Blaize Giangiulio

This paper models the disruption of a maritime chokepoint as a strategic choice rather than an exogenous shock. In an (N+1)-country Armington economy, a disruptor that raises trade costs between other countries improves its own terms of trade even absent any geopolitical motive, and the gain scales quadratically in the size of the disrupted corridor, rationalizing why disruptors target arteries rather than individual routes. I apply the model to the 2023–24 Red Sea crisis using a continuous, geography-determined measure of each country pair's exposure to closure of the Bab el-Mandeb, built by rerouting 39 million city-pair freight paths over a global road and maritime network. PPML gravity estimates show that trade fell on exposed dyads, with the response concentrated among Western-aligned pairs; exact hat algebra delivers the general-equilibrium welfare incidence, with losses increasing in exposure and Iran's own gain negligible. Inverting the disruptor's first-order condition shows that each dollar Iran gains at the margin costs Western economies roughly $200 to $1,200, and that self-interest alone rationalizes the campaign only if it cost Iran less than $42 million per year, well below independent estimates of Iranian support to the Houthis: a positive weight on harming adversaries is required to explain the observed disruption.

We study the spillover effects of the 2025 “Liberation Day” automotive tariffs on the untaxed U.S. used-car market. Using over eight million weekly listings and a VIN-based crosswalk identifying each vehicle's country of assembly, we document an immediate and persistent 0.2–0.6% increase in the prices of foreign-brand used vehicles relative to domestic-brand vehicles. The effect is concentrated among the newest vehicles—the closest substitutes for new cars—consistent with demand-side substitution. We also find that foreign-brand vehicles produced in the U.S. appreciate as much as those assembled abroad, suggesting that the perceived foreignness of used vehicles may be largely independent of true production location. Neglecting tariff pass-through to untaxed product substitutes understates overall consumer incidence.

This paper employs geospatial network data to construct a novel transportation-network-informed distance variable that integrates realistic city-to-city freight routing into the Mayer and Zignago (2011) population-weighted distance formula, rendering bilateral distances that are meaningfully responsive to infrastructure shocks. Using the 2023 Red Sea crisis as a natural experiment, we estimate a saturated gravity model in an event-study framework with importer–exporter–time and country-pair fixed effects on monthly bilateral trade data. Treated country pairs are those whose network-informed distance rises meaningfully following the closure of the Bab el-Mandeb Strait, identified against a control group of unaffected pairs.

We introduce a time-varying measure of distance between and within countries and demonstrate its usefulness in international economics applications. Built from annual population data for 6,162 cities in 235 countries from 1950 to 2024, the measure improves gravity estimation by enabling theoretically consistent distance estimates in the presence of bilateral fixed effects, mitigating endogeneity concerns related to unobservable factors. We confirm the previous finding that the effects of distance have fallen over time and show that this decline has been driven primarily by affluent countries. Within a general equilibrium setting, we demonstrate that a counterfactual reversion to mid-century distances produces large welfare changes ranging from −1.4% to +2.4%, suggesting that geographic population shifts since the mid-twentieth century have benefited some countries and hurt others. U.S. migration to the South and West has had particularly large impacts on trade and welfare in Canada and Mexico.

Do state used-car lemon laws, which often require dealers to attach a short express warranty to vehicles sold below sharp odometer cutoffs (e.g., 125,000 miles in Massachusetts), generate price premiums? Using 9.9 million web-scraped listings and a regression-discontinuity design with rich fixed effects, I find no warranty premium, alongside pronounced bunching of odometer readings just above the Massachusetts cutoff and reallocation of near-cutoff inventory across state lines. I rationalize both facts with a shrouded-attribute model in the spirit of Gabaix and Laibson (2006): when buyers unaware of the mandate are the marginal buyers of covered vehicles, the equilibrium premium is zero regardless of the dealer's cost; dealers have no private incentive to debias buyers because disclosure raises claims costs, cannibalizes service-contract sales, and is a public good among rivals; and the resulting unpriced cost acts as a tax on covered sales that is shifted onto the quantity margin. Calibrated simulations reproduce the observed magnitudes of both the null price effect and the Massachusetts bunching.