The International Trade and Investment Program.

• Author: Feenstra, Robert C.
• Position: Program Report

The research of the International Trade and Investment (ITI) Program, which includes 90 current members, covers a wide range of topics, such as explaining patterns of international trade, foreign direct investment, and immigration, and improving our understanding of the impact of trade policies. In addition, specialized ITI conferences cover such topics as "Globalization and Poverty" and "China's Growing Role in World Trade." (1) These two projects illustrate that a good deal of our research is concerned with developing countries, although that will not be discussed in this summary. Here I focus on a few topics related to trade patterns and trade policy.

The Great Trade Collapse

The financial crisis and great recession of 2008-9 brought with it a "great trade collapse": world trade relative to GDP fell by nearly 30 percent between these two years, exceeding the experience of other post-war recessions. Why did trade fall so much, and why did it recover relatively quickly? The leading explanations stress, in varying degrees, the roles of: inventory adjustment for imports; demand for durable versus non-durable goods; the use of intermediate inputs in trade, which might magnify the impact on trade as "supply chains" are temporarily disrupted; and the role of trade credit, which appears to have dried up temporarily during the crisis.

Beginning with the last of these explanations, Kalina Manova and her co-authors provide the strongest evidence supporting the role of credit constraints on exports. These constraints limit the extensive margin of exports in sectors that are most vulnerable to financial stress. (2) Furthermore, she argues that such sectors faced greater reductions in their exports to the U.S. market during the financial crisis. (3) Tat idea is confirmed for Japan by Mary Amiti and David Weinstein. (4) They find that Japanese exporters faced greater reductions in their sales abroad if they were affiliated with main banks that performed poorly. Focusing on China, my co-authors and I find that firms faced tighter credit constraints on their exports than on their domestic sales, and that exports experienced a significant slowdown because of the 2008 crisis. (5) Ann E. Harrison and her co-authors find that, for the United States, import prices often rose during the crisis, which is inconsistent with falling demand but can arise from a supply constraint, such as a lack of export credit. (6)

Other work casts some doubt on the importance of export credit. George Alessandria and co-authors instead stress the role of inventory adjustment, which can lead to a rapid fall in imports as stocks are adjusted downwards. (7) Andrei Levchenko, Logan Lewis, and Linda Tesar also find a limited role for trade credit in their regression analysis of U.S. trade, but they use an accounting definition of "trade credit" that applies equally well to exports or domestic sales. (8) As an alternative explanation, they find that sectors which are more reliant on imported intermediate inputs suffered more during the crisis, because these supply chains were temporarily disrupted. Fabio Ghironi and his coauthors also stress the importance of imported inputs. They model the different components of aggregate demand (consumption, investment, government spending, and exports) as having different import intensities. (9) They then construct a weighted average of those factors with the weights reflecting their import intensities. Using the resulting variable as an income term, and including an import price, they are able to construct a model that predicts the fluctuations in import demand during the current crisis and earlier episodes much more accurately than do conventional methods that rely on GDP and aggregate prices.

Of course, in the end it will be a combination of factors that explain the great trade collapse: even if inventories or imported intermediates are more important quantitatively, that finding need not detract from the significance of trade credit. Amiti and Weinstein, for example, argue that trade credit can account for about 20 percent of the fall in exports for Japan, so it was not the most important factor, but it was still economically significant. That point is also made for Peruvian exports by Veronica Rappoport and co-authors, who argue that the reduction in loans from banks performing poorly reduced aggregate exports by 15 percent during the crisis. (10) Perhaps the most comprehensive evaluation of the different factors contributing to the great collapse in trade was written by Jonathan Eaton, Sam Kortum, Brent Neiman, and John Romalis. (11) They argue that the relative decline in demand for manufactures was the most important driver of the decline in manufacturing trade, and especially the decline in demand for durable manufactures. These factors account for more than 80 percent of the global decline in trade/ GDP. While they find that trade frictions increased and played an important role in reducing trade in some countries, notably China and Japan, these frictions only had a small impact on global trade.

Offshoring, Wages, And Employment

One of the explanations mentioned earlier for the great trade collapse was that supply chains may have been disrupted during the crisis. While the "supply chain" concept is often mentioned in the social sciences, it has had limited modeling within the international trade context. That shortcoming is being addressed in very recent research. Arnaud Costinot, Jonathan Vogel, and Su Wang model a sequential supply chain in which mistakes potentially occur at each stage in a continuum. (12) There are many countries which differ in their probabilities of making mistakes, and in equilibrium there is a matching between stages of production and countries. Richard Baldwin and Anthony Venables call this type of sequential product chain a "snake" and label the assembly of multiple parts at a central facility a "spider." They provide a partial equilibrium model that illustrates the difficulties of solving for the location of stages in this framework and also make clear that the assignments might be non-monotonically related to transportation costs. (13)

Closely related to the supply chain concept is the role of intermediaries who provide services between buyers and sellers. Examples include large trading houses, such as "Li and Fung" in Hong Kong. Recent research by Costinot and Pol Antras has modeled these intermediation activities. (14) JaeBin Ahn, Amit Khandelwal, and Shang-Jin Wei provide empirical evidence on the role of intermediaries in China. (15)

Also closely related to international supply chains is the fragmentation of production across borders, or offshoring. The most recent theoretical paradigm for offshoring draws on "trade in tasks," which is described in work by Gene Grossman and Esteban Rossi-Hansberg. (16) In this framework, offshoring in low-skilled tasks acts like labor-saving technological progress in that factor. At unchanged prices for goods--as in a small-country framework--increased offshoring raises the wages of low-skilled labor. In contrast, when the prices of goods are endogenous--as in a large-country framework--increased offshoring of less-skilled tasks raises the output of that good and lowers its relative price. This change in relative prices has the expected result of lowering the real and relative wage of less-skilled labor, consistent with earlier work on "trade in inputs" by Gordon Hanson and me. (17) The overall change in wages depends on whether the impact of labor-saving technological change attributable to offshoring dominates the opposite effect of changing international prices, which depends on parameters of production and other features of the economy.

This work on offshoring has been extended by ITI Program members in a number of directions. Richard Baldwin and co-authors integrate the earlier "trade in goods" and "trade in tasks" frameworks, as well as examining the role of heterogeneous firms. (18) Andres RodriguezClare examines the impact of offshoring in a Ricardian model with a continuum of industries. (19) Costinot and Vogel provide the most general treatment of offshoring attributable to factor endowment differences, in a model with a continuum of goods and factors. This leads to a sophisticated matching of factors with goods, for which they provide a complete solution. (20) Antras, Luis Garicano, and Rossi-Hansberg consider the effects of offshoring in a model of multinationals where managers monitor and solve problems for workers. (21) Ariel Burstein and Vogel also consider the role of multinationals that bring technology to the host countries. (22) Based on a quantitative exercise, they argue that the growth of multinationals has been at least as important as the growth of trade in explaining the rising skill premium in the United States. Finally, Grossman and Rossi-Hansberg model offshoring between similar countries, where it is not factor-price differences that determine the location of production, but rather local external economies. (23)

Ann E. Harrison, Margaret S. McMillan, and co-authors provide new empirical studies of offshoring, using data on U.S. multinationals and data from the Current Population Survey (CPS). (24) They find that it is occupations rather than particular industries that are the best unit of analysis for identifying the wage effects of offshoring, which can be significant. Runjuan Liu...