Unemployment inflow and sufficient statistics for welfare analysis of unemployment insurance

Abstract

A common tool to perform welfare analysis of unemployment insurance is the sufficient statistics approach, where the benefit is consumption smoothing across unemployment and employment, and the cost results from lower unemployment outflow. This paper incorporates unemployment inflow into a modified Baily-Chetty model and recovers the cost statistic for local welfare analysis of unemployment insurance. Empirically I estimate this cost statistic using an eligibility variation that took place in Brazil in 2015. I show that unemployment inflow is the main source of moral hazard costs, a feature, by and large, missing in most welfare analysis of unemployment insurance. Performing an empirical analysis at the firm level, I also show that there's no substitution of who is dismissed when eligibility changes, a hypothesis I rely on the local welfare analysis. Finally, I extended the model to allow for individual heterogeneity and show how to recover the parameter of interest even when unemployment inflow and outflow responses are heterogeneous.