Knowledge that Transforms

How does personal exposure to economic conditions affect individual human capital investment choices? Focusing on bachelor’s degree recipients, we find that cohorts exposed to higher unemployment rates during typical schooling years select majors that earn higher wages, have better employment prospects, and lead to work in a related field. Conditional on expected earnings, recessions also encourage women to enter male-dominated fields, and students of both genders pursue more difficult majors. We conclude that economic environments change how students select majors, and we find evidence that students who respond to the business cycle enjoy earnings typical of their new majors.

We analyze the consequences of labor market quotas for the wages of women in high-level positions. Labor market quotas create uncertainty about the reason a woman is promoted. Firms know whether they promoted female employees because of the quota or their ability; their competitors do not. A winner’s curse, reducing competition for women in high-level positions, results. This widens the gender pay gap for these women. Ex ante, women are better off without quotas. Next we investigate how quotas affect incentives for employers to learn women’s abilities to make better job assignment decisions. Then, under specific conditions women may benefit.

Across states, there is little correlation between a state’s marriage rate or cohabitation rate and own population. Within states, there is a positive (no) correlation between a state’s marriage (cohabitation) rate and its population growth rate. The Cobb-Douglas marriage matching function (CDMMF), which extends the Choo-Siow MMF to include peer effects, can rationalize these correlations. The model is easy to estimate. The CDMMF is estimated using panel data across US states from 1990 to 2010. The estimated model replicates the above scale effects. These effects are not sufficient to explain the large recent declines in the gains to marriage.

Survey nonresponse and attrition undermine the validity of many and possibly most econometric estimates. We propose that survey administrators and evaluators proactively create an instrument for observation, for example, by ex ante randomizing participants to differing intensity of follow-up. We illustrate how to apply our proposed methodology using a carefully conducted randomized controlled trial, the Moving to Opportunity demonstration project, which de facto randomly assigned a subset of subjects to more intensive follow-up. The approach yields treatment effect estimates similar to the unbiased estimator based on complete administrative data and has narrower confidence intervals than alternative bounding approaches.

Fewer births occur on major US holidays than would otherwise be expected. We use California data to study the nature and health implications of this birth date manipulation. We document 18% fewer births on the day of and just after a holiday. Cesarean sections account for roughly half of the decline. “Missing” holiday births are displaced to a window of time 11 days before the holiday through 16 days after the holiday. High-risk births are more likely to be rescheduled. Despite the change in timing, we find little evidence of any adverse health consequences for babies born around a holiday.

Using data for three cohorts of women in the PSID, we show that annual earnings inequality fell sharply between the late 1960s and the mid-1990s, with a large decline in the component attributable to the extensive margin. We then fit earnings-generating models that incorporate both intensive- and extensive-margin dynamics to data for the three cohorts. Our models suggest that more than 80% of the decline in female earnings inequality can be attributed to a weakening of the link between family-based factors (i.e., children and the presence and incomes of partners) and the intensive and extensive margins of earnings determination.

The composition of immigrants depends not only on immigrant choice but also on immigration policy, because slots are rationed. Policy determines immigrant attainment, as evidenced by immigrants from Algeria having higher educational attainment than those from Israel or Japan. Theory predicts and evidence confirms that immigrant attainment is inversely related to the number admitted from a source country and positively related to population and education levels at home. A parsimonious specification has only two variables yet explains a majority of the variation in educational attainment of US immigrant groups. The theory and predictions are bolstered by Swedish data.

There is concern that the risky behaviors of teenagers, such as truancy, negatively influence the behaviors of others through their social networks. We use administrative data to construct social networks based on students who are truant together. We simulate these networks to document that certain students systematically coordinate their absences. We validate them by showing that a parent information intervention on student absences has spillover effects from treated students onto their peers. Excluding these effects understates the intervention’s cost-effectiveness by 43%. We show that there is potential to use networks to target interventions more efficiently given a budget constraint.

How does school quality affect health amid multiple behavioral responses? The Rosenwald schools transformed school quality for rural southern African Americans in the early 1900s. Research shows that the schools made black migration northward more likely and that the Great Migration shortened life expectancy for these migrants. Besides the hypothesized health-enhancing effects of school quality, negative health effects might also occur through migration. We disentangle behavioral mechanisms and find complete exposure to the Rosenwald schools increased life expectancy by 2–3 months; a more naive approach finds no relationship. Results are robust to heterogeneous treatment effects and various measurement issues.

In a 2014 paper, Garthwaite, Gross, and Notowidigdo examined the employment impact of the 2005 TennCare contraction. We extend their approach in several directions. First, we use consistent Conley-Taber estimation. Second, we transform their estimates to make them comparable to previous work; the transformed effects have large confidence intervals. We estimate their models using several larger data sets in an attempt to get more precise estimates but find that the results can be quite different. We consider two modifications to account for a major disruption to coverage in 2002, and one of these reduces the differences in the results.