Is the 'Curse of Natural Resources' Really a Curse?

This paper takes a new look at the long-run implications of resource abundance. Using a Schumpeterian growth model that yields an analytical solution for the transition path, it derives conditions under which the curse of natural resources occurs and is in fact a curse, meaning that welfare falls, conditions under which it occurs but it is not a curse, meaning that growth slows down but welfare rises nevertheless, and conditions under which it does not occur at all. An effective way to summarize the results is to picture growth and welfare as hump-shaped functions of resource abundance. The property that the peak of growth occurs earlier than the peak of welfare captures the crucial role of initial consumption, which rises with resource abundance, and is an important reminder that the welfare effect of resource abundance depends on the whole path of consumption, not on a summary statistic of its slope. Growth regressions that ignore the endogeneity of initial income do not provide sufficient information to assess whether resource abundance is bad even if one could prove beyond reasonable doubt that the relation is indeed negative and causal. Recent evidence that the correlation is actually positive should make us even more skeptical of policy advice based on the curse logic.