The effect of unemployment insurance benefits on (self-)employment : two sides of the same coin?

Although a relevant share of firms is created out of unemployment and current active labor market policies in Europe often subsidize unemployed individuals to start their own businesses, little is known about the role of unemployment insurance (UI) generosity for self-employment. By using Spanish administrative data including so far inaccessible information on self-employment, we exploit a reform-driven exogenous cut in UI benefits to identify its causal effect on general employment and decompose it into the effects on self-employment and re-employment. Exploiting a discontinuity in the UI benefit schedule which changed as a result of the 2012 Spanish labor market reform, we estimate the causal reform effects on the extensive margin of (self-)employment and on unemployment duration. We find heterogeneous effects on the extensive margin: while the job-finding rate increases, the startup rate decreases. Over different time horizons, the negative effect on self- employment (35-50%) outweighs the positive effect on employment (5-33%). Therefore, omitting self-employment as a counterfactual outcome might lead to overestimate general employment effects. Our UI benefit duration elasticity estimates indicate that reduced UI benefits extend unemployment duration for individuals transitioning into self-employment but shorten unemployment for individuals finding re-employment. These results might be relevant for the (optimal) design of UI systems.