Regional Aspects of Multinationality and Performance, Volume 13 (Research in Global Strategic Management)

LIST OF CONTRIBUTORS Christian Geisler Asmussen Center for Strategic Management and Globalization, Copenhagen Business...

Author: Alan Rugman

55 downloads 853 Views 2MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

LIST OF CONTRIBUTORS Christian Geisler Asmussen

Center for Strategic Management and Globalization, Copenhagen Business School, Frederiksberg, Denmark

Andreas Bausch

School of Economics and Business Administration, Friedrich Schiller University, Jena, Germany and School of Humanities and Social Sciences, Jacobs University Bremen, Bremen, Germany

Kathrin Boesecke

School of Humanities and Social Sciences, Jacobs University Bremen, Bremen, Germany

Harry P. Bowen

McColl Graduate School of Business, Queens University of Charlotte, NC, USA

Stephen Chen

School of Business, Macquarie University, Australia

Farok J. Contractor

Rutgers Business School, Rutgers University, Brunswick, New Jersey, USA

Fabienne Fortanier

University of Amsterdam Business School, Amsterdam, The Netherlands

Thomas Fritz

School of Humanities and Social Sciences, Jacobs University Bremen, Bremen, Germany

Ajai S. Gaur

Department of Business Policy, National University of Singapore, Singapore

Anthony Goerzen

Faculty of Business, University of Victoria, Victoria, British Columbia, Canada

ix

x

LIST OF CONTRIBUTORS

Walid Hejazi

Rotman School of Management, University of Toronto, Toronto, Ontario, Canada

Alina Kudina

Warwick Business School, The University of Warwick, Coventry, UK

Vikas Kumar

ISEA Istituto di Strategia ed Economia, Aziendale G. Zappa Universita Bocconi, Milano, Italia

Dan Li

Kelley School of Business, Indiana University, Bloomington, IN, USA

Lei Li

Pamplin School of Business Administration, University of Portland, Portland, OR, USA

Alan Muller

University of Amsterdam Business School, Amsterdam, The Netherlands

Lilach Nachum

Baruch College, City University of New York, New York, NY, USA

Chang Hoon Oh

Faculty of Business, Brock University, St. Catharines, Ontario, Canada

Thomas Osegowitsch

Department of Management and Marketing, University of Melbourne, Melbourne, Victoria, Australia

Nicole Richter

Institute of Industrial Management, Hamburg University, Hamburg, Germany

Alan M. Rugman

Kelley School of Business, Indiana University, Bloomington, IN, USA

Andre´ Sammartino

Department of Management and Marketing, University of Melbourne, Melbourne, Victoria, Australia

Nessara Sukpanich

Department of Economics, Thammasat University, Bangkok, Thailand

List of Contributors

xi

Rob van Tulder

Department of Business-Society Management, Rotterdam School of Management, Rotterdam, The Netherlands

Clifford Wymbs

Baruch College, City University of New York, New York, NY

George S. Yip

Capgemini Consulting, London, England

PREFACE I am pleased to acknowledge the invaluable assistance of Mildred Harris and Anne Hasiuk in the preparation of this book. In addition, the authors responded quickly to the invitation to submit original papers and were good natured in their responses to numerous remarks and queries. I am grateful to my editors at Elsevier, Mary Malin and Helen Collins, for their attention to this book. Alan M. Rugman Editor

xiii

INTRODUCTION Alan M. Rugman In the field of international business one of the most basic issues is the relationship between multinationality and performance. Several hundred studies have examined the nature of this relationship, with somewhat inconclusive results. This literature is reviewed and extended in Part B of this book. However, the main contribution of this book lies in Parts A and C which explore the regional dimension of multinationality and performance. In Part A of this book, five original chapters consider how the regional aspects of multinational activity can be incorporated into this large existing empirical literature testing the relationship between multinationality and performance. In the first chapter, Contractor presents a theoretical justification for the 3-stage S Curve. In an interesting twist he argues that the middle stage 2 is consistent with the observations on the regional nature of multinational activity. He also suggests that the final stage 3, where performance suffers due to excessive multinationality, is typically populated by relatively few firms. This fall off in performance beyond a certain threshold of multinationality may correspond to an attempt by some companies to reach a ‘global’ stage. While these propositions remain to be tested, Contractor provides strong new insights into the possible relationship between the regional dimension of multinationality activity and the emerging literature on the S Curve. He also reviews seven generic reasons why studies of multinationality and performance yield different results, and he provides an agenda for research to further test the viability of the 3-stage S Curve concept.

Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 1–8 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13016-3

1

2

ALAN M. RUGMAN

In the next chapter, Oh and Rugman provide new data examining the trends of regional sales between 2001 and 2005. They start with the 2001 benchmark year used in the pathbreaking paper by Rugman and Verbeke (2004) as further developed in the databank reported in the book by Rugman (2005). They find a remarkable degree of stability over time where the ratio of regional-to-total sales averages 75.7 percent. They also provide data for the first time on the regional nature of assets of the world’s largest 500 firms. This averages 76.7 percent again with less than a 1 percent variation over the five-year period. The data in this chapter need to be contrasted with that in the following chapter. Osegowitsch and Sammartino conduct an analysis by taking the set of 380 firms reported in Rugman and Verbeke (2004) and going back to 1991 to find trends in regionalization over the period 1991–2001. One problem is that this yields a greatly reduced sample of 159 firms biased toward the U.S. firms. Indeed, most of their results exclude the Asian firms completely, and represent only 90 North American firms and 36 from Europe. Osegowitsch and Sammartino find that there is a reduction in intra-regional sales for this small set of U.S. and European firms from 84 percent to 76 percent. Over this period they also find that some firms increase their sales outside their home region. They also find that a somewhat larger percentage of firms can be classified as bi-regional contrasted to those found in the Rugman and Verbeke chapter. However, in contrast to their own conclusions their own data actually confirm the two main insights of Rugman and Verbeke (2004). First, the vast majority of firms remain home region oriented over the 1991–2001 period. Second, there are very few global firms. Their chapter is useful in provoking a debate about the trends in regionalization and hopefully a more extensive dataset can be constructed to help reconcile their findings with those of Oh and Rugman. In the penultimate chapter in Part A, Goerzen and Asmussen use a set of Japanese multinationals to test the relationship between regional and global firms. They argue correctly that the performance of a firm is determined by its firm-specific advantages (FSAs). They show that regional firms build more upon location bound FSAs, especially in the marketing area. In contrast, global firms have technological (R&D) FSAs, and these are presumed to be nonlocation bound. They find evidence that the FSAs of global firms are less location bound than the FSAs of regional firms. This work needs to be extended beyond the Japanese dataset in order to test the generalization of these findings. However, their theoretical logic is consistent with that of Rugman and Verbeke (2007), where it is argued that there is a liability of regional foreignness. In other words, the FSAs of multinational enterprises (MNEs) are difficult to deploy and exploit outside of the home region.

Introduction

3

In the final chapter of Part A, Hejazi introduces the logic of international economics and transaction costs to analyze the regional dimension of the activities of MNEs. He uses the well-known gravity model of international economics, which has been used to analyze the importance of geography on the determination of international trade flows. In this chapter, he adapts the gravity model to measure foreign direct investment (FDI) instead of trade flows. The gravity model measures the country-level frictions affecting trade and can only be applied to FDI with some difficulty, as FDI is partly a method to overcome such frictions. Thus the gravity model yields a new type of test of the activities of MNEs, although it is not a direct test of their strategies. Hejazi finds that there is a strong regional bias in the activities of MNEs from the EU, but he does not find this effect for North American MNEs, which is not surprising given the asymmetrical large size of the U.S. market. (The size bias of the U.S. market may also affect the results by Osegowitsch and Sammartino.) While the chapter by Hejazi does not test performance directly it offers new conceptual lenses on the nature of the regional dimension of multinationality. Such work using econometric techniques based on the gravity model needs to be taken up and related to the empirical literature on multinationality and performance. Overall, Part A of this book provides many stimulating ideas on the regional dimension which warrant further research within the context of the literature on multinationality and performance. In Part B of this book, five original chapters reexamine the nature of this basic relationship. They provide new insights into both the theory and empirical aspects of firm performance and the degree of multinationality. Chapters test the S Curve fit, industry effects, moderating role of strategic fit, and impact of global cities. Other chapters conduct a meta-analysis and further explore the theoretical aspects of the basic empirical relationship between multinationality and performance. In the first chapter in Part B, Bowen provides a theoretical critique of the extant empirical work in the multinationality and performance literature. He points out that basic statistical issues have not been resolved, including the issues of endogeneity and nonlinearities in the tests. He argues that the multinationality and performance literature does not take into account the heterogeneous nature of firms, industries, or countries. He argues that the multinationality and performance literature needs to be much better integrated into basic international business theory and that the various ‘‘modes of multinationality’’ (exporting; FDI; outsourcing) can affect measurement. Such variations in types of multinationality are consistent with the observations of Rugman (2005) on the regional nature of multinational

4

ALAN M. RUGMAN

activity. A related criticism of the lack of theoretical underpinnings in the multinationality and performance literature has been advanced by Li, Goerzen, and Verbeke (2007). It is clear that much more effort needs to be put into the development of appropriate theoretical frameworks to model the observed empirical relationship between multinationality and performance. In the second chapter of Part B, Bausch, Fritz, and Boesecke, conduct a meta-analysis of a large set of previously published studies in the literature on multinationality and performance. They confirm a positive relationship between multinationality and performance. They have a broad definition of multinationality and include the traditional type of merger and acquisition (leading to wholly owned subsidiaries through the process of FDI) along with the nontraditional type of alliance formation. This leads them to invent yet another team for multinationality, namely international business combinations. It is unusual to include alliances in this type of work since it is difficult to assess the impact of alliances on firm performance in a direct manner, as can be done with the merger and acquisition mode of multinationality. Some challenging ideas are presented in this chapter, which attempts to extend the field of study by adding the alliance as an additional unit of analysis. It is particularly important to extend these tests to fully address the regional significance of different types of international business combinations. The chapter by Fortanier, Muller, and Tulder offers a cautionary tale for researchers on multinationality and performance. They find that the empirical results testing this relationship are strongly affected by moderating variables. In particular the so-called strategic fit affects performance in a significant and positive manner. Strategic fit moderates the basic aspects of multinationality and performance including the shape, size, and direction of the relationship. In this chapter, strategic fit is based upon the integration and national responsiveness framework (which is also used in the chapter by Li and Li in Part C). Fortanier and her coauthors collect archival data on the chain of ownership of multiple subsidiaries that allows them to conduct more robust econometric analysis on the aspects of integration and responsiveness. They have data for 336 of the world’s largest 500 firms for the year 2002. These variables enter as moderators when testing the basic relationship between multinationality and performance. The results indicate a significant positive relationship between multinationality and return on sales, but this relationship is not a stable one since the strategic fit varies by firm. In other words firms with an integration strategy perform well internationally in industries which are integrated. In contrast, in multidomestic industries only

Introduction

5

firms with a national responsiveness strategy do well internationally, so industry effects matter, as also found by Li and Li. This chapter is interesting as it attempts to introduce organizational structure and strategy issues more explicitly into the literature on multinationality and performance. However, the regional dimension is not explored in this chapter. The chapter by Kumar and Gaur examines the relationship between multinationality and performance within the context of 240 of India’s MNEs, many of which are smaller firms than the world’s largest 500 for which the regional effect has been tested. They find strong evidence of a positive J-shaped exponentially increasing relationship between the internationalization of Indian firms and their performance. They also find that India’s outward FDI differs between developing and developed economies and between manufacturing and service sectors. A key contribution of their chapter is that their data include relatively small and medium-sized firms, not just the world’s largest 500 firms as in several of the chapters in Part C of this book. This helps us better understand the country context in studies of multinationality and performance, as India has many small multinational firms. Usually size of firms is a moderating or control variable, but Kumar and Gaur link it to a country factor for India. The final chapter in Part B, by Nachum and Wymbs, offers an interesting contrast to all others in this book. Their geographic unit of analysis is the city. This is a sub-national unit of analysis, and it can be contrasted with the triad regions developed in Rugman (2005) and tested by others in this book. A very good reason is given for choosing cities – namely that the data tested relates to the financial and professional service industries. These are clustered in the world’s major cities. The authors analyze 673 MNEs in these service sectors that entered New York and London through mergers and acquisitions. They find an interaction between geographic location and the FSAs of these MNEs. It is a useful idea to apply the location decision for an industry and firm at the appropriate geographic level. These findings can be usefully contrasted with several of the chapters in Part C, where regions cross-national borders, rather than being sub-national. In Part C of the book, five original chapters extend the traditional empirical work on multinationality and performance by including the regional dimension. Recent work has shown that the world’s largest firms operate mainly on an intra-regional basis, in terms of both sales and assets (Rugman, 2005). These chapters reexamine these data and relate the traditional literature on multinationality and performance to the new metrics available. The latter include new data on regional sales and on the return on foreign assets (ROFA). The five chapters in Part C take up the

6

ALAN M. RUGMAN

challenge of testing the significance of these regional aspects of multinationality and performance. Various chapters examine regionalization and performance across industries, over time and for various regions of the triad. The first chapter in Part C, by Lei Li and Dan Li, provides an innovative test of the regional aspects of multinationality and performance. The authors use the well-known integration and responsiveness matrix to distinguish between a ‘global’ industry, which has a high degree of economic integration, and a ‘multidomestic’ industry, which is nationally responsive. They choose the computer and office equipment industry as an example of a global industry, and the soap, cleanser, and toilet goods industry as an example of a multidomestic industry. They find significant differences between the two industries in terms of international strategies, which lead to confirmation of the regional dimension in multinational operations. They also test the impact of FSAs in the two industries in terms of both R&D and marketing intensity. Their results indicate that FSAs are largely nonlocation bound in the global industry, but much home region bound in the multidomestic industry. In addition, they show that internationalization pace has a direct positive impact on firm performance in the global industry, but not in the multidomestic industry. The second chapter of Part C is by Rugman, Yip, and Kudina. They introduce a new dependent variable called return on foreign assets (ROFA). They contrast it with the traditional variable, return on total assets (ROTA). They also introduce a regional variable representing regional sales. They test the explanatory power of the regional sales variable with linear, quadratic, and cubic fits. They find that the regional variable explains ROTA in terms of the cubic fit but not ROFA. This data is focused upon a set of 27 U.K. multinationals of which 8 are in manufacturing and 19 are in service sectors; the regional variable represents sales in the EU. The results indicate that the 27 large U.K. MNEs experience strong intra-regional sales and that the regional sales variable is a significant variable affecting firm performance in a positive but nonlinear manner allowing for standard control variables. In a related chapter Sukpanich also includes an independent variable representing intra-regional sales, this time across MNEs in the triad regions of North America and the EU. She has data on 91 firms of which 67 are from North America and 24 from Europe. Of the 91 firms 66 are in manufacturing and 25 in services. She uses the COMPUSTAT database to access data on the FSAs of MNEs. These FSAs include R&D and marketing variables. She finds a strong positive linear relationship between the measure

Introduction

7

of intra-regional sales and performance. Performance is higher for firms based in their home region. This result contrasts with that of Chen as discussed next. Chen conducts a test of the multinationality and performance relationship across some service sector firms in an Asian context. He distinguishes between intra-regional sales and extra-regional sales for this set of service sector firms. He uses the same Osiris database as in the chapter by Rugman, Yip, and Kudina. He does not find support for the hypothesis that intraregional sales are a significant determinant of performance where performance is measured by ROTA. Instead, he finds that extra-regional sales are significant, in contrast to Sukpanich (and others). To some extent this may support some findings in the chapter by Osegowitsch and Sammartino. Further research is required, but it is encouraging to see this focus on the Asian firms. In the final chapter in Part C, Richter tests the importance of a regional sales variable across the UNCTAD set of the world’s 100 largest firms as ranked by foreign assets. This is a somewhat unrepresentative sample as it consists of the world’s most internationalized firms and is thus biased toward finding internationalization (and therefore regionalization). Richter uses the S Curve and finds a significant cubic fit between performance and multinationality where the latter is measured by the UNCTAD transnationality index. When testing the foreign intra-regional sales variable her results are ambiguous with either an S Curve or inverted U Curve supported. However, as with other chapters in this book, she finds that this regional sales variable is a significant determinant of performance. It would be useful to extend this type of research beyond the largest 100 firms (or the 500 largest in Rugman (2005)) to include many more MNEs that are small-to-medium sized. Indeed, it would be useful to test the regional dimension in the emerging literature on international entrepreneurship, some of which is focused upon the internationalization process of small and medium-sized firms. This will require some theoretical adjustments to the assumption that many of these firms are ‘born global’. This literature seems to find relatively fast internationalization of small firms in the information technology and computer sector, but there has not been careful testing of the regional aspects of such internationalization, and other sectors need to be added. Furthermore, the normal metric used is a scope variable dealing with the opening of foreign subsidiaries, whereas, better metric is to use the ratio of foreign to total sales (in this case, regional to total sales). While this work on international entrepreneurship remains to be undertaken, this book provides a very useful starting point in bringing the

8

ALAN M. RUGMAN

regional nature of MNE into the literature on multinationality and performance. The set of 15 chapters in this book exhibits uniformity in showing that the basic relationship between multinationality and performance is beset by issues of heterogeneity across countries, industries, and firms. Yet many of them also show that the regional nature of multinationality can be included in this work in a useful manner. Therefore, the regional dimension of strategy needs to be considered in future work analyzing the relationship between multinationality and performance.

REFERENCES Li, L., Goerzen, A., & Verbeke, A. (2007). Three preconditions for effective research on the multinationality-performance relationship. Management International Review, 47. Rugman, A. M. (2005). The regional multinationals: MNEs and ‘global’ strategic management. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Rugman, A. M., & Verbeke, A. (2007). Liabilities of regional foreignness and the use of firmlevel versus country-level data: A response to Dunning et al. Journal of International Business Studies, 38(1), 200–205.

THE EVOLUTIONARY OR MULTI-STAGE THEORY OF INTERNATIONALIZATION AND ITS RELATIONSHIP TO THE REGIONALIZATION OF FIRMS Farok J. Contractor ABSTRACT This chapter outlines a general theory of international expansion and its effect on the performance of firms. Using the lens of this theory, it addresses the question of why most companies are ‘‘regional,’’ in the sense that their geographical coverage seems to be far from complete. The chapter also treats the perplexing issue of the lack of congruence in empirical findings, over the 30-year history of the Multinationality vs. Performance sub-field in International Business studies. It argues that the apparently contradictory results of past studies are but subsets of the three stages of the general theory. Finally, the chapter indicates fruitful areas for further research.


11

12

FAROK J. CONTRACTOR

INTRODUCTION: THE COSTS AND BENEFITS OF INTERNATIONAL EXPANSION The objective of an international firm is to maximize its profits (or another strategic ‘‘performance indicator’’) by seeking an optimal geographical or spatial configuration of its activities. In the 50-year history of International Business studies, it is not surprising that many authors have indicated the advantages accruing to a firm from international expansion, as well as its costs. The costs of international expansion are felt in early growth outside the company’s home base in what Hymer (1976) described as the cost of foreignness, lately known as the ‘‘liabilities of foreignness’’ literature (e.g., Zaheer & Mosakowski, 1997). At a later stage, some internationally expanding companies may also incur the costs, or a negative net effect on profits or performance, from excessive international expansion beyond an optimal level. Over-expansion, be it in a domestic market, or internationally, is sub-optimal if the incremental costs exceed the incremental benefits of entry into an additional country market. This is the underlying implicit assumption behind the ‘‘regionalization’’ argument put forward by Rugman (2005), which indicates that the overwhelming majority of multinational companies fail to achieve ‘‘global’’ coverage. Rugman’s (2005) calculations, covering 380 of the Fortune 500 firms, showed that only 9 companies could be described as ‘‘global,’’ using his criterion of less than 50 percent of a company’s sales occurring in its home region and more than 20 percent in each of the other ‘‘triad’’ regions. Other scholars, by altering the criteria, have produced somewhat different results, showing that a slightly higher percentage of companies may be labeled as ‘‘global’’ (Osegowitsch & Sammartino, 2006). Nevertheless, the basic contention is apparent, that there are limits to global expansion – and that these limits occur well in advance of the number of countries on the planet – a number that has shown an inexorable increase in the past half century. At the latest count, the number of countries has passed 194 (see Fig. 1). But a casual perusal of annual reports or 10-K filings shows few of even the ‘‘giant’’ multinationals going beyond covering more than 50 nations through controlled affiliates.1 That there should be limits to global expansion is now accepted, although until Rugman’s (2005) analyses, it was not generally known how quickly this limit is reached. However, what is surprising is that, until recently, the sequencing of the benefits and costs over an international expansion path for a company was not articulated in the International Business literature. There was recognition that there would be both benefits as well as costs, of international

The Evolutionary Theory of Internationalization

Fig. 1.

13

Number of Countries Over Time. Source: http://uchicagolaw.typepad.com/ faculty/images/graph_1.jpg

expansion (and even here, the micro or firm-level cultural, organizational, and strategic reasons for the benefits and costs are not yet fully researched). However, the timing of costs and benefits and their effects on profitability or performance as a company expanded abroad, remained imprecise until 2003 when a general 3-stage theory began to emerge (Contractor, Kundu, & Hsu, 2003; Lu & Beamish, 2004; Thomas & Eden, 2004). The objectives of this chapter are (1) to further articulate the 3-stage theory, since it has not yet met complete acceptance, (2) to assert that the theory is indeed a general theory of international expansion, and show how it covers all empirical contingencies, (3) to indicate how seemingly contradictory empirical results from past ‘‘Multinationality-Performance’’ (M-P) studies can be reconciled through the lens of the general theory, and (4) to indicate how the theory relates to the ‘‘regionalization’’ literature in terms of the limits to international growth.

A GENERAL THEORY OF INTERNATIONALIZATION AND PERFORMANCE Over a considerable range of international expansion, incremental benefits exceed the incremental costs of expansion. Otherwise, firms would not venture outside their home base. This assertion that internationalization is beneficial for firms is the bedrock assumption of International Business studies. But this is true for only the middle stage, Stage 2 as shown in Fig. 2. There are two other stages, namely Stages 1 and 3, where incremental

14

FAROK J. CONTRACTOR

Performance

Empirical results when only Stages 2 and 3 statistically valid

General 3-Stage Theory Empirical results when only Stages 1 and 2 statistically valid

Stage 1 DOMESTIC

EARLY INTERNATIONAL

Stage 2

Stage 3

REGIONAL GROWTH

GLOBAL**

Degree of Internationalization

** The term “Global” is only used here to refer to over-internationalized companies and may not conform to the Regional vs. Global classification in Rugman (2005).

Fig. 2.

The General 3-Stage Theory.

international growth produces negative effects on performance, each for very different reasons. The 3-stage theory sequences the incidence of different costs and benefits over the international expansion path of a company. In Stage 1 (Early Internationalization), when the firm is just beginning its initial internationalization foray, there are considerable learning costs and organizational disruption (Doz, Santos, & Williamson, 2001). In several cases, a separate ‘‘international’’ division or department is created, which parallels and duplicates some of the functions of the domestic portion of the enterprise. Such additional fixed costs and duplicative overheads are, at least initially, borne by only one or a few foreign markets. The liabilities of foreignness literature specifically focuses on the additional costs of learning about foreign cultures and markets incurred by an internationally growing firm (Zaheer & Mosakowski, 1997) as well as the costs of seeking legitimacy and acceptance in different institutional environments (Kostova & Zaheer, 1999). Until several more foreign markets are added, or until higher foreign sales are achieved, the incremental costs of initial foreign expansion, per unit of foreign sales, or per country added, tend to be high enough that early internationalization produces a net negative effect on performance. Further international expansion in Stage 2 is labeled in Fig. 2 as ‘‘Regional Growth.’’ Several studies, from the recent ones by Rugman (2005) to venerable literature on internationalization, such as Johanson and Vahlne (1977), strongly suggest that the typical path of internationalization


15

is expansion into psychically, culturally and geographically contiguous markets. Firms tend to follow the least unfamiliar. This means expansion into the same geographical region as the home base of the company. Nevertheless, some are reluctant to use the term ‘‘regional’’ because of the recent emergence of so-called ‘‘born global’’ companies who quickly leap from a domestic focus to a point where they do more business outside their home market, than in it (Oviatt & McDougall, 1997).2 Examples would be Information Technology companies based in India whose foreign sales quickly overtake their domestic Indian business. Moreover, for several such firms, their clientele is not in the Asian region, but in the advanced nations of the EU and the US. But exceptions do not a theory unmake. Most companies pursue a path of incremental internationalization starting in their own region. By Rugman’s (2005) criterion, many so-called ‘‘born global’’ companies would still be described by him as regional, because they fail to meet his criterion of more than 20 percent of sales in two triad regions outside the home base. In Fig. 2, the labels ‘‘regional’’ or ‘‘global’’ are only used loosely, it being understood that the principal focus of the graph, the x-axis, measures ‘‘Degree of Internationalization’’ as a continuous, as opposed to a categorical, variable. In Stage 2, companies enjoy the benefits of international growth resulting from exploitation of idiosyncratic and mobile firm-specific assets in foreign settings (Rugman & Verbeke, 2003). Once in foreign locations, the multinational firm may also access cheaper or better inputs – be they lower cost labor, or knowledge – and transfer these back for the benefit of other company operations (Dunning, 2002). Once international scale is achieved, at least some companies can enjoy lower costs from fuller utilization of installed capacity, or the benefits of disaggregation of the value chain over different nations according to the comparative advantage of each location. The fact of operating, at once, in several nations, may confer on some firms the advantage of greater strategic flexibility in responding to asynchronous business cycles, or supply chain disruptions, and lower foreign exchange volatility, if cash flows are a mix of currencies (Contractor, 2002). Occasionally, a few companies may be able, by virtue of being large and multinational, to accumulate international market power (Kogut, 1985). Finally, by repeated expansion into additional foreign markets, a company may accumulate internationalization experience, in terms of an organizational template to replicate subsidiary organizations (at lower cost than competitors, and certainly at lower cost than during its own early international growth). As a result of these multiple benefits of Stage 2 expansion, the slope of the Performance vs. Degree of Internationalization (DOI) graph in Fig. 2 is

16

FAROK J. CONTRACTOR

positive.3 There remain incremental costs of expanding into each additional country market, or of increasing foreign sales. But these are more than offset by the incremental benefits of further expansion. It is a fact worth noting that, in virtually every one of the over 150 empirical studies on this subject, the results show a positive slope on some portion of the range. In Stage 3, beyond some inflexion point, further internationalization is hypothesized to be sub-optimal, because the benefits of still further expansion are less than the incremental costs. In Fig. 2 this is depicted as a negative slope for the Performance vs. DOI curve. The label ‘‘Global’’ is only used as a shorthand for companies that may have knowingly or unknowingly over-expanded. Under Rugman’s (2005) definition, even highly internationalized companies (with a high Foreign to Total Sales ratio) may be considered by him to be ‘‘regional’’ in terms of their geographical spread. This is only a classification issue and does not detract from the basic theory that excessive internationalization can be sub-optimal, a contention that Rugman would readily accept.

THE LIMITS TO INTERNATIONAL EXPANSION The more interesting question is why internationalization beyond a threshold produces net negative effects on performance. What strategy, or organizational theory considerations can explain the fact that most multinational companies have a rather limited geographical scope? In fact, there are reasons to believe that a majority (and perhaps a substantial majority) of multinational companies have but one foreign affiliate.4 These are generally small or medium-sized enterprises. At the other end of the size spectrum, with giant multinationals, one would expect a very wide, if not universal coverage. But this is not so. Table 1 shows the ‘‘Transnationality Index’’ (TNI) calculated annually by UNCTAD 2004, 2005, 2006. This index is the average percentage of three ratios: Foreign Assets to Total Assets, Foreign Sales to Total Sales and Foreign Employees to Total Employees. For the world’s 100 largest enterprises, the TNI is around 56 percent. As one goes downward in size to the world’s top 1,000 firms and so on, there is a significant downward drop in the TNI index. Only a small fraction of multinational companies have their own managements in more than 50 nations (out of the 200 or so on the planet). Why? The question of limits to international expansion is inadequately researched. One can advance three hypotheses. The first hypothesis is that


Table 1.

17

Average Transnationality Index for the World’s 100 Largest Multinational Companies.

Year

2002

2003

2004

Biggest 100 Firms Together

57.0

55.8

56.8

By Individual Home Nation United States United Kingdom France Germany Japan

43.8 70.4 69.0 46.9 43.6

45.8 69.2 59.5 49.0 42.8

48.2 70.5 62.3 52.2 52.2

Sources: UNCTAD (2006) and UNCTAD (2005). Transnationality Index=Average of (Foreign Assets to Total Assets, Foreign Sales to Total Sales and Foreign employment to Total Employment).

the world is too big, with too many nations, most of which are tiny and peripheral markets. A rank ordering of countries by size of economies reveals a sharp drop-off after rank number 20. The 50th largest economy in 2005 was Hungary, a recent member of the EU, but had a GDP in Purchasing Power Parity terms that was just 1.34 percent of the size of the US economy, and a mere 0.27 percent of the world total economy (U.S. Central Intelligence Agency, 2006). Clearly, more than 160 of the world’s nations are very minor and peripheral markets. The second hypothesis is that much of the world is perceived (correctly or incorrectly) as too risky. Various country risk ratings reveal corporate respondents’ perceptions. The risk rating for the 50th ranked country, and below, is often half, or less, compared to the score for the highest ranked or the ‘‘safest’’ country. The third hypothesis is that cultural distance (Shenkar, 2001) between nations remains great enough that, beyond say the 50th nation in a multinational’s portfolio, cultural differences loom large and impose high costs (Bartlett & Ghoshal, 1990; Sunderam & Black, 1992). Delios and Henisz (2003) and Peng (2003) develop the concepts of institutional and regulatory distance between nations. Hitt, Hoskisson, and Kim (1997) offer an organizational behavior explanation for the limits to international expansion. They suggest that coordination costs and information overload increases with the extent of internationalization. In a schematic diagram, Lu and Beamish (2004) draw ‘‘coordination costs’’ as increasing at an accelerated rate with the degree of internationalization. However, in organizational theory terms, we still have a lot to learn about management practices in multinational companies.

18

FAROK J. CONTRACTOR

What exactly are the costs of cultural, psychic, institutional or regulatory distance between the firm’s home base and each foreign operation? How are these felt by managers, and how are they measured? This remains an area for further research. Unless felt by managers, and measured, how is a company to know that it has over-expanded? Many firms may not know, unless they undertake an empirical study for their industry. The S-Curve depicted in Fig. 2 is not just theory. It is also a management tool. But plotting the position of all firms in a sector on a Performance vs. DOI map, and then statistically fitting a curve with a linear-, quadratic-, and cubic-term for DOI, one can see where one’s own company lies in relation to others and the fitted S-Curve. Alas, the operationalization and testing of the general theory is not simple and often includes methodological problems as we see in the 30-year history of empirical studies in this field.

A 30-YEAR ACCUMULATION OF SEEMINGLY CONTRADICTORY EMPIRICAL STUDIES: THE FAILURE OF INDUCTIVE REASONING For the field of International Business, few questions are more fundamental than the link between firm performance and its DOI. Yet, empirical investigations for at least 30 years have failed to produce agreement. Early works that laid the foundation for the field, such as Caves (1971), Hymer (1976), or Buckley and Casson (1976) simply asserted that multinationality was desirable compared to domestic operations, since presence in several nations enabled the firm to exploit its internalized advantages and achieve economies of scale. But while implicitly accepting the notion of a positive relationship between performance and the degree of internationalization, there was no further examination of the slope, the shape or extent of the link. Subsequently, a plethora of studies, from Severn and Laurence (1974), to Aggarwal (1979), to Siddharthan and Lall (1982), to Grant (1987), to Morck and Yeung (1991), to Hitt et al. (1997), to Riahi-Belkaoui (1998), to Lu and Beamish (2001), to Ruigrok and Wagner (2003) have attempted to empirically trace the relationship – with decidedly mixed results. In the 1980s, some works such as Siddharthan and Lall (1982) found a negative relationship, while others such as Grant (1987) confirmed a positive linear link between performance and DOI. By the late 1980s, scholars were introducing a squared term for DOI, but again finding mixed results. While some studies such as Gongming (1998)


19

found a U-shaped curve, others such as Geringer, Beamish, and Da Costa (1989), Hitt et al. (1997), and Gomes and Ramaswamy (1999) concluded that they had found an inverted-U-shaped relationship. In a 2004 comprehensive survey of a large number of studies, Ruigrok and Wagner (2004) compiled a bibliography that exceeded 180 references, and in their meta-analysis analyzed the results of 62 empirical studies – but without any overall consensus or direction emerging as a general theory from the literature. The lack of consensus in the field, until the year 2000 – a situation that still persists in the mind of some scholars – stemmed from three causes. 1. The absence of a general theory. 2. Considerable variation in the operationalization of both DOI and Performance variables. 3. Contextual variables, in some studies, overwhelming the main link between DOI and Performance.

THE SEARCH FOR A GENERAL THEORY OF INTERNATIONALIZATION VS. PERFORMANCE By the year 2000, it was apparent that the seeming empirical contradictions of past studies might have been the result of not specifying a cubic, or third order, term. Putting a U-shaped and inverted-U-Shaped curve together, produces an S-Curve, which incorporates all stages of international expansion, as shown in Fig. 3.5 Performance

Performance

+


Fig. 3.

Performance

=



The Development of the General 3-Stage Theory.

20

FAROK J. CONTRACTOR

Contractor et al. (2003) enunciated a general ‘‘S-Curve’’ theory (that included all three stages: initial international expansion with negative slope on performance, followed by a second stage of positive effects of further internationalization on performance, and finally a third stage where excessive internationalization again has a negative slope on performance) and found empirical validation in their paper for some sectors. An S-curve combines and reconciles the seemingly contradictory U-shaped and inverted U-shaped results. This was followed a year later by Lu and Beamish (2004), who again partially validated the S-Curve theory for the international expansion of Japanese companies.6 Lacking a general 3-Stage theory, no one prior to 1998 had tested for a cubic term for DOI. Instead, most studies in the 1980s and 1990s only specified the linear and quadratic term for DOI. It was hardly surprising that all empirical studies produced only a U-shaped, or an invertedU-shaped result. A U-shaped statistical fit captures only the first two stages of the 3-Stage theory – as shown in the foregoing Fig. 2. By the same token, an inverted-U-shaped statistical fit captures only the second and third stages of the general 3-stage theory. Thus, depending on their composition, some samples will produce a U-shape fit, while others will produce an invertedU-shape – if the researcher has only specified a first and second order term for DOI. This was the situation until 1998. It is only in recent years, when researchers have begun to test the general theory, by including a (1) linear, (2) quadratic, and (3) cubic specification for DOI, that all three stages are apparent in the results of several studies (e.g., Riahi-Belkaoui, 1998; Contractor et al., 2003; Lu & Beamish, 2004; Thomas & Eden, 2004). Now that we have a complete, or general, 3-stage theory, should we expect S-curve results from all future studies? No, not necessarily. This again depends on the composition of the sample and whether firms in the sample adequately populate all three stages. For instance, in a sample of firms in an older industry, where most firms are already internationally mature (i.e., most have gone beyond the early Stage 1), one would expect most companies to be populating Stages 2 and 3. Some companies might inadvertently, or knowingly, be over-internationalized so that a negative effect on performance is once again seen for such firms in Stage 3. Alternatively, for some firms, or in some sectors, the net costs of initial internationalization in Stage 1 may be low. This could occur, for example, because nations contiguous to them are culturally similar, or they are in a service sector where codified intangible assets can be replicated in foreign locations at low incremental cost. For such firms, benefits even in early Stage 1 may be greater than internationalization costs. In such


21

cases, in statistical testing, the third-order term may not be significant, and the best statistical fit would be an inverted-U shape. This is illustrated in Fig. 2. On the other hand, the statistical fit (despite specifying all three stages in linear, quadratic, and cubic terms for DOI) in other samples might produce a U-shaped curve – as also seen in Fig. 2. For instance, emerging countrybased multinationals will not have had time to sufficiently internationalize. Most firms in such a sample might mainly populate Stages 1 and 2. Few firms in a nation like India can be expected to have over-internationalized, because few Indian companies ventured abroad prior to the mid-1990s. Hence, for an Indian company sample, one would expect few to have reached Stage 3. Alternatively, while all companies have to, or must, pass through the initial internationalization stage (and hazard a negative effect on their overall performance), in the final stage of possible overinternationalization, firms have the option of holding back from excessive internationalization. In practice this is not always possible, as many firms will not know their ‘‘optimum’’ point for international expansion, and may unwittingly cross over into the ‘‘over-internationalized’’ Stage-3 zone. Nevertheless, the sub-population of over-internationalized companies in some samples is likely to be small enough that the last stage is not picked up in the statistical analysis as being significant. For the above reasons, even if a cubic term is introduced, the final result may omit Stage 3 and produce only a statistically significant U-shaped curve. Nonetheless, the general theory now makes available a complete testing for the presence of all the three stages.

LACK OF UNIFORMITY IN THE OPERATIONALIZATION OF DOI AND PERFORMANCE VARIABLES Another reason for the lack of congruence in empirical studies is because the operationalization of measures is very varied. We have, Different Measures for Performance ranging from Return On Assets to Return On Sales in most studies, to Tobin’s Q in a few like Berry (2001). Different Measures for Internationalization from a simple count of numbers of national markets the firm serves to ‘‘Foreign Sales to Total Sales,’’ to ‘‘Foreign Assets to Total Assets,’’ to ‘‘Top Managers’

22

FAROK J. CONTRACTOR

International Experience,’’ to ‘‘Foreign Employees to Total Employees’’ or ‘‘Foreign Subsidiaries to Total Subsidiaries’’ – to more complicated measures such as Herfindahl-like indexes of geographical concentration, as reviewed in Sullivan (1994). Some have constructed composite indexes of DOI from the univariate measures mentioned above. Different Modes of Entry: While most studies measure the sales, assets or other indicators of the FDI affiliates of multinational companies, a few studies include foreign sales achieved by exports, while yet others do not know or reveal the difference.7 A handful of studies also include equity joint ventures. Different Sectors (mostly in manufacturing, but recently also a handful of studies in services such as Contractor et al., 2003). There are a priori reasons to assume that the international expansion path of different sectors will vary. For instance, services – especially knowledge-intensive services requiring relatively little investment in tangible assets – will be markedly different from manufacturing companies. Services can be transmitted at low marginal cost, and unlike manufacturing, many service-sector companies do not need large tangible investments for each expansion. There is bound to be a variation, ceteris paribus, in the difficulty of incremental internationalization and in economic-scale considerations, between services and manufacturing, and from one sub-sector to another, even within manufacturing or services. Different Countries of Origin for the MNEs. Historically, most studies in this field have used data on the internationalization of companies based in the US, which has the world’s largest market and is surrounded by just two immediate neighbors. By contrast, Austria is surrounded by six nations, and is a medium-sized European market. This proximity to other nations, a relatively small home economy in relation to neighbors, and lower cultural or psychic distance (Johanson & Vahlne, 1977) is what ensures a greater likelihood, ceteris paribus, that firms based in a small- or medium-sized European country will be more ‘‘internationalized’’ than American companies. This is seen in the higher Transnationality Index (TNI) for European countries in UNCTAD (2006). Even the largest of European nations have the TNI values greater than the US, as seen in Table 1. Hence, even controlling for sector, the mixing of firms from different home countries does not provide a strictly valid comparison. Firms Based in Emerging Economies: Not only do emerging nations usually comprise a small internal market, but their companies face larger hurdles in international expansion, compared to advanced country-based


23

companies, because of scale, cultural distance, geographical distance, and organizational acumen (Khanna & Palepu, 2000; Yeung, 1999). On the other hand, since emerging country firms have begun to internationalize only recently, few of their firms are likely to have ‘‘over-internationalized’’ beyond an optimum level. Hence the hypothesis that for a Performance vs. DOI study comprising emerging nation multinationals, the statistically-fitted curve would be a U-shape, and the third-order term for DOI would be non-significant. Other Examples of Context-Dependence: Ruigrok and Wagner’s (2004) meta-analysis identifies other contextual variables that would affect firm performance. These include cultural heterogeneity, firm size, mode of foreign entry, and strategy goal. The effect of R&D and advertising intensity as moderating variables was illustrated in Kotabe, Srinivasan, and Aulakh (2002). These could render a sample of firms nonhomogeneous.

CONTEXT-DEPENDENCE AND THE VALIDITY OF A GENERAL THREE-STAGE THEORY IN THE PAST OR FUTURE In general, the accumulated evidence in this field, over more than 30 years, strongly suggests the presence of all three stages of internationalization, (i) Early international expansion producing a negative effect, (ii) Later regional expansion producing a positive effect, and finally, (iii) Some companies inadvertently, or knowingly, having over-internationalized so that a negative effect on performance is seen once again. Hence, we can conclude that an underlying S-shape and three stages exist, but that, In past studies, prior to 1998, no S-curve was observed simply because no empirical analysis had ever specified a third-order term for DOI. In future studies, for reasons detailed above, depending on the context, or characteristics of the firm sample, the best statistical fit may be only a linear, U-shaped, or inverted-U-shaped curve, despite the specification of all three (linear, quadratic, as well as cubic) terms for DOI. This would not invalidate the general theory. Depending on the characteristics of the sample, or context, only two of the three stages may be statistically apparent. Subsets can identify a whole. After all, a U-shape (Stages 1+2) is a subset of the overall S-Curve, as seen in Fig. 2. An inverted-U-curve (Stages 2+3) is also a subset.

24

FAROK J. CONTRACTOR

CONCLUSIONS AND FURTHER RESEARCH The nature of the link between Performance and DOI of the multinational firm is a central question in the field of International Business. It undergirds the field’s main argument that international expansion is beneficial to a firm’s strategy and profits. A plethora of papers, over more than 30 years, has produced seemingly contradictory empirical results. However, when viewed through the lens of the general 3-stage theory, the apparently noncongruent results are seen to be, but different subsets of the theory’s three stages, (i) early internationalization, (ii) regional expansion, and (iii) overexpansion. All three stages have been observed empirically only in recent years, simply because no one, prior to 1998, had tested for all the three. Because of this 30-year history of seemingly incompatible results, the general theory is not yet universally accepted. Even if intuitively reasonable on an a priori basis, reservations about the general theory have been voiced on both theoretical and empirical grounds. Beyond Contractor et al. (2003) and Lu and Beamish (2004), the theory still needs further probing, development, and articulation. We have not yet fully specified the exact nature of the liability of foreignness, or about when, how, and why some companies become over-internationalized (Zaheer & Mosakowski, 1997; Ruigrok, Wagner, & Amman, 2004). The body of empirical evidence in this sub-field is almost entirely comprised of large-sample, cross-sectional studies, using data from secondary sources. What is conspicuously lacking is the studies at the ‘‘micro’’ level, in order to better understand the ‘‘whys’’ of managerial thinking and mind-set that leads to initial and later international growth. By the same token, the field is short on studies (whether at the large-sample or case-level) covering data from the service sector and emerging nations whose firms are becoming increasingly internationalized. Below is a tentative research agenda for areas that are not fully understood: Exploration of the ‘‘micro’’ factors that create the ‘‘liability of foreignness.’’ Zaheer and Mosakowski (1977) presented the liability of foreignness in terms of the additional learning costs of understanding foreign markets and culture, and overcoming the unfamiliar. Kostova and Zaheer (1999) extended this to include institutional differences across nations and the institutional discrimination that non-local companies must suffer. While managers can learn how to do business abroad, institutional differences may persist beyond the initial expansion stage. In general, a


25

further detailed exploration of the costs of doing business abroad, over various stages of international expansion, would add considerably to the cogency of arguments in the field (Eden & Miller, 2004). Why the liability of foreignness is small for some companies but persists in others. The costs of initial expansion for some companies appear to be small, or persist only briefly. This is one possible inference one may draw from inverted-U-shaped results (where statistically speaking, the initial downward effect on performance from early internationalization is not significant). For instance, it may be hypothesized that some service sectors that do not rely on the foreign replication of tangible assets can very quickly reap the benefits of international growth without having to pay significant initial costs. A parallel explanation could be that (contrary to the popular notion that services are ‘‘localized’’ and culture-specific), business process services are actually quite standardized worldwide, for example in Information Technology. For such companies, their liability of foreignness may be low or minimal. Research is needed to probe such details. Will companies based in emerging markets face greater costs of internationalization because of cultural or geographical distance from major markets, and because of the smaller scale of home country markets? On the other hand, the counter-hypothesis is that, because firms are based in a smaller domestic market (be it in India or in Switzerland) the benefits of international growth would be all the more valuable for such companies. We have virtually no studies, thus far, on the performanceinternationalization nexus of companies based in emerging nations – a lacuna all the more glaring because of the incipient internationalization of Chinese, Indian, and East European companies. What determines the Inflexion Points between the three stages of international expansion? Why do some companies over-internationalize? Is it conscious? Is there hysteresis? It is clear from the many cross-sectional inverted-U-shaped results (which have a negative slope on the right hand side of the invertedU) that a minority of firms exhibit excessive internationalization beyond an optimal degree. Do some do so deliberately, for short-term or strategic reasons? If so, what? Or is excessive internationalization inadvertent? Indeed, how many companies are conscious of an optimal DOI as a management or strategy issue? Casual empiricism (based on readings of Annual or 10-K reports) suggests that few of even the largest multinationals venture beyond covering 50 or so nations internally, that is to say, via their own controlled affiliates.

26

FAROK J. CONTRACTOR

Research on the ‘‘limits to internationalization,’’ in terms of a dissection of administrative and coordination costs, is lacking in the International Business literature. Which of the alternative macroeconomic hypothesis for the ‘‘limits to international expansion,’’ is the stronger explanation? That beyond the top 50 countries, the remaining markets are seen as too small, or too risky, or culturally distant, to justify the risks and incremental costs? Or is the limit to international expansion reached because of internal cognitive, information-processing limits? Which of the two, macroeconomic, or behavioral, alternative hypotheses has greater validity? Does this vary by sector, home nation, and company size? More studies on the international path of service-sector companies, firms based in emerging and smaller nations, as well as small-sized firms are needed to redress the bias of US-based, manufacturing sector samples. As internationalization is a process over many years, longitudinal studies are not only critically needed to redress the overwhelming bias in favor of cross-sectional studies, but would also remedy the latter’s methodological limitations. (However, this is a general critique of the relative dearth of longitudinal studies in the field of management, and not particularly directed at the Performance-Internationalization literature).

NOTES 1. One can measure a company’s ‘‘Degree of Internationalization’’ or ‘‘Multinationality’’ based on a variety of indexes. A simple count of the countries in which the firm sells its products or services is the crudest measure. Alternatively, one may count the number of nations in which the firm has Foreign Direct Investment (FDI) affiliates. Other studies measure ratios, such as Foreign to Total Sales, or Foreign to Total Assets. Later in this chapter, I comment on the methodological issue of using different indexes. 2. The speed with which a firm traverses through Stage 1 into Stage 2 is indeed relevant to the shape of the Performance vs. Degree of Internationalization curve. Both theoretically, as well as empirically (e.g., Contractor et al., 2003), companies that make the passage through Stage 1 more quickly suffer less negative effects on performance. For such firms, Stage 1 has a shallower negative slope, or they may avoid Stage 1 altogether. 3. This is not to say that all the benefits of Stage 2 internationalization summarized in the foregoing paragraph will accrue to all firms. But it is a sufficiently large and diverse set of benefits that one or several benefits would accrue to most companies – with the result that most firms would see incremental benefits exceeding incremental costs in Stage 2.


27

4. Data on many such companies are not even picked up by the statistical surveys of governments because their foreign affiliates are considered too small, i.e, foreign sales or assets falling below US $2 million, for example. 5. This idea occurred to the author in a Fall 2000 doctoral seminar, where he was teaching, when he drew the U- and inverted-U-curves on the board, as depicted in Fig. 3. 6. Lu and Beamish (2004) referenced the S-curve idea to their previous article, Lu and Beamish (2001), where the S-curve notion was mentioned in passing at the end of their 2001 article, as a possible way to reconcile past empirical work. Interestingly, neither Lu and Beamish (2004) nor Contractor et al. (2003) cited Riahi-Belkaoui (1998) who actually tested a cubic term in a short article in International Business Review (IBR). Riahi-Belkaoui (1998) does not develop the S-Curve theory but merely outlines the basic notion in a couple of paragraphs. Contractor, Kundu and Hsu were unaware of Riahi-Belkaoui’s work until the middle of 2004, well after their publication in the Journal of International Business Studies in 2003. Presumably, a similar ignorance explained Lu and Beamish’s (2004) failure to cite Riahi-Belkaoui. This failure, however, is revealing. It indicates that IBR is not yet a well-read journal, at least not by North American scholars. Geography still matters. It suggests that the International Business field (and certainly a sub-field such as Internationalization vs. Performance) is scattered over different scholars, from different backgrounds, each plowing their lonely furrows. The fact that Professor Riahi-Belkaoui is in Accounting (which is hardly represented in the Academy of International Business – and not at all in the Academy of Management) is another possible reason for the neglect of his paper. 7. Indeed, in several secondary sources of published data on multinationalcompany operations, the published source is unable to tell whether ‘‘foreign sales’’ were the result of exports or whether they represent the sales of the company’s foreign affiliate in the foreign location. Scholars using such sources have not always bothered to inform readers of this methodological problem.

REFERENCES Aggarwal, R. (1979). Multinationality and stock market valuation: An empirical study of U.S. Markets and companies. Management International Review, 19, 5–12. Bartlett, C., & Ghoshal, S. (1990). Managing across borders: The transnational solution. Boston: Harvard Business School Press. Berry, H. (2001). When does multinationality increase firm value? Evidence from US and Japanese firms, 1974–997. Working paper, University of Pennsylvania. Buckley, P. J., & Casson, M. C. (1976). The future of the multinational enterprise. London: Macmillan. Caves, R. E. (1971). Industrial corporations: The industrial economics of foreign investment. Economica, New series, 38(149), 1–27. Contractor, F. (2002). Strategies for international expansion. In: M. Warner (Ed.), International encyclopedia of business and management (2nd ed., Vol. 7). London: Thomson Learning. Contractor, F. J., Kundu, S. K., & Hsu, C. (2003). A three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34, 5–18.

