Economics

We document that women’s economic mobility improved nearly a century before married women gained broad labor market opportunities. Using Massachusetts marriage registers linked to U.S. censuses (1850–1920), we create new father–child links for women to estimate intergenerational mobility and assortative mating, overcoming a key historical linkage barrier. Estimates from a structural marriage market model suggest assortative mating fell 61% from 1850–1870 to 1900–1920. Counterfactuals imply women’s mobility would have been far lower absent the decline in assortative mating. Had late cohorts faced early cohort sorting, the rank–rank slope between a woman’s father and husband wo..

We conducted a within-subject laboratory experiment in which participants decided whether to experience physical discomfort for charity, with or without additional personal compensation. Acceptance decreased with greater discomfort and increased with both larger charitable donations and personal payments. We show that private monetary incentives and prosocial benefits interact in a less-than-additive way: personal compensation raises participation but attenuates the marginal impact of charitable donations, making the combined impact of private and social rewards smaller than the sum of their separate effects. We also find suggestive evidence that the sequencing of compensated and uncompensat..

We develop a dynamic quantitative model of trade and labor adjustment, incorporating nominal wage rigidity and consumption–saving decisions, to study how China’s currency peg interacted with its rapid growth in shaping the US economy. We show that the peg temporarily boosts China’s export growth by preventing an appreciation of the Chinese currency, thereby amplifying the US labor-market consequences of the China shock. At the same time, the temporary export boom increases China’s savings and leads to a larger US trade deficit. Calibrating the model to match trade and labor-market flow data, we find that China’s currency peg played a quantitatively important role in the US manufact..

Adjustable-rate mortgages (ARMs) transmit monetary policy less directly than often assumed. We exploit quasi-experimental variation in ARM rate reset timing in Portugal—where over 92% of mortgages are indexed to Euribor—around the ECB’s 2022–2023 tightening cycle to estimate responses to mortgage payment shocks. After reset dates, mortgage renegotiations increase by 10 percentage points, lender switching by 4, partial prepayments by 5, and full prepayments by 3, offsetting about 17% of the payment increase implied by policy rates. Responses occur only immediately after resets, consistent with selective inattention, and are largest among younger, more educated, and higher-balance borr..

Dynamic complementarity is the concept that past investments that lead to higher stocks of skill at one age promote the growth of skills from investment at that age. We define and provide evidence on dynamic complementarity using unique Chinese data from a home visiting program for young children targeted to parents in rural China. In addition, we investigate growth in learning due to innate, parental, and environmental factors that occur in the absence of any formal intervention.

NFTs provided an extraordinary real-time laboratory for bubble economics: returns were exceptionally right-skewed, illiquidity pervaded even the most active platforms, and a handful of trades drove aggregate performance. Investors extrapolating from realized returns without recognizing selection bias and survivorship faced a substantial risk of disappointment. As our data and simulations confirm, successful NFT investing during the bubble required an almost perfect confluence of timing, liquidity, and luck.

Frontline workers in public bureaucracies perform multidimensional tasks and rely on supervisors who both monitor performance and supply essential managerial inputs. We study how incentive design should account for these interactions using a conceptual framework and a province-wide randomized experiment with the Punjab Agriculture Extension Department in Pakistan. Across 131 tehsils, we compare: (i) an Objective pay-for-performance scheme tied to digital activity metrics; (ii) a Subjective scheme in which supervisors allocate bonuses; and (iii) a Subjective Plus scheme that preserves discretion but introduces light-touch oversight of supervisors. All three schemes increase outreach, but thro..

Natural disasters undermine economic development, destroying lives, livelihoods, the built and natural environment. Public assistance programs follow, but knowledge of their reach, impact and evolution is limited. We address this gap, analyzing population-representative longitudinal survey data collected in Indonesia before and for two decades after the 2004 Indian Ocean tsunami which was followed by a large recovery program involving multiple major actors. Those who lived, at the time of the tsunami, in communities that were badly damaged, were far more likely to receive aid and received larger amounts than residents of other areas immediately after the tsunami. Those gaps were small and in..

Voting-based collective decisions are typically made either anonymously or publicly. Anonymous voting protects truthful expression but conceals individual behavior; public voting provides information about individual votes, but, when one option is socially stigmatized, it can distort participation and choices. We introduce threshold majority voting, in which voters choose a disclosure threshold determining whether and when their votes are revealed. In an experiment at UC Berkeley on the participation of transgender women in women’s sports, public voting nearly doubles abstention and reduces support for the stigmatized option. Threshold voting eliminates these distortions while revealing on..

