LABOR MARKET POLICY EXPENDITURES EFFECTS: EVIDENCE FROM 15 EUROPEAN UNION COUNTRIES

This paper examines the effects of active (ALMP) and passive labor market policy (PLMP) expenditures on labor market outcomes. The analysis covers the sample of 15 EU countries during the period from late 1990s/early 2000s to 2017. We use a local projection method to estimate linear and nonlinear country-specific effects of LMP expenditures on labor market outcomes during the different phases of the business cycle. The results indicate that in more than half of the analyzed countries ALMP expenditures improve labor market outcomes. This conclusion holds for at least two years after the initial expenditure shock. Furthermore, the (median) effects seem to be independent of the business cycle phase. Concerning PLMP effects on labor market outcomes, the results indicate that in more than half of the analyzed countries PLMPs worsen labor market outcomes by decreasing employment and increasing youth and long-term unemployment. Regarding the overall unemployment rate, the results are inconclusive. In comparison to ALMPs, it seems that the business cycle phase matters in the case of PLMPs. Namely, adverse (median) PLMP effects on employment and long-term unemployment are found to be lower (employment) or insignificant (long-term unemployment) during times of subdued economic activity. Conversely, in the case of the overall and youth unemployment, the PLMP effects are insignificant during expansionary phases but detrimental during slack periods. Still, the finding that PLMP adverse (median) effects on long-term unemployment seem to be insignificant during slacks slightly mitigates usual concerns regarding generous benefit systems and the equilibrium unemployment rate, at least during recessionary times. However, median responses conceal significant differences of LMPs effects on labor market outcomes during distinguishable stages of the business cycle between different countries, i.e. the effects seem to be country, and policy type specific.