Clemente Pignatti (University of Milan)
I study the system of active labor market policies (ALMPs) available to unemployment benefit (UB) recipients in Mauritius, where individuals can choose between passive UB receipt, training or start-up support. I document negative self-selection into training relative to both passive UB receipt and start-up support, based on both observable and unobservable characteristics. Leveraging the random assignment of UB recipients to caseworkers at the time of program registration, I find that participation in training and start-up support has negative short-term effects on formal employment. However, these effects dissipate within one year after job loss. For training, treatment effects become positive three to four years after job loss, particularly for women and younger workers. In addition, both training and start-up support increase the likelihood of re-employment in higher-paying firms. By contrast, I find no effects on outcomes in the informal labor market, where most jobs are of subsistence nature. Finally, I conduct a counterfactual exercise to compare the efficiency of the current assignment rule with alternative strategies. I find that letting people choose their preferred option after a short meeting with a caseworker performs relatively well compared to other commonly used assignment methods.