AEA Poster Session
Poster Session
Saturday, Jan. 3, 2026 7:00 AM - 6:00 PM (EST)
Sunday, Jan. 4, 2026 7:00 AM - 6:00 PM (EST)
Monday, Jan. 5, 2026 7:00 AM - 12:00 PM (EST)
(De)regulating Certifiers: Theory and Evidence from the Environmental Impact Assessment Industry (L5, Q5)
Abstract
Third-party certification and auditing are widely used to assess the compliance of firms with regulation yet often substandard in many developing markets. This paper investigates how entry deregulation—a common reform in failing certification markets—affects service quality and welfare. I develop a theoretical model that shows entry deregulation undermines reputation incentives and exacerbates moral hazard, echoing the classical public interest theory of regulation. Empirically, I analyze a deregulation reform in China's Environmental Impact Assessment (EIA) industry, using machine learning to construct a continuous measure of service quality from a corpus of five million documents. The reduced-form estimates show a persistent decline in service quality and worsening environmental externalities post-reform. Finally, I develop a dynamic model of certifiers, based on a moment-based Markov equilibrium that incorporates reputation concerns and externalities. The counterfactual analysis highlights the interplay between regulatory tools: entry regulation, dynamic enforcement, and penalty disclosure.A Time for the Old or the Young? Age Pay Gap and Technological Disruptions (J3, J1)
Abstract
Unlike other population aging countries where the wage disparity between workers over 55 and those under 35 (hereafter referred to as the "age pay gap") has consistently widened, the United States has experienced a narrowing of this gap since the mid-2010s. This paper examines the evolving trends in the age pay gap by introducing a conceptual framework that considers how experience-augmenting and experience-replacing technologies impact the returns on experience for senior workers across various occupations with differing levels of technological exposure.Utilizing data from the Current Population Survey's Outgoing Rotation Group (CPS-ORG), the study documents that in occupations requiring intensive experience-based tasks, the age pay gap tends to widen when exposed to experience-augmenting technologies, such as robotization. Conversely, in fields where experience-replacing technologies, like Generative Artificial Intelligence (AI), are prevalent, the age pay gap appears to shrink. These findings provide insight into the observed widening of the pay gap between older and younger workers prior to the mid-2010s and its subsequent reversal.
While there are prevalent concerns regarding the negative career implications for the younger generation due to an increased stock of older workers in the labor market, this paper offers the perspective that intergenerational inequality is not an inevitable consequence of an aging population. Instead, the development and integration of experience-replacing technologies may serve to mitigate such disparities, fostering a more equitable labor market across age groups.
Accessing External Validity in Instrumental Variables Settings: Machine Learning and Heterogeneous Treatment Effect approach (C4, C1)
Abstract
In this paper, we investigate the issue of external validity in Instrumental Variables (IV) settings when site selection is non-random. To predict the target Local Average Treatment Effect (LATE) from a sample LATE, we employ two approaches: a heterogeneous treatment effect model suggested by Allcott (2015) and a double/debiased machine learning (DML) model. Given concerns about internal validity and the limited extrapolation capabilities of conventional two-stage least squares (2SLS) estimation, we assess whether DML models outperform 2SLS. We conduct simulation studies to evaluate performance under scenarios where key assumptions are violated. We find that DML outperforms the Allcott (2015) method when the "rich covariates" assumption (Blandhol et al., 2025) is violated—a scenario that undermines the internal validity of 2SLS. When external validity concerns arise due to observable variables, both DML and Alcott’s methods yield similar estimates of heterogeneous treatment effects in the target population. However, both approaches decline in performance as external validity issues intensify, especially with greater variance in observable characteristics. In such cases, DML demonstrates better predictive accuracy under larger covariate shifts between the sample and target populations. Furthermore, when unobservable site selection correlates with observed covariates, the DML model exploits this relationship to enhance out-of-sample predictions. Conversely, if unobserved site selection is independent, both DML and the Allcott (2015) method fail, indicating limitations in both internal and external validity. Particularly for external validity, failure is because of violations of the assumption of conditional external unconfoundedness among compliers—that is, the assumption that treatment effects and site selection are independent, conditional on covariates. By identifying circumstances where critical assumptions break down in IV contexts, we provide insights into the strengths and limitations of DML and the Allcott (2015) extrapolation method. Our analysis informs when it is appropriate to generalize treatment effects from sample populations to target populations.Agricultural Commercialization in Sub-Saharan Africa: Does Farmers’ Gender Matter? (Q1, F0)
Abstract
Female farmers in many developing countries receive fewer resources and face more barriers to the markets than male farmers. I investigate the relationship between farmers’ gender and agricultural sales in Ethiopia, Nigeria, and Tanzania, using three rounds of longitudinal household and agricultural surveys. I use the Two-Way Mundlak estimator to identify gender differences in market participation and the value of crop sales. Findings show that female farmers are less likely to engage in selling agricultural output. When they do engage in agricultural sales, the value of sales they receive is lower compared to male farmers. Access to transportation increases the probability of selling for female farmers, but not the value of sales. There is a need for targeted interventions that improve female farmers’ access to markets without making male farmers worse off. Investments to develop complete and inclusive markets are needed to enhance agricultural productivity and ensure food security in sub-Saharan Africa.Allowance of Storefront Recreational Cannabis Retailers and Cannabis-related Healthcare Encounters: A Local-level Spatial Difference-in-Differences Analysis in California, United States (I1)
Abstract
Background and Aims: While half of the states in the United States have approved statewide retail sales of cannabis, local governments retain the authority to opt in or out of allowing storefront recreational cannabis retailers. This study aimed to examine local-level associations between the allowance of storefront recreational cannabis retailers and cannabis-related healthcare encounters in California, United States.Design: A secondary data analysis of cannabis-related healthcare encounters across all the cities in California from 2010 to 2020. A spatial difference-in-differences model was employed at the city-quarter level to assess both intracity and intercity associations, controlling for time-varying city-level policies and sociodemographic factors while accounting for spatial spillover effects on neighboring cities.
Setting: California, United States.
Participants: All California residents from 2010 to 2020.
Measurements: Three cannabis-related healthcare encounter outcomes were assessed: (1) population-adjusted emergency department visits involving cannabis use disorders, (2) population-adjusted inpatient discharges involving cannabis use disorders, and (3) a binary indicator of any calls to poison centers involving cannabis exposure. The primary policy variable was whether a city allowed storefront recreational cannabis retailers.
Findings: No evidence was found of an intracity association between the allowance of storefront recreational cannabis retailers and cannabis-related healthcare encounters. For intercity associations, the allowance of storefront recreational cannabis retailers was associated with a 6.7 percentage point decrease in the probability of any cannabis-related poison center calls in neighboring cities.
Conclusions: In California, United States, the allowance of storefront recreational cannabis retailers appeared to have spillover effects on neighboring cities rather than directly influencing cannabis-related healthcare encounters within the policy-implementing cities. Further research is needed to explore the underlying causal mechanisms. Continued surveillance is recommended in and around jurisdictions allowing storefront recreational cannabis retailers.
Artificial Intelligence, Knowledge Spillovers, and Growth (O4, O3)
Abstract
Advances in artificial intelligence (AI) have demonstrated significant potential to reshape the global economy. This paper develops an endogenous growth model that distinguishes two key effects of AI on innovation: the intensive margin, where AI improves the processing and application of human-accessible knowledge, and the extensive margin, where AI acquires and processes knowledge beyond human capacity. Our findings suggest that in a social planner economy, AI accelerates economic growth through both margins, although it may reduce the fraction of labor allocated to R&D due to the extensive margin. In a decentralized economy, however, AI could slow economic growth and reduce the fraction of labor allocated to R&D because of the extensive margin, thus exacerbating market inefficiencies. We also examine how optimal tax policies can mitigate these inefficiencies.Assessing the Impact of China’s River Chief System on Surface Water Quality (Q5, H4)
Abstract
Surface water quality plays a critical role in public health, economic development, and environmental sustainability, functioning as a key input for drinking water, agriculture, and industrial production. However, its non-excludable and non-rivalrous nature renders it vulnerable to the classic "tragedy of the commons," often resulting in overuse and pollution. This paper evaluates the effectiveness of China’s River Chief System (RCS), a high-profile, top-down environmental governance reform designed to improve surface water quality by assigning local government officials direct responsibility for water bodies within their jurisdictions. Leveraging the staggered rollout of the RCS across Chinese municipalities, I employ a Staggered Difference-in-Differences (SDID) framework combined with high-frequency ammonia nitrogen (NH₃-N) concentration data from the China Environmental Monitoring Station. The baseline estimates indicate that the implementation of RCS policies leads to a statistically significant decline in NH₃-N levels by approximately 0.01 mg/L, suggesting improvements in surface water quality. However, further heterogeneity analysis reveals unintended behavioral responses. Specifically, disaggregating the data into daytime and nighttime monitoring shows that while daytime NH₃-N concentrations decline post-RCS, nighttime levels exhibit a modest but statistically significant increase. This pattern is consistent with strategic pollution behavior, such as delayed or covert wastewater discharge by local actors seeking to avoid regulatory scrutiny during non-monitoring hours. These findings underscore both the potential and the limits of top-down environmental governance in managing public goods. The RCS improves surface water quality by clarifying accountability and incentivizing local compliance, but may also generate distorted behaviors in response to regulatory pressure. This paper contributes to the growing literature on environmental policy and public goods provision by highlighting the role of institutional accountability in improving environmental outcomes and by documenting the behavioral responses that can emerge under performance-based governance systems.Asylum In The Shadow Of Interstate Tensions: Implications For Recognition Rates (F5, F2)
Abstract
I estimate the impact of interstate tensions on asylum recognition rates recorded in the United Nations High Commissioner for Refugees (UNHCR) dataset. Using the gravity framework, I analyze a global sample from 2000 to 2019. I construct three measures of interstate tensions. First, using the Global Sanctions Database (GSDB), I measure tensions arising from the frequency and duration of imposed sanctions between countries. Second, using the UCDP/PRIO dataset, I construct an interstate conflict variable by measuring the frequency and duration of conflicts. Third, using UN General Assembly voting patterns as a proxy, I measure the ideological differences between countries. My results show that longer sanction histories between countries and larger ideological differences result in higher asylum recognition rates, while interstate conflict, although positive, is insignificant. This research highlights the importance of non-humanitarian factors in granting asylum.Attention and Memory in Stock Sale Decisions (D9, G4)
Abstract
Both extreme positive and negative stock or market 1-day returns lead to a higher propensity to realize stocks. We observe that reactions to extreme events are modulated by experienced realized returns, according to the principles of selective recalling. After extreme positive one-day stock or market returns, investors whose best past realized returns were higher tend to react less. After extreme negative one-day stock or market returns, investors whose worst past realized returns were lower tend to react more. These effects are consistent with a process of memory retrieval guided by similarity, which in turn influences belief formation, with past positive (negative) memories of past realized returns leading to more optimistic (pessimistic) beliefs which mitigate (exacerbate) propensity to sell after extreme positive (negative) 1-day stock or market returns. We also observe that the mitigating effect of positive memories does not decay with time, while the exacerbating effect of negative memories is stronger if these memories are more recent, consistent with biased recalling. Moreover, more attentive investors show both stronger mitigating (exacerbating) effects of positive (negative) memories.Attention Cost of Unfair Treatment (J7, M5)
Abstract
Unfair treatment, whether actual or perceived, is widespread in the labor market. Ruminating over perceived injustices can occupy mental bandwidth that would otherwise be directed towards the task at hand. This paper explores the productivity and pecuniary effect of such cognitive distraction by unfair treatment on individual financial outcomes using an online experiment. I consider a reduction in performance through the channel of attention. Participants engage in an online real-effort performance task for a financial bonus, with some participants able to take the financial bonus from another subject and reallocate it to themselves. The recipients learn if their financial bonus was taken after completing the real-effort task but before completing a series of cognitive pay-for-performance tasks. A control group of participants perform the same tasks without the loss of a financial bonus. I find a significant reduction in performance due to the treatment, especially among participants who perceived discrimination was a factor in the decision to take the financial bonus (especially race discrimination). Victims of unfair treatment performed worse on subsequent tasks, impairing productivity and reducing individual earnings by 5-8 percent (0.4 SD’s). The results suggest that the cognitive cost of unfair treatment can have significant financial and productivity consequences.Automatic Bandwidth Selection for HAC Estimators Under Network Dependence (C1, C2)
Abstract
This paper focuses on the estimation of covariance matrices in the presence of cross-sectional dependence induced by an observed network structure. Existing network heteroskedasticity and autocorrrelation consistent (HAC) variance estimator rely on the choice of a lag truncation parameter, or bandwidth. While the literature provides conditions on the growth rate of the bandwidth parameter that ensure the consistency of network HAC estimators, no results are available regarding its choice for a fixed sample size. To fill this gap, we propose a data-driven approach for selecting the bandwidth for the network HAC estimator. Specifically, we derive the asymptotic mean squared error of network HAC estimator for the sample mean of a \psi - dependent process, leveraging results from Kojevnikov, Marmer, and Song (2021). Based on this, we characterize the asymptotically optimal bandwidth minimizing the asymptotic mean squared error, and propose an estimator of it.Automatic Enrollment and Optimal Defaults in a Second Best Environment: Evidence from Auto-IRAs (D1, D9)
Abstract
Using linked employee-employer data, I examine the effects of automatic enrollment in state auto-IRA programs and find large, persistent increases in retirement saving, with participants retaining their savings even after job separation. Leveraging the automatic escalation feature of auto-IRAs, I show that auto-IRA participants facing higher default rates are more likely to opt out of default rate saving, with many selecting a zero saving rate. To explain these behaviors, I develop a theoretical model of saving that extends previous models by incorporating frictions for deviating from both non-saving and default saving. Using a bunching framework, I estimate annual frictions of 0.81% of income for default saving and 0.79% for non-saving. I then calculate the optimal default rate under various assumptions on whether these frictions reflect real costs or behavioral biases, finding this optimal rate to be between 3.2% and 3.6% under each assumption. In most parameterizations, the results rule out default policies that promote active choice as optimal. The findings recommend setting broadly attractive default rates, even if default effects reflect behavioral biases; in this case, the default rate acts as a "second best" option that mitigates other distortions to saving behavior.Average Inflation Targeting in Open Economies with Sovereign Risk (E3, E5)
Abstract
We study how sovereign default risk affects monetary policy transmission and evaluate the performance of Average Inflation Targeting (AIT) in a tractable, log-linearized small open economy New Keynesian framework. In our novel framework, sovereign risk endogenously raises firms’ working capital costs, which increases marginal costs and inflation, thereby weakening the central bank’s influence on inflation through the traditional cost channel. We show that AIT, by better anchoring expectations than standard Inflation Targeting (IT), significantly reduces macroeconomic instability and welfare losses associated with sovereign risk. These results establish AIT as a powerful monetary policy framework for both developed and emerging economies with high default risk.Bank Credit Supply, Debt Maturity, and Investment over Business Cycles (E5, G1)
Abstract
This paper examines how fluctuations in bank credit supply influence corporate debt maturity and, in turn, firm investment over the business cycle. Using firm-bank level data, we empirically separate credit demand and credit supply and show that shifts in bank credit supply are key drivers of corporate debt maturity decisions during recessions. We then estimate panel regressions to study how changes in debt maturity affect firm-level investment. Our results indicate that firms significantly reduce investment when debt maturity shortens in recessions: a 1% decline in debt maturity in the previous period leads to a 0.92% drop in the investment-to-capital ratio. Motivated by these findings, we develop a dynamic model in which firms endogenously choose debt maturity and default under varying credit conditions. A quantitative analysis of the model highlights how procyclical credit supply shapes firm dynamics through its impact on debt maturity and investment decisions.Bayesian Synthetic Control with a Soft Simplex Constraint (C1, C2)
Abstract
Whether the synthetic control method should be implemented with the simplex constraint and how to implement it in a high-dimensional setting have been widely discussed. To address both issues simultaneously, we propose a novel Bayesian synthetic control method that integrates a soft simplex constraint with spike-and-slab variable selection. Our model is featured by a hierarchical prior capturing how well the data aligns with the simplex assumption, which enables our method to efficiently adapt to the structure and information contained in the data by utilizing the constraint in a more flexible and data-driven manner. A unique computational challenge posed by our model is that conventional Markov chain Monte Carlo sampling algorithms for Bayesian variable selection are no longer applicable, since the soft simplex constraint results in an intractable marginal likelihood. To tackle this challenge, we propose to update the regression coefficients of two predictors simultaneously from their full conditional posterior distribution, which has an explicit but highly complicated characterization. This novel Gibbs updating scheme leads to an efficient Metropolis-within-Gibbs sampler that enables effective posterior sampling from our model and accurate estimation of the average treatment effect. Simulation studies demonstrate that our method performs well across a wide range of settings, in terms of both variable selection and treatment effect estimation, even when the true data-generating process does not adhere to the simplex constraint. Finally, application of our method to two empirical examples in the economic literature yields interesting insights into the impact of economic policies.Big Fish in Small Pond or Small Fish in Big Pond? The Educational Impacts of Tracking. (I0, I2)
Abstract
The practice of allocating students to different schools and classrooms based on their prior performance is widely implemented across various educational systems, including the United States, China, and OECD countries. However, its impact on student outcomes remains debated. This study examines whether academically marginal students benefit more from being lower-performing in high-achieving schools/classes or higher-performing in regular schools/classes. Our context is two representative counties in China, where first-year high school students are separated into high-achieving and regular schools, as well as elite and regular classrooms, based solely on their scores from the provincially standardized High School Entrance Exam (HSEE). Using a unique administrative dataset encompassing 31,406 participants from the standardized National College Entrance Exam (NCEE) between 2016 and 2024, and employing a regression discontinuity design, we find that, despite a significant disparity in overall educational performance between students in high-achieving and ordinary environments, the top 10% "leaders" in ordinary high schools significantly outperform the bottom 10% "laggards" in high-achieving schools by approximately 0.08 SD in total NCEE scores. Similarly, laggards in elite classrooms within each school exhibit significantly lower educational performance compared to leaders in regular classrooms. These results remain robust across various specifications, including adjusting the bandwidth and using scores from each subject. Mechanism analysis indicates that while teacher quality accounts for the superior performance of students in high-achieving environments, self-confidence may also play a critical role in shaping educational outcomes. This is evidenced by the consistent out-performance of leaders in ordinary schools or classrooms over their peers from similar backgrounds in prior mock exams and mid-term assessments. These findings highlight the causal impact of high-achieving environments on students and provide empirical evidence for the influence of self-confidence—a critical yet often overlooked factor inBlades of Despair? The Impact of Wind Farms on Crime (K4, Q4)
Abstract
Wind energy has rapidly expanded over the past two decades, now accounting for a substantial share of new electricity generation capacity in the United States. A growing body of literature has documented the economic benefits of wind energy—such as increased income, GDP, and home values—and its potential health costs. This includes evidence that low-frequency noise from wind turbines may contribute to sleep disruption and increased suicide rates. This paper contributes to this literature by examining the impact of commercial wind turbine installations on local crime rates. Using Uniform Crime Reporting (UCR) data from 1995 to 2020 and a staggered difference-in-differences model, we estimate the causal effect of wind energy development on a range of criminal offenses. To explore potential mechanisms, we leverage the Nielsen Retail Dataset to examine changes in alcohol sales in areas near new wind farms. Our findings indicate that county-level crime rates decrease after the wind turbines are installed in the area. These results are robust to the inclusion of demographic controls. We also investigate spatial heterogeneity to assess more localized impacts.Bundling to Mitigate Adverse Selection: Evidence from Land-Expropriated Farmers in China's Social Security System (I1, H0)
Abstract
In 2003, Zhejiang Province, China, established a low-premium, low-benefit social security system for land-expropriated farmers. Starting in 2011, the province allowed eligible farmers to make a lump-sum payment to transfer to the high-premium, high-benefit Basic Pension Scheme for Employees (BPSE) and Basic Medical Insurance for Employees (BMIE). This policy transition offers a unique opportunity to study adverse selection in social insurance. Using micro-level social insurance data from C City from 2011 to 2020, we identify evidence of adverse selection in both pension and medical insurance. We also extend classic information asymmetry insurance theory (Akerlof, 1970; Einav et al., 2010) from a single-insurance market to a setting with two types of insurance. By exploiting exogenous government-set price variation, we estimate demand as well as average and marginal cost curves for BPSE and BMIE. These curves allow us to quantify the welfare loss caused by adverse selection and evaluate whether bundling pension and medical insurance can mitigate adverse selection through risk pooling.Can Digital Extension Reduce Farmers’ Vulnerability? Randomized Evidence from Rural Sri Lanka (Q1, C9)
Abstract
Smallholder farmers often lack access to timely and tailored information, hindering their ability to adapt and make informed decisions. This paper evaluates the impact of the Smart Extension and Efficient Decision-making (S.E.E.D.) Hub, an integrated e-extension mobile application, on farmer’s vulnerability in Sri Lanka using a randomized controlled trial. Developed by the Food and Agriculture Organization, this app serves as a one-stop shop, providing rural paddy farmers with free, up-to-date, language friendly, location-specific agricultural information, including weather updates, market prices, disaster alerts, and crop management practices.This study involves 220 farmer organizations (FOs) randomly assigned to treatment or control groups. Among 110 treated FOs, 55 received app access (T1), while the other 55 received access plus a three-hour training (T2). Following a multi-stage cluster-stratified design, 2,200 households (ten per FO) were surveyed at baseline (2022), followed by midline (2023) and endline (2024). Attrition rates were low (3%), ensuring a balanced panel of 2,100 households.
Balance tests showed successful randomization: comparison between treatment and control groups accounted for baseline characteristics and unobserved differences at strata, irrigation scheme, and district levels. Local Average Treatment Effect (LATE) was estimated for sub-population of compliers, those who downloaded the app, with intent-to-treat included for robustness. Note that households receiving T2 show a slightly (15%) higher likelihood of compliance relative to T1.
Results showed a 15% increase in information access among treated farmers, rising to 50% for compliers, driven by improved weather and market information. Treated farmers were also more likely to diversify crops (79.5%), increase rice productivity (27.2%), sell at higher prices (33.5%), and generate more revenue (64.6%). Additionally, timely access to information proved particularly valuable during natural disasters, helping farmers mitigate farm and market risks. Paper suggests further complementary interventions to maximizing the impact of digital extension services in Sri Lanka.
Can We Govern the Ungovernable? The Computational Challenge of AI-to-AI Contracts (D7)
Abstract
Artificial intelligence is revolutionizing economic interactions, with autonomous agents now negotiating, modifying, and enforcing contracts at speeds and scales beyond human comprehension. This paper introduces extended Multi-agent Influence Diagrams (eMAIDs)—a novel game-theoretic and computational framework that captures how AI agents not only make strategic decisions but also reshape the very rules governing their interactions through iterative contract modifications. Unlike traditional contract theory or fixed-structure game models, eMAIDs allow agents to propose and implement changes to the game graph itself, modeling real-world AI-to-AI contracting scenarios such as algorithmic trading or smart contract execution on decentralized blockchains.We demonstrate that even under highly simplified assumptions, determining the final outcomes of eMAID-based contract negotiations is PSPACE-hard, and in more general settings, Turing uncomputable. Additionally, evaluating whether a third party can exploit a contract is also PSPACE-hard, raising serious concerns about the feasibility of regulatory oversight. A key result reveals an asymmetry: while AI agents can efficiently compute their best-response strategies, human regulators face intractable or even undecidable problems in attempting to predict or control outcomes.
