A Causal Machine Learning Analysis of Swiss Active Labor Market Policies

Abstract:

Active labor market policies are widely used in Switzerland, enrolling more than half of all unemployed individuals. This paper evaluates the effectiveness of these programs in improving employment and earnings outcomes using rich administrative data on individuals unemployed in 2014 and 2015, which contains their detailed labor market histories and other covariates. The analysis applies causal machine learning to estimate heterogeneous program effects, enabling granular program evaluation and potential for improved program targeting. The results indicate that programs are most effective for individuals with a non-EU migration background, while Temporary Wage Subsidies also benefit those with lower educational attainment. Finally, shallow policy trees provide practical guidance for improving the targeting of program assignments.