digital form and supported by decentralized online networks, cryptocurrencies exemplify the potential for streamlined, borderless financial interaction. However, the accountability dimension presents significant challenges. Cryptocurrencies’ anonymity and lack of centralized oversight make them vulnerable to misuse. They have been criticized as “tax havens” that enable transactions beyond the reach of traditional financial institutions, thereby facilitating tax evasion, money laundering, and other illicit activities. Such risks raise serious concerns for public accountability and regulatory enforcement. Finally, in the realm of public–private partnership, the development of Central Bank Digital Currencies (CBDCs) demonstrates how governments are beginning to collaborate with private entities, such as commercial banks, to integrate digital currencies within sovereign monetary systems. Overall, these five dimensions highlight both the opportunities and governance challenges that cryptocurrencies present. As the proliferation of digital currencies becomes inevitable, modern governments must respond with prudence, innovation, and openness. Adapting regulatory frameworks and governance practices will be essential to ensure accountability and maintain good governance.

It is well documented in the finance literature that stocks identified as underpriced earn significantly higher future returns than those identified as overpriced, a phenomenon known as the mispricing effect. In this paper, we examine the sources driving this effect. We collect all common stocks of U.S. companies listed on the NYSE, AMEX, and NASDAQ exchanges from 1963 to 2024, and employ portfolio analysis and Fama-MacBeth regressions. Relying on the Stambaugh, Yu, and Yuan (2015) mispricing score, our empirical results reveal that the mispricing effect exists primarily among stocks that are far from their 52-week high prices but becomes insignificant among stocks that are near or at their 52-week high prices. This pattern remains robust after adjusting for Fama-French risk factors and controlling for capital gains overhang and idiosyncratic volatility. Our findings suggest that investors consider the 52-week high price as an anchor when they evaluate firm value. When a stock’s price is near or reaches a new 52-week high, it provides a clear value identity and reduces the room for valuation ambiguity, causing the mispricing effect to be attenuated. By contrast, when a stock’s price is far from its 52-week high, the lack of reference value leads to heightened valuation uncertainty. In this high-ambiguity state, psychological biases, such as overconfidence and sentiment, exert greater influence on prices, causing the mispricing effect to be magnified. To validate this anchoring mechanism, we employ the 52-week low as an alternative anchor for robustness tests. Consistent with valuation ambiguity arguments, we find that proximity to the 52-week low also acts as a constraint that dampens the mispricing effect. This confirms that price anchors, whether high or low, reduce valuation uncertainty and the mispricing effect. Overall, this paper contributes to the finance literature by identifying price anchors as regulators of market efficiency. We conclude that mispricing is fueled by the valuation ambiguity associated with drifting far from established anchors.

This paper investigates the relationship between Bitcoin crash risk and investor disagreement, and uses the investor heterogeneity model proposed by Hong and Stein (1999) to explain the asymmetry in the distribution of Bitcoin returns. We find that crash risk, proxied by the negative conditional skewness (NCSKEW) and down-to-up volatility (DUVOL) of 5-minute intraday Bitcoin returns, increases with the level of investor disagreement. However, an increase in Bitcoin crash risk does not lead to a significant change in investor disagreement on the same day; instead, it results in a decline in investor disagreement on the following day. Robustness checks confirm that the findings are robust to changes in control variables and the sampling frequency for high-frequency returns. Furthermore, we explore the moderating roles of market liquidity and investor attention. The crash-magnification effect of investor disagreement is significantly more pronounced in periods of low liquidity and limited investor attention, suggesting that investor opinion disagreement plays a stronger role in the destabilization of the Bitcoin market under such conditions. Furthermore, we find that during the COVID-19 pandemic, as well as in the pre-pandemic and post-pandemic subsample periods, Bitcoin crash risk and investor disagreement are positively correlated.

