Essays in Labor Economics

Thumbnail Image



Journal Title

Journal ISSN

Volume Title

Repository Usage Stats



The essays in this dissertation explore how barriers to advancement may lead to path dependence in economic outcomes. One uses a finite mixture model to demonstrate that some police officers are more likely than others to stop African-American drivers. The conclusion is one that though widely believed has proven challenging to establish empirically. By doing so, the analysis makes two contributions. First, it more closely aligns with the understanding of racial profiling. While disproportional susceptibility to vehicle searches exemplifies profiling, being pulled over is a much more common margin for police interaction, which this paper models a tractable way of identifying. Second, studies of secondary decisions such as searches frequently assume that there is no bias in the initial stop decision. An analysis of traffic stops across eight states questions this assumption, concluding that stopped drivers constitute a selected sample. Although bias is theoretically continuous, average behavior fits well into two distinct groups, with 30–40% of officers in the group that exhibits a relatively high propensity to stop minority drivers. The implication is that race-based policing is more prevalent than the "rotten apples" theory might suggest. The second essay asks how job loss is distributed in an unexpected downturn, examines the equity consequences, and evaluates policy responses. After demonstrating two novel empirical patterns relating job loss risk to earnings, it constructs and estimates a search model with heterogeneous job termination rates to explain the prevailing dynamics, and then uses it to evaluate the impact of large-scale transfers, unemployment insurance, and an earnings subsidy during a simulated downturn. Transfers have a minimal employment effect unless the shock also reduces savings. Raising unemployment benefits by 30% leads to a 0.6 point increase in the unemployment rate. Adding an earnings subsidy to unemployment benefits raises employment by 7 percentage points at a 23% lower level of inequality yet increases costs by only 16-30%. The tradeoff is lower employment at better paying jobs versus higher consumption at subsidized low-wage jobs. In contrast to a compensating differentials framework, the model shows that lower earners tend to be more at risk for layoffs, and that this effect is stronger following a negative aggregate shock. Following a simulated shock, mean consumption falls by 5% yet consumption at the 10th percentile drops by 16%.






Abrahams, Scott (2022). Essays in Labor Economics. Dissertation, Duke University. Retrieved from


Dukes student scholarship is made available to the public using a Creative Commons Attribution / Non-commercial / No derivative (CC-BY-NC-ND) license.