How Central Exams Affect Educational Achievement: International Evidence from TIMSS and TIMSS-Repeat

Abstract

Central exams reduce the informational asymmetry between principals and agents ubiquitous in education systems, thereby improving educational performance. Using international micro data for nearly half a million students, I find that students in countries with central exit-exam systems perform 35 to 47 percent of an international standard deviation in test scores better in their middle-school years in both math and science than students in countries without central exams. The result is robust to restricting the analysis to within-continental variation and to controlling for the general centralization of the education system, other institutional features, and population homogeneity. Central exams equalize opportunities for students from different parental backgrounds. Budgetary and salary autonomy for schools is detrimental in systems without central exams but beneficial in systems with central exams.