These findings highlight a paradigm shift in economic governance. As AI agents increasingly dominate contracting, traditional tools of oversight may prove insufficient. We argue for the need to develop AI-driven regulatory systems capable of interpreting and constraining autonomous contracts in real time. Our work underscores that AI contracting is not merely an evolution of existing systems but a potential rupture—heralding economic structures that may be fundamentally opaque to human understanding.
Capital Market Integration, Labor Market Distortions, and Labor Misallocation (J4, F2)
Abstract
This paper examines the labor market effects of foreign capital liberalization in India, with a focus on labor misallocation and employer market power, a key source of misallocation. We estimate the firm-specific labor markdowns, the gap between MRPL and wage, as a proxy for monopsony power. Leveraging a foreign direct investment liberalization episode as a natural experiment, we use difference-in-differences and event study designs to identify the causal impacts of capital market integration on the labor market. For firms with ex ante high MRPL, liberalization increases employment by 17% and wage bills by 15% and reduces MRPL by 14% and labor markdowns by 15%, with no impact on wages relative to low MRPL firms. These effects are driven by female workers.Central Bank Information Effects and the Transmission of the Global Financial Cycle (E5, F3)
Abstract
This paper identifies central bank "information effects" as key drivers of the global financial cycle using our novel Monetary Policy Statements Database (MPSD), comprising 8,274 monetary policy statements from 51 central banks. We examine instances where equities and policy rates co-move, termed "information shocks," finding this phenomenon is prevalent and associated with text sentiment and verbiage related to macroeconomic conditions and inflation. Statements linked to positive information shocks (i.e., rate hikes with higher equity prices) correlate with dovish sentiment and inflation-focused language, suggesting these sometimes represent pure monetary policy shocks. Conversely, negative information shocks (i.e., rate cuts with lower equity prices) support findings that information effects relate to risk-taking capacity rather than monetary policy reactions. Similar patterns emerge when analyzing policy rate pass-through, CDS, and FX markets. Notably, negative information shocks frequently appear in succession before economic downturns, particularly preceding the 2008 and 2020 crises. These shocks propagate risk-off sentiment (measured through lexical and LLM text analysis methods) prior to global market sell-offs. We confirm this relationship using both original global financial factor data and our newly created daily-frequency proxy for the global financial cycle. Our analysis reveals significant cross-country variations in the intensity and timing of information effects, influenced by central bank credibility, communication strategies, and country-specific macroeconomic contexts. Our findings explain how central bank communications impact global financial markets and contribute to financial cycles.Certification Design for a Competitive Market (D8, L5)
Abstract
A designer chooses market rules for a two-sided market for vertically differentiated products. A common mechanism for this type of rule is to allow for free trade of goods of deterministically certified quality thresholds. We first give conditions under which optimal mechanisms can be implemented as such *certification for a competitive market*. We then show that under these conditions, the design problem reduces to the problem of selling surplus to quasi-assortatively matches buyer-seller pairs, and that a virtual surplus formulation applies. In particular, if designers have greater concern for money, certification will be lower. With sufficient concern for money, certification is weakly lower than under full certifiability. With sufficient concern for quality, certification is weakly higher than under full certifiability. As an application, we discuss the (sub-)optimality of certification design in voluntary carbon markets.Characterizing Returns to STEM (I2, J2)
Abstract
Most developed countries have been fostering education in the STEM field to increase innovation. In this study, we estimate heterogeneous returns to a STEM education in Switzerland based on individual-level data, exploiting the regional distribution of relative distances to technical and cantonal universities as a cost factor driving college major choice. A clear setting in the 1990s where prospective students could freely choose universities and majors allows us to exploit the regional distribution of technical and cantonal universities for identification in an instrumental variable framework. On average, individuals strongly gain in terms of wages from a STEM education. Descending Marginal Treatment Effect (MTE) curves suggest positive selection on gains, implying that individuals with a low resistance for a STEM education gain the most. This positive selection is driven by both heterogeneity in the returns to a non-STEM education as well as a STEM education. By simulating policies that aim to increase STEM enrollment and estimating corresponding policy-relevant treatment effects (PRTEs), we show that the effectiveness of such policies strongly depends on who is affected and, thus, on their observable and unobservable characteristics. We also show how such policies must be designed to increase STEM enrollment and benefit targeted individuals, most importantly, women.China's Expanded Network Coverage and the Mental Health of the Elderly and Adult Population (I1, I3)
Abstract
Due to the rapid growth of China's cities and the imbalance in rural development, more and more people are choosing to relocate to cities. Their parents may live in the city with their children or stay at their original address in the countryside. However, under China's province-based public health insurance system, these older people do not have access to health insurance entitlements in the provinces where their children live. The lack of access to public health insurance makes middle-aged and elderly people worry about the cost of medical care and results in health problems, especially mental health. The national system of direct settlement of inter-provincial hospitalization medical fees was established in 2017 to meet the needs of these ‘out-of-network’ patients in China. This paper examines the impact of the 2017 reform on the mental health of middle-aged and elderly migrants. This study utilises data from the China Health and Retirement Longitudinal Study (CHARLS) and uses the proportion of tertiary hospitals with an inter-provincial direct billing system in place after the 2017 reform as an exogenous change in the three-differential setting. Our analysis finds that the reform improved mental health for middle-aged and elderly people. This study provides empirical evidence on the impact of health insurance coverage.Climate Change and Monetary Policy: A Bayesian DSGE Perspective (E5, Q5)
Abstract
This study utilizes a Bayesian DSGE model with endogenous productivity, calibrated using uniquely rich Australian data—including seasonally adjusted and weather-normalized CO₂ emissions—to examine how monetary policy affects emissions. The availability of such data provides an ideal foundation for assessing the relationship between monetary policy and emissions within a DSGE framework.We find that a 1% cyclical fluctuation in GDP from its trend corresponds to a 0.52% cyclical fluctuation in CO₂ emissions relative to trend. This elasticity is smaller than previous estimates, which range from 0.64 to 0.70. We estimate that a 1 percentage point increase in interest rates leads to a 0.8% decrease in GDP, a 0.4% decline in emission flows, and a 2.1% reduction in the pollution stock, all relative to their trends. The short-run effect on emissions is considerably larger than in earlier studies, while the long-run effect—estimated at just a 0.1% decline in emissions—is notably smaller. Our estimates align more closely with findings in the broader monetary policy literature, where monetary policy has strong short-term effects but modest long-term real effects. Our estimate of the impact on the pollution stock (2.1% relative to trend) is, to our knowledge, entirely novel. This provides new insight into the cumulative environmental consequences of monetary policy.
Historical decomposition shows that while monetary policy shocks contribute to cyclical emissions variability, they often coincide with private demand shocks that move emissions in the opposite direction. Given the long-term nature of climate change and the countercyclical role of monetary policy in emissions, we conclude that climate concerns should not be a central focus of monetary policy.
Closing One Door, Opening Another: Spillovers from Germany’s Western Balkan Regulation (F2, J6)
Abstract
Germany enacted the Western Balkan Regulation in November 2015, creating a new legal pathway for nationals from Western Balkan countries to enter Germany for employment purposes. This regulation was introduced alongside broader asylum reforms that designated Albania, Kosovo, and Montenegro as safe countries of origin. This paper examines the impact of the regulation on flows of both economic migrants and asylum seekers from the Western Balkans to Germany and other EU countries. Using data from 2010 to 2023, the results suggest that the regulation is associated with a significant increase in residence permits issued to Western Balkan nationals for employment in Germany, relative to unaffected countries. At the same time, the regulation and broader asylum reforms are found to be associated with a significant decline in asylum applications received by Germany from Western Balkan nationals post-2015, compared to unaffected countries. This indicates a shift from asylum seeking to employment-based migration as a result of a policy change, known as categorical substitution in migration literature. Beyond Germany, the paper also explores the spillovers effects of this regulation on other EU countries. It examines potential destination substitution, where asylum seekers from the Western Balkans redirect their asylum applications to other EU member states or economic migrants from the region choose to relocate to Germany instead of other EU member states due to the regulation. These findings underscore the significant impact migration policies can have on shaping both the volume and direction of migration flows, particularly when they target specific countries or types of migrants.Collateral, Household Borrowing, and Income Distribution (E3, E0)
Abstract
In this article, we provide key stylised facts about the relationship between household debtand wealth accumulation at both the state and household levels. We then integrate a collateral
constraint into a model with heterogeneous agents to study the effects of collateral on wealth
inequality. We use estimates from US microeconomic data and the simulated time series from our
macro model to predict the wealth accumulation response at the top and bottom of the personal
income distribution. Debt is modelled as collateral-dependent, and its concentration poses a
serious concern. Our results indicate that high collateral requirements benefit high-income more
than low-income households.
Collusion, Elites and Foreign Entities: The Case of Late Tsarist Russia (P0, N1)
Abstract
Political elites can have a substantial impact on economic growth, but if an economy is in the early stages of modern growth, where major structural transformation is required, elites could have an even greater impact by profoundly altering the growth path of that economy. How elites affect growth is still not fully understood, especially for late-industrializers - countries that rely on adopted technology to industrialize. Elites can have a positive effect by navigating through existing barriers, or using their influence to increase their own rents. Furthermore, foreign agents or entities who have financial capital and technology may play an outsized role in a structurally changing economy. It is unclear whether foreigners use their own advantages to inject capital into the economy or increase their own rents. This paper uses the context of Late Tsarist Russia to study how elites and foreign entities affect competition in a late-industrializing economy. In the late 19th and early 20th centuries Tsarist Russia was trying to industrialize. During this time it is not clear to what extent monopolies and collusion permeated the Tsarist economy, and if political elites - the nobility and government officials - and foreign investors influenced anti-competitive behavior. Using a detailed dataset on corporate charters and a new dataset I collected on collusive agreements and organizations, I explore the impact of the presence of political elites and foreign actors on collusive activity, and the subsequent implications for economic growth. I find elites and foreigners have a negative impact on new collusive activity, likely because they hold enough political influence that substitutes for the need to collude compared to other social groups with less influence.Colonialism and Trust: Evidence from Morocco (Z1, O1)
Abstract
This paper demonstrates how exposure of Morocco to different institutional and administrative practices under the French and Spanish protectorates in the 20th century has a long-term impact on interpersonal trust and trust in institutions. French administrators believed that Berbers, with their distinct traditions and customs, were different from Arabs, viewing them as more suitable for assimilation into French culture. I show that the French administration’s Berber policy, aimed at creating ethnic dualism by separating legal systems for Berbers (urf) and Arabs (sharia), led to higher trust levels in the region under the former Spanish protectorate, where such a policy was not implemented. Specifically, by using spatial regression discontinuity design, georeferencing historical maps and constructing trust variables from Afrobarometer survey, I find that people living on the Spanish side of the border have more trust in the parliament and president. I also create a government performance evaluation index, showing that they have more positive views on how the government is managing economic issues. Additionally, those on the Spanish side have better attitudes toward people of other religions and ethnicities. I further confirm that these results are not influenced by wealth or economic activity proxied by nighttime lights, and that they hold under various specifications and bandwidth choices. A falsification test shifting the border south by up to 45 kilometers shows no effect at any of false borders. To the best of my knowledge, this is the first paper exploiting this type of variation in North Africa.Communist Party Building and China’s Enterprise Reform: Structure, Mechanism and Effect from “Mixing” to “Integration” (P2, L2)
Abstract
China is now focusing on gathering strength. Communist party building is one of the most important method, which is also the unique form of organizational capital in Chinese enterprises. This study examines whether communist party building plays a collaborative governance role in deepening mixed-ownership reform from “mixed property rights” to “integrated mechanisms”. We find that it significantly enhances mixed-ownership reform performance, especially during cooperative periods because its stabilizing senior management teams, reducing organizational vulnerability, and achieving effective supervisio. Party organization and party member are the two perspectives we describe, which show a stronger impact of executive level. Following Magne et al. (2021), we use multiple instrumental variables of the distance to the birthplace of China Communist Party and national party organization/member development which is still robust. For the efficiency outcome, communist party building extends and deepens reform effects, boosts state-owned enterprise’s vitality, and enhances private-owned enterprise’s confidence.Competing Information Intermediaries (D8, L1)
Abstract
We study competition among information intermediaries in a market where pricing consists of a participation fee and a disclosure fee. A seller hires an intermediary to test a good’s quality and may disclose the resulting signal to buyers. We derive the intermediaries’ optimal pricing and information structures. In contrast to monopoly, we show that oligopolistic competition reduces disclosure fees and induces more informative signal structures. The presence of multiple intermediaries enhances overall market informativeness due to adversarial disclosure strategies. Specifically, when buyers observe a signal from intermediary A, two scenarios are possible: (a) the seller has only participated in A’s test and disclosed the resulting signal, or (b) the seller has participated in tests from both A and B but chooses to disclose A’s signal because it is more favorable. To compete, intermediary B can influence buyers’ beliefs by making them perceive scenario (b) as more likely. Then, B can construct a more precise information structure, thereby diminishing the informational value of A’s disclosed signal. Buyers, anticipating the possibility of undisclosed unfavorable signals, update their beliefs accordingly, preventing the disclosed good signal from significantly increasing their ex-post beliefs. Beyond price reductions, we show that constructing a more precise information structure is a strategic tool in intermediary competition. Furthermore, we explore an asymmetric environment where intermediaries differ in their commitment power and moral hazard. Our findings are robust: the intermediary with the strongest commitment power dominates the market, with their pricing and information strategy constrained by the second-highest commitment level.Competitiveness and Partner Income: Gender Differences in Mating and Cross-Productivity Effects (J3, J1)
Abstract
Gender differences in competitiveness are typically studied for individual labor market outcomes, yet Becker’s (1973) marriage theory suggests traits like competitiveness affect partners’ incomes via mating and cross-productivity effects. Mating effects stem from pairing with economically enhancing partners, while cross-productivity effects occur when one’s traits boost the other’s income. Using Dutch household panel data, we find that men’s and women’s competitiveness predict their future income, but only women’s competitiveness significantly increases their male partner’s income, even after controlling for personality traits and couple fixed effects with a novel strategy addressing the single measurement of competitiveness. Men’s competitiveness does not affect their female partner’s income. Women’s competitiveness does not increase men’s income via household specialization: only women’s work hours rise with their competitiveness, not men’s. Women’s competitiveness also does not reduce men’s housework or childcare time. Conversely, men’s competitiveness increases women’s housework and decreases men’s childcare. Financial satisfaction moderates the effect of women’s competitiveness on men’s income. Our findings suggest women’s competitiveness may account for 10–31% of the gender income gap among partnered individuals, highlighting its broader economic role.Corporate Bond Refinancing Under Capital Supply Uncertainty (G3, G2)
Abstract
Corporate bond refinancing, which replaces existing debt with new issuances, unlike issuance for new financing needs, faces a fixed maturity deadline. This deadline makes the firm’s bond refinance decision take different responses to capital supply uncertainty. In this paper, I examine the effect of capital supply uncertainty – measured as the average flow volatility of mutual fund investors holding the bond – on firms’ refinancing decisions on the bond. My main finding is that the credit supply uncertainty hasa positive and significant impact on high-yield bond refinancing, and a positive but statistically insignificant effect on investment-grade bond refinancing. To address potential endogeneity concerns, I use a shift-share instrumental variable. These results highlight a novel fact that capital supply uncertainty can make firms’ refinancing decisions sooner rather than later.
Counterfactual Fairness and Explanations in Credit Scoring (M1, C1)
Abstract
Machine learning is widely used in credit risk assessment due to its efficiency and accuracy. However, concerns around interpretability and fairness remain pressing, particularly for individuals denied credit, who are most in need of actionable recourse. Existing methods for counterfactual fairness are largely prediction-focused, offering limited support for those affected by negative outcomes. Moreover, current counterfactual explanation approaches in credit research tend to rely on statistical correlations, often neglecting causal relationships of features, robustness to bias, and the integration of fairness with interpretability. We propose a novel causal framework that integrates counterfactual fairness with counterfactual explanations, focusing on recourse rather than mere prediction. We ask: if a rejected female were counterfactually male, would she need to exert less effort, or receive more diverse and feasible recourse options to achieve approval? Our framework goes beyond minimal recourse cost by also considering the number and diversity of actionable actions, critical dimensions for fairness. We respect causality, encode the structure causal model into the counterfactual generation process and enhance robustness against measurement errors and model misspecification by incorporating multiple techniques, including Gaussian-based methods and conditional variational autoencoders. Our framework supports both point- and distribution-based counterfactuals, and handles both discrete and continuous data, making it particularly applicable to real-world credit datasets. We validate our method on Small and Medium-Sized Enterprises' access to finance and synthetic datasets, showing consistent improvements in fairness, interpretability, and robustness. By ensuring that counterfactual recourse options are equally attainable across groups, our framework provides rejected applicants with fair, feasible, and transparent actions toward acceptance, offering a principled and equitable solution to long-standing challenges in credit risk decision-making.Covariate Selection for the Synthetic Control Method (C5, C2)
Abstract
This paper proposes a new procedure to select appropriate predictors for the synthetic control method (SCM) using the adaptive group LASSO algorithm. Traditional SCM assumes that a set of time-invariant predictors share the same synthetic control weights as the outcome of interest. We consider time-varying covariates, some of which have different synthetic control weights from the outcome of interest. Our method selects the relevant predictors and estimates the synthetic control weights in a single step. We present a data-driven procedure to jointly optimize the selection of the penalty terms and the weights on each covariate. Our method is robust against potential biases from mis-specified predictors and enhances efficiency by fully exploiting the appropriate predictors. We prove that our method inherits the oracle property from LASSO, ensuring consistency in predictor selection and estimation. In Monte Carlo simulations, our method demonstrates strong finite-sample performance across various data generating processes.Credit Strikes Back: The Macroeconomic Impact of the 2022-23 ECB Monetary Tightening and the Role of Lending Rates (E5, E3)
Abstract
This paper examines the transmission of the European Central Bank’s (ECB) 2022-23 monetary policy tightening to the euro-area economy, with a particular focus on the cost of credit for non-financial corporations (NFCs). The analysis employs simple pass-through equations, a Bayesian VAR (BVAR) model, and an estimated dynamic stochastic general equilibrium (DSGE) model with a banking sector. Three key findings emerge from the study. First, during the 2022-23 tightening cycle, the ECB's key interest rates increased significantly more and at an unprecedented pace compared to past tightening episodes, leading to a stronger transmission of monetary policy to the credit market. Second, BVAR model results highlight that banks' risk perceptions - drawn from the Euro Area Bank Lending Survey - played a crucial role in amplifying the transmission of the tightening to credit costs. Finally, simulations based on both the BVAR and DSGE models indicate that the 2022-23 tightening had a substantial impact on euro-area GDP growth and inflation, with the bank lending channel identified as a key driver of these effects.Demand for Income Risk Mitigation (J6, D8)
Abstract
Independent contract work has increased in the 2000s (Collins et al. (2019); Katz and Krueger (2019a, 2019b); Abraham et al. (2020)). While such work may enable people to self-insure against income or expenditure shocks, short-term contracts are also risky. There is uncertainty whether a suitable job is available when needed, and risks of (last-minute) job cancellations or hours being cut. Moreover, independent contractors are often excluded from unemployment insurance, (employer-provided) health insurance and retirement programs, occupational health and safety regulations, and wage and hour laws. Thus, as regular employment contracts decrease, firms are insuring agents against risk less. This begs the question whether there is demand for (supplemental) risk mitigation, specifically of income risk. This paper studies the demand for income risk mitigation using a large ex-ante incentive-compatible survey experiment on a nationwide online labor market platform.Workers using the platform are presented with a series of discrete choices between a job contract that offers less insurance and one that offers more insurance against a specific type of income risk. Workers are randomized into different income risk and pay groups, information treatments and alternative insurance options. The survey also elicits workers’ risk beliefs and preferences, household income and liquidity constraints. Worker choices allow us to estimate the marginal rate of substitution (MRS) and willingness to pay (WTP) for income risk mitigation across risk types, insurance levels and pay levels; the sensitivity of the MRS and WTP to changes in beliefs; and workers’ willingness to trade-off flexibility for insurance.
We find that there is high demand for income risk mitigation. Across risk and pay groups, 40 to 62% of respondents is willing to pay at least 12% of posted pay for insurance. The high demand appears to be driven by binding liquidity constraints, though people also tend to overestimate risks.
Digital Arbitrariness and Centralization of Power, Experiment in China (P0, A1)
Abstract
Historically, information has mediated the relationship between different tiers of government. In the digital age, the question of how digital technology alters the dynamic between levels of government becomes crucial. Specifically, concerns arise regarding potential shifts in power, such as increased centralization and arbitrary decision-making by higher-level authorities when armed with extensive data.To investigate, we designed and implemented an RCT experiment with over 2000 Chinese civil servants. The study centered around a land type monitoring scenario using satellite digital information systems. Respondents were bifurcated into higher-level and lower-level government groups. The higher-level group made decisions informed by various visual data, including satellite imagery, while the lower-level group responded to directives. We gauged decision-making tendencies and responses through multiple metrics.
Our findings are revealing. Officials leveraging digital technology, particularly satellite data, displayed heightened authoritarian and centralized traits. Higher-level respondents with satellite access favored stricter orders. In the lower echelons, subordinates initially attempted to relay the truth but resembled to deception when rebuffed. Heterogeneity analysis disclosed that those lacking grassroots exposure and in influential departments were more prone to digital authoritarianism. Notably, higher-level digital overconfidence spurred lower-level subterfuge, pressuring grassroots workers without directly disrupting local affairs. Moreover, supplying “bad information” to superiors compounded formalistic burdens on the ground. In sum, our research deciphers the micro underpinnings of digital technology’s sway on government governance and illuminates the intricate web of interactions and adverse outcomes among different government levels.
