Essays in Labor and Public Economics

Abstract

This dissertation studies labor and public economics. Chapter 1 is titled “How Does Unemployment Affect Consumer Spending?” and is coauthored with Pascal Noel. We study the spending of unemployed individuals using anonymized data on 210,000 checking accounts that received a direct deposit of unemployment insurance (UI) benefits. Unemployment causes a large but short-lived drop in income, generating a need for liquidity. At onset of unemployment, monthly spending drops by 6%, and work-related expenses explain one- quarter of the drop. Spending declines by less than 1% with each additional month of UI receipt. When UI benefits are exhausted, spending falls sharply by 11%. Unemployment is a good setting to test alternative models of consumption because the change in income is large. We find that families do little self-insurance before or during unemployment, in the sense that spending is very sensitive to monthly income. We compare the spending data to three benchmark models; the drop in spending from UI onset through exhaustion fits the buffer stock model well, but spending falls much more than predicted by the permanent income model and much less than the hand-to-mouth model. We identify two failures of the buffer stock model relative to the data – it predicts higher assets at onset, and it predicts that spending will evolve smoothly around the largely predictable income drop at benefit exhaustion. Chapter 2 is titled “The Incidence of Housing Voucher Generosity” and is coauthored with Rob Collinson. Most housing voucher recipients live in low-quality neighborhoods. We study how changes in voucher generosity affect neighborhood poverty, unit-quality and rents using administrative data. We examine a policy making vouchers more generous across a metro area. This policy had no impact on neighborhood poverty, little impact on observed quality, and increased rents. A second policy, which indexed rent ceilings to neighborhood rents, led voucher recipients to move to higher quality neighborhoods with lower crime, poverty and unemployment. These results are consistent with a model where the first policy acts as an income effect and the second as a substitution effect. Chapter 3 is titled “A Permutation Test for the Regression Kink Design” and is coau- thored with Simon Jaeger. This chapter proposes a permutation test for the Regression Kink (RK) design—an increasingly popular empirical method for causal inference. Analo- gous to the Regression Discontinuity design, which evaluates discontinuous changes in the level of an outcome variable with respect to the running variable at a point at which the level of a policy changes, the RK design evaluates discontinuous changes in the slope of an outcome variable with respect to the running variable at a kink point at which the slope of a policy with respect to the running variable changes. Using simulation studies based on data from existing RK designs, we document empirically that the statistical significance of RK estimators based on conventional standard errors can be spurious. In the simulations, false positives arise as a consequence of nonlinearities in the underlying relationship between the outcome and the assignment variable. As a complement to standard RK inference, we propose that researchers construct a distribution of placebo estimates in regions with and without a policy kink and use this distribution to gauge statistical significance. Under the assumption that the location of the kink point is random, this permutation test has exact size in finite samples for testing a sharp null hypothesis of no effect of the policy on the outcome. We document using simulations that our method improves upon the size of standard approaches.