«edited by Steven Brakman and Harry Garretsen The MIT Press Cambridge, Massachusetts London, England ( 2008 Massachusetts Institute of Technology All ...»
Foreign Direct Investment
and the Multinational
edited by Steven Brakman and
The MIT Press
( 2008 Massachusetts Institute of Technology
All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage
and retrieval) without permission in writing from the publisher.
For information about special quantity discounts, please email email@example.com This book was set in Palatino on 3B2 by Asco Typesetters, Hong Kong.
Printed and bound in the United States of America.
Library of Congress Cataloging-in-Publication Data Foreign direct investment and the multinational enterprise / edited by Steven Brakman and Harry Garretsen.
p. cm. — (CESifo seminar series) Includes bibliographical references and index.
ISBN 978-0-262-02645-1 (hardcover : alk. paper)
1. Investments, Foreign. 2. International business enterprises. I. Brakman, Steven. II.
HG4538.F619198 2008 332.67 0 214—dc22 2007032377 1 Foreign Direct Investment and the Multinational Enterprise: An Introduction Steven Brakman and Harry Garretsen
1.1 Introduction One of the stylized facts about today’s world economy is the importance of foreign direct investment (FDI). Figure 1.1 compares the growth of world gross domestic product (GDP), world trade, and FDI.
What is particularly striking about this ﬁgure is that from 1990 onward, FDI grows far more rapidly than world GDP and world trade.
The sharp decline of FDI growth around 2000 corresponds to the worldwide fall of share prices that not only ended all speculation about the wonders of the new economy but also signaled a (temporary) halt to cross-border mergers and acquisitions, one of the main vehicles for FDI. Figure 1.1 is just one example of the importance of FDI. Similar bursts of rapid FDI growth occured in earlier periods (Eichengreen 2003, Obstfeld and Taylor 2003), and it might be expected that such a salient characteristic of the world economy would have been closely scrutinized by theoreticians and empirical researchers alike. It also seems reasonable to suppose that by now, a vast amount of literature that focuses on FDI—its causes and consequences—would exist. Until quite recently, however, this was not the case. The reason is that it is far from trivial to formalize and analyze FDI and its determinants.
The standard theories of international trade—in the absence of trade costs—have no need for international factor mobility and so do not encompass FDI. In the neoclassical view of the world, factor price equalization (FPE) removes all incentives for international factor mobility.
Indeed, in Heckscher-Ohlin-Samuelson (HOS) types of trade models, trade and factor mobility are perfect substitutes. In the transition period toward a new equilibrium, both trade and factor mobility are equally capable of restoring FPE. The literature traditionally focuses 2 Steven Brakman and Harry Garretsen
Figure 1.1Growth of world GDP, FDI, and tradeNote: 1970 ¼ 100.Source: World Bank (2004).
on the trade channel as the means of restoring equilibrium, because (ﬁnal) goods are assumed to be more mobile than factors of production. Similarly, in situations where a market distortion is present—for example, a tariff resulting in a failure of FPE—trade and capital ﬂows
are substitutes in the sense that capital inﬂow eliminates trade (Mundell 1957). Figure 1.1 suggests, however, that this may not be the case:
FDI growth mirrors the growth of international trade, even though FDI grows much faster than trade.1 This implies that the HOS trade model will not do and that alternative theoretical explanations are required to explain FDI and the presence of multinational enterprises (MNEs).