28

FAROK J. CONTRACTOR

Delios, A., & Henisz, W. (2003). Political hazards, experience, and sequential entry strategies: The international expansion of Japanese firms, 1980–1998. Strategic Management Journal, 24, 1153–1164. Doz, Y., Santos, J., & Williamson, P. (2001). From global to metanational: How companies win in the knowledge economy. Cambridge, MA: Harvard Business School Press. Dunning, J. (Ed.), (2002). Regions, globalization and the knowledge-based economy (New ed.). Oxford: Oxford University Press. Eden, L., & Miller, S. (2004). Opening the black box: Multinationals and the costs of doing business abroad. Working Paper, Texas A&M University. Geringer, J., Beamish, P., & Da Costa, R. (1989). Diversification strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10(2), 109–119. Gomes, L. K., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between multinationality and performance. Journal of International Business Studies, 30(1), 173–188. Gongming, Q. (1998). Determinants of profit performance for the largest US firms 1981–92. Multinational Business Review, 6(2), 44–51. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies. Journal of International Business Studies, 18, 79–89. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management Journal, 40, 767–798. Hymer, S. H. (1976). The international operations of national firms: A study of direct foreign investment. Cambridge, MA: MIT Press. Johanson, J., & Vahlne, J. (1977). The internationalization process of the firm: A model of knowledge development and increasing foreign market commitments. Journal of International Business Studies, 8, 23–32. Khanna, T., & Palepu, K. (2000). Is group affiliation profitable in emerging markets? An analysis of diversified Indian business groups. Journal of Finance, 2, 867–892. Kogut, B. (1985). Designing global strategies: Profiting from operational flexibility. Sloan Management Review, 27, 27–38. Kostova, T., & Zaheer, S. (1999). Organizational legitimacy under conditions of complexity: The case of the multinational enterprise. Academy of Management Review, 24(1), 64–81. Kotabe, M., Srinivasan, S., & Aulakh, P. (2002). Multinationality and firm performance: The moderating role of R&D and marketing capabilities. Journal of International Business Studies, 33(1), 79–97. Lu, J. W., & Beamish, P. W. (2001). The internationalization and performance of SMEs. Strategic Management Journal, 22, 565–586. Lu, J. W., & Beamish, P. W. (2004). International diversification and firm performance: The S-Curve hypothesis. Academy of Management Journal, 47, 598–609. Morck, R., & Yeung, B. (1991). Why investors value multinationality? Journal of Business, 64, 165–187. Osegowitsch, T., & Sammartino, A. (2006). A new perspective on the regionalization debate. Australian Centre for International Business, University of Melbourne, Working Paper no. 11, November. Retrieved from: http://www.ecom.unimelb.edu.au/acib/workpap/ home.html


29

Oviatt, B., & McDougall, P. (1997). Challenges for internationalization process theory: The case of international new ventures. Management International Review, 37(2), 85–99. Peng, M. (2003). Institutional transitions and strategic choices. Academy of Management Review, 28(2), 275–296. Riahi-Belkaoui, A. (1998). The effects of the degree of internationalization on firm performance. International Business Review, 7, 315–321. Rugman, A. (2005). The regional multinationals: MNEs and ‘‘global’’ strategic management. Cambridge, UK: Cambridge University Press. Rugman, A., & Verbeke, A. (2003). Extending the theory of the multinational enterprise: Internalization and strategic management perspectives. Journal of International Business Studies, 34, 125–137. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43, 63–83. Ruigrok, W., & Wagner, H. (2004). Internationalization and firm performance: Meta-analytic review and future research directions. Working Paper, University of St. Gallen. Ruigrok, W., Wagner, H., & Amman, W. (2004). The form of the internationalizationperformance relationship: ‘‘Universal’’ or ‘‘context dependent’’? Paper presented at the AIB/JIBS Workshop, Stockholm, Sweden, July 10. Severn, A. K., & Laurence, M. M. (1974). Direct investment, research intensity, and profitability. Journal of Financial and Quantitative Analysis, 29, 181–190. Shenkar, O. (2001). Cultural distance revisited: Towards a more rigorous conceptualization and measurement of cultural differences. Journal of International Business Studies, 32(3), 519–535. Siddharthan, N. S., & Lall, S. (1982). Recent growth of the largest U.S. multinationals. Oxford Bulletin of Economics and Statistics, 44, 1–13. Sullivan, D. (1994). Measuring the degree of internationalization of a firm. Journal of International Business Studies, 25(2), 325–342. Sunderam, A., & Black, S. (1992). The environment and internal organization of multinational enterprises. Academy of Management Review, 17, 729–743. Thomas, D., & Eden, L. (2004). What is the shape of the multinationality-performance relationship? Multinational Business Review, 12(1), 89–110. UNCTAD (2004). World investment report, 2004. Geneva: United Nations. UNCTAD (2005). World investment report, 2005. Geneva: United Nations. UNCTAD (2006). World investment report, 2006. Geneva: United Nations. United States Central Intelligence Agency (2006). The world factbook, 2006. Washington DC: Central Intelligence Agency. Yeung, H. (1999). The internationalization of ethnic Chinese business firms from Southeast Asia: Strategies, processes and competitive advantage. International Journal of Urban and Regional Research, 23, 11–19. Zaheer, S., & Mosakowski, E. (1997). The dynamics of the liability of foreignness. Strategic Management Journal, 18(6), 439–464.

MULTINATIONALITY AND REGIONAL PERFORMANCE, 2001–2005 Alan M. Rugman and Chang Hoon Oh ABSTRACT The traditional dependent variable in the multinationality and performance literature is the ratio of foreign (F) to total (T) sales, (F/T). This can now be supplemented by a new regional variable, the ratio of regional (R) to total (T) sales, i.e. (R/T). Data are presented on both (F/T) and (R/T) for both sales and assets for a five-year period, 2001–2005. Implications are drawn for future research on multinationality and performance in the light of this regional phenomenon.

INTRODUCTION The relationship between multinationality (M) and performance (P) is a traditional topic in the area of international business. Recently there has been renewed interest, partly due to theoretical advances such as the regionalization thesis of Rugman and Verbeke (2004). This contrasts with the previous assumption of uniform internationalization.


31

32

ALAN M. RUGMAN AND CHANG HOON OH

In this literature, P, as a dependent variable, is broadly determined by the degree of multinationality, M, where M is usually proxied by the ratio of foreign (F) to total (T) sales or assets, i.e. (F/T). There is either a linear, quadratic, or cubic (S-curve) fit, allowing for controls such as size of the firm, industry grouping, and organizational learning effect over time. Recently, this M and P literature has included regional aspects of (F/T) and performance, for example performance has looked at return on foreign assets (ROFA). There is now better and more detailed data on the geographic dispersion of activities. The new accounting standards affecting most of the world’s multinational enterprises (MNEs) now make it possible to adopt both a new dependent variable (for performance) and a new independent variable (for multinationality): 1. Performance can now capture the ROFA, not just the return on total assets (ROTA). 2. Multinationality is now available on a regional basis, i.e. the ratio of regional (R) to total (T) sales – (R/T). This offers better information on the strategic performance of an MNE, in comparison to the traditional metric of the return of foreign to total sales or assets – (F/T).

MULTINATIONALITY AND PERFORMANCE The relationship between M and P has received much attention from diverse disciplines: international business, strategic management, finance, and economics. In general, (F/TS), where (TS) is total sales, and the number of foreign affiliates (NOFA) has been used for the M variable. The P variables have been: ROTA; return on total sales (ROTS); sales growth; market share; Tobin’s q; and abnormal return (AR). Each measure has its own advantages. McWilliams and Siegel (1997) argue that financial performance is the best measure because market performance (sales growth, market share) and accounting performance (ROTS, ROTA) do not reflect a firm’s expected future profits. Also M includes the unaccountable future benefits generated from non-location bound firm-specific advantage (FSA). We summarize some representative studies in the existing literature that have investigated the relationship between M and P using financial performance as a dependent variable in Table 1. We exclude studies using the event study approach because AR from an event study captures a short-term stock market response rather than a long-term equilibrium. Existing findings only weakly support either the positive or the negative relationship between M and P.

Multinationality and Regional Performance, 2001–2005

Table 1. Authors

33

Selected Studies on Multinationality and Financial Performance. Sample

Dependent Variable

Errunza and Senbet (1981) Errunza and Senbet (1984)

US MNE

Excess valuea

F/TS

(+) F/TSb

402 US MNE

Excess valuea

F/TS; ENT; NOFA

Kim and Lyn (1986)

154 US LMNE 1,644 US MNE 152 US LMNE

Excess valuea

F/TS; NOFA

(+) F/TSb; (+) ENT; ( ) NOFAb, (+) F/TSb

Tobin’s q

NOFA; NOFC

Risk Adjusted Return Tobin’s q

GMD; GRD; UD

Morck and Yeung (1991) Kim et al. (1993)

Christophe (1997) Denis et al. (2002) Christophe and Lee (2005)

500 US LMNE 7,520 US MNE 100 US LMNE

Excess valuec Tobin’s q

Independent Variables

F/TS Dummy of increasing F/TS F/TA; F/TS; OS; PD

Results

(+) NOFA; NOFC (+) GMDb; GRDb; (+) UD ( ) F/TSb ( ) F/TSb ( ) F/TAb; PD; (+) F/TS; OS

Note: LMNE, large multinational enterprise; ENT, entropy measure of a firm’s geographical diversification; F/TA, average foreign assets as a percentage of total assets; F/TS, average foreign sales as a percentage of total sales; GMD, global market diversification; GRD, globalrelated diversification; NOFA, number of foreign affiliates; NOFC, number of foreign countries, into which affiliates enter; OS, portion of overseas subsidiaries to total subsidiaries; PD, psychic dispersion of international operations; UD, unrelated diversification. a Excess value (as a proxy for Tobin’s q) is defined as the difference between market value of common equity and net worth normalized by annual sales. b Statistical significance at the o10% level. c Excess value is measured as the log of the ratio of the firm’s actual value to the imputed value of its industrial segments as stand-alone domestic firms.

(F/TS) is the most popular measure of M: six out of these eight particular existing studies have commonly used (F/TS) as one of the proxies for M. (F/TS) is also the most frequently used proxy for M in the literature using accounting performance as well: see Grant (1987), Tallman and Li (1996), and Gomes and Ramaswamy (1999). Except for Christophe and Lee (2005), (F/TS) is usually statistically significant, but the results are inconsistent in the literature. Errunza and

34


Senbet (1981, 1984) and Kim and Lyn (1986) found a significant and positive effect of (F/TS) on P. In contrast, Christophe (1997) and Denis, Denis, and Yost (2002) found a significant and negative effect of (F/TS) on P. It is important to note that the large sample studies (Christophe, 1997; Denis et al., 2002) found negative relations, while the small sample studies (Errunza & Senbet, 1984; Kim & Lyn, 1986) estimated a positive relationship. These mixed results also occur in the studies using accounting and market performances as P variables. Some existing studies include a non-linear relationship between M and P, such as: Inverted U-shape; U-shape; and S-shape relations. This type of literature uses accounting and market performance as a dependent variable, and the results are not conclusive: see Contractor, Kundu, and Hsu (2003) for a literature review on the link between P and M, as well as papers in this volume, in Part A. Recently, some studies have extended the M and P literature to the regional context. Delios and Beamish (2005) use Japanese MNEs data to show that global MNEs do better than home region oriented MNEs. ROTS and Tobin’s q of global MNEs are significantly higher than those of home region oriented MNEs. Rugman, Yip, and Jayaratne (2007) have tested ROFA for P as well as R/T for M and have found that foreign operation increases foreign performance. Further tests are reported in Part C of this volume.

THE REGIONAL DIMENSION OF MULTINATIONALITY AND PERFORMANCE Recent empirical research has established that MNEs operate regionally rather than globally. It was shown by Rugman and Verbeke (2004) that only nine of the world’s 500 largest firms operate globally, i.e. in all three regions of the broad triad of North America, Europe, and Asia Pacific. In contrast, 320 of the 380 firms providing data for year 2001 on the geographic scope of their sales average 80% of such sales in their home region. In Rugman (2005) some 60 cases were examined to establish the robust nature of this regional effect. It was also demonstrated that whenever data on assets were provided by firms these upstream production data also revealed a regional rather than a global effect. The implications are that the MNEs are unlikely to source through a global supply chain, but rather through regional clusters. However, robust testing of the asset data remains to be undertaken.


35

In this chapter we shall now present data on both sales and assets. These data are presented for the five-year time period 2001–2005. The purpose of these data will be to demonstrate that the regional effect is applicable over time and that there appears to be no trend toward globalization. Rather, these data indicate that there is a longitudinal argument that regionalization is now a stable phenomenon. Table 2 reports data on intra-regional sales and assets for the period 2001–2005. This is for a set of the world’s largest 500 firms. Among 500 firms, geographic sales and assets data are available for 386 firms during this period. The data are compiled from the annual reports of these publicly traded companies. These annual reports are now available on the Internet under each company’s name. This table updates and supplements the data for 2001 reported in Rugman and Verbeke (2004) and in Rugman (2005). In Table 2 data are reported for the ratio of regional to total sales (R/TS) and also for the ratio of regional to total assets (R/TA). In addition the table reports the conventional measure of multinationality in previous empirical research. This is (F/TS), i.e. the ratio of foreign (F) to total (T) sales. The table also reports the ratio of foreign to total assets (F/TA). The table reports that the average (R/TS) for the world’s largest firms is 75.7%. There is almost no variation over time. Next the table reports that the average (R/TA) is 76.7%. Again, there is very little variation over time. Table 2.

Foreign and Intra-Regional Sales and Assets of Large Firms, 2001–2005. Sales

Assets

F/TS

R/TS

F/TA

R/TA

2001 2002 2003 2004 2005

33.6 34.9 35.5 35.8 36.4

75.6 75.8 75.8 75.2 75.2

31.2 32.1 32.7 33.2 33.1

77.2 76.9 76.4 76.4 76.5

Average

35.2

75.7

32.5

76.7

Note: Data are for 386 firms listed in 2002 Fortune Global 500. Among 500 largest firms, 386 firms report regional sales and/or assets data. We only use data for firms that reported both intra-regional sales (assets) and foreign sales (assets) in this table. Total numbers of observations are 1,603 and 1,359 for sales and assets, respectively. Numbers of observations are, typically, getting smaller every year because several firms are merged, acquired, or bankrupt during the observation period. Sources: Annual reports of each firm for 2001–2005.

36


Foreign-to-Total %

SALES % 100.0 ROWS

(F/TS) = 35.2

75.7 RORS 64.8

(R/TS) = 75.7 HOMES

Fig. 1. The Distinction between (F/T), (R/T), and Home Sales, 2001–2005. Note: HOMES, sales in the home country; RORS, foreign sales in the rest of the home region; ROWS, foreign sales in the rest of the world; (F/TS), value of foreign (F) to total (T) sales; (R/TS), intra-regional sales (HOMES plus RORS). Total observations are 1,603 for 386 firms listed in 2002 Fortune Global 500.

These data suggest that the world’s largest firms are slightly more regional on assets than on sales. Table 2 also reports that the average (F/T) for sales is 35.2% while the average (F/T) for assets is 32.5%. These data are now further discussed by means of Figs. 1 and 2. Building upon the methodology discussed in Rugman and Verbeke (2007), Fig. 1 shows graphically the distinction between home sales (HOMES) and foreign sales (F/TS). Foreign sales are broken down into sales in the rest of the home region (RORS) and in the rest of the world (ROWS). As shown by Rugman and Verbeke (2007) the conventional measure of multinationality is (F/TS). For the world’s largest firms (F/TS) averages 35.2%. However, a large component of these foreign sales are sales in the RORS. In Fig. 1 this amounts to 10.9% of total sales. Thus the intraregional sales (R/TS) average is 75.7%. We are now in possession of two important statistics for the dependent variable in research on multinationality and performance. We either use the (F/TS) variable or the new (R/TS) variable. From the viewpoint of strategy it has been argued that (R/TS) contains important attributes excluded from the pure multinationality variable. This regional variable (R/TS) has the advantage of including sales in the home country and in the RORS. In explaining the performance of the firm in terms of strategy home country


Assets %

37

Foreign-to-Total %

100.0 ROWA (F/TA) = 32.5

76.7 RORA 67.5

HOMEA

(R/TA) = 76.7

Fig. 2. The Distinction between (F/T), (R/T), and Home Assets, 2001–2005. Note: HOMEA, assets in the home country; RORA, foreign assets in the rest of the home region; ROWA, foreign assets in the rest of the world; (F/TA), value of foreign (F) to total (T) assets; (R/TA), intra-regional assets (HOMEA plus RORA). Total observations are 1,359 for 386 firms listed in 2002 Fortune Global 500.

sales are obviously of immense importance. While we wish to examine the impact of foreign sales on performance it is also important to distinguish between foreign sales in the RORS in comparison with true international sales (ROWS). Fig. 2 repeats the above analysis for assets instead of sales. The definitions are comparable and are explained in the note to Fig. 2. Briefly, Fig. 2 complements Fig. 1 in that the data on home and foreign assets are much the same as the data on home and foreign sales. The average home assets at 67.5% are slightly higher than the average home sales of 64.8%. The regional assets (R/TA) at 76.7% are slightly higher than the regional sales of 75.7%. Finally, the firms’ foreign assets (F/TA) of 32.5% are slightly lower than their average foreign sales of 35.2%. There is an interesting variation by triad region. In Table 3, the (F/TS) for Europe is 55.5% and for assets it is 52.1%. This partly reflects the historical lag in statistical data collection whereby the 27 member states of the EU still record trade, foreign investment, and foreign sales data across country borders. In practice, the EU is now an integrated internal market with common political institutions, a uniform judicial system, and one with a common currency for most of the member states. Thus scholars need to be careful in using data on (F/T) since the international aspect of both sales

38


Table 3. Foreign and Intra-Regional Sales and Assets of Large Firm by Regional Origin, 2001–2005. Sales

Assets

F/TS

R/TS

F/TA

R/TA

N. America Europe Asia

26.9 55.5 27.4

78.0 70.8 77.2

27.0 52.1 21.1

76.5 72.9 81.6

Average

35.2

75.7

32.5

76.7

Note: See notes in Table 2.

and assets is exaggerated for Europe. In contrast, the regional variable is stable across the three regions of the triad. The distortionary effect of Europe does not appear in the average of (R/TS) of 75.7% and (R/TA) of 76.7%. It can be concluded that the (R/T) variable is more stable and, perhaps, more reliable than the traditional (F/T) variable. Both variables are preferable to the ‘‘scope measure’’ which simply counts the number of foreign countries in which an MNE has subsidiaries. This gives only a vague indication of geographic sales (or asset) dispersion and it is probably very misleading as it misses the magnitude of sales (or assets) across countries and triad regions. For example, for a UK firm to have a subsidiary in the United States is much more significant than for it to have 12 subsidiaries in the new member states of the EU. Yet the scope measure would count the UK firm as 12 times more internationalized in Europe than in North America.

DIFFERENCES ACROSS INDUSTRIES In the previous section, we have shown that the M variables are very stable across the last five years. McGahan and Porter (1997), Roquebert, Philips, and Westfall (1996), and Rumelt (1991) also found little annual variation in performance. The literature has found that industry differences explain about 10% of the variation in performance. This literature does not analyze directly the relationship between P and M. Instead, the corporate and business segment effects in the literature probably capture the impact of multinationality. Now we will discuss how M is different across industries and incorporate regional data.


39

Table 4. Foreign and Intra-Regional Sales and Assets of Large Firms by Industry. Sales

Assets

F/TS

R/TS

F/TA

R/TA

A. Manufacturing Industries Aerospace and defense Chemicals and pharmaceuticals Computer, office, and electronics Construction, building materials, and glass Energy, petroleum, and refining Food, drug, and tobacco Motor vehicle and parts Natural resources manufacturing Other manufacturing

47.3 37.6 56.2 53.2 35.5 27.8 39.8 56.6 61.7 49.1

64.6 68.3 53.9 58.0 80.2 77.3 69.4 56.5 71.1 62.0

39.3 41.9 45.6 37.2 25.5 35.5 43.9 44.2 47.4 36.7

68.9 73.9 63.7 69.5 85.0 73.2 62.3 63.3 64.2 68.1

B. Service Industries Bank Entertainment, printing, and publishing Merchandiser Telecommunications and utilities Transportation services Other financial Services Other services

26.0 29.5 39.3 18.1 20.2 23.4 37.8 28.5

84.2 83.2 73.9 89.5 89.3 81.4 79.7 77.9

26.2 35.8 20.3 17.6 23.8 14.5 35.9 32.4

83.8 81.4 89.2 90.8 87.3 88.2 72.6 73.7

Note: See notes in Table 2.

We divide our sample into nine manufacturing industries and seven service industries, as identified in the appendix. Table 4 contains average values for (F/T) and (R/T) of each industry. We find that industry differences are larger than differences across years and triad regions. (F/TS) and (R/TS) of manufacturing firms are 47% and 65% on average and those of service firms are 26% and 84%. Likewise, (F/TA) and (R/TA) of manufacturing firms are 39% and 69% and those of service firms are 26% and 84%. In general, manufacturing firms focus more on their home market and home region compared to those in the service industries. The manufacturing industry has a gap between (R/TS) (or F/TS) and (R/TA) (or F/TA), but the service industry has almost identical values for such variables. Within-industry differences are also very large for manufacturing industries compared to the differences for service industries. Among all manufacturing industries, the chemicals and pharmaceutical industry is the most internationalized industry – it generates 54% of sales and has 64% of

40


assets in foreign regions on average – while the construction, building materials, and glass industry is the most regionalized. In the service industry, the merchandising firms are the most regionalized, they have about 90% of sales and assets in the home region. We think that sales knowledge and logistics might be the FSAs that are hard to be delivering into and exploiting in foreign regions. In contrast, natural resources-oriented industries, i.e. natural resources manufacturing and food, drug, and tobacco industry, and labor-intensive industries, i.e. the chemical and pharmaceutical industry and the motor vehicle and parts industry, are typically more internationalized than others.

CONCLUSION Our main conclusion is that the regional variable (R/T) is an important new measure that supplements the traditional (F/T) measure; indeed (R/T) may be the superior measure. We find evidence that MNEs perform on an intra-regional basis, on the basis of both sales and assets. We find strong intra-regional effects across all industry sectors. We find little support for a trend toward globalization, or the need for a global strategy for MNEs. We conclude that research on multinationality and performance needs to take into account the new metrics available on regional sales and assets, and the ROFA.

REFERENCES Christophe, S. E. (1997). Hysteresis, and the value of the U.S. multinational corporation. Journal of Business, 70(3), 435–462. Christophe, S. E., & Lee, H. (2005). What matters about internationalization: A market based assessment. Journal of Business Research, 58, 636–643. Contractor, F. J., Kundu, S. K., & Hsu, C.-C. (2003). A three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 24, 5–18. Delios, A., & Beamish, P. W. (2005). Regional and global strategies of Japanese firms. Management International Review, 45(Special Issue 1), 19–36. Denis, D. J., Denis, D. K., & Yost, K. (2002). Global diversification, industrial diversification, and firm value. Journal of Finance, 57(5), 1951–1979. Errunza, V., & Senbet, L. (1981). The effect of international operations on the market value of the firm: Theory and evidence. Journal of Finance, 36(2), 401–417. Errunza, V., & Senbet, L. (1984). International corporate diversification, market valuation and size-adjusted evidence. Journal of Finance, 39(3), 727–743.


41

Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of relationship between multinationality and performance. Journal of International Business Studies, 30(1), 173–188. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies. Journal of International Business Studies, 22, 249–263. Kim, W. S., & Lyn, E. O. (1986). Excess market value, the multinational corporation, and Tobin’s q-ratio. Journal of International Business Studies, 17(1), 119–125. Kim, W. C., Hwang, P., & Burgers, W. P. (1993). Multinationals’ diversification and the riskreturn trade-off. Strategic Management Journal, 14(4), 257–286. McGahan, A. M., & Porter, M. E. (1997). How much does industry matter, really? Strategic Management Journal, 18(1), 15–30. McWilliams, A., & Siegel, D. (1997). Event studies in management research: Theoretical and empirical issues. Academy of Management Journal, 40(3), 626–657. Morck, R., & Yeung, B. (1991). Why investor value multinationality? Journal of Business, 64(2), 165–187. Roquebert, J. A., Phillips, R. L., & Westfall, P. A. (1996). Markets vs. management: What ‘drives’ profitability? Strategic Management Journal, 17(8), 653–664. Rugman, A. M. (2005). The regional multinationals: MNEs and ‘global’ strategic management. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Rugman, A. M., & Verbeke, A. (2007). Liabilities of regional foreignness and the use of firmlevel versus country-level data: A response to Dunning et al. Journal of International Business Studies, 38(1), 200–205. Rugman, A. M., Yip, G., & Jayaratne, S. (2007). A note on return on foreign assets and foreign presence for UK multinationals. British Journal of Management, 18. Rumelt, R. P. (1991). How much does industry matter? Strategic Management Journal, 12(3), 167–185. Tallman, S., & Li, J. (1996). Effects of international diversity and product diversity on the performance of multinational firms. Academy of Management Journal, 39, 179–196.

42


APPENDIX. INDUSTRY CLASSIFICATION This Chapter Aerospace and Defense Bank Chemicals and Pharmaceuticals Computer, Office, and Electronics

Construction, Building Materials, and Glass Energy, Petroleum, and Refining Entertainment, Printing, and Publishing Food, Drug, and Tobacco

Merchandisers

Motor Vehicle and Parts Natural Resources Manufacturing Other Financial Services

Fortune Global 500 Aerospace and Defense Banks: Commercial and Savings Chemicals Pharmaceuticals Computers, Office Equipment Electronics, Electrical Equipment Network and Other Communications Equipment Semiconductors and Other Electronic Components Building Materials, Glass Engineering, Construction Metals Energy Petroleum Refining Entertainment Hotels, Casinos, Resorts Beverages Food Consumer Products Food Production Food Services Tobacco Food & Drug Stores General Merchandisers Specialty Retailers Wholesalers: Electronics and Office Equipment Wholesalers: Health Care Wholesalers: Other Motor Vehicles & Parts Forest & Paper Products Mining, Crude-Oil Production Diversified Financials Insurance: Life, Health (mutual) Insurance: Life, Health (stock) Insurance: P & C (mutual) Insurance: P & C (stock) Securities


43

APPENDIX. (Continued ) Other Manufacturing

Other Services

Telecommunication and Utility Transportation Services

Apparel Household and Personal Products Industrial & Farm Equipment Miscellaneous Scientific, Photo, Control Equipment Computer Services and Software Health Care: Insurance & Managed Care Health Care: Other Homebuilders Oil and Gas Equipment, Services Temporary Help Trading Telecommunications Utilities Airlines Mail, Package, Freight Delivery Railroads Shipping

Note: Aetna is the Health Care: Insurance & Managed Care industry in the Fortune Global 500, and it is classified in the Other Service industry. Alcan and Corus Group are in the Metals industry in Fortune Global 500, and it is classified in the Natural Resources Manufacturing industry.

EXPLORING TRENDS IN REGIONALISATION Thomas Osegowitsch and Andre´ Sammartino ABSTRACT In this chapter, we revisit the empirical findings of Rugman and coauthors concerning the overwhelming home-regionalisation among the world’s largest firms. Using a longitudinal research design and continuous measures of internationalisation, we observe a number of secular trends. Among other, we find that sales growth beyond the home region is faster than sales growth within the home region. We use our empirical results to critique and augment existing regionalisation theory. In particular, we raise doubts about the sharp distinction in the literature between expansion in the home region and expansion in host regions.

INTRODUCTION In an earlier publication, the authors of this chapter explored Alan Rugman and coauthors’ regionalisation research (Osegowitsch & Sammartino, 2007). Specifically, we tested the robustness of their classification system, which, based on firms’ sales profile across the triad regions, allocates them into the three principal categories of home-regional, bi-regional and global (e.g. Rugman & Verbeke, 2004; Rugman & Collinson, 2005; Rugman, 2005).


45

46

THOMAS OSEGOWITSCH AND ANDRE´ SAMMARTINO

We concluded that their key finding of overwhelming home-regionality is sensitive to the defined thresholds underpinning the various categories. We demonstrated that manipulating classification thresholds based on theoretical and empirical arguments leads to changes in results. The alternative systems of classification we explored revealed a significant share of bi-regional companies. Global companies remained rare, although their numbers were also up, substantially compared with Rugman’s original classification scheme. Notwithstanding the discrepancies between our own results and those of Rugman and coauthors, their key empirical insight – that sales of many Fortune Global 500 firms are concentrated in their home region1 – stands and warrants explanation. The IB research community needs to urgently confront the question why, in an age of purported globalisation, many of the world’s largest firms appear to have barely ventured beyond the confines of their home region. One means to enhance our theoretical understanding is an awareness of trends over time. In this chapter, we refine the preliminary longitudinal analysis presented in Osegowitsch and Sammartino (2007). We begin with a summary of the home-regionalisation view as put forward by Rugman and coauthors. We then introduce a data set that captures the extent of internationalisation for 159 of the world’s largest firms between 1991 and 2001. Analysis shows that the fastest growing portion of their sales occurs outside the home region. In light of this finding, we challenge some theoretical arguments underpinning the regionalisation thesis. We also claim that, in line with the theoretical arguments discussed, the ratio of nondomestic home-region sales (ROR/T) is an under-utilised measure in this type of research.

EXTANT THEORY Rugman and coauthors have begun to address the main research question that arises from their empirical work in a series of articles and books (e.g. Rugman & Verbeke, 2004, 2007; Rugman, 2005). In short, they explain their finding of overwhelming home regionality on the basis of severe limits to the transferability and acceptability of firms’ firm-specific advantages (FSAs) beyond the home region (Rugman & Verbeke, 2004, p. 6). These limits to transferability and acceptability – which apply irrespective of whether FSAs are embodied in exports, transferred to

Exploring Trends in Regionalisation

47

licensees, or transferred to subsidiaries – are captured in the concept of (home-)region boundedness of FSAs (Rugman & Verbeke, 2004, p. 13).2 In a subsequent paper, Rugman and Verbeke (2007) elaborate their argument by introducing the related concepts of liability of intra-regional expansion and liability of inter-regional expansion. The lack of FSA leveraging beyond the home region is explained by the significant cost differential between the two. [T]he liability of intra-regional expansion appears to be much lower than the liability of inter-regional expansion: the additional costs of doing business abroad are often much higher when venturing into other regions of the world than when expanding intraregionally, in the home triad region. (Rugman & Verbeke, 2007, p. 201, original emphasis)

The additional costs of doing business in host regions are principally those associated with the development of complementary FSAs that are typically necessary to successfully deploy the firm’s main FSAs. These complementary FSAs are required to make the main, imperfectly nonlocation-bound FSAs ‘saleable’ in host regions. As a result, the MNE’s operations in the host regions function on the basis of an overall FSA bundle that may differ significantly from the one deployed in the home region. The difference between both reflects the liability of inter-regional foreignness. (Rugman & Verbeke, 2007, p. 204)

While trends over time are not investigated (although see Rugman & Oh, 2007), Rugman (2005, pp. 2, 63) argues that home regionality has been a defining characteristic of large firms and will remain so into the foreseeable future. The cost penalty associated with inter-regional expansion is persistent in view of concomitant efforts at intra-regional integration and inter-regional protectionism (barriers to regional entry, trade wars, etc.). Rugman concludes that the end result is the persistence of MNEs that will continue to earn 80% or more of their income in their home triad region. (Rugman, 2005, p. 63)

RESEARCH DESIGN The home-regionalisation thesis may be subjected to testing in a strictly cross-sectional format (e.g. Rugman, 2005; Goerzen & Asmussen, 2007). Yet the thesis, at its core, is dynamic. It argues that the expansion path for

48


MNEs will be primarily within their home regions. As a result, we can also assess the regionalisation thesis by studying developments over time. Surprisingly, the large volume of empirical work that followed in the wake of Rugman’s (2000) initial book is almost entirely cross-sectional. An exception is our preliminary longitudinal analysis (Osegowitsch & Sammartino, 2007; but see also Rugman & Oh, 2007). Utilising Rugman’s broad categories (home-regional, bi-regional and global companies) based on defined thresholds, we identified a trend towards greater internationalisation beyond the home region. While individual results varied depending on which particular system of thresholds was employed, significant increases were observed for the bi-regional and global categories between 1991 and 2001. The growth in the number of companies with a significant sales presence in one other triad region (bi-regional category) or two other triad regions (global category) was at the expense of firms in the home-regional category. These initial findings cast some doubt on Rugman’s pronouncements on change over time. In this chapter, we expand upon our earlier investigation and present more detailed and nuanced evidence of developments over time. Importantly, we do so by moving away from the use of broad classification categories and instead employ continuous measures. This choice is informed by two arguments. First, as acknowledged in our original work, in the context of this research there is no one set of classification thresholds that can lay claim to be the ‘true’ set; an element of judgement pervades all of them. Second, the use of thresholds in the construction of broad classification categories inevitably results in a loss of data and such classification systems may be too coarse-grained to pick up on important nuances. In response to such concerns, we revisit the data using continuous measures of geographic spread. We agree with Rugman and coauthors that sales represents an accurate metric of the successful deployment of FSAs and consequently rely on five related measures: (i) Foreign sales as a share of total sales (F/T). (ii) Home-regional sales as a share of total sales (R/T). (iii) The share of ‘rest of region’ sales (ROR/T), i.e. home-regional sales ex domestic sales. (iv) The share of sales in the ‘rest of the world’ (ROW/T), i.e. sales occurring outside the home region, which can be simply calculated as 100 percent-R/T. (v) The share of sales in the two relevant host triad regions: North America (NA/T), Europe (E/T) and the Asia-Pacific (AP/T).


49

Initially, these measures allow us to gauge the extent to which firms have grown their international sales over time. In addition, they allow us to compare the rates of home-regional expansion vis-a`-vis the rates of interregional expansion. Finally, they allow us to detect specific internationalisation patterns across firms from each of the three triad regions. We apply the above measures to a subsample of firms for which international sales data was available for the decade prior to Rugman’s 2001 snapshot. We started with his original sample (Rugman, 2005), which comprised 365 firms from the Fortune Global 500. These 365 were those for which sufficient information was available on the geographic spread of their 2001 sales. Using Datastream, we attempted to access the relevant 1991 and 1996 data for each of the 365 entities. We were able to extract a dataset of 159 firms for which we could ascertain the extent of internationalisation between 1991 and 2001. The limited availability of 1991 (and 1996) data reflects several factors, including the less-stringent reporting requirements in the early 1990s and the impact of mergers and acquisitions over time. All 159 firms in our sample provided sufficient information to calculate F/T ratios for 1991, 1996 and 2001. Data on home-region sales (R/T) was consistently available for the sampled firms from North America and Europe, but there was a notable lack of 1991 (and to a lesser extent 1996) R/T data for firms from the Asia-Pacific. This reflects the historical inclination of Japanese firms – which constitute the overwhelming majority of Asia-Pacific firms – to merely report ‘‘domestic sales’’ and ‘‘foreign sales’’. Due to such data shortcomings, we confine our discussions of trends in R/T (as well as ROR/T and ROW/T) to North American and European firms. Further disaggregation of sales in each of the three triad regions – North America (NA/T), Europe (E/T) and the Asia-Pacific (AP/T) – are also hampered by less than comprehensive data. Accounting standards provide firms with considerable discretion in reporting geographic breakdowns, which, in a significant number of cases, prevented us from obtaining the relevant ratios.3 Our findings concerning these ratios must be interpreted with the data constraints in mind. Details of missing data are disclosed in the relevant table notes. Our sample contains 90 firms from North America (56.6 percent), 36 from Europe (22.2 percent) and 33 from the Asia-Pacific (20.8 percent). In comparison, the distribution of workable firms in Rugman’s 2001 sample was 48.8 percent from North America, 31.3 percent from Europe and 19.7 percent from the Asia-Pacific. Our sample contains a markedly larger ratio of manufacturing companies (89) vs. service companies (70) than Rugman’s sample. Rugman and Oh’s (2007) results show that manufacturing


50

companies tend to report higher F/T and lower R/T ratios than service companies. The corresponding ratios for our sample, however, are only immaterially different from Rugman’s. The mean share of foreign sales (F/T) in 2001 for our sample was 33.9 percent, which is very similar to the F/T for Rugman’s larger set (35.2 percent, see Table 2 in Rugman & Oh, 2007). The mean share of home-region sales (R/T) for our full sample was 75.6 percent in 2001, which is virtually identical with that for Rugman’s sample (75.7 percent, see Table 2 in Rugman & Oh, 2007).

RESULTS The 1990s saw a marked increase in the sampled firms’ internationalisation. The mean share of foreign sales for the 159 firms in our full sample increased from 22.7 percent in 1991 to 33.9 percent in 2001 (Table 1). This represents an approximate 50 percent surge in the importance of foreign sales for these large and growing corporations. Read differently, on average, the rate of growth in foreign sales outstripped growth in domestic sales by almost 2:1.4 Table 1 also provides a more disaggregated picture of this trend. It shows that the number of firms in our sample with zero overseas sales dropped from 28.3 percent in 1991 to only 15.7 percent in 2001. In turn, there was a commensurate leap in the portion of firms with more than 50 percent of their sales outside their home country, from 13.8 percent in 1991 to 27.0 percent in 2001. Clearly, foreign sales became more important to this subset of the world’s largest firms during the 1990s. Fewer could afford to sell their products and services solely at home, and for more than a quarter of them the majority of revenues came from outside their home country by 2001. Table 1.

Foreign Sales Shares: Means and Breakdowns – All Firms (in Percent). Foreign Sales Share (F/T)

F/T Means

1991 1996 2001

22.7 27.3 33.9

F/T Brackets 0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

28.3 25.8 15.7

12.0 10.0 11.1

13.8 12.0 12.6

12.6 11.3 8.1

10.0 9.5 18.3

9.5 10.0 8.2

13.8 21.4 27.0

Note: Sample size is 159 firms.


Table 2.

1991 1996 2001

51

Levels of Internationalisation – North American and European Firms (Mean Levels, in Percent). Foreign Sales (F/T)

Home Region Sales (R/T)

Rest of Region Sales (ROR/T)

Rest of World Sales (ROW/T)

24.4 29.3 34.5

84.2 80.4 76.2

8.5 9.7 10.6

15.8 19.6 23.8

Note: Sample size is 126 firms.

We now turn to a more detailed analysis of the documented expansion in the share of foreign sales. Due to the lack of complete data, we dropped Asia-Pacific firms from this part of the investigation and used only North American and European firms.5 Table 2 confirms the dominant share of home-region sales (R/T) for these firms, in line with Rugman and coauthors’ findings. Table 2 also shows, however, that the mean share of home-region sales did not remain static but fell from 84.2 percent to 76.2 percent between 1991 and 2001. The subsequent column reveals that the share of sales in the ‘rest of region’ (ROR/T), i.e. home-regional sales net of domestic sales, grew from 8.5 percent to 10.6 percent over the same period. This makes clear that the decreasing R/T ratio is entirely due to the diminishing share of domestic sales. It also highlights the need for a home-region sales metric net of domestic sales (ROR/T) to complement R/T. Recall that Rugman and coauthors’ main argument is that extant FSAs in the home country are more readily leveraged into other home-region countries than into countries in other regions. We believe it is more accurate to test this proposition – and the corresponding differential between intra-regional and inter-regional liabilities of foreignness – by stripping domestic sales from home-region sales to obtain ROR/T. The fall in R/T for our European and North American subsample was, of course, mirrored by an increase in the ‘rest of world’ sales share (ROW/T=1 R/T) from 15.8 percent to 23.8 percent (Table 2). This is difficult to reconcile with Rugman and coauthors’ argument that the preferred path to an increased international presence is to expand intraregionally. Firms from outside the region are seen to be at a competitive disadvantage since deployment of their FSAs is hampered by an interregional cost penalty. Yet the data suggests that during the 1990s, a growing portion of large firms were able to overcome such presumed impediments and ventured beyond the home region in search of markets for their products and services.


52

ROW/T, 15.8% ROR/T, 8.5%

ROW/T, 19.6% ROR/T, 9.7%

D/T, 75.6%

D/T, 70.7%

1991

1996

Fig. 1.

ROW/T, 23.8% ROR/T, 10.6%

D/T, 65.5%

2001

Distribution of Firm Sales – North American and European Firms.

It is also instructive to note that, by 2001, mean ROW/T stood at 23.8 percent compared to 10.6 percent for mean ROR/T (see Fig. 1). This further questions claims regarding the comparative difficulty in deploying FSAs outside the home region. It also reinforces the case for our adjusted measure of home-region sales share, ROR/T. The increase in inter-regional sales by 2001 was the result of impressive growth rates during the 1990s. Let us take the stylised average firm, whose total sales grew in line with Fortune 500 ranked companies (10.4 percent per annum) and whose distribution of sales (F/T, R/T, ROR/T and ROW/T) followed the trends captured in Table 2. For that stylised firm, the rate of compound annual growth in inter-regional sales would have been 15.1 percent. By comparison, the annual rates of growth for ROR/T and R/T would have been 12.9 percent and 9.4 percent, respectively, while domestic sales would have increased at an even slower rate of 8.9 percent. On the whole, these figures point to firms grasping more rather than fewer opportunities outside their home region. A further breakdown of non-home-region sales shares among North American and European firms is given in Table 3. As shown, the portion of firms with no sales beyond the home region dropped from 38.1 percent in 1991 to 20.6 percent in 2001. At the other end of the spectrum, the portion of firms that earned more than 50 percent of sales beyond the home region almost doubled over the period of investigation. In 2001, this bracket contained corporations such as Motorola, ING Group and McDonalds. More broadly, by 2001, more than half of the firms (the 51.6% of companies populating brackets 4, 5, 6 and 7) had achieved ROW/T greater than


Table 3.

53

Breakdown of Rest of World Sales Shares – North American and European Firms (in Percent). Rest of World Sales (ROW/T) Brackets

1991 1996 2001

0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

38.1 31.0 20.6

11.9 11.9 15.9

19.0 14.2 11.9

8.2 14.3 13.7

11.9 9.6 19.1

4.8 8.7 6.3

7.1 10.3 13.5

20 percent. This would seem to defy the view of the home-region boundary as a significant barrier to further internationalisation. In the following section, we are testing whether the insights and trends uncovered thus far hold for firms from each of the triad regions. Analysing separate subsamples of North American, European and Asia-Pacific firms also allows for a tentative comparison of their distinctive internationalisation patterns.

North American Firms North American firms constitute the largest group in our full sample, with 80 firms from the United States and 10 from Canada. Not surprisingly, these firms’ mean foreign sales ratio (F/T) sat below the average for the full sample in both 1991 and 2001.6 The trend, however, was clearly towards greater internationalisation, with F/T increasing from 17.3 percent to 23.4 percent over the decade (Table 4). While it was still possible, for at least the US corporations in the sample, to grow large while only serving the domestic market, more firms stepped into the international arena during the 1990s. This is confirmed in Table 5, with the number of firms with no foreign sales declining from 41.1 percent in 1991 to 23.3 percent in 2001. Rugman (2005, p. 26) found North American firms to be the most homeregion focused. Our data supports this claim, but also highlights the changes afoot. On average, home-regional sales accounted for a dominant 86.5 percent of total sales in 1991; by 2001, this figure had dropped to 81.3 percent (Table 4). As documented in Table 5, almost half of the firms (45.6 percent) had no sales outside the home region in 1991, but by 2001 this group had shrunk to 27.8 percent. On the other hand, the proportion of firms with more than 20 percent of their sales outside North America increased from 28.8 percent in 1991 to 43.3 percent in 2001.


54

Table 4.

Extent of Internationalisation – North American Firms (Mean Levels, in Percent).

Foreign Sales (F/T)




Europe Sales (E/T)

AsiaPacific Sales (AP/T)

17.3 21.1 23.4

86.5 83.4 81.3

3.8 4.5 4.7

13.5 16.6 18.7

9.4 11.9 12.1

2.5 3.4 4.3

1991 1996 2001

Note: Sample size is 90 firms. E/T only reported for 75 firms in 1991, 70 in 1996 and 55 in 2001. AP/T only reported for 61 firms in 1991, 59 in 1996 and 49 in 2001.

Table 5. Breakdown of Foreign Sales and Rest of World Sales Shares – North American Firms (in Percent). Foreign Sales (F/T) Brackets

1991 1996 2001

0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

41.1 35.6 23.3

10.0 8.8 14.5

10.0 8.9 14.4

12.2 14.5 8.9

12.3 8.9 21.1

6.6 11.1 6.7

7.8 12.2 11.1

Rest of World Sales (ROW/T) Brackets

1991 1996 2001

0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

45.6 36.7 27.8

12.2 11.1 18.9

14.4 13.3 10.0

6.7 15.6 12.2

14.3 8.9 17.8

3.4 8.8 6.6

4.4 5.6 6.7

The decline in R/T for the North American firms (Table 4) is countered by a rise, albeit off a very small base, in their ROR/T, again highlighting the need for such an adjusted measure of home-regional sales. The ‘‘small base’’ clearly is an artefact of both the dominant share of US firms in the North American subsample and the dominant size of the US economy within the North American region. As discussed earlier, the data on the individual host regions (Europe and the Asia-Pacific in the case of North American firms) are not comprehensive and our findings are only tentative.7 Table 4 shows that the share of both European (E/T) and Asia-Pacific sales (A/T) grew over the period of


55

investigation. While E/T surged from 1991 to 1996 and recorded only modest growth during the subsequent five-year period, the A/T ratio grew at an even pace throughout. Overall, Europe has remained a more attractive sales location than the Asia-Pacific. This is plausible on cultural, economic and institutional grounds, and may also reflect the attraction of the EU’s advanced internal integration. These factors would seem to have outweighed the faster economic growth enjoyed by Asian economies during the 1990s.

European Firms The 36 European firms in our sample come from 10 different countries, with France (12 firms), the United Kingdom (10) and Germany (5) the most common domiciles. Not surprisingly, given the level of economic integration within the EU and the smaller size of the home-country economies of these firms, European firms are the most internationalised in our dataset. As shown in Table 6, European firms, on average, earned 42.0 percent of their revenue (F/T) from outside of their home country in 1991. In that same year, only 11.1 percent of the European firms had no recorded foreign sales (Table 7). By 2001, the average F/T had risen to 62.3 percent and only a single firm in our dataset remained solely domestic. Similarly, in 1991, 41.7 percent of firms relied on foreign markets for more than half of their total sales. By 2001, this proportion had climbed to 72.2 percent. A major motivation behind the formation and the continued ‘‘deepening’’ of the EU was to enable European businesses to expand their scope and to integrate their activities across the member countries. This would allow Table 6.

1991 1996 2001

Extent of Internationalisation – European Firms (Mean Levels, in Percent).

Foreign Sales (F/T)




North America Sales (NA/T)

AsiaPacific Sales (AP/T)

42.0 49.8 62.3

78.3 72.9 63.3

20.3 22.7 25.4

21.7 27.1 36.6

12.3 13.9 22.8

7.6 9.3 10.3

Note: Sample size is 36 firms. NA/T only reported for 25 firms in 1991, 26 in 1996 and 31 in 2001. AP/T only reported for 21 firms in 1991, 20 in 1996 and 21 in 2001.


56

Table 7. Breakdown of Foreign Sales and Rest of World Sales Shares – European Firms (in Percent). Foreign Sales (F/T) Brackets

1991 1996 2001

0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

11.1 8.3 2.8

13.98 8.4 2.8

2.8 2.7 5.5

2.8 2.8 0

8.3 8.4 11.1

22.2 13.8 2.8

41.7 55.6 72.2


1991 1996 2001

0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

19.4 16.7 2.8

11.2 13.9 8.3

27.7 16.6 16.7

11.1 11.1 13.9

8.4 11.1 22.2

8.3 8.4 5.5

13.9 22.2 30.6

firms to attain larger economies of scale and scope in order to compete more effectively against the vast corporations from the United States and Japan. The relatively high F/T ratio (42.0 percent) for European firms in 1991, coupled with their high level of intra-regional sales (78.3 percent) in that year, indicate that home-regional expansion was well underway at the start of the 1990s. By 2001, in the wake of further policy measures designed to deepen integration within the EU, the large European firms in our sample had further expanded their international sales presence. Remarkably though, much of that overseas growth had come from outside the region and the share of home-regional sales declined substantially from 78.3 percent to 63.1 percent. Our ROR/T measure again allows us to flesh out the details of this phenomenon. As can be seen in Table 6, the portion of intra-regional sales net of domestic sales did in fact grow significantly over the period of examination, but obviously could not offset the impact of slower domestic sales to maintain R/T at 1991-levels. The need to seek growth outside the home country was most pronounced with the European corporations: for the stylised average European corporation, domestic sales growth for the period under examination was only 5.8 percent per annum. By comparison, annual growth rates for R/T, ROR/T and ROW/T were 8.1 percent, 13.0 percent and 16.5 percent, respectively. The host-region sales shares reported in Table 6, while reliant on incomplete data, point to most of that inter-regional expansion occurring in North America (NA/T). As such, they mirror the results obtained for the North American subsample. It would appear that European firms targeted


57

primarily the large US market (and to a lesser extent the Canadian market) to build sales beyond the home region. While in 1991, only 5 of the 25 valid observations in our European subsample had North American sales shares (NA/T) in excess of 20 percent, their numbers had grown to 15 (out of 31 valid observations) by 2001. In the same year, 5 firms (AstraZeneca, BP, ING Group, Skanska and Wolseley) reported NA/T ratios in excess of 40 percent. Sales growth in the Asia-Pacific region was somewhat less impressive, but the mean host-regional sales share had broken through the 10 percent mark by 2001.8 At that time, only 4 firms (out of the 21 valid observations) had attained AP/T ratios above 20 percent. In general, the preference for the North American market is likely to reflect the perception of lower cultural, institutional and economic ‘‘distance’’ relative to the Asia-Pacific. Asia-Pacific Firms Analysis of our Asia-Pacific subsample must be prefaced by a reminder that this subset suffers from data limitations.9 The 33 Asia-Pacific firms in our sample are overwhelmingly Japanese (30 companies), complemented by 2 Australian and 1 Korean firm. As a group, these firms were the most aggressive internationalisers over the decade. On average, Asia-Pacific firms earned only 16.0 percent of their revenue from outside the home country (F/T) in 1991. By 2001, F/T had almost doubled to reach 31.6 percent (Table 8). For the stylised average Asia-Pacific firm, this translates into a formidable compound growth rate in F/T of 18.2 percent per annum, 10 percent higher than the domestic growth rate. Table 8.

1991 1996 2001

Extent of Internationalisation – Asia-Pacific Firms (Mean Levels, in Percent).

Foreign Sales (F/T)

HomeRegion Sales (R/T)



Europe Sales (E/T)

North America Sales (NA/T)

16.0 19.9 31.6

NA NA 73.5

NA NA 5.2

NA NA 26.5

NA NA 7.7

NA NA 16.5

Note: Sample size is 33 firms. E/T only reported for 26 firms in 2001. NA/T only reported for 32 firms in 2001.


58

Table 9. Breakdown of Foreign Sales and Rest of World Sales Shares – Asia-Pacific Firms (in Percent). Foreign Sales (F/T) Brackets

1991 1996 2001

0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

12.1 18.2 9.1

18.2 15.1 6.1

36.4 30.3 15.1

34.2 12.2 15.2

6.1 12.1 21.2

3.0 3.0 15.1

0 9.1 18.2


2001

0%

0.1–10%

10.1–20%

20.1–30%

30.1–40%

40.1–50%

W50%

12.1

12.1

21.3

15.1

18.2

6.0

15.2

As shown in Table 9, in 1991 two-thirds of the firms (66.7 percent) had an F/T equal to or less than 20 percent. By 2001, roughly the same portion of firms (69.7 percent) earned more than 20 percent of their revenue offshore. If Asia-Pacific firms were testing the waters of foreign expansion in 1991, by 2001 they were swimming confidently. Too few reliable data points exist to present R/T, ROR/T and ROW/T for 1991 and 1996 with any confidence. By 2001, the inter-regional sales share (ROW/T) for our Asia-Pacific subsample outstripped ROR/T by a ratio of 5:1 (Table 8). The same data limitations prevent us from supplying host-regional sales shares for the earlier years. By 2001, most of the firms in our Asia-Pacific subsample were reporting sales for Europe and North America. While the mean European sales share of 7.7 percent was still modest, the North American share had reached a more impressive 16.5 percent (Table 8).

DISCUSSION The empirical investigation presented in this paper generates a number of important insights. While some of our findings are based on less than perfect data, the overall picture emerging from our analyses is clear. To recap, the sampled firms experienced the following trends from 1991 to 2001: (1) a substantial decrease in domestic sales as a share of total sales; (2) a decrease in the share of home-regional sales (R/T);


59

(3) an increase in the share of home-regional sales net of domestic sales (ROR/T); and (4) a substantial increase in the share of sales outside the home region (ROW/T). These results also hold for separate analyses of our North American and European subsamples. Analysis of our Asia-Pacific subsample was hampered by incomplete data. For this particular subset of firms, we can only confirm the first of the above trends with confidence. (We note that our findings for 1991 to 2001 are at variance with the results reported in Rugman and Oh (2007). Using a different sample, they find no changes in the intra-regional sales ratio for the period 2001–2005.)

