International trade and investment.

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

The International Trade and Investment (ITI) Program does research on patterns of international trade and foreign direct investment; policies designed to influence the level of trade and investment; and their consequences on importing and exporting countries, such as for their wages, growth, the environment, and so on. Empirical work in the Program has benefited from several new datasets covering both U.S. and global trade at a detailed level; these are now available from the NBER (see NBER Service Brings You New Data for Free). (1) In this review, we cover work completed since last the Program Report in Winter 2000/2001, beginning with a new research area dealing with the microeconomics of the trading firm.

Microeconomics of The Trading Firm

Traditionally, theories of international trade have explained trade patterns by appealing to differences in the factor endowments found in various countries or to cross-country differences in industry productivity. That type of research continues, extending earlier models to allow for multiple industries, factors of production, and countries. (2) However, a new line of research digs deeper into the determinants of trade by allowing for differences across firms and recognizing that only the most productive firms will become exporters. That theoretical prediction receives strong empirical confirmation; generally, the new theory allows for a rich exploration of firm-level differences in datasets for the United States and other countries. Here are summaries of several research areas within this broad topic:

Firm Heterogeneity

The first way that firms can differ is in terms of productivity. Jonathan Eaton and Samuel Kortum introduced a Ricardian model with heterogeneous firms. (3) In that model, firms receive random productivity draws and compete with other firms producing the identical product, so that only the most productive firm survives within each country. Across countries, however, firms face differing transportation costs on their sales to external markets, so that multiple firms can be producing the same product and selling to different markets.

A second model with heterogeneous firms is attributable to Marc Melitz. (4) His work builds on an earlier model of monopolistic competition and trade in which goods are differentiated. In contrast to other researchers, Melitz allows the firms within an industry to be heterogeneous in their productivities. Each firm has to pay a fixed cost (for example, to develop its differentiated product), so that only the more productive firms will end up being profitable, while the least-productive firms exit the market. Furthermore, Melitz assumes that there is an additional fixed cost of exporting (for example, to market the product abroad), so that only the most productive firms find it profitable to export. This model has been extended to allow for multiple industries with differentiated products in each. (5)

These ideas have been applied to datasets on U.S. and French firms, as well as some for developing countries. (6) For the United States, Andrew B. Bernard, J. Bradford Jensen, and Peter K. Schott have studied the factors leading to the exit of manufacturing firms, including competition from low-wage countries and declining trade barriers. (7) Generally, exits occur less frequently at multi-product plants, at exporters, and at plants paying above average wages. In addition, productivity growth is faster in industries with falling trade costs, and plants in industries with falling trade costs are more likely to die or become exporters. Eaton, Kortum, and Francis Kramarz find that in France, firms differ substantially in export participation, with most firms only selling at home, and that markets in which French firms have a large share are also those where many more firms are exporting. (8)

These empirical applications depend on having firm-level datasets, which are not always available. An alternative is to use product-level trade data. This approach does not allow for the measurement of firm heterogeneity, but does allow for the entry and exit of products across years, as analyzed by Thomas I. Prusa. (9) David Hummels and Peter J. Klenow decompose the growth of world trade into that part attributable to countries exporting new products --what they call the "extensive margin"--and that part attributable to countries exporting more of the same products--the "intensive margin." They find that extensive margin accounts for two-thirds of the greater exports of larger economies, and one-third of their imports. (10) Hiau Looi Kee and I estimate the impact of new goods on productivity growth for the exporter, and find that export variety accounts for 13 percent of within-country productivity growth. (11) Conversely, Christian Broda and David E. Weinstein measure the impact of new goods on the welfare of the importer. For the United States, they find that upward bias in the conventional import price index (because of ignoring product variety) is approximately 1.2 percent per year, implying that the welfare gains from cumulative variety growth in imports are 2.8 percent of GDP in 2001. (12)

Incomplete Contracts

Aside from firm heterogeneity, a second theoretical innovation has been to take partial-equilibrium models of incomplete contracts between firms and apply them to a general equilibrium setting with trade. One example of this approach is the work by Pol Antras dealing with the well known "product cycle," originally studied by Raymond Vernon. Under this story, new products are developed in advanced countries like the United States or Europe, and only later do these technologies diffuse to developing countries where wages are lower. What factors explain this diffusion? While earlier research on growth models has stressed the imitation of products by developing countries, Vernon's story instead had the technologies voluntarily transferred abroad, either within the multinational firm or between firms. How are we to explain the decision of the firms to transfer their production abroad?

Antras specifies that contracts between a firm and its subsidiaries are incomplete, and shows how the dynamics of the product cycle can be described effectively. (13) In particular, he models the Northern firm as having two activities: R and D and production. It is more difficult to write contracts to specify and compensate the R and D activity, and more difficult still to write these contracts in the South. Over time, however, R and D becomes less important relative to production. With this framework, Antras solves for the equilibrium time at which the Northern firms will shift production to the South, and for whether the firm will engage in multinational activity there, or arms-length contracts that license its technology to unrelated firms. In other work, Antras finds that capital-intensive industries are more likely to engage in intra-firm trade across borders, and he offers an incomplete-contracting explanation for this finding. (14)

There are many other papers that explore incomplete contracts and outsourcing. Gene M. Grossman and Elhanan Helpman develop general equilibrium models of outsourcing building upon either the property-rights approach or the incentive-systems approach to the theory of the firm. (15) Gordon Hanson and I test these two approaches using data for processing trade in China, while Keith Head, John pies, and Barbara J. Spencer examine vertical networks in Japan, (16) and Deborah L. Swenson considers U.S. offshore assembly. (17) Motivated by evidence on the importance of incomplete information and networks in international trade, James E. Rauch and Joel Watson investigate the supply of "network intermediation." (18) They provide both empirical evidence and a theoretical explanation for this activity. Finally, Diego Puga and Daniel Trefler examine how the tension between innovation and the control over this activity shapes the organization of the firm. (19)

Foreign Direct Investment

Both the monopolistic competition model with heterogeneous firms and the incomplete contracting model can be used to analyze foreign direct investment (FDI). The challenge in the FDI literature has been to also explain why firms need to have ownership in their foreign subsidiaries, rather than just exporting or licensing their technologies abroad. By modeling this decision as being made by heterogeneous firms under incomplete con tracts, one can obtain new insights into the determinants of FDI.

Helpman, Melitz, and Stephen R. Yeaple model the decision of heterogeneous firms to serve foreign markets either through exports or FDI. (20) These modes of market access involve different relative costs, some of which are sunk while others vary with sales volume (such as transport costs and tariffs). Relative to investment in a subsidiary, exporting involves lower sunk costs but higher per-unit costs. In equilibrium, only the more productive firms choose to serve the foreign markets, and the most productive among this group will further choose to serve the overseas market via FDI. Testing their predictions on data of U.S. affiliate sales and exports, they confirm that having more productive firms leads to significantly more FDI relative to export sales. Likewise, Head and Pies confirm this prediction for Japanese multinationals. (21)

Prior literature on multinationals has distinguished two main reasons for FDI to occur: "vertical" investment, which takes advantage of lower factor prices abroad; and "horizontal" investment, which takes advantage of proximity to foreign markets by operating abroad. Recent literature has recognized that the rationale for FDI is more complex, though. Grossman, Helpman, and Adam Szeidl expand the set of choices available to the firm to include production of intermediate goods and assembly performed at home, in another Northern country, in the low-wage South, or in several of these locations. (22) Notice that these...