Analyzing investment decisions is fraught. Theory assumes expectations formed ex ante are about ex post outcomes that are risky. Yet (Knightian) uncertainty sullies real-world outcomes so pervasively that even the state space is rarely known. We examine the casino game roulette, and establish several statistical propositions to formally characterize its play. In parallel to sports economics, casino games offer a natural laboratory in which more aspects of behavior are transparent and readily observable. Unlike many other games, roulette has features reminiscent of investment decision-making. It is a basal, stripped-down environment in which most players form portfolios; risk-return tradeoffs are present; decision possibilities are limited, making analysis easier; Knightian uncertainty about the state space, probabilities, and payoff schedule is absent; and portfolio-choice is cleaner. In short, there is little scope for hand-waving away the mismatch between ex ante and ex post information, because the mapping is fully specified. Even so, a unique dataset of real-world play shows that anomalous portfolio-choice behavior is ubiquitous in roulette. Monte Carlo simulations of a variety of heuristic models of the gambling experience fail to reproduce the anomalies (similarly, Blavatskyy 2024 argues that there are puzzles here, because a variety of theories of choice cannot explain roulette play in a virtual environment). In particular, people avoid portfolios in roulette with mean-variance-skewness characteristics that are superior across a broad range of standard preferences in favor of portfolios with something we dub “vanity diversification”. This is our descriptive label for a widespread pattern of idiosyncratic choices consistent with a preference for seemingly non-instrumental diversification (i.e., unjustified by standard mean-variance-skewness tradeoffs), revealed by players’ actions that forgo expected value (implicitly paying an expected-value premium) to do so. We interpret this as indicative of preference primitives for variety and/or broader scope that enter utility directly. It suggests an explanation for why real-world investors frequently retain assets that do not improve their overall portfolio, or forgo reallocations to make themselves better off.

Assuming the existence of an arbitrage equilibrium, from an empirically observed set of test asset returns and test factors, I derive an optimal implicit underlying test asset returns space transformation for the existence of the equilibrium, generalizing the arbitrage pricing theory (APT) of Ross (1976) and Chamberlain and Rothschild (1983). The empirical tests show that the space measure transformation is economically and statistically significant, with an increase in the portfolio’s Sharpe ratios of 0.25 across known factor models, with industry concentration (HerfBE), and assets under management (AM) as the most relevant economic proxies in the space transformation.

The contributions by artificial intelligence (AI) technologies to the economy can be in general attributed to rises in factor productivity and enhancements in task efficiency in production chains, which could significantly lead to increases in GDP. Nonetheless, AI technologies have the potential to worsen employment and income distribution via job displacement by AI-driven automation especially for lower-skilled workers and income inequality by AI-induced investments on capital and skilled labor. Using vertically-linked general-equilibrium models with goods and factor markets, this paper examines theoretically the effects of developing AI technologies on employment and income inequality, and then verifies the theoretical findings empirically. This paper further considers firm dynamics in the long run induced by adoption in AI technologies in production. AI technologies can raise or lower profits, which leads firms to enter into or exit from the downstream industrial sector in the long run. Thus, income inequality in the short run could be worsened or mitigated in the long run by the firm-entry effect.

This paper provides some of the first aggregate-level evidence on the labor market effects of generative artificial intelligence across U.S. local labor market areas (LMAs) following the release of ChatGPT. Using data from the American Community Survey and occupation-level AI exposure scores developed by OpenAI and Anthropic, I con- struct an LMA-level measure of AI exposure by combining pre-ChatGPT occupational employment shares with exposure scores. Employing a difference-in-differences and event study design, I exploit geographic variation in AI exposure to identify impacts across U.S. local labor markets and individuals, with effects varying across industries and demographic groups. The findings show that AI exposure does not affect local economic activity but significantly reduces the likelihood of labor force participation, employment, and full-time work. These effects are most pronounced among women and non-college-educated individuals. Furthermore, the findings suggest that the negative effects of AI intensify as the technology advances and firms increasingly integrate AI into workplace tasks.

This paper explores the effects of Nepal's legalization and the subsequent roll-out of abortion services nationwide after 2002, on the outcomes of women and children in 2010-11 and 2022-2023. We adopt a variety of difference in difference strategies to estimate the effects: i) a simple district-year level analysis making comparisons between cohorts of mothers and children across districts with different timing of facility access, ii) within-family comparisons of siblings born before and after the year in which first facility opened in the district, and iii) a rural-only analysis utilizing the exact village, month of opening, and imputed month of conception to define treatment most precisely. We detect no effects on women's schooling or employment in the short run, and no change in the sex composition of births. We observe positive effects in the long run, as cohorts born after a facility opened are less likely to have a chronic disease, less likely to have a disability, and more likely to hold a school scholarship, while household consumption and poverty are unchanged. Our longer run results suggest that the benefits of abortion access accrue gradually and may operate through children's health and public schooling support rather than through the immediate composition of births. However, the mixed evidence for investment channels operating closer to birth remains a puzzle.