Becoming an MNE has obvious disadvantages: the need to set up a foreign plant or a sales network, to try to overcome cultural and legal differences, to bear the risk of expropriation, and, of course, exchange rate risks. Models that explain the existence of MNEs must highlight the potential beneﬁts of production in foreign markets and show that these are larger than, for instance, the costs of setting up a plant in a foreign market. An early attempt to do so is the so-called OLI approach of Dunning (1977). The O refers to ownership advantage. A ﬁrm must have a product or asset that is uniquely associated with this ﬁrm because of a patent, a brand name, a special production process, or some other characteristic unique to it. This provides the ﬁrm with market power since it supplies a product that is different from others Foreign Direct Investment and the Multinational Enterprise 3 in the market. The L refers to location advantages. Instead of exporting, a MNE chooses instead to produce in a foreign country because it is more proﬁtable. The additional proﬁts come from the fact that by setting up a foreign plant, the ﬁrm is able to avoid barriers to trade, like tariffs or transportation costs, which reduce competitiveness should the ﬁrm choose to export. A different location-related motive for FDI is to beneﬁt from lower factor costs in foreign markets. The I refers to the internalization advantage. It recognizes that even if the O and the L conditions are satisﬁed, a ﬁrm does not necessarily need to set up a foreign plant. It may simply choose to license a foreign ﬁrm to produce, that is, it might outsource part of its production. However, this could reduce long-run proﬁts if the foreign partner decides to defect on the original arrangement and start for itself (after gaining knowledge of the production process). In-house production reduces these risks. Assuming that the location issue has been solved and the ﬁrm opts for foreign production, the internalization issue is basically about whether this foreign production should be in the form of FDI or outsourcing.
Although interesting, Dunning’s approach is more an organizing framework than a model. It is useful because it identiﬁes elements that should be the ingredients for any full-ﬂedged model of the MNE and FDI, such as imperfect competition (the O of OLI), barriers to trade like transportation costs (the L of OLI), and internalization aspects (the I of OLI). The seminal contribution of Dixit and Stiglitz (1977), which provides an elegant and tractable way of incorporating these elements in a formal model, paved the way for the recent burst of MNE and FDI research activity in the ﬁeld of international economics. The development of the modern theory of MNEs resembles the development of other trade (related) theories, like the new trade theory and the new economic geography, where the workhorse model of Dixit and Stiglitz (1977) proved to be important as well because of the need to model imperfect competition. The elegant formalization provided by the DixitStiglitz model allows the analysis of increasing returns, imperfect competition, and product differentiation, elements that are crucial to understanding intraindustry trade. The addition of transportation costs in this model leads to the famous home market effect, which is fundamental to the explanation of agglomeration (see the contributions in Brakman and Heijdra 2004).
Helpman (1984) is one of the ﬁrst attempts to apply the Dixit-Stiglitz framework to MNEs. It is a two factor of production model with monopolistic competition in the sector that can potentially locate 4 Steven Brakman and Harry Garretsen headquarters activities in a different country from where production is carried out. In this model, it is assumed that headquarter services and production are characterized by different factor intensities. This gives rise to multinational behavior if headquarters and production can be separated. As usual in a three-commodity, two-factor model, the trade pattern is ambiguous, because many trade patterns are consistent with full employment. Allowing multinational behavior also implies that the employment of resources by a country might differ from its endowments, indicating that the so-called factor-price-equalization set is larger in the case of multinational production than in a world without MNEs. Helpman’s model applies to vertical FDI—production is located in only one country. Horizontal FDI is not possible by assumption, which is a serious drawback of the model, as most FDI is in the form of horizontal FDI. Furthermore, the bulk of FDI is between developed countries, implying that FDI is mostly market seeking rather than factor-cost seeking (Markusen 2002).
The integration of imperfect competition and horizontal FDI was the central element of the research program Markusen started in the 1980s and is summarized in Markusen (2002). If a ﬁrm decides to set up two plants in different countries and each plant sells to only the local market, a critical element is transportation costs. A ﬁrm has an incentive to become multinational if the additional costs of setting up a foreign subsidiary—the plant-speciﬁc ﬁxed costs—are offset by avoiding costs associated with barriers to trade. This implies that transportation or trade costs become an essential element of these models. The most general model in Markusen (2002), called the knowledge-capital model, combines both vertical and horizontal multinational behavior at the same time. Not surprisingly, given the prevalence of horizontal FDI in the data, tests of the knowledge-capital model reveal that the horizontal FDI model is empirically more relevant than the vertical FDI model (Carr, Markusen, and Maskus 2001).
