We use 130 years of data on historical migrations to the United States to show a causal effect of the ancestry composition of US counties on foreign direct investment (FDI) sent and received by local firms. To isolate the causal effect of ancestry on FDI, we build a simple reduced-form model of migrations: Migrations from a foreign country to a US county at a given time depend on (i) a push factor, causing emigration from that foreign country to the entire United States, and (ii) a pull factor, causing immigration from all origins into that US county. The interaction between time-series variation in origin-specific push factors and destination-specific pull factors generates quasi-random variation in the allocation of migrants across US counties. We find that doubling the number of residents with ancestry from a given foreign country relative to the mean increases the probability that at least one local firm engages in FDI with that country by 4 percentage points. We present evidence that this effect is primarily driven by a reduction in information frictions, and not by better contract enforcement, taste similarities, or a convergence in factor endowments.