Optimal Unemployment Insurance in a Model With Skill Loss and Match Quality Uncertainty

Abstract

This dissertation makes a contribution to the question of how best to set the rate of unemployment compensation. Previous research on this topic has emphasized the behavioral response of non-workers to various incentives created by unemployment insurance. Recent work has emphasized two new features. One is the importance of including savings in the model, and the other the recognition that skills tend to rise during employment and fall during unemployment spells. This thesis seeks to combine all three features, search incentives, savings, and skill change effects. The strategy is to develop an unemployment model with these features and to obtain parameters values from a variety of sources, including SIPP data and research by other authors on related questions. The model is then simulated for various ranges of policy choices. The primary policy choice is the benefit replacement ratio, a number that determines the actual level of unemployment compensation. Taxes are set under different assumptions. In some cases, taxes are set to achieve budget balance. In other cases, taxes are set independently of benefit levels. This feature assumes the possibility of a subsidy from other sources, but it allows for a study of the independent incentive effects of benefits and tax rates. Results from the simulations using the most likely parameter specification indicate that a replacement ratio of 58\% is best. A replacement ratio slightly higher than the optimal ratio can lead to a large decrease in average utility, and is problematic. When human capital changes are relatively less responsive to unemployment and employment duration, longer unemployment spells are more desirable as they lead to better matches. When the effect of taxes and benefits are looked at separately, the benefit ratio aspects matters more than the tax rate.