This article has two main goals. The first is to assess the effectiveness of social policy in Serbia by identifying its intensity, coverage, adequacy and targeting performance. The second goal is to compare the time and income group impact of different social policy programmes on income inequality, in order to derive policy recommendations. To this end, using statistical data from the Household Budget Survey 2006–09 which cover more than 4500 households per year, we have conducted the first research of this kind in Serbia. Various statistical and econometric means were used: Gini coefficients for income inequality, the Coady–Grosh–Hoddinott (CGH) index for targeting performance, the Newey–West (HAC) procedure and recursive regression, Error Component Two Stages Least Squares Method (EC-2SLS) for panel estimation. The confirmed hypotheses are that, in general, income inequality is decreasing as a consequence of certain social policy measures, except in the last observed year, due to the economic crisis, but that these effects are different in different income groups. In the conclusion certain policy recommendations are drawn concerning specific vulnerable groups and the need for redistribution of resources towards better targeted programmes.