Abstract

While studies have shown that unemployment benefit sanctions can increase reemployment rates (see e.g. van den Berg et al. 2022; van den Berg et al. 2013; Boockmann et al. 2014), they have also revealed the occurrence of negative sideeffects e.g. on the quality of future employment or withdrawal from the labour force (see e.g. van den Berg et al. 2022; van den Berg & Vikström 2014; Wolf 2021). Based on those heterogeneous effects we analyze the effect of unemployment benefit sanctions (UB II) on future employment outcomes. In contrast to existing studies, we use the date of the sanction event as our first treatment, since behavior of individuals might change from this date onwards. In addition, we consider the characteristics of the sanction (e.g., type and amount) as well as employment histories and other sociodemographic characteristics of the sanctioned person. Since past employment histories might lead to a biased estimation of the sanction effect, we specify a dynamic model including lagged information on employment state, sanctions, timevarying characteristics as well as timeconstant unobserved heterogeneity. We use a Bayesian Markov Chain Monte Carlo (MCMC) method (see Fitzenberger et al. 2023), which is especially well suited to estimate high dimensional models.