Surprisingly little is known about the impact of natural resource booms on income inequality in resource rich countries (Ross, 2007). This paper develops a theory, in the context of a two-sector growth model in which learning-by-doing drives growth, to explain the time path of inequality following a resource boom. Under the condition that the nontraded sector uses unskilled labor more intensively than the traded sector, we find that income inequality will fall in the short run immediately after a boom and will then increase steadily over time as the economy grows, until the initial impact of the boom on inequality disappears. Using dynamic panel data estimation for 90 countries between 1965 and 1999 and exploiting variation in world commodity prices to identify resource booms, we find evidence in support of the theory, especially for oil and mineral booms. We also find that uncertainty about future commodity export prices significantly increases long-run inequality.