Explanation of Current Home-Regionality Rugman and coauthors (e.g. Rugman & Verbeke, 2004; Rugman, 2005) are correct when they state that, at present, sales of many of the world’s largest firms are concentrated in their home region. The arguments underpinning their regionalisation thesis were presented at the beginning of this paper. In essence, many FSAs are home-region bound: they can be deployed within the home region with relative ease, but inter-regional deployment is stifled by significant additional costs.10 Assuming further intra-regional integration and continuing inter-regional protectionism, it is argued that intra-regional sales have been and will remain the norm and inter-regional sales the exception. The results based on our sample for the years 1991, 1996 and 2001 lead us to a somewhat different interpretation. At the outset, it is important to qualify Rugman and coauthors’ important empirical discovery. First, homeregional sales dominate total sales chiefly because of the magnitude of home country sales. Once home country sales are removed, the share of home-regional sales (ROR/T) is markedly smaller than the share of ‘rest of world’ sales (see Fig. 1, see also Fig. 1 in Rugman & Oh, 2007). Second, longitudinal trends from 1991 to 2001 suggest that the share of sales outside the home region grew faster than the share of sales within the home region – irrespective of whether or not domestic sales are removed (again see Fig. 1). Both of these findings are difficult to reconcile with Rugman and coauthors’ theoretical arguments and their pronouncements on (the lack of) change. As argued elsewhere (Osegowitsch & Sammartino, 2007), we see the dominance of home-regional sales (including domestic sales) as subject to change over time, a position which finds support in our empirical results.

60


Based on the observed (international) sales pattern at a given point in time, Rugman and coauthors draw conclusions regarding the transferability of FSAs. We agree with their proposition that sales is the decisive indicator of FSA transferability or reach:11 (international) sales reflect (international) customers’ judgement about the attractiveness of the firm’s goods or services and, ultimately, the relevance of the underlying FSAs. But if we accept this proposition we must also accept that (international) sales over time may reveal the evolution of FSA reach. Based on the empirical evidence presented in this chapter, we must then conclude that FSA reach expanded from 1991 to 2001 and that FSAs were increasingly extended beyond the home region. It would seem that a gradual convergence process diminished ‘‘distance’’ (Ghemawat, 2001) across countries and regions, thus expanding the transferability and acceptability of many FSAs. While the phenomenon may have faded since 2001 (see Rugman & Oh, 2007), during the previous decade firms successfully expanded their sales outside the home-region boundary. They managed to leverage their FSAs in foreign markets within the home region, but, to an even greater extent, leveraged these FSAs in the rest of the world. A competing explanation for our finding of superior sales growth in host regions would be that the home region is simply reaching saturation point and companies are turning to other regions in a last-ditch effort to keep up growth. We discount this argument for two reasons. First, the share of ROR/T is growing at a healthy rate, suggesting that growth opportunities in the home region are far from depleted. Second, ROR/T ratios are actually rather small. In 2001, for the North American and European subsample, we calculated a domestic sales share of 65.3 percent. By comparison, the share of sales in the rest of the home region (ROR/T) accounted for a mere 10.6 percent (see Fig. 1).12 In other words, domestic sales are six times larger than sales in all the other countries of the home region. While these numbers may be inflated by the strong presence of US firms (80 out of 126 observations), for our dedicated European subsample the corresponding numbers are 37.7 percent and 25.4 percent. Here, domestic sales are 50 percent larger than sales to all the other countries in the home region. These results lead us to reject the home region saturation thesis. We further contend that actual FSA deployment is plagued by significant lag effects. In other words, a company’s realised FSA reach may not match its potential reach at a given point in time, due to factors such as Penrosian constraints13 and significant lead-times for assembling the required co-specialised assets (Teece, 1986). These lag effects may also help explain why, at present, home-regional sales account for a very large portion of total


61

sales and why Rugman (2005) and Osegowitsch and Sammartino (2007) (still) find the home-regional category – however defined – the dominant category. The longitudinal trends charted for the period 1991–2001 suggest that firms do eventually close the gap to the potential reach of their FSAs. Correspondingly, a significant number of erstwhile home-regional firms had already migrated into the bi-regional category and, to a lesser extent, the global category by 2001 (see Osegowitsch & Sammartino, 2007). If the trends presented in this chapter persist, the home-regional category will be further weakened in years to come as more companies assume bi-regional and possibly global status.

CONCLUSION As mentioned earlier, we hope that charting historical trends in firm internationalisation will facilitate our understanding of the status quo and, ultimately, allow us to glimpse into the future. In view of the strong trends we document from 1991 to 2001 – strong in terms of their magnitudes and in terms of their universality – it is difficult not to take issue with Rugman’s assertion that nothing will change and that the overwhelming home-region orientation of the world’s largest firms will persist (2005, p. 63). For our subsample of the Fortune Global 500, we have identified a decline in the home-regional sales share (R/T) from 1991 to 2001, which is the result of a falling domestic sales share overwhelming a growing ‘rest of region’ share. More generally, we have demonstrated a trend towards inter-regional expansion for the decade prior to 2001, with inter-regional sales growth outstripping growth rates in all other categories. Apart from the dynamics, the regionalisation thesis is also challenged by the fact that in 2001, ROR/T amounted to only 40 percent of ROW/T for our full sample (not shown). For our European and North American subsample, the corresponding figure is 45 percent.14 In other words, FSAs were more than twice as likely to be leveraged in foreign countries beyond the home region than within the home region. The picture of the home region painted by Rugman and coauthors is one of an almost insurmountable barrier that only a handful of exceptional competitors – those in the rare bi-regional and global categories – will be able to cross. The empirical results obtained for our sample for the period 1991–2001 challenge this view (but see also Rugman & Oh, 2007). At a general level, the mean share of non-home-regional sales and the corresponding growth rates suggest that a sizeable number of companies were able to

62


successfully venture beyond the home region, and that their ranks are growing. The breakdowns of ‘rest of world’ internationalisation presented herein (Tables 3, 5, 7 and 9) allow us to dismiss the notion that mean ROW/T ratios may have risen solely on account of a small number of exceptional firms that registered phenomenal increases in ROW/T. Instead, we observed a comprehensive migration of firms from the lower ROW/T brackets into the higher ROW/T brackets. What emerges from the longitudinal evidence assembled in this chapter is that there is no sharp delineation between the home region and host regions. It is not only after all the opportunities in the home region are exhausted that firms venture further afield, as a kind of ‘‘last resort’’ option to keep up growth. Instead, some firms simultaneously leverage their FSAs within and beyond the home region.15 This suggests that the home-region has a rather permeable border and what divides companies in the home-regional, bi-regional and global categories are differences of degree rather than differences in kind.

NOTES 1. Even under the most ‘‘relaxed’’ system of thresholds presented in Osegowitsch and Sammartino (2007), some 55 percent of the world’s largest firms are classified as homeregionals. An alternative, more demanding standard – requiring at least 15 percent instead of 10 percent of sales in a host region or host regions (to attain bi-regional and global status, respectively) – results in a home-regional share of 64 percent. Under Rugman’s original classification, some 88 percent of firms are designated home-regionals. 2. Any FSAs that may have been developed in overseas subsidiaries face the same kind of limitations in terms of transferability and acceptability since ‘‘they cannot be readily deployed and exploited beyond the host country’s borders or a perhaps somewhat broader region’’ (Rugman & Verbeke, 2007, p. 202). 3. Not all corporations use a triad scheme to disaggregate their sales. Quite a number of them aggregate across triad regions and/or include data from non-triad areas such as the Middle East, Africa or Latin America among their triad figures (see also Rugman, 2005, Ch. 2 on this issue). 4. By way of comparison, total sales for the firms ranked in the Fortune Global 500 increased at a compound rate of 10.4 percent per annum between 1991 and 2001. A stylised average firm whose total sales and F/T developed in line with the norm would have recorded a compound annual growth rate of 8.7 percent in domestic sales and 15.0 percent in foreign sales. 5. The F/Ts for North American and European firms in 1991, 1996 and 2001 (Table 2) are very similar to the corresponding F/Ts for the full sample of 159 firms (22.7 percent, 27.3 percent and 33.9 percent, respectively). This would indicate that dropping Asia-Pacific firms has no significant impact on the result obtained. 6. The North American F/T averages are boosted considerably by the 10 Canadian firms in the sub-sample. Their F/T ratios stood at 37.2 percent in 1991 and 55.5 percent in 2001.


63

7. By default, firms that reported 100 percent North American sales were indicating that they had zero sales in Europe and the Asia-Pacific. As the number of such firms fell between 1991 and 2001, there was a corresponding drop in the overall availability of reliable data for E/T and AP/T. 8. We would expect more recent data to show an expanding focus on the Asia-Pacific, driven principally by the booming Chinese economy. 9. Apart from missing data, the Asia-Pacific is also the most dynamic region, which generates its own set of difficulties. For instance, there were 11 Chinese firms in the Fortune Global 500 (2002) list that was used as a sampling frame by Rugman and coauthors. Unfortunately, none of these firms feature in our Asia-Pacific subsample since a geographic breakdown of their 1991 sales is not available. As such, (sub)-sample bias is likely. 10. Strictly speaking, their argument applies to any FSA the firm is trying to deploy outside the FSA’s ‘‘region of origin’’, irrespective of whether the FSA was created in the home region or in host regions (Rugman & Verbeke, 2007, p. 202). 11. The notion of FSA region boundedness (or country boundedness) implies an exogenous phenomenon, determined entirely by environmental forces. As indicated elsewhere (Osegowitsch & Sammartino, 2007), we view the boundaries of effective FSA leverage as also governed by firm-specific aspects and as partly controlled by management. Hence, we prefer the term FSA reach. 12. The corresponding number for our full sample is even lower, at 9.5 percent. 13. The ‘‘Penrose effect’’ simply suggests that there are strict limits to a firm’s growth rate due to dynamic adjustment costs that are incurred by firms trying to grow their productive resources. Penrose (1959) focussed on one major source of dynamic adjustment costs, namely those attributable to the expansion of management resources. She insists that a firm’s expansion requires the services of experienced internal managers. Hiring new managers is an inadequate solution since only seasoned internal managers can undertake the coordination task inherent in firm expansion. As a result, the rate of growth is limited by the rate at which the firm can develop internal managers. Not only domestic expansion but also international expansion is subject to the Penrose effect (e.g. Rugman & Verbeke, 2002; Tan & Mahoney, 2005). 14. The corresponding share for Rugman’s larger sample is 47 percent (10.4 percent/23.3 percent=0.467, see Fig. 1 in Rugman & Oh, 2007). 15. In passing, we note that this is one of the limitations of this chapter and the broader research stream on regionalisation. The identification of regional sales outcomes and regional trends may conceal significant within-region variation. Our results – showing firms growing their sales in the home region (ROR/T) as well as host regions (ROW/T) – coupled with anecdotal evidence would suggest that firms are predominantly making country-by-country choices in selecting new markets for their FSAs. Aggregating these choices to the regional level, and inferring that firms are making region-by-region choices may be preventing us from getting closer to actual firm strategies.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the assistance of Eric Quintane and Catherine Lambiris in compiling the data used in this chapter.

64


REFERENCES Ghemawat, P. (2001). Distance still matters: The hard reality of global expansion. Harvard Business Review, 79(8), 137–147. Goerzen, A., & Asmussen, C. G. (2007). The geographic orientation of multinational enterprises and its implications for performance. In: A. M. Rugman (Ed.), Research in global strategic management: Vol. 13. Regional aspects of multinationality and performance. Amsterdam: Elsevier. Osegowitsch, T., & Sammartino, A. (2007). Reassessing regionalisation. Journal of International Business Studies, 38 (in press). Penrose, E. (1959). The theory of growth of the firm. New York: Wiley. Rugman, A. M. (2000). The end of globalization. London: Random House. Rugman, A. M. (2005). The regional multinationals: MNEs and ‘‘global’’ strategic management. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Collinson, S. C. (2005). Multinational enterprises in the new Europe: Are they really global? Organizational Dynamics, 34(3), 258–272. Rugman, A. M., & Oh, C. H. (2007). Multinationality and regional performance, 2001–2005. In: A. M. Rugman (Ed.), Research in global strategic management: Vol. 13. Regional aspects of multinationality and performance. Amsterdam: Elsevier. Rugman, A. M., & Verbeke, A. (2002). Edith Penrose’s contribution to the resource-based view of strategic management. Strategic Management Journal, 23(8), 769–780. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Rugman, A. M., & Verbeke, A. (2007). Liabilities of regional foreignness and the use of firmlevel versus country-level data: A response to Dunning et al. (2007). Journal of International Business Studies, 38(1), 200–205. Tan, D., & Mahoney, J. T. (2005). Examining the Penrose effect in an international business context: The dynamics of Japanese firm growth in US industries. Managerial and Decision Economics, 26, 113–127. Teece, D. J. (1986). Profiting from technological innovation. Research Policy, 15(6), 285–305.

THE GEOGRAPHIC ORIENTATION OF MULTINATIONAL ENTERPRISES AND ITS IMPLICATIONS FOR PERFORMANCE Anthony Goerzen and Christian Geisler Asmussen ABSTRACT Diametrically opposed views exist on the nature of global strategic management, the existence of global multinational enterprises (MNEs), and the performance implications of regional and global orientation. However, these divergent opinions on the nature of global strategy ‘‘should be considered a starting point for introducing systematically a regional component in international business research’’ (Rugman & Verbeke, 2004a, p. 5). Our aim in this chapter, therefore, is to examine the geographic orientation (i.e., regional versus global) of multinational firms to provide new insights into some of the important characteristics that distinguish between these MNE archetypes. Our findings suggest that the interaction between the MNE’s organizational characteristics and its geographic orientation is associated with MNEs performance. By arguing for a contingency perspective on regional and global strategy, we thus attempt to bridge the gap between these two opposing viewpoints. Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 65–83 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13004-7

65

66

ANTHONY GOERZEN AND CHRISTIAN GEISLER ASMUSSEN

INTRODUCTION Foreign direct investment (FDI) has been the fundamental driver underlying the current unprecedented trend of growth in the global economy (UNCTAD, 2000). Advances in information, communication, and transportation technologies have been suggested as some of the basic reasons why political boundaries are becoming less significant as investment barriers to multinational enterprises (MNEs). Scholars from a variety of disciplines have for some time been interested in these developments, often subsumed under the label ‘‘globalization.’’ Trends in economic globalization and their implications for MNEs have led to the concept of global strategy as an emerging major area of international business scholarship (Govindarajan & Gupta, 2001; Jeannet, 2000; Yip, 2002). Recently, however, some researchers have begun to question the basis of global strategy, arguing that global MNEs are virtually non-existent and that globalization, as generally envisioned, is not relevant to scholars or practitioners (Rugman, 2000, 2001; Rugman & Brain, 2003; Rugman & Hodgetts, 2001; Rugman & Verbeke, 2004a). By examining micro-level data on sales patterns, these researchers have indicated that MNEs predominantly internationalize within rather than across the Asian, European, and North American economic regions. That is, ‘‘global’’ business is actually driven by MNEs which derive their revenues predominately within their own home regions. This work has led to the perspective that ‘‘not only is globalization a myth; global strategy is a myth’’ (Rugman, 2001). Viewing competition from this region-centric lens has led to the normative assertion that MNEs should design strategies and adopt structures that focus on markets close to their countries-of-origin. Yet, so far, scholars have not been able in empirical studies to converge on a clear causal relationship from regional or global orientation to performance. Rugman and Sukpanich (2006) find that home region orientation increases firm performance but that the effect is moderated by other organizational characteristics. Delios and Beamish (2005), on the other hand, find that global firms have higher performance compared to regional firms. It appears, therefore, that there are diametrically opposed views on and mixed evidence of the nature of global strategic management, the existence of global MNEs, and the performance implications of regional and global orientations. As highlighted by Rugman and Verbeke (2004a, p. 5), however, these divergent opinions on the nature of global strategy is based on an examination of sales distributions among the triad economic regions

Geographic Orientation of MNEs and Its Implications for Performance

67

and this ‘‘should be considered a starting point for introducing systematically a regional component in international business research.’’ This chapter, therefore, is intended to inform this debate by making several significant contributions. Our first aim is to evaluate whether regional MNEs are indeed a predominant organizational form, using different data and methods than those used by previous researchers. By examining a sample of 580 Japanese MNEs, we determine the extent to which the international investments of these firms can be considered regional or global, thereby testing the generalizability of Rugman and Verbeke’s (2004a) findings. Upon evidence that there are in fact significant groups of MNEs that are global or home region-oriented, the second contribution of this research is to provide new theoretical and managerial insights into some of the important characteristics distinguishing between these MNE archetypes. In particular, we explore the extent to which regional and global firms differ in terms of size, multinationality, and configuration of intangible assets, and we show how each of these firm characteristics interact with the MNE’s geographic orientation in determining the firm’s performance.

GEOGRAPHIC ORIENTATION AND MNE PERFORMANCE International management scholars have for many years tried to uncover the relationship between international investment and performance. Typically, the nature of an MNE’s international posture is conceptualized as the extent of foreign operations, often measured by the share of foreign to total sales (Hitt, Hoskisson, & Kim, 1997). This makes some theoretical sense as the costs and benefits of internationalization, by definition, accrue only to the extent the firm ventures outside its home market. Yet recent research suggests that it is not as much the degree of internationalization as it is the pattern – i.e., the distribution of foreign activities among foreign countries – that matters to performance (Goerzen & Beamish, 2003). To illustrate, one MNE may spread its operations among diverse country environments in all corners of the world, while another may focus all of its investments in a closely integrated set of proximate countries; while these two firms may be equally internationalized in terms of FDI as share of total capital invested, they are clearly completely different in terms of spatial organization and, therefore, they arguably face different opportunities and challenges (Goerzen, 2005a, 2005b; Goerzen & Beamish, 2005).

68


One particular internationalization pattern, which has garnered increased attention recently, is that of the regionally oriented multinational firm that operates primarily within its home region. Using sales data, Rugman and Verbeke (2004a) show that the regional MNE is, in fact, a predominant organizational form. In fact, they find that only nine out of the world’s 500 largest MNEs have managed to achieve a global sales distribution with a significant presence in all three regions of the economic triad (i.e., Europe, North America, and Asia). A number of other empirical studies support this result in different industries and country settings and using assets as a measure of upstream globalization (see e.g., Rugman & Girod, 2003). The authors of these studies propose two stylized facts to explain this phenomenon. First, firm-specific advantages (FSAs) may be more location dependent than previously assumed and MNEs may, therefore, reach the limits of organization, exhausting the value of their proprietary resources and capabilities, before obtaining a truly global scope (Rugman & Verbeke, 2004a, 2004b). For example, as illustrated by Wal-Mart’s recent withdrawal from the German market, a US firm with a proprietary advantage in supply chain management and logistics may experience significant cultural adaptation problems if it were to implement its managerial systems in Europe. Second, they contend that MNEs also encounter a liability of interregional foreignness (Rugman & Brain, 2003) as they venture outside their home region, leading them to incur relatively higher costs than those they would face in foreign markets closer to home. For example, a Japanese firm may find it somewhat more difficult to adapt to local networks and business practices in France than would a Spanish firm – whereas both would be at a disadvantage compared to a French firm. This inter-regional liability of foreignness can thus be defined as the difference between the costs of doing business outside the MNE’s home region relative to the costs of doing business in foreign countries within the home region. Taken together, these arguments suggest that most MNEs should be neither completely overwhelmed by the liability of foreignness, nor able to fully overcome it by means of their FSAs, but rather caught in a state of ‘‘semi-globalization’’ (Ghemawat, 2003).

A REASSESSMENT OF THE NATURE OF GEOGRAPHIC ORIENTATION As is natural for an emerging stream of research, there is still little consensus of how to operationalize the construct of regional strategy (see, e.g., the


69

critique by Aharoni, 2006). The approach of most previous studies has been to decompose the total operations of the firm into a home and a host region component. However, this means that any firm with a low degree of internationalization (in fact, any firm which has less than 50% of its operations abroad) will automatically be classified as home region-oriented, regardless of the regional distribution of its international operations. Hence, we cannot say if a given home region-oriented firm is so classified merely because it has a low degree of internationalization, or because it has, in fact, chosen to internationalize within its home region – two possibilities with quite different theoretical implications. Therefore, while the arguments for regional strategy are compelling, we believe there is a need for further empirical corroboration of the phenomenon. To control for the degree of internationalization, this study follows the approach of Delios and Beamish (2005) in using a more direct measure of the international orientation of the MNE, in which we decompose the foreign operations of the firm into a home and host region component. In this way, we operationalize a regional multinational firm, which has a large share of its foreign operations within its home region. An example of such a firm is Kawasho Corporation, a Japanese steel company, which has 49 of its total 57 foreign subsidiaries located in the Asia-Pacific region, primarily in Korea, Malaysia, Thailand, and the Philippines. Hence, Kawasho can be said to be highly home region-oriented in its approach to internationalization given that 86% of its foreign ventures are within its home region. Thus, to measure the home region orientation of each of the firms in our sample, we divide the number of foreign subsidiaries in the Asia-Pacific region with the total number of foreign subsidiaries of that firm. If this value is above 50%, the firm is classified as regional and otherwise it is classified as global in its internationalization approach. Fig. 1 provides a graphic representation of home region orientation among the firms in our sample. Clearly, when looking at the composition of foreign operations, the predominance of regional over global MNEs is not nearly as large as has been suggested by previous studies. In fact, there seem to be a wide range of firms of both high and low regional orientations; based on the 50% criteria, for example, there are 358 regional firms and 222 global firms. There is also a high degree of heterogeneity, as some firms have virtually all of their subsidiaries in their home region while others have very few. This points to the question of what leads some of these firms to pursue a global strategy, while others focus on their home region? While many factors may shape the direction of a given firm’s FDI, we believe a central contingent factor is the MNE’s configuration of FSAs. Internalization

70

ANTHONY GOERZEN AND CHRISTIAN GEISLER ASMUSSEN 7

6

Percent

5

4

3

2

1

0 −0.01

0.06

0.13

0.2

0.27

0.34

0.41

0.48

0.55

0.62

0.69

0.76

0.83

0.9

0.97

Share of subsidiaries in Home Region

Fig. 1.

Distribution of Home Region Orientation.

theory suggests that firms undertake FDI in order to internalize the exploitation of FSAs in foreign countries (Buckley & Casson, 1976; Hennart, 1982; Rugman, 1981). Furthermore, since MNE’s face a liability of foreignness leading them to incur higher costs in host markets than those faced by their local competitors (Hymer, 1976; Zaheer, 1995), they need to possess proprietary advantages in order to compensate for those costs and make multinational expansion a worthwhile endeavor. The degree to which these FSAs are regional or global in reach (Rugman & Verbeke, 2004a), and whether they are strong enough to compensate for the increased liability of foreignness outside the home region (Rugman & Brain, 2003), are crucial determinants of the viability of a global strategy. We focus on two of the most commonly cited sources of proprietary FSAs: technological assets and marketing assets, respectively proxied by the ratio of R&D and marketing expenditures to sales (Morck & Yeung, 1991). These FSAs may affect not only the MNE’s degree of multinationality per se, as suggested by prior research, but also the pattern of its allocation of international activity (Rugman & Sukpanich, 2006). Another important determinant of geographic orientation is the MNE’s cognitive capacity to handle the increased complexity resulting from


71

operations with global reach. According to the process theory of internationalization, MNEs internationalize in small incremental steps in order to match their tolerance for risk and uncertainty (Johanson & Vahlne, 1977, 1990; Johanson & Wiedersheim-Paul, 1975). International experience facilitates learning by managers and employees, enabling the MNE to pursue an increasingly wide geographic orientation. We will examine the impact of international experience in both a relative and an absolute sense. The relative sense pertains to the MNE’s degree of internationalization, measured as the foreign capital invested as a share of the total capital invested, and the absolute sense to the size of the foreign operations of the firm as given by the number of foreign subsidiaries.

METHODS Description of the Data The primary source of data used in this study was a 1999 survey of 13,529 subsidiaries of 580 Japanese MNEs that, following Stopford and Wells (1972) definition of an MNE, had operations in six or more countries. The survey results were published by Kaigai Shinshutsu Kigyou Souran, a publication of Toyo Keizai Shinposha (Toyo Keizai, 1999). Toyo Keizai (which translates to Oriental Economist) was formed in 1895 and currently publishes more than 100 volumes annually as well as a variety of data covering economic conditions, stock markets, and Japanese corporations. The 13,529 surveys, which were sent to the subsidiaries through their parent firms, were completed by the subsidiary general managers with a response rate of 60%. The survey requested basic facts such as subsidiary location, industry, annual revenue, number of employees, and capital invested. These subsidiary-level data, aggregated to the corporate level using FORTRAN, were then augmented with corporate and industry details from Compustat and the Analysts’ Guide (Daiwa Institute of Research, 1999). Variable Measurement Geographic Orientation Earlier empirical work on regional strategies (Rugman, 2005; Rugman & Verbeke, 2004a) classified MNEs into four groups – home-regional, hostregional, bi-regional, and global – while Rugman and Sukpanich (2006)

72


later combined the three latter groups into one to move toward a regional– global dichotomy for two reasons. First, the host-regional, bi-regional, and global MNEs are arguably more different in degree than in kind, as they all share the theoretically important property that they have significant activity outside their home region. On a more practical level, Rugman and Sukpanich (2006) found each of the three non-home region-oriented groups to be too small for meaningful empirical analysis. Our study follows the approach of Rugman and Sukpanich (2006) distinguishing simply between home region-oriented (henceforth ‘‘regional’’) and non-home regionoriented (‘‘global’’) MNEs. On this basis, we define regional MNEs as those firms that have half or more of their subsidiaries located within their home region while global firms have more than half of their operations outside their home region. The 50% cut off point is both intuitive and it has previously been used in studies of large MNE’s regional internationalization patterns (Rugman, 2005; Rugman & Verbeke, 2004a). However, where previous studies used total sales, we classify our firms based on the distribution of FDI. We use FDI as share of total capital invested to measure the firm’s degree of multinationality. We then define home region orientation as the share of the foreign subsidiaries located within the home region (in this case the AsiaPacific region, given that our sample is of Japanese MNEs). This variable was then used to classify the firms in the sample. We argue that there are two advantages to our classification approach. First, we believe that, as compared to sales, capital invested better captures the organizational complexity associated with a given geographic orientation, thus going to the heart of internalization theory, the dominant perspective on the multinational firm. Second, looking at international rather than total operations, we control for the home country effect so that firms are not classified as regional merely because they have a low degree of multinationality (Aharoni, 2006). Only the distribution of the foreign operations tells us something about the internationalization strategy of the MNE – i.e., whether it is regionally or globally oriented.1 Economic Performance Although accounting-based performance measures have been common in strategic management research, the strategy literature is increasingly using market-based measures that adjust for levered and unlevered market risk (Farjoun, 1998) since they are more ‘‘forward looking’’ as compared to accounting-based measures that are retrospective, based on historical information (Meyer, 1994). The performance of the MNE is proxied,


73

therefore, by three measures. First, Jensen’s Alpha (Jensen, 1968) is defined as ai ¼ r¯ i ½¯rf þ bi ð¯rM r¯f Þ where ai is firm i’s ‘‘excess’’ return over and above that predicted by the Capital Asset Pricing Model, r¯i is firm i’s average stock market return (i.e., capital gains/losses plus dividends), r¯ f is the risk-free rate of return defined by the 10-year Japanese Corporate Bond Benchmark Rate, bi is the firm’s b (derived from the firm’s stock price variance), and r¯m is the average Nikkei Stock Exchange return, all over the sample period. The second market-based measure of Economic Performance is Sharpe’s Measure (Sharpe, 1966), calculated as follows: ð¯ri r¯ f Þ=si where r¯i is firm i’s average rate of return and r¯f is the risk-free rate of return, and si is the firm’s standard deviation of returns. Following prior research (e.g., Nayyar, 1993), the third measure of market-based economic performance is the market-to-book ratio. All three measures are marketbased in the sense that they incorporate information about the stock market’s reaction to the firms’ strategies. We performed a confirmatory factor analysis on the three items and the eigen value criterion confirmed that they can indeed be considered as loading on the same construct. Cronbach’s alpha for the scale is 0.72 indicating adequate reliability. Proprietary Assets Internalization theory attributes the growth of MNEs, in both number and size, to their abilities to organize, control, and transfer proprietary assets within the firm rather than through external markets (Buckley & Casson, 1976; Rugman, 1981). In effect, inefficient markets encourage firms to appropriate the value from these assets through internal use and development. While proprietary assets can take a variety of forms, they are most often conceptualized as either technical knowledge leading to superior production methods (i.e., lower costs and/or superior quality) or market knowledge including special skills in product styling or promotion (Caves, 1996). Thus, proprietary assets will be evaluated in this chapter using measures of technological and marketing assets. Following prior research (e.g., Delios & Beamish, 1999; Kogut & Chang, 1991; Morck & Yeung, 1991), proprietary technological assets will be measured by R&D intensity and proprietary marketing assets will be measured by advertising intensity. Both of these measures will be obtained from Datastream. International Experience Experience has been shown to enable firms to make better use of internal capabilities (Pennings, Barkema, & Douma, 1994) and, therefore, is included in our model. Consistent with prior research, we use two measures to access

74


the concept of international experience including the number of foreign subsidiaries as well as foreign capital invested (Goerzen, 2001).

RESULTS Table 1 provides descriptive statistics and bivariate correlations of home region orientation, international experience, proprietary assets, and performance. Several surprising relationships emerge immediately from the correlations. First, the MNE’s intangible assets (R&D and advertising intensity) are apparently not related to the firm’s degree but rather to its pattern of internationalization in that global firms have higher levels of these FSAs than do regional firms. According to conventional wisdom, intangible assets are essential prerequisites to internationalization; our results suggest that global firms have particularly high levels of these resources and capabilities. Second, while internationalization and the scale of foreign operations are correlated, neither variable is related to the MNE’s home region orientation. Hence, contrary to expectation, global firms do not seem to have more foreign operations than do regional firms. That is, they are not simply larger firms but rather they appear to have only a different spatial distribution of operations. Finally, home region orientation seems to be a stronger predictor of performance than internationalization. The negative correlation between ratio of FDI in the Asia-Pacific region and performance means that, on average, global firms perform better than do regional firms. Of course, correlations only show bivariate relationships that may not reflect causality. Since our interest is in the way that MNE’s geographic orientation interacts with other key variables and their associations with Table 1. Variable

Descriptive Statistics and Correlations. Mean SD

1

2

3

4

5

Number of subsidiaries 29.4 60.4 FDI as share of capital invested 0.14 0.14 0.26 Ratio of FDI in the Asia-Pacific region 0.54 0.19 0.01 0.02 R&D intensity 0.02 0.03 0.01 0.06 0.29 Advertising intensity 0.01 0.02 0.06 0.09 0.18 0.12 Market performance 0.00 1.00 0.04 0.04 0.14 0.22 0.12 po0.05. po0.01. po0.001.


75

performance, we segment our sample firms into groups based on their values of their proprietary assets (i.e., R&D intensity, advertising intensity), the size of their foreign operations, and their geographic orientation. For each variable, firms above and below the medians were classified as ‘‘high’’ and ‘‘low,’’ respectively, and each of those categories was further subdivided among the regional–global axis. For each variable, this resulted in four categories of firms in a two-by-two matrix. We then estimated the performance within each group and performed the Analysis of Variance (ANOVA) to determine group differences.2 The next section reports the results for each of these analyses.

PROPRIETARY ASSETS, GEOGRAPHIC ORIENTATION, AND PERFORMANCE Technological Assets Table 2 shows the interactive performance effect of R&D intensity and geographic orientation. In general, firms with higher relative R&D spending appear to have superior performance than do firms with lower R&D intensity, confirming the expectation that technological know-how is an integral part of a firm’s FSAs. However, the difference is significant only for global firms, whereas regional MNEs do not seem to reap the full benefits of their R&D investments. Furthermore, the performance differential between regional and global firms, while insignificant in all subgroupings in this study, is weakly significant statistically (i.e., po0.1) for high R&D intensity firms. Hence, for firms with strong technological FSAs there seems to be a performance benefit associated with global orientation. Table 2. Archetype

Regional MNEs Global MNEs F statistic po0.05. po0.01. po0.001.

Mean

0.018 0.031 13.82

R&D Intensity (RDI). SD

0.037 0.021

Market Performance Low RDI

High RDI

0.02 0.05 0.03

0.11 0.41 3.81

F statistic

0.58 7.43

76


Taken together, this indicates that technological FSAs enable or support a higher geographic reach and may be an important driver of globalization at the firm level. Indeed, of the nine global companies identified by Rugman and Verbeke (2004a), seven of them are operating in high technology markets such as computers, electronics, or telecommunications equipment.3 Arguably, the most important proprietary assets in these industries are relatively nonlocation bound, since the individual products (e.g., MP3 players or LCD screens) share the same underlying technology yet are still capable of being modified to suit the local standards and preferences of customers in various foreign markets. Therefore, if an important part of the firm’s proprietary assets are technological in nature we can expect the reach of its FSAs to be generally high. This also explains why, as shown in Table 2, global firms have significantly higher R&D intensity than do regional firms. Marketing Assets Table 3 shows the interactive effect of marketing intensity and geographic orientation. First and foremost, this analysis suggests that marketing assets, similar to technological assets, are indeed beneficial as high advertising intensive firms have superior performance than do firms with lower relative spending marketing. However, the difference here is significant only for regional firms. Hence, marketing FSAs are apparently important determinants of the performance of regional firms but less important in discriminating between high and low performing global firms. This may reflect the relative difficulty of transferring downstream compared to upstream assets across national boundaries, supporting Anand and Delios’ (1997) earlier findings. Unlike technology, brand equity is closely associated with normative, cognitive, and affective processes in the minds of the customer. When Table 3. Advertising Intensity (ADI). Archetype

Regional MNEs Global MNEs F statistic po0.05. po0.01. po0.001.

Mean

0.007 0.015 16.56

SD

0.013 0.025

Market Performance Low ADI

High ADI

0.19 0.00 1.72

0.15 0.23 0.29

F statistic

5.62 2.19


77

cultural differences exist it may therefore be more difficult to exploit this type of asset in other countries. Consistent with this argument, Rugman and Verbeke (2004a) argue that regional barriers to expansion are higher in downstream activities than in upstream activities. Hence, the marketing expenditures of an MNE are likely to provide the greatest payoffs within the home region. Even though marketing FSAs do not seem to have as clear an association with global firms’ performance, the average marketing intensity of global firms’ is actually significantly higher than that of the regional firms. While this paradox is not easily explained, we offer two suggestions as to why this may be the case. First, there could be diminishing returns to brand investments. For regional firms operating with low levels of brand equity, advertising may be an effective way to differentiate oneself from the competition and therefore an enhancer of economic performance. Conversely, such a strategy may be less effective for global firms, which appear to have stronger brand assets at the outset. This suggests that acquiring a certain level of marketing assets is simply a cost of playing the game for global firms – without such assets global orientation may not be feasible – but once acquired, further investments in marketing assets yield low returns. A second explanation (which questions the direction of causality between intangible assets and performance) is that the success of global firms (e.g., Caterpillar, IBM, Sony), based on their technological strength, enables them to invest more heavily in their brands. This explanation is compatible with the signaling theory of advertising (Kihlstrom & Riordan, 1984) and could also explain the positive correlation between R&D intensity and advertising intensity as it implies a complementarity between the two types of assets. Expressed simply, global firms – as well as strong regional firms – may signal their superior technological capabilities to the market by investing heavily in advertising. Still, we can say relatively little more about this given that our data is cross-sectional and further research into this paradox, therefore, is warranted.

INTERNATIONAL EXPERIENCE, GEOGRAPHIC ORIENTATION, AND PERFORMANCE Number of Foreign Subsidiaries Table 4 shows the impact of geographic orientation and the firm experience in international markets as measured by the number of foreign subsidiaries.

78


Table 4. Archetype

Regional MNEs Global MNEs F statistic

Mean

35.7 22.3 5.68

Number of Subsidiaries (NS). SD

80.8 17.3

Market Performance Low NS

High NS

0.12 0.04 1.08

0.05 0.14 2.78

F statistic

0.29 0.53

po0.05. po0.01. po0.001.

We can see here that regional firms have larger foreign operations than do global firms. Regional firms, for example, have on average 35.7 subsidiaries compared to only 22.3 subsidiaries for their global counterparts. This is surprising as the organizational complexity argument suggests the opposite; we have argued, with reference to the process theory of internationalization, that firms with large international experience would find it more viable to pursue a global strategy. However, it may also be the case that firms who choose a regional profile for their internationalization strategy subsequently find it easier to expand their foreign operations because they encounter a lower liability of foreignness and learn faster about the environments of their new markets. This explanation, in which international experience succeeds rather than precedes geographic orientation, would account for the observed difference between regional and global firms. However, the difference is only weakly significant statistically and, when we look at the performance effects, there is actually no statistical difference between the four groups.

Foreign Capital Invested Finally, in Table 5 we report the results for international experience as measured by the extent of the firms’ foreign versus total capital invested. The results reflect those for size of international operations, except here none of the group differences are significant statistically. This indicates that experience and home region orientation are two independent dimensions of geographic orientation and that the choice of whether to pursue a regional or a global strategy is not contingent of the MNE’s degree of internationalization.


Table 5. Archetype

Regional MNEs Global MNEs F statistic

79

FDI as Share of Capital Invested (FT).

Mean

0.139 0.141 0.02

SD

0.137 0.148

Market Performance Low FT

High FT

0.09 0.10 1.96

0.08 0.11 1.86

F statistic

0.00 0.00

po0.05. po0.01. po0.001.

DISCUSSION It is well established that proprietary assets underpin an MNE’s capability to overcome the added costs of cross-border activities (Hymer, 1976). A corollary to this stylized fact is that they need stronger proprietary capabilities as they expand into areas where these costs are higher. Globally oriented firms that distribute their FDI across a variety of far-flung locations face a liability of foreignness potentially more serious than regional firms do, and would therefore have a larger incentive to invest in FSAs. First and foremost, the analyses presented in this study suggest that the intangible assets of global firms are much stronger than those of regional MNEs. We believe this may reflect the high entry costs in the global marketplace, caused by the incremental liability of foreignness encountered as soon as the firm ventures outside its home region. Our study also supports the idea that the performance returns to marketing and technological capabilities depend on whether the MNE pursues a regional or a global strategy. We have seen that the returns to R&D are highest for global firms, and the returns to advertising are highest for regional firms. Interestingly, our results contradict those of Rugman and Sukpanich (2006), who find that the returns to R&D are higher for regional firms than they are for global firms, and that advertising returns are the same. Some of these differences may be explained by the different measurement approaches of the two studies. Their definition of home region orientation is based on sales while our definition is based on FDI, and they include the home country in intra-regional sales while we focus on the distribution of foreign operations between home and host regions to control for home country orientation. Also, Rugman and Sukpanich (2006) use accounting data to assess performance while we use stock market-based measures. Taken together,

80


however, the two studies strongly suggest that regional orientation is an important moderator in the relationship between intangibles assets and economic performance. We believe that careful studies investigating this issue with different samples and methodologies would be beneficial. While on the one hand intangible assets clearly matter, we find on the other that international experience has surprisingly little, if any, effect on home region orientation and performance. On average, regional firms have slightly more subsidiaries than do global firms, but this seems not to reflect any performance differential. Furthermore, multinationality seems to be completely unrelated to home region orientation. In combination, these two analyses tell us that international experience is not an important contingency determining the feasibility of a global strategy. Even firms with limited experience in foreign markets can apparently pursue a global strategy without suffering a performance penalty – if anything, doing so increases their performance. This tells us that the relationship between home region orientation and performance is not moderated by the size of the foreign operations. These findings have implications for the process theory of internationalization (Johanson & Vahlne, 1977; Johanson & Wiedersheim-Paul, 1975), which suggests that MNEs should pursue initial international expansion among countries with low psychic distance which, in most cases, would translate to countries within their home region. Our findings suggest that some firms in fact skip this initial stage and pursue a global strategy even with very little international experience. Our results also have implications for the empirical research into the relationship between internationalization and performance. In particular, it suggests that previous studies may have paid too much attention to the degree of internationalization and too little attention to firms’ FDI patterns. Some measures used in this literature, such as FDI as share of total capital invested, capture only the degree of internationalization without distinguishing between different internationalization patterns. Other measures, such as entropy scores, contain information on both the degree and pattern of internationalization and therefore confound these two dimensions. By separating out the pattern of internationalization we have shown that where the firm’s FDI is matters as much as, if not more than, how much of it there is.

CONCLUSION The extant theory of regional multinationals is essentially a deterministic framework that does not explain the observed heterogeneity in geographic


81

orientation among firms. For example, we are left with the empirical question of whether the analysis of the global MNEs in the Fortune 500 are temporal anomalies – observed at a given point in time to have often overextended their global reach for fear of being left behind – or whether they are inherently different from other firms in a way that enables them to take advantage of the forces of economic globalization. This study lends support to the latter view, by uncovering the role of FSAs as a prerequisite of globalization and a driver of MNE performance.

NOTES 1. Note that a firm can have a low degree of internationalization (FDI share of total capital invested) but still be globally oriented (have a low ratio of FDI in the Asia-Pacific region) if most of its limited international investment is outside its home region. 2. As Tables 1–4 show, global firms consistently outperform regional firms within all subgroups, a finding which is consistent with the previously discussed negative correlation between home region orientation and performance. Yet the difference is not significant here, which may be because the ANOVA loses half of the observations and hence does not have sufficient power to detect it. 3. The seven companies are IBM, Sony, Philips, Nokia, Intel, Canon, and Flextronics. The remaining two global companies are Coca-Cola and LVMH.

REFERENCES Aharoni, Y. (2006). The regional multinationals MNEs and global strategic management [Review of the book]. International Business Review, 15, 439–446. Anand, J., & Delios, A. (1997). Location specificity and the transferability of downstream assets to foreign subsidiaries. Journal of International Business Studies, 28, 579–603. Buckley, P., & Casson, M. (1976). The future of the multinational enterprise. London: Macmillan Press. Caves, R. (1996). Multinational enterprise and economic analysis (2nd ed.). Cambridge, UK: Cambridge University Press. Daiwa Institute of Research. (1999). Analysts’ guide. Tokyo, Japan: Daiwa Institute. Delios, A., & Beamish, P. (1999). Geographic scope, product diversification, and the corporate performance of Japanese firms. Strategic Management Journal, 20, 711–727. Delios, A., & Beamish, P. (2005). Regional and global strategies of Japanese firms. Management International Review, 45(Special Issue 1), 19–36. Farjoun, M. (1998). The independent and joint effects of the skill and physical bases of relatedness in diversification. Strategic Management Journal, 19, 611–630. Ghemawat, P. (2003). Semiglobalization and international business strategy. Journal of International Business Studies, 34, 138–152. Goerzen, A. (2001). Network diversity and multinational enterprise performance. Unpublished dissertation. University of Western Ontario, London, ON, Canada.

82


Goerzen, A. (2005a). Managing alliance networks: Emerging practices of multinational corporations. Academy of Management Executive, 19, 94–107. Goerzen, A. (2005b). Networks and location: Organizing the diversified multinational corporation for value creation. New York: Palgrave Macmillan. Goerzen, A., & Beamish, P. (2003). Geographic scope and multinational enterprise performance. Strategic Management Journal, 24, 1289–1306. Goerzen, A., & Beamish, P. (2005). The effect of alliance network diversity on multinational enterprise performance. Strategic Management Journal, 26, 333–354. Govindarajan, V., & Gupta, A. (2001). The quest for global dominance. San Francisco, CA: Jossey-Bass/Wiley. Hennart, J.-F. (1982). Theory of multinational enterprise. Ann Arbor, MI: University of Michigan Press. Hitt, M., Hoskisson, R., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management Journal, 40, 767–798. Hymer, S. (1976). The international operations of national firms: A study of direct foreign investment. Cambridge, MA: The MIT Press. Jeannet, J. (2000). Managing with a global mindset. London: Financial Times Prentice Hall Pearson. Jensen, M. (1968). The performance of mutual funds in the period, 1945–1964. Journal of Finance, 23, 389–416. Johanson, J., & Wiedersheim-Paul, F. (1975). The internationalization of the firm: Four Swedish cases. Journal of Management Studies, 12(3), 305–322. Johanson, J., & Vahlne, J.-E. (1977). The internationalization process of the firm: A model of knowledge development and increasing foreign market commitments. Journal of International Business Studies, 8(1), 23–32. Johanson, J., & Vahlne, J.-E. (1990). The mechanism of internationalisation. International Marketing Review, 7(4), 11–24. Kihlstrom, R., & Riordan, M. (1984). Advertising as a signal. The Journal of Political, 92, 427–450. Kogut, B., & Chang, S. (1991). Technological capabilities and Japanese foreign direct investment in the United States. Review of Economics and Statistics, 73, 401–413. Meyer, M. (1994). Measuring performance in economic organizations. In: N. Smeker & R. Swedberg (Eds), Handbook of economic sociology (pp. 556–578). Princeton, NJ: Princeton University Press. Morck, R., & Yeung, B. (1991). Why investors value multinationality? Journal of Business, 64, 165–187. Nayyar, P. (1993). Performance effects of information asymmetry and economies of scope in diversified service firms. Academy of Management Journal, 36, 28–57. Pennings, J., Barkema, H., & Douma, S. (1994). Organizational learning and diversification. Academy of Management Journal, 37, 608–640. Rugman, A. M. (1981). Inside the multinationals: The economics of internal markets. London: Croom Helm. Rugman, A. M. (2000). The end of globalization. New York: AMACOM-McGraw Hill. Rugman, A. M. (2001). The end of globalization: Why global strategy is a myth & how to profit from the realities of regional markets. New York: AMACOM-McGraw Hill.


83

Rugman, A. M. (2005). The regional multinationals. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Brain, C. (2003). Multinational enterprises are regional, not global. Multinational Business Review, 11(1), 3–12. Rugman, A. M., & Girod, S. (2003). Retail multinationals and globalization. European Management Journal, 21, 24–37. Rugman, A. M., & Hodgetts, R. (2001). The end of global strategy. European Management Journal, 19, 333–343. Rugman, A. M., & Sukpanich, N. (2006). Firm-specific advantages, intra-regional sales and performance of multinational enterprises. The International Trade Journal, 20, 355–382. Rugman, A. M., & Verbeke, A. (2004a). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35, 3–18. Rugman, A. M., & Verbeke, A. (2004b). Towards a theory of regional multinationals: A transaction cost economics approach. Management International Review, 44(Special Issue), 3–15. Sharpe, W. (1966). Mutual fund performance. Journal of Business, 39, 119–138. Stopford, J., & Wells, L. (1972). Managing the multinational enterprise. New York: Basic Books. Toyo Keizai. (1999). Kaigai Shinshutsu Kigyou Souran-Kuni Betsu. Tokyo: Toyo Keizai Ltd. UNCTAD. (2000). World investment report, 2000. New York: United Nations. Yip, G. (2002). Total global strategy. Upper Saddle River, NJ: Prentice-Hall. Zaheer, S. (1995). Overcoming the liability of foreignness. Academy of Management Journal, 38, 341–363.

THE REGIONAL NATURE OF MNE ACTIVITIES AND THE GRAVITY MODEL Walid Hejazi ABSTRACT It has been demonstrated by Rugman and his colleagues that a majority of the activities undertaken by the world’s largest 500 MNEs, such as sales, assets, and employment, are regional in nature. This evidence has also been extended to trade and FDI patterns of OECD countries. Given the costs associated with doing business in foreign and distant markets, one may expect there to be a regional concentration in such activities. That is, the concentration of MNE activities in regional markets may be consistent with a transactions cost model. The objective of the analysis undertaken in this paper is to measure the extent to which the concentrations of OECD MNE activities can be explained by a formal transactions costs model (the gravity model in this case). These results are important for two main reasons. To the extent the concentrations are consistent with a formal model, then, first, this would provide further theoretical arguments in support of Rugman’s hypotheses, and second, this would indicate that MNE managers have it right – that is, the activities of the corporations they manage are as global as they should be. On the other hand, if the activities are not fully explainable by a transactions cost model, the implications would be quite different. Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 85–109 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13005-9

85

86

WALID HEJAZI

The results indicate that although some activities can be explained by a gravity model, many dimensions of OECD MNE activities, especially within the EU, are not explainable using a gravity model. That is, many of the activities of EU MNEs are more regionally concentrated than would be predicted by transactions costs.

1. INTRODUCTION Rugman and his colleagues have demonstrated that the activities of the world’s largest 500 MNEs are predominantly regional in nature (Rugman, 2005; Rugman & Verbeke, 2004; Rugman & Moore, 2003; Rugman & Girod, 2003; Rugman & Brain, 2003; Rugman, 2000). This result is in sharp contrast to the perceptions held by many policy makers, academics, and the popular press that the activities of the world’s largest MNEs are global in nature. In a recent paper, Hejazi (2007) uses data on all US MNEs over the period from 1980 to 2000 to establish that in fact regional biases in US MNE activities are consistent with a gravity model – that is, given the costs associated with doing business in distant and foreign markets, US MNE managers allocate a large share of their activities/assets nationally and regionally. That is, the hypothesis put forth by Rugman on the regional nature of MNE activities is consistent with a transactions cost interpretation. The evidence therefore indicates that, in the process of maximizing profits, managers of US MNEs have allocated their activities properly when it comes to Asia and Europe. A puzzle that remains is that there is ‘‘too much’’ US MNE activity within the US market itself than can be explained by the model. It should be noted also that there are similar home biases in many dimensions of economics, including where savings are invested (the Feldstein–Horioka puzzle) (Feldstein & Horioka, 1980), the holding of domestic equities (French & Porterba, 1991), and the preference for goods produced in the domestic market (the Armington bias) (Blonigen & Wilson, 1999). This evidence indicates therefore that in the case of the US, the regional bias in US MNE activities are driven by the national (or home) bias. Given the size of the US economy relative to that of Canada and Mexico, it is unclear whether the behavior of US MNEs with respect to North America and other regions would generalize to the behavior of European MNEs. This is the objective of the current paper. Using data on the activities of several OECD MNEs, the paper documents the regional

The Regional Nature of MNE Activities and the Gravity Model

87

character of that activity. The paper then uses a gravity model to test whether the regional distributions seen in the activities of OECD MNEs are consistent with transactions costs, or whether those activities are more regional than is consistent with predictions of the model. It has also been shown that both North American and EU trade flows are more regional than would be predicted by a gravity model, but there are differences in the patterns for FDI. More specifically, the evidence indicates that intra-EU FDI patterns are far more regional than would be predicted by a gravity model, whereas there is less intra-regional FDI locating in North America than would be predicted by a gravity model (Hejazi, 2005). It is important to point out that in looking at trade and FDI data, the domestic dimension is not considered, and this is an important point to make. In looking at trade and FDI patterns, there are no national equivalents to consider (or at least such data are not available). This paper uses data from the OECD publication entitled Measuring Globalization: Volume 1 (Manufacturing). This publication provides data on production, employment, turnover, exports, imports, as well as intra-firm exports and imports, for several OECD countries over the period 1994– 1999, although the sample coverage varies by country and variable. The data used for this study will focus on the inward activities into each host OECD economy by foreign MNEs from other OECD counties. That is, for each OECD country, we have the amount of activities undertaken by foreign MNES from each other OECD country. This paper documents patterns in the regional and global distribution of these activities, and measures the extent to which these patterns are consistent with a gravity model. The data description below shows clearly that there is a regional bias in the activities of OECD MNEs – that is, there is far more activity undertaken by MNEs in their home region than would be predicted from a simple benchmarking exercise. However, we need to test whether these concentrations are explainable by a transactions cost model. The results indicate that there is significantly more foreign MNE employment, turnover and exports between EU countries than can be predicted by a gravity model. That is, each EU country has more of these foreign activities locating inside its country originating from other EU countries than a transactions cost model would predict. For some other dimensions of MNE activity, namely imports, and intra-firm exports and imports, there does not seem to be a regional bias. Together this evidence indicates that Rugman’s observation on the regional nature of EU MNE activities survives the estimation of a formal

88

WALID HEJAZI

transactions cost model. In the case of North America, this regional distribution is driven by the national dimension – that is, the large share of US MNE activities inside the US drives the regional concentrations seen in North American MNE activities. However, as the evidence below indicates, this is not the case for the EU. Specifically, for several dimensions of MNE activity, there is more intra-EU MNE activity than can be explained by a transactions cost model. Although the analysis here does not account for the activities by MNEs in their home market, that would not affect these results. There is more intra-EU activity than can be explained by a gravity model, and this provides further support for Rugman’s hypothesis. The format of this paper is as follows. Section 2 discusses the hypotheses to be tested in this paper. Section 3 reviews the regional distribution of OECD MNE activities. Section 4 discusses the gravity model. Section 5 presents the empirical evidence, and Section 6 concludes.

