This paper analyses determinants of public sector efficiency in 15 Central Eastern and South Eastern European (CESEE) countries. Using Data Envelopment Analysis (DEA) and panel Tobit regressions we analyse the efficiency of public expenditure on health care, education, and public administration. This paper represents the first analysis of public expenditure efficiency for this region. The countries in the sample share a similar economic transition background and transition related challenges in terms of government size and efficiency. In this paper we use input-oriented DEA to calculate efficiency scores and apply Tobit regressions to identify the determinants of public expenditure efficiency for each category. According to our results, the health care sector shows the most room for improvement. With an average input efficiency score of 0.87, the analysed countries could attain the same output with an average 13% fewer resources. The results show that Bulgaria, Estonia, and Slovenia are the most efficient, whilst Serbia, Croatia, and Bosnia and Herzegovina are the worst performers. GDP per capita has a positive effect on efficiency in the health care sector, the rising share of the elderly population reduces the efficiency coefficient, whilst a higher level of education has a positive effect on the efficiency score, as recorded in earlier studies. In education, our results suggest that countries waste an average 10% of their resources. The Czech Republic, Estonia, Poland, Romania, and the Slovak Republic represent the frontier, while the worst performers are Slovenia, Lithuania, and Latvia. The results show that GDP per capita has a positive effect on efficiency in the education sector, whilst a higher unemployment rate leads to lower efficiency scores. Regarding public administration, the average efficiency score of the analysed countries is 0.9, which indicates that countries are wasting an average 10% of their resources. Estonia and Romania are fully efficient, whilst Montenegro, Bosnia and Herzegovina, and Hungary are the most inefficient countries in the sample. As expected, GDP per capita and institutional quality have positive effects on the efficiency of public administration.