Do Employment Subsidies Work? Evidence from Regionally Targeted Subsidies in Turkey

This paper studies the e ects on registered employment, earnings, and number ofregistered establishments of two employment subsidy schemes in Turkey. We implementa di erence-in-di erences methodology to construct appropriate counterfactuals forthe covered provinces. Our ndings suggest that both subsidy programs did lead tosigni cant net increases in registered jobs in eligible provinces (5%{13% for the rstprogram and 11%{15% for the second). However, the cost of the actual job creationwas high because of substantial deadweight losses, particularly for the rst program(47% and 78%). Because of better design features, the second subsidy program hadlower, though still signi cant, deadweight losses (23%{44%). Although constrainedby data availability, the evidence suggests that the dominant e ect of subsidies wasto increase social security registration of rms and workers rather than boosting totalemployment and economic activity. This supports the hypothesis that in countrieswith weak enforcement institutions, high labor taxes on low-wage workers may lead tosubstantial incentives for rms and workers to operate informally.