Quasi-Experimental Evidence on the Effects of Unemployment Insurance from New York State

This paper examines unemployment duration and the incidence of claims following a 36 percent increase in the maximum weekly benefit in New York State. This benefit increase sharply increased benefits for a large group of claimants, while leaving them unchanged for a large share of claimants who provide a natural comparison group. The New York benefit increase has the special features that it was unexpected and applied to in-progress spells. These features allow the effects on duration to be convincingly separated from effects on incidence. The results show a sharp fall in the hazard of leaving UI that coincides with the increase in benefits. The evidence is also consistent with a substantial effect of the benefit level on the incidence of claims and with this change in incidence biasing duration estimates. The evidence further suggests that, at least in this case, standard methods that identify duration effects through nonlinearities in the benefit schedule are not badly biased.