Despite the large common net benefits of climate mitigation, broad-based political consensus for large-scale policy action remains elusive. We hypothesize that financial exposure to energy stocks central to the green transition can induce learning and greater support for climate mitigation policies. We conduct a RCT which randomizes both the presence of financial market exposure to the energy sector, as well as which type of portfolio—fossil-fuel (brown) or renewable energy (green)—is given to an individual. Treatment increases support for mitigation action and intent to undertake adaptation, with positive support caused by ownership of both green and brown assets. The effects are partic..

This paper tests whether the arrival of mass media triggers a decline in fertility, a central prediction of modernization theory. Using a field experiment, we vary exposure to mass media and its content in a quarter of Burkina Faso. We provide radios to 1,600 women without previous access to mass media. Half live in status quo areas and half in areas where the local radio station was randomly selected to air a science-based family planning campaign. Contrary to modernization theory and previous literature, gaining access to status quo mass media decreases contraception use by 14 percent and reinforces traditional gender norms. In contrast, receiving a radio in campaign areas boosts contracep..

Can questions about life satisfaction be used to measure parental preferences for daughters versus sons? Daughter preference has rarely been documented in the literature, even in matrilineal settings. One possible reason is that the commonly used measures of parental gender preference, such as fertility-stopping rules and sex ratio at birth, are ill-suited to high-fertility settings. We instead assess maternal preferences in Malawi by examining the life satisfaction of women who currently have one child, comparing those with a daughter to those with a son. We find that in matrilineal (but not patrilineal) households, having a daughter increases mothers' life satisfaction, relative to having ..

Women’s employment remains highly sensitive to childcare constraints, making childcare availability a critical lever for supporting mothers’ labor force attachment. We study the effects of expanded full-day programming in Head Start, using the 2016 federal funding initiative that targeted grantees with low full-day enrollment. Linking administrative program data, geo-coded center locations, and household data on employment, we estimate a difference-in-differences design by comparing mothers of young children in treated and untreated areas. The policy increased full-day enrollment by 19 percent and raised single mothers’ employment (1.9%), hours (2.5%), and earnings (6.5%). Results show..

Gender norms—shared expectations about appropriate behavior by gender—shape the outcomes of men and women across societies, and are correlated with GDP per capita. This chapter surveys the literature on gender norms and economic development, focusing on the pervasive and traditional ‘male breadwinner norm’ that assigns men to market work and women to domestic responsibilities. We review empirical approaches to measuring norms, including direct survey questions on attitudes, second-order beliefs about others' views, and behavioral proxies. Establishing causal effects of norms on behavior poses significant challenges, and we review a range of approaches to identify this link. We then p..

We present the GABRIEL software package, which uses GPT to quantify attributes in qualitative data (e.g. how “pro innovation” a speech is). GPT is evaluated on classification and attribute rating performance against 1000+ human annotated tasks across a range of topics and data. We find that GPT as a measurement tool is accurate across domains and generally indistinguishable from human evaluators. Our evidence indicates that labeling results do not depend on the exact prompting strategy used, and that GPT is not relying on training data contamination or inferring attributes from other attributes. We showcase the possibilities of GABRIEL by quantifying novel and granular trends in Congress..

It is widely believed that procrastination contributes to under-saving. Using household survey data from the United States and Singapore, we find little evidence to support this view. In both countries, indicators of procrastination have only weak, if any, explanatory power for saving regret. About half of respondents aged 60–74 expressed regret about not having saved more, but this sentiment was primarily associated with exposure to economic shocks rather than procrastination. A larger share of U.S. respondents reported saving regret, reflecting their greater exposure to adverse shocks. In contrast, Singapore’s system of mandatory saving for retirement, housing, and health expenses appe..

We present the first representative international data on firm-level AI use. We survey almost 6000 CFOs, CEOs and executives from stratified firm samples across the US, UK, Germany and Australia. We find four key facts. First, around 70% of firms actively use AI, particularly younger, more productive firms. Second, while over two thirds of top executives regularly use AI, their average use is only 1.5 hours a week, with one quarter reporting no AI use. Third, firms report little impact of AI over the last 3 years, with over 80% of firms reporting no impact on either employment or productivity. Fourth, firms predict sizable impacts over the next 3 years, forecasting AI will boost productivity..

Investment in equipment and structures is one of the most cyclical components of GDP, a fact often associated with a negative response to heightened uncertainty in recessions. R&D investment, by contrast, is only mildly procyclical. We show that this difference could arise because of a positive response of R&D investment to uncertainty - promoting more research, development, and experimentation - a feature of most recessions but most notably during COVID. Both effects are distinct manifestations of real options, one in which costly reversibility delays investment, and the other in which investment enhances resolution of uncertainty.