Distance Frictions and Resource Reallocation in Multi-establishment Firms: Theory and Evidence from Airline Route Expansions (D2, F0)
Abstract
We study how reductions in distance-related frictions with headquarters affect labor allocation and establishment organization in multi-establishment firms. Using matched employer-employee data from Brazil and a difference-in-differences design, we estimate the effects of new airline routes connecting branch establishments to their headquarters on branch-level outcomes. In contrast to prior studies, we find that improved connectivity reduces non-managerial employment at the branch and increases the likelihood of branch closure. To interpret these results, we develop a model in which firms weigh two types of distance-related frictions: monitoring costs, which make it harder to oversee local labor, and market delivery costs, which make it more difficult to serve consumers remotely. Improved connectivity lowers both. By enabling better oversight, it can expose and correct inefficient labor allocation—leading to employment reductions or exit, depending on the relative importance of each cost. The model highlights how the impact of reduced frictions depends on pre-existing distortions and can account for the mixed effects of connectivity documented across contexts.Do Central Bank Reforms Lead to More Monetary Discipline? (E5, C2)
Abstract
This paper investigates the impact of reforms altering legal central bank independence (CBI) on monetary policy discipline and credibility, two key mechanisms shaping price stability. Using a sample of 155 countries over more than 50 years (1972–2023), we show that reforms improving CBI strengthen monetary discipline and the credibility of central banks. Larger reforms enhance monetary discipline with a lag, achieving their full effect after ten years. Central bank reforms have a greater impact on monetary discipline in countries that have not reversed earlier reforms. CBI reforms have the strongest impact in democratic countries, countries with flexible exchange rates, and those without a monetary policy strategy. The effects of CBI on monetary discipline and credibility are amplified when public debt-to-GDP ratios are high. These findings underscore the crucial role of CBI as a key factor influencing price stability and highlight the risks associated with weakening institutional autonomy.Doctors With(out) Borders: Labor Supply Responses to Occupational Licensing For Foreign Physicians (N3, J2)
Abstract
Many countries implement policies to attract immigrant healthcare workers in response to the healthcare workforce shortage. However, little is known about whether such policies generate sizable gains in labor supply or whether the gains are persistent. I tackle these questions in the context of medical licenses for physicians in the US during 1960-1970, where soon after the introduction of Medicare and Medicaid, many state medical boards removed naturalization and visa requirements for foreign-born physicians to obtain license. Using a difference-in-differences identification strategy and a newly digitized individual-level data of all the medical degree owners (MDs) in the US spanning five years between 1966 to 1981, I examine the short and long-run impact of this policy on the stock and flow of local physician market, and the licensing status, full-time employment, and retention of individual physicians. County-level findings suggest removing naturalization requirement increases the number of recently licensed foreign physicians in a county but did not change the total number of foreign physicians, indicating an immediate take-up in licensing but no changes in the total size of the physician workforce. I find no evidence that the policy crowds out the licensing of domestic physicians. Individual-level analysis shows the policy accelerates foreign physicians' licensing process in the short run but have no impact on their full-time employment in the long-run. Heterogeneity analysis suggests the effect is driven by European physicians but not Asian physicians, indicating the naturalization mainly precludes Europeans from being licensed. This paper sheds light on the importance of naturalization in labor market institutions for high-skill immigrants.Does Access Mean Success? Connection to Policy-Makers and Lobbying Success of Political Actors (P0, D7)
Abstract
This article investigates the relationship between direct access to policymakers and lobbying success. I collect large-scale, unique textual data to capture the content of lobbying activities and track subsequent changes in 1,041 European Union regulations, from the draft to the final adopted version. I build two alternative measures to identify lobbying success of comments written by interest groups on a draft regulation: one based on plagiarism detection and the other from a large language model. I measure direct access to policymakers from meetings held between the executive and interest groups. I find that access to policymakers is associated with a 22 to 29 percent higher likelihood of lobbying success, using a balanced sample of comments from organizations with and without access to policymakers. This effect is stronger for comments from organizations with more meetings, or with access to top-level officials. Finally, I exploit variation in meeting timing and policymakers turnover across mandates to explore the underlying mechanisms. I find evidence that political connections are the primary driver of the effect of access on lobbying success, outweighing the influence of information transmission, institutional knowledge, or the intrinsic quality of the organization.Draining the Well: Bank Lending in the Era of Quantitative Tightening (E5)
Abstract
We use two years of Quantitative Tightening (QT) in post-Covid UK and confidential loan-level data to test conflicting predictions about the impact of QT on credit supply. We present evidence that QT reduces lending through a liquidity self-insurance motive. When a bank's reserve holdings (self-insurance) fall due to QT, its lending declines relative to that of other banks. This reflects a threefold adjustment in banks' balance sheets. On the asset side, banks rebalance away from loans towards lower-yielding but more liquid securities. On the liabilities side, banks shift away from wholesale deposits towards more stable, albeit costlier, longer-term funding sources. However, within asset classes with the same liquidity risk, banks reallocate towards higher-yielding assets, thereby increasing interest rate risk. The effects of QT are symmetrically opposite to those observed during Quantitative Easing (QE). Consequently, QE does not necessarily result in a ratcheting up in liquidity risk-taking.Economic Detox, Recession or Stagflation? A New Approach (E3, E5)
Abstract
Talks of “economic detox,” recession and stagflations are getting louder by the day. Consequently, the economic policy uncertainty index has hit the highest level in the post-pandemic era. Predicting detox (or slowdown), recessions and stagflation are vital for decisionmakers, as a distinct set of monetary policy actions would be needed to fight stagflation than those policy stances which are adopted during economic slowdowns.This study develops a new framework to predict probability of stagflation, recession, and soft-landing. We believe our proposed framework would help decisionmakers set an appropriate policy stance by incorporating probabilities of the three scenarios as potential outlooks of the economy in their existing decision-making toolkit.
First, we characterize episodes of soft-landings and stagflation for the U.S. economy. Our work suggests that there are thirteen episodes of stagflation and thirteen periods of soft landings in the post-1955 era. The NBER suggested that there are ten recessions during the same period.
Second, we build an ordered Probit framework to generate one-year out simultaneous probability of stagflation, recession, and soft-landing. In a simulated real-time out-of-sample analysis, our framework accurately predicted all three scenarios in the post-1976 period. Moreover, the persistently higher probabilities of all three growth scenarios of the past few quarters may cloud the near-term policy path.
Furthermore, our framework accurately predicted episodes of policy pivots in the post-1990. In our view, accurately predicting periods of monetary policy pivots is vital, as a rate cut that comes too soon or too late would be harmful to the economy and damage the FOMC's reputation. In conclusion, using our framework to predict potential risks to the economic outlook and utilizing those probabilities to forecast policy pivots as well as the pace of adjustments to the policy stance would help decision makers to design effective policies.
Efficiency and Game Theory Models for Cross-border Bank Resolution (K2)
Abstract
Bank resolution plays a central role in global financial stability because, without effective resolution regime, bank failures may cause systemic risk and the public will lose confidence in the banking industry. In the past, many governments spent substantial resources to resolve banks in crises. Banks should be resolved with minimised costs from taxpayers and society. Cross-border bank resolution must improve and become more efficient.However, bank authorities struggle to achieve higher efficiency. This research defines productive efficiency of cross-border bank resolution as a situation when the cross-border bank resolution costs are minimised to preserve the continuation of bank services and critical functions. This paper argues that greater cooperation can improve the productive efficiency of cross-border bank resolution, because it can reduce resolution input from the perspectives of time cost, asset discount and externalities, and improve resolution output in terms of preserving the continuation of the critical functions and services of bank.
This research then designs three original Game Theory models to represent cross-border bank resolution scenarios. The graphic and quantifiable Game Theory models might help authorities to make efficient decisions quickly through negotiation and binding arrangements. Additionally, the Game Theory models could illustrate the importance of cooperation and the results of this research might help authorities to adopt more cooperative measures in future cross-border bank resolution.
Based on Game Theory analysis, this paper recommends how to improve cross-border cooperation through more enforceable cooperative agreements and mechanisms. This research proposes that authorities should introduce ex-post review and resolvability rating mechanism for countries, supplemented by monetary sanctions and reputational sanctions. Monetary sanctions may include confiscation of collaterals and increased contribution to a common deposit insurance scheme or a common resolution fund. The Importance of a country’s reputation in international banking market can be a strong incentive for them to cooperate.
Encouraging Online Knowledge Contributions - Evidence from a Field Experiment (M0, D9)
Abstract
With the prominence of user-generated content platforms, online knowledge platforms have experienced substantial growth. However, it is unclear why individuals voluntarily contribute knowledge, and how platform strategies can facilitate individuals’ contributions through non-monetary incentives. I develop a stylized model of individual knowledge contributions where individuals have social motivation to gain reputation and instrumental motivation to obtain functionality privileges on the platform. Based on the model, I design and implement a large-scale field experiment involving 12,182 individuals on one of the largest online question-and-answer platforms. I sample and manually treat participating platform individuals daily over the course of four and a half months. The treatment gives one anonymous upvote to eligible answers, exogenously shifting individuals’ social and instrumental motivations. I then track comprehensive data on individuals’ subsequent behavior on the platform daily for four months. I find that the treatment significantly increases an individual’s probability of contributing additional answers by around 15% of the baseline, and the difference between the control and treatment groups persists over time. The treatment effect is the most pronounced for individuals with low-to-moderate answering experience or reputation and is slightly stronger for those who are close to obtaining additional privileges after the treatment. The overall quality and effort of future answers remain stable. Using data from the field experiment, I structurally estimate the model of contribution decisions to quantify the relative importance of social and instrumental motivation. Simulation results suggest that social motivation is more important, and platform strategies that boost social motivation are more effective in encouraging contributions.Estimating the Impact of Public Policy with Unobserved Variation (H0, J2)
Abstract
In recent years, many sub-state jurisdictions in the United States have introduced local economic policies previous only seen at the state or federal level. Because most publicly available data does not identify individuals at these local levels, estimating the impact of these policies is difficult as local variation will bias state-level estimates. In this paper I propose a solution that combines an intention-to-treat approach with two-sample IV, where aggregates are sufficient for the first-stage estimation, and that identifies the local average treatment effect (LATE) of the policy. Using the recent prevalence of local minimum wage changes as a motivating case, I show this method provides statistically distinct results, when compared to standard methods, though differences are economically small and estimates are qualitatively similar to previous studies.Extreme Temperatures and Energy Consumption in Bangladesh (Q5, O1)
Abstract
While extensive research has been conducted on the relationship between temperature and energy consumption in high-income countries, much less is known about these dynamics in poorer regions. This study addresses that gap by analyzing energy expenditure patterns in rural Bangladesh, where households depend on a mix of fuel types ranging from solid fuels to electricity. The study utilizes data from the Bangladesh Integrated Household Survey (BIHS), a nationally representative dataset spanning three rounds (2011–2012, 2015, and 2018–2019), and integrates it with temperature and precipitation data from the ERA5 reanalysis dataset, prepared by the European Centre for Medium-Range Weather Forecasts (ECMWF). By employing a temperature bin approach, the study estimates the effects of extreme temperatures on household energy expenditures in a flexible manner. The econometric strategy controls for village and year-fixed effects, ensuring that variations in energy expenditures are mainly driven by temperature fluctuations rather than broader economic or infrastructural factors. The analysis reveals distinct patterns of energy consumption: expenditures on solid fuels increase significantly during colder periods, while electricity expenditures decline in hotter temperatures. The robustness of these findings is confirmed through multiple econometric specifications, including degree days and quadratic models. These results suggest that rural households encounter severe constraints in adjusting their energy usage, which may stem from affordability, lack of infrastructure, or challenges in behavioral adaptation. This conclusion is crucial, given that this inability to adapt to temperature extremes can have serious consequences for household well-being, including increased health risks and mortality.ywordsFactor-Biased Efficiency Gains from Exporting: Evidence from Colombia (L6, O3)
Abstract
This study examines whether exporting enhances efficiency and favors specific inputs. We develop a production function model within a dynamic exporting and investment framework, capturing factor-biased technical changes. Using Kalman filtering to address measurement error and propensity score matching to control for self-selection into exporting, we analyze Colombia’s manufacturing sectors from 1981 to 1991. New exporters achieve a 4% annual increase in labor-augmenting and unskilled labor relative productivity, with no change in Hicks-neutral productivity. Unskilled labor-augmenting productivity grows by 8% annually, aligning with machinery asset expansion, while TFP rises by 3% per year.Family Matters: How Globalization Reshapes Firm Management and Productivity (F1, O1)
Abstract
How does globalization affect firm management and productivity? I investigate this question using a product-specific import competition shock in India, focusing on family-managed firms-- the predominant form of corporate governance in the developing world. Utilizing a novel manager-firm matched dataset, I analyze tenure records and family ties of over 6 million company directors. Employing an event-study approach, I find that firms exposed to import competition replace family managers with unrelated professional executives, increasing firm productivity by 20 percent. To evaluate the aggregate implications, I construct a model of industrial equilibrium where family firms balance non-pecuniary private benefits of family management against higher profits achievable through professional management. Consistent with empirical findings, the model predicts that import competition drives the least productive family firms to adopt professional management to avoid exit. These results underscore managerial restructuring as a critical mechanism linking competition to productivity gains.Firm-Age Markup Dynamics: Facts, Model, and Policy Implications (E0, L1)
Abstract
We utilize comprehensive administrative and commercial datasets from the UnitedKingdom and Finland covering the firm population to explore the nature of firm markup
growth as firms age. Our analysis reveals that markups increase with firm age, and
the growth and higher-order moments of the firm-markup distribution can be effectively modeled using a mixed-normal distribution, as applied by Guvenen, Karahan,
Ozkan, and Song (2021) to the life-cycle earnings of workers. We then develop a firm
dynamics model with heterogeneous firm markups based on the Homothetic Single
Aggregator’s well-defined properties (Matsuyama, 2025). We fit the model to the data
of the two economies to analyze how the firm entry, exit and markup dynamics shape
the firm-age-markup distribution. As markups entail social costs, the findings suggest
that the nature of markup dynamics and the initial heterogeneity (Sterk, Sedlacek, &
Pugsley, 2021) and growth play a critical role in determining the optimal design of industrial policies, with key differences observed across countries. This study advances
the literature on the social costs of markup (Baqaee, Farhi, & Sangani, 2023; Edmond,
Midrigan, & Xu, 2023) heterogeneity, offering new insights for policymakers to design
context-sensitive interventions.
Baqaee, D. R., Farhi, E., & Sangani, K. (2023). The darwinian returns to scale. The Review
of Economic Studies, 91(3), 1373-1405.
Edmond, C., Midrigan, V., & Xu, D. Y. (2023). How costly are markups? Journal of Political
Economy, 131(7), 1619-1675.
Guvenen, F., Karahan, F., Ozkan, S., & Song, J. (2021). What do data on millions of u.s.
workers reveal about lifecycle earnings dynamics? Econometrica, 89(5), 2303-2339.
Matsuyama, K. (2025). Homothetic non-ces demand systems with applications to monop-
olistic competition. Annual Review of Economics, forthcoming.
Sterk, V., Sedlacek, P., & Pugsley, B. (2021, February). The nature of firm growth. American
Economic Review, 111(2), 547–79.
From Text to Verdict: Predicting IP Litigation Outcomes with Machine Learning (O3, G2)
Abstract
This study evaluates the effectiveness of various machine learning models in predicting the outcomes of intellectual property (IP) litigation, leveraging data from China Judgement Online, which contains over 140 million lawsuits spanning 2010 to 2021. We hypothesize that the analysis of textual content should be contextualized by the company characteristics. Our findings indicate that the Structural Topic Model (STM), which integrates both financial and textual data, significantly enhances the accuracy of predictive models compared to those relying on a single input type. The results reveal critical financial implications of IP litigation. Specifically, the STM model exhibits strong predictive performance across diverse cases involving both plaintiffs and defendants, as well as different IP types such as copyrights, trademarks, and patents. Furthermore, winning IP litigation is significantly associated with a plaintiff’s return on assets (ROA) and higher levels of innovation. These findings underscore the importance of robust legal protection for IP in driving innovation and bolstering financial performance.Gender Disparities in Divorce Laws: Incentivizing Women’s Education (J1, I2)
Abstract
Redistributive policies within households, such as property division upon divorce, can affect marital gains differently for women and men. Despite the extensive literature on the effects of divorce laws and property division rules on married couples' behavior, little empirical work exists about how forward-looking single women respond to changes in marital returns. In 2011, China changed the property division upon divorce from an equal to a title-based regime for housing property in certain conditions, leading to some women losing intrahousehold property ownership. In this paper, I explore how single females choose education as a self-insurance against the expected decrease in the share of the surplus they can extract in the marriage. I develop a two-period model to illustrate how changes in property division rules influence single women's incentives to invest in education. Data supports the model's prediction that reduced marital gains in the second period motivate women to increase educational investment in the first period. I employ a difference-in-differences framework to quantify the heterogeneous impact of the national property division reform, leveraging geographic variation in property ownership by gender. The results demonstrate that education is an insurance mechanism, mitigating the costs of an unwanted divorce. In response to the relatively higher property owned by males, more single women enroll in college before marriage, particularly in four-year college programs. Overall impacts of the policy, aggregated from the dosage difference-in-differences, further support the insurance mechanism. This paper highlights how shifts in the marriage market drive female educational investment, offering insight into why the gender gap in college enrollment has widened in favor of women in China since 2010. It also reveals an unintended consequence of property division reforms: their impact on the human capital accumulation of future generations—an important consideration for policymakers when designing redistributive household policies.Gender Inequality in Smoking: The Impact of West German Television on East Germany (I1, P3)
Abstract
In this paper, we estimate the rise in smoking prevalence in East Germany caused by exposure to West German TV, with a focus on gender differences. We leverage survey data on smoking prevalence from 1989 (before the fall of the Berlin Wall) and 2002 (twelve years following reunification) and exploit exogenous variation in exposure to the West German TV signal to identify the causal effect. We document that exposure to West German TV increases smoking prevalence, but this rise is concentrated entirely among women, with a negligible impact on men. Our estimates imply a 10.7 percentage point increase in the probability of smoking among women and a 68% increase in the number of cigarettes smoked. These findings are robust across various tests and suggest that the convergence in smoking rates between men and women will likely lead to an increase in women's mortality and healthcare costs.Gender Quotas in Academia: Evidence from South Korea's Affirmative Action Policy (J7, J2)
Abstract
This study investigates a unique incentive-based affirmative action program in South Korea’s academic labor market. Specifically, I analyze the effects of a 2018 policy requiring all national and public universities in South Korea to ensure that neither gender constitutes less than 25% of tenure-track or tenured faculty at the university level, with noncompliance resulting in reductions in government administrative and financial support. Using department-level faculty composition data for the entire universe of Korean universities and a difference-in-differences methodology, I find that the implementation of gender quotas is associated with an increased share of female new hires in national and public universities. Furthermore, using web-scraped individual-level data on tenure-track and tenured professors, I explore the differential impact of gender quotas on entry barriers for newly hired professors by gender, focusing on research performance. Analysis of entry barrier-related outcomes is still in progress.Geopolitical Risks and Foreign Institutional Investors: Evidence from the Taiwan Stock Market (F5, P0)
Abstract
In this study, we examine the impact of geopolitical risks on the trading behavior of foreign institutional investors in the Taiwan stock market during the outbreak of the Russian-Ukrainian War. Defining firms with foreign ownership in the top (bottom) 30% as the treated (control) firms, we use a difference-in-difference analysis to test this issue. Our findings demonstrate that treated firms suffered larger losses in stock return compared to control firms after the war outbreaks. Moreover, treated firms are associated with selling pressure, downside risk, and turnover than control firms after the shocks. This effect is also stronger for treated firms with lower operating performances, higher volatility, and higher market liquidity. Overall, we find that treated firms suffered more after the war outbreaks due to the increasing threat of geopolitical risks.Global Investment and Local Perceptions Amid Great Power Competition: Evidence from Chile (F2, O5)
Abstract
How does Chinese FDI shape public opinion in Latin America amid U.S.–China competition? Using Chile as a case, we examine the causal effect of Chinese investment on local attitudes toward both China and the United States.Applying a Synthetic Difference-in-Differences (SDID) model, we estimate the impact of Chinese FDI across Chilean provinces from 2002 to 2018. Combining georeferenced Latinobarómetro survey data with project-level FDI records, household census data, electoral outcomes, and hand-collected interviews, we construct a novel, spatially matched panel dataset. We find highly heterogeneous effects on public sentiment toward China. Notably, private-sector agricultural investments—despite lower investment volume—led to significant improvements in local attitudes, likely due to fewer legal conflicts, lower environmental impact, and strong ties between farmers and communities. In contrast, mining and energy investments had no significant effect, despite narratives of resource nationalism. Additionally, provinces with lower poverty rates, larger populations, and more left-leaning governing parties exhibit more favorable attitudes toward China, regardless of FDI presence.
Beyond China, we examine whether Chinese FDI crowds out U.S. influence. Contrary to the common zero-sum narrative, we find no evidence that Chinese investment reduces pro-U.S. sentiment. Instead, local perceptions of China and the U.S. are positively correlated, suggesting that economic engagement with China does not come at the expense of U.S. favorability.
This study makes four contributions. First, it offers the first project-level causal evaluation of Chinese FDI’s political effects in Latin America, addressing a critical gap in the literature. Second, it applies SDID to international investment analysis, demonstrating its advantage combining traditional Synthetic Control and Difference-in-Difference methods. Third, it introduces a novel dataset on the political affiliations of Chilean communes, compiled through the author’s face-to-face interviews with Chilean high-rank politicians. Finally, it highlights sectoral heterogeneity in FDI effects, offering insights for policymakers navigating economic development and geopolitical alignment.
God, Save the Tsar! How Churches Boost Electoral Support Beyond Economy and Fear (N3, Z1)
Abstract
The role of religion in shaping political behavior remains theoretically underdeveloped (Becker et al., 2024; Rubin, 2025) especially in authoritarian setting (Campante & Yanagizawa-Drott, 2015; Montero & Yang, 2022). We study how religious infrastructure affects formal electoral outcomes using data on several thousand of Orthodox churches constructed between 2012 and 2018. We employ a difference-in-differences design (Borusyak, Jaravel, & Spiess, 2024), comparing municipalities that received new churches to those that did not, and estimate the causal effect of construction on formal voting behavior.Our findings indicate that the erection of a church leads to a 2–5 percentage point increase in the vote share for incumbent president Vladimir Putin, while having no discernible impact on voter turnout. This implies that the effect is likely driven not by changes in political mobilization but rather by shifts in the voters’ preferences. The result is robust to multiple theoretically-justified model specifications and matching techniques.
We next evaluate if the result can be explained by and rule out two prominent mechanisms. First, in contrast to Gruber’s (2005) result, we show that the effect cannot be attributed to local economic stimulation associated with church construction, as we find no change in employment, infrastructure, or investment indicators. Second, we find no support for the “projecting strength” mechanism common to autocracies, whereby costly, visible projects serve to intimidate or signal dominance (Egorov & Sonin, 2024). To disentangle the latter, we leverage on the fact that a significant fraction of the recently erected churches we built in restricted locations with no access for the general public such as military bases or prisons.
Taken together, our results nuance the role of religion in sustaining political order, suggesting that its influence in authoritarian regimes operates through less visible and less strategic channels than commonly assumed.
Harnessing Machine Learning for Price Index Construction in Trade Using Customs Data (C4, F1)
Abstract
In index number theory, unit values have long faced skepticism for their use in broadly defined product categories, yet recent scholarship highlights their effectiveness when applied to narrowly specified products, accurately capturing price movements. This study introduces an innovative method to construct price indices by exploiting the detailed granularity of administrative customs transaction data, which includes precise product attributes and transaction details, offering a robust alternative to conventional survey-based approaches.We utilize Gaussian Mixture Models (GMM), a probabilistic clustering technique, to group transaction IDs—unique identifiers derived from attributes such as country of origin, sender, importer, recipient, and order size—based on their price movement vectors. This clustering adheres to Hicks’ composite commodity theorem, enabling the aggregation of commodities with stable relative prices into a single unit for index construction. GMM’s probabilistic framework optimizes cluster allocations, ensuring theoretical consistency and temporal stability by penalizing excessive reassignments of transaction IDs across clusters.
