Essays On Racial Inequality In The Labor Market

Abstract

This dissertation studies issues related to racial inequality in the labor market. The first two chapters focus on the impact of racially segregated referral networks on inequality and aggregate welfare, while the final chapter focuses on differences in returns to work experience and the supply of labor between black and white workers in recent decades. The first chapter shows that there are racial differences in the composition of referral networks and the use of referral networks by occupation. In particular, non-college black and white workers in the United States who obtain a job via referral display substantial social segregation, using same-race contacts around 90% of the time. While non-college black and white workers use referrals at a similar rate overall, black workers use referrals for higher-skill and higher-paying occupations at a lower rate than white workers. I also document racial differences in occupational choice, with white workers sorting into higher-skill occupations. The following chapter connects and rationalize these observations by incorporating a referral-based matching function into a standard search and match model with occupational choice, heterogeneous ability levels, free entry, and wages determined by Nash bargaining. Social segregation can lead to differences in occupational choice by race, and thus wage and employment inequality, in the steady state. After calibrating the model to examine black and white workers in the United States, the estimates show that racially biased networks alone can generate a black-white wage gap of 1.66 percent and an employment gap of 0.74 percentage points. Moving from the segregated to the desegregated steady state harms the majority white workers while helping the minority black workers, resulting in a decrease in aggregate welfare. In the final chapter I utilize individual fixed effects combined with an instrumental variables approach to document the extent to which returns to work experience differ for black and white workers; I then use a life-cycle model with a learning-by-doing human capital production function to assess the consequences of these differences for the supply of labor. Returns to an extra thousand hours of work experience for the typical white worker are 23 cents per hour in 2012 USD (amounting to an additional $478 per year of full time work), compared to 12 cents for an otherwise identical black worker (amounting to an additional $250 per year). Using a life-cycle model, differences in returns to experience combined with simulated differences in choices of hours worked can account for approximately 10 percent of the measured difference in average wages over the life-cycle between black and white workers.