Do institutions rule when explaining cross-country divergence? By employing regression tree analysis to uncover the existence and nature of multiple development clubs and growth regimes, this paper finds that to a large extent they do. However, the role of ethnic fractionalization cannot be dismissed. The findings suggest that sufficiently high-quality institutions may be necessary for the negative impact on development from high levels of ethnic fractionalization to be mitigated. Interestingly, I find no role for geographic factors-neither those associated with climate nor physical isolation-in explaining divergence. There is also no evidence to suggest a role for religious fractionalization.
