In many countries, the government is using state subsidies to promote local authorities to improve, e.g., social welfare services. But evaluations of these initiatives often show small effects compared to the politicians´ goals. In the paper we present results from a Swedish study including eight larger national programs aimed at promoting local authorities’ social services. The aim is to analyze and discuss if success or failure to achieve the national goals of the programs can be related to unrealistic political ambitions, to the implementation process, to limitations or deficiencies in the evaluations, or to all three. The first two issues are analysed within a political science model, and in the latter case data availability and evaluation designs are focused. In the light of the results, program theoretical, ethical and qualitative issues in evaluation are discussed.