2. HYPOTHESIS DEVELOPMENT Multinationals are firms that by definition do business across international borders. As such, the result indicating that most multinationals are not global in nature, that is their activities are significantly concentrated in their home region, may seem like a paradox. To the contrary, however, it is expected that there will be, to some extent, a concentration of MNE activity not only in the home country, but also in the home region. This section of the paper will discuss the theory underlying this prediction, and its application to OECD MNE activities. In a world where there are no transportation costs or other costs associated with doing business in distant and foreign markets, then one may expect that MNE activities will be distributed globally in proportion to GDP. Therefore, given that North America is 26% of world GDP (in the year 2000), then one may expect MNEs to concentrate about 26% of their activities in the North American market. Similarly, Europe is about 23% of world GDP and the Asia Pacific is about 20%, and hence these regions should receive about 23% and 20% of MNE activities proportionately (see Fig. 1). The evidence presented by Rugman and his colleagues clearly indicates that the distributions of MNE activities are not so clearly explained by GDP. This evidence is predominantly based on the activities of the world’s largest 500 MNEs, but has also been shown to be the case for patterns of trade and FDI.


89

North America GDP share: 26.34% Share of U.S. MNE Sales: 74.56% Net Income: 69.93% Assets: 73.68% Employment: 77.38% Employee Compensation: 82.36%

Europe

Asia Pacific

GDP share: 23.39%

GDP share: 20.20%

Share of U.S. MNE Sales: 15.12% Net Income: 18.60% Assets: 17.11% Employment: 12.00% Employee Compensation: 11.54%

Share of U.S. MNE Sales: 5.64% Net Income: 5.03% Assets: 4.10% Employment: 5.00% Employee Compensation: 3.10%

Fig. 1. The Regional Distribution of U.S. MNE Activities. Source: Reproduced from Hejazi (2007). Note: These GDP shares are calculated against the 120 countries that were used in the regression analysis.

Given the extent of the US data, I will discuss patterns of US MNE activities to motivate the hypotheses to be tested below. US MNEs have a much larger concentration of their activities in North America than would be predicted by simply benchmarking off of GDP. The concentration of US MNE activity in North America is something on the order of three times the GDP share (Fig. 1). More specifically, although North America makes up about 26% of world GDP, 74.56% of US MNE Sales are in North America, 69.93% of Net Income, 73.68% of Assets, 77.38% of Employment, and 82.36% of Employee Compensation. There are many economic reasons to expect the activities of US MNEs to be concentrated inside North America. There are significant costs associated with operating in foreign and distant markets. Therefore, in addition to GDP, other factors that must be factored into the explanation of the distribution of MNE activity would include distance between economies, exchange rates, language similarities, free trade agreements, as well as institutional differences/similarities. Using this Gravity Model framework, Hejazi (2007) demonstrates that US MNE mangers have in fact distributed their activities in each of these regions according to transactions cost

90

WALID HEJAZI

considerations. That is, the amount of US MNE activity locating in Europe and the Asia Pacific are consistent with the amounts that would be predicted, given the costs associated with US MNEs operating in these foreign and distant markets. The results of Hejazi (2007) also indicate that much of the home-region bias seen in the US MNE data are in fact driven by the home bias – that is, although three-quarters of US MNE activity is located in North America, most of this is inside the US proper. That is, the regional concentration documented is in fact a national bias. Furthermore, the amount of US MNE activity inside the home region (North America) but outside the US is in fact consistent with the predictions of a gravity model. This raises some important unanswered questions. The US is a large market, especially in relation to the size of North America. That is, Canada and Mexico are quite small relative to the United States. In sharp contrast, the relative sizes of countries within Europe are far less asymmetric. Therefore, although the MNEs of other countries have been documented to have large home-region concentrations, would the extent of the concentration of these activities be consistent with a transactions cost model such as the gravity model? That is, would the activities of Canadian, European or Asian MNEs be similar to those of US MNEs? This is the major question addressed in this paper. There is one important difference between the analysis described in Hejazi (2007) and the current study. The data on US MNEs used in Hejazi (2007) considers the distribution of US MNE activity across 50 countries, where the US itself is used as a location – that is, in addition to the foreign activities of US MNEs, that analysis also takes into account the activities undertaken by US MNEs inside the US itself. In contrast the data available for OECD MNEs does not include activities undertaken by MNEs in their home country. As such, we are unable to test whether the regional distribution of MNE activities for OECD MNEs are driven by the national dimension. The data used in this analysis relate to the activities by foreign MNEs locating inside each OECD country, and whether the distribution of these activities is consistent with the predictions of a gravity model. This gives rise to the hypotheses to be tested in this paper. Hypothesis 1. The foreign MNE activities locating inside each OECD country should be more regionally concentrated than would be predicted by only considering GDP. That is, the regional share of an MNEs activities located inside its home region should be higher than that region’s share of world GDP.


91

Hypothesis 2. Once the costs associated with operating in foreign and distant markets are taken into account, MNE activities should not be regionally concentrated. That is, MNE activities should be consistent with the transactions costs associated with operating across markets. The first hypothesis indicates that we expect, for example, EU host countries to receive a share of their foreign activities from other EU countries that is larger than the EU’s share in world GDP. Using France for example, it has a given amount of foreign MNE activity. These activities are undertaken by EU, North American, and other MNEs. The question addressed is, given that the EU is about 23% of the world’s GDP, would the share of foreign activity inside France by other EU countries be higher than 23%. If the answer to this question is yes, then that would be evidence in favour of Hypothesis 1. Hypothesis 2, on the other hand, relates to whether that concentration of EU activity inside France is consistent with predictions of the gravity model. Given that France is inside the EU, and hence has proximity to other EU countries, has a free trade agreement, and many other factors that would explain a high degree of integration, then the question becomes whether these regional concentrations can be explained by these transactions-cost considerations. These hypotheses are tested below. First, the regional distribution of foreign activity locating inside OECD countries is documented. Second, a gravity model is used to test whether these concentrations can be explained by a transactions-cost motivation.

3. DATA DESCRIPTION: THE REGIONAL DISTRIBUTION OF OECD MNE ACTIVITY To test these hypotheses, data are used on the foreign activities of OECD MNEs. That is, we consider the distribution of MNE activities operating in each foreign market for which data are available. The test will be whether these activities are distributed as would be predicted by a transactions cost model or, alternatively, are these activities more regionally concentrated than would be predicted by a gravity model. From the OECD publication, ‘‘Measuring Globalization: The Role of Multinationals in the OECD Economies’’, Volume I (Manufacturing), we have assembled data on the variables listed in Table 1. These data are available for OECD MNEs, on a bilateral basis, into many countries. The availability of these data vary for each country by both activities and by years. The tests

92

WALID HEJAZI

Table 1. Inward Inward Inward Inward Inward Inward Inward

Inward MNE Activities for which Data are Available (Manufacturing).

Production by country of origin in the manufacturing sector Number of employees by country of origin in the manufacturing sector Turnover by country of origin in the manufacturing sector Total Exports by country of origin in the manufacturing sector Total Imports by country of origin in the manufacturing sector Intra-Firm Exports by country of origin in the manufacturing sector Intra-Firm Imports by country of origin in the manufacturing sector

Table 2.

Canada.

Inward Production by Country of Origin in the Manufacturing Sector

1994 1995 1996 1997 1998

US

EU

Europe

Asia (non-OECD)

71.47 73.22 72.27 70.82 72.75

18.86 17.03 17.95 19.30 17.49

20.65 18.68 19.55 20.73 19.42

0.80 1.08 1.11 1.11 1.19

undertaken therefore use all data that are available given the test undertaken. The focus here is on manufacturing and for inward MNE activity into each OECD country. The data for outward activity is far too sparse for analysis. Table 2 describes production in Canada by foreign MNEs. Over 70% of that production is undertaken by US MNEs, with about 20% being undertaken by European MNEs. Only about 1% takes place by non-OECD MNEs from Asia. This is certainly consistent with the view that the activities of MNEs are indeed regional. This result also points to an enormous asymmetry between Canada and the US. Although around 70% of US MNE activity locates inside the US, only about 3–7% of US MNE activity locates in Canada and Mexico combined. However, the data in Table 2 indicate that although the US is less reliant on Canada as a destination for its activity, that activity is enormously important for Canada. These data certainly confirm Hypothesis 1 for Canada: Canada receives a far larger share of its activity from North America relative to North America’s GDP share. The major objective of this paper, however, is to test whether the results documented for North America hold for European countries. There is far less asymmetry within Europe, and hence the relationships documented in North America may be a special case. We therefore review the regional


93

distribution of MNE activity for several major European countries and Japan. Information on the US is also considered. Table 3 provides data on the share of employees in each host economy by home region. This is done for seven European countries and the USA. In the case of Finland, in the year 1999, 51.17% of all employment inside Finland by foreign MNEs was by EU MNEs. Over 80% of that employment was by European MNEs, and less than 2% by non-OECD Asian MNEs. US MNEs employed 15.36% of all employees in Finland, working for foreign MNEs. The figures for Sweden are quite similar to that for Finland, whereas those for the Netherlands, France, Germany, Italy, and the UK are less dependent on the EU and more on the US. Nevertheless, the UK has the least dependence on EU MNE employment inside its borders, but even here, the share is 32.39%. What the results in this table indicate is that countries within Europe have a relatively large share of foreign employment inside their domestic economies undertaken by other EU MNEs. Table 3 also provides the distribution of foreign MNE employment within the US. Here, the reliance on the European MNEs is significantly higher than is the case for North American MNEs. More specifically, in 1998 less than 10% of foreign employment inside the US was done by Canadian and Mexican MNE, whereas 65% was undertaken by European MNEs and 58.86% by EU MNEs. It must be stated that together, Canada and Mexico make up less than 10% of world GDP, and hence the share of foreign employment within the US from other North American economies exceeds those other economies share in world GDP. But nevertheless, the reliance of the US on other North American MNEs is far lower than is the case for EU countries. This makes the US quite different than all the other countries discussed so far: Canada, like the EU countries, has a very large reliance on intra-regional MNEs for domestic employment. The US does not. We will see below that Japan is similar to the US in this respect, depending almost entirely on the US and Europe, and very little on other economies in its region. Table 4 consider turnover by foreign MNEs in each host economy listed, Table 5 considers total exports by foreign MNEs locating in each economy, Table 6 considers total imports, Table 7 intra-firm exports, and Table 8 intra-firm imports. Patterns of turnover by foreign MNEs locating inside each country are quite similar to those documented for employment: Sweden and Finland have a very high dependence on European MNEs and less of a dependence on US MNEs, although the dependence on the US for turnover is higher than was the case for employment. Trade intermediated by MNEs have somewhat different patterns (Tables 5, 6, 7, and 8). Exports (Table 5) from France by foreign MNEs are higher

Inward Number of Employees by Country of Origin in the Manufacturing Sector.

US

EU

Europe

Asia (non-OECD)

US

Finland 1995 1996 1997 1998 1999

11.15 15.87 18 16.92 15.36

31.63 31.3 31.7 30.58 29.66

46.25 46.82 45.41 46.53 51.17

81.91 76.82 78.53 80.68

NA NA NA NA 1.51

1995 1996 1997 1998

29.56 30.65 33.05 36.21

42.71 45.49 44.27 35.46 34.2

53.17 53.41 53.63 52.87 53.08

62.33 62.33 62.04 54.21 63.97

US

41.5 41.32 40.14 45.41 47.67 EU

0.32 0.46 0.67 0.19 0.09

1994 1995 1996 1997 1998

9.99 15.17 17.55 17.23 18.15

58.44 55.25 56.66

53.82 50.82 49.96 48.81

61.73 59.39 58.79 55.65

NA NA NA NA

60.46 55.5 52.49 54.47 54.02

87.99 83.24 80.91 80.01 79.16

NA NA 0.69 0.98 1.00

55.87 NA NA 61.22 63.21

0.61 NA 0.23 0.26 0.26

Europe

Asia (non-OECD)

1995 1996 1997 1998

44.97 46.54 44.51 45.39

28.98 32.83 33.54 32.39

NA 39.93 40.03 36.83

NA 0.56 0.76 0.73

Canada+Mexico

EU

Europe

Asia (non-OECD)

62.34 62.86 64.44 65.04 65.23

2.15 2.38 2.04 2.10 NA

USA 66.17 63.09 63.54

0.45 0.42 0.39

1994 1995 1996 1997 1998

16.44 13.96 12.49 9.62 8.93

50.27 55.06 56.84 57.3 58.86

WALID HEJAZI

29.63 31.84 30.72

Asia (non-OECD)

UK

Italy 1995 1997 1999

Europe

Sweden

Germany 1995 1996 1997 1998 1999

EU Netherlands

France 1994 1995 1996 1997 1998

94

Table 3.

US

EU

Europe

Asia (non-OECD)

US

Finland 1995 1996 1997 1998 1999

13.02 17.4 19.87 19.29 18.23

43.78 43.38 43.38 43.22 46.86

NA 77.53 71.37 72.5 74.75

36.55 35.3 35.56 34.07 31.95

49.2 49.86 49.69 49.77 51.86

58.02 58.38 57.85 50.82 61.19

NA NA NA NA 1.12

1995 1996 1997 1998

43.28 45.88 53.43 54.31

0.25 0.29 0.43 0.11 0.05

1994 1995 1996 1997 1998

10.64 15.02 17.27 18.52 19.64

52.02 54.89 52.21 39.99 35.8

39.94 39.37 37.59 43.64 46.93

49.67 NA NA 58.36 NA

0.26 0.04 0.11 0.16 0.17

US

EU

Europe

Asia (non-OECD)

35.82 35.84 33.31

52.88 51.28 54.28

38.49 34.19 31.93 33.27

49.18 45.35 39.19 38

NA NA NA NA

59.32 56.72 52.27 53.32 54.01

87.42 83.47 81.42 79.42 77.98

NA NA 0.48 0.26 1.01

UK 1995 1996 1997 1998

55.69 57.22 55.16 55.36

22.57 25.84 26.51 27.64

NA 31.17 31.51 22.13

NA 0.35 0.4 0.29

Canada+Mexico

EU

Europe

Asia (non-OECD)

62.35 61.65 63.15 63.3 66.49

2.02 2.22 2.04 1.92 NA

Italy 1995 1997 1999

Asia (non-OECD)

Sweden

Germany 1995 1996 1997 1998 1999

Europe

Netherlands

France 1994 1995 1996 1997 1998

EU


Table 4. Inward Turnover by Country of Origin in the Manufacturing Sector.

USA 60.24 59.31 61.47

0.24 0.29 0.17

14.10 9.01 8.84 7.64 6.88

49.86 53.32 55.08 55.98 60.46

95

1994 1995 1996 1997 1998

96

WALID HEJAZI

Table 5.

Inward Total Exports by Country of Origin in the Manufacturing Sector. US

EU

Europe

Asia (non-OECD)

NA NA NA 66.83 71.95

NA NA NA NA 1.53

52.03 51.88 50.92 45.09 55.42

0.35 0.41 0.72 0.2 0.1

NA 40.65 38.25 35.18

NA NA NA NA

87.43 81.51 79.53 79.64 80.59

NA NA 0.17 0.02 0.05

Finland 1995 1996 1997 1998 1999

NA NA NA 23.98 21.05

NA NA NA 33.43 37.95 France

1994 1995 1996 1997 1998

42.11 40.74 40.84 39.53 37.08

43.06 43.75 43.48 43.92 47.09 Netherlands

1995 1996 1997 1998

NA 52.25 54.17 56.88

NA 31.27 32.01 31.53 Sweden

1994 1995 1996 1997 1998

8.2 17.58 19.53 18.71 18.13

61.97 52.43 52.24 58.83 61.83

to Europe than to the US, but the difference is far less pronounced than for the activities discussed above. In fact for the Netherlands, the share going to the US is higher than that going to Europe. On the other hand, in the case of Sweden, as in the case of the other activities, Europe is far more important a destination for Swedish exports mediated by foreign MNEs than is the US. The results for imports (Table 6) are quite similar to those for exports. The data for intra-firm trade are most limited, and are reported in Tables 7 and 8. There, it is shown that again, for the Netherlands, the US is a more important destination for intra-firm trade (both exports and imports) than is Europe. In contrast, Sweden, as above, has a larger reliance on Europe than on the US.


97

Table 6. Inward Total Imports by Country of Origin in the Manufacturing Sector. US

EU

Europe

Asia (non-OECD)

NA NA NA 67.33 NA

NA NA NA NA NA

NA 38.57 34.1 33.67

NA NA NA NA

NA 83.52 83.48 79.55 77.23

NA NA 0.86 0.69 0.83

Finland 1995 1996 1997 1998 1999

NA NA NA 9.11 NA

NA NA NA 38.18 NA Netherlands

1995 1996 1997 1998

NA 53.37 58.6 59.55

NA 30.88 28.04 29.66 Sweden

1994 1995 1996 1997 1998

NA 14.88 15.43 18.04 20.61

NA 55.58 54.4 55.3 55.58

Table 7. Inward Intra-Firm Exports By Country of Origin in the Manufacturing Sector. US

EU

Europe

Asia (non-OECD)

NA NA 36.49 34.11

NA NA NA NA

89.49 79.72 77.93 76.76 78.21

NA NA NA 0.43 0.03

Netherlands 1995 1996 1997 1998

NA NA 56.96 60.67

NA NA 30.59 29.21 Sweden

1994 1995 1996 1997 1998

9.55 20.01 21.78 21.91 21.07

60.32 50.27 50.11 51.6 56.25

98

WALID HEJAZI

Table 8.

Inward Intra-Firm Imports By Country of Origin in the Manufacturing Sector. US

EU

Europe

Asia (non-OECD)

NA NA 29.06 34.04

NA NA NA NA

Netherlands 1995 1996 1997 1998

NA NA 65.21 62.78

NA NA 23.52 28.84

Finally, we turn to a discussion of activities locating inside Japan (Table 9). These data indicate the very strong reliance of Japan on the US for foreign activities locating inside its borders. In 1998, 69% of all foreign employment in Japan was undertaken by US MNEs, 75.87% of turnover, 86.7% of exports, 74.4% of imports, 84.4% of intra-firm exports, and 73% of imports – of course, these are shares of activities undertaken by foreign MNEs operating in Japan. The importance of Europe is of the order of a third or less of the US’s importance. The importance of MNEs from nonOECD Asian economies is less than 1%. Therefore, Japan is more like the US, having a much higher reliance on MNEs from other regions. What is interesting is that the US has a heavy reliance on European MNEs for foreign activities locating inside its market, whereas Japan has a heavy reliance on US MNEs. This discussion on the regional nature of foreign MNE activities locating in each of these economies highlights two very important insights. First, as Rugman has documented elsewhere, there is a large regional concentration/ bias in these activities. However, as a refinement, the evidence presented here is that these concentrations, with respect to inward activities, are more prevalent for EU countries and Canada, but are less prevalent for the US and Japan, and to a lesser extent the UK. For comparison sake, I have reproduced a table from Hejazi (2005) (Table 10), which considers the intra-regional export and outward FDI shares for several OECD countries. It should be noted that in the case of exports, this would include all exports, which are those that include domestic and foreign MNEs as well as non-MNEs in each country. In contrast, the outward FDI would reflect the activities of just MNEs, but as in the case of exports, may include both domestic and foreign MNEs in each economy. As in the data described above on the activities of OECD MNEs locating in other OECD countries, there is a large regional concentration in these

US

EU

Europe

Asia (non-OECD)

Inward Number of Employees by Country of Origin in the Manufacturing Sector. 1994 1995 1996 1997 1998

72.59 65.23 64.83 63.95 69.40

17.86 19.08 18.73 22.57 21.73

NA 27.56 29.82 28.63

Japan.

0.41 0.48 0.46 0.19 0.21

US

71.09 66.81 68.90 71.37 75.87

21.95 15.88 16.33 18.16 17.61

NA 22.06 21.87 22.21

1.18 0.89 1.40 0.10 0.13

1994 1995 1996 1997 1998

76.86 71.91 74.08 76.33 86.70

11.11 11.51 10.12 10.89 9.68

15.18 15.32 11.88

6.24 4.40 5.06 0.38 0.16

Asia (non-OECD)

58.31 67.37 68.71 66.21 74.43

31.54 15.15 14.13 15.24 19.51

23.20 18.37 24.90

2.70 1.48 2.13 0.11 0.47

Inward Intra-Firm Exports By Country of Origin in the Manufacturing Sector 1994 1995 1996 1997 1998

Inward Total Exports by Country of Origin in the Manufacturing Sector 1994 1995 1996 1997 1998

Europe

Inward Total Imports by Country of Origin in the Manufacturing Sector

Inward Turnover by Country of Origin in the Manufacturing Sector

1994 1995 1996 1997 1998

EU

72.73 73.35 42.21 75.21 84.48

10.87 11.18 25.32 15.66 11.64

16.52 16.71 39.56 18.93 14.77

10.31 0.21 0.62 0.30 0.16

Inward Intra-Firm Imports By Country of Origin in the Manufacturing Sector 1994 1995 1996 1997 1998

56.94 70.77 68.95 77.68 73.34

31.37 14.25 14.71 14.83 17.50

39.03 24.73 17.56 18.67 25.66


Table 9.

3.62 0.32 0.60 0.15 0.67

99

100

WALID HEJAZI

Table 10.

Regional Export and Outward FDI Shares for OECD Countries. Regional Export Shares

Austria Bel Lux Denmark Finland France Germany Greece Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland UK Canada USA

Regional Outward FDI Shares

1980

1985

1990

1995

1998

1982

1985

1990

1995

2000

60 77 66 58 57 60 51 81 55 74 48 66 51 59 61 47 69 24

60 74 56 51 55 58 55 72 52 75 64 69 53 55 59 50 85 29

68 79 64 61 64 63 68 78 63 77 72 81 71 63 64 54 81 29

63 74 62 56 63 56 59 72 57 73 70 81 70 58 62 55 82 30

64 75 60 55 62 55 52 67 56 68 70 82 70 57 62 58 87 35

NA NA NA NA NA 35 NA NA NA NA NA NA NA NA NA NA 68 23

NA NA NA NA NA 57 NA NA 78 44 NA NA NA 71 64 21 70 23

76 NA NA 75 NA 63 NA NA 64 51 82 NA NA 67 52 28 61 20

49 NA NA 51 NA 57 NA NA 70 54 85 57 NA 61 48 47 53 16

44 NA 54 58 NA 55 17 67 67 54 73 73 NA 56 46 53 51 15

Source: Reproduced from Hejazi (2005).

data. Hejazi (2005) demonstrates that there is more intra-regional trade than can be predicted by a transactions cost model, and this is strongly the case in both North America and Europe. In the case of FDI, Hejazi finds that although there is more regional FDI between EU countries, this is not the case for North American MNEs. The objective of the current study is to test whether the activities described above on the inward activities into each OECD country are explainable, using a transaction cost model, or whether there is ‘‘too much’’ regional activity, a task to which we now turn.

4. THE GRAVITY MODEL The discussion above, confirms the notion that MNE activity locating inside OECD countries have a relatively large participation by MNEs from the


101

local region. That is, regional MNEs have a larger share of their activities in their own region relative to the region’s share of world GDP. The important question that must be asked is whether these concentrations are consistent with economic theory. Given the costs associated with doing business in distant and unfamiliar markets, with different languages, institutions, and the costs associated with working outside local business networks, as well as explicit and implicit restrictions on trade and FDI, it is expected there would be more activity inside the home region. The objective of this paper is to use one benchmark, namely that predicted by the gravity model, as a test of whether the observed regional concentrations of MNE activities locating in OECD host economies are consistent with economic theory. The idea underlying the gravity model for trade is that, countries of similar size and per capita GDP have similar needs both in terms of intermediate inputs (Ethier, 1982) and consumption patterns. Also, two countries’ trade should be positively related to these countries’ incomes, and countries that are close together and have similar languages will have smaller transactions costs of doing business and correspondingly larger levels of bilateral trade. Trade flows are also sensitive to movements in the exchange rate. Dummy variables are included for several regional groupings, and measure persistent patterns of trade within regional areas, which are not captured by the gravity variables. It is well-known that the gravity model explains trade flows well, but what is relatively less well-known is that there are theoretical foundations for the gravity equation. The early contributions to these theoretical developments include Bergstrand (1985, 1989, 1990); Leamer (1974); Anderson (1979). Helpman (1987) interpreted the success of the gravity model as evidence in favor of the monopolistic competition model. This was based on the belief that the gravity model was consistent with that model and not with the Hechscher–Ohlin model. However, Deardorff (1998) established that the gravity model is indeed consistent with both the Hechscher–Ohlin and monopolistic competition models of international trade, but his result was restricted to a bilateral world. More recently, the approach of Head and Ries (2004) have provided additional theoretical underpinnings for a gravity-like model. The gravity model has been used to explain bilateral trade flows among large groups of countries and over long periods of time (Feenstra, Markusen, & Rose, 2001; Hejazi & Trefler, 1996; Frankel, Stein, & Wei, 1995). The gravity model has also been used to explain patterns of the FDI (Brainard, 1997; Grosse & Trevino, 1996; Grubert & Mutti, 1991; Hejazi & Safarian, 2005; Lipsey & Weiss, 1981, 1984; Stein & Duade, 2001).

102

WALID HEJAZI

In contrast to that for trade, the gravity model for the FDI has not been given theoretical foundations. As such, much of the empirical evidence must be qualified. That is, without a theoretical model underlying its derivation, the gravity model applied to the FDI is a reduced form analysis whose results must be interpreted carefully. As in the case of trade, however, ongoing research has been developing empirical models for the FDI that do have theoretical foundations, although these tend to be quite difficult to work with at present (Markusen, 2002). Following the international business literature, we use the gravity model to explain patterns of MNE activities, although unlike other studies, the model will be used to explain several dimensions of the foreign MNE activities locating inside OECD countries. The estimating equation for MNE activities is written as follows: lnðActivityijt Þ ¼ a0 þ a1 lnðGDPHOMEit Þ þ a2 lnðGDPHOSTjt Þ þ a3 lnðDISTANCEij Þ þ a4 lnðXRATEijt Þ þ a5 ðLANGUAGEij Þ þ a6 ðADJACENCYij Þ þ a7 ðNAi Þ þ a8 ðEUi Þ þ eijt

ð1Þ

where ln is the natural logarithm. Activitiesijt represent the activities undertaken in a host economy by foreign MNEs, and represent the seven activities listed in Table 1. The index i indicates the host economy, and the index j indicates the home country, and t indexes year. The sample covers 28 OECD countries, both as home and host, and covers the sample 1994–1999, although this sample coverage also varies across countries. GDP measures real Gross Domestic Product of the home and host countries, respectively. Distance is a measure of the physical distance between countries. XRATE is the purchasing power parity exchange rate between countries i and j in year t. The standard gravity model is extended to include regional dummies for North America (NA) and the European Union (EU). These dummy variables pick up any persistent deviations between the model’s predictions and MNE activities trade with each region. In Eq. (1), a7 and a8 capture whether countries within North America and the European Union, respectively, have more foreign MNE activity than would be predicted by the gravity model. If the estimated values of a7 and a8 equal zero, the amount of foreign MNE activity observed is consistent with predictions of the model. On the other hand, if the estimated values are


103

positive, these regions have more foreign MNE activity than the benchmark as predicted by the gravity model. The hypothesis to be tested in this paper is whether the foreign MNE activities locating in OECD countries are consistent with this transactions costs model, or whether there is more of such activity from other regional partners than the model would predict. To test this hypothesis, two additional dummy variables are created to capture when two trading partners are within the same region. BNA is set equal to one when both trading partners are in North America, and zero otherwise. (This is similar to the strategy used by McCallum (1995) in showing that trade between the provinces in Canada is far greater than trade between them and states in the US.) Similarly, BEU is set equal to one when both countries are in the European Union, and zero otherwise. The above regression is then extended to include these dummies. lnðActivityijt Þ ¼ a0 þ a1 lnðGDPHOMEit Þ þ a2 lnðGDPHOSTÞ þ a3 lnðDISTANCEij Þ þ a4 lnðXRATEijt Þ þ a5 ðLANGUAGEij Þ þ a6 ðADJACENCYij Þ þ a7 ðNAi Þ þ a8 ðEUi Þ þ a9 ðBNAij Þ þ a10 ðBEUij Þ þ eijt

ð2Þ

The Rugman hypothesis that is being tested would predict that a9 and a10 are both positive. Eq. (2) is estimated using bilateral inward MNE activities listed in Table 1 for a sample of 28 OECD countries.

5. EMPIRICAL EVIDENCE The regression results are reported in Table 11. The model estimates the gravity-model factors that explain the inward activities of OECD MNEs into each OECD country. Host countries get a larger share of foreign MNE employment, turnover, and exports from countries that share the same language, whereas the amount of foreign production, imports, and intrafirm trade (exports and imports) are unrelated to language similarities. Adjacency turns out for the most part not be important in explaining the inward activities of MNES (there are marginally significant positive relationships to turnover, imports and less so for exports). Distance, on the other hand is strongly negatively related to the seven measures of

Language Adjacency Distance GDP home GDP host Exchange rate Both in NA Both in EU EU dummy NA dummy Constant

Gravity Model Results..

Production

Employment

Turnover

Total exports

Total imports

Intra-firm exports

Intra-firm imports

0.20 (0.33) 0.64 (0.75) 2.60 (5.05) 2.66 (5.43) 0.82 (6.64) 0.00 (0.01) 5.06 (3.26) (Dropped)

0.63 (2.33) 0.00 (0.98) 1.22 (13.50) 0.02 (0.31) 0.80 (12.14) 0.06 (2.93) 5.56 (8.48) 1.62 (7.92) 1.56 (7.71) 1.18 (4.20) 4.59 (1.89) 0.50 695

0.74 (2.52) 0.00 (1.95) 1.22 (12.36) 0.35 (4.53) 0.94 (12.86) 0.11 (3.02) 5.55 (7.63) 2.21 (8.19) 2.15 (8.60) 1.01 (3.19) 1.00 (0.36) 0.62 594

1.04 (1.82) 0.00 (1.46) 0.85 (7.81) 0.24 (2.60) 0.75 (8.27) 0.78 (10.57) (Dropped)

1.39 (1.42) 0.66 (1.62) 0.98 (6.93) 0.05 (0.52) 0.86 (9.47) 0.54 (8.00) (Dropped)

0.48 (0.31) 0.64 (0.95) 1.58 (4.25) 1.51 (5.13) 0.59 (3.61) 0.07 (0.33) (Dropped)

0.91 (0.54) 1.34 (1.16) 0.84 (1.65) 6.10 (0.84) 0.82 (3.72) 2.64 (0.70) (Dropped)

1.07 (3.24) 0.82 (2.67) 0.20 (0.56) 11.76 (3.53) 0.59 360

0.12 (0.36) 0.50 (1.78) 0.11 (0.30) 9.89 (2.94) 0.56 271

0.65 (0.72) 0.80 (1.15) 1.23 (1.88) 42.57 (4.82) 0.67 143

0.59 (0.46) 1.17 (1.46) 0.03 (0.03) 150.67 (0.79) 0.51 89

1.08 (2.87) 3.00 (3.71) 62.72 (5.93) 0.46 147

Notes: NOB is the number of observations used in the regression. t-statistics provided in parentheses.

WALID HEJAZI

Adjusted R2 NOB

104

Table 11.


105

intra-firm activity analyzed, with the exception of intra-firm imports, where negative relationship is marginally significant. The coefficient on the GDP of the host country is positive and highly significant in all cases. There are mixed results on the GDP of the home, where the coefficient is positive and significant in two cases (Production and Total Exports), negative in two cases (Turnover and Intra-firm Exports) and statistically insignificant in the other cases. There is similar mixed evidence on the coefficient on the exchange rate. It is positive only in the cases of Employment and Turnover, and negative in the cases of Total Exports and Total Imports, and statistically insignificant in the other cases. The coefficients of interest however, are the estimated relationships when both the countries involved are in the same region (North America or Europe). As indicated above, if there is no regional bias, then the fact that both countries are in the same region should not drive their having more bilateral activity, once the gravity model factors have been taken into account. The results presented here indicate that this clearly is not the case for several aspects of MNE activity. In the case of Employment, Turnover, and Total Exports, there is significantly more bilateral activity between European countries than can be predicted by this transactions cost model. The amounts of imports by MNEs into host economies, as well as Intra-firm Imports and Exports appear to be consistent with transactions costs. There is not enough data to test this hypothesis for production. These results are in sharp contrast to the results documented for North America. In that case, it is found that there is less Production, Employment, and Turnover between North American economies than is predicted by the gravity model. There is not enough data in the database used here to test the other four dimensions within North America. Something should also be said about the North America and EU dummies. The test is whether there is more activity within these regions than predicted by the model. This of course is after the gravity-model factors have been accounted for, including an accounting for whether the two partners are in the same region. These regional dummies indicate that there tends to be less inward MNE activity within the EU than is predicted by the model, whereas the estimates for North America tend to be positive or insignificant. These coefficients are not directly related to the theory being tested in this paper – these regional coefficients test the amount of activity in locating in these regions, but not whether MNEs are overly concentrated in their home region.

106

WALID HEJAZI

It is also important to note the differences and similarities between these results and those reported in Hejazi (2005). Hejazi (2005) considers the outward FDI and exports of OECD countries and finds that there are more exports between EU and NA countries than can be explained by a gravity model. The export data used in the 2005 study includes total exports and not just those undertaken by MNEs, as is the case here. The results for exports in this paper are those undertaken by MNEs locating into a host (EU) economy and exporting, and the results indicate that more of these exports go to other EU partners than can be predicted by a gravity-type model. The exports from a local economy can be broken down into three components: those undertaken by domestic MNEs, those undertaken by foreign MNEs locating in the local economy, and those undertaken by non-MNEs. From Hejazi (2005), we know that the sum of these is more regionally concentrated (both in the US and the EU) than can be explained by a gravity model. The Hejazi (2005) do not indicate which of the three components is driving the aggregate results. The empirical results documented here show that foreign MNEs locating in EU countries have a regional bias in where they export to – and this is an entirely new result. This result does not hold for imports. There is not enough data in the data set used here to test these hypotheses for North America. The Hejazi (2005) study also found that there was a regional bias in intraEU patterns of outward FDI, but not for NA. The current study does not consider outward FDI – recall from above that there was insufficient outward data in the data set used to test hypotheses on the outward activities of OECD MNEs.

6. CONCLUSIONS There has emerged significant evidence of a regional bias in the distribution of MNE activities. That is, a majority of the world’s largest MNEs have a large share of their activities in their home region, and in this sense this majority cannot be said to be globally oriented. That is, MNEs are regional in nature. Economic theory would predict, however, that there should be both national and regional biases in the activities of MNEs. Since there are costs associated with operating in foreign and distant locations, especially those that do not share similar languages and institutions, then MNEs are more likely to locate in closer and culturally similar markets.


107

Previous evidence, discussed above, finds that these concentrations using US MNE activities abroad are consistent with a transactions cost model, although there remains a significant home (national) bias in the activities of US MNEs. Another study using data on outward FDI and exports indicates that there is a significant regional bias in exports among OECD countries, whereas in the case of outward FDI, there is a significant regional bias in Europe although not in North America. The current study considered the activities of foreign OECD MNEs locating in other OECD host economies. The distribution of these foreign activities in each host economy is analyzed. The results indicate that activities by foreign MNEs locating in host economies have a regional bias in the EU. That is, European MNEs have more activity within their region than can be predicted by a transactions cost model. The data used have limited information on North America, although for the data that are available, there is no such bias within North America. The Rugman hypothesis therefore continues to hold most strongly within the EU context, even after controlling for the transactions costs associated with operating in foreign and distant markets.

REFERENCES Anderson, J. E. (1979). A theoretical foundation for the gravity equation. American Economic Review, 69, 106–116. Bergstrand, J. H. (1985). The gravity equation in international trade: Some microeconomic foundations and empirical evidence. Review of Economics and Statistics, 67, 474–481. Bergstrand, J. H. (1989). The generalized gravity equation, monopolistic competition, and the factor proportions theory of international trade. Review of Economics and Statistics, 71, 143–153. Bergstrand, J. H. (1990). The Heckscher–Ohlin–Samuelson model, the Linder hypothesis and the determinants of bilateral intra-industry trade. Economic Journal, 100(403), 1216– 1229. Blonigen, B. A., & Wilson, W. W. (1999). Explaining Armington: What determines substitutability between home and foreign goods? Canadian Journal of Economics, 32(1), 1–20. Brainard, S. L. (1997). An empirical assessment of the proximity-concentration trade-off between multinational sales and trade. American Economic Review, 87(4), 520–544. Deardorff, A. V. (1998). Determinants of bilateral trade: Does gravity work in a neoclassical world. In: J. A. Frankel (Ed.), The regionalization of the world economy (pp. 7–32). Chicago: University of Chicago Press. Ethier, W. (1982). National and international returns to scale in the modern theory of international trade. American Economic Review, 72, 389–405.

108

WALID HEJAZI

Feenstra, R. C., Markusen, J. R., & Rose, A. K. (2001). Using the gravity equation to differentiate among alternative theories of trade. Canadian Journal of Economics, 34(2), 430–447. Feldstein, M., & Horioka, C. (1980). Domestic savings and international capital flows. The Economic Journal, 90, 314–329. Frankel, J. A., Stein, E., & Wei, S. J. (1995). APEC and regional trading arrangements in the Pacific. In: W. Dobson & F. Flatters (Eds), Pacific trade and investment: Options for the 1990s (pp. 289–312). Kingston, ON, Canada: John Deutsch Institute. French, K. R., & Porterba, J. M. (1991). Investor diversification and international equity markets. American Economic Review, Papers and Proceedings, 81, 222–226. Grubert, H., & Mutti, J. (1991). Taxes, tariffs, and transfer pricing in multinational corporate decision-making. Review of Economics and Statistics, 73(2), 285–293. Grosse, R., & Trevino, L. (1996). Foreign direct investment in the United States: An analysis by country of origin. Journal of International Business Studies, 27(1), 139–155. Head, K., & Ries, J. (2004). Judging Japan’s FDI: The verdict from a dartboard model. Working Paper, University of British Columbia. Hejazi, W. (2005). Are regional concentrations of OECD exports and outward FDI consistent with gravity? Atlantic Economic Journal, 33(4), 423–436. Hejazi, W. (2007). Reconsidering the concentration of US MNE activity: Is it global, regional or national? Management International Review, 47(1), 5–27. Hejazi, W., & Safarian, A. E. (2005). NAFTA effects and the level of development. Journal of Business Research, 58, 1741–1749. Hejazi, W., & Trefler, D. (1996). Canada and the Asia Pacific region: Views from the gravity, monopolistic competition, and Heckscher–Ohlin models. In: R. Harris (Ed.), The Asia Pacific region and the global economy: A Canadian perspective (pp. 47–86). Calgary: University of Calgary Press. Helpman, E. (1987). Imperfect competition and international trade: Evidence from fourteen industrial countries. Journal of the Japanese and International Economies, 1, 62–81. Leamer, E. E. (1974). The commodity composition of international trade in manufactures: An empirical analysis. Oxford Economic Papers, 26, 350–374. Lipsey, R. E., & Weiss, M. Y. (1981). Foreign production and exports in manufacturing industries. The Review of Economic Statistics, 63, 488–494. Lipsey, R. E., & Weiss, M. Y. (1984). Foreign production and exports of individual firms. The Review of Economics and Statistics, 66, 304–308. Markusen, J. R. (2002). Multinational firms and the theory of international trade. Cambridge, MA: MIT Press. McCallum, J. (1995). National borders matter: Canada–US regional trade patterns. American Economic Review, 85(3), 615–623. Rugman, A. M. (2000). The end of globalization. London: Random House. Rugman, A. M. (2005). The regional multinationals. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Brain, C. Multinational enterprises are regional, not global. The Multinational Business Review, 11(1), 3–12. Rugman, A. M., & Girod, S. (2003). Retail multinationals and globalization: The evidence is regional. European Management Journal, 21(1), 24–37.


109

Rugman, A. M., & Moore, K. (2003). Canadian multinationals are regional, not global. Policy Options, (Aug.), 44–46. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Stein, E., & Duade, C. (2001). Institutions, integration and the location of foreign direct investment: New horizons for foreign direct investment. In: OECD global forum on international investment (pp. 101–128). Paris: OECD. http://www.oecd.org/document/ 23/0,2340,en_2649_34893_2405399_1_1_1_1,00.html

THE EMPIRICS OF MULTINATIONALITY AND PERFORMANCE Harry P. Bowen ABSTRACT In this paper, I address issues concerning the empirical estimation of a relationship between firm performance and its degree of multinationality. I argue for greater delineation of the underlying nature of firms’ multinationality and point to several statistical issues regarding estimation that appear to need resolution, but which appear to have been largely neglected in the literature that has examined for a multinationality– performance relationship. Among these are endogeneity of the multinationality construct in the performance relationship and the likelihood that the multinationality–performance relationship is heterogeneous across firms.

MULTINATIONALITY AND PERFORMANCE: BASIC ISSUES It was recently suggested that what determines the international success or failure of firms is the ‘‘one big question’’ that has, and should remain, at the Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 113–142 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13006-0

113

114

HARRY P. BOWEN

core of international business research (Peng, 2004). True to this dictum, whether and how a firm’s performance depends on its degree of ‘‘multinationality’’ remains one of the most researched questions in the international business literature. As readers of this volume are likely aware, conflicting empirical results produced over the past 25 years have precluded reaching a consensus regarding the answer to this key question. However, recent research has made important inroads into the analysis of this key question as it attempts to reconcile prior results. A seemingly important direction has been to explore the functional nature of the multinationality– performance relationship; in particular, whether it is nonlinear and if so, what is the nature of this nonlinearity. The results of this strand of work appear to suggest that the relationship is indeed non-linear. While recent efforts have sought to refine the nature of an empirical multinationality–performance, there are several issues that appear to have been largely neglected in the broad literature on this subject. In this chapter, I reflect on these issues and offer thoughts for improvement and directions for further work. Some of the issues raised are statistical, while others deal with conceptualizations of the multinationality–performance relationship, and ultimately how estimation of this relationship may allow us to better understand the ‘‘one big question.’’ In this regard, I reconsider the underlying notions of multinationality and their associated costs and benefits to suggest that attention could usefully be directed toward the different forms in which multinationality manifests itself, what I call ‘‘modes of multinationality’’ (MoMs). This call for differentiating different aspects of multinationality parallels Rugman and Verbeke’s (2004) observation that relatively few ‘‘global’’ companies appear to be truly global in scope; they instead appear to concentrate on exploiting their advantages in their closer regional markets.

From Theory to Empirics For over 25 years, considerable intellectual effort has been directed toward understanding the basis for the multinational firm. The predominant theory ascribes the existence of the multinational firm to ownership, location and internalization advantages (Dunning, 1981; Rugman, 1981), with more recent work stressing the multinational firm as an organic mechanism for the spatial transmission of tacit knowledge regarding the resources and capabilities that underlie a firm’s competitive advantages (Kogut, 1997; Kogut & Zander, 1993). These theoretical frameworks serve to explain the governance structure represented by the multinational firm, and to suggest

The Empirics of Multinationality and Performance

115

why such firms might be expected to outperform their equivalent but purely domestic rivals. Based on these theoretical frameworks, an extensive empirical literature (for an exhaustive summary see Hitt, Tihanyi, Miller, & Connelly, 2006) has examined for a relationship between the extent or degree of multinationality and firm performance; the degree of multinationality is most often captured by either the firm’s foreign sales ratio (share of a foreign sales in total-firm sales) or by a measure of the geographic distribution of firm sales. The key elements for hypothesizing a relationship between a firm’s degree of multinationality and its performance rests primarily on arguments about the behavior of the two components of a firm’s profit (performance): revenue and cost, and their relationship to the extent of a firm’s international presence. On the revenue side, gains from venturing into new markets arise from exploiting market imperfections (Rugman, 1979) and firm-specific advantages that are in turn linked to firm’s intangible assets (Caves, 1996). Firms may also gain from opportunities to engage in price discrimination if markets can be spatially segmented. On the cost side, the spreading of the (quasi-fixed) costs of investments in creating intangible assets (e.g., R&D and advertising) over a larger customer base confers economics of scale, and the ability to deploy intangible assets having public good characteristics (i.e., knowledge capital) into the different markets in which the firm operates can confer economics of scope. Increased geographic reach and an attendant increase in the size of the firm can also confer advantages, and hence cost savings, over suppliers, distributors, etc. Firms may also use their geographic reach to reconfigure value-chain activities to arbitrage difference in factor input costs across markets. However, these advantages come at a cost: international expansion into new markets will incur (fixed) costs associated with the ‘‘liability of foreignness’’ (Hymer, 1976). These entry costs depend on the nature of the firm’s products, the extent to which they need adaptation to local conditions, and the firm’s relative unfamiliarity with local business practices. All these factors suggest that a firm will enter and be present in a foreign market if it perceives that its inherent capabilities will overcome the initial costs associated of the liability of foreignness. As is discussed later, that a firm initially incurs additional costs to enter a foreign market suggests that its performance would be expected to improve the longer the firm is present in a given market since the liability of foreignness and its attendant costs would be expected to decline as the firm learns about its host market(s). The above summarizes the essential theoretical elements that have led researchers to hypothesize a positive relationship between a firm’s degree of

116

HARRY P. BOWEN

multinationality and its performance. In what follows, I take the underlying theoretical rationale for a relationship between performance and multinationality as given and instead focus on various issues that arise when seeking to empirically detect the nature of the hypothesized relationship.

Measuring Multinationality To a large extent, the choice of a measure of multinationality has been data bound, meaning that while scholars may have a wish list of variables to capture their idea of what is meant by a firm’s degree of multinationality, data constraints have generally prohibited such wishes from becoming reality. The measures that have been used are generally of two types. The first captures the degree of a firm’s international orientation or involvement in international markets. The most common measure of this type, and also the most commonly used measure in empirical studies, is the foreign sales ratio, defined as the share of a firm’s total sales that derive from outside its home market. This measure therefore captures multinationality in terms of the importance of international transactions to the firm, and it can often include exports as well as sales by foreign affiliates. Other measures of this type would include foreign assets to total assets, foreign employment to total employment, and a count of the number of foreign affiliates. The second class of measures instead attempts to capture the diversity of a firm’s international involvement. The most common measure of this type is an entropy-type diversity index based on a firm’s sales shares across different geographic markets, although the Herfindahl-type index has also been used.1 Other share variables that can be used to construct such diversity indexes are affiliate asset shares and affiliate employment shares across different countries or regions. Another diversity measure is a count of the number of markets in which a firm is active. Generally, the first type of measure is thought to capture the depth of a firm’s involvement in international markets, while the second type of measure is thought to capture the breath of the firm’s involvement in international activities. Some researchers rightly see multinationality as a multidimensional construct that encompasses both of the above types of a firm’s international involvement (and perhaps others) and that the measure of multinationality should therefore also have a multidimensional flavor (Sullivan, 1994). Researchers have therefore often constructed composite measures such as the average of the foreign sales ratio and the ratio of affiliate employment to total employment (e.g., Lu & Beamish, 2004). In this context, the UNCTAD


117

routinely publishes as part of its World Investment Report a ‘‘Transnationality Index’’ defined for each firm as the average of three ratios: foreign assets to total assets, foreign sales to total sales and foreign employment to total employment. Others (e.g., Contractor, Kundu, & Hsu, 2003) have used data reduction methods such as factor analysis or principle components to combine different multinationality measures into a single composite. While the desire for a single measure that can capture all the seemingly diverse aspects of firms’ international activity is perhaps understandable, collapsing different dimensions of multinationality would appear the wrong direction to take if one is to gain a better understanding of how multinationality impacts performance. Instead, if we are to move closer to answering Peng’s (2004) ‘‘one big question,’’ it would seem desirable to have a greater understanding and identification of the individual sources of higher (lower) performance that may arise from multinationality. This argues for disaggregation rather than aggregation, for example, to distinguish the involvement of a firm in one geographic region versus another or the nature of its affiliate activity (e.g., production versus distribution) in different locations. Rather than aggregate different dimensions of multinationality it would seem worthwhile to consider them separately, and to seek to obtain measures that better capture the underlying nature of a firm’s involvement in international markets. As discussed in the next section, recent research on the multinational-performance relationship suggests that more clarity on the different bases or forms of multinationality is warranted.

