Author ' s personal copy Airplanes and comparative advantage ☆

a r t i c l e i n f o Airplanes are a fast but expensive means of shipping goods, a fact which has implications for comparative advantage. The paper develops a Ricardian model with a continuum of goods which vary by weight and hence transport cost. Comparative advantage depends on relative air and surface transport costs across countries and goods, as well as stochastic productivity. A key testable implication is that the U.S. should import heavier goods from nearby countries, and lighter goods from faraway counties. This implication is tested using detailed data on U.S. imports from 1990 to 2003. Looking across goods the U.S. imports, nearby exporters have lower market share in goods that the rest of the world ships by air. Looking across exporters for individual goods, distance from the US is associated with much higher import unit values. These effects are large, which establish that the model identifies an important influence on specialization and trade. Countries vary in their distances from each other, and traded goods have differing physical characteristics. As a consequence, the cost of shipping goods varies dramatically by type of good and how far they are shipped. A moment's reflection suggests that these facts are probably important for understanding international trade, yet they have been widely ignored by trade economists. In this paper I focus on one aspect of this set of facts, which is that airplanes are a fast but expensive means of shipping goods. The fact that airplanes are fast and expensive means that they will be used for shipping only when timely delivery is valuable enough to outweigh the premium that must be paid for air shipment. They will also be used disproportionately for goods that are produced far from where they are sold, since the speed advantage of airplanes over surface transport is increasing in distance. In this paper I show how these considerations can be incorporated into the influential Eaton and Kortum (2002) model of comparative advantage. In this general model, differences across goods in transport costs (both air and surface) and the value that consumers place on timely delivery interact with relative distance to affect global trade patterns. The model of the paper delivers two empirical implications that I test using highly disaggregated data on all U.S. imports from 1990 to 2003. The first implication is that nearby trading partners (Canada and Mexico) should have lower …