The method is tested using Switzerland’s customs database, enriched with importer-declared free text descriptions and categorical variables. Fruit and vegetable products —heterogeneous, seasonal, and vital to food security policy yet often overlooked in trade models— provide an ideal case study. We assess the approach by evaluating relative price stability within clusters, the interpretability of product groupings, the confidence of cluster assignments, and the preservation of importer data confidentiality.
If successful, this method offers a cost-effective, representative alternative to survey-based indices, leveraging customs data’s richness. This advancement promises to enhance economic measurement and inform policy analysis in international trade.
Health IT Diffusion and Physician Density (O3, I1)
Abstract
This paper examines how the diffusion of advanced health information technology (HIT) affects the aggregate supply of hospital-based (HB) physicians, who deliver direct patient care under hospital contracts. Leveraging sharp increases in county-level HIT adoption rates driven by federal incentives, we compare physician supply per 100k population in counties with rapid diffusion (treatment group) to those with slower or no uptake during our sample period (control group). We construct a novel dataset from multiple sources, including Area Health Resources Files, HIMSS Analytics Database for HIT adoption, CMS Medicare Provider and Service Data, and Healthcare Cost Report Information System from 2006-2018, with one-year lagged measures of HIT adoption. We employ an event-study approach following East et al. (AER 2023) to identify treatment and control groups based on sharp increases in adoption rates, and we use the staggered difference-in-differences (DiD) estimator proposed by De Chaisemartin and d'Haultfoeuille (2024) to address potential biases in traditional two-way fixed effects models.We find that HIT diffusion led to a 10.3% increase in HB physician rate in treated counties relative to control counties, with medical and surgical specialties accounting for most of the increase. This growth is further concentrated among early-career physicians and in physician shortage areas. Mechanism tests suggest that physicians benefit financially from practicing in treated counties, with higher Medicare reimbursement, more Medicare patients, and increases in hospital profits. Counties with moderate pre-period surgical volumes experienced the largest increases in physician supply and outpatient surgeries, suggesting greater capacity for demand expansion post-treatment. Our models control extensively for potential confounders, including county-level market concentration and healthcare consolidation. Various robustness checks, including alternative DiD estimators and sample restrictions, support the validity of our results. Our findings suggest that strategic HIT investments can attract physicians, expand care capacity, and reduce geographic disparities in access to health care.
Hiding Lemons among Peaches: Optimal Retention and Promotion Policy Design (J3, M5)
Abstract
How should an employer design its retention and promotion policy when there is competition for its employees? In an environment where the incumbent employer learns more about its workers than other potential employers, we show that the incumbent’s optimal policy always over-retains workers inefficiently retaining workers who would be more productive elsewhere to dampen the positive signalling effect of retention. In contrast, the optimal policy may over- or under-promote workers. The incumbent’s incentive to distort its policy is driven by the technologies of competing firms in the labour market. We also demonstrate the extent to which the firm can implement the optimal policy without the ability to commit and study the incentive for the firm to manipulate the signalling effect via designing jobs. Our results shed light on the role of (possibly vacuous) job titles and provide a novel rationale for the Peter Principle.Higher Education Policies and Spatial Inequality Across US States (I2, R1)
Abstract
To what extent have the federal and state higher education policies—which have affected the supply and demand of college education—contributed to the growing cross-state wage inequality since 1980? To address this question, we develop and estimate a quantitative spatial model with endogenous college education and migration. The model features skill-biased technical changes and agglomeration effects, which means that a state’s average wage rises as more residents attain college degrees or as college-educated workers migrate into the state. Counterfactual analyses reveal that tuition discounts and public appropriations by the federal and state governments have become more progressive. In the baseline, the standard deviation of log mean wages across states increased from 0.087 in 1980 to 0.117 in 2019. In contrast, had the higher education policies remained at their 1980 levels, the standard deviation would have risen by 73% more to 0.139 by 2019. Agglomeration effects played a significant role in this result: if they had been half as strong as estimated, the standard deviation would have increased by 17% more (instead of 73%) in the counterfactual than in the baseline.Household Bargaining and Fertility Decisions Under China’s Two-Child Policy (J0, I0)
Abstract
This paper studies how the introduction of China's Two-Child Policy affected intrahousehold resource allocation and female bargaining power. Using micro-level data from the China Family Panel Studies (CFPS), I estimate a structural household demand model based on the collective framework developed by Browning, Chiappori, and Lewbel (2013). My results indicate that the Two-Child Policy substantially reduced women's resource shares within households from an average of 38% to 25%—a decline of approximately one-third. This deterioration was particularly pronounced among families with firstborn daughters, consistent with intensified son-seeking behavior following the policy reform. The policy also weakened the relationship between women's relative income and their bargaining power, suggesting that traditional economic levers of household influence were undermined. Complementary reduced-form evidence reveals significant increases in fertility among previously constrained households and a corresponding rise in women's housework burden. These findings reveal a troubling paradox: fertility liberalization intended to expand reproductive choice may actually worsen gender inequalities within households unless accompanied by targeted complementary policies.How Genetically Engineered Crops Influence U.S. Crop Yields, Climate Sensitivity and Production Locations (Q1, Q5)
Abstract
Climate change is reshaping agricultural production across the United States. Rising temperatures, shifting precipitation patterns, and increasing climate variability are altering crop suitability across regions, prompting farmers to adapt through changes in regional crop mix. While prior research has identified climate as a key driver of these spatial shifts, other factors—including technological innovation—also contribute. In particular, the widespread adoption of genetically engineered (GE) crops may be changing the sensitivity of agricultural systems to climate stress.This paper investigates how GE seed adoption influences both crop yield outcomes and land use decisions under changing climate conditions. We develop an empirical framework that analyzes the intensive (yield-enhancing) and extensive (land-expanding) margins of GE adoption. Using U.S. county-level data on corn, soybean, and cotton from 1974-2020, we estimate (1) effects of GE adoption and climate variables on mean crop yield and crop yield variability and (2) climate effects on crop mix and regional crop migration.
We find that GE adoption significantly increases yields and reduces yield variability across crops. For instance, a 10 percent increase in GE corn adoption is associated with a 2.1% increase in corn yield and a statistically significant reduction in crop yield variance. On the extensive margin, a 1 percentage point increase in GE corn adoption raises corn land use share by 0.029 percentage points and dampens the observed northward movement in the incidence of corn acres. Interaction terms indicate that GE adoption allows crops to more successfully be produced in their current regions. Simulations show that GE adoption slows the climate-induced northward migration of corn and soybean, moderating the pace and direction of change.
These findings offer new insight into how biotechnology is another tool in the quest for climate adaptation and carries implications for crop mix, sustainability, and long-term agricultural policy design plus research and development thrusts.
How Interest Rate Swaps Reshape the Yield Curve: Evidence from China's Derivative Market Liberalizing (G0, F3)
Abstract
This paper provides the first causal evidence on how interest rate derivatives reshape the yield curve. Exploiting China’s recent derivative market liberalization, Swap Connect, and the specific demand of foreign investors in China’s bond market as an identification strategy, I show that the introduction of interest rate derivatives contributes to a flattening of the yield curve. Using daily bond trading data, I find that this policy, by enhancing foreign investors' ability to manage interest rate risk, reduces the term spread between 10-year and 3-month Chinese government bonds by 15 basis points, an 11% decline from its initial level. These findings reveal the previously underappreciated role of interest rate derivatives in shaping the yield curve and highlight the broader importance of derivative markets in the financial system.How Many Years Does It Take for AI Adopting Firms to Realize Productivity Effects? (O3, L2)
Abstract
Despite recent rapid progress in AI technologies, we are not seeing a surge in economic growth, yet. It is well known that a fundamental technological change, such as AI, takes time to realize its productivity effect. Then, how many years does it take for the realization? This paper investigates the lagged effects of AI adoption among South Korean firms by implementing an event analysis utilizing a panel survey.I found adoption of AI had no significant contemporaneous effects on firm productivity. AI adopting firms are significantly larger and more productive than non-adopting firms in South Korea, but such a productivity differential is declining over time and becomes insignificant as firm characteristics, such as firm size, age or intangible investments, are controlled for.
Since the year of first adoption, the labor productivity of AI adopting firms did not rise significantly faster than that of non-adopting firms over time, and remained insignificant up to 5 years after. Adopting firms’ output and employment level grew significantly faster, but they were already faster growing firms even before AI adoption.
It is also tested whether AI adoption is associated with a surge in complementary innovation activities. New business entry significantly increased since AI adoption, but R&D and intangible investment didn’t. Other types of innovation activities, including patents and trademarks, did not rise significantly either. Considering recent attention to AI technologies, such a lack of acceleration in innovation activities is rather puzzling and could have been a reason for the lack of productivity effect.
How Wealth and Age Interact to Affect Entrepreneurship (J1, D2)
Abstract
A large literature shows that entrepreneurial activity is driving economic growth and job creation. Examining the interaction of wealth and age on entrepreneurship is important because the population is aging in many countries, with ever larger fractions of the population reaching advanced age. In addition, on average, wealth is concentrated among older individuals. Despite these trends, little is known about the different effects of additional wealth on business activity of old versus young individuals.To examine this issue, we use comprehensive matched employer-employee tax record data and variation in lottery prize amounts as a plausibly exogenous increase in individuals' wealth. The lottery data consist of all lottery wins, provided by the lottery corporation of a Canadian province. For each of these lottery wins, we match the lottery winners to their tax-based administrative data, which include an employer-employee matched dataset. These administrative data allow us to observe whether winners choose to enter or exit business ownership, how their businesses fare, how much labor supply they provide (as measured by wage income), and whether they choose to exit wage labor. We implement a stacked difference-in-differences (DID) empirical methodology to assess the relative effects of an additional dollar of wealth on individuals' economic decisions.
Our main finding is that additional wealth leads older individuals (aged 55 and above) to reduce ownership of corporations. Further, with additional wealth, older individuals reduce their wage labor income and the number of jobs they have. These individuals also exit both wage labor supply and firm ownership, which is consistent with them retiring. In contrast to our findings on older individuals, younger individuals (aged 55 and under) increase their entrepreneurial activities with extra wealth and decrease wage labor supply. Thus, additional wealth allows them to transition from wage labor to entrepreneurship.
Identification of Treatment Effect Heterogeneity Using Prognostic Variables (C1, C2)
Abstract
A key challenge in causal inference is that only one potential outcome—either treated or untreated—can be observed for each unit. This one-sided nature of observations makes it infeasible to analyze policy-relevant causal measures of treatment effect heterogeneity, such as the distribution of treatment effects, without imposing strong structural assumptions. I propose a new identification strategy that recovers the joint distribution of treated and untreated potential outcomes by leveraging variation in prognostic variables—variables that predict untreated outcomes and do not directly affect treatment effects. The proposed framework identifies the joint distribution by solving a system of equations implied by pairs of marginal distributions of treated and untreated potential outcomes whose relationship is invariant to the prognostic variables. As a corollary, the distribution of treatment effects is identified. In an empirical application, I revisit a lending experiment in Egypt to analyze the program's trade-offs by examining the proportion of firms that potentially gain or lose in terms of business growth due to increased access to credit.Increasing the Statistical Power of Correlation-Based Likert Scale Data Analysis Using a Compositional Data Approach (M0, C1)
Abstract
Likert scales are one of the most widely used tools in economic research and other fields involving psychological constructs, such as attitudes, opinions, perceptions, traits and states. Their popularity stems from their versatility in capturing both the direction and intensity of individuals’ responses across various contexts. They enable researchers to detect subtle shifts in attitudes that binary or open-ended questions cannot reveal. However, detecting reliable correlations between psychological constructs often requires large samples, creating challenges for researchers facing budget constraints or targeting niche populations. This paper investigates how a compositional data approach using isometric log-ratio (ilr) transformation can increase the statistical power of correlation tests when analyzing Likert scale data. Building on previous research demonstrating the effectiveness of this approach with normal, heavy-tailed, skewed, and bimodal distributions, we extend these findings to platykurtic and leptokurtic distributions, both of which are common in economic (big) data. Through a series of simulations, we demonstrate that the ilr approach universally enhances statistical power across all unimodal distributional forms, particularly for small to moderate correlations (≤ 0.4). This approach substantially reduces the required sample sizes, with savings (depending on the effect size) ranging from only a few to more than 100 participants while maintaining similar or even larger statistical power. This efficiency is achieved by treating Likert scale data as compositional data, as these scales inherently capture both the level of agreement and disagreement towards an item assertion. The approach is particularly valuable for scales with small item numbers (≤ 30) where the normal approximation of item means fails. Our findings offer researchers a practical, cost-effective tool to enhance statistical power without increasing sample sizes. This benefits both academic research and commercial research where time and resource constraints are important considerations.Increasing Wealth Inequality Under Capitalism (P1, F5)
Abstract
This paper explores whether wealth concentration in capitalist countries is increasing as Karl Marx, Thomas Piketty, and many scholars warned. We argue that capitalist countries experience wealth concentration through utilizing the OECD countries’ data. The OECD countries’ top 1% share ratio increased 2.44% from 1995 to 2022 with a 99% confidence level according to our t test’s result. Also, the average top 1% share increases every year, and some of them are significant through using panel data analysis. We conclude that capitalism has a property to enhance wealth concentration.Indirect Deterrence Effects from IRS Filing and Payment Compliance Programs (H2, H3)
Abstract
This study investigates how IRS enforcement actions affect taxpayers who were previously compliant, with a focus on sustaining voluntary compliance. We examine three forms of intervention: mailed collection letters to taxpayers with unpaid balances, reminder notices sent to non-filers, and in-person visits by IRS officers to resolve outstanding debts. While these actions are primarily directed at individuals who have fallen behind, they may also deter noncompliance among others through broader signaling effects.To estimate these indirect effects, we implement a two-stage least squares (2SLS) approach. We instrument local enforcement intensity with exogenous variation in IRS staffing levels across years and regions. To capture how enforcement awareness diffuses through social networks, we construct exposure measures weighted by the Social Connectedness Index (SCI), which quantifies the strength of social ties across geographic areas.
Our results indicate that IRS enforcement actions have economically and statistically significant spillover effects. A 10% increase in mailed collection letters reduces new tax delinquencies among previously compliant taxpayers by 16%, equivalent to approximately $3.2 billion. Reminder notices to non-filers also yield meaningful reductions, while in-person visits have more limited indirect reach. These effects are amplified in communities with stronger social ties, suggesting that peer networks reinforce perceived enforcement risk and promote compliance.
This study contributes to the tax compliance literature in four key ways. First, it shifts the focus from correcting noncompliance to sustaining compliance among those already meeting their obligations. Second, it develops a network-based framework to capture how enforcement signals diffuse across socially connected populations. Third, it exploits plausibly exogenous variation in IRS staffing to identify causal effects. Finally, our findings suggest that conventional evaluations of enforcement may understate its full impact by neglecting spillovers to compliant taxpayers. Accounting for these indirect effects is essential for designing effective, resource-constrained enforcement strategies.
Inflation, Market Power, and Markups: Evidence from the Airline Industry (L2, R4)
Abstract
The link between market power and pricing is central to industrial organization research, yet the effects of macroeconomic conditions on this relationship remain understudied. Estimating a structural model of airline demand and supply, this work finds that inflation reduces consumer price sensitivity and allows airlines to exploit this reduced sensitivity to increase product-level markups, particularly in more concentrated markets. A one percentage point increase in inflation is associated with a 2.72 percentage point increase in product-level markups in the full sample. This effect is substantially larger in concentrated markets, where the markup response to inflation rises to 3.41 percentage points, compared to 2.19 percentage points in more competitive markets. These findings demonstrate how firms can leverage inflation-induced changes in consumer behavior to enhance profitability, and how market concentration amplifies inflationary pressures by enabling firms to implement larger price increases than would be possible in more competitive environments.Infrastructural Shocks and University Success in the United States (O3, O2)
Abstract
Does educational infrastructural resources lead to long-run university success? Using newly digitized data on historical college enrollments, the spatial distribution of World War II military surplus, and institutional-level research output, I leverage quasi-random variation in access to military property surpluses from the postwar demobilization to estimate the impact of infrastructural inputs on enrollment and other institutional performance metrics. Constructing a balanced panel of colleges spanning each decade from the 1930s to the 2010s, I find that exposure to military real property surplus enabled U.S. institutions, particularly public universities, to expand enrollments in the postwar years and boost research output in the long run, with the recruitment of senior scientists emerging as a key mechanism. These findings underscore the enduring benefits of investing in higher education infrastructure.Integrating Asset Pricing with the Information Market (G1, E7)
Abstract
We study asset prices and portfolio choices in an exchange economy with an information market. Investors make information choices to trade in financial markets and information producers decide the quality to compete for market share and maximize profits. Our model explicitly solves for the equilibrium number of information producers and the level of information quality. In our baseline model with homogeneous producers, we show that technological advancements increase the number of producers and the quality levels, while reducing market disagreement. We explore three extensions to this model. First, when a spillover effect exists in information production, it weakens competition but has an ambiguous impact on equilibrium quality. Second, with heterogeneous producers in terms of technology, non-experts pursue non-quality competition to differentiate themselves from experts, leading to a more concentrated market. Finally, when investors can learn from both public and private signals, technological improvements increase market disagreement and can exacerbate consumption misallocation.Intergenerational Mobility over Nine Generations: Evidence from Poland, 1800-1984 (N3, J6)
Abstract
In this paper, we utilize the mass genealogical data to measure social mobility in Poland over the last two centuries. To do so we digitized the unique dataset of elite biographies and imputed the nine generations of the Polish elite to the Poland’s largest genealogical database. We measure direct family links between the top 0.01% of the Polish society across nine generations. We find that intergenerational mobility was low and stagnant until WW I, then it gradually increased in the 20th century, especially after WW II. In the 19th century, 35-39% of the elite were direct descendants of the elite of the previous generation and 50% of the elite retained the social status for their children. In 1984, descendants of the previous generation of the elite still formed 17% of the current elite. Family formation strongly influenced the persistence of social status, especially in the case of outsiders, joining the elite as the first in their families. Our results also show that religious institutions were an important engine of social mobility.Intra-Household Bargaining, Endogenous Divorce, and Female Labor Supply Over the Life Cycle (J2, H3)
Abstract
This paper examines the effects of intra-household bargaining and endogenous divorce on female labor force participation over the life cycle. I develop a structural life-cycle model with a limited commitment framework, incorporating household heterogeneity in education, non-labor income, and female bargaining power. The model features gender-education-specific marital preference shocks, making divorce risk an additional source of uncertainty alongside earnings risk.Calibrated to Japanese survey data, the benchmark model closely replicates male life-cycle earnings profiles but exhibits deviations for females, highlighting potential mechanisms driving gender disparities. To address this, I extend the model to incorporate social norms influencing the gender division of housework and the career penalties associated with motherhood due to increasing returns to experience.
I apply this framework to evaluate the impact of alternative taxation systems on household labor supply and welfare. In particular, I show that progressive income taxation at the household level imposes higher marginal tax rates on secondary earners, discouraging female labor force participation. The model aims to inform the design of an optimal tax system that balances welfare maximization with equity considerations.
It’s a Lady’s Business: Investigating the Effects of the Menstrual Cycle on Entrepreneurial Behavior (L2, I1)
Abstract
This paper investigates how menstrual cycles affect entrepreneurial behavior. The menstrual cycle influences the functioning of the hormones in the human body, which in turn impact both psychological and physiological aspects of an individual. Given the majority of women experience menstrual cycles until the start of the menopause, this reality also holds for numerous female entrepreneurs. In this research, we aim to determine how hormone levels in three specific phases of the menstrual cycle affect entrepreneurial behavior, measured in terms of innovativeness, risk-taking, pro-activeness, passion and persistence. During the menstrual phase (or menses), the level of progesterone and estrogen are lowest. In the follicular phase, the level of progesterone remains low, but the amount of estrogen increases until it reaches a maximum level coinciding with ovulation. In the luteal phase, both progesterone and estrogen maintain high levels, with progesterone reaching the highest levels. Our findings suggest that female entrepreneurs (n = 135) show the highest levels of innovativeness during the follicular phase, whilst risk taking, proactiveness and passion are highest during the luteal phase. We find no evidence of differences in the level of persistence across the three phases. By combining knowledge from the field of medicine and biology with entrepreneurship and entrepreneurial psychology, our research adds to the Biological Perspective in Entrepreneurship by investigating the influence of progesterone and estrogen, two hormones that have not been linked to entrepreneurial behavior before. Our findings benefit female entrepreneurs by providing greater insight into the strengths and weaknesses associated with their human nature, enabling them to make better informed decisions for their entrepreneurial ventures.Like Magic: Reducing Consumer Credit Risk with Incentive Contracts (G2, G5)
Abstract
Incentive structures can expand access to consumer credit while improving repayment performance. We analyze a quasi-natural experiment in Brazil, where a private lender offers two types of unsecured personal loans: a standard contract with a market interest rate, and an incentive contract with a lower rate but a powerful repayment mechanism—installments must be paid via the borrower's electricity bill, with nonpayment resulting in service disconnection. This setting provides a rare opportunity to examine the causal effects of a non-pecuniary, yet enforceable, penalty within voluntary credit contracts.We develop a stylized model predicting that high-risk borrowers have stronger incentives to select the incentive contract and exert repayment effort. Using a proprietary dataset of over 17,000 approved loans from 2021 to 2023, and an additional 11,000 rejected applications, we find that borrowers who select the incentive contract are significantly riskier at origination—exhibiting around 30% lower credit scores—yet become 20% less likely to default ex post than standard borrowers. This effect is especially pronounced for ex-ante riskier individuals and single female borrowers. Over time, the lender shifts from offering standard to primarily incentive contracts, attracting a riskier borrower pool but achieving improved repayment rates.
Our findings are robust to matching techniques, controls for observable heterogeneity, and selection corrections using rejected applications. The results highlight how incentive contracts can mitigate moral hazard and help borrower manage their fund better, by leveraging essential service continuity as a commitment device. In doing so, they promote financial inclusion and expand credit access for high-risk borrowers—without raising default risk. Our study contributes to contract theory, household finance, and development economics by documenting how non-price, non-pecuniary mechanisms can shape borrower behavior and credit market outcomes in emerging markets. The findings also offer practical insights for designing sustainable lending products in environments characterized by weak enforcement and limited collateral.
