In this paper we estimate a model of the determinants of economic and social development that takes seriously three of the criticisms of panel data models in the existing growth literature: that long run coefficients are biased because the lagged dependent variable is not strictly exogenous; that they are biased because of slope coefficient heterogeneity; and that they are biased because of explanatory variable endogeneity. The model indicates that there are strong causal relationships in both directions between on the one hand, economic development, and on the other, social and political development.