Beyond treatment effect estimation; new developments in policy evaluation

To what extent are hiring incentives targeting a specific group of vulnerable unemployed people more effective with respect to generalized incentives without a definite target? Do targeted policies have negative side effects on the labor market that are too important to accept them? Which are the channels guaranteeing a long-lasting effect of European Structural Funds? Even though there is vast literature on hiring subsidies and European Structural Funds, these questions remained unanswered. To answer them, it is necessary to go beyond the simple treatment effect estimation. We did it following two different paths. To answer the first two questions we compared the impact of two similar hiring policies, one oriented towards a target group of long-term unemployed people and one generalized, implemented on the Italian labor market. We considered administrative data on job contracts and workers, and applied counterfactual analysis methods. The results show that only the targeted policy has a positive and significant impact, while the effects of the generalized policy on the vulnerable group are negligible. Moreover, we did not detect any indirect negative side effect of the targeted policy on the local labor market. To answer the third question we introduced a new methodology, namely the Mediation Analysis Synthetic Control (MASC) and applied it to investigate on the impact of European Structural Funds reduction in Abruzzi region. MASC is a generalization of the Synthetic Control Method (SCM) that allows decomposing the total effect of the intervention into its indirect effects and its direct effect when data on only one treated and few control units are available. The results show that the negative impact of European Structural Funds reduction on economic growth is mainly driven by the indirect effect passing through employment reduction. Instead, only a small portion of the negative impact is due to the indirect effect passing through investments reduction.