The Nature of the Multinationality–Performance Relationship Early studies of the multinationality–performance relationship were directed at uncovering the nature of the relationship with the null hypothesis being that the relationship is positive. The investigations were therefore directed more at uncovering an association as opposed to a causal relationship, and the usual linear specification was adopted. While several studies found evidence of a positive relationship, others found no relationship or a negative relationship (for reviews, see Contractor et al., 2003 and Hitt et al., 2006). These conflicting findings have generally prevented reaching a consensus about the nature of the relationship. Some of the most recent efforts (e.g., Gomes & Ramaswamy, 1999; Contractor et al., 2003; Lu & Beamish, 2004) have sought to resolve the conflicting results of past studies by giving greater consideration to the underlying costs associated with international expansion (either in terms of

118

HARRY P. BOWEN

depth or breath) and what this implies for the nature of the multinationality–performance relationship. The result of these efforts is to suggest that a potential explanation for the previously mixed results is that the multinationality–performance relationship is nonlinear. Initial results indicated that the relationship follows an inverted U-shape (Gomes & Ramaswamy, 1999; Geringer, Tallman, & Olsen, 2000). The most recent analyses propose that the relationship is more complex, and in particular, have posited that the relationship is S-shaped. Researchers have interpreted the hypothesized S-shape to mean that firms pass though (three) stages of international expansion, with the performance implication of expanded multinationality being different at each stage (Contractor et al., 2003; Lu & Beamish, 2004). I offer some reflections on this evolving strand of the literature below. The rational for postulating a nonlinear relationship (quadratic, cubic, etc.) derives primarily from a deeper consideration of the behavior of costs as the importance of international activity to the firm rises. Since performance (profit) relates to the difference between revenue and cost, the behavior of revenues should also play a part in any explanation of the multinationality–performance relationship. However, the literature has thus far focused mainly on the behavior of costs and largely ignored the behavior of revenues; the latter are essentially assumed to rise at a decreasing rate as the degree of multinationality rises. I note in passing that the arguments commonly advanced in the literature equate a firm’s degree of multinationality (either depth or breath) with the passage of time. While plausible, equating the passage of time to the importance of international activities to the firm could be questioned. Moreover, most studies derive their estimates in essentially a cross-sectional framework, which then raises the (difficult) question of the whether the cross-sectionally estimated ‘‘time path’’ between performance and multinationality is indicative of the time path applicable to a single firm. For now, I put aside this issue and instead proceed to consider the reasoning advanced for expecting a nonlinear relationship between multinationality and performance. To fix ideas and as a framework for discussion, let the unit of observation be firm ‘‘i’’ at time ‘‘t’’ and assume that the relationship between this firm’s performance (Pit) and its (degree of) multinationality (Mit) at time ‘‘t’’ can be written as Pit ¼ Fi ðM it ; Z it ÞRi ðPit ; Qit Þ Gi ðM it ; Z it ÞC i ðQit Þ ¼ Oi ðPit ; Qit ; M it ; Z it Þ

(1)

In this expression, revenue Ri ( ) depends on the price (Pit) and quantity (Qit) of shipments (level of activity), while cost Ci ( ) depends only on the


119

quantity of shipments. The functions Fi( ) and Gi( ) ‘‘modify’’ a firm’s revenue and cost. These functions depend on the firm’s degree of multinationality (Mit) and a set of firm specific characteristics (Zit) at time t.2 The function Oi( ) is therefore the firm’s profit function which, by definition, depends on price, quantity, degree of multinationality and firm specific characteristics. The revenue, cost, and modifier functions, and hence the profit function, have an ‘‘i’’ subscript to indicate that they are firm specific, but they have no ‘‘t’’ subscript to indicate the assumption that the nature of these functions does not vary over time.3 The revenue and cost functions can be thought to relate to a single product, produced and sold in one market or to an ‘‘aggregate’’ relationship that combines several products and markets. Written in this way, the multinationality variable Mit takes the role of a choice variable that the firm can use to influence its revenue and cost, and hence its performance.4 The multinationality variable appears in both Fi( ) and Gi( ) to capture that multinationality is likely to have a separate influence on revenue versus cost.5 Given this characterization, I now consider the basic arguments advanced for expecting that the firm’s profit function Oi( ) will be a nonlinear function of the degree of multinationality. The essential arguments advanced are these. First, in the early stages of internationalization, the firm will incur costs associated with the liability of foreignness and is therefore likely to incur losses. As the firm’s international presence grows (over time) the firm learns, adapts, and becomes more informed about selling/operating outside its domestic market. This learning or experience implies that the initial entry costs diminish in importance (over time) and at some point profits will turn positive. However, as the firm’s degree of multinationality rises so also do costs associated with higher complexity and coordination. As the degree of multinationality rises, these costs may begin to dominate and as a result the firm’s profits decline and may even become negative. These arguments imply that the relationship between the level of a firm’s profit and its degree of multinationality, holding fixed all other determinants of profit, would exhibit an inverted U-shape. In this explanation, the inverted U-shape relationship emerges because the firm bears a (fixed) cost of entering foreign markets, where the relevant fixed cost is associated with the liability of foreignness (Gomes & Ramaswamy, 1999). While this explanation deals primarily with the behavior of the level of profit as the degree of multinationality rises, it also has implications for the expected incremental effect of higher multinationality: the inverted U-shape implies that the incremental effect is at first positive, but it then becomes negative as the degree of multinationality passes the point at which the level

120

HARRY P. BOWEN

of profit is maximized, and the firm’s degree of multinationality enters the region of ‘‘excess multinationality.’’ Recent literature (Contractor et al., 2003; Lu & Beamish, 2004) has gone beyond the inverted U-shape to argue that the relationship is instead Sshaped, which is algebraically represented by a cubic relationship. Fig. 1 illustrates the nature of the relationship suggested. In Fig. 1, the degree of multinationality has been divided into three stages: I, II and III, corresponding to negative profit, positive and rising profit, and positive but declining profit. As indicated in Fig. 1, the key difference between asserting an S-shape rather than an inverted U-shape concerns the behavior of incremental profit in Stage I (at a low level of multinationality). Along the dashed segment of the profit relationship, the incremental effect of increased multinationality on performance is negative and declining until a point of minimum profit is reached, after which the incremental effect becomes positive, although the level of profit may still be negative. The behavior of costs that could produce this initial sloping segment of the profit curve is that the incremental cost (marginal cost) of international expansion first falls at a decreasing rate, but then subsequently rises at an increasing rate. In Fig. 1, the point where this switch in the rate of change in incremental cost occurs, is where the dashed segment of the profit relationship ends and the Profit

Stage I

Stage II

Stage III

0 Degree of Multinationality

Fig. 1.

Relationship between Performance (Profit) and Multinationality.


121

solid segment begins.6 As noted, the main theoretical rationale offered to explain the behavior of profits during Stage I is that firms incur fixed costs associated with the liability of foreignness in the initial phases of international expansion. The profit relationship depicted in Fig. 1 bears a noticeable resemblance to the standard economics’ textbook depiction of a firm’s profit function, where the variable on the horizontal axis would be the firm’s output per period (level of activity). In this context, economics would offer two explanations for falling marginal costs at low levels of firm activity. The first arises in the context of the short run, when some inputs the firm employees are considered fixed. In this case, marginal cost initially falls with rising firm activity due to an underutilization of fixed inputs at low levels of firm activity. As the firm’s level of activity rises these fixed inputs eventually become ‘‘fully employed,’’ after which further expansion by the firm entails diminishing (marginal) returns and hence marginal cost begins to rise. Moving to the long run when all inputs are variable, falling marginal cost at low levels of a firm’s activity instead arise from economies of scale (declining long-run average costs). The theoretical explanation advanced for the 3-stage hypothesis appears to mix, and in fact do not make clear, whether the contemplated behavior of costs relates to the short-run or long-run perspective. In treating the liability of foreignness as a fixed cost of entry, the theoretical arguments take on a decidedly short-run flavor, whereas the implicit treatment of time in the analysis suggests instead the long-run view involving the effects of scale. However, there is another conceptual issue to be considered. In Fig. 1, the initial dotted segment of Stage 1 is associated with negative profit. If profit were instead always positive then it is not possible to construct a (sensible) cost relationship that would give rise to a profit relationship in which the initial segment has a negative slope; to do so would require total costs to fall as the degree of multinationality rises. This suggests one quick way to assess, in a given set of data, whether an S-shaped relationship arising from the behavior of costs is likely to be a sensible specification: one need only look for a threshold value of multinationality below which firms earn negative profits. As a quick check of this idea, Table 2 shows for each year between 1984 and 1999 the average rate of return on assets (ROA) in a sample of firms at deciles of the foreign sales ratio.7 For example, in 1984, the average ROA was 13.8% for firms whose foreign sales ratio fell in the 10th percentile, while the average ROA of firms in the next decile of the distribution was 12.2% Of note is that the average ROA values in any given year do not differ appreciably across deciles of the foreign sales ratio and, as indicated

122

HARRY P. BOWEN

by the values in the last column of Table 1, the average ROA across all firms declines over time. The last two rows in Table 1 show respectively, the rate of return averaged over all years in each decile, and the average rate of return computed using the data pooled over all years. As seen, the average of the ROA values in a given column do not differ appreciably from the average ROA values computed from the sample pooled over all years. In Table 1, the average ROA is consistently positive in the lower deciles of the foreign sales ratio; the two negative average ROA values are only in the last decile and only in the last two years. This casts doubt on an S-curve relationship arising from the assumed behavior of costs, at least in this set of data. For completeness, Fig. 2 plots the average rate-of-return values shown in the second to last row of Table 1 against the average foreign sales ratio in each decile, and it shows fitted linear, quadratic, and cubic relationships along with a table of estimation results. The adjusted R-square values rise with the number of included regressors, with the cubic model evidencing a good overall fit as judged by the overall F-statistic. However, while the estimates in the linear and quadratic specifications are significant at the 5% level, the coefficient estimates in the cubic specification are not significant. For the cubic specification, the significant F-statistic but lack of significant coefficient estimates is one signal of a collinearity problem. Indeed, the correlation between the level and squared values is 0.961, the correlation between the level and cubed values is 0.899, and the correlation between the squared and cubed values is 0.984. Obviously, the relatively low-average ROA value in the 10th decile has an important influence on the results, although this ‘‘outlier’’ can be explained by an appeal to the hypothesis that ‘‘excessive multinationality’’ will be associated with low or negative performance. The bottom line of this admittedly simple inquiry is that there is no evidence that the average firm earns negative returns at low levels of multinationality (in data derived from COMPUSTAT, a widely used source), and that ‘‘outliers’’ may figure importantly in estimation. While the above suggests refining the basis for the presumed behavior of incremental costs, the more narrow economics perspective suggests one is describing the behavior of a firm’s profit function as its level of foreign involvement rises. If so, then this raises some concerns about the usual measures of multinationality. As noted, one of the most commonly used measures is the foreign sales ratio. However, it is the firm’s total profit function that is being depicted in Fig. 1. This in turn depends on a firm’s total sales, both domestic and foreign. If one thought to be depicting profits on foreign activity then the variable on the horizontal axis would be the level of foreign sales, and the relationship depicted would be drawn holding fixed

Average Rate of Return on Assets (in %) at Deciles of the Foreign Sales Ratioa. Deciles of the Foreign Sales Ratio (Upper Limit of Decile in %)b

Year

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Mean (Column) All years a

1 (7.2) 13.8 14.2 11.1 9.3 6.5 11.6 10.2 8.1 7.6 8.1 10.7 11.5 9.4 6.0 5.2 4.4 9.2 8.9

2 (11.3) 12.2 12.1 9.5 11.8 12.3 12.0 9.9 11.0 10.6 8.6 10.9 9.0 7.6 8.0 6.4 5.8 9.9 9.5

3 (15.2) 13.9 12.0 11.8 9.2 10.9 11.8 7.4 7.1 9.4 11.2 10.8 10.7 10.3 7.9 8.2 6.1 9.9 9.7

4 (19.6) 14.2 10.6 13.1 10.7 11.5 10.2 10.1 9.7 8.6 10.5 11.3 9.8 9.9 8.0 8.2 8.2 10.3 10.0

5 (24.6) 13.1 10.1 7.8 8.6 7.7 9.1 10.0 7.7 9.5 8.3 9.8 8.3 5.2 10.2 9.7 4.0 8.7 8.5

6 (30.1) 13.3 9.6 8.2 10.4 9.7 8.2 7.7 8.2 9.7 7.3 11.7 9.5 9.7 9.3 10.1 9.0 9.5 9.5

7 (36.5) 13.0 11.0 9.0 8.8 9.4 8.8 8.2 7.8 7.3 7.7 7.6 9.9 10.2 11.5 8.1 8.2 9.2 9.1

8 (44.3) 12.1 9.2 8.4 10.8 11.8 10.3 8.0 5.2 10.3 7.5 9.9 7.7 9.9 8.6 6.6 4.4 8.8 8.6

9 (56.0) 14.0 11.1 8.8 10.9 11.8 8.4 8.8 8.6 5.1 7.2 9.0 8.4 10.1 7.5 1.2 5.7 8.5 8.1

Mean (Row) 10 (72.1) 7.3 3.0 3.8 6.7 6.0 6.7 4.6 6.9 3.9 2.1 4.0 2.4 2.1 1.1 3.4 4.4 3.3 2.7

12.69 10.29 9.15 9.71 9.77 9.71 8.48 8.01 8.19 7.85 9.58 8.73 8.43 7.81 6.03 5.16 8.72 8.45


Table 1.

Based on sample of 16,498 observations (approximately 1,000 firms in each year). Upper limit is for data pooled over all years; value for last decile is the mean value.

b

123

124

12

Average Return on Assets (%)

10

8

6

4

2

Variable

Linear

Quadratic

Cubic

FSR FSR Squared FSR Cubed Intercept

-8.08**

8.76* -22.89**

11.06**

8.96**

-2.10 13.67 -32.46 9.70**

Adj. R-square F-Statistic Observations

0.649 17.62** 10

0.918 51.55** 10

0.924 37.52** 10

** P < .01

* P < .05

0

Fig. 2.

0.1

0.2

0.3

0.4 Multinationality

0.5

0.6

0.7

Relationship between Average Return on Assets and Multinationality (Foreign Sales Ratio).

0.8

HARRY P. BOWEN

0


125

the level of domestic sales. This line of thinking suggests a relationship between profit on foreign activities and the level of foreign sales, with domestic sales included as a control variable. This view lends credence to Rugman and Verbeke’s (2004) suggestion that one should focus on the rate of return on foreign assets in relationship to the foreign sales ratio. In this respect, preliminary work by Rugman (2006) in a sample of U.K. firms indicated support for an inverted U-shape relationship, but no support of an S-shaped relationship, between the rate of return on foreign assets and the share of home-region sales in total firm sales. Returning to the characterization of the multinationality–performance relationship given in expression (1), it is clear that the relationship embeds a number of elements that could affect the nature of relationship. First, there are different channels through which multinationality influences firm performance, suggesting that ‘‘catch all’’ characterizations of multinationality only serve to confound the potentially different effects that multinationality may have on revenue versus cost. In particular, different ‘‘MoMs’’ may be important for the nature of the multinationality– performance relationship. For example, if a firm’s multinationality relates to off-shore production facilities intended to take advantage of locationspecific factors such as lower labor costs this might affect primarily the cost side, and one would expect a measure of this type of MoM to evidence a positive relationship with performance. Conversely, if a firm’s MoM relates to having distribution/sales activities in foreign markets, but not production, then this MoM might be expected to affect both revenues and costs, with an uncertain net effect. More broadly, expression (1) suggests that the spatial dimension of a firm’s revenue or cost is another element of the relationship, suggesting, for example, that foreign sales divided by the number of markets in which the firm operates could be an appropriate measure. What all this suggests that different MoMs are likely to have different implications for the multinationality–performance relationship, and hence that efforts should be directed to separately identifying and analyzing different MoMs.8 In my view, the recent search for higher-order terms to characterize the multinationality–performance relationship is unlikely to significantly expand our understanding of the underlying basis for the relationship, that is, the aspects of multinationality that would give rise to superior performance. In terms of functional form, it seems both sufficient and parsimonious to adopt a quadratic specification since it captures the essential notion that in the early stages of internationalization firms incur higher costs associated with the liability of foreignness and that there exists

126

HARRY P. BOWEN

some ‘‘optimal’’ degree of multinationality, beyond which multinationality becomes excessive. Although the existence of significance scale economies at low levels of internationalization would appear to be a better basis for expecting a cubic relationship, the risk of over-parameterizing the model with additional powers of the multinational construct (which also introduces variables with high inter-correlations) seems unwarranted. As previously noted, the recent explorations of potential nonlinearity of the multinationality–performance relationship have mainly focused on the behavior of costs with only limited attention given to how the behavior of revenues might contribute. But an inverted U-shaped profit function can arise under several different specifications of the revenue and cost relationships. For example, an inverted U-shape can arise when there are constant costs of international expansion, but the revenue function has an inverted U-shape (the case in which a firm has market power). An inverted U-shaped profit function can also arise when incremental revenue is constant (no market power) and costs are U-shaped. A third possibility is that the incremental cost of international expansion is constant or rising but the revenue relationship evidences a point of inflection (a different kind of S-shape), that is, in the early stages of international expansion revenues rise at an increasing rate but then subsequently rise at a decreasing rate as further international expansion encounters diminishing returns (i.e., market saturation). Clearly, still more configurations are possible if the revenue and cost relationships have even higher-order nonlinearities.9 All this is to suggest that a deeper understanding of the empirical behavior of both revenue and cost may be needed before a better understanding of the relationship between multinationality and performance can be reached. Stated differently, when examining the relationship between multinationality and performance, the ability to untangle the ways in which multinationality affects revenue and cost is problematic. A separate focus on the behavior of revenue and cost may therefore be needed to identify the dimensions of multinationality that are most important.10 Does multinationality (or specific types of multinationality) mainly affect a firm’s ability to generate revenues or does it mainly affect costs, and if so how? As a final remark on the current literature that investigates for nonlinearities in the multinationality–performance relationship, a close look at some of the recent literature suggests that some cautionary remarks are in order regarding the interpretation of results. As an example, the first column in Table 2 reproduces results from a recent study that estimated a linear, quadratic and cubic multinationality–performance relationship. The multinationality measure was a composite measure that averaged the


127

Table 2. Analysis of Estimates of a Multinationality–Performance Relationship. Form of the model

Level Quadratic Cubic a

Coefficient Estimatea Level (M )

Square (M2 )

0.17 0.30 0.38

0.26 0.75

Cube (M3 )

0.50

Incremental Return (=Slope of Relationship)

Implied Value of Multinationality (M ) that Maximizes Performance

0.17 0.30+0.52M 0.38+1.5(MM2)

M=0 M=0.577 M=0

All estimates significant at the 0.01 level.

number of affiliates and number of geographic areas in which firms operated (as per my remarks above, I would have wished for these two elements to have been examined separately). The study concluded that its results strongly supported the hypothesis of the cubic relationship depicted in Fig. 1. However, using the study’s results, the column in Table 2 labeled ‘‘Incremental Return’’ reveals that the estimated cubic relationship is far from that shown in Fig. 1. For both the level and cubic models, the implied value of the multinationality variable (which ranged between 0 and 1) at which firm performance is maximized is zero, that is, when there is no multinationality. This arises because, as indicated in Table 2, the slope of the estimated multinationality–performance relationship in each case is negative over the entire range of permissible values of the multinationality variable. Hence, despite statistically significant evidence of a cubic relationship, the estimated multinationality–performance relationship is in fact negatively sloped over the entire range of the multinationality variable. This negatively sloped relationship is further revealed in the estimates of the quadratic specification which imply a U-shaped and not an inverted-U shaped relationship; in this instance the estimated coefficients imply that firm performance is minimized when the multinationality variable takes a value of about 0.58. The study reported that the multinationality variable had a sample mean of 0.04 and standard deviation 0.07. This indicates that the value of the multinationality variable that minimizes firm performance in the quadratic model greatly exceeds 3 standard deviations above its mean, suggesting that for virtually all firms in the study’s sample, the multinationality–performance relationship is negative. The study’s claim to present evidence in support of an S-curve hypothesis is true in the sense that the cubic model ‘‘fits the data,’’ but the estimated relationship is opposite that hypothesized. These remarks are simply intended to caution analysts

128

HARRY P. BOWEN

that when fitting nonlinear models considerable care must be taken to explore the nature of the estimated relationship.

MULTINATIONALITY AND PERFORMANCE: STATISTICAL ISSUES Data and Estimation Strategies The expression for firm performance given in (1) specifies the unit of analysis to be a single firm at a given point in time. In particular, the underlying revenue, cost and ‘‘modifier’’ functions are all firm-specific. Given the firmspecific nature of the relationship, it would seem that the proper method for estimating the effect of multinationality on firm performance would be to conduct a time-series analysis for a given firm. However, the traditional approach has been to estimate the relationship in a cross-section of firms, where the cross-sectional data are for a single year, or where the crosssectional data has been ‘‘expanded’’ in the time-dimension to form longitudinal data. The data sets frequently used in the literature contain many firms relative to the number of time periods and in this regard they are perhaps best considered to fall under the heading of panel data. Panel data are increasingly being used, given both the short-comings of single-year cross-sectional analysis (Bowen & Wiersema, 1999) and the increasing availability of data on firms over several years. One virtue of panel data is that it can allow for modeling dynamic elements, such as including lagged values of the dependent or independent variables in a model, although this has not been the focus of past work. But the primary virtue of panel data is that it allows one to take account of idiosyncratic differences among the cross-sectional units (firms); such differences are commonly referred to as (firm-level) heterogeneity. That heterogeneity is likely to be an issue for the multinationality–performance relationship is immediately evident from expression (1) since it explicitly indicates that, at a minimum, the revenue and cost functions are likely to be firm-specific. The potential importance of firm-level heterogeneity is further suggested by the common use in the literature (Hitt et al., 2006) of a firm-level variable to serve as a moderator for the multinationality variable since, by definition, an interaction variable formed using a firm-level characteristic is a recognition that the relationship between multinationality and performance differs across firms.11 However, this type of firm-level heterogeneity is only one dimension of potential heterogeneity; heterogeneity can also exist at the industry and at the country


129

level.12 Several statistical methods exist to incorporate and test for such heterogeneity, and how this is done depends on what the researcher is willing to assume about the nature and source of heterogeneity. Below, I note some of these estimation strategies in terms of the issues that might be expected to arise in an analysis of the multinationality–performance relationship. Is this regard, the discussion is meant to point to issues that appear to not have been adequately addressed in the literature, and that may serve to point to other reasons for the mixed results obtained in the literature. Anticipating the conclusion of this discussion, what appears needed is a more systematic approach to, and investigation of, potential heterogeneity in the multinationality–performance relationship.

Omitted Firm Specific Characteristics One underlying premise of the multinational-performance relationship is that the performance benefit arising from operating across international markets relates to the presence of firm-specific intangible assets. Data permitting, the presence of such intangible assets has been captured by variables such as a firm’s expenditure on R&D or advertising (both are usually measured per dollar of firm sales); such variables have generally been found to be significant control variables. However, it is likely that other firm-specific characteristics (managerial effort, ability, management’s embedded knowledge of foreign markets) have been omitted from consideration. Such omitted variables represent one source of heterogeneity in the level of performance across firms, and their omission can lead to inconsistent estimates if these omitted variables are correlated with any of the included variables in a model. Given this, failure to take account of this source of heterogeneity may be one source of the mixed results obtained in the literature. Panel data allow one to control for this source of heterogeneity using one of the two techniques: fixed effects and random effects (Greene, 2003). The choice between the two approaches hinges on whether the omitted firmspecific characteristics are assumed to be correlated (fixed effects) or uncorrelated (random effects) with the one or more of the variables included in the relationship. Since some of the omitted factors may themselves be antecedents of multinationality (Hitt et al., 2006), the fixed effect specification would seem preferred. Moreover, even if the omitted factors are not correlated with any model variables, the fixed effects specification is still appropriate. The main benefit of the random effects specification is that fewer ancillary parameters need to be estimated and hence it conserves on

130

HARRY P. BOWEN

degrees of freedom. However, since most of the panel datasets employed in recent literature have well over 1,000 observations, conservation of degrees of freedom is a less-convincing argument for the choice of the random effects specification. In any event, test procedures are available to assess the appropriateness of the random effects assumption of no correlation between omitted and included variables (Greene, 2003). A point worth noting is that both the fixed and random effects specifications assume that the excluded firm characteristics are constant over time. For this reason, these techniques prevent estimation of the effect of any variable included in a model that is constant over time, such effects are simply absorbed in estimating the firm-specific effects. Since most specifications of multinational-performance relationship use time varying variables, this would not seem to be a limiting factor. One instance where this could be important is for models that include industry-specific dummy variables (e.g. Contractor et al., 2003), if the industry membership of firms does not change over time. Despite these concerns, the limitations faced by researchers in obtaining data on potentially important firm-specific characteristics suggest that such omitted characteristics may be an important source of heterogeneity in the multinationality–performance relationship, and hence also a potential source of inconsistent estimates, and hence invalid inferences. This strongly argues for panel data and the use of either the fixed or random effects specification as standard aspects of any estimation strategy regarding the multinationality–performance relationship.

Model Heterogeneity Heterogeneity arising from the omission of firm-specific but time-invariant characteristics is a potential source of heterogeneity that may bias inference. However, this issue addresses heterogeneity only in terms of the level of performance across firms (i.e., the intercept); it does not address issues of potential heterogeneity in the relationship itself, that is, differences in (slope) coefficients across firms. Virtually all studies using panel data estimate the ‘‘fully-pooled model’’ that assumes that all model coefficients are the same across firms and constant over time. The assumption that slope coefficients are homogeneous across the cross-sectional units is a common assumption in applied work, and in the case of single-year cross-sectional data is in fact necessary if estimation is to proceed. But while the assumption of coefficient homogeneity is common, one can question whether it is an appropriate assumption in the context of the multinationality–performance relationship.


131

If one takes expression (1) to be a reasonable characterization of a multinationality–performance relationship then the assumption of coefficient homogeneity is not likely to be appropriate. As previously noted, the empirical multinationality–performance literature itself provides evidence of heterogeneity via investigations that consider whether the coefficient on the multinationality variable varies with (is moderated by) some selectively chosen firm-level variable (e.g., firm R&D intensity). While the use of an interaction variable goes some way toward addressing the issue of coefficient heterogeneity, such analysis takes too narrow a perspective on the more general issue of whether the coefficients in the multinationality–performance relationship can be assumed to be the same across the cross-sectional units (or over time). In this regard, estimation methods exist that can admit the hypothesis of coefficient heterogeneity without committing, as does the use of an interaction variable, to any particular source of heterogeneity. Of course, the source of heterogeneity may ultimately be of interest to researchers. However, before beginning the search for potential sources of heterogeneity the first order of investigation should be to examine and test for the presence of heterogeneity. General approaches for detecting whether heterogeneity is an issue fall under the heading of random coefficient models (Greene, 2003), and their use to date appears absent from the multinationality–performance literature. In its simplest form, a random coefficients model assumes that coefficient heterogeneity arises from stochastic (random) variation, that is, there is an underlying common coefficient so that differences in a given coefficient across firms arises from random variation. Like the ‘‘random effects’’ model described earlier in the context of omitted firm-specific characteristics, the random coefficients model assumes that the stochastic variation underlying the heterogeneity is constant over time. However, these assumptions can be relaxed to admit, for example, heterogeneity that depends on firm-level covariates.13 Space limitations prevent a full discussion of random coefficients methods, but the nature of the multinationality–performance relationship and the firm-level data often used to estimate this relationship suggest that variants of the random coefficient method may be particularly appropriate. In one sense, the issue of heterogeneity can be viewed not as an ‘‘if ’’ question, but rather at what level does heterogeneity arise. For example, it may be reasonable to assume that the revenue and cost conditions of firms within the same industry are the same, suggesting that the effect of multinationality on performance is homogenous for firms within the same industry but that it may differ between industries. Virtually, all studies of the

132

HARRY P. BOWEN

multinationality–performance relationship use industry control variables to capture differences in performance levels across firms that arise from differences in industry membership, but no systematic effort has been directed to assessing whether there are significant differences in the multinationality coefficient across industries.14 One hint that heterogeneity at the industry level is important comes from Contractor et al. (2003) who found that the multinationality–performance relationship differed for capital-intensive service sectors (air transport, restaurants, hotels, etc.) versus knowledge-intensive service sectors (advertising, market research, etc.). If differences across industries are significant, this may go a long way to explain the mixed results obtained in the literature, since the results that have been obtained represent a kind of ‘‘average’’ effect of multinationality on performance across industries. Hence, if the effect for some industries is negative (due to excessive multinationality) but positive for other industries, the average effect may be of any sign, or may simply not be statistically significant. Returning to the notion of different MoMs , it seems reasonable that one MoM may be relevant (and a source of advantage) in one industry, but that another MoM is relevant (and a source of advantage) in another industry. For example, the nature and advantages of multinationality to textile firms is likely to be completely different from those of pharmaceutical firms. It is even possible that controlling for heterogeneity may serve to obviate the question of whether the multinationality–performance relationship is nonlinear.

Endogenous Regressors The characterization of the multinationality–performance relationship given in expression (1) assigns to multinationality the role of being a choice variable for the firm. In this respect, it seems reasonable to assume that a firm’s degree of multinationality is indeed a choice under the control of the firm’s managers. But acknowledging that a firm’s degree of multinationality is a choice variable raises an important concern: this choice is likely to be endogenous with respect to firm performance. This arises since manager’s decisions regarding the depth or breath of the firm’s international presence are linked to the expected performance outcome of such decisions. More generally, the concurrent nature of the relationship between managers’ decisions and expected performance means that in any decision–performance relationship the decision variable and performance are jointly determined, that is, the decision variable is an endogenous variable in the performance


133

relationship.15 This raises problems for making valid statistical inferences since, in statistical terms, this source of endogeneity means that the error term in the multinationality–performance relationship will be correlated with the multinationality variable. Such correlation violates one of the basic assumptions of least squares estimation and hence estimation using OLS (or pooled OLS in the case of panel data) will yield inconsistent estimates. Recognizing and addressing the issue of endogeneity is therefore important for making valid inferences. Yet, on this issue, a review of recent studies of the multinationality–performance relationship (Bowen & Wiersema, 2007) indicated only one instance (Lu & Beamish, 2004) in which the issue of endogeneity was acknowledged. However, rather than adopt as their core model the one in which multinationality was treated as an endogenous regressor they chose only to remark that the results derived (but not reported) assuming endogeneity were qualitatively similar to the reported OLS estimates, that is, the endogeneity corrected estimates served only as a ‘‘robustness check.’’ In addition, no details were provided on how the issue of endogeneity was addressed. Several statistical methods exist to deal with this form of endogeneity and the attendant bias it induces. The most commonly used methods are based on the method of instrumental variables, which effectively uses a variable that is highly correlated with the endogenous right-hand-side variable, but uncorrelated with the error term in the relationship. Computationally, the instrumental variable technique is relatively straightforward and is implemented in most statistical packages; the crucial problem is selecting an appropriate variable to serve as an instrument. Perhaps the most widely known and used instrumental variables technique is two-stage least squares (2SLS). This technique uses predicted values of an endogenous right-hand-side variable as the instrumental variable. However, to implement this procedure one must specify a relationship between the endogenous variable and a set of variables, some of which differ from those already included in the performance equation.16 For the performance-multinationality relationship this means specifying a relationship for the multinationality variable.17 Developing such a relationship clearly requires that the researcher think more deeply about the underlying determinants of multinationality. In this regard, a useful starting point is the Hitt et al. (2006) summary of those studies that have investigated the antecedents of multinationality. The issue of endogeneity arising from simultaneity between a decision variable and firm performance is a potentially serious and often neglected issue in most empirical management research. A central problem is that the

134

HARRY P. BOWEN

underlying theoretical frameworks often offer limited assistance in specifying a set of systematic relationships that could be used as a foundation for specifying a model that could address the issue of causality. Structural equations modeling (SEM) and 2SLS are close cousins, and both seek to address issues of endogeneity. Few studies have approached the analysis of the multinationality–performance relationship with such methods, and hence the role that endogeneity bias may play in explaining the mixed findings of prior studies awaits further analysis.

Other Dimensions of Firm Scope The Product Dimension The IB literature (unlike the strategic management literature) has long recognized in its empirical investigations that firm scope comprises two key dimensions: the product markets in which a firm is active (product diversification) and the geographic reach of its activities (international diversification). In particular, all major studies of multinationality– performance relationship in the IB literature include some measure of product diversification as a control variable, and in some cases this variable has been used to investigate if a firm’s product diversification moderates the multinationality–performance relationship (Hitt et al., 2006). However, if one accepts that multinationality is a choice variable for the firm and that it is therefore endogenous in the performance relationship, then the same must also be accepted for decisions regarding a firm’s product diversification. That is, both product diversification and international diversification are likely to be endogenous variables in the multinationality–performance relationship. Moreover, these two choice variables are unlikely to be independent. To my knowledge, no peer-reviewed study has yet dealt simultaneously with both these issues.18 That both the degree of multinationality and product diversification are likely to be endogenous in the performance relationship, adds further concerns regarding the validity of the inferences drawn from past studies of the multinationality–performance relationship. In this regard, endogeneity bias associated with firms’ product diversification strategies can be added as another potential explanation for the mixed results obtained in prior research. The Country Dimension Few studies of the multinational-performance relationship have incorporated the country dimension (Hitt et al., 2006). Most studies have used U.S.


135

or Japanese data, the latter due to its richer availability of data on the geographic distribution of FDI by Japanese firms. However, home country characteristics are also likely to be an important determinant of firms’ international success, as are the characteristics of the countries that are the target of a firm’s international expansion or the location of its foreign activities. One insight into the country of origin dimension comes from Ruigrok and Wagner (2003), who estimated a multinationality–performance relationship for German firms and found evidence of a U-shaped relationship. They suggested their finding reflected that international expansion by German firms is, by virtue of their location, initially into markets that are both smaller and less familiar (Switzerland and Austria) than are those available to, for example, U.S. firms (e.g., Canada) in the early stages of internationalization. As a result, they initially face higher costs associated with the liability of foreignness that would say U.S. firms. More generally, the framework of Porter (1990) makes central that homemarket characteristics can have important implications for the success of a nation’s firms in international markets. By restricting the sample to firms from only one country, this important dimension is relegated to the constant term of the relationship (which is often estimated to be negative – a curious outcome). The potential importance of home country effects can be thought to be another layer of the issue of heterogeneity of the multinationality– performance relationship. While the use of data on firms from only a single country obviates this concern, future analysis should seek to incorporate the country dimension by expanding the country coverage of the data. Doing so would then permit one to examine for heterogeneity in the multinationality– performance relationship at three levels: the firm, the industry, and the country. The issue of the country of destination of a firm’s international activities may also be important and can in part be seen as delineating more carefully the nature of firm’s international involvement, that is, in delineating alternative modes of multinationality. It can also be seen as an attempt to incorporate elements of ‘‘distance’’ into the relationship. In this respect, the spatial dimension of firms’ international activities is for the most part conspicuously absent in most of the multinationality–performance literature. The widespread use of total foreign sales in computing the ratio of foreign to total sales, or even the geographic distribution of a firm’s sales, fails to take account of the spatial distance across markets. This ‘‘flat earth’’ perspective goes against many of the underlying premises for expecting international diversification to be associated with higher firm performance.

136

HARRY P. BOWEN

That the location of firm’s international involvement may be important is suggested by the work of Pantzalis (2001), who separated the geographic scope of MNC activity into two regions: developed and developing country markets. He found significantly higher performance for firms active in developing country markets, but no significant performance difference for firms operating in developed country markets. This suggests that where firms operate may be a further, and important, piece of the puzzle. In one respect, the attempts by researchers to explore the role of cultural differences in determining differential firm performance is one attempt to incorporate ‘‘distance’’ into the multinationality–performance model, but to date this has mainly been directed at assessing if such differences influence the mode of foreign market entry (Tihanyi, Griffith, & Russell, 2005). However, the focus of the current volume on the regional activities of MNCs can be considered a step in the direction of delineating the importance of where (and perhaps how) firms operate for understanding the impact of multinationality on performance. In addition, narrowing the scope of where firms originate, and where they operate, in terms of their international activities to relatively homogenous geographic regions may serve to obviate the influence of potential heterogeneity in the multinationality–performance relationship that relates to differences in the international locations where firms operate.

Sample Selection Bias A further issue potentially affecting the reliability of estimates of the multinationality–performance relationships is sample selection bias. If firms select to become multinational because they have superior performance (needed to overcome the liability of foreignness) then internationally diversified firms would be expected to evidence higher performance. Effectively, the sample consists of higher-performing firms who have, by definition, chosen to become multinational precisely because they possess advantages sufficient to offset the higher costs of operating in foreign markets. The statistical problem that arises is that the sample of firms can no longer be considered a random sample. The sample selection issue is a concern in many branches of statistical inquiry, and has been the subject of a large literature dealing with the evaluation of treatment programs (Greene, 2003). A common approach for addressing this concern is to employ the ‘‘Heckman’’ sample selection model (Greene, 2003). In the present context, this model would involve two


137

equations; one (the selection equation) models the decision of a firm to become multinational which, among other variables, is likely to include firm performance, and a second equation that specifies the relationship between multinationality and firm performance. The estimates derived from the selection equation are then used in the multinationality–performance equation to take account of an essentially omitted variable (Greene, 2003).

Cross-Sectional Variation and Time-Series Behavior In theorizing the nature of the relationship between multinationality and performance the arguments advanced implicitly, if not explicitly, refer to the evolution of a firm’s performance as its degree of multinationality changes over time. Consideration of expression (1) suggests that rising multinationality can logically be separated from the level of a firm’s activity (production). In this sense, the degree of multinationality acts as a kind of ‘‘input’’ that imparts positive or negative effects on the firm’s revenue and cost. However, the potentially firm-specific nature of these effects raises the problematic question of whether estimation using cross-sectional or panel data on firms with different degrees of multinationality can be interpreted as indicative of the time path that would be followed by an individual firm. This is a complicated question for which I have no ready answer. However, one item of note in this regard is that the timing and scale of firms’ entry into international markets could result in the finding from a crosssectional analysis to differ markedly from that obtained from a time-series analysis. For example, some firms may have expanded rapidly and on a large scale into international markets while others have expanded more slowly and at a lower scale (this may be a function of the industry of the firm, a reference to my earlier remarks regarding industry-level heterogeneity). If multinationality is measured by the foreign sales ratio, it is likely that the ‘‘slower’’ firm will also have a lower foreign sales ratio, so the timing (and scale) of firms’ international involvement in a cross-sectional analysis would tend to correlate with the size of firms’ international sales. However, a measure such as foreign assets to total assets, or number of affiliates, may assign a higher degree of multinationality to the newer, larger scale, entrant. In this case, the cross-sectional relationship between performance and degree of multinationality may be found to be negative, since the newer, but larger scale entrant, is more likely to be incurring a loss. What this calls for are measures of multinationality that can better capture not only the depth or breath of international involvement, but also the time span of such involvement.

138

HARRY P. BOWEN

Finally, another aspect related to the time dimension is that knowledge acquisition and learning are seen as important for the performance gains from operating internationally. And, as previously noted, a key aspect underlying the expectation of a nonlinear relationship is that the (fixed) costs associated with international expansion declines the longer the firm’s presence in the market. This suggests that the extent of firms’ learning would be expected to figure prominently in the multinationality–performance relationship. At present, such learning is implicitly thought to be captured by the relative extent of a firm’s involvement in foreign markets (e.g., via the foreign sales ratio). However, this may or may not accurately reflect accumulated knowledge. If not, then accumulated experience would be an important omitted variable, and may be another source of heterogeneity in the relationship. Since learning can be expected to evolve over time, the fixed or random effects specifications are not suited to capturing the influence of sequential learning. Mimicking the learning curve literature, a variable that may better capture accumulated learning with respect to operating in foreign markets would be a firm’s accumulated sales in overseas markets, where the accumulation sales variable could also be distinguished by country or region so as to capture the potentially different kinds of knowledge being learned.

CONCLUDING REMARKS In this paper, I have considered some issues that appear relevant to the empirical investigation of a hypothesized relationship between firm performance and a firm’s degree of multinationality.19 In this regard, I have considered certain ‘‘technical’’ details regarding the proposition that the relationship may be nonlinear, and I have suggested that a quadratic specification is both parsimonious and sufficient to capture the key aspects of this proposition. I have also noted issues regarding specification and estimation of the relationship, with an eye toward resolving issues that if properly addressed may help explain the mixed findings heretofore presented in the literature. Among the issues raised, I consider two as central: first is the possibility of inconsistent estimates arising from the multinationality construct being endogenous in the performance relationship. The second is that the underlying relationship may be heterogeneous across firms (or industries or countries). Exploring these issues constitute important directions for further applied research. However, doing so will require researchers to


139

commit to developing more specific models of multinationality, and to employing and exploiting techniques that make use of the richness of longitudinal data. I have also suggested that more clarity regarding the multinationality– performance relationship would be obtained by a better delineation of the different modes in which multinationality manifests itself, so that we can begin to untangle the different sources of performance that are encapsulated in the underlying theoretical explanations.20 The theoretical arguments advanced for the relationship point to a number of contributing factors, but the analysis to date offers little understanding of the relative importance of the alternative factors. At one level, it is almost tautological that multinationality and performance are positively related, at least if we believe firms undertake international activities with the (expectation) of performance gains. However, decisions that appear appropriate ex ante may be bad ex post, and it is the ex post outcomes that are captured by the data. Ultimately, the role of expectations may become an important component of analysis. However, for the moment there appears to be a sufficiently rich set of issues regarding both data and estimation that can occupy our attention.

NOTES 1. For a given share variable Si, (e.g., share of region i’s sales in total firm sales) P the Entropy measure is E ¼ N i¼1 S i lnðS i Þ, while the Herfindahl measure is PN computed as H ¼ i¼1 ðS2i Þ. Higher values of E indicate more similarity in share values, and E reaches it maximum value when all shares are equal. Higher values of H instead indicate less similarity across share values so the inverse of the H values is often used. In each case, greater similarity across the share values is taken to indicate greater diversity. 2. Of course, a firm’s activity level may also depend on these firm-specific factors and hence they could also be included as arguments in the revenue and cost functions. I separate these factors to capture notions of firm-specific attributes such as brand name, firm specific knowledge, etc. However, where these factors appear in expression (1) is not germane to the basic issues I will discuss. 3. It can be questioned whether the ‘‘modifier’’ functions should be considered to be firm specific. 4. As will be discussed later in the paper, that the degree of multinationality is a choice variable for the firm implies that this variable is endogenous in the multinational-performance relationship. 5. One could also write (1) as Pit ¼ Yi ðM it ; Zit Þ½Ri ðPit ; Qit Þ C i ðQit Þ, in which case the influence of multinationality and firm-specific characteristics would have the same or ‘‘neutral’’ effect on revenue and on cost.

140

HARRY P. BOWEN

6. More precisely, the dashed segment of profit relationship arises if the rate at which incremental cost falls is greater than the rate at which incremental revenue falls. 7. The data were derived from COMPUSTAT as described in Bowen and Wiersema (2007). 8. This suggests seeking a taxonomy of MoMs similar in spirit to the strategic management literature’s search to understand the performance implication of product market diversification, a search that eventually produced consensus that the key distinction is between related versus unrelated diversification. 9. If the multinationality-performance relationship is viewed as describing the time path of a firm’s revenue and cost in a given foreign market, then an S-shaped profit relationship can arise if revenues evolve along an S-shaped path that converges to some upper limit, while costs evolve along a sloping path that converges to some lower limit. This implies an S-shaped profit relationship that also approaches an upper bound. 10. For example, if the driving force on the cost side is the ability of the multinational firm to create and benefit from economies of scope, then an approach that focuses directly on the estimation of the cost function, and that examines how this behaves in relation to say the extent of geographic dispersion of a firm’s activities, could allow for shaper conclusions about the underlying benefit of this type of multinationality. 11. In using an interaction variable, which variable is considered the moderator and which is the focus variable is somewhat problematic. For example, Lu and Beamish (2004) use firm-level R&D to moderate the relationship between multinationality and performance. But one could take the opposite view: the effect of firm-level R&D on performance is moderated by a firm’s degree of multinationality. This ultimately raises questions of causality, which unfortunately cannot be answered by the use of interaction variables. 12. Country-level heterogeneity could potentially relate to both a firm’s country of origin as well the foreign countries in which the firm operates. 13. This effectively reduces to using interaction variables. However, important is the one can use the random coefficients model as a basis to first explore the assumption of coefficient homogeneity, without committing to any particular source of the homogeneity. 14. This could easily be examined by interacting the multinationality variable with a set of industry dummy variables and testing for equality of the estimated coefficients. 15. Hamilton and Nickerson (2003) make the same point in discussing the empirical strategic management literature. 16. This requirement is one method for resolving the ‘‘identification problem’’ that can arise in a multiple-equation models (Greene, 2003). 17. Bowen and Wiersema (2007) present a model to address the issue of endogeneity with respect to the multinationality–performance relationship in exactly this way. 18. Bowen and Wiersema (2007) developed a model in which both international and product diversification are endogenous in the performance relationship. Controlling for endogeneity as well as firm-level heterogeneity arising from omitted firm-specific characteristics, they find that international diversification (multinationality) and


141

product diversification both evidence an inverted U-shape with firm performance, suggesting that the nonlinear specification is robust to the issues of endogeneity and firm heterogeneity. Their framework also allowed assessment of the nature of the direct relationship between international and product diversification. They concluded that these scope decisions represent substitute strategies within the firm. 19. One item omitted is discussion of the performance measure. While the choice of measure is often dictated by the data on hand, my preference is for market-based measures since they in principle capture the expected long-term performance consequences of firms’ current decisions. 20. This echoes Osegowitsch & Zalan’s (2005) concern that the multinationality– performance literature has ‘‘a persistent problem of under-specification.’’

REFERENCES Bowen, H. P., & Wiersema, M. F. (1999). Matching method to paradigm in strategy research: Limitations of cross-sectional analysis and some methodological alternatives. Strategic Management Journal, 20, 625–636. Bowen, H. P., & Wiersema, M. F. (2007). International and product diversification: Their interrelationship and impact on firm performance. Vlerick Leuven Gent Management School Working Paper no. 2006/6. Caves, R. E. (1996). Multinational enterprise and economic analysis (2nd ed.). New York: Cambridge University Press. Contractor, F., Kundu, S., & Hsu, C. C. (2003). A three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34, 5–18. Dunning, J. H. (1981). International production and the multinational enterprise. London: George Allen & Unwin. Geringer, J., Tallman, S., & Olsen, M. (2000). Product and international diversification among Japanese multinational firms. Strategic Management Journal, 21, 51–80. Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between multinationality and performance. Journal of International Business Studies, 31, 173–188. Greene, W. H. (2003). Econometric analysis (4th ed.). New Jersey: Prentice-Hall. Hamilton, B., & Nickerson, J. (2003). Correcting for endogeneity bias in strategic management research. Strategic Organization, 1, 51–78. Hitt, M., Tihanyi, L., Miller, T., & Connelly, C. (2006). International diversification: Antecedents, outcomes and moderators. Journal of Management, 32, 831–867. Hymer, S. (1976). The international operations of national firms. Cambridge, MA: MIT Press. Kogut, B. (1997). The evolutionary theory of the multinational corporation: Within and across countries. In: B. Toyne & D. Nigh (Eds), International business: An emerging vision (pp. 470–488). Columbia, SC: University of South Carolina Press. Kogut, B., & Zander, U. (1993). Knowledge of the firm and the evolutionary theory of the multinational corporation. Journal of International Business Studies, 24, 625–645. Lu, H., & Beamish, P. (2004). International diversification and firm performance: The S-curve hypothesis. Academy of Management Journal, 47, 598–608.

142

HARRY P. BOWEN

Osegowitsch, T., & Zalan, T. (2005). Two decades of multinationality-performance research: The persistent problem of under-specification. Working Paper no. 5, Australian Centre for International Business. Pantzalis, C. (2001). Does location matter? An empirical analysis of geographic scope and MNC tihany market valuation. Journal of International Business Studies, 32(1), 133–155. Peng, M. (2004). Identifying the big question in international business research. Journal of International Business Studies, 35, 99–108. Porter, M. E. (1990). The competitive advantage of nations. New York: Macmillan. Rugman, A. (1979). International diversification and the multinational enterprise. Lexington, MA: Heath. Rugman, A. (1981). Inside the multinationals: The economics of internal markets. New York: Columbia University Press. Rugman, A. (2006). The regional dimension of UK multinationals. Academy of Management Proceedings. (Panel 21), mimeo. Rugman, A., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35, 3–18. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43, 63–84. Sullivan, D. (1994). Measuring the degree of multinationality of a firm. Journal of International Business Studies, 25, 325–342. Tihanyi, L., Griffith, D., & Russell, C. (2005). The effect of cultural distance on entry mode choice, international diversification, and MNE performance: A meta-analysis. Journal of International Business Studies, 36(3), 270–283.

PERFORMANCE EFFECTS OF INTERNATIONALIZATION STRATEGIES: A META-ANALYSIS Andreas Bausch, Thomas Fritz and Kathrin Boesecke ABSTRACT Our meta-analysis of 92 international samples, with a total sample size of 8,491, demonstrates that firms following internationalization strategies by means of external growth modes can realize a significant positive performance impact on firm performance (¯r ¼ 0:156). This performance effect is significantly stronger than for firms using external growth strategies in their home country (117 samples, with a total sample size of 29,998, r¯ ¼ 0:077). Moderating effects are found for the type of international business combination (mergers and acquisitions versus alliance) and the internationalizing firm’s region of origin, whereas the relatedness of the firms and the region entered show no moderating impact.

INTRODUCTION Over the last decades, internationalization has become a more and more important strategic option for firms to exploit, sustain, and create competitive advantages. Measured by the development of foreign direct investments (FDI), which represent a firm’s strongest commitment of Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 143–176 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13007-2

143

144

ANDREAS BAUSCH ET AL.

resources in foreign markets, the internationalization of firms has been spurred by cross-border mergers and acquisitions (M&A) since the late 1980s (Fig. 1). At the same time, international alliances have become an important mode for entering foreign countries (e.g. Beamish & Kachra, 2004; Geringer & Louis, 1989; Meschi, 2004; Kang & Sakai, 2000; Yeheskel, Newburry, & Zeira, 2004). The relevance of these two internationalization strategies has resulted in a wide range of empirical studies testing the performance implications of international M&A or alliances. However, the results of these studies are relatively heterogeneous giving no clear indication about the true performance effect of international M&A and alliances. In this chapter, we therefore summarize past research on the impact of M&A and alliances in foreign markets on firm performance with metaanalytic techniques and look at the potential moderating variables for this relationship. We contrast our results with a sample of purely national business combinations. For our purposes, M&A and alliances are merged together under the term business combinations. A business combination arises when two or more firms combine their activities in specific areas of business such that their joint activities are based on a sustainable interaction leading to a change in the economic autonomy of at least one partner. In the case of M&A, at least one of the transaction partners involved looses its economic autonomy. Alliances are associated only with a reduction in the economic 1600 FDI Inflows (in billion U.S. dollars) 1400

Cross-border M&A Volume (in billion U.S. dollars)

1200 1000 800 600 400 200

Fig. 1.

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

1987

0

Worldwide FDI Inflows and Cross-Border M&A Volume. Source: UNCTAD FDI database.

Performance Effects of Internationalization Strategies

145

autonomy of at least one partner. The alliances incorporated into this metaanalysis include contractual alliances as well as equity joint ventures. The latter ones involve the creation of a new organizational and legal entity. Meta-analysis allows the integration of findings from a large number of primary empirical studies by calculating an average effect size, weighted by sample size. In our analysis, we integrated studies analyzing the performance effects of international and national business combinations. To be classified as an international business combination, at least one of the partners in the transaction had to be located in a country other than that of the other transaction partner(s). The performance impacts were assessed with regard to three dimensions: (1) capital market (cumulative abnormal returns), (2) management surveys (managerial assessment), and (3) accounting (particularly return on assets and return on investments). The integration of 92 international samples, with a total sample size of 8,491, demonstrated that firms following internationalization strategies by means of external modes of growth can realize a significant positive impact on firm performance (¯r ¼ 0:156; 95% confidence interval ¼ 0.121:0.191). Additionally, this performance effect was shown to be significantly stronger than for firms that use external growth strategies in their home country (117 samples, with a total sample size of 29,998; r¯ ¼ 0.077). Furthermore, the observed wealth creation effects were stronger for international alliances than for international M&A. Our results also indicate that international business combinations by Asian firms lead to a stronger performance impact than international business combinations pursued by US or European firms, for which no significant performance differences were observed. At the same time, however, our findings revealed no significant differences between international business combinations tying into Asia, Europe or the United States. In the next sections, we derive hypotheses on the performance effects of international and national business combinations as well as of potential moderating variables, introduce our sample and applied statistical techniques, and then present and discuss our findings. The final section gives a conclusion and shows limitations of our approach.

THEORETICAL BACKGROUND AND HYPOTHESES The different rationales for business combinations are explained by several theoretical approaches and can be divided into seven categories: (1) market power, (2) operational efficiency, (3) financial arguments, (4) transaction

146


costs, (5) access to and transfer of resources, (6) organizational knowledge, and (7) agency problems.

