School choice and accountability have become popular educational policies in the US and the UK. In Europe, such policies are less often applied. Moreover, even when school choice exists, information on school quality is rarely provided, leading to pupil sorting by observable characteristics. In this paper, we estimate how the regional intensity of school choice and the regional percentage of selective schools relate to student test scores in math, reading and science at age 15. We estimate an education production function exploiting regional aggregation in 8 European countries to reduce potential endogeneity bias. We find that both the regional intensity of school choice and that of school selectivity are correlated with significantly higher student test scores.