Essays in Labor Economics

This dissertation consists of three essays. The first chapter is coauthored work with Richard Blundell and James P. Ziliak. We estimate the full distribution of life cycle wages for cohorts of men and women in the US using a quantile selection model to account for systematic differences in employment by gender and education group. Although common within-group time effects are shown to be a key driver of labor-market inequalities across gender, important additional differences by birth cohort emerge with more recent cohorts of women delaying child rearing, and by implication the onset of child penalties in wages. These cross-cohort differences help account for the stalling of progress in gender wage gaps over the past quarter century. In the second chapter I analyse the accuracy of the Baily-Chetty formula for the study of unemployment insurance. In particular, I consider (i) how the approximate measure of consumption smoothing relates to the true consumption smoothing benefit and (ii) the role of heterogeneity for welfare analysis. I present theoretical results that show that certain underlying characteristics of agent preferences (prudence and temperance) can lead to underestimation of these welfare calculations. I then measure the extent to which preference approximations and heterogeneity affect welfare estimates in an empirical application. The third chapter is coauthored work with Toru Kitagawa and Jeff Rowley. We propose a novel method to estimate individualised treatment assignment rules which can be applied to the design of labor market policies such as job training programs. The method is designed to find rules that are stochastic, reflecting uncertainty in estimation of an assignment rule and about its welfare performance. Our approach is to form a prior distribution over assignment rules, not over data generating processes, and to update this prior based upon an empirical welfare criterion, not likelihood. The social planner then assigns treatment by drawing a policy from the resulting posterior. We show analytically a welfare-optimal way of updating the prior using empirical welfare; this posterior is not feasible to compute, so we propose a variational Bayes approximation for the optimal posterior. We characterise the welfare regret convergence of the assignment rule based upon this variational Bayes approximation, showing that it converges to zero at a rate of $\ln\left(n\right)/\sqrt{n}$. We apply our methods to experimental data from the Job Training Partnership Act Study to illustrate the implementation.