Description:
ABSTRACTHanushek and Kimko's analysis of the relationship between growth and schooling quality, as measured by scores in international tests, suffers from potential endogeneity as schooling quality is not always measured at a date strictly prior to the observed growth. To address this problem we treat the data as a panel, relating growth only to test scores at earlier dates. The estimates of the effect of schooling quality on growth are similar to those obtained from cross‐section regressions.