The Empirics of Growth: An Update

We argue that both growth accounting and growth regression analysis can yield useful insights into the growth process, particularly when used to study a large sample of countries over a long period. Much of the apparent variability in conclusions of previous studies arises from measurement problems, differences in data or definitions, and failure to include other conditioning variables. Using data for eighty-four countries over 1960-2000, and focusing attention on issues of measurement and consistency, we conclude that both capital accumulation and efficiency gains are central to the growth process. We find, at best, a weak correlation between growth and increased schooling. And we find that most of the variation in cross-national economic growth is well explained by differences in initial conditions and institutions. However, these variables provide little insight into the changes in growth rates before and after 1980, and policy differences play a minor role.