A common finding of the optimal unemployment insurance literature is that the optimal UI replacement rate is around 50%, implying that current levels in the US are close to optimal. However, a key assumption in the existing literature is that unemployment benefits are the only government spending activity. In this paper I show that recommendations for optimal UI levels are dramatically reduced when one incorporates the fact that UI spending is a small part of overall government spending. This occurs because the negative impact of UI on income tax revenues implies added welfare costs, a mechanism that I refer to as a fiscal externality. Using both a calibrated structural job search model and a "suffcient statistics" method that relies on reduced-form elasticities, I find that the optimal replacement rate drops to zero once fiscal externalities are incorporated. However, I also consider the possibility that more generous UI could increase reservation wages and thus potentially increase the tax base, and I show that this second fiscal externality could have important effects on the results, with an optimal replacement rate which could rise above 70%.
