Explaining welfare recidivism: what role do unemployment and initial spells have?

The question of high welfare re-entry rates has attracted great attention from economists and policymakers. Using a very rich administrative data set for the minimum income program of the Madrid Government (over 50,000 spells), this paper aims to broach various questions arising from the issue of welfare re-entry. We try to identify what factors determine observed differences in the durations of the first off-welfare spell. We analyze the combined effects of the length and type of exit of the first spell, unemployment, and sociodemographic characteristics. The experience of the first spell and, to a lesser extent, employability can contribute toward lengthening the time spent outside the program. Our results also show that off-welfare spells of households leaving the program in periods of low economic growth will be longer than those that do so during economic expansions. With the exception of unemployment effects, our estimates yield very similar results to those obtained in the U.S. studies, suggesting that U.S. welfare policies and analyses are not entirely irrelevant for European discussions of welfare policies.