In a single generation, most of the world’s population has experienced a significant rise in living standards. However, there are twenty-seven nations – 14% of all nations - where economic development measured by per-capita income has stalled resulting large proportions of their populations living in desperate poverty. In a few of these nations, living standards are lower now than a generation ago. Why? Almost all of the "Missing 14%" countries are characterized by serious levels of animosity/conflict, bureaucracy, and/or corruption. As a result, analysis of economic development in these countries is difficult due to the absence of reliable data. Data on even basic economic variables such as real growth, inflation, and employment is either non-existent, distorted for political reasons, or applicable only some areas such as the capital city. The unreliability of data motivates a focus on the evolution of political-economic institutions. Specifically, the characteristics of the "Missing 14%" countries' institutions result from and facilitate animosity, bureaucracy, and corruption. As expected, there is a symbiotic relationship between conflict and corruption but the relationship between conflict and private sector development is more complex. This study of the "Missing 14%" countries ends with recommendations for future research.

Abstract We investigate the effect of off-balance sheet (OBS) intangibles, defined as internally generated knowl-edge and organizational capital, on stock price informativeness. We build on the q-theory framework of Bai et al. (2016) by incorporating both book and OBS intangible assets, allowing us to derive testable predictions about the firm’s information environment. Our analysis shows that OBS intan-gibles provide incremental predictive power beyond the book value of assets for future fundamentals, including cash flows, R&D expenditures, patents, and stock returns. Forecast encompassing tests and principal component analysis indicate that no single proxy fully captures the information content of OBS intangible capital. Consistent with the valuation uncertainty associated with OBS intangibles, firms with higher intangible intensity exhibit significantly lower price informativeness. In contrast, firms with high physical capital intensity exhibit greater price informativeness, reflecting the relative transparency of tangible assets. Overall, our results demonstrate that asset composition, specifically the balance between tangible and intangible capital, significantly affects how effectively stock prices reflect information about future firm fundamentals.

This paper studies whether sustained, high-intensity police deployments generate durable crime deterrence and how such effects evolve over time and space. We examine Costa Rica’s Megaoperativos—large, mobile, multi-agency police operations implemented repeatedly across districts between 2018 and 2022—and ask whether their impacts arise immediately through enforcement or gradually through cumulative deterrence. Using a balanced weekly district-level panel covering the Greater San José metropolitan area, we estimate both instantaneous and cumulative effects of police exposure on reported crime. Our empirical strategy exploits variation in the timing, intensity, and repetition of deployments, allowing us to distinguish short-run mechanical detection effects from longer-run behavioral responses. We additionally assess spatial spillovers to neighboring districts to evaluate whether deterrence diffuses geographically. We find that short-term effects are small or positive, consistent with increased reporting and enforcement activity during operations. In contrast, cumulative exposure to Megaoperativos leads to large, statistically significant, and persistent reductions in crime. These effects grow with repeated deployments and extend beyond treated districts, indicating substantial spatial spillovers rather than displacement. The magnitude and persistence of the estimates suggest that deterrence operates through learning and belief updating rather than one-time enforcement shocks. To interpret these findings, we develop a simple theoretical framework in which perceived apprehension risk evolves with repeated and nearby enforcement, even under imperfect crime measurement. The model rationalizes why deterrence may emerge slowly yet accumulate over time and space. Together, the results contribute new evidence on the dynamic and spatial nature of police deterrence, highlighting the importance of sustained visibility and repetition in crime-control policies rather than isolated, short-lived interventions.

