A Geography of Investment Under Uncertainty: Trade Policy Shocks and the Subnational Reallocation of U.S. Foreign Direct Investment in Mexico

Abstract

Leveraging a Bartik-style shift-share treatment variable, this thesis asks whether unanticipated innovations in U.S. trade policy uncertainty (TPU) trigger a subnational reallocation of U.S.-origin foreign direct investment across Mexican states and industries. Causal identification follows the shock-exogeneity framework of Borusyak et al. [2022]. Using a balanced panel spanning 32 states, 20 industries, and 23 years from Mexico’s Secretaría de Economía, I estimate that a one-standard-deviation increase in the Bartik treatment raises intensive-margin FDI (asinh-transformed) by 0.360 units (wild cluster bootstrap p = 0.017), corresponding to an approximate 43% increase in investment volume for observations in the log-approximating region of the asinh transformation. The effect is contemporaneous, dissipates completely within a year, and does not accumulate over the medium run. The extensive margin yields a directionally consistent but statistically fragile estimate. Grounded in the real-options framework of Investment Under Uncertainty (Dixit and Pindyck [1994]), I interpret these findings as evidence of a flight-to-familiarity mechanism, whereby spikes in TPU lead U.S. multinationals to concentrate their remaining investment in state-industry pairs with the strongest pre-existing operational ties. Heterogeneity analysis shows that the effect is strongest in manufacturing and in Mexico’s border states, and a comprehensive battery of robustness checks supports the causal interpretation.