Linguistic Bridge: The Impact of New Official Languages on International Trade (Z1, F1)
Abstract
This paper investigates the causal impact of official language adoption on international trade. Using a panel dataset of 211 countries from 1996 to 2019 and exploiting exogenous variation from recent language policy changes, we find that adopting a shared official language increases bilateral exports by an average of 18.77%. This effect is robust to various sensitivity tests, including those addressing endogeneity concerns. Crucially, we document significant heterogeneity in the trade impact. The effect is strongest where communication is most vital and when the adopted language signals a targeted commitment to specific trading partners. A counterfactual general equilibrium analysis suggests that the effect is equivalent to a 5.45% reduction in non-tariff barriers. Our results emphasize the importance of linguistic policy in reducing trade barriers and promoting economic integration, revealing that language barriers constitute significant obstacles to international trade.Liquidity Regulation and Shadow Banking in a Dynamic Equilibrium Model (G2, G0)
Abstract
We develop a dynamic macro-finance model of an economy with both traditional and shadow banking sectors to examine how Basel III liquidity requirements affect bank lending, capital structure, and systemic stability. Traditional banks and shadow banks compete for deposits and loans, with deposit-financed institutions transforming liquidity through a portfolio of loans, liquid bonds, and interbank lending. Our model captures business cycles through shocks that affect banks’ net worth, deposit flows, and market interest rates. Within this framework, imposing Basel-style Liquidity Coverage Ratio (LCR) and Net Stable Funding Ratio (NSFR) constraints induces shifts in the relative size and portfolio choices of traditional and shadow banking, alters the riskiness of deposits, and reshapes overall credit supply to the real sector. In particular, tighter liquidity requirements can unintentionally lower bank capital ratios and intensify systemic fragility, especially for the interbank lenders due to the lowered income from interbank trading. Taken together, our results underscore the importance of calibrating liquidity regulations in a way that internalizes the interconnected balance sheets of diverse intermediaries and provides robust system-wide protection against liquidity shocks. Our findings suggest that a revision of current Basel-style frameworks, complemented by broader support mechanisms or prudential tools encompassing shadow banks, is necessary for effective macro-prudential oversight. More importantly, the liquidity regulation should be designed in a more macro-prudential manner given the existence of shadow banking and should coordinate with other available tools, such as capital requirements. Calibrated to the US economy, the optimal level of liquidity requirements should be set around 85% and should be made specific to banks based on their liquidity positions. An appropriate level of capital should also be required and should vary positively according to banks’ overall liquidity positions.Load it Up: Economies of Scale, Market Power, and Natural Disasters in Ground Freight Transportation (Q5, L9)
Abstract
This study examines how sudden increases in marginal costs—triggered by natural disasters such as landslides—affect freight transportation prices and market structure. Using a comprehensive dataset of over 4 million truck movements, more than 1 million detailed route calculations, and geolocated landslide occurrences in Colombia during 2019, I employ staggered difference-in-differences and instrumental variables strategies to address the endogeneity inherent in the freight market. The findings indicate that landslides lead to a modest reduction in overall quantities shipped and a notable decline in the number of trucks operating on affected routes, especially on shorter routes. Moreover, the departure of additional trucks is associated with considerable rises in both trip prices and the price per unit of kilogram transported. The analysis also reveals that economies of scale are critical, as evidenced by an increase in the average load per truck and a shift in market share favoring larger vehicles. Price hikes are driven primarily by incumbent firms that usually charge higher rates rather than by deliberate price manipulation, while companies that continue operating boost profits by deploying additional high-capacity trucks to capture shipping volumes that competitors cannot match. This suggests a shift toward market concentration and elevated markups when marginal costs surge, driven mainly by the ability to exploit economies of scale with large trucks and operational liabilities rather than market power. Finally, I show counterfactual scenarios of possible price responses under policies that reduce marginal cost or increase competition using a Nash-Bertrand competition model.Machine Learning Valuations and Cross-Linguistic Information Asymmetry in Two-Stage Domain Auctions: Evidence from a Randomized Experiment (L1, D4)
Abstract
We investigate how machine learning–based valuations transform bidding behavior and outcomes in a two-stage online domain-name marketplace, where an English auction precedes a Dutch auction for unsold items. Leveraging a large-scale randomized controlled trial at GoDaddy, we randomize the disclosure of these algorithmic appraisals across millions of domains. Our reduced-form analysis reveals that while disclosing machine learning valuations increases sale probabilities and final prices in the English auction, it induces more cautious bidding in the Dutch auction.To rigorously quantify these effects, we develop two structural frameworks that integrate both common and private value signals while accounting for endogenous entry. In the first model, we treat the machine learning valuation as an additive component in bidders’ payoffs; in the second, we incorporate it as a seller-side signal within the common-value framework. By exploiting monotone quantile regression, our approach overcomes a critical identification challenge arising from the co-movement between the mean of the common value and the private value differential. Together, these models demonstrate that although algorithmic valuations can substantially boost revenue in English auctions, they may also intensify the winner’s curse in Dutch auctions with higher bidder participation.
Additionally, we document a persistent cross-linguistic undervaluation bias using natural language processing. A Random Forest tokenizer identifies domains containing Chinese Pinyin tokens, which standard machine learning tools systematically misprice, resulting in disproportionately low sale prices. This finding underscores how overlooked linguistic features can create hidden arbitrage opportunities for informed bidders. Overall, our results illustrate that while machine learning–based valuation signals can reshape auction equilibria, they may also exacerbate informational asymmetries—highlighting the need for careful design to mitigate unintended distortions and cross-linguistic inequities.
Macroeconomic Effects of Temperature Distributional Shocks (C5, Q5)
Abstract
This paper introduces a methodology to examine the macroeconomic impacts of temperature distributional shocks, that is, shocks to the average and the unconditional quantiles of the temperature process. Distributional shocks are defined as persistent deviations from the long-run trend in the global and local temperature characteristics and are identified using the Hamilton (2018) filter. Since these shocks are inherently correlated, we employ a Dynamic Factor Model (DFM) to extract the common sources of variation. Our empirical analysis at the global, United States, and Euro Area levels consistently detects three uncorrelated factor shocks, each with significant macroeconomic implications. The first factor, capturing a general distributional effect, leads to persistent declines in output and total factor productivity, aligning with existing studies that treat average temperature as a sufficient statistic for climate change. The key contribution of this paper to the literature is to uncover two additional factors that do not move the average temperature but instead influence different parts of the distribution in particular ways. These factors induce additional macroeconomic impacts that have not been documented in previous literature. For instance, the third factor—which affects the lower and the upper parts of the global or local temperature distributions in the same direction—reduces output and productivity, particularly in the Euro Area. Our findings underscore the usefulness of analyzing the effects of shocks to the full temperature distribution and carry important implications for estimating the social cost of carbon and assessing climate-related economic risks.Mandatory vs. Voluntary a priori Investment in Information Acquisition in Procurement Auctions (D4, D8)
Abstract
Procurement auctions are essential for securing goods or services at competitive prices, yet bidders often face high uncertainty regarding future costs – a challenge that is particularly prominent in renewable energy auctions. This uncertainty is mitigated by measures such as site and environmental investigations, project specifications and permits. What these costly measures have in common is that they have to be done by the auction winner anyway. The question is when, before or after the auction. Many auctions mandate these measures as prequalification, requiring bidders to invest in information acquisition prior to entering the auctions. While this mandatory setting reduces uncertainty, it imposes sunk costs that may reduce participation and exclude potential bidders. In contrast, a voluntary setting allows bidders to decide whether to invest prior to the auction or to participate without prior investment, accepting the uncertainty and investing after winning the auction.Our paper develops a theoretical model that identifies five types of symmetric equilibria in the mandatory and voluntary settings. We compare these two settings in terms of expected participation, optimal reserve price, expected profits and overall efficiency (expected welfare). Our analysis shows that the voluntary setting always equals or exceeds the mandatory setting in terms of expected participation and expected welfare. Moreover, we derive locally and globally optimal reserve prices that differ between the two settings. Given the optimal reserve price, participants expect higher profits in the voluntary setting – provided that the information cost is sufficiently high to exclude potential bidders – while the auctioneer’s expected profit may favor one setting over the other depending on specific parameters.
Overall, our paper demonstrates advantages of voluntary over mandatory a priori information acquisition in procurement auctions. By providing theoretical insights into the comparisons, this study contributes to the broader discussion on auction design and procurement policy.
Media, Political Networks, and Power Centralization: How TV Questioning Empowers Local Leaders in China (P5, P3)
Abstract
Government propaganda and information control are central to governance, particularly in centralized systems where state media play a key role in shaping political narratives. However, in China—despite strict censorship—local governments in several cities have introduced televised political questioning programs, where journalists and citizens publicly interrogate and even criticize officials. Typically promoted by municipal party secretaries and disciplinary commissions, these programs present a paradox: why would governments voluntarily foster public scrutiny? While existing research focuses primarily on how propaganda shapes public perceptions, little attention has been paid to why governments actively promote such critical programs.We propose a political network-based explanation, arguing that municipal party secretaries on the periphery of local political networks are more likely to adopt these programs as a tool to consolidate control. Using data on televised political questioning and a novel dataset on political relationship networks at the prefecture level, we measure secretaries’ network positions based on shared hometown, work, and educational ties with other officials. Employing network metrics such as degree centrality and betweenness centrality, we find that secretaries with weaker political ties—particularly those estranged from deputy mayors and key bureaucrats—are more inclined to promote these programs.
Further analysis suggests two possible mechanisms. First, public questioning enhances the secretary’s authority, as indicated by an increase in officials citing the secretary’s statements. Second, it provides a platform for patronage, facilitating the promotion of political allies. However, we find no evidence that such programs improve policy implementation or innovation. Our findings shed light on the strategic use of media-driven accountability in autocratic governance.
Medium- and Long-term Impacts on School Attendance and Learning through Videos and Mobiles: Experimental Evidence from Northern Nigeria (I2, D1)
Abstract
Using a cluster-randomized trial in Sharia-governed communities, this study examines edutainment screenings supplemented by distributing preloaded smartphones to 40 percent of invited households. After twelve months, screenings alone lowered the out-of-school rate by 43 percent but did not improve learning for the target 6–9-year-olds. Consistent with an intra-household specialization framework, older siblings (6–12) began working earlier and saw declines in schooling and learning.When smartphones were added, these adverse spillovers reversed. Foundational literacy and numeracy for target children rose by 0.46 and 0.63 standard deviations, respectively, while older siblings’ learning increased by 0.34–0.47 SDs, alongside a 15 percent drop in teenage parenthood and marriage, and a 14 percent reduction in under-age labor. At thirty-six months, out-of-school rates remained 13 percent lower, though learning gains were mainly in basic literacy among smartphone households. Teenage parenthood and marriage fell by 18 percent at follow-up.
Likely, smartphone-based resources gave older siblings flexible study options, offsetting opportunity costs and strengthening support for schooling. Mechanisms included higher parental aspirations, greater confidence in children’s abilities, and reduced time children spent on chores or paid work. The combined interventions were cost-effective, with no significant differences by gender, age, poverty status, or parental literacy.
Meet My Family: Women in Leadership and Gender Stereotypes in the Media (J1, G4)
Abstract
Women continue to be heavily underrepresented in corporate leadership positions. An understudied contributor to this may be stereotypes around female leaders, e.g. in the media. We analyze more than 45,000 newspaper articles from three major German newspapers and show that the coverage of female CEOs and board members focuses more on family and care whereas male CEOs are associated with leadership and power. With no difference in article length, this representation likely reinforces gender stereotypes while in German survey data female and male managers are equally likely to have a family. In a representative and incentivized survey experiment (N=3000), we then study whether the stereotypical representation of female CEOs affects readers' expectations and financial decisions towards the CEO's firm. Based on real CEOs and existing newspaper coverage, we vary whether a female or male CEO's family is either not mentioned, neutrally mentioned or framed in a successfully navigated trade-off with their career. Respondents expect firm stocks to perform differently for female or male CEOs and they invest less money into stocks of a female-led firm, especially when the family of a female CEO is neutrally mentioned. Instead, there is no punishment for mentioning the (successful) trade-off between career and family. These patterns are particularly driven by female respondents and parents. Using quantitative analysis of free-text reasoning, we identify family and gender as strong considerations for respondents' answers. In particular, highlighting a successful navigation of family and career increases the positive sentiment towards family and gender which suggests that these female leaders are seen as a positive role model. Male CEOs never face a penalty for having a family.Micro-Enterprise Saturation: Critical Mass or Overcrowding? (O1, D5)
Abstract
Should entrepreneurship support programs to extremely poor people be scaled by saturating a few villages, or maintaining low saturation and enrolling more villages? On the one hand, saturation may create a critical mass of entrepreneurs whose income generation and associated spending multipliers make each other viable. On the other hand, encouraging the start-up of many small firms in close proximity may be a recipe for overcrowding and artificially fierce competition in a small village economy. In a large randomized control trial in Malawi, I exogenously vary the saturation of an entrepreneurship support program to assess whether business performance increases or decreases in the share of treated neighbors. I also test a simple intervention to mediate overcrowding by encouraging entrepreneurs-to-be to coordinate which sectors they enter. One year after the start of the intervention, the number of successful businesses scales linearly with the number of supported entrepreneurs to be, but the overall earnings do not appear to grow as saturation increases. Instead, earnings are split among the larger number of supported entrepreneurs. The coordination intervention is not affecting business choice.Monetary-Fiscal Interactions During Large-Scale Asset Purchase Programs (E5, E4)
Abstract
This paper examines the effects of asset purchase programs (APPs) that were implemented in a number of countries during the COVID-19 pandemic in concert with large fiscal stimulus plans. We identify APP policy shocks for 14 advanced and emerging market economies using high-frequency identification techniques, and clean them of central bank information effects.Next, we estimate panel local projections, finding that APPs tend to stimulate output, but decrease prices. If we compare our results with those obtained without removing information effects, we find that APPs were partially interpreted as revealing bad news about economic fundamentals.
Since the launch of APPs during COVID-19 recession was accompanied by substantial fiscal expansion, we next conduct a Kitagawa-Blinder-Oaxaca decomposition of the local projection estimates, utilizing data on government spending. This allows us to separate the average monetary policy impact across the sample from its country-specific component, capturing the degree of domestic fiscal accommodation. Our analysis reveals significant cross-country variation in government spending, which drives heterogeneous responses in GDP, consumption, and prices to APP shocks. Remarkably, we find that higher government purchases during that period crowded in private consumption, and had a large effect on inflation.
We show that our empirical findings, some of which are inconsistent with a standard New Keynesian framework, can be rationalized in a simple general equilibrium model with segmented asset markets and fiscal dominance. These include a fall in inflation in response to an expansion in central bank assets, crowding-in of private consumption by government spending, as well as its high influence on inflation.
Overall, our combined empirical and theoretical analysis can be treated as indirect evidence that the monetary-fiscal policy mix observed during the COVID-19 recession could be characterized as fiscal dominance.
Moral Hazard Transmission Due to the Unexpected Rescue of a Private Bank: Evidence from India (G2, G3)
Abstract
A vast body of literature highlights how bank bailouts, capital infusions, and government safety nets can incentivize risk-taking behavior by amplifying moral hazard. While much of the research focuses on the bailed-out institutions, the broader implications of such interventions on other banks remain underexplored. This paper investigates the spillover effects of an unexpected bank bailout on the risk-taking and lending behavior of other banks in the Indian banking sector. Using granular corporate loan-level data and a dynamic difference-in-differences (DID) approach, I compare private banks (treated group) to government banks (control group) following the unexpected rescue of a private bank. I find that private banks significantly increase non-performing assets (NPA) and stock return volatility relative to government banks post-bailout. These findings suggest an increase in risk-taking driven by the “moral hazard” channel, as the bailout is expected to raise the perceived rescue likelihood for other private banks. The effect is particularly pronounced for smaller private banks and those with lower capital ratios. Loan-level analysis shows that affected private banks not only increase their overall lending at aggregate level but also expand lending to risky borrowers and engage in greater loan restructuring, relative to government banks. The results hold after controlling for demand factors through industry-time and/or borrower-time fixed effects. This study contributes to the literature by exploiting an unexpected bailout as an exogenous shock to causally link changes in risk-taking and lending to revised bailout expectations. It disentangles the “competition" and “moral hazard" channels, providing robust evidence of increased risk-taking by private banks driven by moral hazard. The findings underscore the broader negative externalities of bank bailouts, highlighting how such interventions can amplify sector-wide risk-taking behavior.Non-neutral Technological Change in Chinese Manufacturing (D2, O3)
Abstract
Between 1998 and 2008, during a period of state-owned enterprises reforms to improve efficiency, firms in Chinese manufacturing experienced rapid revenue growth and significant shifts in input factor usage and cost shares, providing strong evidence of biased technological change. This paper develops a novel method for estimating production function with firm-level factor-augmenting productivity, for capital, labor, and materials inputs. The estimates indicate an elasticity of substitution of 0.32 in Chinese manufacturing. Labor-augmenting productivity increased fastest at an annual growth of 12.2% and demonstrated a high level of persistence, followed by capital-augmenting productivity at 4.9% and material-augmenting productivity at 1.4%. In addition, substantial heterogeneity in the rates of technological change is observed across manufacturing sectors. Decomposition results show that survivors drive most of the aggregate factor-augmenting productivity changes, while entrants positively impact capital productivity and exit enhance labor productivity by removing inefficient firms.Occupations, Disability Insurance, and Career Choices (H5, J2)
Abstract
Work-limiting disabilities are a major source of income risk for older workers. While public disability insurance (DI) coverage is nearly universal, the probability of becoming dependent on benefits differs across occupations. This paper studies the value of public DI across occupations using administrative data from Germany. I exploit a large-scale policy reform in 2001 that eliminated own-occupation disability insurance (ODI) for younger cohorts, reducing DI generosity substantially. Using a triple-difference design based on birth cohort, occupational specialization, and the reform cutoff, I find that the reform led to higher earnings but reduced full-time employment, primarily due to a decline in manual jobs. The effects are concentrated among men and workers previously employed in non-manual occupations, suggesting changes in occupational transitions and reduced switches into manual work.To interpret these findings and analyze broader labor market impacts, I build a dynamic structural life-cycle model of employment, occupational choice, and retirement under uncertainty. The model incorporates occupation-specific health risks, switching frictions, social insurance programs, and is estimated using administrative labor market and pension data. In counterfactual experiments, I evaluate how adjustments to the DI and retirement system affect the labor supply of different occupational groups.
On the Shoulders of Giants: Financial Spillovers in Innovation Networks (O3, G1)
Abstract
Do markets price knowledge spillovers? We show that patent grants influence the stock returns of firms that are connected through technological knowledge dependencies. Using directed patent citations among publicly listed companies in the United States, we construct a granular measure of each firm’s exposure to new patents granted to its technologically upstream firms. Patents granted to these upstream companies significantly boost its abnormal stock returns during the week of the grant. We find that these financial spillovers are predominantly localized within a firm’s immediate technological connections. Additionally, we provide a novel empirical decomposition of financial spillovers generated from patent grants, by distinguishing those spillovers emerging from sources of technological knowledge, from those emerging from product market rivals (negative effect) and suppliers (positive effect). Our findings are robust to alternative specifications and placebo tests, and they suggest that technological knowledge spillovers create important market-priced ties between firms that are not fully captured by traditional product market relationships.Optimizing Policy Targeting with Machine Learning: Evidence from Pakistani Audits (H2, O2)
Abstract
We provide a framework for optimizing the within-policy distribution of treatment by combining machine learning techniques for the estimation of individualized causal responses with sufficient statistics for the relative welfare implications of alternative treatment distributions. This framework is applied to the setting of audit policy optimization in Pakistan. Specifically, we define a model which derives the Marginal Value of Public Funds (MVPF) in terms of three estimable causal effects of individuals in response to an audit: the net-present value of long-run tax liabilities, taxpayer burden from audit compliance, and government expenditures from the audit. With the universe of individual income tax returns in Pakistan from 2012 – 2020, we employ generalized random forests to estimate these individualized causal effects. Next, we determine the optimal distribution of audits by minimizing the MVPF with stochastic gradient descent and genetic algorithms. We find that the welfare cost per-dollar of revenue raised can be reduced by between 40% – 57% with no sacrifice to total revenue. Alternatively, by considering the inverse problem (maximize revenue subject to a maximum social welfare cost), recouped revenue can be more than doubled with no additional cost to social welfare by simply reallocating audits according to our framework.Partisan Supply Chains: The Impact of Political Ideology on Global Sourcing (F5, H7)
Abstract
This study examines how political ideology influences firms' global sourcing strategies. Using a quasi-experimental design centered on foreign elections, we find that increased ideological distance between U.S. firms and foreign governments leads to reduced imports from those countries. R&D-intensive firms show greater sensitivity to ideological shifts, while firms with established supply chains demonstrate resilience. The effect is amplified in countries with strong institutions where ideological preferences translate more effectively into policies. Firms that continue sourcing from ideologically divergent countries face heightened ESG risks. Our findings suggest political ideology serves as a leading indicator of policy risks that firms consider when making global sourcing decisions.Personal Inflation Rates in the Euro Area (E3, D8)
Abstract
Inflation affects households differently depending on their income, spending habits, portfolio choices, and consumption patterns. While headline inflation measures provide a useful aggregate perspective, they often mask substantial variation in individual inflation experiences (Kaplan and Schulhofer-Wohl, 2017). This paper explores personal inflation rates — the inflation households actually experience based on their unique expenditure compositions — and examines how these experiences shape inflation expectations across different demographic and economic groups in the euro area and over time.We leverage detailed non-public data from the ECB Consumer Expectations Survey (CES) on households' inflation perceptions, expectations, and self-reported consumption patterns. The survey structure allows us to construct personal inflation rates based on spending shares across major consumption categories, covering approximately 85% of the basket used for the Harmonized Index of Consumer Prices (HICP). We find that during the period of low inflation (April 2020 – April 2021) and the period of sticky inflation (January 2024 – October 2024), homeowners, high-income households, and older individuals experienced lower inflation. However, during the inflation surge (July 2021 – October 2023), this pattern reversed as rising energy and food prices disproportionately affected these groups, outweighing their lower spending shares in these categories.
Personally experienced inflation accounts for a significant share of the variation in inflation perceptions, inflation expectations, and broader macroeconomic expectations, including personal income expectations. Moreover, these relationships differ notably across inflation regimes.
Finally, to highlight the impact of personal inflation rates on expenditure inequality, we compute the interquartile range (IQR) of real total expenditures, deflated using both the headline HICP and personal price levels. We find suggesting that expenditure inequality intensifies when personal inflation rates are taken into account.
Our findings underscore the need for more targeted monetary policy communication, recognizing the role of individual experiences in inflation perceptions and real spending.
