Knowledge that Transforms

This paper estimates the causal impacts of workers’ compensation income benefits on workers’ health and welfare outcomes. Using claims data from 2004 to 2016, I explore the variation in benefits due to a reform of New York workers’ compensation that increased the maximum weekly benefits. I find that a $77 increase in the weekly benefits led to an additional 3.4 days off work. Medical utilization did not increase. Each extra day off work decreased the reinjury likelihood by 2.9%. The current benefit level in New York is close to optimal in balancing payer cost and worker health outcomes.

As wages in migrant-sending countries catch up with those in destinations, migrants adjust on several margins, including their duration of stay, the number of migrations they undertake, and the amount saved while abroad. This paper combines Mexican and US data to estimate a dynamic model of consumption, emigration, and remigration, accounting for financial constraints. An increase in Mexican household earnings shortens migration duration but raises the number of trips per migrant. For lower-income migrants, a rise in Mexican wages leads to a more than proportional effect on consumption expenditure in Mexico, arising from repatriated savings.

Analyses of survey data highlight observable person characteristics, such as education and occupation, as critical factors for rising earnings inequality, while industry has an offsetting effect. In contrast, analysis of administrative data highlights that rising between-firm earnings dispersion and, in turn, between-industry earnings dispersion dominates the rise in earnings inequality. We construct a novel integrated dataset based on CPS micro data linked with LEHD administrative records. We find that most of the rise in earnings inequality is accounted for by rising between-industry inequality. This finding reflects a substantial contribution of increased sorting and segregation of observable person characteristics between industries.

We estimate cyclicality in labor’s user cost allowing for cyclical fluctuations in the quality of worker-firm matches and wages that are smoothed within employment matches. To do so, we exploit a match’s long-run wage to control for its quality. Using 1980–2019 National Longitudinal Survey of Youth data, we identify three channels by which recessions affect user cost: they lower the new-hire wage and wages going forward in the match, but they also result in higher subsequent separations. We find that labor’s user cost is highly procyclical, increasing by more than 4% for a 1 percentage point decline in unemployment.

This paper develops measures of the occupational homogeneity of employers as indicators of outsourcing. Findings are threefold. First, wages are strongly related to occupational homogeneity, particularly for workers in low-wage occupations. Second, by some measures, workers—particularly those in higher-wage occupations—saw their employing establishments become more occupationally homogeneous during 2004–19. Third, changes in the occupational homogeneity of workplaces contributed to growing wage inequality among workers over the first part of this period. The growing sorting and segregation of workers by occupation into different employers is an important part of wage inequality.

We study the effect of hearing cases alongside female judicial colleagues on the probability that a federal judge hires a female law clerk. Federal judges are assigned to judicial panels at random and have few limitations on their choices of clerks. Using a unique dataset of federal case records merged with judicial hiring information, we find a significant effect of the fraction of copanelists who are female on a male judge’s likelihood of hiring a female clerk. This finding suggests that increases in the diversity of the upper rungs of a profession can create opportunities at the entry level.

Many Americans continued some forms of social distancing after the pandemic. This phenomenon is stronger among older persons, less educated individuals, and those who interact daily with persons at high risk from infectious diseases. Regression models fit to individual-level data suggest that social distancing lowered labor force participation by 2.4 percentage points in 2022, 1.2 points on an earnings-weighted basis. When combined with simple equilibrium models, our results imply that the social distancing drag on participation reduced US output by $205 billion in 2022, shrank the college wage premium by 2.1 percentage points, and modestly steepened the cross-sectional age-wage profile.

We investigate the impact on work absences of a large reduction in paid sick leave benefits in Spain. Our results highlight substantial decreases in frequency (number of spells) mostly offset by increases in duration (length of spells). Overall, the policy did reduce the number of days lost to sick leave. For some, however, return to work was premature, as we document large increases in both the proportion of relapses and the number of working accidents. Displacement toward this unaffected benefit scheme cancels out almost two-fifths of the gains in terms of estimated absence reductions from the sick leave benefit cut.

We explore how access to Head Start impacts maternal labor supply. By relaxing child care constraints, public preschool options like Head Start might lead mothers to reallocate time between employment, child care, and other activities. Using the 1990s enrollment and funding expansions and the 2002 Head Start Impact Study randomized control trial, we show that Head Start increases short-run employment and wage earnings of single mothers. The increase in labor supply does not appear to reduce quality parent-child interactions. Viewing Head Start as a bundle of family-level treatments can shed new light on the impacts of the program beyond children.

This paper assesses the importance of school and neighborhood peers in shaping educational achievement, adolescent misbehavior, and adult crime. Using cohort variation within Charlotte-Mecklenburg County, we focus on the impact of peers whose parents have been arrested, which is strongly and independently predictive of worse outcomes. Results indicate that a 5 percentage point increase in school peers linked to parental arrest reduces educational achievement by 0.016 standard deviations and increases adult arrest rates by 5%. Additional evidence indicates that peer effects are primarily driven by interactions in schools rather than in neighborhoods.