| Abstract: | This paper analyzes how biased beliefs about employment prospects affect the optimal design of unemployment insurance. Empirically, I find that the unemployed greatly overestimate how quickly they will find work. As a consequence, they would search too little for work, save too little for unemployment and deplete their savings too rapidly when unemployed. I analyze the use of the “sufficient‐statistics” formula to characterize the optimal unemployment policy when beliefs are biased and revisit the desirability of providing liquidity to the unemployed. I also find that the optimal unemployment policy may involve increasing benefits during the unemployment spell. |