This study investigates whether sustainability disclosure practices reflect substantive environmental, social, and governance (ESG) performance among financial institutions in Taiwan. As sustainability reporting standards and third-party assurance have become central components of regulatory and market-based ESG frameworks, a key underlying assumption is that greater disclosure signals stronger sustainability performance. Whether such practices translate into real ESG outcomes or instead serve primarily symbolic or legitimacy-seeking purposes remains an open empirical question, particularly within the financial sector. Using panel data on 40 Taiwanese financial institutions from 2016 to 2022, this study examines overall TESG scores and their environmental, social, and governance components to assess the real effects of sustainability disclosure. Fixed-effects and correlated random-effects models are employed to control for unobserved institutional heterogeneity. To address potential endogeneity arising from reverse causality and omitted variables, instrumental variable (IV) estimations are implemented using peer disclosure adoption rates as instruments. Descriptive evidence shows that institutions issuing sustainability reports -- especially those subject to mandatory disclosure - exhibit higher average TESG scores than non-reporting institutions, suggesting a positive association between disclosure and ESG performance at first glance. However, once firm size, institutional type, governance characteristics, and workforce composition are explicitly controlled for, the regression results reveal a markedly different pattern. Both mandatory and voluntary sustainability disclosure are negatively associated with overall TESG performance relative to non-reporting institutions, with the negative effect of voluntary disclosure being statistically significant. These findings suggest that higher unconditional ESG scores among reporting institutions are largely driven by scale and organizational capacity rather than the intrinsic impact of sustainability disclosure. Consistent with this interpretation, asset size emerges as the most robust and positive determinant of ESG performance across all model specifications, highlighting the critical role of resources and governance capacity in achieving substantive sustainability outcomes. Overall, the evidence points to a systematic decoupling between sustainability disclosure practices and actual ESG performance in Taiwan’s financial sector. Sustainability reporting appears to function more as a symbolic or legitimacy-oriented mechanism than as a reliable indicator of genuine sustainability achievement, underscoring the limitations of disclosure-based ESG assessments and the need to place greater emphasis on outcome-based evaluation frameworks.

This paper evaluates the effectiveness of the K-ETS in reducing GHG emissions using firm-level data for covered industrial firms over 2008–2024. The analysis focuses not only on whether aggregate emissions declined relative to a counterfactual benchmark, but also on how heterogeneous firm responses generated or offset the realized effectiveness of the scheme. Because comparable emissions data for non-covered firms are limited and K-ETS participation is determined by emissions and energy-use thresholds, this paper constructs a model-implied business-as-usual (BAU) benchmark within the covered-firm population rather than relying on a treated-versus-untreated design. The BAU benchmark is estimated from pre-K-ETS emissions-intensity relationships over 2008–2014. Emissions intensity is defined as emissions per unit of real sales, allowing the model to separate changes in firm activity from changes in carbon efficiency. The empirical specification relates log emissions intensity to real sales, a pre-policy time trend, fuel-specific relative energy-price exposure, fuel-use status, firm fixed effects, and latent response-group heterogeneity. To capture heterogeneous firm responses, the paper applies a regression-cluster framework that groups firms according to their pre-K-ETS emissions-intensity dynamics and allows slope coefficients to differ across latent firm groups. The results show that actual emissions over 2015–2024 were 8.86 percent below the model-implied BAU benchmark. This aggregate effectiveness, however, is not generated by a uniform response across covered firms. Five of the seven estimated clusters emit below their BAU paths, while two clusters exceed their model-implied benchmarks. The cluster-level results further reveal distinct response types, including scale-responsive groups, fuel-price-responsive groups, large fuel-intensive contributors, and BAU-exceeding groups. Standard-industry aggregation shows where effectiveness is concentrated, but it also masks substantial within-industry heterogeneity. These findings suggest that the effectiveness of emissions trading should be assessed not only by aggregate emissions outcomes, but also by the firm groups through which those outcomes are produced. The paper contributes a heterogeneity-aware BAU evaluation framework for the K-ETS and provides evidence that firm-group diagnostics can complement industry-level reporting in designing more targeted and transition-oriented carbon-pricing policy.

