This paper diagnoses the symptoms of the Dutch disease in a two-sector stochastic endogenous growth model. A productive, low skill-intensive primary sector causes the currency to appreciate in real terms, thus hampering the development of a high skill-intensive secondary sector and thereby reducing growth. Moreover, the volatility of the primary sector generates real exchange rate uncertainty, and may thus reduce investment and learning in the secondary sector and hence also growth. Cross-section and panel regressions based on data for 125 countries in the period 1960–92 confirm a statistically significant inverse relationship between the size of the primary sector and economic growth, but not between the volatility of the real exchange rate and growth.