This thesis includes three empirical studies that analyse driving forces and outcomes of active labour market policies (ALMP) in comparative perspective. Whereas previous comparative research has largely relied on broad conceptualisations of ALMP, often considering policies such as public employment services and training programs as one single type of intervention, the studies in this thesis instead analyse more fine-grained categories and classifications of ALMP. By doing so, the studies contribute with nuance to the discussion of what has shaped labour market policies, and how these influence labour market outcomes.

Study I: This study focuses on determinants of spending on public employment services (PES). Contrary to common assumptions, which state that PES is largely determined by structural factors, the main hypothesis is that partisan politics is a relevant predictor of PES spending. The analysis includes 17 welfare states between 1985 and 2011. The results indicate that partisan politics and electoral competition are related to the development of PES spending, but that the turn to activation in many welfare states in the late 1990s reduced the salience of partisan politics.

Study II: This study analyses whether different types of active labour market policies are related to reduced unemployment, and especially if there is evidence for interdependencies between policies. The study distinguishes between public employment services (PES), training, and job creation programs. PES is singled out as a crucial factor that mediates the effects of other labour market policies, both active and passive. The study examines 19 welfare states between 1985 and 2012 and the results indicate that training programs reduce unemployment in the short run, whereas long-term effects are less apparent for all policy types. Increased investment in PES is found to strengthen the impact of training and job creation programs.

Study III: This study examines the link between active labour market policies and transitions to employment from unemployment and inactivity across the European Union between 2003 and 2013. The study distinguishes between three types of interventions: core programs, directed towards the registered unemployed; peripheral programs, targeting the inactive part of the working age population; and mixed programs, targeting both groups. The results indicate that ALMP have beneficial effects on employment transitions among both the unemployed and the inactive, but that effects differ depending on the type of intervention. Unemployed seem to benefit from increased efforts on ALMP regardless of the type of intervention, whereas the inactive seem to have increased chances to transition into employment if programs target both groups, but not if interventions solely target the inactive.