Digital platforms commonly operate on a “free for data” exchange, offering zero price access in return for user data monetized through targeted advertising. While this model expands participation regardless of affordability and supports platform profitability, it imposes non price costs including privacy loss, advertising exposure, and behavioral manipulation that are likely to be unevenly distributed across income groups. This paper asks: How do alternative exchange mechanisms in digital markets affect welfare, inequality, and platform viability, and can pricing or subsidy policies improve equity without undermining revenue? We develop a microeconomic model of consumer choice in a digital platform market with heterogeneous users who differ in income, privacy preferences, and sensitivity to advertising. Consumer utility incorporates both monetary prices and non monetary costs from data exposure and advertising, and is used to simulate adoption of free (ad supported) and paid (ad free) tiers. The model is calibrated using two complementary data sources: individual level survey data capturing income, privacy concern, and ad annoyance, which are mapped into preference parameters, and platform level metrics including advertising revenue per user, ad intensity, subscription prices, and user base size drawn from publicly available SEC filings and company reports for major U.S. digital platforms. We simulate three policy scenarios: (i) a zero price ad supported model, (ii) a dual tier model with optional paid access, and (iii) an income based subsidy for privacy preserving paid access. Our findings are expected to show that the zero price model generates regressive welfare effects, and that dual tier pricing likely exacerbates inequality by allocating ad free access to higher income users while concentrating advertising and data extraction on lower income users, who are least able to benefit from targeted marketing. In contrast, income based pricing or subsidies are expected to increase total welfare, reduce inequality, and maintain platform profitability. These anticipated results contribute to the economics of digital markets by demonstrating how differentiated pricing can enhance equity in data for access exchanges without sacrificing efficiency, providing insights for platform design and policy interventions in digital services.

The recent literature on competition in higher education has offered a number of theoretical foundations for the emerging impact of online higher education programming on the quality of education, price discrimination among various student groups, and social welfare. These models envision a market consisting of a combination of online universities, face-to-face public universities, and face-to-face private universities. We propose a distinct stochastic spatial model that focuses on a particular aspect of the higher education industry in the presence of online providers. In particular, the model allows to formally study how the online option impacts the tuition discrimination imposed by an in-state face-to-face university on the in-state and out-of-state students. Some partial preliminary results are presented.

This paper examines the effects of unemployment shocks on stock prices using a proxy structural vector autoregression (proxy-SVAR) framework. While standard economic intuition suggests that stronger labor market conditions should support equity valuations, historical evidence points to a counter-cyclical relationship in which stock returns are often higher during periods of elevated unemployment. Identifying the causal impact of unemployment shocks on financial markets, however, is challenging due to simultaneity and endogeneity between macroeconomic and financial variables. To address this issue, we construct an external instrument for unemployment shocks by exploiting the timing structure of the Survey of Professional Forecasters (SPF) and official unemployment data releases. Because professional forecasters submit their expectations before the realization of current unemployment, the exogenous component of forecast errors provides a valid and strong instrument for unexpected movements in unemployment. We use this instrument within a proxy-SVAR framework to identify unemployment shocks without relying on restrictive recursive timing assumptions that are problematic in models including financial variables. Using quarterly U.S. data from 1970 to 2017, we find that an unanticipated increase in unemployment leads to a statistically significant rise in stock prices over subsequent quarters. At the same time, unemployment shocks generate declines in industrial production, inflation, and the federal funds rate, consistent with an expansionary monetary policy response. These results contrast sharply with estimates from conventional Cholesky-identified VARs, which yield insignificant responses in the stock market, highlighting the importance of external instruments for credible identification. Our findings support the view that monetary policy transmission plays a central role in the positive response of stock prices to adverse labor market shocks. By lowering interest rates and increasing liquidity, policy easing reduces discount rates and raises equity valuations even as real economic activity weakens. The paper contributes to the literature by providing new causal evidence on the unemployment–stock market relationship and by demonstrating the usefulness of survey-based instruments for identifying labor market shocks in macro-financial settings.

I study the macroeconomic consequences of minimum-wage hikes and the optimal monetary-policy response. Motivated by Japanese prefecture-level evidence that minimum-wage increases are significantly associated with a higher price level, lower employment, and fewer firms, I develop a New Keynesian model with firm dynamics and a statutory minimum wage, in which minimum-wage hikes affect firms' marginal costs and thereby create cost-push pressures. Comparative-statics exercises show that macroeconomic responses are larger when the wage-bill share of minimum-wage labor is higher or when a larger mass of firms is located near the operating cutoff. Under optimal policy, across a range of reduced-form central-bank preferences commonly used in quantitative policy analyses, the model suggests that the policy response to these cost pressures tends to be more aggressive than that implied by the benchmark Taylor rule.

