Background Labour market policies in Sweden is a state responsibility. Even though the municipalitieshave created their own units for labour market issues. This has resulted in parallel systems,and the distribution of responsibility is in many ways unclear. Several studies within the topichave observed that the effects of the activation programs disposed by the municipalities aredifficult to study, dependent there is not enough comparable data. Therefore statistics is nowcollected by Sveriges kommuner och Landsting. Purpose The purpose of this study is to examine if the statistics that are available today can be used tomeasure the effects of labour market actions. Method The study is based on the theory about Governance as decentralization. The hypothesis is thatthe amount of money that the municipalities spend on labour market cations should affect theeffect. With effect means, in this study, that the participants reach work or studies. The studyis made on two Swedish municipalities, Robertsfors and Ljusnarsberg. Those municipalitiesare selected therefor they are the municipalities in Sweden that 2015 spent most and lessmoney on labour market actions. In an attempt to try if the statistics that are available todaycan be used, the study is using a quantitative method. Interviews with representatives from themunicipalities have also been made in order to explain choices in submitting of the statistics. Results The results show that the statistics that are available not can be held for true. Both from thecalculation example as from the interviews that is clear there are uncertainties with thestatistics. From the theory of Governance, this study comes to the conclusion thatdecentralization and standardization are extremes.