Physical vs. Institutional Public Goods Provision: Evidence from China (P0, H4)
Abstract
This paper argues that the level of political and market concentration explains why developing economies often underinvest in institutional infrastructure and legal capacity. Economic growth challenges this equilibrium and incentivizes rulers to invest in institutional infrastructure complementary to physical infrastructure. Rulers make joint investments to expand market entry and size if they can secure greater rents and preserve institutions favoring concentration. Using provincial shares of the national coal reserve as an instrument for market concentration levels, a difference-in-difference analysis of Chinese data from 1997 to 2006 demonstrates that the fiscal expenditure ratio of physical to institutional infrastructure rose 78\% faster in provinces with market concentration indexes in the top quartile in 2000 (the year before China acceded to the World Trade Organization). The paper also presents a theoretical model that proposes that investment in physical infrastructure increases faster than in institutional infrastructure when market concentration levels increase.Plantations and Civil Conflict: Evidence from the Indonesian Palm Oil Boom (D7, O2)
Abstract
The rapid expansion of plantation crops has transformed rural landscapes, particularly in low- and middle-income countries. While often viewed as an economic driver, plantation growth has also raised concerns about land disputes and conflict. This paper examines the impact of plantation crop expansion on civil conflict, using Indonesia’s palm oil industry as a case study. Combining high-resolution satellite imagery with conflict data, we apply a two-way fixed effects model and an instrumental variable approach to address endogeneity. Our findings show that a 1 percentage point increase in plantation crop area leads to a 1.6% rise in conflict incidents, with instrumental variable estimates suggesting a slightly larger effect. These effects persist for at least six years. In contrast, an increase in world market palm oil prices reduces conflict in areas with existing plantations, likely due to higher employment and wages. The opposing effects of land expansion and price changes suggest distinct mechanisms: while price increases improve economic opportunities, expansion fuels disputes over land rights, economic exclusion, and environmental degradation. Further analysis reveals that separatist movements and election-related disputes are particularly sensitive to plantation growth. This study contributes to the literature by highlighting that land expansion influences conflict differently from price fluctuations and that its effects are persistent. These findings underscore the need for conflict-sensitive policies in regions undergoing agricultural transformation. Governments should weigh the long-term risks of plantation expansion against its economic benefits and implement strategies to mitigate conflict. This issue extends beyond Indonesia, as many countries in the tropical belt, such as Malaysia, Colombia, Nigeria, and Brazil, have also experienced rapid plantation crop expansion, particularly in palm oil, rubber, and soy. The lessons from this study hold broader relevance for regions facing similar trade-offs between agricultural development and social stability.Policy Spillovers Across Generations: The Unintended Effects of Schooling Reform and Sugar Rationing (I1, J1)
Abstract
High sugar intake is associated with adverse health and socioeconomic outcomes, yet overconsumption has been widespread for decades. As early life conditions are particularly important for child development and can shape long-term outcomes, exposure to high sugar levels in utero may have lasting effects. This paper shows that a compulsory schooling reform in the UK may have unintentionally contributed to poorer health and employment outcomes in the next generation, by increasing the likelihood that some children were exposed to a sugar-rich diet before birth.Using data from the English Longitudinal Study of Ageing and variation from two natural experiments - a 1947 school reform and the end of sugar rationing in 1953 - I trace these effects more than 50 years later. The schooling reform led many students who were unlikely to stay in school to complete additional years of education. Using a regression discontinuity setting, I show that this altered the family planning of mothers, delaying the birth of their first child by about nine months on average and leading to an increased likelihood of children being in utero just after the end of sugar rationing in the UK, when sugar became much more available, exposing them to much higher sugar levels than they had otherwise experienced.
Comparing children in utero just before and just after the end of rationing in a second regression discontinuity setting, I estimate that in-utero exposure to a sugar-rich diet negatively affected later-life outcomes for those whose mothers were affected by the schooling reform. Between ages 50 and 65, these individuals have higher probabilities of diabetes and obesity, lower labor market participation, and are more likely to retire early than children of mothers not exposed to the schooling reform.
Political Alignment and Household Risk-Taking: Evidence from U.S. Mortgage Market (G5, R3)
Abstract
This paper investigates the impact of political alignment with the U.S. president on household risk-taking in the mortgage market. Using a sample of 37 million mortgages acquired by Fannie Mae and Freddie Mac from 2000 to 2023, we show that borrowers in politically aligned zip codes take on greater financial risks, as indicated by higher leverage ratios, loan amounts, and interest rates, relative to borrowers in misaligned areas with similar risk profiles and economic conditions. We find no evidence of lax screening from lenders, suggesting that borrower-side factors such as heightened optimism and risk tolerance drive this partisan risk-taking effect. We further provide aggregate evidence of increased mortgage origination volumes in politically aligned areas, and these loans experience higher delinquency rates. Our findings reveal an important yet underexplored connection between partisanship and household financial decisions, with potential implications for financial stability.Power Change and Sustainable Development: A Study on the Impact Mechanism of the Chairman Turnover on Corporate ESG Performance (M1, M1)
Abstract
Based on the panel data of Chinese A-share listed corporations in 2015-2022, this paper empirically examines the impact mechanism of chairman turnover on corporate ESG performance. We find that chairman turnover significantly constrains corporate ESG performance by weakening the stability of the executive team and reducing the scale of green investment; corporations can mitigate the constraints of chairman turnover on corporate ESG performance by separating the chairman from the general manager and extending the chairman's term of office; and the heterogeneity results show that the constraints on corporate ESG performance are more pronounced for non-state-owned corporations, large-scale corporations, and corporations from central and western regions evidenced from china.Powerful Politicians and Their Economic Impact (P0, F0)
Abstract
Motivated by the recent Trump 2.0 election, this paper analyzes a unique political arrangement among Chinese cities to investigate the economic effects of concentrated political power. We find that cities governed by powerful politicians with both legislative and executive authority experience lower economic growth compared to those with a separation of powers. This decline is attributed to the misuse of fiscal policies, where local leaders prioritize the state sector over the private sector, leading to resource misallocation. While this concentration of power results in higher debt levels and borrowing costs, it also facilitates more decisive responses during economic downturns, potentially spurring growth in crisis situations.Product Purchasing Contest: A Strategy to Leverage Competition among Consumers (D4, L1)
Abstract
In industries driven by loyalty and status—such as music, luxury goods, gaming, live-streaming and crowdfunding—firms increasingly use product purchasing contests, where consumers compete for a premium good by repeatedly purchasing regular goods. Unlike conventional contests that rely on monetary investments, these contests are based on consumption-based investment where consumers gain utility from purchased goods, while firms incur production costs. This unconventional structure raises concerns over wasteful excessive consumption and consumer exploitation, yet little is known about the strategic implications and welfare outcomes of such contests.To address this, I develop a new framework that extends the Tullock contest framework to capture the distinctive welfare implications of consumption-based investments. The firm gains control over contest structures by adjusting attributes of regular goods (e.g., price, quality). These structural changes make the consumer’s investment function endogenous and reshape its curvature, allowing the firm to affect how investment translates into competition and welfare. Two key policy questions guide the analysis: Should product purchasing contests be prohibited, and if not, how should they be fine-tuned?
First, outright bans for consumer protection can have unintended consequences. The optimal product purchasing contest not only increases firm profit but also achieves the first-best outcome—mirroring non-linear pricing results, but here implemented through linear pricing via contest design. However, this gain is not Pareto-improving: consumer surplus is fully extracted. This trade-off motivates the fine-tuned regulatory approach.
Second, even partial changes to the contest design significantly affect welfare outcomes. The traditional equivalence across mechanisms breaks down: the optimal Tullock contest dominates all-pay and winner-pay auctions.
Unlike conventional approaches that focus on reward structures or winner-selection rules, this paper highlights the strategic role of designing the competition process itself. This structural shift reshapes market outcomes and enables novel comparisons across mechanisms, uncovering strategic and welfare implications that conventional frameworks fail to capture.
Protected Land Rights, Misallocation and Dead Capital (O5, N9)
Abstract
Protected Land Rights (PLRs) for marginalized communities encompass a quarter of land resources across 90 countries. Yet, the impact of PLRs on structural transformation and economic development is poorly understood. Theoretically, protection could support local entrepreneurship and industry. On the other hand, such measures might lead to the economic isolation of targeted areas. Drawing on newly collected historical and high-resolution data in the context of a spatial regression discontinuity design, this paper studies the long-run impact of PLRs. Land transfer restrictions between targeted and out-group members led to 17% higher agricultural income share, 15% lower firm density in non-primary sectors, and 8% lesser housing capital in PLR areas relative to adjacent non-PLR areas. Higher transaction costs in land markets led to land misallocation and impeded efficient land use conversion and agglomeration economies. Reducing constraints on market transactions for marginalized groups may thus foster entrepreneurship and growth.Protection for Power: How Political Influence Shapes Industrial Policy (D7, L5)
Abstract
This paper examines how political power and motivations of politicians to implement policies shape industry interventions in the United States. Political power is constructed based on whether politicians such as congressmen share the same party affiliation as the majority party in Congress or as the president. We identify two key drivers of policy decisions: industries with a high local labor share and those that contribute significantly to campaign financing. By combining these political forces, we construct the shift-share political power exposure variables to analyze its impact on U.S. industry policies. Our findings indicate that political power exposure generally increases the likelihood and number of industry interventions, although the specific effects vary by intervention type. Additionally, we find evidence of a substitution effect between the two channels, suggesting that reliance on one mechanism may reduce dependence on the other.Providing Certainty (D8, D0)
Abstract
It is understood that uncertainty about future policy can incentivize agents to delay investment, while they wait for information to arrive. However, little attention has been paid to how a policymaker should optimally provide certainty to encourage early investment. In a dynamic principal-agent model, we study the endogenous provision of certainty in an environment where information about an uncertain state of the world arrives over time. We study the tradeoff between the agent's benefit from policy certainty and the principal's cost from reduced policy flexibility, and analyze how this tradeoff is distorted by strategic behavior of the players.In our model, a principal chooses a policy at a future date, and wishes to match the policy to an uncertain state of the world. An agent chooses when to make an irreversible investment, and wishes to invest only if he expects the policy will be favorable. We show that, compared to a benchmark with contractible investment time, the agent's investment is delayed. We also show that the principal provides certainty about her future choice of policy inefficiently -- no matter how much the agent values certainty, the principal provides too much of it, so that the agent's benefit from certainty is outweighed by the principal's cost from reduced policy flexibility. Our results apply to green energy subsidies and procurement of vaccines.
Public Policy and Private-Sector Prosocial Motives: The Case of Greenhouse Gas Emissions (Q5, G3)
Abstract
This paper examines how the effectiveness of public policies in promoting corporate prosocial behaviour depends on their interaction with the private sector's intrinsic prosocial motives. Given the lack of theoretical exploration of this interaction, I extend individual-level prosocial behaviour theories to the corporate context and hypothesise that public policies may crowd out private-sector intrinsic motives through both signalling (Bénabou and Tirole, 2006) and replacement (Andreoni, 1988) channels. I test this hypothesis using the 2019 adoption of greenhouse gas (GHG) emissions reduction targets in nine U.S. states as a natural experiment. Comparing firms in adopting and non-adopting states, I find that the former experience reduced shareholder pressure to cut emissions post-policy, as reflected in fewer emissions-related shareholder proposals and lower voting support. Furthermore, I find no evidence that these policies reduce corporate emissions. Consistent with my theoretical framework, facilities more susceptible to the crowding-out effect tend to increase emissions post-policy, while those less affected reduce them. Methodologically, this paper highlights the distinct economic interpretations of regressions using log-transformed versus non-transformed outcome variables. While focusing on GHG emissions, these findings have broader implications for the interaction between public policies and private-sector prosocial efforts: under certain circumstances, public interventions can inadvertently crowd out private-sector intrinsic prosocial motives, undermining their intended effects.Public Transportation Investments and the Rise of the Labor Movement (N7, D7)
Abstract
This paper investigates the determinants of the rise of the labor movement in nineteenth-century Germany, using station openings as shocks to labor demand and labor supply that subsequently affected migration flows, labor market outcomes, and labor political participation. In particular, we provide causal evidence on the effect of station openings on the rise of political participation and election outcomes. To address endogeneity concerns, we employ complementary identification strategies, including a two-way fixed effects approach, a straight-line instrument, and a control function approach following the work of Borusyak and Hull (2023). Using rich historical data for the German Empire in the period from 1871 to 1912, our results indicate that constituencies with access to the rail network experience statistically significant higher vote shares for the Socialist and Conservative parties. We also examine several transmission channels. Again, the empirical results support the hypothesis that the development of the German railway network supported the rise of the labor movement in the 19th century.Quantile Structural Vector Autoregression (C3, C5)
Abstract
Standard impulse response functions measure the effects of shocks on the expectation of response variables. We introduce a framework to measure the effects of shocks on the entire distribution of response variables, not just on the mean. Various identification schemes are considered: short-run and long-run restrictions, external instruments, and their combinations. Asymptotic distribution of the estimators is established. Simulations show our method is robust to heavy tails. Empirical applications reveal causal effects that cannot be captured by the standard approach. For example, the effect of oil price shock on GDP growth is statistically significant only in the left part of GDP growth distribution, so a spike in oil price may cause a recession, but there is no evidence that a drop in oil price may cause an expansion. Another application reveals that real activity shocks reduce stock market volatility.Reclaiming Sovereignty: A Judicial Experiment in Law Enforcement and Residential Real Estate Prices on Tribal Soil (O1, R2)
Abstract
In 2020, the U.S. Supreme Court held that the five tribes, forcibly relocated to Oklahoma by the American government, had sovereign jurisdiction over crimes involving Native Americans on tribal lands. In 2022, the Court recognized that Oklahoma and the federal government had concurrent jurisdiction over offenses committed by non-tribal citizens against members of the tribes. While the former ruling had little effect on home prices on tribal lands, the latter decision showed a marked increase in residential real estate prices, implying improved social and cultural conditions. The ruling pioneered a nuanced public policy approach to Native Americans reclaiming tribal sovereignty.Redistribution of Opportunities (D6, I3)
Abstract
Redistributing resources between members of societies has been common throughout history. In modern developed economies, one accepted approach is to gather resources from economically advantaged members through progressive income taxation and transfer them to disadvantaged members. These transfers are predictable. Many share the burden of taxes and the benefits of the transfer. This paper studies another way to redistribute resources: by reallocating opportunities. Affirmative actions in hiring, government contracts, and college admissions are but a few examples. The redistribution of opportunities differs from income redistribution: the outcomes depend on the joint production function of the match. It introduces uncertainties in the transfer ex-ante; only a few share the burden and benefits ex-post. It is also generally not progressive. This paper evaluates society's welfare under different policy regimes in a revenue-neutral environment. It finds that redistributing opportunities is uniformly worse than tax-and-transfer due to the uncertainty that the former introduces. However, opportunities, such as education, cannot be easily "taxed." A compromised policy can be expanding the class that receives the better opportunities, albeit with lower benefits than the status quo, effectively serving as a tax. The social welfare of such policy increases relative to the reallocation policy but is still worse than tax-and-transfer. The advantaged members would prefer expansion over reallocation, but the disadvantaged would prefer the opposite. Although no single policy dominates others in all metrics, the analysis elucidates the trade-offs to stakeholders for constructive conversations. For example, planning a small number of apartment buildings in wealthy subdivisions is a form of expansion of opportunities. Raising taxes to improve primary and secondary education is a form of tax and transfer.Reducing Petty Corruption in Health Care with Better Informed Patients: Experimental Evidence from in Ghana (I1, H0)
Abstract
We study how providing health insurance information to patients affects out-of-pocket health expenditures (OOPE) related to unauthorized provider fees (petty corruption) and patient behavior in Ghana’s primary healthcare system. We conducted a randomized controlled trial in 41 health facilities in Ghana, involving 1,807 patients who received one of four information interventions or a placebo message. Providing insured patients with information on their insurance benefits reduced OOPE by 23%, primarily by decreasing petty corruption in the form of top-up payments. The effect was more pronounced in larger health facilities, where OOPE decreased by 46%, suggesting that patient anonymity enhances bargaining power. The intervention did not compromise healthcare quality or access, ruling out supply-side constraints as a potential explanation. Additional changes in health staff behavior, particularly a reduction in drug provision, were observed only in larger facilities. Despite these improvements, 65% of insured patients continued to report OOPE, highlighting the limitations of patient education alone. Furthermore, the four information messages were equally effective, indicating that the delivery mode was less important than the priming effect. While health insurance literacy improved, it remained at a low level, indicating the need for further efforts to enhance patient understanding of their entitlements. Our findings contribute to the literature on financial protection in healthcare by demonstrating that low-cost information interventions can reduce petty corruption. However, complementary enforcement mechanisms are needed to ensure compliance with NHIS policies. Future research should investigate the long-term sustainability of these interventions and their broader effects on patient-provider interactions and health system accountability.Regression Discontinuity Design with Distribution-Valued Outcomes (C1, C2)
Abstract
This article introduces Regression Discontinuity Design (RDD) with Distribution-Valued Outcomes (R3D), extending the standard RDD framework to settings where the outcome is a distribution rather than a scalar. Such settings are common when treatment is assigned at a higher level of aggregation than the outcome, for example, when a subsidy is given to firms below a revenue cutoff and the outcome of interest is the distribution of employee wages. Since standard RDD methods are not directly applicable to this setting, I propose a novel approach based on discontinuities in the conditional average of random quantiles, with ``local average quantile treatment effects'' as the target estimands. To estimate these, I introduce two related estimators: one using local polynomial regression on random quantiles and another based on local Fréchet regression, a form of functional regression. For both estimators, I establish asymptotic normality and develop uniform, debiased confidence bands and a data-driven bandwidth selection procedure. These theoretical results are supported by simulations. Using the proposed methods, I estimate the effects of gubernatorial party control on the within-state income distribution in the US based on a close-election design. This produces evidence for a classical equality-efficiency tradeoff under Democratic governorship, with reductions in income above the median and slight improvements at the bottom quantiles.Regulation-Driven Innovations: A Textual Analysis of U.S. Patents and Federal Regulations (O3, O4)
Abstract
Some innovations are developed to comply with or circumvent legal and regulatory requirements. While these regulation-driven innovations can generate societal benefits, they may also incur unintended economic costs. This paper explores this unique type of innovation and examines its relationship with firm dynamics, creative destruction, and economic growth. I present a simple Schumpeterian model demonstrating how regulation-driven innovations can serve as a strategy for firms to achieve higher growth, deter competitors, and reduce the rate of creative destruction. Guided by the model’s implications, I identify regulation-driven innovations from U.S. patents issued between 1976 and 2020 by measuring the degree of alignment between patents and federal regulations. I construct this measure by estimating textual similarities between patent documents and regulatory texts using natural language processing techniques.Linking the measure with patent- and firm-level data, I find that innovation-regulation alignment is positively associated with the economic value of patents and the growth in size and market power of innovating firms. At the aggregate level, however, the static gains for innovating firms fail to offset the dynamic social costs from reduced reallocation and competition.
Relative Income and Gender Norms: Evidence from Latin America (J1, J2)
Abstract
Using data from 1.5 million dual-earner households in Brazil and Mexico, we provide the first evidence of discontinuities in the distribution of households by relative income in middle-income countries. As in high-income settings, we find a sharp drop at the 50% threshold, where the wife earns more than the husband, but the discontinuity is up to five times larger and has grown over time. These patterns are robust to dropping equal earners, minimum wage earners, and self-employed individuals. We also find that primary-earning women still supply more non-market labor than their husbands, though the gender gap is narrower. Using same-sex couples, unmarried couples, and roommates, we also find discontinuities when the younger or non-household head partner earns more. Our findings suggest that gender norms may play a key role in shaping household income dynamics, but sorting and other mechanisms may also play a role.Remitting Against Poverty: Long-Term Impacts of Introducing Mobile Banking to Migrants in Bangladesh (O1, G5)
Abstract
Poor, rural households see migration as a critical way to raise incomes, often involving migration to cities by younger family members while others remain at home. Cheaply and safely sending money home can be a challenge for migrants, but digital solutions like mobile banking are well-suited to sending money across long distances. Mobile banking may lead to economic impacts in rural areas through four channels: consumption, liquidity, investment, and general equilibrium effects.In 2014/15, a sample of 815 households in villages in northwest Bangladesh was introduced to mobile banking via an RCT (Lee et. al, AEJ Applied, 2021). Simultaneously, relatives who had migrated to work in factories in Dhaka were also trained on the technology. A year later, the intervention, which cost $12 on average, had increased mobile banking use by 48 percentage points and had increased urban-to-rural remittances by 26 percent. Among other results, we find that rural expenditures increased, and extreme poverty fell, relative to a control group, together with more reliable consumption in the lean season.
We report on a long-run follow-up with 662 migrant-household pairs eight years later in 2023. In line with our investment hypothesis, we find long-run impacts on assets: early adoption by the original treatment group led to improved housing quality and a higher value of agricultural landholdings than the control group.
Despite low levels of literacy and relatively high levels of poverty, by 2023 the control group had caught up to the treatment group’s usage of mobile money. In line with the wider technological adoption, impacts on rural household consumption, income, poverty, and financial outcomes are no longer detectable in treatment-control comparisons.
Rents, Landlord Heterogeneity, and Residential Mobility (R2, J6)
Abstract
Over the past decade, rent costs have risen steadily, while renters have become increasinglyless mobile. At the same time, the ownership composition of rental properties has shifted,
with businesses and partnerships increasingly replacing individual, ’mom-and-pop’ investors as
landlords for rental units. Using a housing panel from the American Housing Survey (AHS),
this study examines how changes in landlord ownership composition have influenced renter
mobility in response to rising housing costs.
This study builds on the initial work of Hanushek Quigley (1979) 1 and Venti Wise (1984) 2
by constructing an empirical model of renters’ mobility choices. The estimated model tests
the differences in renters’ mobility probabilities, with respect to location-specific rental price
fluctuations. The purpose of the study attempts to answer the question: Do renter households
move less during market rental increases when renting from an institutional landlords rather
renting from a personal investor landlord? Preliminary results show evidence in support for
this question.
I use the American Housing Survey (AHS) to construct a biennial panel for housing units
over the period 2015 to 2023. Through household characteristics, I am able to construct
a mobility variable tracking households that stay over the 2-year period between surveys.
Location-specific rental price data will be collected from multiple sources. For the landlord
variable, a proxy will be used initially of building structure size. It is my intention to use the
Rental Housing Finance Survey (RHFS) to match landlord information (hard to find), to the
AHS panel.
Repression and Redistribution: Elite Control and Strategic Purges in Qing China (N4, P4)
Abstract
Understanding how dictators strategically allocate political power and economic rents within elite coalitions is central to the political economy of authoritarian regimes. While existing theories emphasize coalition stability as a key to autocratic survival (Acemoglu et al., 2008), the strategic logic behind the sequencing of political purges remains underexplored. We develop a formal framework in which dictators employ sequential purges to address internal commitment and incentive compatibility problems. The model highlights a fundamental trade-off: purging rival factions helps consolidate power, but may also foster elite unity and rebellion. We show that dictators optimally initiate purges by targeting peripheral elites, exploiting informational asymmetries to inhibit coordination and deter credible defection.We test the theory using the case of early 18th-century China. The Yongzheng Emperor’s centralization campaign and targeted purges may have marked a turning point in the country's institutional trajectory, contributing to the long-run divergence from the West (Pomeranz, 2000). Confronted with entrenched elites who held significant fiscal and administrative power, Yongzheng lacked the capacity for direct confrontation. Instead, he adopted a phased purge strategy: first targeting peripheral figures to fragment elite alliances and reallocate rents, then gradually moving against core challengers. Beyond coercion, he used information control and selective redistribution to reshape elite expectations and preserve regime stability.
Our findings show that the sequencing of purges functions not merely as repression, but as a strategic mechanism for sustaining elite cohesion under informational constraints. Targeted repression enables dictators to prevent internal fragmentation without sacrificing stability. More broadly, the study illustrates how elite conflict at critical junctures shapes long-run institutional trajectories, offering new insights into the historical roots of regime durability and barriers to political lib.