These two examples of FDI and MNE modeling by Helpman and Markusen have been important for the development of modern MNE theory, but neither of these two approaches considers the question why the foreign plant has to be internalized (the I of the OLI approach) and why outsourcing will not do. Alternatives to full ownership are, for example, a joint venture or licensing to a foreign ﬁrm. The basic question is thus whether to insource or outsource. The existence of market failures implies that this is a nontrivial decision for the ﬁrm to make. The topic goes back to Coase (1937) and was elaborated by WilForeign Direct Investment and the Multinational Enterprise 5 liamson (1975, 1985). A few issues stand out: the hold-up problem, the asset speciﬁcity problem, the principal-agent problem (whether it can be expected that a foreign agent reveals the true nature of the foreign market), and various matching problems (see Rauch 2001 for the latter). The recent literature now also addresses these problems (Helpman 2006).
Although this introduction is far from complete, it sketches the background against which the chapters in this book were written. The book consists of two parts. In the chapters in part I, ‘‘Theory,’’ the modern theory on FDI and MNE as outlined above is taken as a starting point, and the common denominator is to show how the basic framework of this theory could or should be extended. In the second part of this book, ‘‘Empirics,’’ several of the empirical hypotheses concerning the determinants and effects of FDI associated with the modern theory of FDI and MNEs are tested.
1.2 Part I: Theory
In the chapter 2, Neary identiﬁes an empirical puzzle that he solves theoretically. As noted, horizontal FDI is more prevalent than vertical FDI. And remarkably, the increased integration of the European Union (EU)—or ongoing globalization for that matter—continues hand-inhand with ever increasing FDI. This is puzzling because according to theory, a decrease of FDI is expected: a fall in trade costs should be accompanied by a decrease in horizontal FDI. Neary suggests two solutions for this puzzle. The ﬁrst is the existence of hubs or export platforms for FDI. Foreign ﬁrms still jump over trade barriers to gain access to an integrated market that has low internal trade costs, markets like the EU, and will select a host country from where they will export goods to the rest of that market. The second solution to the puzzle comes from the application of his so-called GOLE (General OLigopolistic Equilibrium) model. In a series of papers, Neary has developed a model that, unlike the Dixit and Stiglitz (1977) framework, allows strategic interaction between ﬁrms within a general equilibrium framework (see Neary 2003, 2007, 2004). Interestingly, this model implies that increased integration leads to cross-border mergers and acquisitions. And since most horizontal FDI takes place through cross-border mergers and acquisitions, this does offer a solution to the puzzle. In this approach, Neary addresses the O of the OLI framework. In chapter 6 Steven Brakman and Harry Garretsen 3, Hoffman and Markusen extend the modern MNE literature by explicitly combining Markusen’s (2002) knowledge-capital model with elements of the new economic geography approach, and thereby deal explicitly with the L of the OLI framework. The authors focus on the effects of investment liberalization and ﬁnd, using simulation experiments, that over a wide range of parameters, headquarters tend to agglomerate but plants tend to spread. Headquarters become more concentrated in countries that are relatively well endowed with skilled labor or countries that are large. The increased spread of plants alters the general conclusion of new economic geography models—that the symmetric or spreading equilibrium is unstable for a wide range of parameter settings, since spreading now becomes the stable equilibrium.
The I of the OLI approach is addressed in chapter 4 by Naghavi and Ottaviano. Innovation does not take place in isolation; it is a global phenomenon. The central idea is that while outsourcing increases transaction costs, it also lowers the costs of governance. In this chapter, the static framework of Grossman and Helpman (2002) is reformulated as a dynamic framework. Naghavi and Ottaviano show that product innovation and matching probability are strongly interrelated.