Overall Performance Effects of National and International Business Combinations and Differences between International and National Business Combinations When a business combination results in an increase in negotiation and market power which can be used against buyers or sellers and thus in a reduction in price competition, this will be expected to lead to higher performance; horizontal M&A and alliances, in particular, can boost market power. Increases in operational efficiency can result from economies of scale or scope, experience curve effects, network externalities, and lower transaction costs (Casti, 1995; Coase, 1937; Scherer & Ross, 1990; Williamson, 1985). In financial theory, diversification is emphasized as an argument in favor of business combinations (Markowitz, 1952). Depending on the characteristics of a transaction with respect to uncertainty, asset specificity, and frequency, decreased transaction costs can lead to higher performance if they are fully or partly internalized (Williamson, 1975). Internalization typically goes along with lower transaction cost when it allows decreasing the potential of opportunistic behavior and behavioral uncertainty due to comparable value systems and mutual trust. The resource-based view (RBV) within strategic management research considers firm resources and capabilities as a main source of competitive advantages (Mitchell, 1994; Rumelt, 1984; Wernerfelt, 1984). In the presence of imperfect factor markets, business combinations can be a performance increasing instrument if they allow to bypass these market imperfections and thus to gain access to resources of the partner or to transfer own resources to the partner (Mitchell, 1994; Rumelt, 1984; Wernerfelt, 1984). Further arguments in favor of positive performance effects of business combinations are provided by the theory of organizational knowledge, which views them as a means by which firms can learn and develop new skills or retain existing capabilities (e.g. Haspeslagh & Jemison, 1991; Johanson & Vahlne, 1977, 1990; Kogut & Chang, 1991; Kogut & Zander, 1991). In agency theory, it is argued that business combinations will show a negative performance effect due to opportunistic behavior of managers in combination with information asymmetries (Ross, 1973). Furthermore, Jensen (1986) suggests in his free cash flow thesis a negative influence, as managers prefer investments rather than repayment of capital regardless of


147

an investments’ net present value. Besides these rationales for business combinations, costs associated with them have to be considered. First and foremost, transaction, financing, and integration costs play a decisive role when reflecting on the performance implications of business combinations. Only when the positive performance implications compensate more than the additional cost, positive net performance effects can be generated. The above-described general rationales for business combinations apply for both national and international transactions. However, theory has highlighted additional benefits for international business combinations. They enable a firm to directly exploit existing monopolistic advantages in a foreign country (Hymer, 1976). Internationalization theory researchers (Dunning, 1973; Hymer, 1976; Rugman, 1981) suggest that FDI allow firmspecific idiosyncratic resources to be transferred to other regional markets. Imperfections in the markets for intangible goods (e.g. immobile, intangible resources, or limited information) often lead to lower transaction costs in the case of internalization as compared to pure market transactions. To take advantage of the potential value that existing intangible resources can have on foreign markets, firms often internalize fully or in parts the market for these resources when internationalizing their business. The learning theory (Johanson & Vahlne, 1977, 1990) highlights that internationalization is an incremental learning process not available to domestically operating firms (Barkema & Vermeulen, 1998; Hamel, 1991). The knowledge acquired in the process of internationalizing a firm may allow generating competitive advantages vis-a`-vis national competitors. International portfolio diversification offers another argument in favor of international business combinations: An international context can offer additional advantages over a purely national diversification if capital markets are imperfectly integrated (Agmon & Lessard, 1977; Errunza & Senebet, 1981; Grubel, 1968). International business combinations can also allow firms to gain cheaper access to resources (Hood & Young, 1979), expand into new markets (Doukas & Travlos, 1988), bypass trade barriers (Danbolt, 2004), increase market power (Grant, 1987; Hamel & Prahalad, 1985), and leverage international tax differences (Scholes & Wolfson, 1990). However, negative performance effects associated with international business combinations might at least partly compensate the benefits of these transactions. Proponents of the learning theory argue that heterogeneity in markets increases the complexity of managing widespread business units and, thus, may exhaust managerial capacity (Jones & Hill, 1988; Roth & O’Donnell, 1996; Williamson, 1975). In fact, empirical studies show that a geographic dispersion of business activities often goes along with

148


communication, coordination, and motivation problems (Hofstede, 1980). In addition, an increased internationality of firms typically enhances the exposure to financial and political risks such as currency fluctuations, government regulations, and trade laws (Boddewyn, 1988; Brewer, 1981; Reeb, Kwok, & Baek, 1979). Despite the above-described possibly negative performance effects in national and international surroundings, existing corporate governance mechanisms and the market for corporate control (Jensen & Ruback, 1983; Sundaram & Black, 1992) should assure that the majority of business combinations pursued by firms increase firm performance. Thus, the discussed general rationales for business combinations led to our Hypothesis 1: H1. There exists a positive relationship between (national and international) business combinations and firm performance. Our discussion above shows that international business combinations compared to national may result in both additional positive performance effects and additional negative performance effects. In general, integration complexity is the main factor influencing negative performance effects in business combinations. Integration complexity is particularly a result of cultural differences which can stem from differences in national and corporate culture. As the latter are already relevant in national business combinations, the additional negative performance effects in international business combinations should not compensate the additional positive performance effects. We therefore further hypothesize that: H2. The business combination–performance relationship is stronger for international business combinations than for national business combinations.

Mergers and Acquisitions versus Alliances as Strategies of Internationalization For successful implementation of an internationalization strategy, consideration must be given to the choice of entry mode (Tallman & Yip, 2001). Acquisitions and alliances are different types of entry modes with varying degrees of control, commitment, risk, and different transaction costs associated with the commitment of resources (Calvet, 1984; Caves, 1982). Alliances are believed to have slower and less efficient control mechanisms and are therefore said to be more costly because they allow the alliance partners greater freedom to act opportunistically, which Hennart (1988)


149

argues will result in high negotiating and monitoring costs. However, these transaction costs may be reduced when alliance partners share the ownership of a separate legal entity or have a joint organizational entity, in which case the incentive for opportunistic behavior is likely to decrease (Pisano & Teece, 1989) because of a stronger alignment of interests, monolithic control, and diminished performance ambiguity (Das & Teng, 1996). When a firm uses an acquisition as a foreign market entry mode, it may have to deal with the problem of restructuring the acquired organization. Particularly in international acquisitions, the integration of different organizational cultures and management styles can cause increased restructuring costs (Bhaumik & Gelb, 2005). Differences in organizational culture may also exacerbate management control between two merging entities. These differences may limit the effectiveness of behavior-based control mechanisms that rely upon trust, value congruence, and respect. In turn, the acquiring company may only be able to use a restricted set of control mechanisms, which consequently may decrease the implementation efficiency of the organizational control process and increase the risk of opportunistic action by the acquired company’s work force (Woodcock, Beamish, & Makino, 1994). The costs of controlling for organizational differences result from the searching costs prior to the acquisition that are required to differentiate appropriate from inappropriate organizational cultures and subsequent to the acquisition, when a variety of management and organizational integration techniques may have to be used to merge the two cultures (Woodcock et al., 1994). The newly purchased firm may furthermore be resistant to change and impede organizational integration, thus preventing potential merger benefits from being fully or immediately realized (Ennew, Wong, & Wright, 1992). The resulting opportunity costs may be significant. Managers may induce agency costs when they become reluctant to release resources under their control even after an acquisition has diminished the need for them (Jensen, 1986). Further potential costs that may occur in acquisitions can be attributed to overpayment as acquisitions are characterized by the existence of information asymmetry between acquirer and target (Csiszar & Schweiger, 1994). The target company has an information advantage due to its superior knowledge about its industry, its internal resources, and the market for these resources and may thus set a price higher than the value of the business or the acquirer may overbid. Overpayment may also arise when managers overestimate potential synergies and their abilities to integrate and manage

150


the target firm (Roll, 1986) and consequently agree to pay an acquisition premium that is far too high, which, in turn, reduces firm performance. In alliances, the risk of overpayment is significantly less. For example, in a joint venture the risk of paying too much for resources is limited because all partners face a potential information asymmetry problem. Thus, it is not in the interest of any partner to induce the other partner to overpay or over commit because of the possibility of reciprocal retaliatory action. Reuer and Koza (2000) empirically analyzed the impact of information asymmetry on joint venture performance and found that joint ventures are particularly valuable in situations of information asymmetry. Balakrishnan and Koza (1993) state that when the costs of valuing and acquiring complementary assets are high and information asymmetries are severe – which seems more likely in international business combinations – alliances may be superior to other modes of governance between markets and hierarchies. Alliances can be a direct and co-operative source of local knowledge and may therefore be more effective in reducing liabilities of foreignness (Hymer, 1976; Kindleberger, 1969; Zaheer, 1995) which firms typically encounter in the process of internationalization (Beamish, 1999; Zacharakis, 1997). In light of the above arguments, our third hypothesis becomes: H3. The type of business combination moderates the effect international business combinations have on firm performance, such that the relationship between international business combinations and performance is stronger for international alliances than for international M&A.

Horizontal versus Heterogeneous International Business Combinations The degree of relatedness of combining firms has been a key issue in strategy research. Horizontal business combinations are mergers, acquisitions, and alliances that take place in the same industry or production stage whereas in heterogeneous business combinations the two transaction partners are from different industries. Strategy researchers have proposed that a higher degree of relatedness between combining firms should correspond with a higher firm performance (Rumelt, 1974). Horizontal business combinations may be driven by increased competition, globalization of product markets, deregulation, increasing convergence of consumer preferences, a desire to access a portfolio of international brands, and difficulty in establishing new brands (Capron & Hulland, 1999).


151

Especially in international horizontal acquisitions or alliances a dominant motive might be to acquire or align with well-established brands in fastgrowing segments of the same industry. Furthermore, a firm might be better able to fully exploit location advantages such as cheaper factor costs, because it already has the knowledge on how to utilize these factors efficiently in their home country (Dunning, 1980). In general, potential sources of value creation in horizontal business combinations are assumed to include collusive as well as operational synergies, e.g. increased market power through collusion or economies of scale in production and distribution or synergistic advantages through the exploitation of complementary strengths. Moreover, the exploitation of asymmetric managerial skills may also achieve synergetic effects. While it can be argued that this can also occur in heterogeneous acquisitions and alliances, this cannot be said for operational and collusive synergies (Ansoff, 1965; Lubatkin, 1983; Salter & Weinhold, 1979). Furthermore, the common basis upon which the sharing of skills and know-how can be developed is significantly smaller in heterogeneous business combinations. If the involved firms are horizontally related, they have a greater understanding of the operational context and of the need for certain courses of action and consequently may be able to work together more effectively. Because of the risks inherent with going abroad, firms might be better off to pursue investment opportunities in business endeavors and operational areas with which they are familiar (Doukas & Travlos, 1988). Heterogeneous business combinations might thus result in further transaction costs that may eventually exceed management capabilities and therefore offset the benefits attributable to internationalization (Tallman & Li, 1996). This seems inevitable as a firm already faces liabilities of foreignness in international business combinations; when entering a heterogeneous acquisition or alliance, managers are simultaneously faced with diversity in two dimensions: the new country they are operating in and the new business segment or production stage, thus creating a further draw on management resources. Finally, as international diversification is more complex than a horizontal business combination with a foreign partner, governance costs may also be higher. Our fourth hypothesis: H4. Relatedness of transaction partners moderates the effect international business combinations have on firm performance, such that the relationship between international business combinations and performance is stronger for horizontal than for heterogeneous international business combinations.

152


Region of Origin and Region Entered As macro-level factors differ between countries and regions, the benefits of international business combinations might depend on both the region of origin of the internationalizing firm and the region entered with the transaction. Firms with a higher degree of internationalization have already gained a certain internationalization experience (Johanson & Vahlne, 1977, 1990) e.g. due to dealing with foreign customers, unfamiliar government regulations, and trade laws. Such experiences in foreign markets should allow firms to increase the benefits of their international activities (Gomez-Mejia & Palich, 1997; Johanson & Vahlne, 1977, 1990). In fact, the latest World Investment Report of the United Nations underlines once again that European firms are on average operating at a broader international scope than firms from the United States or Asia (UNCTAD, 2006). Furthermore, empirical studies found that the internationalization–performance relationship varies between regions (compare, e.g. Geringer, Beamish, & Da Costa, 1989). At the same time, the cost of internationalization can also differ depending on the degree of similarity between the environment in the home country/region and the environment of the region entered. With increasing differences in the contextual settings, the complexity and cost of operating in foreign countries also increase (Kostova & Zaheer, 1999). Analogous to industry-specific entry barriers highlighted in industrial organization economics (Bain, 1959), cultural and institutional factors may form entry barriers to certain regions not allowing a transfer of competitive advantages across national borders (Kogut, 1985). These factors include heterogeneity in politico-regulatory environments (Delios & Henisz, 2000), levels of economic development (Woodward & Rolfe, 1993), and cultural conditions (Chang & Rosenzweig, 2001). Johanson and Vahlne (1977, 1990) posit in their internationalization process model two patterns in the internationalization of the firm. The first is that firms get engaged in specific country markets according to an establishment chain. The second postulates that firms enter new markets with successively greater psychic distance (Johanson & Wiedersheim-Paul, 1975). The reasoning behind both patterns is to cope with lack of experiences in foreign markets. Due to the outlined importance of country-specific and region-specific contextual factors the benefits and costs associated with internationalization are likely to vary depending on both the region an internationalizing firm is


153

originated in and the region entered by an internationalization step. Therefore, we hypothesize: H5. The region of origin moderates the effect international business combinations have on firm performance, such that the relationship between international business combinations and performance differs for firms from different regions. H6. The region entered moderates the effect international business combinations have on firm performance such that the relationship between international business combinations and performance differs for business combinations into different regions. Sample This meta-analysis integrates studies analyzing the relationship between business combinations and firm performance. In order to be included in the population, a primary investigation had to be published between 1980 and 2004, in English or German, and had to address the relationship between business combinations and performance. The literature search done to identify the primary studies was conducted both electronically and manually. A computer-based search was conducted using the Business Source Premier, ABI/Inform, and WISO I databases; the results of this search revealed not only primary investigations, but also narrative reviews, and two meta-analyses on the performance effects of M&A. The following search terms were used to identify the initial set of studies: ‘merger’, ‘acquisition’, ‘alliance’, ‘joint venture’, ‘M&A’ in combination with ‘performance’, ‘wealth creation’, ‘value creation’, ‘success’, ‘wealth effect’, and ‘shareholder value’. We also used the appropriate German translations of each of these terms to identify further studies. In a second step, the past issues of journals with a relatively high number of relevant studies were scrutinized and the reference sections of already identified studies and reviews were searched for bibliographic references leading to further relevant primary investigations. Finally, an Internet-based search was conducted using several search engines in order to reveal unpublished studies. These procedures offered reasonable assurance that we had identified all relevant studies. Note that samples used in multiple studies have been excluded. As mentioned before, we also wanted to compare the performance effects of international business combinations with the performance effects of

154

ANDREAS BAUSCH ET AL. Accounting (5%)

Survey (22 %)

Capital Market (73%)

Accounting (9%)

Survey (6 %)

Capital Market (85%)

Empirical Method in Primary Investigations (International Sample) Empirical Method in Primary Investigations (National Sample)

Monographs (4%) Working Papers (6%)

Journals (90%)

Publication Sources (International and National Sample) JFE (9%) SMJ (5%) MIR (5%) IBR (4%) Journals JBFA (11%) JIBS (6%) JF (5%) FM (5%)

Other (50%)

Fig. 2. Sample Description. Note: 92 correlations from 62 studies in the international sample and 117 correlations from 64 studies in the national sample; Survey, management questionnaires; Accounting, accounting-based analyses; Capital Market, event studies; JBFA, Journal of Business Finance & Accounting; JFE, Journal of Financial Economics; JIBS, Journal of International Business Studies; SMJ, Strategic Management Journal; JF, Journal of Finance; MIR, Management International Review; FM, Financial Management; IBR, International Business Review; Other, other academic journals.

national business combinations (Hypothesis 2). We therefore conducted a hierarchical meta-analysis. We first separated international from national business combinations and then divided both samples further for potentially relevant moderators. Ultimately, 62 relevant primary studies for international business combinations and 64 for national business combinations were identified. Our international sample yielded 92 correlations for the meta-analysis with a total sample size of N ¼ 8,491. The national sample has 117 correlations and a total sample size of N ¼ 29,998. Our samples comprise international and national M&A and alliances that took place between 1948 and 2002. Due to the methodical heterogeneity in the primary investigations, our samples comprise different types of empirical analyses (Fig. 2). Apparently, studies based on capital market data clearly dominate the research in this field (73% and 85% for the international and national samples respectively),


155

followed by management surveys (22% and 6%), and accounting-based analyses (5% and 9%). Combs, Crook, and Shook (2004) point out that organizational performance is a multidimensional construct and hence researchers should clearly differentiate between them. Therefore, we will also investigate whether the distinct measures of performance moderate the results in our sample. In the coding process, a relatively narrow event period was selected from the included event studies so that any influence of M&A and alliances on the stock market return shown be relatively precise and free of extraneous influence from other events. Performance measures using accounting-based analyses included return on investment, return on equity, and return on assets. When studies reported multiple performance measures, the average of these measures was included in the meta-analysis. Fig. 2 also gives an indication of the most important sources for the basic data. Most of the included studies stem from journals, whereas monographs (book chapters and dissertations) and working papers add only 4% and 6%, respectively. In Fig. 3, the heterogeneity of the results in the international and national samples and thus in the existing primary investigations can be seen. Although the results appear to suggest a significant positive business combination–performance relationship in both samples (58% and 53%, respectively), the performance effect can only be statistically quantified by means of a meta-analysis, taking into consideration effect measures, sample sizes, and sampling error.

METHOD The term meta-analysis was first introduced by Glass in the 1970s: ‘‘Metaanalysis refers to the analysis of analyses. I use it to refer to the statistical analysis of a large collection of analysis results from individual studies for the purpose of integrating the findings’’ (Glass, 1976, p. 3). In other words, meta-analysis statistically integrates existing empirical studies on a single research topic, thus allowing one to reach consistent overall conclusions. Unlike a narrative review, meta-analysis allows to not only systematically quantify the relationship being analyzed, but also account for sampling error, an important source of artificial variance. The meta-analytic methods applied in this study are based on the procedures proposed by Hunter, Schmidt, and Jackson (1982) and Hunter and Schmidt (2004). Most meta-analyses in strategic management research follow these techniques (compare, e.g. Dalton, Certo, & Roengpitya, 2003;

156

ANDREAS BAUSCH ET AL. International sample: 92 correlations from 62 studies; 5% level of significance Negative − Significant (3%) Positive − Non Significant (27%)

Negative − Non Significant (12%)

Positive − Significant (58%) National sample: 117 correlations from 64 studies; 5% level of significance Negative − Significant (18%) Positive − Non Significant (18%) Negative − Non Significant (11%)

Positive − Significant (53%)

Fig. 3.

Significance of Results in Primary Empirical Studies.

Dalton, Daily, Ellstrand, & Johnson, 1998; Daniel, Lohrke, Fornaciari, & Turner, 2004; Gooding & Wagner, 1985). The main purpose of combining and integrating existing empirical results is to determine an average effect size across the studies. Under the assumption that all studies originate from one population, the best estimate for the population correlation (r) is the weighted average correlation (¯r) in which each correlation is weighted by the individual study size: P ½N i ri r¯ ¼ P (1) Ni


157

with ri as the correlation in study i and Ni as the number of observations in study i. In calculating the r-statistic from primary investigations, a variety of procedures have been used. Where data on effect sizes (e.g. zero-order correlations or d-statistics) were not provided but the results of statistical tests were reported (e.g. t-test, F-test), formulas given by Glass, McGraw, and Smith (1981) and Hunter et al. (1982) were used to transform the significance tests into the r-statistic. Analogous to the weighted average correlation, the variance across studies (s2r observed variance) is determined by the weighted average squared difference between the observed correlations and the weighted average correlation: P N i ðri r¯Þ2 2 P sr ¼ (2) Ni Hunter and Schmidt (2004) ascertained eleven artifacts that can influence effect sizes. Due to a lack of data, we were only able to deal with sampling error – which, incidentally, accounts for most of the variability in effect sizes resulting from artifacts. This is assuming a reliability of 1.0 and no range restriction. To correct for the sampling error and to calculate the variance of the population (sr, residual variance) we subtracted the sampling error variance (s2e ) from the observed variance. When the observed variance can be entirely attributed to sampling error, the homogeneity of the sample is obvious. However, a residual variance often remains in the sample; this can be either a result of heterogeneity in the sample and thus an indicator of the existence of a different population or a result of remaining uncorrected artifacts. Therefore, it is necessary to test for homogeneity. Credibility intervals and the 75% rule are commonly accepted tests of homogeneity (Hunter & Schmidt, 2004). Credibility intervals are generated around the weighted corrected average correlation using the corrected standard deviation (sr). If the interval is large or includes zero, there is a high probability that several subpopulations exist; correspondingly, small credibility intervals not including zero indicate that the weighted average correlation is the best predictor of a single homogenous population. Koslowky and Sagie (1993) suggest on the basis of an empirical test a threshold of 0.11 to separate small from large credibility intervals. The 75% rule tests the homogeneity of the included studies by comparing the sampling error variance to the observed variance. If the sampling error variance is larger than 75%, the source of the remaining unexplained 25%

158


of the observed variance can be expected to be uncorrected artifacts and thus the population can be assumed to be homogenous (Hunter & Schmidt, 2004). When the total sample is found to be heterogeneous, a search for moderators is initiated. These moderators are derived from the underlying theories relevant to the topic under consideration or from characteristics of the primary investigations. In a next step, the total sample is divided into subsamples according to the moderators and a separate meta-analysis is performed for each subgroup. A moderating variable can be confirmed when the weighted average correlations differ in the two subgroups and the average residual variance is smaller than in the total sample. We verified the significance of the differences between the two subgroups via a z-test (significance levels of 1%, 5%, and 10%). To check the significance of our results we calculated 95% confidence intervals around the weighted average correlations. A 95% confidence interval that does not include zero is an indicator that there is a true relationship between the variables.

RESULTS AND DISCUSSION Overall Performance Effects of National and International Business Combinations and Differences between International and National Business Combinations Both international and national business combinations demonstrate, on average, a positive performance effect (Table 1). The significantly positive weighted average correlations of r¯ ¼ 0:156 and r¯ ¼ 0:077 confirmed Hypothesis 1. To test the robustness of our results, we additionally performed a file drawer analysis and calculated Failsafe N according to Rosenthal (1979). This is the number of new, filed, or unretrieved studies with a correlation of zero that is required to bring the overall probability to any desired level of significance (we applied a level of p ¼ 0.05). According to the calculated Failsafe N, 9,972 (18,801) studies with an effect size of zero would be necessary to make our findings for the international (national) sample insignificant. Comparing the weighted average effect size of the international business combinations with our second sample of purely national business combinations revealed a significantly stronger performance effect for the international business combinations (z ¼ 2.619), supporting Hypothesis 2.


159

Table 1. Results of Meta-Analysis. Sample/Subsample International BC National BC M&A Alliances Horizontal Heterogeneous From Asia From Europe From the US Into Asia Into Europe Into the US Accounting Survey Capital Market

K

T

r¯

s2r

s2e

s2r

s2e =s2r

92 117 38 54 5 7 19 16 25 9 6 26 5 20 67

8,491 29,998 2,703 5,788 163 340 1,406 1,558 2,962 441 320 1,518 516 2,082 5,893

0.156 0.077 0.116 0.174 0.089 0.156 0.274 0.138 0.095 0.2 0.207 0.289 0.25 0.187 0.136

0.029 0.069 0.034 0.026 0.015 0.038 0.02 0.04 0.095 0.038 0.021 0.085 0.011 0.042 0.025

0.01 0.004 0.014 0.009 0.03 0.02 0.012 0.01 0.008 0.019 0.017 0.014 0.009 0.009 0.011

0.019 0.065 0.020 0.017 0 0.018 0.009 0.03 0.087 0.019 0.004 0.07 0.003 0.033 0.014

0.356 0.056 0.408 0.341 1.957 0.518 0.571 0.249 0.087 0.499 0.80 0.17 0.772 0.215 0.445

95% Credibility

95% Confidence

0.112:0.424 0.121:0.191 0.421:0.576 0.03:0.125 0.161:0.392 0.058:0.174 0.081:0.43 0.132:0.217 0.089:0.089 0.064:0.241 0.109:0.421 0.012:0.3 0.091:0.456 0.225:0.322 0.2:0.477 0.09:0.187 0.483:0.674 0.06:0.131 0.07:0.469 0.11:0.289 0.079: 0.336 0.102:0.312 0.231:0.809 0.243:0.335 0.151:0.348 0.168:0.33 0.167:0.541 0.098:0.277 0.093:0.366 0.099:0.174

Failsafe N

Z

9,972 18,801 517 4,952 24 1,124 207 441 48 31 2,194 90 631 3,986

2.619,a 1.586,b 0.073c 2.272,d 0.542e 2.553,f 0.087g 0.988h 1.036i 2.22,j 0.946k 1.026l

P Note: K, number of correlations; T, total sample size ( Ni); r¯, weighted average correlation; s2r , observed variance; s2e , sampling error variance; s2r , residual variance; International/National BC, international/national business combinations; M&A/Alliances, international mergers & acquisitions/alliances; Horizontal/Heterogeneous, horizontal/heterogeneous international business combinations; From Asia/From Europe/From the US, international business combinations conducted by Asian/European/US firms; Into Asia/Into Europe/Into the US, international business combinations with a (target) company in Asia/Europe/the US; Accounting, accountingbased analyses; Survey, management questionnaires; Capital Market, event studies. Level of significance 0.01. Level of significance 0.05. Level of significance 0.1. a International vs. National BC. b M&A vs. Alliances. c Horizontal vs. Heterogeneous. d From Asia vs. From Europe. e From Europe vs. From the US. f From Asia vs. From the US. g Into Asia vs. Into Europe. h Into Europe vs. Into the US. i Into Asia vs. Into the US. j Accounting vs. Capital Market. k Accounting vs. Survey. l Capital Market vs. Survey.

The above-described results make clear that business combinations are in general connected with positive performance effects. Thus, despite the arguments of agency theory (Ross, 1973), the free cash flow thesis (Jensen, 1986), and the costs associated with these transactions, existing

160


corporate governance mechanisms seem to assure that in general business combinations with positive performance effects are conducted. However, international business combinations realize a stronger performance effect than national transactions. This suggests that further advantages can be realized in an international context, in particular regional exploitation of monopolistic advantages (Hymer, 1976), international transfer of idiosyncratic resources in conjunction with market imperfections (Coase, 1937), competitive advantages resulting from acquiring knowledge in the process of internationalization (Johanson & Vahlne, 1977, 1990), international portfolio diversification (Agmon & Lessard, 1977), cheaper access to resources (Hood & Young, 1979), entrance into new markets (Doukas & Travlos, 1988), increases in market power (Grant, 1987), and leverage of international tax differences (Scholes & Wolfson, 1990). Thus, despite enhanced complexity of widespread business units (Hofstede, 1980; Jones & Hill, 1988; Roth & O’Donnell, 1996; Williamson, 1975) and increased exposure to financial and political risks (Boddewyn, 1988; Brewer, 1981; Reeb et al., 1979), additional advantages realizable in international transactions seem to assure on average a net increase of the performance impact compared to national business combinations. The meta-analyses by Bausch and Krist (2007) and Wagner and Ruigrok (2004) on the relationship between the degree of internationalization and performance find a comparatively weaker but still significantly positive weighted average correlation of r¯ ¼ 0.059 and r¯ ¼ 0.04. Liability of foreignness (Hymer, 1976; Kindleberger, 1969; Zaheer, 1995) offers a potential explanation for our much stronger performance impact. Theory discusses the need for firm-specific advantages like organizational or managerial capabilities in order to overcome liabilities of foreignness (Buckley & Casson, 1976; Caves, 1982; Dunning, 1977; Hennart, 1982). However, especially business combinations may be an effective way to quickly overcome these obstacles since they allow firms to gain quick access to the market-specific resources and knowledge of the transaction partner and, thus, to reduce economic disadvantages compared to national competitors (Agarwal & Ramaswami, 1992; Asiedu & Esfahani, 2001; Teece, 1986). Both the 75% rule and the 95% credibility interval indicated that our sample of international business combinations was heterogeneous. Thus, other variables seem to moderate the relationship considered. We therefore conducted moderator analyses.


161

Mergers and Acquisitions versus Alliances as Strategies of Internationalization As stated in Hypothesis 3, international alliances demonstrate, on average, a stronger performance effect than do international M&A (¯r ¼ 0.174 versus r¯ ¼ 0.116). The difference in the performance effects is significant at the 10% level (z ¼ 1.586). Thus, the integration of the merging entities and associated restructuring costs in international M&A (Bhaumik & Gelb, 2005) probably hamper value creation. Agency costs (Jensen, 1986), costs related to change resistance of the acquired firm (Ennew et al., 1992), and high acquisition premiums might furthermore reduce firm performance in international acquisitions. Our results indicate that alliances provide greater opportunities for enhancing firm performance through internationalization than do M&A. Entering foreign markets through alliances reduces liabilities of foreignness by employing a cooperative approach to acquiring local market knowledge and minimizing resource overpayment and retaliatory action (Woodcock et al., 1994). In our national sample, we also found a stronger performance effect for alliances than for M&A (¯r ¼ 0:170 versus r¯ ¼ 0:070). The difference is significant at the 5%-level (z ¼ 2.157). Information asymmetry between bidder and target and the payment of high acquisition premiums may also occur in national M&A and may thus decrease performance.

Horizontal versus Heterogeneous International Business Combinations In a further effort to investigate potential moderators, we divided our sample into horizontal and heterogeneous business combinations. Wealth creation was still significantly positive in both subsamples (¯r ¼ 0.089 for horizontal and r¯ ¼ 0.156 for heterogeneous business combinations). However, performance effects did not differ significantly. Consequently, we were unable to identify relatedness of the firms involved as a moderator in the relationship between international business combinations and firm performance. However, these results need to be interpreted carefully, as the sample size was relatively small (N ¼ 5 for horizontal and N ¼ 7 for heterogeneous international business combinations). Moreover, as all studies in our horizontal and heterogeneous subsamples used the cumulated abnormal return as performance measure, this result has to be interpreted as the expectation of the capital market.

162


Particularly in the M&A literature, one finds many studies with conflicting empirical results concerning the impact of the target–bidder relatedness (most often defined as resource or product-market similarity). King, Dalton, Daily, and Covin (2004) found in their meta-analysis on postacquisition performance that relatedness of the combining firms does not significantly explain variance in post-acquisition performance. They found only a significant negative result for conglomerate mergers but only for the 1- to 60-month event window and not for day 0. We also investigated whether there are any differences between horizontal and heterogeneous business combinations in our sample of national transactions. But again, relatedness of the involved firms could not be confirmed as a moderator and thus confirms the findings of our international sample as well as the findings of King et al. (2004). While horizontal business combinations may profit from collusive and operational synergies, heterogeneous business combinations probably benefit from cheaper access to capital (Steiner, 1975) and improved income stability (Higgins & Schall, 1975). Moreover, in heterogeneous international business combinations, management presumably would look more closely at potential synergies and the planning process of such a transaction is probably more detailed. In contrast to international diversification where the management is most likely aware of the complexity of integrating geographically and product-diversified firms, managers often habitually assume a high synergy potential per se in horizontal transactions.

Region of Origin and Region Entered Moderating our sample with regard to the home region where the internationalizing company is situated allowed us to distinguish between international business combinations conducted by firms located in Asia, Europe, and the United States. In all three subsamples, the relationship between international business combinations and performance was significantly positive (¯r ¼ 0:274, 0.138, and 0.095). The z-values showed that the performance effects for international business combinations by Asian firms are significantly stronger than for European and US firms. No significant difference was observable between US and European firms. Although these results confirmed Hypothesis 5 in general, the significantly stronger performance effect for business combinations conducted by Asian firms was quite surprising. In fact, Asian firms are still relatively focused on their home region (Rugman & Collinson, 2006). Such a low average degree


163

of internationalization going along with less internationalization experiences should lead to a weaker performance effect compared to other regions with a higher average degree of internationalization (Gomez-Mejia & Palich, 1997; Johanson & Vahlne, 1977, 1990). Additionally, the low average degree of internationalization in conjunction with the major cultural differences Asian firms face when internationalizing outside Asia, should result in a high psychic distance these firms have to cope with in foreign markets (Johanson & Vahlne, 1977, 1990; Johanson & Wiedersheim-Paul, 1975). Rugman and Collinson (2006) argued that Asian firms are less internationalized as an outcome of firm-specific advantages tied to the home region of these firms. International business combinations allowing firms to gain access to firm-specific advantages of the target or partner can therefore be particularly valuable for Asian firms as compared to European or US firms. This effect may offer an explanation for the significantly higher performance impact observed for Asian firms. In addition to the firm-specific advantages associated with a specific region, the comparatively low degree of internationalization of Asian firms per se offers a rationale for our findings. Firms with a low degree of internationalization and therefore less experience in other market environments (Johanson & Vahlne, 1977, 1990) might particularly benefit from the market knowledge acquired in international business combinations (Asiedu & Esfahani, 2001; Teece, 1986). This effect might also reduce the psychic distance they face in foreign markets. In contrast to the home region, the target region of an international business combination did not moderate the observed overall relationship. Thus, we could not confirm Hypothesis 6. Among international business combinations venturing into Asia, Europe, and the United States, no significant performance differences could be observed. In other words, the performance effects of international business combinations are, on average, not influenced by the region entered. Unfortunately, the included empirical studies neither allowed distinguishing between market entry in high-growth and low-growth countries nor enabled us to analyze other regions than the above described.

Empirical Method of Primary Investigations Our moderator analysis of the empirical method used in the primary investigations showed a significant difference between the performance effects found in studies using accounting and capital market data (z ¼ 2.22). The subsample of studies applying accounting data showed significantly

164


stronger performance effects (¯r ¼ 0.249) than did the subsample of capital market studies (¯r ¼ 0.136). This result is surprising, as one might assume that accounting-based analyses, due to their short-term perspective, would show a smaller firm performance effect than studies using capital market data. The latter are based on the capital market’s overall unbiased assessment of the present value of the future benefits of the business combination – and thus also allow a consideration for longer-term impacts. An explanation for this result might be that the stock market views external growth modes in an international setting as being far more complex and is, thus, more critical in its evaluation of the potential for synergy. Furthermore, investors may also have difficulties in judging the potential synergies of an international business combination as they might have less knowledge of foreign markets. We could not observe any significant differences between studies using accounting data and studies using managerial judgment as a measure of firm performance; neither was there a significant difference between capital market studies and management surveys. The discussion of which performance measures should be used has a long tradition in management research (compare, e.g. Combs et al., 2004; Keats, 1990; Varadarajan & Ramanujam, 1986; Venkatraman, 1989; Venkatraman & Ramanujam, 1986). As can be seen from our results, different measures of firm performance may show differences in the amount of the performance effect, although the direction of the performance impact stays the same. Therefore, it is advisable to use more than one performance dimension, where possible, in order to get a better understanding of the actual performance effect. Nevertheless, our results have to be interpreted with caution as the sample size was relatively small, with N ¼ 5 for studies using accounting data.

CONCLUSION AND LIMITATIONS The intention of our analysis was threefold: (1) to examine the performance effects of international and national business combinations on a meta-analytic basis, (2) to compare the performance effects observed for international business combinations with the performance effects of national business combinations, and (3) to identify moderating variables in the relationship between international business combinations and firm performance. By meta-analytically integrating 92 samples of international business combinations we were able to show that internationalization via business combinations has a significant positive performance effect. The comparison


165

with the meta-analytic results for 117 samples consisting only of national business combinations revealed significantly stronger performance effects in an international context, although the overall effect is also significantly positive for national business combinations, suggesting that further performance-enhancing effects are realizable in the case of international business combinations. The moderating effects for both type of business combination and country of origin were confirmed by our analysis. We found significantly stronger performance effects for alliances than for M&A and a significantly stronger performance effect for international business combinations conducted by Asian firms as compared to those of US or European firms. Neither the relatedness of the transaction partners nor the target country had a moderating effect on our overall results. With regard to the applied statistical methods in the primary investigations, we only found a significant difference between accounting and capital market data. For studies using accounting data, we observed a significantly stronger performance effect. Although our results resolve contradictions in past empirical research on the performance effects of international growth strategies by means of M&A and alliances, they can only be a starting point for further analyses. An analysis of further moderators influencing the performance effects of international business combinations might lead to a finer grained understanding of the contextual factors firms face when pursuing external growth strategies in an international arena. At the same time, a comparison of the performance implications of other internationalization strategies with international business combinations might allow further conclusions on the suitability of specific internationalization strategies in specific environments. An interpretation of our results must be attentive to the limitations of meta-analytic approaches. Meta-analyses relying on bi-variate relationships cannot demonstrate causality per se. But as the event studies, accountingbased analyses, and management questionnaires included in our sample explicitly focused on the effect that business combinations have on performance, a reverse effect is rather unlikely. As with all meta-analytic studies, our rules for inclusion can be subject to question. The inclusion criteria used in this paper were consistent with the literature; however, studies adopting broader or narrower rules for inclusion might yield different results. Further limitations arise from the small sample size in subsamples as compared to the overall sample. Unfortunately, with the exception of sampling error, none of the included studies reported the information necessary for a correction of artifacts; our results can therefore only be interpreted as conservative estimates of the true

166


relationship. Moreover, a number of studies had to be omitted from our metaanalysis because they did not contain sufficient information for the computation of effect sizes. Finally, more detailed reporting in the primary investigations would have allowed further moderation of the sample. This highlights the need for more complete reporting of empirical data in published articles on this topic. In the future, statistical tests should be included or, at a minimum, zero-order correlations should be reported. More detailed reporting of research results can only lead to greater accuracy in our interpretation of empirical data and our ability to compare and draw conclusions.

REFERENCES Agarwal, S., & Ramaswami, S. N. (1992). Choice of foreign entry mode: Impact of ownership, location, and internalization factors. Journal of International Business Studies, 23, 1–27. Agmon, T., & Lessard, D. (1977). Investor recognition of corporate international diversification. The Journal of Finance, 32, 1049–1055. Ansoff, H. J. (1965). Corporate strategy. New York: McGraw-Hill. Asiedu, B. J., & Esfahani, H. S. (2001). Ownership structure in foreign direct investments. Review of Economics and Statistics, 83, 647–662. Bain, J. S. (1959). Industrial organization. New York: Wiley. Barkema, H. G., & Vermeulen, W. K. (1998). International expansion through start-up or acquisition: A learning perspective. Academy of Management Journal, 41, 7–26. Bausch, A., & Krist, M. (2007). The effect of context-related moderators on the internationalization–performance relationship: Evidence from meta-analysis. Management International Review, 47, 1–29. Beamish, P. W. (1999). The role of alliances in international entrepreneurship. In: R. Wright (Ed.), Research in global strategic management (Vol. 7), International entrepreneurship (pp. 43–61). Stamford, CT: JAI Press. Beamish, P. W., & Kachra, A. (2004). Number of partners and JV performance. Journal of World Business, 39, 107–120. Bhaumik, S. K., & Gelb, S. (2005). Determinants of entry mode choice of MNCs in emerging markets. Emerging Markets Finance & Trade, 41, 5–24. Boddewyn, J. J. (1988). Political aspects of MNE theory. Journal of International Business Studies, 19, 341–363. Brewer, H. L. (1981). Investor benefits from corporate international diversification. Journal of Financial and Quantitative Analysis, 16, 113–124. Buckley, P. J., & Casson, M. C. (1976). The future of the multinational enterprise. London: Macmillan. Calvet, A. L. (1984). A synthesis of foreign direct investment theories and theories of the multinational enterprise. Journal of International Business Studies, 12, 43–59. Capron, L., & Hulland, J. (1999). Redeployment of brands, sales forces, and general marketing management expertise following horizontal acquisitions: A resource-based view. Journal of Marketing, 63, 41–54.


167

Casti, J. L. (1995). The theory of networks. In: D. Batten, J. L. Casti & R. Thord (Eds), Networks in action (pp. 3–24). Berlin: Springer. Caves, R. E. (1982). Multinational enterprise and economic analysis. Cambridge, UK: Cambridge University Press. Chang, S. J., & Rosenzweig, P. M. (2001). The choice of entry mode in sequential foreign direct investment. Strategic Management Journal, 22, 747–776. Coase, R. H. (1937). The nature of the firm. Economica, 4, 386–405. Combs, J. G., Crook, T. R., & Shook, C. L. (2004). The dimensionality of organizational performance and its implications for strategic management research. In: D. J. Ketchen & D. D. Bergh (Eds), Research methodology in strategy management (Vol. 2, pp. 259–286). Greenwich, CT: JAI Press. Csiszar, E. N., & Schweiger, D. M. (1994). An integrative framework for creating value through acquisition. In: H. E. Glass & B. N. Craven (Eds), Handbook of business strategy (pp. 93–115). New York: Warren, Gorham, & Lamont. Dalton, D. R., Certo, S. T., & Roengpitya, R. (2003). Meta-analysis of financial performance and equity: Fusion or confusion? Academy of Management Journal, 25, 13–26. Dalton, D. R., Daily, C. M., Ellstrand, A. E., & Johnson, J. L. (1998). Meta-analytic reviews of board composition, leadership structure, and financial performance. Strategic Management Journal, 19, 269–290. Daniel, F., Lohrke, F. T., Fornaciari, C. J., & Turner, R. A. (2004). Slack resources and firm performance: A meta-analysis of their relationship with firm performance. Journal of Business Research, 57, 564–574. Das, T. K., & Teng, B.-S. (1996). Risk types and the inter-firm alliance structures. Journal of Management Studies, 33, 827–843. Delios, A., & Henisz, W. (2000). Japanese firm’s investment strategies in emerging economies. Academy of Management Journal, 43, 305–323. Doukas, J., & Travlos, N. (1988). The effects of corporate multinationalism on shareholder’s wealth: Evidence from international acquisitions. The Journal of Finance, 43, 1161–1175. Dunning, J. H. (1973). The determinants of international production. Oxford Economic Papers, 25, 1161–1175. Dunning, J. H. (1977). Trade, location of economic activity and the MNE: A search for an eclectic approach. In: P. O. Hesselborn & P. M. Wijkman (Eds), The international allocation of economic activity (pp. 395–418). New York: Holmes & Meier. Dunning, J. H. (1980). Toward an eclectic theory of international production: Some empirical tests. Journal of International Business Studies, 11, 9–31. Ennew, C., Wong, P., & Wright, M. (1992). Organizational structures and the boundaries of the firm: Acquisition and divestment in financial services. Service Industries Journal, 12, 478–496. Errunza, V. R., & Senebet, L. T. (1981). The effects of international operations on the market value of the firm. The Journal of Finance, 36, 401–417. Geringer, J. M., Beamish, P. W., & Da Costa, R. C. (1989). Diversification strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10, 109–119. Geringer, J. M., & Louis, H. (1989). Control and performance of international joint ventures. Journal of International Business Studies, 20, 235–254. Glass, G. V. (1976). Primary, secondary and meta-analysis of research. Educational Researcher, 5, 3–8.

168


Glass, G. V., McGraw, B., & Smith, M. L. (1981). Meta analysis in social research. Beverly Hills, CA: Sage. Gomez-Mejia, L. R., & Palich, L. E. (1997). Cultural diversity and the performance of multinational firms. Journal of International Business Studies, 30, 309–335. Gooding, R. Z., & Wagner, J. A., III. (1985). A meta-analytic review of the relationship between size and performance: The productivity and efficiency of organizations and their subunits. Administrative Science Quarterly, 30, 462–481. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies. Journal of International Business Studies, 18, 78–89. Grubel, H. G. (1968). International diversified portfolios: The welfare gains and capital flows. American Economic Review, 58, 1299–1314. Hamel, G. (1991). Competition for competence and inter-partner learning within international strategic alliances. Strategic Management Journal, 12, 83–103. Hamel, G., & Prahalad, C. K. (1985). Do you really have a global strategy? Harvard Business Review, 63, 139–148. Haspeslagh, P. C., & Jemison, D. B. (1991). Managing acquisitions: Creating value through corporate renewal. New York: Free Press. Hennart, J.-F. (1982). A theory of multinational enterprise. Ann Arbor, MI: University of Michigan Press. Hennart, J.-F. (1988). A transaction costs theory of equity joint ventures. Strategic Management Journal, 9, 361–374. Higgins, R. C., & Schall, L. D. (1975). Corporate bankruptcy and conglomerate mergers. Journal of Finance, 30, 93–113. Hofstede, G. (1980). Culture’s consequences: International differences in work-related values. Beverly Hills, CA: Sage. Hood, N., & Young, S. (1979). The economics of multinational enterprise. New York: Longman. Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting error and bias in research findings (2nd ed.). Thousand Oaks, CA: Sage. Hunter, J. E., Schmidt, F. L., & Jackson, G. B. (1982). Meta-analysis: Cumulating research findings across studies. Beverly Hills, CA: Sage. Hymer, S. H. (1976). The international operations of national firms: A study of direct foreign investment. Cambridge, MA: MIT Press. Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance and takeovers. American Economic Review, 76, 323–329. Jensen, M. C., & Ruback, R. S. (1983). The market for corporate control: The scientific evidence. Journal of Financial Economics, 11, 5–50. Johanson, J., & Vahlne, J. E. (1977). The internationalization of process of the firm: A model of knowledge development and increasing foreign market commitments. Journal of International Business Studies, 8, 23–32. Johanson, J., & Vahlne, J. E. (1990). The mechanism of internationalization. International Marketing review, 7, 11–24. Johanson, J., & Wiedersheim-Paul, F. (1975). The internationalization of the firm: Four Swedish cases. Journal of Management Studies, 12, 305–322. Jones, G. R., & Hill, C. W. L. (1988). Transaction cost analysis of strategy-structure choice. Strategic Management Journal, 9(2), 159–172. Kang, N.-H., & Sakai, K. (2000). International strategic alliances: Their role in industrial globalization. STI Working Paper, 2000/5. OECD Working Paper, 38. Paris: OECD.


169

Keats, B. W. (1990). Diversification and business economic performance revisited: Issues of measurement and causality. Journal of Management, 16, 61–72. Kindleberger, C. (1969). American business abroad. New Haven, CT: Yale University Press. King, D. R., Dalton, D. R., Daily, C. M., & Covin, J. G. (2004). Meta-analyses of postacquisition performance: Indications of unidentified moderators. Strategic Management Journal, 25, 187–200. Kogut, B. (1985). Designing global strategies: Profiting from organizational flexibility. Sloan Management Review, 21, 27–38. Kogut, B., & Chang, S. J. (1991). Technological capabilities and Japanese direct investment in the United States. Review of Economics and Statistics, 73, 401–443. Kogut, B., & Zander, U. B. (1991). Knowledge of the firm, combinative capabilities, and the replication of technology. Organization Science, 3, 338–387. Koslowky, M., & Sagie, A. (1993). On the efficacy of credibility intervals as indicators of moderator effects in meta-analysis. Journal of Organizational Behaviour, 15, 695–699. Kostova, T., & Zaheer, S. (1999). Organizational legitimacy under conditions of complexity: The case of the multinational enterprise. Academy of Management Review, 24, 64–81. Lubatkin, M. (1983). Mergers and the performance of the acquiring firm. Academy of Management Review, 8, 218–225. Markowitz, H. M. (1952). Portfolio selection. The Journal of Finance, 7, 77–91. Mitchell, W. (1994). The dynamics of evolving markets: The effects of business sales and age on dissolution and divestitures. Administrative Science Quarterly, 39, 575–602. Pisano, G., & Teece, D. J. (1989). Collaborative arrangements and global technology strategy: Some evidence from the telecommunication equipment industry. In: R. S. Rosenbloom & R. A. Burgelmann (Eds), Research on technological innovation, management, and policy (pp. 227–256). Greenwich, CT: JAI Press. Reeb, D. M., Kwok, C. C. Y., & Baek, Y. H. (1979). Systematic risk of the multinational cooperation. Journal of International Business Studies, 29, 263–279. Reuer, J. J., & Koza, M. P. (2000). Asymmetric information and joint venture performance: Theory and evidence for domestic and international joint ventures. Strategic Management Journal, 21, 195–197. Roll, R. (1986). The hubris hypothesis of corporate takeovers. Journal of Business, 59, 197–216. Rosenthal, R. (1979). The ‘‘file drawer problem’’ and the tolerance for null results. Psychological Bulletin, 86, 638–641. Ross, S. A. (1973). The economic theory of agency: The principal’s problem. American Economic Review, 63, 134–149. Roth, K., & O’Donnell, S. (1996). Foreign subsidiary compensation strategy: An agency theory perspective. Academy of Management Journal, 39, 678–703. Rugman, A. M. (1981). Internalization as a general theory for foreign direct investment: A reappraisal of the literature. Weltwirtschaftliches Archiv, 116, 365–375. Rugman, A. M., & Collinson, S. C. (2006). Asian business is regional; not global. In: M. Fratianni (Ed.), Regional economic integration (pp. 167–199). Amsterdam: Elsevier. Rumelt, R. P. (1974). Strategy, structure and economic performance. Cambridge: Harvard University Press. Rumelt, R. P. (1984). Toward a strategic theory of the firm. In: R. Lamb (Ed.), Competitive strategic management (pp. 556–570). Englewood Cliffs, NJ: Prentice-Hall. Salter, M. S., & Weinhold, W. A. (1979). Diversification through acquisition. New York: The Free Press.

170


Scherer, F. M., & Ross, D. (1990). Industrial market structure and economic performance. Boston: Houghton Mifflin. Scholes, M. S., & Wolfson, M. A. (1990). The effect of changes in tax laws on corporate reorganization activity. Journal of Business, 45, 31–48. Steiner, P. O. (1975). Mergers. Ann Arbor, MI: University of Michigan Press. Sundaram, A. K., & Black, J. S. (1992). The environment and international organization of multinational enterprises. Academy of Management Review, 17, 729–757. Tallman, S. B., & Li, J. (1996). Effects of international diversity and product diversity on the performance of multinational firms. Academy of Management Journal, 39, 179–196. Tallman, S. B., & Yip, G. S. (2001). Strategy and the multinational enterprise. In: A. M. Rugman & T. L. Brewer (Eds), Oxford handbook of international business (pp. 317–348). Oxford, NY: Oxford University Press. Teece, D. J. (1986). Transaction cost economics and the multinational enterprise. Journal of Economic Behavior and Organization, 7, 21–45. UNCTAD. (2006). World investment report, 2006: FDI from developing and transition economies: Implications for development. New York: United Nations. Varadarajan, N., & Ramanujam, V. (1986). Measurement of business performance in strategy research: A comparison of approaches. Academy of Management Review, 11, 801–814. Venkatraman, N. (1989). Strategic orientation of business enterprises: The construct, dimensionality, and measurement. Management Sciences, 35, 942–962. Venkatraman, N., & Ramanujam, V. (1986). Measurement of business research: A comparison of approaches. Academy of Management Research, 1, 801–814. Wagner, H., & Ruigrok, W. (2004). Internationalization and firm performance: Meta-analytic review and future directions. Paper presented at the Academy of International Business 46th Annual Meeting, Stockholm, Sweden. Wernerfelt, B. (1984). A resource-based view of the firm. Strategic Management Journal, 5, 171–180. Williamson, O. E. (1975). Markets and hierarchies. New York: Free Press. Williamson, O. E. (1985). The economic institutions of capitalism. New York: Free Press. Woodcock, C. P., Beamish, P. W., & Makino, S. (1994). Ownership-based entry mode strategies and international performance. Journal of International Business Studies, 25, 253–273. Woodward, D. P., & Rolfe, R. J. (1993). The location of export-oriented foreign direct investment in the Caribbean Basin. Journal of International Business Studies, 24, 121–144. Zacharakis, A. L. (1997). Entrepreneurial entry into foreign markets: A transaction cost perspective. Entrepreneurship: Theory & Practice, 21, 23–39. Zaheer, S. (1995). Overcoming liability of foreignness. Academy of Management Journal, 38, 341–363.

FURTHER READING Studies employed in the meta-analysis are given below. Not all of these studies are discussed in the chapter. Albrecht, S. (1994). Erfolgreiche Zusammenschlussstrategien: Eine empirische Untersuchung deutscher Unternehmen. Wiesbaden: Gabler-Verlag.


