An evaluation of the Swedish trainee replacement schemes

This paper estimates the employment effects of the Swedish trainee replacement schemes (an active labour market program that was in operation during the 1990s). The empirical analysis exploits a large and rich administrative data set, and we control for observed and unobserved selection bias by using a multiple equation model and the maximum likelihood estimation method. The estimation results point at a selection of participants having a high ex ante probability of employment. In addition, the results suggest that participation in replacement schemes increased the re-employment probability by 5 to 10 percentage points.