Evaluating Asymmetries in Active Labour Market Policies: the Case of Italy

This paper aims at analyzing whether Labour Market Programs (ALMP) could have different effects on unemployment and employment dynamics according to the particular region where the program is implemented. To this end, the research analyses alternative theoretical and econometric models thought to capture the possible effects that active labour market policies might have on labour forces dynamics. The econometric methodologies implemented are the Generalized Method of Moment (GMM) and the Panel Vector Autoregression (P-VAR). The evidence emerging from the GMM models suggests that the effects of ALMP on unemployment are dissimilar across the Italian regions. It follows that some active programs are likely to exert a greater effect in the South than in the North. The results of the P-VAR estimated models are synthesized in the impulse response analysis and the forecast error variance decomposition. The impulse response analysis suggests that an increase in ALMP lead to: (i) a decrease in the unemployment rate, and (ii) significant increase in labour force participation. More interestingly, results obtained from the error-variance decomposition analysis show that unemployment movements are not driven by shocks in the ALMP and that, especially in the northern regions, atypical contracts shocks account for a substantial portion of unemployment dynamics.