The paper estimates the correlates of student performance across countries using student-level data from the PISA-2000 study. There is a substantial relationship of international student performance in math, science, and reading with institutional features. In particular, there are positive associations with accountability through external exit exams, school autonomy in personnel-management and process decisions, and private operation of schools. By contrast, there is no positive association with private school funding, and a negative association with autonomy in areas with scope for decentralized opportunistic behavior. Student performanceshows a positive relationship with school autonomy particularly where external exit exams are in place, highlighting the role of external exams as currency of school systems. Positive associations of student performance with resource endowments of schools are mainly related to resource quality. While there is a strong association with family background, computer availability at home shows even a negative relationship with student performance on basic skills, after holding other effects constant. The explanatory power of the model between countries is remarkable.