Little is known about the impact on small-scale enterprises of a large negative shock that destroys assets and disrupts local markets. We document the short- and long-run impacts of the 2004 Indian Ocean tsunami using longitudinal household survey data. We leverage topography-driven variation in exposure to the tsunami in coastal Aceh and North Sumatra, Indonesia. There are large short-run declines in business ownership, real profits, and real business assets among those exposed to the tsunami relative to comparison individuals who were not directly exposed. The gap in ownership rates disappears within two years in the non-agricultural sector but persists for 15 years in the agricultural sec..

The implementation of the North American Free Trade Agreement (NAFTA) in 1994 resulted in a great restructuring in industry composition in the US, with substantial heterogeneity across local areas. In this paper, we investigate the effects of NAFTA on the mortality rates of the working-age population. We implement event studies and difference-in-difference analyses to examine dynamic changes in mortality rates in different years relative to NAFTA and in areas with differential exposure to NAFTA. Comparing areas with high versus low trade exposure measures, we find a 2.1 percent rise in the mortality rate of those aged 25-55. A back-of-the-envelope calculation suggests a 7.3 percent rise in m..

Work-related burnout and stress-related sickness absence have become increasingly prevalent, but evidence on which workplace features shape workers’ mental health remains limited. Using population-level Swedish register data covering all lower- and upper-secondary teachers from 2006–2024, we show that schools serving more disadvantaged students exhibit substantially higher rates of sickness absence, particularly for stress-related diagnoses. Exploiting within-teacher variation across student cohorts, we separate sorting from exposure and find that a one standard deviation increase in student disadvantage raises overall and stress-related sick leave by 3.6% and 8.7%, respectively. Survey ..

We examine how workers perceive the trade-offs of freelancing using a novel survey design that explores the nature of workers' perceptions of their own jobs and the implications of work arrangements for their take-home pay. We find that, across several alternative classifications of freelance work, workers in such arrangements make less per hour than traditional employees, but report having greater control of when, where, and how they work. We find that on average, self-employed workers spend an additional 5 to 8 percentage points of gross pay covering unreimbursed expenses relative to traditional employees. However, when asked about expectations of net pay in freelance and traditional emplo..

Using confidential transaction-level data on overnight tri-party Treasury repo from 2018–2025, we study liquidity flows within bank holding companies (BHCs) between primary dealers and affiliated entities. Affiliates, especially depository institutions with abundant reserves, provide a large share of dealers’ repo funding, yet affiliated repo borrowing is systematically more expensive: dealers pay an affiliation premium of about 4–5 bps relative to otherwise comparable unaffiliated trades, even after controlling for trade characteristics and market conditions. We interpret this premium as an internal transfer price. Because consolidated leverage regulation nets intra-BHC positions and ..

We examine how performance changes when teachers transfer across very different school contexts. The Talent Transfer Initiative program created a rare natural experiment to study such transfers by randomly assigning low-achieving schools the ability to offer high-performing teachers at higher-achieving schools a $20,000 transfer stipend. Forecast tests show that these high-performing teachers’ prior value added is only moderately predictive of their effectiveness in low-achieving schools. Using a difference-in-differences framework, we estimate that incentivized-transfer teachers’ value added dropped by 0.12 student standard deviations. This decline appears to be driven by lower match qu..

Allocation mechanisms in M&A deals are complex, but a main feature is that a target board controls who to invite to the sale. In a theoretical model, we show that it is optimal for the target to invite fewer potential acquirers when they are more homogeneous (i.e., when their values for the target are more correlated). Furthermore, greater correlation (and hence a smaller optimal bidder pool) yields the target a higher surplus from the sale (i.e., higher premium). We test the model empirically and show that M&A deals with smaller bidder pools are associated with higher target returns. This is not a result of synergies in the deals: the target's share of the surplus is simply higher in deals ..

This paper compares the impacts of critical mineral price and oil price on an economy in a unified neoclassical growth model. Unlike oil price shocks, which affect the cost of utilizing existing capital (e.g., cars), critical mineral price shocks influence the cost of creating new capital (e.g., electric vehicles) without altering the cost of existing capital. We find that both types of shocks ultimately reduce output and welfare. However, oil-price increases are systematically more contractionary for the economy. Mineral-price increases generate comparatively larger adjustments in investment, capital, and external borrowing but smaller and more gradual losses in output and welfare, and in c..

We use frontier advancements in Artificial Intelligence and machine learning to extract and classify the part of key economic agents’ behaviors that are predictable from past behaviors. Even the agents themselves might view these as novel (innovative) decisions; however, we show in strong contrast that a large percentage of these actions and behaviors can be predicted—and thus mimicked—in the absence of these individuals. In particular, we show that 71% of mutual fund managers’ trade directions can be predicted in the absence of the agent making a single trade. For some managers, this increases to nearly all of their trades in a given quarter. Further, we find that manager behavior i..