This paper revisits the cross-country growth empirics debate using a novel Limited Information Bayesian Model Averaging framework to address model uncertainty in the context of a dynamic growth model in panel data with endogenous regressors. Our empirical findings suggest that once model uncertainty is accounted for there is strong evidence that initial income, investment, life expectancy, and population growth are robustly correlated with economic growth. We also find evidence that debt, openness, and inflation are robust growth determinants. Overall, the set of our robust growth determinants differs from those identified by other studies that incorporate model uncertainty, but ignore dynamics and/or endogeneity. This underscores the importance of accounting for model uncertainty and endogeneity in the investigation of growth determinants.
More here
