The paper explores the optimal monetary policy reaction to productivity shocks in an open economy. Whereas earlier studies assume that countries specialize in producing particular goods, I enrich the analysis by allowing for incomplete specialization. I confirm the finding of Obstfeld and Rogoff (2000)—who build on Friedman (1953)—that a flexible exchange rate is highly valuable in delivering the optimal response to country- specific shocks. Its value is, however, much smaller when shocks are sector-specific, because exchange rate fluctuations then lead to misallocations between different firms within a sector. The limitation on the value of flexibility is sizable even when specialization is high.