Return Predictability with Macroeconomic Variables (C5, G1)
Abstract
Consider a predictive regression with a financial asset, such as a stock 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 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. 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 regression.
The second contribution of the paper is an IV 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 estimator's consistency, 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 common macroeconomic predictor variables to predict the aggregate stock market. I show that the standard OLS return predictability regression fails to show evidence of return. In contrast, the proposed IV estimator yields economically and statistically significant evidence of return predictability.
Reviews or Ratings: Quantifying Information Loss from Coarsening (D8)
Abstract
This paper quantifies the amount of information that is lost from coarsening signals, in a rate of learning sense. I show how a platform seeking to learn an unknown state should trade off between the informativeness and frequency of submission of reviews (signals). As is the case in many online platforms, I model signals as the realized utility of past consumers, where their average utility is dependent on the unknown quality of a product (state). I study how taste heterogeneity for the product impacts the performance of these systems, especially binary review systems. Information loss from coarsening is increasing in the homogeneity of the population (in terms of their idiosyncratic taste shocks). Moreover, optimal asymmetry of review systems follows the same pattern: as homogeneity increases the platform increasingly isolates extreme signals on one side of the distribution. Regardless of the degree of homogeneity, when the optimal threshold is used a binary review is preferred to a full signal if it is at least 3.25 times as frequent. Finally, I illustrate how information loss can be visualized in posterior space.Route To Cities: Natural Endowments Under Varying Institutions (N9, R1)
Abstract
According to economic geography literature, the location and size of cities are influenced by natural endowments (first nature) and human decisions (second nature). This study examines the role of natural endowments and their interaction with human decisions in the pre-industrial era. By examining the location of cities in China over the past 2000 years and linking natural endowments to historical transportation networks, my analysis reveals that natural endowments on routes generally had a positive and statistically significant influence on urban location. However, the magnitude of this effect was not constant over time but fluctuated across different dynastic cycles. This suggests that the value added to natural endowments shifted in response to changing institutional contexts. Mechanism tests indicate that governance strategies (direct vs delegated governance) and taxation structures (indirect vs direct taxation) were key institutional factors influencing the varying impact of natural endowments. Additionally, the arrival of Arab and European traders spurred the development of the maritime Silk Road and international trade, leading to a long-term, stable shift of cities towards coastal port areas.Sanctions and Currencies in Global Credit (F3, F5)
Abstract
This paper examines the impact of financial sanctions on the global dominance of the US dollar in credit markets. Using confidential bank-level data covering all global banks based in the UK, we analyse the aftermath of the 2014 Russian invasion of Crimea. Our findings reveal that banks with continued exposure to Russia shifted their claims from US dollars to euros. We attribute this "euro-isation" of global claims to an increased relative credit risk associated with dollar-denominated transactions, driven by extraterritorial sanctions leveraging the US dollar payment system. To contextualize these results, we develop a three-country model with financial intermediaries, where sanctions introduce both jurisdictional and currency-circuit frictions.Shades of Disparity: Employer-Ascribed Race and Labor Market Outcomes in Brazil (J7, J1)
Abstract
This paper studies the fluid nature of racial classification in Brazil’s labor market, where employers rather than workers record race. Linking administrative employer-employee data (RAIS) to data on self-declared race, we find that 30% of job changes coincide with changes in recorded race, and firm effects explain nearly one-third of the variation in ascribed race. Examining firm characteristics, we show that whitening firms are smaller, high-paying establishments with high-skilled and unambiguously Black or Brown peer groups, consistent with social-comparison mechanisms in racial ascription. We also demonstrate how estimates of the wage gap depend on the measure of race: relying on racial classifications by current employers yields smaller differentials (21–26 percent), while person effects, which capture the joint perception of all employers over a worker’s career, produces larger gaps (up to 42 percent). Linking data on self-declared race indicates that some of these differences reflect a divergence between employer-ascribed and worker-identified race. Overall, our findings suggest that racial categories emerge through interactions between workers and firms, with significant implications for measuring and addressing labor market inequalities.Smart Homes: Impact on Sustainable Consumer Behavior (G5, Q5)
Abstract
We examine the causal impact of smart home digitalization on sustainable consumer behaviors using a Difference-in-Differences (DiD) approach applied to a pilot smart home initiative in Singapore's public housing estates. Our analysis reveals significant reductions in household utility consumption, private car usage, and notable increases in sustainable dietary and energy-related behaviors. Heterogeneous impacts are evident, with stronger behavioral shifts among younger, wealthier, and higher-usage households. Mechanism analysis suggests that changes are driven both by increased economic savings and heightened environmental awareness resulting from real-time feedback provided by smart technologies. Furthermore, we document significant spillover effects on transportation choices, highlighting the importance of social visibility and economic incentives in fostering broader sustainability impacts. These findings underscore the role of targeted digital interventions in shaping sustainable urban consumption behaviors and provide actionable insights for policy design.Sovereign Credit Risk, U.S. Monetary Policy, and the Role of Financial Intermediaries (F3, G1)
Abstract
International asset prices are strongly interconnected and U.S. monetary policy plays an important role in driving global financial markets (e.g., Miranda-Agrippino and Rey (2020)). In this paper we study the impact of U.S. monetary policy movements on sovereign credit markets, with a particular focus on the role of financial intermediaries.The intermediary channel serves as an economically motivated source of time variation. As sovereign credit markets are intermediated via dealers, shocks to the U.S. interest rate curve affect both the fundamental value of foreign credit assets and, additionally, the ability of constrained financial institutions to intermediate with other interested parties in sovereign debt. As a result, U.S. monetary policy can have multi-dimensional effects on the pricing of sovereign credit risk.
Using sovereign credit default swap (CDS) spreads from developed and emerging market economies, we provide strong evidence that the health of global intermediaries serves as an amplification mechanism for policy-related interest rate shocks. Measures of intermediary stress available through commercial data (e.g., He, Kelly, and Manela (2017)) and regulatory data on dealer and country-specific long and short CDS positions, suggest that credit spread sensitivities to US monetary policy are significantly elevated when financial intermediaries display a reduced capacity to take on risk prior to an FOMC announcement.
We rationalize our empirical findings and shed new insights in a general equilibrium model with emerging market economies that borrow sovereign debt from constrained financial institutions in an advanced economy (similar to the U.S.). The advanced economy contains price rigidities, and shocks to U.S. interest rates, via a Taylor Rule, affect the pricing of global credit as financial institutions determine the credit supply curve. When intermediaries are constrained, a positive shock to interest rates severely affects their stochastic discount factor, reducing credit supply and adversely affecting equilibrium credit spreads.
Student Loans and Labor Supply Incentives (G5, I2)
Abstract
We develop a dynamic household finance model showing that student loans – non-dischargeable in the U.S. bankruptcy – alleviate the well-documented debt overhang
in labor supply decisions. Non-dischargeability mutes opportunities for households to
strategically reduce labor supply at the expense of creditors, thus correcting incentive
distortions. This corrective effect, however, is partially undone by Income Driven
Repayment (IDR) plans, which set student loan payments formulaically regardless of
outstanding balance. IDR thus allows households to pseudo “discharge” student debt
and re-activates debt overhang. We supplement our model with empirical analyses and
uncover potentially unintended consequences of proposed reforms in student loans.
Sun, Sand, and Services: Tourism and Household Welfare in Jamaica (O1, Y4)
Abstract
Many developing economies have pursued specialization in services, such as tourism. However, there are questions about the scope of service sector industries to improve household welfare across the income distribution. I answer the question: “Are increases in the size of a nation’s tourism industry welfare improving for a nation’s citizens?” in Jamaica, a country with a 4.4-billion-dollar tourism industry, welcoming over 4 million tourists per year, and comprising over 30% of GDP.I employ two unique datasets spanning the period 2001-2021. I use a nationally representative survey of households that provides a rich repeated cross-section of household expenditures, demographics, and location. I combine this with government exit surveys completed by random samples of tourists from each year of my study. These surveys provide granular data on locations visited, expenditures, and tourist demographics.
I contribute to areas of the literature studying the relationship between service sector growth and economic development, the spatial characteristics of economic activity, and the relationship between trade and structural change.
I instrument for municipality tourism levels with a Bartik instrument, exploiting variation in where tourists from different countries choose to vacation over my study period.
The benefits of tourism accrue mostly to urban households in non-tourism related sectors, with a 5 million dollar increase in accommodation spending results on average in a 1% increase in per-capita household expenditures, while an increase of 23 million dollars decreases the likelihood of urban poverty by 1%. Among rural households, only the wealthiest benefit.
I show that tourism can be welfare improving for some households, and that sectoral spillovers are possible. However, the benefits may exacerbate existing inequalities and may not extend across space to rural populations. I examine the underlying mechanisms of my findings and draw conclusions about their implications for specialization in tourism, and services more generally.
Supervisor from afar: Panacea or Slow Posion? (H7, P0)
Abstract
Rotation has long been regarded as an effective design to mitigate principal-agent problems, especially regrading multi-level governance practice. However, leveraging newly developed data on county-level government overtime, this paper uncovers its hidden institutional costs. Specifically, by lowering administrative costs, rotation can intensify a “race to the bottom,” exacerbating distortions in government intervention and deepening bureaucratic conflict. Drawing on China’s cross-regional promotion system, empirical evidence shows that while rotated officials significantly increase the working hours of local civil servants, this does not lead to measurable improvements in administrative efficiency or increased public support. On the contrary, rising dissatisfaction, bureaucratic inertia, and even instances of passive resistance and work stoppages among civil servants have been observed. Building on these findings, this paper contributes to a rethinking of the conventional wisdom on rotation, institutions, and the political economy of principal-agent problems.Tech Openness: Corporate Culture in Times of AI-Transformation (G3, O3)
Abstract
Motivated by recent technological advances and the importance of corporate culture for organizational transformation, we investigate the role of tech open corporate culture in a firm's capital market performance during technological disruptions. We develop and validate a measure of corporate tech openness culture based on Glassdoor employee reviews. Using the launch of a major large language model as our setting, we show that tech openness is positively related to firms' stock market performance following disruptive technological advances. A one-standard deviation increase in tech openness is associated with a 1 percentage point increase in cumulative abnormal return (CAR) in a 3-day event window. A tech openness long-short portfolio yields 1.4 percentage points of daily abnormal return during the event. Consistent with corporate culture driving the effect, our baseline results are stronger when investors are more aware of firms' culture. These findings suggest that capital markets recognize tech open corporate culture as a valuable intangible asset, highlighting the importance of human factors in technological transformation.Technology, Trade, and the Skill Premium in a Model with Skill Flexibility (F1, J3)
Abstract
This paper introduces skill flexibility--a recent empirical finding in labor economics--into the trade and wage inequality literature. First, it proposes a new mechanism that can cause an increase in wage inequality between high- and low-skilled labor through trade. The mechanism is that when the variety of imported goods increases, the diversity of inputs handled by workers increases, which increases the demand for high-skilled labor with greater skill flexibility, and widens the wage inequality between high- and low-skilled labor. Second, using the calibrated model, this paper quantitatively tests the importance of this mechanism by comparing it with the well-known mechanism that widens wage inequality through skill-biased technological change. Specifically, we answer the following three questions: (1) How much will wage inequality widen if both technological change and trade occur? (2) How much will wage inequality widen if only technological change occurs? (3) How much will wage inequality widen if only trade occurs?The After Party: Consequences of Party Bans (P0, N4)
Abstract
Political parties are often banned to suppress their political presence. We argue that political party bans may have unintended consequences. We assemble a comprehensive database of political party bans across the world since 1900. First, we show that party bans are often incomplete: over 60% of banned parties regroup or merge with an existing party. Across all regimes, factions of banned parties recover 50% of the banned party's vote share; in mature democracies, banned factions experience no loss in vote share following a ban. Post-ban, parties endogenously change their platforms to resemble banned parties. In equilibrium, this increases the probability that banned factions are represented in government. We use a structural model of voter demand and party supply to decompose these equilibrium effects of party bans, highlighting the importance of both factors in producing backfiring. Finally, we turn to case studies of modern day Belgium and 1950s Germany, showing that similar mechanisms hold when examining candidate-level results and parliamentary speech.The Dissociative Relationships of Corruption and Democracy on High-Tech Entrepreneurs versus Low-Tech Entrepreneurs (P0, L2)
Abstract
Corruption is a major barrier to investment and economic growth, particularly in entrepreneurship, which is vital for development. It undermines institutional effectiveness, erodes trust, and increases uncertainty, stifling innovation and entrepreneurial activity. While corruption hampers entrepreneurship, democracy plays a key role in fostering it by creating an environment that encourages innovation, competition, and risk-taking. In democratic societies, transparent regulatory frameworks provide access to essential resources such as a fair legal system, financial services, and strong property rights protections.Democracy and corruption are often linked, but they are not mutually exclusive. Democracies can experience corruption, and authoritarian regimes may have relatively low levels of corruption. Given the multifaceted nature of democracy—encompassing electoral processes, political participation, and civil liberties—it is important to examine how democracy affects entrepreneurship. Corruption can coexist with democracy, and non-democratic regimes can also exhibit low corruption levels.
This study examines how corruption and democracy affect high-tech and low-tech entrepreneurs differently. High-tech entrepreneurship, often global in scope, relies on human capital, financial markets, and international networks, which are more common in democratic countries, and is less affected by local corruption. High-tech sectors, with higher profit margins, are also better able to absorb the costs of corruption. In contrast, low-tech ventures, which depend more on local markets and government regulations, are more vulnerable to corruption and its associated costs while less directly influenced by democracy.
Using a dataset of 102 countries, the paper distinguishes between high-tech and low-tech entrepreneurship and addresses conflicting evidence on the impact of corruption and democracy on entrepreneurship. The findings reveal that high-tech entrepreneurship is less affected by corruption and more positively associated with democratic institutions, while low-tech ventures are more vulnerable to corruption and less influenced by democracy. This research contributes to understanding the complex relationships between institutional quality, entrepreneurship, and economic growth.
The Economic Costs of Oligopoly on the Markets of Rare Earth Elements (Q3, D4)
Abstract
Abstract. In this research, the costs of rare earth elements (REEs) due to the oligopolistic market structure are estimated. On the supply side, the China exercise a dominant position on the international REE market, because controls more than 60 percent of deposits, and more than 80 percent of REEs traded (Baskaran, 2024, Fan et al., 2023). Other countries assume an ancillary role in international markets for these minerals, which will increasingly play a crucial role in economic growth in the coming decades (REEs are considered the minerals of technology). Using the fringe oligopoly model (Benchekroun et al., 2023, Gilbert, 1978, Lewis and Schmalensee. 1980, Newbery, 1981), the costs due to this market structure for economies less endowed with deposits of these minerals are estimated. To perform this type of analysis, the Herfindal-Hirshman index was calculated (Masson. and Shaanan, 1984, Newbery, (1981) and a new index of dependence on REEs (vulnerability) was constructed. Our database covers fifty years and sixty-two countries. Most of the data used comes from the World Bank database.Using various econometric models, we are able to calculate the economic costs to countries with lesser REEs caused by the market power exerted by the country acting as the leader in this market. In the research we propose a simple way to estimate the loss for each individual country involved in the international trade of these minerals. Finally, the role of technological progress Countries dependent on REEs (belonging to the oligopolistic market fringe) in various ways to reduce these costs is explained. For example, by reducing the amount of REEs used in the production process, by exploiting the recycling of minerals contained in exhausted products, and finally by finding substitute inputs.
The Effects of Armed Conflict on Women's Empowerment in Burkina Faso (C1, I0)
Abstract
Background:Previous studies have examined the consequences of armed conflict on women's wellbeing. Less work, however, examines how armed conflict affects women's empowerment. We analyze the effects of armed conflict in Burkina Faso on multiple domains of women's empowerment measured in the project-level Women's Empowerment in Agriculture Index.
Data:
We use data from the evaluation of a gender- and nutrition-sensitive poultry production intervention (SELEVER) in Burkina Faso, which took place during a period of increased conflict (2017-2020) and combine these data with Armed Conflict Location and Event Data project database.
Methodology:
To evaluate the effect of conflict on empowerment, we estimate a difference-in-difference model, separately for women and men, across multiple empowerment indicators, in which the primary explanatory variable indicates whether conflict moved closer to the village during this time period. Then, to determine if the SELEVER program had a protective effect in the context of increased insecurity, we estimate a triple difference model. As robustness-check, we use a novel approach proposed for difference-in-differences using continuous treatment (de Chaisemartin et al. (2024)), where we define conflict exposure as the distance to conflict (in km).
Results:
Increased exposure to conflict negatively affected intrahousehold decisions in terms of women's input into livelihood decisions and women's control over use of income. Interpreting these findings along with trends suggesting increased acceptability of intimate partner violence suggests an alarming shift in intrahousehold dynamics. Somewhat surprisingly, we also found that increased proximity to conflict led to an increase in access to and decisions on credit, as well as an increase in women's self-efficacy. These findings may be related to how humanitarian support was being delivered. In examining whether SELEVER had any protective effect, we only find evidence of a protective effect in the area of work balance on both women and men.
The Impact of 3G Network Coverage on Fertility Decisions and Infant Mortality in Nigeria (I1, O3)
Abstract
This study examines the causal relationship between 3G mobile network coverage, fertility decisions, and infant mortality in Nigeria. Using geo-referenced data from Nigerian Demographic and Health Surveys (2013-2018) matched with mobile coverage information, we implement two-way fixed effects and staggered difference-in-differences approaches for fertility analysis, alongside a sample selection model for infant mortality. Results show that increased 3G coverage significantly reduces birth rates, with effects approximately twice as strong for adolescent women (15-19 years). The spatial gradient of effects—stronger at closer proximity (20km) and diminishing with distance(40km)—supports a causal interpretation, and our staggered DID estimator confirms robustness to concerns about heterogeneous treatment timing. For infant mortality, our selection-corrected models reveal no statistically significant direct relationship between 3G coverage and child survival outcomes after accounting for fertility decisions. These findings indicate that mobile connectivity primarily influences demographic outcomes through fertility decisions rather than through direct effects on child survival, suggesting telecommunications infrastructure investments may yield substantial demographic benefits primarily through reduced fertility rates, particularly among adolescents. The study advances our understanding of how information access through digital technologies shapes reproductive health choices in developing contexts.
The Impact of Repealing Certificate-of-Need Laws on Healthcare Spending: A Causal Analysis (I1, L5)
Abstract
Certificate-of-Need laws in 35 US states require healthcare providers to obtain government approval before building new facilities or expanding services. Their stated goal is to ensure a genuine community need and avoid redundant services, with the aim of controlling healthcare costs and distributing resources fairly, although their effectiveness is debated. Our study differs from previous literature as we presentthe causal effects of repealing Certificate-of-Need laws on various healthcare expenses using modern difference-in-differences techniques. Our findings suggest that repealing Certificate-of-Need laws can reduce the burden of healthcare spending.
The Impact of the "Forced Labor Dirty List" on Agricultural Transactions in Brazil (Q1)
Abstract
The cattle sector is a major perpetrator of labor exploitation in Brazil, accounting for nearly a third of the 60,000 workers ever rescued from modern slave labor (MSL) in the country since 1995. The Ministry of Labor maintains the “Dirty List”, a public registry of employers that have been found to have used MSL. The list is available online and is updated every six months, allowing consumers and companies to boycott firms with known cases of MSL. In the cattle sector, 75% of export-certified slaughterhouses signed sustainable supply chain Cattle Agreements (CAs) promising not to purchase cattle from ranchers on the Dirty List.We used an event study approach and data from millions of cattle transactions in the Brazilian Amazon to address the question: How did the public Dirty List combined with private-sector sustainable sourcing commitments impact market access and cattle sales for ranchers who employed MSL?
We found that ranchers did not change their total sales volumes when they went on the Dirty List, but significantly altered their destinations of sale. Dirty Listed ranchers halved direct sales to CA slaughterhouses, indicating that these buyers monitored and blocked non-compliant direct suppliers. However, Dirty Listed ranchers continued to sell cattle to non-CA slaughterhouses at pre-listing levels (representing around 25% of total sales volume). They also increased sales to other ranchers by the same magnitude as they decreased sales to CA slaughterhouses. Many of the ranchers who purchased from Dirty Listed ranchers then quickly went on to sell cattle to CA slaughterhouses, suggesting that sanctioned ranchers laundered cattle through clean ranches to continue to access CA slaughterhouses. Finally, with export-level shipment data from a subset of our study years, we demonstrate that both domestic and international supply chains were likely contaminated with cattle raised by Dirty Listed ranchers.
The Impact of the United States Supreme Court's Ruling in Janus v. AFSCME (K3, J4)
Abstract
In 2018 in the case, Janus v. AFSCME, the United States Supreme Court overruled one of its prior decisions that had allowed public-sector unions to deduct dues from every employee even if they chose not to be associated with a union. In the first academic study on the impact of this landmark case, I quantify the impact of this decision on union membership. Using data generated from several hundred Open Records Requests, I find a 12 percent decline in the number of dues-paying members over the subsequent three-year period. About two-thirds of that reduction occurs in the very first year itself suggesting that the Janus decision represented a distinct shock to the legal environment governing public-sector unions.The Impact of Unemployment on Hiring Costs following Recessions (E2, E3)
Abstract
Empirical evidence suggests that after a recession the rate of recovery in the labor market is proportional to the severity of unemployment that precedes it, irrespective of other aspects of the recession. This proportionality can only be partially explained by the high number of job seekers relative to available work in the post-recession labor market, indicating other feedback mechanisms linking unemployment to hiring costs. This paper develops and estimates a flexible model of hiring costs to firms using data from the Survey of Income and Program Participation and the Economic Census. The relative importance of different theoretical sources of hiring costs during recoveries are compared using the estimation results.The Impacts of Reduced Access to Legal Assistance: Evidence from England andWales (H5, K4)
Abstract
In 2013, England and Wales implemented a sweeping legal aid reform that drastically reduced publicly funded legal assistance for low-income households facing social welfare issues. The 80% funding cut led to uneven provider closures and increased congestion, restricting legal assistance to immediate court actions while eliminating support for early interventions. This paper examines the reform’s impact on access to justice and socioeconomic outcomes for vulnerable populations.Constructing panel data on provider activity from 2011 to 2023, this paper examines the reform’s impact on legal aid availability, eviction and debt court cases, housing market tension, healthcare services use, and mortality. We adopt a dual strategy: first, a difference-in-differences approach leveraging spatial and temporal variations in access to providers, measured by changes in distance; and second, a Bartik instrument to address differential provider resilience to the reform and predict shifts in legal aid flows. We quantify the cumulative impact of reduced access to free, in-person legal assistance on outcomes with lasting socioeconomic implications.
Preliminary findings suggest that the legal aid cuts increased the average distance to the nearest provider by 3.2 km. This reduced access led to localized rises in eviction filings and orders, as well as higher mortality over the decade. This study highlights an overlooked intervention targeting households at risk of homelessness and over-indebtedness. Using a Marginal Value for Public Funds framework, it shows how a cost-savings reform initiated by the central government shifted welfare costs onto local authorities, offering empirical insights into the unintended socioeconomic and public health consequences of cutting legal aid post-recession.
