Social Assistance and Crime

Jeder, der an dem Kolloquium teilnehmen möchte und noch nicht über den erforderlichen Link verfügt, kann diesen per E-Mail an colloquia-a [at] mzes.uni-mannheim.de anfordern.

Abstract:

Daniel Auer, Michaela Slotwinski, Dominik Hangartner, Selina Kurer, Alois Stutzer, Stefanie Kurt

A popular concern in the discussion about optimal social policy is that generous welfare state provisions discourage recipients from labor market participation. This argument is particularly salient in the discussion about social benefits for recent immigrants, including asylum seekers. Standard reservation wage models, however, suggest that welfare cuts may have harmful consequences for recipients and society if employment effects are nonexistent. In this case, cuts may drive people into illegal endeavors as alternative sources of income. To test whether a cut in social assistance is associated with more crime, we draw on Swiss administrative data linking records of all foreign individuals to official criminal records. We provide empirical evidence based on three different quasi-experimental settings that arise from the autonomy of Swiss cantons in setting social assistance rates. First, we use a classical natural experiment in which asylum seekers are distributed quasi-randomly across states (cantons) and thus across locations with the same general welfare regime but with different levels of welfare provision. Second, we exploit the variation in welfare jumps associated with status changes across cantons. While in general the changes associated with a change in status from asylum seeker to recognized refugee are homogeneous and determined at the national level, the associated change in welfare provision varies across cantons. Third, we estimate a difference-in-differences model following a significant policy change for provisionally admitted persons in the canton of Zurich. All specifications provide evidence for a strongly negative causal relationship between social assistance and crime.