Research

Among U.S. college graduates, the long workweek has largely disappeared. Since 2000, the share working 50 or more hours a week has fallen by half, from 28 to 14 percent, while the share working exactly 40 has risen from 54 to 75 percent. Using the CPS, SIPP, and NLSY, I document four facts. First, workers moved from working long weeks to working exactly 40. Second, the decline is a calendar-time shift that reached every age at once, with younger workers cutting first. Third, the within-worker price of long hours did not fall. Fourth, long hours stopped leading to promotion around 2000 and stopped leading to faster earnings growth after 2008. To explain these facts, I build a life-cycle directed search model in which hours are a term of the contract, working 50 or more hours raises the probability of promotion, and hours beyond 40 carry a fixed cost. The loss of the promotion channel accounts for about half of the decline in long hours and explains why the young led. The collapse lowers lifetime utility on average, with losses concentrated among workers who preferred long hours.

We leverage a historical natural experiment to provide new evidence on the long-run causal effects of parental wealth. Beginning with the Land Rush of 1889, settlers raced to claim homesteads in Oklahoma Territory, unaware of oil fields hidden beneath their feet. Two decades later, wells began dotting the landscape. Homesteaders who previously earned a living from the land began to receive sizable production royalties. We link land patents, oil wells, and census records to follow homesteaders and their children over decades, and leverage highly local variation to compare homesteaders who likely received royalties with their neighbors who did not. We find that oil wealth shocks caused homesteaders to retire and move to towns and cities at higher rates. Their children received an additional year of education and were 16 percentage points more likely to graduate from high school. Children owned more valuable homes and may have worked in higher-earning occupations. We further identify that children in close proximity to oil extraction but without a parental wealth shock worked in higher-earning occupations and married earlier, but owned less valuable homes and may have died earlier than children farther from oil.

In this paper, we build a model of the labor market with search frictions in which workers choose how much to invest in a multidimensional set of skills before entering the labor force. We show that search frictions have important consequences for skill allocation, distorting skill choices and affecting the quality of matches in the labor market. Workers become less specialized as a way to insure against labor market risk, reducing the match value between workers and firms. This induces a novel type of cost of search frictions, which we call skill distortion. We propose a method to separately identify the output cost of skill distortion in our model. We show it is quantitatively relevant and that search frictions account for around one third of total mismatch when compared to preference heterogeneity. We also show that pre-labor market skill accumulation can have important consequences for the design of unemployment insurance policy by endogenously reducing mismatch in response to more generous insurance benefits.

Measuring occupational mobility from the Current Population Survey using retrospective or longitudinal methods generates substantially different outcomes, in both levels and trends. Using a generalized method of moments technique, we estimate the level of occupational mobility and the measurement error in both of these measures for 1981–2018. We estimate that occupational mobility has been trending down, particularly since 2000, consistent with retrospective measures of occupational mobility. However, estimated mobility is 2–3 percentage points or 60%–70% higher than retrospective measures. Measurement error in longitudinal measures is large and has been worsening over time.