The opportunity cost of holding international reserves plays a central role in all models of optimal demand for foreign exchange. This cost is conventionally defined as the difference between the yield on reserves and the marginal productivity forgone from an alternative investment in fixed capital. Most empirical studies have failed to find a significant opportunity-cost effect, since none of them measure it in accordance with its theoretical definition. The results for Israel show that, when this cost is measured properly, it turns out to be a crucial determinant of reserve demand. Copyright 1992 by MIT Press.