The Geography of Multinational Firms

Over recent decades, foreign direct investment (FDI) has become a major force in the global economy. The geographical pattern of capital formation, trade and technological spillovers across countries and regions, are to an increasing extent determined by the strategies chosen by multinational firms (MNFs). Between 1982 and 1994, the rate of growth of the global FDI stock was more than twice that of gross fixed capital formation, the growth of sales by foreign affiliates of multinational firms well exceeded that of world exports, and, by 1994, the MNFs accounted for approximately 6 percent of world output (United Nations, 1997, pp. xv-xvi). The overall mechanisms behind this rapid internationalization in terms of multinational production have been attributed to the dismantling of trade barriers and the deregulation of capital markets, together with the advances in information technology that have facilitated the coordination and monitoring of internationally dispersed production. This development carries two important implications: First, firms operate in markets characterized by much tougher competition than only a decade ago, and, second, countries and regions are involved in competition for production to a much larger extent than before.

International economics contains a number of important sub-areas which have long been disjoint from one another. One of these is the traditional general-equilibrium theory of international trade, which constituted the core of both graduate and undergraduate trade courses for decades. Quite separately, another area of study was industrial organization aspectsof international trade, practiced by a separate group of economists with little overlap between them and the trade theorists. Each area had its strengths and weaknesses. Trade theory imposed an important general-equilibrium discipline on researchers, but was almost exclusively conducted under the twin assumptions of perfect competition and constant returns to scale. Any analysis of large national or multinational firms was excluded by definition, as was the possibility of gains from trade through the capture of scale economies or pro-competitive effects. The industrial organization approach to trade was often partial-equilibrium in nature and indeed often not even at that level, focussing instead on the international organization of individual firms. Yet that literature was able to examine important empirical phenomenon ignored by the more formal trade theory.

In recent decades the world has witnessed a remarkable growth in overseas production by foreign affiliates. An increasing number of firms and countries are involved in this process and in 1996 as much as one third of all outward foreign direct investments were directed to less developed countries (UN, 1997). Annual FDI flows almost doubled between 1991 and 1996 and for several years, the aggregate output of affiliates located in foreign countries has outweighted exports as the dominate means of servicing foreign markets (UN, 1997).

It has long been recognized that, at the level of the firm, the decision to export and the decision to invest abroad are interrelated. They both concern different ways of supplying foreign markets. However, it was not until the mid 1980s that any attempts were made to incorporate FDI and multinational firms into general equilibrium trade models (e.g., Helpman, 1984, 1985; Markusen, 1984; Ethier, 1986). As was explained in , recent contributions in this area of research have stressed the interplay between proximity advantages and concentration advantages in determining whether firms choose to become multinationals or national exporting firms. If transport costs or other trade costs such as tariffs are high, the firm will have an incentive to locate production directly on the market where the good is sold. On the other hand, if there are strong scale economies at the level of the plant, the firm will have an incentive to concentrate production in a few sites, thus making it more likely that a foreign market will be supplied through exports (Brainard, 1993a; Markusen and Venables, 1996).

Trade across nations is either explained by differences in the endowments of factors of production, the production technology, or, in the so called “new” trade literature, by economies of scale and consumer preferences characterized by “love for variety”. All of these explanations embark from the assumption that trade flows occur between firms in one country and consumers in other countries. However, empirical evidence points to an increasing share of trade being channeled through foreign affiliates of multinational firms (MNFs). As much as one third of world trade is estimated to take place within firms (UN, 1995).

Until recently, locational analysis in international economics was mainly carried out by applying John Dunning’s so called eclectic - or OLI - approach.1 Dunning’s contribution (1977) was to provide a taxonomy that related micro- and macroeconomic variables in a consistent way to the pattern of foreign direct investment (FDI). However, as shown in the preceding chapters, the substantial theoretical advances made in the last decade enables the location of economic activities to be modelled as an endogenous process (Krugman, 1991; Brainard, 1993, 1997; Venables, 1996; Markusen, 1995).

Larger multinational firms commonly undertake R&D at home as well as in their foreign affiliates. However, MNFs still perform the major part of their R&D at home. This has been attributed to factors such as scale economies in R&D, the need of proximity to the company headquarters, and the desire to maintain strategic knowledge within firms (x7Caves, 1996, ).

From the end of the 1960s to the early 1990s, yearly outflows of foreign direct investment from Sweden were between three and five times as large as inflows. Between 1991 and 1995 inflows exceeded outflows each year. It may be too early (by the end of 1997) to determine whether this shift in net investment flows is a temporary phenomenon or a structural change toward more balanced flows in the future. However, the shift in investment flows is interesting enough to warrant some attention to inflows of foreign direct investment into Sweden. Short term variations in the aggregate inflow of FDI into any country may be related to “macro” variables, such as changes in the level of economic activity during the business cycle and temporary misalignment of the real exchange rate. It may also be related to policy variables such as changes in corporate tax rates and the level of subsidies. The long-term changes in the ownership pattern of a country’s industrial assets, on the other hand, are more likely to be related to changes in structural determinants at the “micro” level. Factors that affect the location of production in different industries are also likely to affect its ownership pattern. Political attitudes toward foreign ownership also change over time and such changes may influence FDI inflows through changes in legislation governing foreigners rights to own domestic assets. In this chapter, our analysis of the pattern of FDI inflows into Sweden will focus on such long-term structural issues.

Entry through FDI can either take the form of acquisitions of existing firms, or by setting up a new plant, i.e., greenfield investment.1 The choice of entry mode has several implications for the investing MNF as well as for the host country. Greenfield establishments tend to be more integrated with the parent company than acquired firms.2 Larger intra-firm trade volumes between parent companies and greenfield affiliates, indicate that transfers of embodied technology play a more important role when the chosen entry mode is greenfield (Andersson et al., 1996). Previous research has also shown that more technologically advanced firms prefer greenfield investments (Caves, 1996). Acquired firms, on the other hand, are characterized by their own corporate culture and connections with local subcontractors, and the ties to the parent company are usually much looser. Moreover, acquisitions may lead to knowledge transfers in either direction, since a takeover is also a means by which the parent increases its knowledge base.

Multinational production and market concentration are closely related. The empirical evidence shows that foreign subsidiaries often operate in highly concentrated markets. For example, (1958) and (1973) found that a majority of foreign subsidiaries in the UK operated in concentrated markets. (1981) reported similar results for UK, France and West Germany. (1978) showed that the share of activity carried out abroad by US firms was positively related to concentration. Thus, the correlation between market concentration and multinational activity is fairly well established in the empirical literature.