Using a quantitative search-and-matching framework, we examine how fiscal policy transmits through labor markets under fundamentally different crisis environments, comparing outcomes during the Great Recession with those of the COVID-19 pandemic. We introduce public health frictions as reduced-form wedges that constrain labor utilization and consumption, thereby generating underutilization and forced savings without relying on explicit epidemiological SIR models. This framework preserves a common labor market structure while allowing shocks to operate either through aggregate demand, as in the Great Recession, or through utilization constraints, as observed during the COVID-19 episode. Calibrated to the U.S. economy, the results show that the large-scale fiscal transfers implemented under the CARES Act and the American Rescue Plan Act (ARPA) stabilized household income and prevented persistent unemployment but also increased reservation wages and reduced search intensity. In a supply-constrained environment, such aggressive fiscal support generates more persistent inflation and lower matching efficiency, producing elevated vacancy rates as firms compete for scarce workers. These results highlight that fiscal interventions that are effective in demand-driven downturns may exacerbate labor market tightness and inflationary pressures when supply is constrained. By holding the labor market structure fixed, the paper isolates how the underlying shock regime fundamentally reshapes fiscal transmission.

Out-of-sample simulations with small-scale time-series macro models that include only the money stock and shocks to potential output---but not supply chains or fiscal policy---explain most of the post-COVID U.S. inflation surge. Two novel findings underlie this result: 1) the data reject the conventional (but untested) output gap restriction, so relaxing it makes time-varying potential output independent and stochastic; 2) the short-run Taylor Rule and long-run Quantity Theory of Money (QTM) equations can be included simultaneously with proper treatment of I(1) and I(0) variables. Without money, the model explains very little of post-COVID U.S. inflation.

This paper presents the details of a game-based learning module for upper division undergraduate Labor Economics classes. The game, A Quest for a Modest Wage, takes students through the lifecycle of an average worker in the U.S. (high school through late career). The game uses simplified Dungeons and Dragons (D&D) mechanics to show students the contrast between choices they can control and outcomes they can’t–such as a bad roll of the die, income status of parents, or recessions. I ask students to reflect post-game on the diverse reality of economic life in the U.S. and the fairness/unfairness of the system I created. After playing the game, students should be able to describe the economic approach to analyzing social issues and the labor market. Additionally, students should be able to propose changes to game mechanics and explain how the outcomes of the game would change, simulating changes to a model. Along with fulfilling curriculum and course outcomes, Quest also helps students bridge the gap between their background and the backgrounds of their peers through gameplay and discussion, creating an environment where students can build empathy for diverse backgrounds. Preliminary student feedback on the game will be shared with IRB approval (CHRP/IRB 2425-82).

Mass layoffs are widespread in the United States, with 65% of counties experiencing at least one mass layoff event annually, even during times of economic growth. They are also unexpected and unrelated to the local community's socio-economic conditions. This study assesses the effect of mass layoffs, an exogenous labor demand shock, on entrepreneurship in the United States. We link county-level residence-based data on mass layoffs from the Bureau of Labor Statistics to measures of proprietorship from the Bureau of Economic Analysis as well as Business Formation Statistics, collected by the U.S. Census, which tracks business creation at the county level. Preliminary results indicate that an increase in the number of laid-off workers in the local area boosts entrepreneurship, with the effect driven by urban communities, but independent of the county's economic prosperity. To shed light on the mechanisms underlying the observed effect and to develop policy implications, we examine the socio-economic characteristics of involuntary displaced workers using data from the American Community Survey and the Current Population Survey.