171

Andrade, G., Mitchell, M., & Stafford, E. (2001). New evidence and perspectives on mergers. Journal of Economic Perspective, 15, 103–120. Baird, I. S., Lyles, M. A., Ji, S., & Wharton, R. (1990). Joint venture success: A Sino-U.S. perspective. International Studies of Management & Organization, 20, 125–134. Balakrishnan, S., & Koza, M. P. (1993). Information asymmetry, adverse selection and joint ventures. Journal of Economic Behavior and Organization, 20, 99–117. Baradwaj, B. G., Dubofski, D. A., & Fraser, D. R. (1991). Bidder returns in interstate and intrastate bank acquisitions. Journal of Financial Services Research, 5, 261–273. Beitel, P., Schiereck, D., & Wahrenburg, M. (2002). Explaining the M&A-success in European bank mergers and acquisitions. Institute for Mergers and Acquisitions, Universita¨t Witten/Herdecke, Witten, Germany. Bo¨hmer, E., & Lo¨ffler, Y. (1999). Kursrelevante Ergebnisse bei Unternehmensu¨bernahmen: Eine empirische Analyse des deutschen Kapitalmarktes. Zeitschrift fu¨r betriebswirtschaftliche Forschung und Praxis, 51, 299–324. Bradley, M. (1980). Interfirm tender offers and the market for corporate control. Journal of Business, 53, 345–376. Bradley, M., Desai, A., & Kim, E. H. (1988). Synergistic gains from corporate acquisitions and their division between stockholders of target and acquiring firms. Journal of Financial Economics, 21, 3–40. Brooks, L., Feils, D. J., & Sahoo, B. K. (2000). The impact of shifts in forecasted earnings and systematic risk on acquiring firm shareholders wealth in domestic and international acquisitions. Journal of Financial and Strategic Decisions, 13, 1–6. Bu¨hner, R. (1990). Erfolg von Unternehmenszusammenschlu¨ssen in der Bundesrepublik Deutschland. Stuttgart: C.E. Poeschel Verlag. Bu¨hner, R. (1992). Aktiona¨rsbeurteilung Grenzu¨berschreitender Zusammenschlu¨sse. Zeitschrift fu¨r betriebswirtschaftliche Forschung und Praxis, 44, 445–459. Byrd, J. W., & Hickman, K. A. (1992). Do outside directors monitor managers? Evidence from tender offer bids. Journal of Financial Economics, 32, 195–221. Cable, J. R., Palfrey, J., & Runge, J. (1980). Economic determinants and effects of mergers in West Germany, 1964–1974. Zeitschrift fu¨r die gesamte Staatswissenschaft, 136, 226–248. Cakici, N., Hessel, C., & Tandon, K. (1991). Foreign acquisitions in the United States and the effect on shareholder wealth. Journal of International Financial Management and Accounting, 3, 39–60. Campa, J. M., & Hernando, I. (2004). Value creation in European M&As. European Financial Management, 10, 47–81. Carow, K., Heron, R., & Saxton, T. (2004). Do early birds get the returns? An empirical investigation of early-mover advantages in acquisitions. Strategic Management Journal, 25, 563–585. Cebenoyan, A. S., Papaioannou, G. J., & Travlos, N. G. (1992). Foreign takeover activity in the U.S. and wealth effects for target firm shareholders. Financial Management, 21, 58–68. Chakrabarti, A., Hauschildt, J., & Su¨verkru¨p, C. (1994). Does it pay to acquire technological firms? R&D Management, 24, 47–56. Chang, S.-C., & Chen, S.-S. (2002). The wealth effect of domestic joint ventures: Evidence from Taiwan. Journal of Business Finance & Accounting, 29, 201–222. Chang, S.-C., & Huang, P. C. (2002). Corporate multinationalism, organizational learning, and market reaction to international joint ventures: Evidence from Taiwan. Global Finance Journal, 13, 181–194.

172


Chatterjee, S. (1986). Types of synergy and economic value: The impact of acquisition on merging and rival firms. Strategic Management Journal, 7, 119–130. Chen, H., Hu, M. Y., & Shieh, J. C. F. (1991). The wealth effect of international joint ventures: The case of U.S. investment in China. Financial Management, 20(4), 31–41. Chen, S.-S., Ho, K. W., Lee, C. F., & Yeo, G. H. H. (2000). Investment opportunities, free cash flow and market reaction to international joint ventures. Journal of Banking & Finance, 24, 1747–1765. Cheng, L. T. W., & Davidson, W. N., III. (1993). Reinterpretation of the shareholder gains in sell-off transactions. Journal of Business Finance & Accounting, 20, 599–611. Cheng, L. T. W., Fung, J. K. W., & Lam, K. (1998). An examination of the determinants of stock price effects of US–Chinese joint venture announcements. International Business Review, 7, 151–161. Choi, D., & Philippatos, G. C. (1983). An examination of merger synergism. The Journal of Financial Research, 6, 239–256. Chung, I. Y., Koford, K. J., & Lee, I. (1993). Stock market views of corporate multinationalism: Some evidence from announcements of international joint ventures. The Quarterly Review of Economics and Finance, 33, 275–293. Conn, R. L., & Connell, F. (1990). International mergers: Returns to U.S. and British firms. Journal of Business Finance & Accounting, 17, 689–711. Cordeiro, J. J. (1993). The role of director interlocks, governance, and management incentives in explaining stockholder gains from joint ventures. Academy of Management Best Paper Proceedings, 12–16. Cornett, M. M., & Tehranian, H. (1992). Changes in corporate performance associated with bank acquisitions. Journal of Financial Economics, 31, 211–234. Crutchley, C. E., Enyang, G., & Hansen, R. S. (1991). Stockholder benefits from Japanese–U.S. joint ventures. Financial Management, 20(4), 22–30. Danbolt, J. (1995). An analysis of gains and losses to shareholders of foreign bidding companies engaged in cross-border acquisitions into the United Kingdom, 1986–1991. The European Journal of Finance, 11, 279–309. Danbolt, J. (2004). Target company cross-border effects in acquisitions into the UK. European Financial Management, 10, 83–108. Datta, D. K., & Grant, J. H. (1990). Relationship between type of acquisition, the autonomy given to the acquired firm and acquisition success: An empirical analysis. Journal of Management, 16, 29–44. Davidson, W. N., III., & Cheng, L. T. W. (1997). Target firm returns: Does the form of payment affect abnormal returns? Journal of Business Finance & Accounting, 24, 465–479. DeLong, G. L. (2001). Stockholder gains from focusing versus diversifying mergers. Journal of Financial Economics, 59, 221–252. Dennis, D. K., & McConnell, J. J. (1986). Corporate mergers and security returns. Journal of Financial Economics, 16, 143–187. Dhanaraj, C., Lyles, M. A., Steensma, H. K., & Tihanyi, L. (2004). Managing tacit and explicit knowledge transfer in IJVs: The role of relational embeddedness and the impact on performance. Journal of International Business Studies, 35, 428–442. Dickerson, A. P., Gibson, H. D., & Tsakalotos, E. (1997). The impact of acquisitions on company performance: Evidence from a large panel of UK firms. Oxford Economic Papers, 49, 344–361.


173

Ding, D. Z. (1997). Control, conflict, and performance: A study of U.S.–Chinese joint ventures. Journal of International Marketing, 3, 31–45. Draper, P., & Paudyal, K. (1999). Corporate takeovers: Mode of payment, returns and trading activity. Journal of Business Finance & Accounting, 26, 521–558. Dussauge, P., & Garrette, B. (1995). Determinants of success in international strategic alliances: Evidence from the global aerospace industry. Journal of International Business Studies, 26, 505–530. Eger, C. E. (1983). An empirical test of the redistribution effect in pure exchange mergers. Journal of Financial and Quantitative Analysis, 18, 547–573. Etebari, A. (1993). Market impact of announcements of joint ventures between U.S. firms and eastern and Central European countries: Early evidence. Global Finance Journal, 4, 103–123. Eun, C. S., Kolodny, R., & Scheraga, C. (1996). Cross-border acquisitions and shareholder wealth: Tests of the synergy and internationalization hypothesis. Journal of Banking & Finance, 20, 1559–1582. Fatemi, A. M., & Furtado, E. P. H. (1988). An empirical investigation of the wealth effects of foreign acquisitions. In: S. J. Khoury & A. Gosh (Eds), Recent developments in international banking and finance (Vol. 2, pp. 363–379). San Francisco, CA: Jossey-Bass. Finnerty, J. E., Owers, J. E., & Rogers, R. C. (1986). The valuation impact of joint ventures. Management International Review, 26, 14–26. Floreani, A., & Rigamonti, S. (2001). Mergers and shareholders’ wealth in the insurance industry. Milano: Universita` Cattolica del San Cuore. Franks, J. R., & Harris, R. S. (1989). Shareholder wealth effects of corporate takeovers: The U.K. experience, 1955–1985. Journal of Financial Economics, 23, 225–249. Fro¨hls, M. A., Keown, A., McNabb, M., & Martin, J. (1998). Growth opportunities, corporate governance and the market value of multinational joint ventures. Managerial and Decision Economics, 19, 13–29. Geringer, J. M., & Hebert, L. (1991). Measuring performance of international joint ventures. Journal of international Business Studies, 22(2), 249–263. Gerpott, T. J., & Hartmann, I. (2000). IntegrationsmaXnahmen und Erfolg von Unternehmensakquisitionen in Ostdeutschland. Die Betriebswirtschaft, 60, 97–111. Glaister, K. W., & P.J., B. (1999). Performance relationships in UK international alliances. Management International Review, 39, 139–147. Gleason, K. C., Mathur, I., & Wiggins, R. A. (2003). Evidence on value creation in the financial services industries through the use of joint ventures and strategic alliances. The Financial Review, 38, 213–324. Gregory, A. (1997). An examination of the long run performance of UK acquiring firms. Journal of Business Finance & Accounting, 24, 971–1002. Hannan, T. H., & Wolken, J. D. (1989). Returns to bidders and targets in the acquisition process: Evidence from the banking industry. Journal of Financial Service Research, 3, 5–16. Harris, R. S., & Ravenscraft, D. (1991). The role of acquisition in foreign direct investment: Evidence from the U.S. stock market. The Journal of Finance, 46, 825–844. He, L. T., Myer, F. C. N., & Webb, J. R. (1997). The wealth effects of domestic vs. international joint ventures: The case of real estate. Journal of Real Estate Research, 13(3), 349–358. Healy, P. M., Palepu, K. G., & Ruback, R. S. (1997). Which takeovers are profitable? Strategic or financial? Sloan Management Review, 38, 45–57.

174


Holl, P., & Kyriazis, D. (1997). Wealth creation and bid resistance in UK takeover bids. Strategic Management Journal, 18, 483–498. Houston, J. F., James, C. M., & Ryngaert, M. D. (2001). Where do merger gains come from? Bank mergers from the perspective of insiders and outsiders. Journal of Financial Economics, 4, 285–331. Isobe, T., Makino, S., & Montgomery, D. B. (2000). Resource commitment, entry timing, and market performance of foreign direct investments in emerging economies: The case of Japanese international joint ventures in China. Academy of Management Journal, 43, 468–484. Kaplan, S. N., & Weisbach, M. S. (1992). The success of acquisitions: Evidence from divestures. The Journal of Finance, 47, 107–138. Kennedy, V. A., & Limmack, R. J. (1996). Takeover activity, CEO turnover and the market for corporate control. Journal of Business Finance & Accounting, 23, 267–285. Kim, K., & Park, J.-H. (2002). The determinants of value creation for partner firms in the global alliance context. Management International Review, 42, 361–384. Krishnan, R., & Noorderhaven, N. G. (2003). Effects of structural conditions and process characteristics on alliance outcome. Academy of International Business Conference Proceedings, Tilburg University, Tilburg, The Netherlands. Kuipers, D. R., Miller, D., & Patel, A. (2003). The legal environment and corporate valuation: Evidence from cross-border takeovers. Working paper, Texas Tech University, Lubbock. Lee, C., & Beamish, P. W. (1995). The characteristics and performance of Korean joint Ventures in LDCs. Journal of International Business Studies, 26(3), 637–654. Litz, R. A., & Stewart, A. C. (1998). Franchising for sustainable advantage? Comparing the performance of independent retailers and trade-name franchisees. Journal of Business Venturing, 13, 131–150. Loderer, C., & Martin, K. (1990). Corporate acquisitions by listed firms: The experience of a comprehensive sample. Financial Management, 19, 17–33. Lowinski, F., Schiereck, D., & Thomas, T. W. (2004). The effect of cross-border acquisitions and shareholder wealth: Evidence from Switzerland. Review of Quantitative Finance and Accounting, 22, 315–330. Lu, Q. (2001). Do mergers destroy value? Evidence from failed bids. Department of Finance, Kellogg School of Management, Evanston, IL. Lubatkin, M. (1987). Merger strategies and stockholder value. Strategic Management Journal, 8, 39–53. Lui, S. S., & Ngo, H.-y. (2004). The role of trust and contractual safeguards on cooperation in non-equity alliances. Journal of Management, 30, 471–485. Lummer, S. L., & McConnell, J. J. (1990). Stock valuation effects of international joint ventures. In: S. G. Rhee & R. P. Chang (Eds), Pacific-basin capital markets research, (pp. 531–546). New York, NY: Elsevier Science Publishers. Luo, Y. (1997). Partner selection and venturing success: The case of joint ventures with firms in the People’s Republic of China. Organization Science, 8, 648–662. Luo, Y. (2002). Contract, cooperation, and performance in international joint ventures. Strategic Management Journal, 23, 903–919. Lyles, M. A., & Baird, I. S. (1994). Performance of international joint ventures in two Eastern European countries: The case of Hungary and Poland. Management International Review, 34(4), 313–329.


175

Makino, S., & Delios, A. (1996). Local knowledge transfer and performance: Implications for alliance formation in Asia. Journal of International Business Studies, 27(5), 905–927. Malatesta, P. H. (1983). The wealth effect of merger activity and the objective functions of merging firms. Journal of Financial Economics, 11, 155–181. Maquieira, C. P., Megginson, W. L., & Nail, L. (1998). Wealth creation versus wealth redistribution in pure stock for stock mergers. Journal of Financial Economics, 48, 3–33. Markides, C. C., & Ittner, C. D. (1994). Shareholder benefits from corporate international diversification: Evidence from U.S. international acquisitions. Journal of International Business Studies, 25, 343–366. McConnell, J. J., & Nantell, T. J. (1985). Corporate combinations and common stock returns: The case of joint ventures. Journal of Finance, 11, 519–536. Merchant, H., & Schendel, D. (2000). How do international joint ventures create shareholder value? Strategic Management Journal, 21, 723–737. Meschi, P.-X. (2004). Valuation effect of international joint ventures: Does experience matter? International Business Review, 13, 595–612. Meschi, P.-X., & Cheng, L. T. W. (2002). Stock price reactions to Sino–European joint ventures. Journal of World Business, 37, 119–126. Min, J. H., & Prather, L. J. (2001). Tobin’s Q, agency conflicts, and differential wealth effects of international joint ventures. Global Finance Journal, 12, 267–283. Mohanram, P., & Nanda, A. (1996). When do joint ventures create value? Working Paper No. 96-028, Harvard Business School, Boston, MA. Morck, R., & Yeung, B. (1992). Internationalization: An event study test. Journal of International Economics, 33, 41–56. Niden, C. M. (1993). An empirical examination of white knight takeovers: Synergy and overbidding. Financial Management, 22, 28–45. Park, J.-H., & Kim, K. (1999). International alliance formation and stock market responses: Industry globalization as an environmental contingency. Journal of International Management, 5, 257–284. Park, S. H., & Kim, D. (1997). Market valuation of joint ventures: Joint venture characteristics and wealth gains. Journal of Business Venturing, 12, 83–108. Pettway, R. H., Sicherman, N. W., & Spiess, D. K. (1993). Japanese foreign direct investment: Wealth effects from purchases and sales of U.S. assets. Financial Management, 22, 82–95. Picken, L. G. (2003). Unternehmensvereinigungen und shareholder value. Frankfurt am Main: Lang. Piske, R. (2002). German acquisitions in Poland: An empirical study on integration management and integration success. Human Resource Development International, 3, 295–312. Porrini, P. (2004). Can a previous alliance between acquirer and a target affect acquisition performance. Journal of Management, 30, 545–562. Prather, L. J., & Min, J. H. (1998). Testing of the positive multinational network hypothesis: Wealth effects of international joint ventures in emerging markets. Multinational Finance Journal, 2, 151–165. Raffeé, H., & Eisele, J. (1993). Erfolgsfaktoren des Joint Venture-Management: Grundlagen und erste Ergebnisse einer empirischen Untersuchung. Working Paper No. 97, Institut fu¨r Marketing, Universita¨t Mannheim, Mannheim, Germany. Ravichandran, R., & Sa-Aadu, J. (1988). Resource combinations and security price reactions: The case of real estate joint ventures. Journal of the American Real Estate & Urban Economics Association, 16, 105–122.

176


Saxton, T. (1997). The effects of partner and relationship characteristics on alliance outcomes. Academy of Management Journal, 40, 443–461. Scanlon, K. P., Trifts, J. W., & Pettway, R. H. (1989). Impact of relative size and industrial relatedness on returns to shareholders of acquiring firms. Journal of Financial Research, 12, 103–112. Schipper, K., & Thompson, R. (1983). Evidence on the capitalized value of merger activity for acquiring firms. Journal of Financial Economics, 11, 85–119. Seth, A., Song, K. P., & Pettit, R. (2000). Synergy, manageralism or hubris? An empirical investigation of motives for foreign acquisitions of U.S. firms. Journal of International Business Studies, 11, 387–405. Shaked, I., Michael, A., & McClain, D. (1991). The foreign acquirer bonanza: Myth or reality? Journal of Business Finance & Accounting, 18, 431–447. Sicherman, N. W., & Pettway, R. H. (1987). Acquisition of divested assets and shareholders’ wealth. The Journal of Finance, 42, 1261–1273. Sim, A. B., & Ali, Y. (1998). Performance of international joint ventures from developing and developed countries: An empirical study in a developing country context. Journal of World Business, 33, 357–377. Sudarsanam, S., Holl, P., & Salami, A. (1996). Shareholder wealth gains in mergers: Effects of synergy and ownership structure. Journal of Business Finance & Accounting, 23, 673–697. Tatoglu, E., & Glaister, K. W. (1998). Performance of international joint ventures in Turkey: Perspectives of Western firms and Turkish firms. International Business Review, 7, 635–656. Travlos, N. G. (1987). Corporate takeover bids, methods of payment and bidding firms’ stock returns. The Journal of Finance, 42, 943–963. Wang, Y., & Wu, C. (2004). The wealth effect of strategic alliances in Taiwan’s high-tech electronics industry. International Journal of Management, 21, 361–369. Wansley, J. W., Lane, W. R., & Yang, H. C. (1983). Shareholder returns to USA acquired firms in foreign and domestic acquisitions. Journal of Business Finance & Accounting, 10, 647–656. Weir, C. (1997). Corporate governance, performance and take-overs: An empirical analysis of UK mergers. Applied Economics, 29, 1465–1475. Yeheskel, O., Newburry, W., & Zeira, Y. (2004). Significant differences in the pre- and post-incorporation stages of equity international joint ventures and international acquisitions, and their impacts on effectiveness. International Business Review, 13, 613–636.

INTERNATIONALIZATION AND PERFORMANCE: THE MODERATING ROLE OF STRATEGIC FIT Fabienne Fortanier, Alan Muller and Rob van Tulder ABSTRACT Recent research on the internationalization–performance (IP) relationship has suggested that many of the different results can be explained by the role of moderating factors. This paper explores the hitherto underemphasized role of strategic fit between organizational structure on the one hand and industry pressures towards integration and responsiveness on the other hand. We suggest a new way of measuring organizational structure (and consequently strategic fit), based on archival data rather than questionnaires, and include these measures in our regression analysis on a sample of 332 Fortune companies.We find that strategic fit positively affects performance and moderates the shape, size and direction of the internationalization–performance relationship.


177

178

FABIENNE FORTANIER ET AL.

INTRODUCTION The relationship between internationalization and firm performance (the ‘IP relationship’) remains a much researched and fervently debated issue (Lu & Beamish, 2004; Contractor, Kundu, & Hsu, 2003; Hitt, Hoskisson, & Kim, 1997; Ruigrok & Wagner, 2003). Over the past three decades, theoretical explanations have proposed different balances between the costs and benefits of internationalization. The explanations have found empirical substantiation in a range of relationships, including inverted J-curves (Geringer, Beamish, & Da Costa, 1989), U-curves (Ruigrok & Wagner, 2003), inverted U-curves (Hitt et al., 1997), and most recently, S-curves (Contractor et al., 2003; Lu & Beamish, 2004). As a result of the divergent and sometimes conflicting outcomes of these contributions, no single theory or relationship has thus far been hailed as definitive. Several explanations have been brought forward to account for the different findings. In addition to methodological issues such as sample- and home-country bias, attention has been focused on the role of factors that moderate the relationship between internationalization and performance, such as the presence of intangible assets that can be exploited abroad (Lu & Beamish, 2004; Kotabe, Srinivasan, & Aulakh, 2002) or the (geographic) dispersion of international activities (Vachani, 1991; Goerzen & Beamish, 2003). The role of the MNE’s organizational structure, however, has remained underemphasized as moderating factor. Although the importance of organizational issues has been referred to in the context of the IP relationship (Lu & Beamish, 2004), and in international business literature more generally (Foss & Pedersen, 2004), their moderating role in the IP relationship has not yet been theorized specifically nor tested empirically. Yet the way in which a firm organizes and coordinates its overseas subsidiaries greatly influences the cost side of internationalization, and hence if (and at what point) each individual firm may encounter its internationalization ‘threshold’ (Sullivan, 1994; Ruigrok & Wagner, 2003). In addition, earlier contributions such as Bartlett and Goshal (1989) and Porter (1986), also described how the organizational configuration of a firm’s internationalization strategy can positively influence its performance and competitiveness. Especially, the extent of correspondence between a firm’s organizational structure and the external pressures it is exposed to (towards on the one hand, global integration, and on the other, local responsiveness) is argued to be critical for performance (Prahalad & Doz, 1987). Thus far, however, much of these analyses of organizational

Internationalization and Performance

179

coordination strategies in MNEs has been perceptual and case-based, and the link with performance remains unclear (cf. Harzing, 2000). This paper aims to contribute to the debate on internationalization and performance by analyzing the role of ‘strategic fit’ as a moderating factor, where strategic fit is defined as the consonance of a firm’s organizational structure with the integration and responsiveness pressures it is exposed to at the industry level. To do so, this paper proceeds as follows. First, the following section of the paper explores in more detail the costs and benefits of internationalization put forward in the existing literature. Consequently, the role of organizational structure in internationalization strategies is elaborated on, and hypotheses are developed on how the strategic fit between organizational structure and sectoral context can contribute to profitable internationalization. In the section on methodology and data, we introduce our dataset and explain in detail how we measure the indicators necessary to test our hypotheses. This paper not only contributes to the debate on the IP relationship, but also represents a new suggestion on how to operationalize the organizational response to industry-level pressures for integration and local responsiveness. More specifically, we first use industry-level data on R&D and advertising intensity to develop a measure of relative pressures towards integration and local responsiveness (as suggested by e.g. Prahalad & Doz, 1987 and Kobrin, 1991). Consequently, we construct a 3-item scale of organizational structure (ranging from ‘integrated’ at one extreme to ‘multidomestic’ on the other, where the individual items are measures of the ownership linkages between all of the firm’s individual subsidiaries (our sample of 336 firms together controlled over 60,000 majority-owned subsidiaries in 2002). Finally, comparing the indicator of organizational structure to the relative industry-level pressures produces the indicator of ‘strategic fit’, which is introduced as the moderating factor in the IP relationship using OLS regression. The results of the regression analyses are presented in the fourth section, while the final section of the paper discusses the consequences of the findings, addresses some of the shortcomings of the study and provides suggestions for future research.

LITERATURE REVIEW Costs and Benefits of Internationalization The relationship between firm internationalization and performance has been addressed in a large number of studies (for overviews, see e.g. Hitt et al.,

180


1997; Ruigrok & Wagner, 2003). Early theorists such as Hamal and Prahalad (1985) focused primarily on the benefits of internationalization. From the late 1980s onwards, more emphasis was placed on the costs that accompany international expansion, and by the 1990s a debate ensued on ‘the shape of the curve’, where divergent curvilinear relationships were identified, based on the different emphases authors placed on a range of theoretical explanations for the ‘sum’ of costs and benefits at various stages of internationalization. On the benefit side, the company-internal exploitation of market imperfections in (international) product and factor markets is generally considered the key motive for internationalization (Rugman, 1981). MNEs are also assumed to have intangible assets (Barney, 1991; Hitt et al., 1997), ownership advantages (Dunning, 1993) or firm-specific advantages (Rugman & Verbeke, 1992), that help in achieving international economies of scale and scope (Caves, 1996). In addition, internationalization enables risk spreading (Kwok & Reeb, 2000), expands market opportunities (Bu¨hner, 1987), and, via operational flexibility, increases leverage with respect to competitors and governments (Kogut, 1989). Finally, organizational learning theory emphasizes how firms can not only exploit existing capabilities across borders, but also through internationalization, gain new firm-specific advantages (Ruigrok & Wagner, 2003; Noria & Ghoshal, 1997; Cantwell & Narula, 2001). The costs associated with internationalization stem from transaction costs (Jones & Hill, 1988; Teece, 1986; Williamson, 1967) and agency costs (Roth & O’Donnell, 1996). Also, foreign entrants are faced with the ‘liability’ of a relative lack of local market knowledge (Zaheer, 1995). The notion that eventually the costs of international coordination would outweigh the benefits derived from the international exploitation of proprietary assets found empirical backing in an inverted-J relationship between internationalization and performance, symbolizing an ‘internationalization threshold’ (Geringer et al., 1989; Gomes & Ramaswamy, 1999). In its more extreme form, this argument led to an inverted U-curve (Hitt et al., 1997). The argument builds on the concept of incremental internationalization (Johanson & Vahlne, 1977). It asserts that firms start internationalization in familiar markets that are geographically and culturally close, and are hence able to largely off-set the costs of internationalization by using similar management systems and marketing techniques as in the home market. It is only after encompassing everbroader geographic markets that the costs of internationalization escalate and profits are reduced.


181

This approach has been criticized though for its deterministic perspective on firm behavior and for disregarding proactive management (Sullivan, 1994). In contrast, organizational learning perspectives have generated assumptions and evidence of a standard U-form (Ruigrok & Wagner, 2003), where the first phases of internationalization are accompanied by the costs associated with the liability of foreignness, initial learning costs and insufficient economies of scale (as summarized by Contractor et al., 2003). Only after those difficulties have been overcome by adjusting organizational settings can internationalization become profitable (Ruigrok & Wagner, 2003). The S-curve hypothesis, first introduced by Sullivan (1994), has received significant recent attention (Contractor et al., 2003; Lu & Beamish, 2004) as an attempt to integrate these two approaches by combining the negative performance effects of the ‘liability of foreignness’ in the early stages of internationalization with learning effects, economies of scale and scope and transaction cost internalization in the second stage (positive performance effects) and finally the internationalization threshold based on the prohibitive coordination costs of ‘overstretch’ in the final stage. This diversity of empirical results has induced researchers to look for explanations for the divergence. On the methodological side, Contractor et al. (2003) suggest that some of the confusion could be related to the ‘phase’ of internationalization in which the majority of firms in a particular sample find themselves. Similarly, another explanation is that studies often take single-country (or single-sector) samples, while it has been shown that both the degree of internationalization as well as performance differ importantly across countries of origin (Harveston, Kedia, & Francis, 1999; Wan & Hoskisson, 2003). More recently, the discussion has expanded to include the role of moderating variables that can account for the contradictory findings. For example, a firm’s ability to exploit intangible assets abroad has been found to positively influence the extent to which international activity is profitable (Lu & Beamish, 2004; Kotabe et al., 2002). Furthermore, a considerable number of studies added product diversification to international diversification in examining multinational performance (Hitt et al., 1997; Doukas & Lang, 2003). Finally, Vermeulen and Barkema (2002) highlight that the speed at which subsidiaries are established and the irregularity of the internationalization process negatively moderate how much a firm benefits from its international operations. Another moderating factor in the internationalization-performance relationship that has however received less attention is the role of organizational structure. Several authors have already mentioned that the

182


way in which a firm organizes its international activity importantly influences the extent to which internationalization is profitable (Lu & Beamish, 2004). However, the role of organizational structure as moderating in the IP relationship has not yet received extensive theoretical or empirical attention. This section elaborates on this idea and develops concrete hypotheses that will be tested in the empirical part of the paper.

The Role of Organizational Structure The role of organizational structure per se in enhancing performance (i.e., outside the context of the IP relationship) has traditionally been an important issue in the management literature. Especially the role of strategic fit, or contingency between a firm’s external environment and its organizational structure for firm performance has been stressed (Chandler, 1962; Lawrence & Lorsch, 1967; Rumelt, 1980; Venkatraman & Camillus, 1984; Porter, 1986). Much of the early literature emphasized the fit between a firm’s strategy and internal organization. But the rapid international expansion of many firms precipitated a shift toward analyzing firms’ organization with respect to their external and international environments (Prahalad & Doz, 1987; Bartlett & Ghoshal, 1989; Porter, 1986; Morrison, 1990; Ohmae, 1990). The emphasis on external environment as a driver of firm organization led to a range of ‘ideal-type’ organizational configurations by which a firm could organize internal processes and command structures given the characteristics of its industry and its pattern of internationalization. In particular, two sets of contrasting external international pressures are deemed to be critical in designing the optimal organizational structure of international firms. On the one hand, firms face pressure to integrate activities globally in order to optimally exploit economies of scale and survive under conditions of ‘global’ competition (Kobrin, 1991). These pressures toward global integration induce a company to manage its foreign subsidiaries rather tightly from headquarters, defined variously as the necessity for ‘control’, ‘coordination’, or ‘centralization’ (Cray, 1984; Porter, 1986; Doz & Prahalad, 1984; Bartlett & Ghoshal, 1989; Ruigrok & van Tulder, 1995; Roth, Schweiger, & Morrison, 1991; Egelhoff, 1988). On the other hand, consumer needs can differ importantly across host countries, as can host government policies, and markets and business environments more generally. Such heterogeneity pressures firms to


183

differentiate their strategy across countries and to be responsive to local conditions. As a result, firms may be best served by decentralization, thus allowing their subsidiaries considerable autonomy, capitalizing on their ability to navigate in the host country environment (Birkinshaw & Morrison, 1995). The pressures driving these organizational issues were brought together in the integration–responsiveness (IR) grid developed by Prahalad and Doz (1987). The IR grid is intended to assess ‘the relative importance of the two sets of conflicting demands’ (Prahalad & Doz, 1987, p. 22; italics original; see also Tallman & Yip, 2001). It is argued that firms predominantly experiencing pressures towards global integration of activities should be organized with a centralized (‘hub and spoke’) structure, whereas firms mainly subject to local responsiveness pressures should be organized as decentralized federation. Firms experiencing both pressures should organize as ‘multifocal’ or ‘transnational’ firms, characterized by a balance between centralized decision making and local subsidiary autonomy.

Strategic Fit as a Moderator of the IP Relationship Recent work exploring the integration–responsiveness grid and related typologies has largely been aimed at objectively measuring the incidence of global, multidomestic and multifocal strategies among real firms (Harzing, 2000; Leong & Tan, 1993). Other studies have investigated whether the fit of such strategies with their industry pressures results in the theoretically expected performance surpluses, with mixed results (see Grein, Craig, & Takada (2001) as well as Roth et al. (1991) who found positive evidence, while Roth & Morrison (1992) did not). We propose to include strategic fit as a moderator in the IP relationship. We expect that the success of an MNE’s international strategy will to a large extent depend on how well the organizational challenges of internationalization are resolved, given the pressures to which the MNE is exposed. On the one hand, the fit between organizational structure with a firm’s environment affects the coordination and transaction costs of internationalization (Jones & Hill, 1988), and thus whether a firm can limit its ‘liability of foreignness’ (Zaheer, 1995) and increase its internationalization ‘threshold’ (cf. Sullivan, 1994). On the other hand, the way in which a firm organizes its overseas subsidiaries, and the extent to which these are integrated and coordinated, also gives MNEs the possibility to capitalize on their firm-specific advantages in a strategy of ‘cross-border arbitrage’, and

184


achieve sustainable competitive advantage (Ghemawat, 2003; Porter, 1991). Thus, strategic fit could be expected to both reduce the costs and increase the benefits of internationalization. This beneficial effect of strategic fit should be more critical at higher levels of internationalization. For less international firms, the overall cost of ‘misfit’ should be smaller; and for these firms it might even be more profitable to maintain a sub-optimal organization rather than change it. Therefore, we hypothesize that not only will fit have a direct positive effect on performance, but also an indirect effect by positively moderating the IP relationship. Hypothesis 1. Strategic fit positively affects performance. Hypothesis 2. Strategic fit positively moderates the internationalization– performance relationship. While the aim of this paper and testing these hypotheses is to gain more insight into the complex nature of the IP relationship, an interaction effect between the degree of internationalization (DOI) and strategic fit could also help explaining the diverging outcomes of the studies on the relationship between fit and performance cited above, that have tended to take DOI is taken as given, without considering how variations in DOI across strategy types or sectors may affect the outcome. From the perspective of both debates – on internationalization and performance, and on fit and performance – gaining more insight into the link between integration and responsiveness pressures on the one hand and internationalization and performance on the other should be worthwhile, given that these have not yet been firmly established (Harzing, 2000).

DATA AND METHODOLOGY Sample To test the hypotheses developed above, we collected data for a sample of Fortune Global 500 firms, for the year 2002. The sample included all nonfinancial triad-based firms (n=336). Financial firms were eliminated from the sample on the grounds that their internationalization and activities more generally do not refer to ‘production’ and thus are subject to a different dynamic.


185

Data and Variable Definition Measuring strategic fit requires information on both the relative external pressures toward integration and responsiveness (IR pressures), and on the organizational structure of the firm. In addition, both variables need to be measured on the same (kind of) scale in order to calculate the difference between them as a proxy of correspondence between organizational structure and external pressures. Integration and Responsiveness Pressures We first calculate our measure of relative IR pressures. We follow the suggestions by Prahalad and Doz (1987, p. 36), who argue that ‘functions such as R&D, manufacturing, marketing and service may be used to identify pressures for global integration and local responsiveness’, and Kobrin (1991) who finds that in particular R&D intensity and advertising intensity are appropriate to measure global integration and responsiveness, respectively. Hence, we use industry level R&D and advertising intensity as measures for integration pressures and responsiveness pressures. Advertising intensity was measured as the ratio of purchased advertising services to total industry output, calculated from the US Census input– output tables (latest year available, 1997). R&D intensity was measured as the ratio of R&D expenditure to total sales of US multinational firms (parents and foreign affiliates combined), calculated from the US Bureau of Economic Analysis’ (BEA’s) financial and operating statistics of US MNEs (latest year available, 2002). All data were collected at the industry level as defined by the US North American Industry Classification System (NAICS). Complete data were available for 67 NAICS industries (at the 3- or 4-digit level, as detailed as possible) that together covered the whole range of primary, secondary and tertiary sectors. After classing our firms in these NAICS sectors (based on their Fortune sector descriptions), only 37 of these 67 sectors contained (one or more) firms. The sectors that included the largest number of firms are positioned in Fig. 1 on axes representing standardized advertising and R&D intensity. The overall majority of sectors is clearly either dominated by responsiveness or integration pressures. In addition, the placement of the industries is very similar to the suggestions that other authors have occasionally given as examples of industries dominated by either integration or responsiveness pressures (e.g., food and beverages, or retail, as responsiveness; automotive, or electronics, as integration) (Kobrin, 1991, p. 21; Prahalad & Doz, 1987, p. 37), which further validates using

186

FABIENNE FORTANIER ET AL. Pharmaceuticals [0.7; 2.57] Integration pressures dominate Other transport

R&D Intensity (Z-scores)

.5

Computers & related

Automotive Chemicals Machinery

0.0

Elec.Equipment

Soaps & toilet prep

Plastic & rubber

-.5

Nonmet.minerals Primary metals Petroleum Construction

-1.0 -1.0

Fig. 1.

Responsiveness pressures dominate

-.5

Food Beverages & Tob.

Wholesale Telecom

Utilities Transportation

0.0 .5 Advertising intensity (Z-scores)

Retail [2.5; -0.6]

1.0

Industry Pressures Toward Integration (R&D Intensity) and Responsiveness (Advertising Intensity).

R&D and advertising intensity as proxies for integration–responsiveness pressures. While the IR grid essentially depicts the pressures toward integration and responsiveness as orthogonal, they are more often seen as ‘relative’ pressures (toward either more integration or more responsiveness) (Prahalad & Doz, 1987; Tallman & Yip, 2001), and as inversely related (see also the results of Kobrin (1991), who finds that advertising is negatively related to integration, while R&D is positively related to integration). From this perspective, it is possible to combine both pressures to a single scale, with integration and responsiveness on both ends, and a balance between them in the middle. Such a scale can be constructed at the sector level using the standardized data used in Fig. 1. By subtracting the standardized R&D scores from the standardized advertising scores, a scale is created on which negative values represent sectors where integration pressures dominate, and positive values sectors where pressures towards responsiveness are relatively more important. Values around zero then represent a balance between pressures. In addition, higher numbers (both


187

negative and positive) are an indication of higher importance of the particular pressure. Organizational Structure After thus developing the indicator of relative pressures towards integration and responsiveness, the second dimension necessary to measure strategic fit can be constructed: organizational structure. Firms that can be characterized as being either locally responsive (or ‘multidomestic’) or globally integrated (or ‘global’) differ across many dimensions (Harzing, 2000). Still, one of the key discriminating characteristic that qualifies firms as being locally responsive or globally integrated, is their organizational structure, and more specifically, whether the locus of decision making in the MNEs is centralized (at headquarters), or rather more decentralized at the level of foreign subsidiaries (Pugh, Hickson, Jinnings, & Turner, 1968; Cray, 1984; Martinez & Jarillo, 1989). Measuring a decentralized (‘locally responsive’) versus centralized (‘integrated’) organizational structure has traditionally been done through questionnaires directed at HQ or subsidiary managers, or both, asking them about the level of decision-making authority the subsidiary has in a range of areas (Harzing, 2000; Bartlett & Ghoshal, 1989). The advantage of such measures is that they represent a relatively precise and detailed way to operationalize and measure the concept of centralization. However, measuring centralization through questionnaires has also some important disadvantages: the indicators are perception based, generally collected for only a few subsidiaries per MNE and (partly due to response rates) for a relatively small number of MNEs, and finally, such studies generate results that are difficult to replicate or compare over time or across studies. In this paper, we propose an alternative measure of the centralization of organizational structure, based on archival data on the organizational ‘trees’ of MNEs that consist of the chain of ownership in which a subsidiary can itself be a parent to one or more other subsidiaries of the same firm. Using such data, we can calculate our measures taking into account the structure of the entire MNE (which can include information on more than 1,000 affiliates per firm, rather than just a few). While using less fine-grained information than can be obtained through questionnaires, the measures we propose are not dependent on the views of one particular manager and hence arguably more objective. And in addition, the measures are much easier to replicate, and can be collected and compared for a larger set of firms as well as over time. The advantages and disadvantages of our

188


indicator hereby mirror the ones of questionnaire-based measures, and can hence be an important addition to existing research in this area. The data that we use for our indicator are drawn from Dun & Bradstreet’s (D&B) Who owns Whom (2002) database. From this database, information was obtained on the number and location of the subsidiaries of each firm, as well as on the hierarchical ‘tree’ of ownership relationships between the subsidiaries. Together, our sample of 336 firms had an aggregate majoritycontrolled subsidiary base consisting of over 60,000 affiliates. Using the information of the subsidiaries and the linkages between them, we can construct measures of organizational structure (and later on, also internationalization) that allow for a ‘systemic’ perspective (cf. Gupta & Govindarajan, 2000) of the subsidiary network at the MNE level. While the ‘corporate tree’ of legal (ownership) relationships between subsidiaries does not necessarily represents also the ‘real’ organizational structure, it can serve as a good proxy for it. For example, Prahalad and Doz (1987) emphasize that the essence of global competition centers on the management of strategic coordination, ‘even when global coordination across subsidiaries in terms of product flows does not take place’ (p. 40). Moreover, the legal relationship constitutes a form of interdependency and a transmission channel (Gupta & Govindarajan, 2000) along which intangible assets can be transferred. Finally, the potentially ‘arduous relationship’ (cf. Szulanski, 1996) between two units that can complicate knowledge transfer may be reduced if the linkages are institutionalized through ownership relationships. Taking these considerations into account, distinctive patters of linkages between subsidiaries and parents can be expected to emerge across MNEs (Harzing, 2000), even if subsidiary roles within MNEs can be quite diverse (Jarillo & Martinez, 1990; Birkinshaw & Hood, 1998). We used the D&B information on the design of the international hierarchy of ownership ties to construct three measures of the degree of decentralization, that represent the degree of autonomy which subsidiaries in host-country contexts are able to exercise, also described as ‘independence’ (Harzing, 2000), or ‘subsidiary choice’ (Birkinshaw & Hood, 1998). 1. AVLEV. The average of the hierarchical level on which each subsidiary within a firm finds itself; where headquarters are at level 0, the subsidiaries directly owned by headquarters are at level 1, their subsidiaries in turn at level 2 and so on. This indicator is included since in the organizational design literature, a larger number of organizational levels is traditionally seen as an indicator of a more decentralized


189

organization (Schollhammer, 1971). In addition, since decentralization can be defined as the degree to which authority is delegated to lower organizational levels, a higher average number of organizational levels should indicate that decisions are on an average taken lower in the organization. 2. NONHOME. The number of foreign subsidiaries whose ‘immediate parent’ is not located in the MNEs home country (as percentage of total subsidiaries). This indicator is included as a proxy for decentralized decision-making. When immediate parents are foreign subsidiaries, and thus coordinate the activities of at least one other (foreign) subsidiary, organizational ‘sub-trees’ are created in host countries, which can be seen as a sign of embeddedness in the host economy. 3. INTER. The number of foreign subsidiaries that are owned by an immediate parent in a country not their own (and not the MNEs home), as percentage of total subsidiaries. It is based on the combination of ideas that decentralization implies that decisions are taken at lower levels in the organization, and that internationalization (and the coordination of internationalization) is one important area in which such decisions can be made. Therefore, if foreign subsidiaries instead of headquarters coordinate international activities, this can be considered as a sign of decentralization. The three indicators are all highly correlated, as shown in Table 1. We aimed to combine the information on these three organizational dimensions of (de)centralization into one indicator that can be compared with our measure of industry pressures. To do so, we conduced a factor analysis (principal components) to both explore whether the data indeed do measure only one dimension of organizational structure, and to generate factor scores. The results of the factor analysis indicated that all indicators indeed Table 1.

Factor Analysis: Pearson Correlations and Factor Scores for Factor 1. Correlation Coefficients 1

1 2 3

AVLEV NONHOME INTER

po0.01.

0.536 0.359

Factor Analysis

2

Loadings

0.700

0.735 0.912 0.839

190


loaded (see also Table 1) on a single factor (eigenvalue 2.08), which explained 69.2% of the variance caused by the three indicators. The scale’s Cronbach’s alpha is 0.77, again an indication that the three indicators measure a single dimension. The factor scores of organizational structure (or decentralization) form a standardized variable (mean of 0, standard deviation of 1) just like the variable of relative pressures. Therefore, we can calculate our measure of fit (FIT) as the absolute difference between the firm’s organizational structure score and its value of relative sector pressures. This difference was multiplied by 1, to ensure that high scores thus represent a good fit, and that we can expect the coefficient for the FIT variable to be positive in relation to performance. Other Variables in the Model In addition to FIT, the other key variable in our analysis is the degree of a firm’s internationalization. This can be measured in terms of its operation, ownership or orientation (Annavarula & Beldona, 2000). Since the internationalization–performance debate focuses primarily on the profitability of international production, the measures developed in this paper emphasize the operational dimension of internationalization. Instead of using the popular sales- or asset-based indicators (cf. Ruigrok & Wagner, 2003), we follow the increasing number of studies that develop measures of international production based on subsidiaries (Lu & Beamish, 2004; Goerzen & Beamish, 2003; Ietto-Gilles, 1998). The degree of internationalization (DOI) is measured as the ratio of foreign subsidiaries to total subsidiaries (also calculated from D&B data). These independent variables (FIT and DOI) are related to the dependent variable, corporate performance, which is measured as Return on Sales (ROS) (operating income/sales) (see Ruigrok & Wagner (2003) for an overview of similar studies). Additionally, the relationship between the independent variables and performance is controlled for several other variables. The first is firm size, which as measure of scale economies is virtually always included in estimations of the IP relationship. Firm size is calculated as the logarithmic of total sales (LogSales) and its coefficient is expected to be positive. In addition, a set of dummy variables for the various home countries (regions) of the MNEs in the sample is included to capture country (region) of origin effects (see Hitt, Hoskisson, & Ireland, 1994 and Contractor et al., 2003). Dummy variables are created for the USA, Europe (with separate dummies for French, German, British and ‘Other European’ firms) and Asia (predominantly Japan).


191

Estimation We analyzed the effects of the organizational fit on the relationship between internationalization and performance using OLS regression. Both squared and cubed values of DOI are included to test for the nature of the non-linear relationship between internationalization and performance. The entire model can be specified as follows: ROS ¼ b0 þ b1 LOGSALE þ b2 dFRA þ b3 dGER þ b4 dUK þ b5 dRESTEUR þb6 dASIA þ b7 FIT þ b8 DOI þ b9 DOI 2 þ b10 DOI 3 þ

We estimated this model for the entire sample as well as for three equally sized sub-samples, that were defined on the basis of the variable FIT (i.e., subsamples with firms whose organizational structure either shows a ‘good’, ‘moderate’, or ‘poor’ fit with industry pressures), to see if the costs and benefits of internationalization for performance differ by the level of strategic fit.

RESULTS Table 2 reports the descriptive statistics of the variables included in the analysis and their correlations. The table shows that at a bivariate level, only DOI is significantly positively correlated with ROS; the correlation coefficients between ROS and LOGSALES and ROS and FIT are not significant (although with the expected sign). Our measure of FIT is negatively related to firm size, indicating that the organizational structure of larger firms show a relatively poor fit with industry pressures, which may be a sign of organizational ‘overstretch’. There is, as expected, a positive relationship between FIT and DOI, which indicates that firms that are more Table 2.

1 2 3 4

Descriptive Statistics and Pearson Correlations (n=336).

ROS LOGSALES DOI FIT

po0.10. po0.05. po0.01.

Mean

S.D.

1

2

3

5.96 4.21 0.44 1.57

8.21 0.45 0.27 1.20

1.00 0.08 0.10* 0.03

1.00 0.21*** 0.19***

1.00 0.11**

192


international, have an organizational structure that on average better matches the industry pressures they are exposed to. The relationship among DOI, FIT and ROS is further explored with regression analysis. With the first regressions, a few outliers were observed that scored either exceptionally high (W30%) or very low (o25%) on the ROS variable (which also explains the relatively high standard deviation of ROS in Table 2). These were removed, leaving us with a sample of 332 firms. The results of the OLS regressions are displayed in Table 3. The first four models in Table 3 show the results of the estimations for the entire sample of 332 firms. All the regional dummy variables have significantly negative coefficients, with the exception of the UK, indicating that firms from all these countries and regions (except UK) perform on average worse than American firms (the reference category). The results for the variable FIT support our first hypothesis, namely that a good fit between structure and strategy increases performance. Interestingly, firm size (logsales) is not significantly related to performance. This may be due to our selection of firms that included the largest firms worldwide by revenue, implying that any additional differences in size among them do not matter much. Finally, the results indicate the presence of a U-shaped curvilinear relationship between internationalization and performance, and do not support the S-hypothesis. However, these findings change importantly if the sample is split up into three equally-sized subgroups of firms, according to their score on the FIT ’ variable. As models 5–16 in Table 3 show, the significant results for FIT in explaining performance have disappeared (as to be expected when splitting the sample with this variable); only in the group of firms that exhibit the worst alignment of their organizational structure with IR pressures, does an increase in FIT have significant performance implications. Again, the coefficient for size (LogSales) is insignificant in all model specifications and across all groups. As far as the region/country dummy variables are concerned, German firms and to a lesser extent Asian firms show lower ROS values than their American counterparts. Our key interest concerns the shape and direction of the relationship between internationalization and performance at various levels of fit. For the group that displayed the best fit between organizational structure and IR pressures, we find clear evidence for a positive and linear relationship (models 6–8). This indicates that a good alignment between organizational structure and sector pressures toward integration or responsiveness, offsets both the costs of internationalization associated with the liability of foreignness at lower levels of internationalization, and the costs of ‘overstretch’ (the internationalization threshold) at high levels of internationalization.

Table 3. Regression Results, for Entire Sample and 3 Sub-Samples (Split by Fit). Entire Sample 1 Constant

5.45 1.24 D_ASIA 5.50*** 5.05 D_FRA 4.25*** 2.93 D_GER 9.13*** 6.00 D_UK 1.49 1.00 D_RESTEU 3.52*** 2.63 FIT 0.88*** 2.64 LogSALES 1.33 1.34 DOI DOI 2 DOI 3

R2 Adjusted R2 F n

0.138 0.120 7.426*** 332

po0.10. po0.05. po0.01.

2

3

Good Fit 4

6.15 6.99 8.49* 1.40 1.60 1.91 5.34*** 5.02*** 4.85*** 4.90 4.64 4.48 4.80*** 4.30*** 4.47*** 3.25 2.93 3.05 9.63*** 9.38*** 9.42*** 6.26 6.16 6.20 1.72 0.64 0.50 1.16 0.42 0.33 3.90*** 3.68*** 3.68*** 2.89 2.76 2.77 0.78** 1.04*** 1.02*** 2.30 3.01 2.97 0.84 1.22 1.09 0.83 1.21 1.07 2.93* 13.68** 30.30** 1.90 2.41 2.53 18.55*** 66.15** 3.03 2.14 35.19 1.57 0.148 0.127 7.001*** 332

0.171 0.148 7.403*** 332

0.178 0.152 6.941*** 332

5

6

2.68 0.36 6.26*** 3.39 4.14 1.60 8.41*** 3.65 3.97 1.58 3.31 1.43 1.07 0.42 1.56 0.95

2.09 0.29 5.36*** 2.86 4.08 1.60 9.04*** 3.95 4.40* 1.77 2.80 1.22 1.34 0.53 0.97 0.59 5.64** 1.98

0.260 0.210 5.172*** 111

0.287 0.232 5.143*** 111

7

Moderate Fit 8

1.52 5.65 0.21 0.71 5.24*** 4.97** 2.77 2.62 4.00 4.47* 1.56 1.73 9.03*** 9.18*** 3.93 4.00 4.69* 4.49* 1.85 1.77 2.69 3.45 1.16 1.45 1.31 1.22 0.52 0.49 1.37 0.98 0.77 0.54 1.11 30.23 0.10 1.20 7.30 88.10 0.62 1.39 60.33 1.29 0.290 0.227 4.587*** 111

0.302 0.232 4.323*** 111

9

10

6.08 0.69 2.80 1.32 4.44 1.64 9.70*** 3.23 5.43* 1.84 1.84 0.70 1.51 0.54 1.27 0.70

6.46 0.73 2.49 1.10 4.51 1.66 9.81*** 3.24 5.37* 1.81 1.91 0.72 1.42 0.50 0.98 0.51 1.36 0.43

0.117 0.056 1.928* 110

0.118 0.049 1.696 110

11

Poor Fit 12

12.31 20.19** 1.37 2.18 1.95 2.31 0.88 1.07 3.35 4.37 1.24 1.65 9.49*** 9.66*** 3.21 3.36 3.04 2.93 1.00 0.99 1.64 1.84 0.63 0.73 3.72 5.27* 1.28 1.82 0.88 0.22 0.46 0.12 24.53** 78.73*** 2.23 3.37 30.62** 187.22*** 2.46 3.05 116.41** 2.61 0.169 0.094 2.254** 110

0.222 0.143 2.825*** 110

13

14

14.24 1.66 7.39*** 4.19 3.37 1.46 8.50*** 3.06 2.60 1.13 4.88** 2.35 2.23*** 3.21 0.42 0.23

12.90 1.50 7.84*** 4.38 4.74* 1.88 9.03*** 3.23 3.23 1.38 5.90*** 2.68 1.83** 2.42 0.22 0.12 3.44 1.33

0.232 0.180 4.454*** 111

0.245 0.186 4.148*** 111

15

16

13.98 12.86 1.61 1.50 7.47*** 8.06*** 4.12 4.44 4.46* 4.21* 1.76 1.68 8.52*** 8.87*** 3.01 3.17 2.58 3.27 1.07 1.36 5.63** 6.19*** 2.54 2.81 1.96** 2.17*** 2.56 2.85 0.30 0.29 0.17 0.16 6.39 27.79 0.70 1.41 10.64 83.33* 1.12 1.69 66.87* 1.95 0.255 0.188 3.834*** 111

0.282 0.21 3.925*** 111

194

10

5

0

ROS

0

0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95

1

-5

-15 Total Moderate Fit

Good Fit Poor Fit

-20 DOI

Fig. 2.

