Does trade cause capital to flow? Evidence from historical rainfall

We use a historical quasi-experiment to estimate the causal effect of trade on capital flows. We argue that fluctuations in regional rainfall within the Ottoman Empire capture the exogenous variation in exports from the Empire to Germany, France, and the U.K., during the period of 1859–1913. The identification is based on the following historical facts: First, only surplus production was allowed to be exported from the Empire (provisionistic policy). Second, different products grown in different regions were subject to variation in regional rainfall. Third, different bundles of products were exported to Germany, France, and the U.K. by the Empire. Using the export-bundle-weighted regional rainfall as an instrument for Ottoman exports to each country, our instrumental variable regression suggests the following: When a given region of the Empire received more rainfall than others, the resulting surplus production was exported more to countries that historically imported more of those products, and this leads to higher foreign investment by those countries in the Empire. Our findings support theories predicting complementarity between trade and finance, in which causality runs from trade to capital flows.