The COVID-19 pandemic disrupted labor markets in developing countries, but there is still limited study on how firms recover in the medium- to long-term, especially when comparing formal and informal firms. This study investigates labor market changes and recovery in Ghana using data from 3,602 firms across three survey waves: May–June 2020, August–September 2020, and September 2021. A key advantage is that the first wave's employment data is anchored to January 15, 2020, before Ghana's first COVID-19 case, providing a true pre-pandemic baseline. The analysis uses high-dimensional firm fixed-effects models with survey weights and cluster-robust standard errors at the firm level. Formality is based on registration status (62.8% formal, 37.2% informal). The study also examines regional differences, emphasizing areas affected by lockdowns (Greater Accra and Ashanti regions), though these results are interpreted with caution due to timing and measurement issues. The findings show that informal firms had strong employment rebounds in Wave 2 (12.4% to 15% above January 2020) and returned to baseline in Wave 3, while formal firms only partly recovered in Wave 2 (4.4% to 5.3%) and then experienced a huge decline in Wave 3 (−21.5% to −22.6%). Both types of firms cut wages and hours at similar rates, so resilience is mainly about worker retention and firms staying open, not just adjusting labor costs. The effects of the lockdown are more pronounced among formal firms, with higher job losses and wage cuts in lockdown areas, while informal firms mainly adjusted by reducing hours. The results also show that micro firms (1-5 workers) of both types increased employment by Wave 3, whereas small firms (6-30 workers) saw employment decline. Ghana’s informal sector acts as a shock absorber. The findings suggest that relief programs, such as cash support, are more effective when delivered via mobile money than through formal banking, which excludes informal firms that have proven to be the most resilient. Further, regulatory compliance should be adaptable to firm size and, if necessary, suspended during crises.

Panel or grouped data are often used to allow for unobserved individual heterogeneity in econometric models via fixed effects. In this paper, we discuss identification of a panel data model in which the unobserved heterogeneity both enters additively and interacts with treatment variables. We present identification and estimation methods for parameters of interest in this model under both strict and weak exogeneity assumptions. The key identification insight is that other periods’ treatment variables are instruments for the unobserved fixed effects. We apply our proposed estimator to matched student-teacher data used to estimate value-added models of teacher quality. We show that the common assumption that the return to teacher quality is the same for all students is rejected by the data. We also present evidence that No Child Left Behind-era school accountability raised the effectiveness of teacher quality for lower performing students.

This study investigates the dynamic responses of industrial markups to exchange rate fluctuations in the Korean manufacturing sector. Under the foundational principles of cost minimization, we derive a theoretically integrated multi-input markup equation that incorporates capital, labor, and both domestic and imported intermediate inputs. Using a Panel ARDL model, we identify a significant and persistent 'markup squeeze' following currency depreciation. Our findings reveal that the short-run impact (-0.040) intensifies and stabilizes at a long-run equilibrium of -0.010. This paper demonstrates that considering a comprehensive cost structure is critical for capturing the pricing-to-market behavior in small open economies, providing a more robust buffer analysis against external cost shocks.

This paper examines whether small business owners' adoption of retirement plans (SEP IRA, Solo 401k) increases risk-taking in their ventures. We hypothesize that securing personal retirement reduces background financial anxiety, enabling more aggressive business strategies. Using multiple data sources including household surveys with detailed balance sheet information and administrative records linking businesses to retirement plan offerings we measure risk-taking through financial leverage, capital investment, and wealth concentration. Our source of data may include but not limited to SCF, ERISA/DOL filings, Census Bureau and other relevant data soources. After controlling for comprehensive owner and business characteristics, our analysis reveals that retirement plan adoption is associated with significantly higher levels of calculated risk-taking. The effects are strongest among previously retirement-insecure owners and persist across various empirical specifications. Results suggest retirement security policies may have the co-benefit of stimulating small business dynamism, highlighting an important intersection between household financial planning and entrepreneurial economic growth. Our findings will highlight how household-level financial planning decisions shape entrepreneurial strategy, offering new evidence at the intersection of household finance and entrepreneurship.

Local governments frequently consider the long-term impact of spending influxes when making policy decisions. Sectors that are perceived to be more transitory, such as construction, are therefore viewed with a degree of skepticism because the projects of these sectors must end at some point. This study analyzes contemporaneous and lagged effects of increased construction spending on wages, establishments, and employment for counties across the United States. Using data from the Industrial Info Resources for 2010-2019, we find that construction spending significantly affects all three county-level outcomes, with effects intuitively occurring contemporaneously or two years into the future. These findings suggest that increased spending, even in sectors considered to primarily affect short-term outcomes, yield regional benefits through time.