This paper investigates the effect of introducing a massive workfare program on property crimes. In order to circumvent the endogeneity problem common to joint factors determining crime and demand for workfare we make use of instrumental variables. We exploit two special features. First, the program was assigned according to political criteria which were trying to attract provinces and/or counties who were aligned with the national government. Second, the program was grant in mid-2002 and closed afterwards, so there were no new-comers to the program. We use political affiliation of different level of governments as instrument for the number of workfare programs per capita and find that the program reduced property crime but had no effect on other kinds of crime. The paper represents a contribution to the crime literature, since this issue has not been explored. If workfare programs have an effect on crime, then the welfare effect is different from the one often calculated in the literature.