Although there is an extensive literature on the short-term effects of unemployment benefits, little is known about their medium-term implications. In this paper, I use rich and novel administrative data to study the effects of potential benefit duration on aggregate outcomes over 4 years after layoff. To obtain causal estimates, the identification exploits an age at layoff rule which determines a 4 months increase in potential benefit duration if the worker is fired after turning 50. Longer potential benefit duration leads to longer periods receiving unemployment benefits and time spent in not employment before finding a new job with a small increase in the probability of finding a permanent contract. Over the 4 years following layoff, the overall difference in time spent in not employment is substantially reduced. A higher probability of employment in the period following the first reemployment date contributes to explain this discrepancy. These findings are important from a policy perspective as they suggest that classical measures of the cost of unemployment benefits tend to overestimate the negative externalities of potential benefit duration.