Transportation costs and trade clusters: Some empirical evidence from U.S. customs districts

a r t i c l e i n f o This paper assesses the impact on transportation costs of the mix of goods imported from Canada into the United States. The analysis focuses on merchandise imports processed through eight U.S. customs districts over 1990–2013. The extent to which transportation costs are influenced by the nature of " trade clusters " is examined for various factors that may give rise to trade cluster effects. A trade cluster is defined as a geographical concentration of traded commodities that are in the same industry or in closely related industries. Trade clusters may exist in groups of distinct industries if some other factor provides positive spillover effects in trade. The paper first presents and discusses the average costs of transporting imports for each of eight sample customs districts. These estimates of transportation costs are derived from documents that require importers to report the value of imports and the cost of freight and insurance. Next, measures of concentration and similarity such as Herfindahl-Hirschman and Grubel-Lloyd indices are constructed and related to trade clusters. These measures are used in fixed-effects regressions for transportation costs, and the results suggest that trade cluster effects matter. The paper concludes with a discussion of policy implications that follow from the regression findings. Much has been written over the years about the phenomenon of spatial clustering and agglomeration effects. Specifically, there is an abundant literature focusing on the potential external economies of scale that participants in specific activities can capture by being physically located in close proximity to one other. Krugman (1991) provides a seminal analysis of the conceptual advantages and disadvantages of geographical clustering by producers, while Malmberg and Maskell (2002) review the empirical literature. There is also substantial empirical evidence identifying the determinants and consequences of geographical clustering across a range of industries and economic activities, such as production, sourcing of labor and other factor inputs, and innovation (see, for example, Bell (2005), and Globerman, Shapiro, and Vining (2005).) Most of the literature on clustering tends to focus on the benefits of firms co-locating " upstream " value chain activities such as research and development and production in a given geographic location, although the precise boundaries of the locations vary across studies. In most studies, the presence and strength of agglomeration economies are identified at the level of the city or metropolitan area, although regional clusters (e.g. Silicon …