Essays on Social Insurance and Labor Markets

Abstract

This dissertation consists of three chapters. In the first chapter, I study the macroeconomic and welfare consequences of introducing a universal healthcare system to replace the existing employer-based set-up in the U.S., paying particular attention to the reform's labor market effects. To study the policy, I develop an incomplete asset markets model with labor market frictions and medical expenditure risk over the life cycle. First, I compare the model-implied partial equilibrium employment responses to public health insurance generosity to existing empirical evidence. The model partially reconciles the puzzlingly wide range of estimates found in three microeconomic experiments conducted in Tennessee, Oregon, and Wisconsin. Next, I use the model to understand the general equilibrium effects of switching to universal healthcare. I find that it results in higher reservation wages and a corresponding reduction in firm vacancy creation, both of which lead to a quantitatively large decline in the job finding rate. The negative impact of the lower job finding rate outweighs the insurance benefits of generous public health coverage, resulting in substantial welfare losses among low-wealth households for whom employment is most valuable. In the second chapter (joint with Serdar Birinci), we investigate how unemployment insurance generosity should vary with the business cycle. We find that the optimal policy is countercyclical. Not only does the policy smooth the consumption of job losers, but also provides insurance against aggregate risk by reducing the need for excess savings during recessions. Meanwhile, the moral hazard effects of generous benefits are attenuated in recessions because jobs are scarce and thus, the forgone value of job search is low. In the third chapter (joint with Anmol Bhandari, Serdar Birinci, and Ellen McGrattan), we study survey data used for measuring business income and valuations. We document large inconsistencies between survey data and aggregated administrative data for statistics such as the level and distribution of business income, and the number of returns. These inconsistencies are attributable to both non-representative samples and measurement errors. Non-representativeness results from undersampling businesses with low income owners. Measurement errors emerge because respondents do not use relevant financial documents as basis for their responses while some survey questions suffer from framing problems.