The Long-Run Effects of Public Spending and Investment on Regional Economies: Multi-Century Evidence from the US Civil War (R1, N9)
Abstract
This paper examines government-driven economic expansion resulting from industrial mobilization in the U.S. North during the Civil War to investigate whether short-term regional government intervention improves long-term outcomes for local economies and individuals. I analyze the period from 1820 to 2000 and compare counties with government procurement and manufacturing facilities (Quartermaster (QM) and Ordnance) to other counties that had similar population levels, infrastructure, and manufacturing capacity in 1860. I find that counties with QM sites and comparable counties without them exhibited similar trends in outcomes before 1860 and showed no differences in outcome levels in the decades leading up to the Civil War. However, after the Civil War, the presence of these facilities had a large and persistent impact on local development, characterized by higher population growth, manufacturing employment, and average manufacturing wages—effects that endured for more than a century. Additionally, individuals in treated counties experienced greater upward mobility than those in the counterfactual group. Men born between 1850 and 1860 who lived in treated counties in 1860 achieved significant improvements in their occupational scores as adults compared to their household backgrounds, with the most substantial gains among those from households with low occupational scores in 1860. Post-Civil War, treated counties also saw a significant increase in innovation. The nature of the site mattered significantly for long-run persistence of treatment effects: counties with QM sites exhibited sustained effects over the long term, while Ordnance sites did not. Furthermore, I present a simple conceptual framework to explain these findings, arguing that agglomeration-related technological spillovers are the core mechanism driving long-run path dependence. The pattern of spillover effects across county boundaries into neighboring counties aligns with the agglomeration shadows conceptualized in Fujita, Krugman, and Mori (1999).The Political Economy of Decontaminating Public Lands (Q5, H7)
Abstract
We examines how investment in pollution remediation and abatement differs between different types of federal land in Canada. When we isolate Indigenous territory, known as a reserve in Canada's Indian Act, we find that, compared to other types of federal land, they are more likely to be contaminated and that the progress of pollution remediation is slower. We also find that less is invested in the remediation process in polluted sites in Indigenous territories than on other types of federal land. Our results are robust to the type of pollutant and the location within Canada of federal land. We propose a number of possible explanations for our results, from systemic racism on the part of the Canadian government, to principal-agent issues in pollution abatement, to the tragedy of the anti-common.The Power of Persuasion: Evidence from Frederick Douglass's Oratory (N4, D7)
Abstract
This paper examines the impact of Frederick Douglass’s powerful oratory on political attitudes, collective behaviors and racial inequality. Leveraging newly compiled data on the locations and timing of Douglass’s addresses, we document three main findings. First, Douglass’s speeches substantially reduced support for the pro-slavery Democratic Party in towns he visited. Second, towns exposed to his speeches experienced higher rates of Union Army enlistment during the American Civil War (1861–1865). Third, inequality faced by African Americans decreased in the aftermath of his visits. A mechanism analysis of contemporary newspapers indicates that Douglass’s oratory fundamentally altered local media portrayals of African Americans, reshaping public opinion and facilitating community-wide support for antislavery reforms. These results highlight the pivotal role of individual leadership and persuasive speech in driving social and political change and encouraging high-risk collective action even without tangible material incentives.The Power of Social Networks: Information Elites and the Spread of Politically Sensitive Information under Media Censorship (P3, D7)
Abstract
This paper empirically examines how politically sensitive information spreads through social networks under strict media censorship in China. Using a unique dataset from Sina Weibo (China’s version of Twitter), we map large-scale anonymized users’ online social networks and identify key network nodes—referred to as “information elites”—who can circumvent censorship to access uncensored content. We analyze users’ public posts to determine whether they align with the Chinese government’s propaganda on three key issues, which may reflect their sources of information: (1) COVID-19: Omicron remains fatal; (2) the Russia-Ukraine war: Russia is fighting for justice; and (3) Japan’s discharge of nuclear wastewater into the ocean: Japan is extremely selfish and irresponsible. Our findings show that: (i) users connected to information elites are significantly more likely to disagree with government propaganda than those who are not connected; (ii) even users who initially agree with the propaganda are more likely to shift their beliefs over time if connected to information elites. These results suggest that information elites may significantly reshape their friends’ beliefs by sharing updated, uncensored information. Our findings highlight the power of social networks in undermining the effectiveness of media censorship in the information age, with information elites playing a pivotal role in disseminating uncensored content among citizens.The Private Cost of Legal Uncertainty: Evidence from the Unified Patent Court (K4, K4)
Abstract
“Legal certainty” signifies the certainty of economic agents about the meaning and enforceability of legal rules and the procedures for resolving legal disputes. It has been considered a cornerstone of market economies for centuries, yet it remains challenging to quantify. This paper leverages the introduction of the Unitary Patent system in the European Union as an exogenous shock that exposed patentees to uncertainty about the new Unified Patent Court (UPC). While the reform aims to reduce the cost of maintaining patents across multiple European countries, the established system of national validations remains available. Patentees can thus “pay” to avoid UPC jurisdiction by opting for national validations. I calculate that patentees who retained the national route despite possible savings forfeited approximately EUR 100 million, while those who accepted simultaneous UPC jurisdiction left EUR 65 million in savings "on the table." On a per-patent basis, this equates to around EUR 10,000 per grant and EUR 25,000 per patentee. For patentees with a long history of maintaining European patents, I estimate that the average patentee requires fee savings of EUR 4,000-6,000 per patent to be indifferent between the Unitary Patent and the traditional routes. I disaggregate part of these costs of the UPC route into expectations about litigation costs, litigation outcomes, and the importance of patent licensing, and find considerable heterogeneity in the remaining baseline cost between patentees from different countries participating in the Unitary Patent.The Risks of Generative AI: Data Quality and Its Impact on Economic Growth (O4, H4)
Abstract
Generative AI has recently emerged as a pivotal technological driver of economic growth, leading to a substantial surge in AI-generated content. However, the widespread adoption of this technology raises concerns regarding potential risks, particularly those stemming from excessive dependence on AI. This paper investigates both the benefits and drawbacks of Generative AI from the perspective of data quality. Specifically, we develop a semi-endogenous growth model in which production relies on two types of data: AI-generated data and producer data, with the latter representing real-world information. Although AI-generated data are significantly cheaper to produce, their use entails a trade-off in the form of lower data quality, resulting in higher error rates in production. Our analysis demonstrates that firms, operating under competitive equilibrium, tend to overutilize AI-generated data relative to the optimal allocation. This finding suggests that AI-generated data should serve as a complement to, rather than a substitute for, real-world data in practical applications.The Role of Merit-Based Aid in College Enrollment and Selection: Evidence from Georgia’s HOPE Scholarship (I2, J0)
Abstract
As concerns over college affordability and financial aid policies grow, recent research has focused on assessing whether financial aid effectively targets students in need. While merit-based scholarships provide financial support, they may inadvertently widen disparities between low-income and affluent students, as they lack income-based eligibility requirements. In Georgia, the merit-based HOPE Scholarship has been shown to increase overall college enrollment rates compared to other Southern states. However, to accurately evaluate its impact on students from diverse backgrounds, it is crucial to examine the choices of students near the eligibility threshold and analyze effects across different student subgroups. This study investigates the impact of financial aid on high school students’ college enrollment and choices, with a particular focus on Georgia’s HOPE Scholarship from the 2007 to 2024 school years. Using student-level administrative data from large public school districts in Atlanta—which account for 35% of Georgia’s high school graduates—this study employs a regression discontinuity design to estimate how eligibility, determined by a weighted GPA cutoff of 3.0, affects students’ college enrollment decisions. By linking high school transcripts with National Student ClearingHouse data, it also tracks students’ postsecondary enrollment, persistence, and degree completion. Additionally, the study explores how financial aid influences college type, major selection, and institutional selectivity. Through a cost-benefit analysis across overall students and subgroups, particularly those from low-income backgrounds and minorities, this research aims to identify which groups benefit the most from financial aid. The findings will provide policymakers with insights to refine financial aid programs, enhancing both effectiveness and equity in higher education.The role of shocks in poverty (O1)
Abstract
In any narrative of persistent, extreme poverty, shocks loom large. Households living below the poverty line face an incessant onslaught of shocks to their health and livelihoods. But their frequency, paired with recall bias makes them hard to measure and study. I test whether a successful and widely studied poverty alleviation program, the graduation model, operates by lowering the arrival rate of shocks or by improving households' ability to cope with shocks. To do so, I conduct a large RCT with a unique panel structure tracking households across about 40,000 visits to explore the relationship between shocks and poverty. High-frequency data overcomes the measurement challenge of long recall periods in classical RCT designs. I find that one year into the intervention, shock arrival rates are nearly identical between the treatment and control group, but coping ability is dramatically ameliorated. I then investigate the short-run impacts of shocks on households in the study using an event-study design that allows for multiple treatment doses, as well as the long-run impacts in a structural model that also allows me to investigate a generalization of the typical concept of a poverty trap (analysis pending). In this model, shocks from one domain of life, say agriculture, can exacerbate a households ability to cope with shocks in another, say health, thus generalizing previously studied domain-specific poverty traps.Time-Varying Salience Effect (G4, G1)
Abstract
I combine an empirical analysis and an online experiment to examine how selective attention, framed by salience theory, influences financial markets across business cycles. Salience theory suggests investors disproportionately focus on extreme stock returns, affecting trading behavior and asset prices. Traditional asset pricing models assume full investor rationality, yet empirical evidence reveals bounded rationality and selective attention leading to systematic biases.Using U.S. stock data from NYSE, Nasdaq, and Amex (1931–2015), I classify stocks into deciles based on their salience—measured by deviations of individual stock returns from the market average—and construct a zero-cost strategy (long upside-salient stocks, short downside-salient stocks). Univariate portfolio sorts, firm-level Fama-MacBeth regressions, and subsample analyses across business cycles (identified by NBER classifications) control for firm characteristics including market beta, size, book-to-market ratio, momentum, illiquidity, and idiosyncratic volatility.
Results confirm stocks with high upside salience exhibit lower future returns, consistent with investor overvaluation of salient returns. Crucially, this salience effect is stronger in recessions (-1.29% per month) than expansions (-0.53% per month), suggesting heightened investor attention constraints during downturns. Investors disproportionately focus on upside salient (growth) stocks during recessions, intensifying overvaluation and subsequent price reversals. The relationship remains robust after controlling for other known predictors.
The online experiment simulates economic states, confirming participants allocate greater attention to upside salient stocks during downturns, mirroring historical mispricing patterns. These findings underscore the cyclical nature of salience-driven mispricing, providing valuable insights for asset pricing models, portfolio management strategies, and policymakers aiming to mitigate behavioral biases.
Trade Ties That Bind: Strategic Trade in the Age of Geoeconomic Fragmentation (F1, F5)
Abstract
Recent work estimates that the costs of geoeconomic fragmentation will be large if the world fragments into zones, but that these costs depend importantly on the structure and composition of the zones. Instead of taking zones as given, this paper examines the endogenous formation of zones if countries interact strategically. We use a quantitative trade model to simulate the formation of zones when countries take trading partners’ decisions into account in both simultaneous and sequential games. The model’s algorithm then tests the stability of the equilibrium. Our findings reveal that almost all countries will prefer to remain nonaligned when forced to choose between zones centering around the United States and China. The European Union acts as a ‘swing state’, pulling a substantial share of other countries with it if it decides to choose a US- or China-centered zone, or form its own zone, respectively. Attempts to punish nonaligned countries, or to retaliate against the main zones lead to minimal changes in the composition of zones but lead to significantly higher global costs of fragmentation.Turning Information into Action: Bridging the Policy–Practice Gap through a Scalable School Leadership Intervention (I2, C9)
Abstract
We study whether light-touch reminders paired with actionable guidance can convert standardized test report cards into school actions and higher achievement. In a nationwide randomized controlled trial with 2,647 schools in the Dominican Republic (2019), principals were sent instructions to retrieve their full school report and to convene structured workshops with leadership teams, teachers, and families—supported by templates and guides—while follow-up reminder calls varied in content and intensity across treatment arms. The intervention markedly increased intermediate engagement: email open rates rose by up to 26 percentage points, report downloads more than doubled, and documented workshop completion increased by about 30 percentage points.COVID-19 delayed national exams originally planned for 2020–2021, yielding 2023 tests for 3rd grade and 2024 tests for 6th grade. Using these outcomes, we estimate learning effects with a difference-in-differences/2SLS design that instruments observed school actions with random assignment to teacher-workshop reminder calls, which boosted improvement-plan submissions. We find no effects in 3rd grade. For 6th grade public schools, we detect statistically significant gains in mathematics but no effects in Spanish.
Taken together, the results indicate that scalable, reminder-based implementation—coupled with simple, actionable guidance—can move schools from available assessment results to coordinated practice and, in some settings, higher learning.
Understanding the Linkages between Multiple Greenhouses, Forest Cover, and Economic Prosperity (Q2, Q5)
Abstract
The Global Forest Resources Assessment 2020 report indicates considerable variability in both the direction and magnitude of the change of forest cover across Asia, Africa, Europe, North America, Oceania, and South America. In this paper, I develop a two-part time series model to examine first the link between three correlated environmental indicators, area of forest cover, carbon dioxide emissions, and methane emissions. The model further examines the relationship between this set of environmental indicators and economic prosperity (income level and degree of income equality) of a country while accounting for covariates such as trade openness and rate of change in urban-rural population. Following Pesaran (2006, 2007), the model incorporates the possibility of cross-sectional dependence across the panels.Using data from the World Development Indicators, the model is applied to a diverse panel of 78 countries from all six continents for the years 1992 through 2021. The dataset includes countries that have consistently remained top emitters of both carbon dioxide and methane over the time frame and include the largest forested areas on the planet.
Preliminary results indicate that while multiple countries are making progress on one dimension of improving environmental quality, such as reducing annual carbon dioxide emissions, they are losing ground on another dimension, such as losing forest cover rapidly or experiencing a steady rise in methane emissions. The research findings emphasize the need to focus on correlated environmental indicators while designing policies that are aimed to improve environmental quality to meet climate change related goals.
Unemployment Insurance, Informality, and Job Search Behavior: Evidence from Longitudinal Data (J6, J4)
Abstract
This paper examines how unemployment insurance (UI) design affects job search behavior in Chile, where a hybrid system uniquely combines individual savings accounts (ISA) with a collective solidarity fund (SF). Leveraging novel longitudinal administrative data that links job applications from Chile’s public job posting platform (BNE) to UI payment histories and formal employment records, the analysis tracks job search intensity, target wages, and transitions to formal and informal employment throughout unemployment spells.To the best of our knowledge, this is the first study to examine job search behavior using UI-linked administrative application data in a middle-income country, and only the second overall, following Marinescu and Skandalis (2022) in France. A distinctive feature of the Chilean context is the availability of data on informal job postings via the BNE, enabling a detailed analysis of how informal employment interacts with UI receipt.
First, the paper documents differences in search behavior between ISA-funded and SF-funded beneficiaries, testing whether self-financing reduces moral hazard compared to collectively funded benefits. Second, it exploits a 2015 UI reform that introduced front-loaded benefits for ISA recipients and increased generosity for SF recipients. Using a difference-in-differences strategy, we estimate the reform’s impact on job search intensity, reemployment outcomes, and informal job uptake.
Complementing the reduced-form analysis, the paper develops and estimates a structural job search model incorporating formal and informal labor markets, sector-specific search costs, and unobserved heterogeneity. The model captures the trade-offs faced by UI recipients who supplement their income through informal work while maintaining benefit eligibility. The results uncover key mechanisms behind observed behavioral differences in job search and reemployment, offering evidence to inform UI policy design in developing countries, where informality plays a central role in shaping labor market dynamics.
US Tariffs in a Model with Trade and FDI (F1, F4)
Abstract
The new US administration aims to reduce imports and attract FDI by imposing tariffs and using the proceeds to support domestic investment. This paper develops a dynamic two-country model (US vs. RoW) with monopolistically competitive firms making export and FDI decisions. The model incorporates international supply linkages and economies of scale in FDI sales, considering both cost and demand effects. We use trade elasticities and empirical estimates of how tariffs affect the ratio of imports to inward FDI sales. A key focus is how the use of tariff revenues—either as household transfers or investment subsidies—alters the macroeconomic effects of tariffs. We find that unilateral US tariffs combined with transfers increase US consumption, reduce imports, and raise inward FDI. However, higher production and investment costs lower total US investment. A real dollar appreciation mitigates the impact on RoW exporters but raises costs, generating negative spillovers abroad. When tariffs are used to fund investment subsidies, US investment turns positive and expansionary effects are stronger, especially for FDI inflows. The resulting US investment boom raises global interest rates, intensifying negative spillovers to the RoW. The way tariff revenues are recycled also shapes the outcome under retaliation. With transfers, the US is more adversely affected due to higher openness raising production and investment costs. With subsidies, this ranking reverses: greater openness leads to higher tariff revenues relative to GDP, enabling more investment support and cushioning the effects of retaliation.Using Dynamic Financial Incentives to Form Physical Activity Habits (D9, I1)
Abstract
One major challenge of using financial incentives to induce behavioral change is that people often revert when the incentives are removed. This paper investigates if robust habits can be formed by offering personalized goals, coupled with daily incentives sustained over a long period of time. Five hundred adults were given wearable step trackers and were incentivized to walk 2,500 more steps per day than they normally would. Participants were randomly assigned to a no-incentive control condition, or one of four treatment conditions that offered up to SG$14 (US$10) per week if step goals were met every day. Two conditions offered a fixed SG$2 incentive per day to achieve the step goals, in either gain or loss framing. Third condition offered incentives that dynamically increased for each consecutive day the goal was met. Fourth condition offered incentives that dynamically increased for each consecutive day the goal was not met.The incentives significantly increased the frequency of goal attainment during the 36-week intervention period. The effect sizes were similar across the four treatment conditions. Most importantly, we found almost perfect persistence of goal attainment rates during the 12-week post-intervention period, with just 1 percentage-point decline in the average goal attainment rate. There is compelling evidence of habit formation: the dynamic incentive conditions incentivized participants to exercise in unique path-dependent patterns during the intervention period, and these participants continued to exercise in the respective patterns after the incentives were removed.
Vertical Integration and Consumer Behavior: The Impact of the OptumRx-Catamaran Merger on Medicare Part D (I1, L1)
Abstract
In recent years, the growing trend of vertical integration between insurers and pharmacy benefit managers (PBMs) has raised concerns about its impact on market competition and consumer welfare. This paper examines how the 2015 merger between UnitedHealth’s OptumRx and Catamaran, the last major stand-alone PBM, affected consumer behavior in the Medicare Part D market. PBMs play a crucial role in negotiating drug prices, managing formularies, and administering prescription drug benefits. While vertical integration has the potential to improve efficiency and reduce costs, it may also reduce competition, leading to higher prices and fewer choices for consumers. Using individual-level data from the 2010–2022 waves of the Health and Retirement Study (HRS), we analyze changes in beneficiaries’ enrollment decisions, prescription drug spending, and healthcare utilization following the merger. Our findings show that the merger led to a decline in enrollment in stand-alone prescription drug plans (PDPs) and increased out-of-pocket (OOP) drug expenditures among remaining enrollees, particularly those with higher medication needs. These effects were more pronounced among lower-income and less healthy individuals. Beyond prescription drug use, the merger also influenced healthcare-seeking behavior—PDP enrollees reduced outpatient visits, and overall Medicare beneficiaries became less likely to enroll in Medicare Advantage Prescription Drug (MAPD) plans. We identify higher PDP premiums, lower plan quality, and more restricted drug coverage after the merger as key factors driving these behavioral changes, highlighting the potential consumer harms associated with vertical integration in the PBM sector.Water Contamination and Real Estate Prices: The Effect of Newark’s Lead-in-Water Crisis (R3, Q5)
Abstract
I use the lead-in-water crisis in Newark, New Jersey, as a natural experiment to assess the impact of water contamination on real estate prices. Leveraging publicly available real estate transaction data with unique property identifiers, I merge it with shapefiles containing geospatial coordinates for each property. The resulting dataset is then further merged with a shapefile delineating Newark boundaries. The treatment group comprises properties in the city's western section, served by the Pequannock water system, while the control group consists of properties in areas served by the Wanaque water system.Using a difference-in-differences approach, I analyze the variation in real estate prices in contaminated areas relative to uncontaminated areas, before and after March 2016. This date marks the identification of elevated lead levels in the drinking water of several public schools served by Pequannock water system, a contamination that continued to exceed the federal action level of 15.0 micrograms per liter (μg/L) in the Pequannock-served areas until the end of December 2019. These violations were the first recorded breaches of federal lead standards following nearly two decades of compliance. The findings indicate an approximate 13 percent decline in property values within affected areas.
This paper makes two primary contributions. First, it is the first study to examine the effects of water contamination in Newark on real estate prices. Although the Newark crisis shares lead contamination similarities with the Flint crisis, its cause differs; while Flint was triggered by a change in water supply, Newark resulted from an inadequate corrosion control system at the Pequannock Water Treatment Plant. Second, given that the real estate and leasing sector contributes over 16 percent to New Jersey’s economy, these findings have broad implications for understanding the economic consequences of environmental crises on local housing markets.
When Refinancing meets Monetary Tightening: Heterogeneous Impacts on Spending and Debt via Mortgage Flexibility (E2, G5)
Abstract
How does monetary tightening affect borrowers’ consumption when they can leverage mortgage flexibility to smooth its impact? We study the UK’s post-2021 monetary tightening, where periodic refinancing exposed mortgagors to sharp payment shocks. The UK’s short fixed-rate mortgage terms create strong incentives for households to refinance rather than revert to higher Standard Variable Rates, providing a natural setting to examine refinancing behavior under tighter monetary policy.Using high-frequency transaction data linked to loan-level mortgage records, we analyze households’ monthly balance sheets around refinancing episodes. We find that one-in-three households actively re-optimizes borrowing when facing higher interest rates, either by extending loan terms to reduce monthly payments or through collateral-driven borrowing—leveraging property wealth to increase borrowing via home equity extraction. This response was particularly pronounced among those who benefited from prior house price appreciation, allowing them to offset cash-flow constraints by borrowing against property wealth. Households with low liquid savings or low incomes were also more likely to extract home equity to sustain spending and rebuild financial buffers.
Our findings reveal a trade-off between interest costs and liquidity: despite rising borrowing costs, some households used mortgage flexibility—particularly collateral-driven borrowing via home equity extraction—to smooth consumption and consolidate unsecured debt. However, this mechanism was not universally available. Those unable or unwilling to adjust their borrowing terms experienced sharp consumption declines of around 3%, consistent with the cash-flow channel of monetary transmission.
In aggregate, we show that mortgage flexibility weakened the pass-through of monetary policy to spending. This uncovers a previously undocumented mechanism through which refinancing structures and house price dynamics shape household responses to interest rate shocks, highlighting the critical role of mortgage design in determining economic resilience. The effectiveness of monetary tightening depends not only on borrowing costs but also on households’ ability to adjust mortgage terms.