Essays in Labor Economics

This dissertation consists of three chapters. In the first two chapters I study factors that influence the entry of individuals into military service. In the last chapter, my co-authors and I look for evidence of peer effects in marriage and fertility decisions among junior enlisted soldiers in the U.S. Army.

In Chapter one I use variation in advertising exposure across the universe of U.S. television markets over the period 2010-2019 to identify the effect of advertising by components of the Department of Defense (DoD) on the quantity and quality of U.S. Army applications and hires. To estimate advertising effectiveness, I employ two identification strategies. In the first strategy, I use variation in advertising exposure driven by different levels of viewership across TV markets. In the second strategy, I isolate variation in advertising exposure along the borders of TV markets. I estimate long-run own-advertising elasticities of applications and hires of 0.007-0.013 and 0.006-0.008, respectively. These estimates imply a marginal cost per applicant of $1,027-$1,953 and a marginal cost per hire of $3,155-$5,014. However I find consistent evidence of negative spillovers in advertising across DoD components. On the margin, every dollar spent on advertising by the Army can be undone by an additional dollar of expenditure by another DoD service. This suggests that TV advertising primarily shifts applicants among DoD components without substantially expanding overall labor supply. The existence of combative advertising by DoD components contributes to an inefficiently high level of public expenditure on DoD TV Advertising. (JEL L10, J30, M31, M37)

In Chapter two I consider how the demographic composition of recruiters affects the racial and gender composition of applicants and hires. Using detailed administrative data covering all applications and enlistments into the U.S. Army from September 2005-September 2019 (1.5 million applicants, 847 thousand enlistees, and 29 thousand recruiters), I show that frictions in the selection and assignment of Army recruiters to recruiting stations leads to plausibly random within-recruiting station variation in the presence of a minority-race or female recruiter after conditioning on variables that could influence the recruiter assignment decision. I find that Black and Hispanic recruiters have modest positive effects on same-type applications and enlistments, but no effect on total enlistments. I find no evidence that female recruiters increase female applications or enlistments. I confirm my results by estimating an unmatched and matched difference-in-difference to identify the effect of recruiter race and gender around recruiter arrivals and departures. I find evidence that the same-type shift in applications and enlistments is strongest upon arrival, with only Black recruiters having a statistically significant effect at departure. Over the 15 years of the sample, I estimate that Black recruiters increased the fraction of Army hires that were Black by 0.24 percentage points (1.2 percent). (JEL J23, J30, J71)

Chapter three is a joint project with Susan Carter and Abigail Wozniak. A large literature links marriage to later life outcomes for children and adults, but U.S. marriage rates have declined markedly in recent decades, particularly for those with less education. This connection raises concerns that declining marriage might exacerbate inequality in outcomes like children's achievement and adult longevity. In this paper, we provide causal evidence on one mechanism for marriage rate patterns: peer effects. We use exogenous assignment to peer groups for 18-24 year-olds in the U.S. Army to identify the impact of peer marriage on individual marriage choices in our sample. We find that peer choices have statistically significant but economically small impacts on own marriage. These precisely estimated, small impacts are robust to a range of sample restrictions focusing on different demographic groups. We argue that our U.S. Army sample can potentially identify upper bounds on peer impacts on marriage. Our precisely estimated, small peer impacts suggest that other factors are likely more important for explaining the decline in marriage rates. (JEL J12, J13)