What We Have Learned about Policy and Growth from Cross-Country Regressions?

Economists have been seeking to comprehend why some countries are rich and others poor for well over 200 years. A better understanding of the national policies associated with long-run growth would both contribute to our ability to explain cross country differences in per capita incomes and provide a basis for making policy recommendations that could lead to improvements in human welfare. Recently, economists have used cross-country regressions to search for empirical linkages between longrun growth and indicators of national policies (e.g., Roger Kormendi and Philip Meguire, 1985; Robert J. Barro, 1991). The large cross-country growth literature has identified various fiscal, monetary, trade, exchange-rate, and financial-policy indicators that are significantly correlated with longrun growth. Yet, Levine and David Renelt (1992) show that many of these findings are fragile to small alterations in the conditioning information set. That is, small changes in the right-hand-side variables produce different conclusions regarding the relationship between individual policies and growth. In this paper, we take stock of what the profession has learned from cross-country regression studies of policy and long-run growth. I. Limitations