Internationalization and Performance, by Strategic Fit.


-10


195

These costs clearly do play a role for firms with a moderate fit between organizational structure and industry pressures. Here, we find strong evidence for a horizontal S-shaped relationship between internationalization and performance (models 10–12). These firms do experience the learning costs at early stages of internationalization, and are better-off not crossing the internationalization threshold at an internationalization level of about 80 percent. Finally, for the one-third of firms that showed the worst degrees of strategic fit, we also find evidence of an ‘S-like’ curvilinear relationship, but here the shape should best be described as an ‘inverse’ horizontal S, where an increasingly negative relationship between internationalization and performance is present from early stages of internationalization onward, only countered at the highest levels of internationalization (models 14–16). Especially this last increase in performance (or better, reduced cost) is surprising, if only barely significant (at 10% level). A possible (and speculative) explanation for this result could be that those firms that are highly international and show a poor fit, are the ones that are trying to ‘redefine the rules of the game’ (Prahalad & Doz, 1987) in the industry, and purposely organize in a way that contradicts industry pressures. These firms may have seen chances where others have not (yet), and are therefore able to achieve some performance benefit (or at least reduced costs). To compare the different relationships between internationalization and performance, Fig. 2 shows the marginal impact of DOI on firm performance for both the entire sample and each of the three sub-sets of firms.

DISCUSSION AND CONCLUSIONS A wealth of previous research has explored the costs and benefits of internationalization for multinational enterprises. Although, in general, most studies point to some degree of a positive relationship, results have been largely conflicting with respect to the exact (curvilinear) shape of the curve. We hypothesized that strategic fit, measured as the congruence between firms’ organizational structures and extant integration or responsiveness pressures at the industry level, moderates the relationship between internationalization and performance. This could explain for some of the divergence in results in previous studies. To test these hypotheses, we suggested a new way of calculating measures of organizational structure (and consequently, strategic fit) based on archival data rather than questionnaires. We found considerable support for our hypotheses that strategic fit increases performance (H1), and that strategic fit positively

196


moderates the IP relationship (H2). More specifically, the results showed that the relationship between internationalization and performance is quite complex and differs in size, shape and direction across various levels of fit. Our results suggest that firms with a good fit between organizational characteristics and the predominant pressures in their industry (integration or local responsiveness) tend to profit more, and increasingly so, from internationalization. This supports the early strategic management literature that anticipated a linear, positive relationship between internationalization and performance. Since these were the same authors that developed the literature on competitive and industry pressures, they may have implicitly or explicitly focused on the expectation that good strategy meant a good understanding of the competitive environment (i.e., good fit), and therefore that internationalization would allow firms to benefit while minimizing some of the associated costs. On the other hand, the ‘moderate fit’ group may be representative of ‘average’ firms, and thus that sample characteristics may explain why other studies have found a horizontal S-curve for the relationship between internationalization and performance. It may be that firms with average fit levels are less equipped to manage their environment more generally, and as a result are forced to incur considerable costs to make internationalization successful in the early stages. As learning effects begin to develop, those costs are reduced until environmental complexity begins to outweigh the ability of the firm to manage that environment through learning. Finally, it may be that the group with the worst fit is the least attuned to its environment and that, as a result, increasing environmental complexity continues to raise the costs of doing business without the firm being able to exploit the potential benefits that a more complex environment has to offer. This paper aims to shed light on the complexity of the relationship between internationalization and performance. In it, we explicitly take into account (and operationalize) the configuration of the international activities of a sample of the world’s largest MNEs, and relate those indicators to the balance between integration and local responsiveness pressures at the sector level, and finally to performance. The empirical findings confirm our theoretical expectations that strategic fit is important if firms want to benefit from internationalization, and as such can serve as recommendation to managers to ensure that their organizations are (better) aligned with sectoral pressures. Still, the research results presented in this paper have their limitations, which implies that certain important questions for further research remain. One of these questions relates to the conceptualization of the variables of international structure, which hinges on the legal relationships between


197

subsidiary ‘members’ in the corporate tree. While we gave several arguments as to why this could be a good proxy for a firm’s ‘real’ organizational structure, and our empirical results are in line with theoretical expectations and therefore also do not immediately raise questions as to the validity of our measurement, it can still not necessarily be directly inferred that e.g. knowledge, products or coordination always follow such legal relationships. From this point of view, measuring concepts such as decentralization through surveys would yield more detailed and more precise data. Yet the advantage of the dataset used in this paper is that it leads to research that is based on a wider subsidiary base per firm than generally possible through questionnaires, and is more easily replicable. Given these advantages and disadvantages of both methods, it would be worth exploring in future research whether our measures of organizational structure correspond with those based on questionnaires. Such a comparison may also shed light on an additional issue that has not yet been fully explored: the option that managerial perceptions of IR pressures may diverge from more ‘objective’ pressures – and hence lead to strategies that are ill suited to the environment. This may be a reason why the (perceptionbased) studies reviewed above have not always been able to establish the theoretically expected positive relationship between organizational structure and environmental pressures. Finally, future research could extend the analysis of this paper over time, to examine in more detail the causal relations (rather than only correlations) between organizational structure, internationalization and performance. Similarly, follow-up studies could include more alternative measures of performance, such as return on assets or market value, to explore whether the conclusions from this study hold in other contexts.

REFERENCES Annavarula, M., & Beldona, S. (2000). Multinationality–performance relationship: A review and reconceptualization. International Journal of Organizational Analysis, 8, 48–67. Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17, 99–120. Bartlett, C., & Ghoshal, S. (1989). Managing across borders. Cambridge, MA: Harvard University Press. Birkinshaw, J., & Hood, N. (1998). Multinational subsidiary evolution: Capability and charter change in foreign-owned subsidiary companies. Academy of Management Review, 23, 773–795. Birkinshaw, J., & Morrison, A. (1995). Configurations of strategy and structure in subsidiaries of multinational corporations. Journal of International Business Studies, 26, 729–753.

198


Bu¨hner, C. (1987). Assessing international diversification of West German corporations. Strategic Management Journal, 8, 25–37. Cantwell, J., & Narula, R. (2001). The eclectic paradigm in the global economy. International Journal of the Economics of Business, 8, 155–172. Caves, R. (1996). Multinational enterprise and economic analysis. Cambridge, UK: Cambridge University Press. Chandler, A. (1962). Strategy and structure. Boston, MA: MIT. Contractor, F., Kundu, S., & Hsu, C. C. (2003). A three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34, 5–18. Cray, D. (1984). Control and coordination in multinational corporations. Journal of International Business Studies, 25, 85–98. Doukas, J., & Lang, L. (2003). Foreign direct investment, diversification and firm performance. Journal of International Business Studies, 34, 153–172. Doz, Y., & Prahalad, C. (1984). Patterns of strategic control within multinational corporations. Journal of International Business Studies, 15(2), 55–72. Dunning, J. (1993). Multinational enterprises and the global economy. New York: Addison-Wesley. Egelhoff, W. G. (1988). Organizing the multinational enterprise: An information processing perspective. Cambridge, MA: Ballinger. Foss, N. J., & Pedersen, T. (2004). Organizing knowledge processes in the multinational corporation: An introduction. Journal of International Business Studies, 35, 443–455. Geringer, J. M., Beamish, P. W., & Da Costa, R. C. (1989). Diversification strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10, 109–119. Ghemawat, P. (2003). Semi-globalization and international business strategy. Journal of International Business Studies, 34(2), 138–152. Goerzen, A., & Beamish, P. (2003). Geographic scope and multinational enterprise performance. Strategic Management Journal, 24, 1289–1306. Gomes, L. K., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between multinationality and performance. Journal of International Business Studies, 30, 173–188. Grein, A., Craig, S., & Takada, H. (2001). Integration and responsiveness: Marketing strategies of Japanese and European automobile manufacturers. Journal of International Marketing, 9, 19–30. Gupta, A., & Govindarajan, V. (2000). Knowledge flows within corporations. Strategic Management Journal, 21, 473–496. Hamal, G., & Prahalad, C. K. (1985). Do you really have a global strategy? Harvard Business Review, 63(4), 139–148. Harveston, P., Kedia, B., & Francis, J. (1999). MNE’s dependence on foreign operations and performance: A study of MNEs from the ‘‘triad’’ regions. International Business Review, 8, 293–307. Harzing, A. W. (2000). An empirical test and extension of the Bartlett and Ghoshal typology of multinational companies. Journal of International Business Studies, 31, 101–120. Hitt, M., Hoskisson, R., & Ireland, R. (1994). A mid-range theory of the interactive effects of international and product diversification on innovation and performance. Journal of Management, 20, 297–327.


199

Hitt, M., Hoskisson, R., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management Journal, 40, 767–798. Ietto-Gilles, G. (1998). Different conceptual frameworks in the assessment of the degree of internationalization: Empirical analysis of various indices for the top 100 TNCs. Transnational Corporations, 7(1), 17–39. Jarillo, J. C., & Martinez, J. I. (1990). Different roles of subsidiaries: The case of multinational corporations in Spain. Strategic Management Journal, 11, 501–512. Johanson, J., & Vahlne, J. E. (1977). The internationalization process of the firm. Journal of International Business Studies, 8, 23–32. Jones, G., & Hill, C. (1988). Transaction cost analysis of strategy-structure choice. Strategic Management Journal, 9, 159–172. Kobrin, S. (1991). An empirical analysis of the determinants of global integration. Strategic Management Journal, 12, 17–31. Kogut, B. (1989). A note on global strategies. Strategic Management Journal, 10, 383–389. Kotabe, M., Srinivasan, S., & Aulakh, P. (2002). Multinationality and firm performance: The moderating role of R&D and marketing capabilities. Journal of International Business Studies, 33, 79–97. Kwok, C., & Reeb, D. (2000). Internationalization and firm risk: An upstream–downstream hypothesis. Journal of International Business Studies, 31, 611–629. Lawrence, P., & Lorsch, J. (1967). Organization and environment. Boston, MA: Harvard Business School. Leong, S., & Tan, C. (1993). Managing across borders: An empirical test of the Bartlett and Ghoshal organizational typology. Journal of International Business Studies, 24(3), 449–464. Lu, J., & Beamish, P. (2004). International diversification and firm performance: The S-curve hypothesis. Academy of Management Journal, 47, 598–609. Martinez, J., & Jarillo, C. (1989). The evolution of research on coordination mechanisms in multinational corporations. Journal of International Business Studies, 20, 489–514. Morrison, A. J. (1990). Strategies in global industries. New York: Quorum Books. Noria, N., & Ghoshal, S. (1997). The differentiated network: Organizing multinational corporations for value creation. San Francisco, CA: Jossey-Bass Publishers. Ohmae, K. (1990). The borderless world. London: Fontana. Porter, M. (1986). Competition in global industries: A conceptual framework. In: M. Porter (Ed.), Competition in global industries (pp. 15–60). Boston, MA: Harvard University Press. Porter, M. (1991). Towards a new theory of strategy. Strategic Management Journal, 12, 95–117. Prahalad, C., & Doz, Y. (1987). The multinational mission: Balancing local demands and global vision. New York: Free Press. Pugh, D., Hickson, D., Jinnings, C., & Turner, C. (1968). Dimensions of organizational structure. Administrative Science Quarterly, 13, 65–105. Roth, K., & Morrision, A. (1992). Business-level competitive strategy: A contingency link to internationalization. Journal of Management, 18, 473–487. Roth, K., & O’Donnell, S. (1996). Foreign subsidiary compensation strategy: An agency theory perspectives. Academy of Management Journal, 93, 678–703.

200


Roth, K., Schweiger, D., & Morrison, A. (1991). Global strategy implementation at the business unit level. Journal of International Business Studies, 22, 369–403. Rugman, A. (1981). Inside the multinationals: The economics of internal markets. New York: Columbia University Press. Rugman, A., & Verbeke, A. (1992). A note on the transnational solution and the transaction cost theory of multinational strategic management. Journal of International Business Studies, 23, 761–771. Ruigrok, W., & Van Tulder, R. (1995). The logic of international restructuring. London: Routledge. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43, 63–83. Rumelt, R. P. (1980). Evaluating business strategy. In: H. Mintzberg, J. B. Quinn & S. Ghoshal (Eds), The strategy process (Rev. European ed.) (pp. 91–100). London: Prentice-Hall. Schollhammer, H. (1971). Organization structures of multinational corporations. Academy of Management Journal, 14, 345–365. Sullivan, D. (1994). Measuring the degree of internationalization of a firm. Journal of International Business Studies, 24, 325–342. Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practices within the firm. Strategic Management Journal, 17(Winter Special Issue), 27–43. Tallman, S. B., & Yip, G. S. (2001). Strategy and the multinational enterprise. In: A. M. Rugman & T. L. Brewer (Eds), Oxford handbook of international business (pp. 317–348). Oxford: Oxford University Press. Teece, D. (1986). Transactions cost economics and the multinational enterprise. Journal of Economic Behavior and Organization, 7, 21–45. Vachani, S. (1991). Distinguishing between related and unrelated international geographic diversification: A comprehensive measure of global diversification. Journal of International Business Studies, 22, 307–322. Venkatraman, N., & Camillus, J. (1984). Exploring the concept of ‘fit’ in strategic management. Academy of Management Review, 9, 513–525. Vermeulen, F., & Barkema, H. (2002). Pace, rhythm, and scope: Process dependence in building a profitable multinational corporation. Strategic Management Journal, 23, 637–654. Wan, W., & Hoskisson, R. (2003). Home country environments, corporate diversification strategies, and firm performance. Academy of Management Journal, 46, 27–45. Who owns whom. (2002). [Data file] Manchester, UK: Dun & Bradstreet. Williamson, O. (1967). Hierarchical control and optimum firm size. Journal of Political Economy, 75, 123–138. Zaheer, S. (1995). Overcoming the liability of foreignness. Academy of Management Journal, 38, 341–363.

INTERNATIONALIZATION OF INDIAN FIRMS: REGIONALIZATION PATTERNS AND IMPACT ON PERFORMANCE Vikas Kumar and Ajai S. Gaur ABSTRACT We investigate the internationalization pattern and performance of Indian firms. We first discuss the regionalization trend evident in the internationalization of Indian manufacturing and service firms over time. Next, we empirically test the impact of degree of internationalization on firm financial performance of Indian firms. We also test the moderation effect of business group affiliation on the internationalization–performance relationship. We find that Indian outward foreign direct investment has been shifting from developing to developed economies over time. Also, firm performance of Indian firms is positively related to the degree of internationalization and that service firms profit more than manufacturing firms from internationalization. Business group affiliation reduces the positive effect of internationalization on firm performance.


201

202

VIKAS KUMAR AND AJAI S. GAUR

INTRODUCTION Understanding the new developments of globalization is one of the big research questions in the field of international business (Buckley, 2002). Rugman and Verbeke (2004), in their study of the largest so-called global firms, demonstrate that majority of the firms have only a regional presence. According to them, most of the Fortune Global 500 multinational enterprises (MNEs) have majority of their sales in only one, usually their home leg, of the ‘triad’ markets, namely in North America, the European Union or Japan. In essence, the notion of globalization and, in particular, the extent of globalization of firms that international business (IB) scholars understand is being put under renewed investigation. At a more micro level, understanding the internationalization of firms and predicting the determinants of international success or failure of these firms is a critical issue in IB research (Buckley & Ghauri, 2004; Peng, 2004). While the ‘regionalization’ research (Rugman, 2000) is quite revealing, any generalization to other contexts is prone to error, given that the largest MNEs constitute for only a very small proportion of the firms that operate internationally. As per UNCTAD (2003), there were over 64,000 firms operating internationally, controlling at least 870,000 foreign affiliates. Most of these firms are significantly smaller, in terms of sales, compared to the Global 500 MNEs, and yet are active in the international arena. Many of such firms are based in the emerging markets like India, China, Brazil, and Russia, and are very different from the conventional MNEs in terms of resources, origins, growth, organizational forms, and internationalization patterns and strategies (Mathews & Zander, 2007). Our understanding of the internationalization process, path, and performance of emerging market firms is very limited and in light of the ‘regionalization’ theory of Rugman and Verbeke (2004) calls for a systematic exploration. In this paper, we focus on the internationalization pattern of firms from the emerging market of India. Through descriptive analysis of archival data on foreign investments of Indian firms, we explore the nature of regionalization prominent in the internationalization process of Indian firms. Hence, we address questions such as the following. Have Indian firms internationalized primarily by entering into certain regions or blocks of countries? If true, does this regionalization trend vary by time or sector or both? What are some possible theoretical explanations of such a regionalization approach to internationalization? We answer the above through data obtained from the India Investment Center, UNCTAD publications, Indiastat.com (2005), and from a few India-centric studies.

Internationalization of Indian Firms

203

Next, we address a specific and the more relevant research question for company managers, pertaining to the relationship between the degree of internationalization and firm performance in the unique context of an emerging market, India. While this relationship has been studied and empirically tested extensively using multinationals from the developed economies (Hitt, Bierman, Uhlenbruck, & Shimizu, 2006), there seem to be only a few studies exploring this phenomenon from an emerging market context (Nachum, 2004). Company managers of newly internationalizing Indian firms will be interested in finding out the profit impact of increasing internationalization of their company’s operations. We use international data on Indian manufacturing and service firms for a period of 5 years, from 1997 to 2001, to test the relationship between internationalization and firm performance in the context of India. Through this study, we extend the research stream on the relationship between internationalization and firm performance for emerging market firms. Given the goals of this study, an Indian context is appropriate for several reasons. First, India is the second largest emerging economy after China and is currently witnessing a phenomenal internationalization of its firms (Economist Intelligence Unit & Columbia Program on International Investment, 2006). Exports and international sales for the Indian economy have been achieved mainly by Indian firms going abroad rather than through subsidiaries of foreign MNEs. At the aggregate level of analysis, regional sales of Indian firms depict trends in contrast to what has been shown for developed economies. As such, the Indian context of this study can shed light on the regionalization aspect of emerging economy firms. Second, Indian firms have arrived at the international stage only lately. An investigation into these will help understand the effect of internationalization for latecomers from other emerging markets. Third, India has been relatively understudied by international researchers as compared to other emerging economies, particularly China. Finally, the use of an Indian sample could provide a good complement to developed market(US, Europe, and Japan) and emerging market-based (Mexico) samples on which most of the prior research on internationalization–performance has been conducted (e.g. Contractor, Kundu, & Hsu, 2003; Ruigrok & Wagner, 2003; Lu & Beamish, 2004; Thomas, 2006).

REGIONALIZATION STRATEGY OF MNES Rugman (2000) and Rugman and Verbeke (2004) have empirically demonstrated that the largest MNEs in the world pursue regional strategies,

204


with a focus on their home triad region. An astounding revelation of their study, in particular for management scholars propagating ‘global strategy of firms’, is that only 9 out of the 500 largest companies in the world have sales and operations distributed equally in all the 3 major triad markets and can hence be called truly global companies. All the others follow a regional strategy or at the best a bi-regional strategy and hence are regional MNEs. Evidence of companies pursuing regional rather than global strategies has also been shown in the context of emerging economies. Yin and Choi (2005) demonstrate that most inward foreign direct investments (FDI) into China emanate from other Asian countries, Hong Kong being the biggest source destination, and that the conventional triad economies constitute only a minor share in China’s total inward FDI. Moreover, the story of Chinese enterprises’ international expansion is even more supportive of the regionalization phenomenon. Out of the 11 largest state-owned enterprises that are now part of the Global 500 companies, only 5 have ventured into international territories, that too mainly in Asia, the Middle East, and parts of Africa. The remaining six companies have only domestic operations. The theoretical rationale for pursuing a regional strategy derives mainly from the transaction cost economics perspective (Rugman & Verbeke, 2005). Successful international expansion by a firm requires that it makes investments in order to better utilize its firm-specific resources and capabilities in the given international location, which might further be characterized by certain specific locational advantages. These investments may take multiple forms like adapting of products and processes, building social capital, gaining legitimacy in eyes of the new stakeholders, and learning cultural and institutional norms and values. Most of these and other similar investments usually become larger, as a firm moves to locations that are distant from its home region. Hence, firms usually expand sales and related operations primarily in their home regions and have only sparse operations in non-home regions. Benefits from such regional expansion will be enhanced where institutional and economic differences between countries of the region are limited and where mutual understanding between governments in the form of economic unions or trading blocks exist. However, a firm’s home country context has a significant bearing on firm strategy pertaining to internationalization path, process, and even performance (Wan & Hoskisson, 2003). Emerging economies have highly uncertain business environments and many of these countries have recently undergone a massive institutional and economic transformation (Choi, Lee, & Kim, 1999). Due to varying institutional contexts, conclusions drawn in one context may not apply equally well in another context (Khanna & Rivkin, 2001).


205

The uniqueness of the emerging economy context can be gauged from the error-prone risk assessments for the emerging economies done by credit rating agencies (McNamara & Vaaler, 2000). It is these unique aspects that make the study of internationalization of firms from the emerging economies interesting and useful for understanding future outward investments. Emerging economies and the BRIC (Brazil, Russia, India, and China) countries in particular are being given special attention as firms from these countries are now also becoming a significant source of outward FDI (O-FDI). Indian O-FDI stock has grown from $0.6 billion in 1996 to $6.6 billion in 2004 (UNCTAD, 2005), and these O-FDI flows are increasing consistently. Most of the Indian O-FDI has been in the form of cross-border mergers and acquisitions, mainly in telecommunications, energy, pharmaceuticals, software, and offshore outsourcing service sectors (Economist Intelligence Unit et al., 2006). Somewhat surprising is the fact that many of the recent cross-border mergers and acquisitions have been conducted in the developed economies, evidenced by the aggressive acquisition of Corus (UK steel producer) by Tata Steel (largest private Indian steel producer) in 2007, buyout of Betapharm Arzneimittel (German pharmaceutical firm) by Dr. Reddy’s Laboratories in 2006, purchase of Eve Holding (a Belgium industrial firm) by Suzlon Energy (Indian energy firm) in 2006, and similar other O-FDIs from India. Besides these, Indian firms such as Infosys, Wipro, Satyam, and Tata Consultancy Services from the information technology/ software/outsourcing sector; ICICI and HDFC from the financial services sector; and Ranbaxy Laboratories, Sun, Cipla, Cadila, Wockhardt, and Cipla from the pharmaceutical sector have caught the attention of academic scholars (e.g. Khanna & Palepu, 2006, Ramamurti, 2004) and business consultants (Sinha, 2005) due to their international success.

REGIONALIZATION OF INDIAN O-FDI Given the difficulties in data availability in emerging economies (Hoskisson, Eden, Lau, & Wright, 2000), we were not able to find regional dispersion of sales at the firm level. However, we did find data on the aggregate country level which depicts the nature of regionalization in India’s O-FDI and more interestingly the changing pattern of it in terms of time and sector. Table 1 presents the sectoral and regional dispersion of Indian firm’s internationalization in terms of number of O-FDI made and the equity of O-FDI, in three time periods (1975–1990, 1991–1995, and 1996–2001). The figures for

206


Table 1. Sectoral and Regional Distribution of Outward Foreign Direct Investments (O-FDI) from Indiaa. 1975–1990

1991–1995

1996–2001

Number of O-FDI (% of total) Manufacturing Services Developing Developed

55.65 43.04 75.02 27.95

53.85 45.75 58.22 41.77

45.82 53.95 40.52 59.47

Equity of O-FDI (% of Total) Manufacturing Services Developing Developed

65.28 32.91 86.09 13.89

55.37 44.41 60.28 39.7

36.06 62.24 36.2 63.79

Source: The above data are compiled from published reports of India Investment Centre, a Government of India organization, and from database on Indiastat.com a Exploration and refining of oil, minerals, and precious stones is treated as separate from Manufacturing and Services, and accounts for the remaining percent of O-FDIs.

both, number of O-FDIs and the equity of O-FDI, show a shift from a focus on developing countries in the 1975–1990 time period to a relatively greater focus on the developed countries in the 1991–1995 and 1996–2001 time periods. Developing countries include Southeast and East Asia, South Asia, Pacific Islands, Africa, West Asia, Central Asia, Central and East Europe, Latin America, and the Caribbean. Developed countries include Western Europe, North America, and a few other developed countries. Similarly, both the number of O-FDIs and the equity of O-FDI of Indian service sector have surpassed that of the manufacturing sector in the latter two time periods. A rather small, yet critical observation to make is that the increase in the equity of O-FDI for the service sector has been greater than the increase in the number of O-FDIs for the manufacturing sector over the three time periods. This not only implies that Indian service firms have been investing abroad more than their manufacturing counterparts, but that their individual investment size is becoming bigger in recent times. Looking at both these trends, of manufacturing vs. service sector O-FDI and developed vs. developing O-FDI, together, we can infer that the Indian service firms have been more aggressive players in the international arena since the historic liberalization drive of the Indian economy in the early 1990s. One can also infer, with a considerable level of certainty that


207

a significant part of the O-FDI of Indian service firms has been in the developed economies. Given the macro nature of data available, we cannot make inferences on the degree of regionalization of investments by individual Indian firms. We can, however, observe regionalization at the country level of analysis. This regionalization pattern is different from what we see in the case of Western MNEs and is also dynamic. Our data pertaining to top Indian O-FDI destinations suggest that the United States is the most important destination accounting for 19% of the total O-FDI flows from 1996 to 2003. The Russian Federation stands at number 2 position with a 17.6% share. Bermuda, British Virgin Islands, and Mauritius are the other top destinations for India’s O-FDI, solely because of tax reasons. The United Kingdom and France are the other developed countries for which India is becoming an important investor. Overall, our data suggests that during the 1975–1990 time period majority of the O-FDIs were in developing economies and hence indicative of avid regionalization, assuming India’s home region to be primarily developing (South Asia, Southeast Asia, and Central Asia). Since the 1990s most of India’s O-FDI has been directed toward the developed economies (Western Europe and North America), which in absolute terms on a world level is still very miniscule, is indicative of a nonregional approach to internationalization. While the objective of the study is not to challenge and refute the regionalization thesis, which given the recent internationalization of Indian firms is beyond the scope of this study, we do provide some basic explanations for such a trend in India’s O-FDI. The home markets of emerging economy firms are very small when compared to that of advanced economies (Ray, 2004). Consequently, most of the emerging market firms tend to be much smaller, even in the large emerging nations such as India. Scholars such as Lall (1983) and Wells (1983) studied developing country multinationals and qualified them as typically small and resource-deficient. International expansion powerfully complements local market size and enables them to achieve scale economies. Since the major markets, attributable to the higher level of income, lie in the developed economies of Western Europe and North America, many emerging economy firms are enticed to venture into these distant markets even before they have expanded in their home-region. Emerging economy firms typically operate in relatively higher-risk home economies subject to uncertain structural changes (Nachum, 2004). Diversifying into more stable international markets, which usually are in the developed economies, significantly minimizes the impact of market failure in focusing on only one single market, their home country, for

208


accessing resources and selling their products (Rugman, 1979). Emerging economies like India are heavily regulated, with constraints on private enterprise (Kumar & McLeod, 1981). However, these economies are undergoing rapid institutional and economic transformation, allowing not only MNEs to enter their home markets but also providing a solid platform for their home-grown firms to enter foreign markets. Given the high incentives to internationalize for emerging economy firms, many of them are ready to risk entering into geographically and culturally distant markets of the developed economies. Delios and Beamish (2005) conclude that global strategies can be pursued by firms if they have high R&D, advertising, and export intensities as they result in non-location bound firm-specific assets which can be easily deployed in different contexts. By this token, emerging economy firms will be restricted within their home regions, since most lack the resources and capabilities that the developed economy firms possess (Hitt, Dacin, Levitas, Arregle, & Borza, 2000). In fact, this has been the case and as a result most emerging economy firms continue to only have domestic operations. Some of them, though, have been able to come up with organizational and strategic innovations compensating for their lack of financial and managerial capabilities (Mathews, 2006). However, increasing number of home-grown Indian firms like the Tata group of companies, ICICI, and Infosys have been improving their ownership-specific advantages, including financial and managerial capabilities. Moreover, the harsh environmental conditions such as a weak institutional context, demanding yet price-sensitive consumers, and challenging distribution networks in their home markets, instead of acting as impediments, have helped emerging economy firms to develop unique competencies, to be later used to compete successfully in foreign markets (Sinha, 2005). For example, Ranbaxy Laboratories by serving the need of cheap drugs for the huge Indian masses developed into one of the world’s leading generic drug manufacturer. The rise of companies such as Samsung from Korea and Acer from Taiwan into powerful global giants clearly demonstrates this phenomenon. More importantly, these unique competencies are equally non-location bound and can hence be deployed in an effective and speedy manner even in the developed economies. Kostova and Zaheer (1999) proposed that with increasing institutional distance between the home and host countries, transfer of strategic routines and establishment of legitimacy becomes more challenging. By this token, emerging economy firms will face an uphill task of establishing offices in developed economies and profiting from them. India, however, with its colonial past and the widespread use of English language in the corporate


209

sector occupies a rather favorable position with regards to Indian firms expanding and culturally assimilating faster in developed markets like US (McManus & Floyd, 2004). Moreover, the social and relational capital from the significant Indian diaspora in developed economies has been very effective in providing know-how, market access, capital, and overall guidance in the United States and Europe (Kapur & Ramamurti, 2001). More generally, emerging economy firms that enter into developed economies face a much better institutional framework than what they are used to in their home economies, which makes linking of their firm resources to the location-based advantages not extremely difficult. In certain instances, they benefit more from such internationalization. In addition to the above explanations of India’s recent O-FDI going into developed economies, the case of Indian service sector, software in particular, deserves a special mention. The recent dramatic international growth seen in Indian service companies, particularly, in areas such as computers and software (NASSCOM, 2001) has been achieved mainly in the US and Western European markets. The biggest factor has been the huge untapped market in these developed economies and the cost-competitiveness of the Indian players vis-a`-vis their developed market counterparts. In general, Indian service-sector firms have been able to pursue aggressive international expansion because of (i) lower initial capital requirements, (ii) the ability to separate the value chain across international boundaries between core activities (e.g. back-end software program development performed in India) and supplementary activities (such as front-office consulting and marketing activities, performed in Europe or the USA), and (iii) significantly lower government controls and interference over online business activities. In addition, the service sector has benefited tremendously from the advances made in education in India. Over the last 15 years, the number of academic institutions for higher learning more than doubled from 6,761 in 1991 to 14,143 in 2003. This has resulted in a much larger stock of qualified personnel in India. The total number of doctoral degrees awarded in science and engineering disciplines increased by 81% from 5,049 in 1996 to 9,131 in 2000 (www.indiastat.com). All this has helped the Indian service sector, predominantly knowledge-intensive in nature (software, IT, engineering, and healthcare) to expand internationally. Also, such knowledge-intensive services are less prone to severe adaptation or cultural assimilation owing to the standard nature of these service offerings and thus can be offered in a variety of contexts with minimal incremental costs. In the next section, we empirically test the impact of the increasing degree of internationalization of Indian firms on their financial performance. Li (2005),

210


in his study of 574 US service firms for the period from 1997 to 2001, found that a home triad-based regional strategy was financially more rewarding. That is, US service firms that pursued a regional strategy outperformed the ones that pursued a more global strategy. Due to paucity of data, we cannot test for the moderating effect of regional vs. global strategy on the internationalization–performance relationship of Indian firms. However, we have data on both services and manufacturing firms for a similar time period (1997–2001), and given our descriptive analysis of the internationalization strategy of Indian services (more globally oriented) and manufacturing (more regionally oriented) in the recent past, we can make certain inferences on the performance implications of regionalization in the context of Indian firms. Since the prevalence of large diversified business groups is an important aspect of many emerging economies the world over (Yiu, Bruton, & Lu, 2005) and is a sign of responding to and coping with the institutional voids in the weak institutional contexts of emerging economies (Chang & Choi, 1988; Khanna & Palepu, 1997, 2000a, 2000b; Leff, 1978), we control and test for the impact of this key variable on the internationalization–performance link.

INTERNATIONALIZATION–PERFORMANCE LINK OF INDIAN FIRMS Sample We implemented our investigation on a sample of Indian firms belonging to manufacturing and services sectors. We derived our list of Indian firms from the annual database, CAPITALINE, compiled by Capital Market Publishers India Private Limited. This database is built on raw data obtained from reports that all companies file with the Registrar of Companies, a federal agency. We obtained data from 1997 to 2001 to develop a longitudinal data set covering five years. For each year, we had complete information on foreign sales, total sales, return on sales, and age for 240 firms. We obtained the information about business group affiliation of these firms from ISI Emerging Markets database. Variables Dependent Variable We chose return on sales (ROS) as the dependent variable. ROS is one of the accounting-based measures of performance, commonly used in literature on firm performance (Li, 2005).


211

Explanatory Variables The degree of internationalization (DOI) and group affiliation (Group Affiliation) are the two independent variables used in this study. We measure degree of internalization by the ratio of foreign sales to total sales. We measured group affiliation by an indicator variable, which took the value of 1 if the firm was affiliated to a business group and 0 if it was not. Control Variables We controlled for size of the firms (Sales), age of the firm (Age), and industry fixed effects. We measured firm size by the natural logarithm of total sales. We measured age of the firm by total number of years since inception. To control for industry effects we used an industry dummy variable. Modeling Procedure We examined the performance implications of DOI and business group affiliation using a firm-year unit of analysis over the five-year period from 1997 to 2001. With firm-year records for performance analysis, we used general linear square (GLS) random-effects models to test the hypotheses. Random effects estimator is appropriate when unobserved heterogeneity is not correlated with the independent variables. We tested for this assumption using Hausman test (Baltagi, 1995, p. 68) and found no significant difference between random effects and fixed effects estimates (w2=9.93, d.f.=15, p=0.82). Therefore, we chose the random effects estimation for both the equations. Results Table 2 provides the descriptive statistics and correlations. As we can see, 40% of the firms in the sample were business group affiliated. Average age of the firms in the sample was 25.45 years in the year 2001. The average DOI for these firms was 21.58% (logit transformed value= 1.29). Table 3 presents the results of regression with ROS as the dependent variable for full sample as well as for the manufacturing- and service-sector subsamples. We built the models in a hierarchical manner. Model 1 has firm size (Sales), age, manufacturing dummy, and DOI besides industry indicator variables. We introduced a square term of DOI in Model 2. Model 3 has the business group affiliation variable and Model 4 has the interaction between DOI and business group affiliation. Model 5 has the interaction between manufacturing dummy and DOI. Model 6 has all the effects together.

212


Table 2. Variables 1. 2. 3. 4. 5. 6.

Descriptive Statistics and Correlationsa. Mean S.D.

4.22 1.49 Salesb Age 25.45 18.19 Degree of internationalizationc 1.29 2.82 Group affiliation 0.40 0.49 Manufacturing dummy 0.51 0.50 Return on sales 5.94 11.36

1 – 0.231 0.075 0.286 0.143 0.230

2

– 0.381 0.227 0.046 0.106

3

– 0.204 0.062 0.343

4

– 0.031 0.060

5

6

– 0.110 –

N=240; Correlation W0.14 and o 0.14 are significant at p=0.05. a Based on data for year 2001. b Logarithmic transformation of sales (Indian national rupees, in million). c Logit transformation of foreign sales to total sales.

We conducted Wald tests on the significance of the inclusion of each additional variable. As shown in the Wald test w2 statistics, the inclusion of the square term significantly improved the model fit (w2=34.9, d.f.=1, po0.001). In Model 3, inclusion of group affiliation does not improve the model fit (w2=0.50, d.f.=1, pW0.10). However, inclusion of interaction terms between DOI – group affiliation and DOI – manufacturing sector dummy in Model 4 (w2=18.4, d.f.=1, po0.001) and Model 5 (w2=10.8, d.f.=1, po0.001) improved the model fit significantly. The coefficient of DOI is positive and significant in all the models (Model 6: b=2.16, po0.001). Coefficient of the square term of DOI is also positive and significant (Model 6: b=0.13, po0.01). This suggests that firm performance improves with increasing DOI for Indian firms. To test for relative performance of manufacturing- and service-sector firms, we look at the coefficient of interaction term between manufacturing dummy and DOI in Models 5 and 6. In both the models, this term takes a negative and significant sign (Model 6: b= 0.88, po0.001), suggesting that internationalization has stronger positive impact on firms belonging to service sector than those belonging to manufacturing sector in the Indian context. Assuming that the Indian services have followed a more globally oriented strategy for internationalization as compared to the Indian manufacturing, this finding is interesting and needs to be further examined for plausible explanations deep routed in theory. Next, we look at the coefficient of group affiliation in Models 3, 4, and 6. The coefficient of group affiliation is negative and significant in Models 4 and 6 (Model 6: b= 2.15, po0.01). Thus, in the Indian context, costs of group affiliation seem to outweigh the benefits, resulting in poorer


213

Table 3. Regression of Firm Performance on Degree of Internationalization and Group Affiliationa. Variables

Full Sample Model 1

Salesb Age Manufacturing dummy Degree of internationalizationc (Degree of internationalization)2 Group affiliation

0.65** (0.21) 0.01 (0.02) 4.55*** (1.25) 0.91*** (0.11)

Group affiliation DOI Manufacturing dummy DOI Number of cases (per year) Adjusted R2 Wald w2 Wald test w2 (1)

Model 2 Model 3 Model 4 Model 5 Model 6 0.74*** (0.21) 0.02 (0.02) 5.57*** (1.25) 1.34*** (0.14) 0.15*** (0.03)

0.78*** (0.22) 0.02 (0.02) 1.39 (2.03) 1.34*** (0.14) 0.15*** (0.03) 0.43 (0.62)

0.80*** (0.22) 0.02 (0.02) 6.06*** (1.02) 1.64*** (0.16) 0.14*** (0.03) 1.77** (0.70) 0.80*** (0.21)

0.78*** (0.22) 0.02 (0.02) 7.05*** (1.34) 1.68*** (0.18) 0.15*** (0.03) 0.49 (0.61)

0.65** (0.22) 240 0.18 264***

0.82*** (0.22) 0.02 (0.02) 8.12*** (1.34) 2.16*** (0.21) 0.13*** (0.03) 2.15** (0.71) 0.98** (0.21) 0.88*** (0.22)

240 240 240 240 240 0.20 0.20 0.21 0.21 0.23 298.9*** 299.4*** 317.7*** 310.2*** 337.2*** 34.9*** 0.50 18.4*** 10.8*** 37.8***

Note: Coefficient estimates for industry indicator variables not presented in the table; upper number in a cell is a parameter estimate; numbers in the parentheses are standard errors; young firms (average ageo16 years), old firms (average age W=16 years), 16 years was the median age. **po0.01; ***po0.001. a Return on sales in the dependent variable. b Logarithm of sales used in regression. c Logit transformation of degree of internationalization used in regression.

performance for group-affiliated firms. To test for the moderating effect of group affiliation, we look for the coefficient of interaction term between group affiliation and DOI in Models 4 and 6. This term has a negative and significant coefficient in both the models (Model 6: b= 0.98, po0.001), giving indirect support to the regionalization theory. Firms that have responded to the institutional context of a particular location by being affiliated to a business group are not able to link their resources effectively in distant international markets and hence have inferior performance compared to those internationalized firms that are not affiliated to a business group.

214


DISCUSSION AND CONCLUSION We examined the relationship between degree of internationalization and firm performance in the Indian context. We found that a positive and exponentially increasing curve was the best statistical fit for the internationalization-performance link of Indian firms. Theoretically, Indian internationalizing firms should also go through the initial decline in performance due to initial liabilities of foreignness and newness, cultural and institutional differences, learning and adaptation costs, etc. One possible explanation of the absence of the initial negative impact of internationalization could be due to the nature of the sample of firms selected for empirical analysis. We found significant coefficients for some of the industry fixed effects, which suggest that the shape of the internationalization–performance curve will depend on the level of internationalization of the various subsectors within the broad manufacturing and services sector. That is to say, if firms in a sub-sector are almost equally present in the early as well as in the growth stages of internationalization, then the shape of the statistically fitted curve could be U-shaped. If almost all firms in a sub-sector are in the early stage of internationalization, the shape of the fitted curve would simply be a downward sloping linear curve. By contrast, sub-sectors that are mainly in their growth stage, and include a number of relatively large internationalized companies, would have an upward sloping linear curve. Another explanation of such a relationship could emanate from the unique home institutional context of Indian firms. In India, relatively few firms indulge in international operations. The ones, which have international operations, have some really superior capabilities such as cost competitiveness against their competitors due to cheaper manpower, exclusive rights over cheaper raw material etc. Moreover, as discussed earlier, the small home market and high economic and political risk of the Indian economy act as complementary forces for greater benefits from internationalization. Emerging economy firms also get lots of incentives from local governments in their internationalization efforts. These factors lead to consistently increasing performance with increasing degree of internationalization. Next, based on archival data we focused on the issue of regionalization in the internationalization of Indian firms. We analyzed the pattern of India’s O-FDI in terms of sectors and recipient destinations over three distinct time periods (1975–1990, 1991–1995, and 1996–2001). We found that while majority of O-FDI projects were in the manufacturing sector and in other developing economies in the first time period (1975–1990), the


215

pattern changed dramatically in the latter two time periods (1991–1995 and 1996–2001). Service firms accounted for a greater share in the total O-FDI and the recipient destination also shifted from developing to developed economies over time. We provide some arguments for India’s O-FDI increasingly being directed toward developed economies, primarily based on factors such as large markets and better institutional contexts in developed economies and due to the compelling need of Indian firms to internationally diversify to more stable markets. Linking the two trends of changing industrial sector from which O-FDIs emanate and changing destination of India’s O-FDI, we conclude that Indian service sector have pursued a more global strategy by expanding into distant markets of the developed economies, while the Indian manufacturing sector has gone through a slower internationalization characterized by dominant regional strategy. Assuming that this has been the case in reality, we test for which strategy – more globally oriented or more regionally oriented – has translated to greater financial performance. We found that service firms have outperformed the manufacturing firms in deriving better performance from internationalizing their sales and operations. To gain insight into the differential effect of internationalization for manufacturing- and service-sector firms, we plotted the interaction between DOI and manufacturing dummy in Fig. 1. The slope of line for servicesector firms is steeper than that for manufacturing-sector firms suggesting that as DOI increases, performance of service-sector firms improves more than manufacturing-sector firms. This can be related to the phenomenal

Manufacturing

Services

Performance (ROS)

20 15 10 5 0 5%

Fig. 1.

15% 25% 35% 45% 55% 65% 75% 85% 95% Degree of internationalization (Foreign sales to total sales)

Differential Impact of Internationalization on Performance of Manufacturingand Service-Sector Firms.

216


growth in the Indian software sector in particular, along with the growth in other knowledge-based services. Interestingly, anecdotal evidence suggests that significant growth in India’s knowledge-intensive services has been in the developed economies. This opens room for renewed theoretical inspection of the advantages that a regional strategy has to confer to internationally expanding firm from an emerging economy. Further, we investigated the effect of business group affiliation on firm performance and how group affiliation interacts with the DOI to affect firm performance. We found a significant moderating role of business group affiliation on the internationalization–performance relationship. To gain insight into this moderating effect, we plotted the interaction effect of group affiliation and DOI on firm performance. Fig. 2 depicts this interaction effect. The relative slopes of the two curves for group affiliated and unaffiliated firms show this effect. The line for unaffiliated firms crosses the line for group affiliated firms and goes above, as the level of internationalization increases. This suggests that as the DOI increases, affiliation to a business group becomes detrimental for firm performance. Highly internationalized firms perform better if they are not affiliated to some business group. An alternative interpretation, as mentioned earlier, pertains to group affiliated firms being too ingrained in the domestic context in terms of their resources and capabilities, which makes synchronization with distant foreign locational advantages very difficult to achieve.

Affiliated

Unaffiliated

Performance (ROS)

15 13 11 9 7 5 3 5%

15%

25%

35% 45% 55% 65% 75% Degree of internationalization (Foreign sales to total sales)

85%

95%

Fig. 2. Moderating Effect of Business Group Affiliation on Internationalization– Firm Performance Relationship.


217

Our study advances knowledge of emerging economy firms by studying their internationalization patterns and performance. Prior studies show significant differences in firm behavior in emerging markets and implicitly address the need to study internationalizing firms from such economies separately (Wells, 1998). Most research has already focused on companies from developed economies, especially the United States, and little is known about international firms from emerging economies (Luo & Peng, 1999; Peng & Luo, 2000). This, coupled with the unique nature of emerging economies, makes this study useful and interesting. Moreover, this study also serves as a critical extension to the research on internationalization and firm performance by focusing on a newer context and bringing out the role of regionalization.

ACKNOWLEDGMENT This research was supported by the Bocconi University DIR research grant (no. 601862) and by an Asia Research Institute grant.

REFERENCES Baltagi, B. (1995). Econometric analysis of panel data. New York: Wiley. Buckley, P. (2002). Is the international business research agenda running out of steam? Journal of International Business Studies, 33, 365–373. Buckley, P., & Ghauri, P. (2004). Globalization, economic geography, and the strategy of multinational enterprises. Journal of International Business Studies, 35(2), 81–98. Chang, S. J., & Choi, U. (1988). Strategy, structure, and performance of Korean business groups: A transactions cost approach. Journal of Industrial Economics, 37(2), 141–158. Choi, C. J., Lee, S. H., & Kim, J. B. (1999). A note on counter-trade: Contractual uncertainty and transaction governance in emerging economies. Journal of International Business Studies, 30(1), 189–202. Contractor, F. J., Kundu, S. K., & Hsu, C. C. (2003). A three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34(1), 5–19. Delios, A., & Beamish, P. (2005). Regional and global strategies of Japanese firms. Management International Review, 45(Special Issue 1), 19–36. Economist Intelligence Unit & Columbia Program on International Investment. (2006). World investment prospects to 2010: Boom or backlash? London: EIU. Hitt, M. A., Bierman, L., Uhlenbruck, K., & Shimizu, K. (2006). The importance of resources in the internationalization of professional service firms: The good, the bad, and the ugly. Academy of Management Journal, 49(6), 1137–1157.

218


Hitt, M. A., Dacin, M. T., Levitas, E., Arregle, J. L., & Borza, A. (2000). Partner selection in emerging and developed market contexts: Resource-based and organizational learning perspectives. Academy of Management Journal, 43(3), 449–468. Hoskisson, R. E., Eden, L., Lau, C. M., & Wright, M. (2000). Strategy in emerging economies. Academy of Management Journal, 43(3), 249–268. Indiastat.com (2005). Indiastat.com, revealing India y statistically. Retrieved on December 20, from, www.indiastat.com Kapur, D., & Ramamurti, R. (2001). India’s emerging competitive advantage in services. Academy of Management Executive, 15(2), 20–31. Khanna, T., & Palepu, K. (1997). Why focused strategies may be wrong for emerging markets? Harvard Business Review, 75(4), 41–51. Khanna, T., & Palepu, K. (2000a). Is group affiliation profitable in emerging markets? An analysis of diversified Indian business groups. Journal of Finance, 55(2), 867–892. Khanna, T., & Palepu, K. (2000b). The future of business groups in emerging markets: Long run evidence from Chile. Academy of Management Journal, 43(3), 268–285. Khanna, T., & Palepu, K. (2006). Emerging giants: Building world-class companies in developing countries. Harvard Business Review, 84(10), 60–69. Khanna, T., & Rivkin, J. W. (2001). Estimating the performance effects of business groups in emerging markets. Strategic Management Journal, 22, 45–74. Kostova, T., & Zaheer, S. (1999). Organizational legitimacy under conditions of complexity: The case of the multinational enterprise. Academy of Management Review, 24, 64–81. Kumar, K., & McLeod, M. G. (Eds). (1981). Multinationals from developing countries. Lexington, MA: Lexington Books. Lall, S. (1983). The new multinationals: The spread of third world enterprises. Chichester, UK: Wiley. Leff, N. (1978). Industrial organization and entrepreneurship in the developing countries: The economic groups. Economic Development and Cultural Change, 26, 661–675. Li, L. (2005). Is regional strategy more effective than global strategy in the US service industries? Management International Review, 45(Special Issue 1), 37–57. Lu, J. W., & Beamish, P. W. (2004). International diversification and firm performance: The S-curve hypothesis. Academy of Management Journal, 47(4), 598–609. Luo, Y., & Peng, M. W. (1999). Learning to compete in a transition economy: Experience, environment, and performance. Journal of International Business Studies, 30(2), 269–296. Mathews, J. A. (2006). Dragon multinationals: New players in 21st century of globalization. Asia Pacific Journal of Management, 23, 5–27. Mathews, J. A., & Zander, I. (2007). The international entrepreneurial dynamics of accelerated internationalization. Journal of International Business Studies, doi:10.1057/palgrave. jibs.8400271. McManus, J., & Floyd, D. (2004). A macro and micro perspective of the global software industry with specific orientation to India, China and the Philippines. Journal of Marketing and Logistics, 16(4), 52–65. McNamara, G., & Vaaler, P. M. (2000). The influence of competitive positioning and rivalry on emerging market risk assessment. Journal of International Business Studies, 31(2), 337–347. Nachum, L. (2004). Geographic and industrial diversification of developing country firms. Journal of Management Studies, 41(2), 273–294. NASSCOM. (2001). The Indian software industry report. New Delhi: NASSCOM.


219

Peng, M. W. (2004). Identifying the big question in international business research. Journal of International Business Studies, 35(2), 99–108. Peng, M. W., & Luo, Y. (2000). Managerial ties and firm performance in a transition economy: The nature of a micro–macro link. Academy of Management Journal, 43(3), 486–501. Ramamurti, R. (2004). Developing countries and MNEs: Extending and enriching the research agenda. Journal of International Business Studies, 35(4), 277–283. Ray, S. (2004). Performance implications of corporate strategic behavior of firms in an emerging economy during economic liberalization. Problems and Perspectives in Management, 2, 246–256. Rugman, A. M. (1979). International diversification and multinational enterprise. Lexington, MA: Heath. Rugman, A. M. (2000). The end of globalization. London: Random House and NewYork: Amacon McGraw-Hill. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Rugman, A. M., & Verbeke, A. (2005). Towards a theory of regional multinationals: A transaction cost economics approach. Management International Review, 45(Special Issue 1), 5–17. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43(1), 63–83. Sinha, J. (2005). Global champions from emerging markets. McKinsey Quarterly, 2, 26–36. Thomas, D. E. (2006). International diversification and firm performance in Mexican firms: A curvilinear relationship? Journal of Business Research, 59(4), 501–507. UNCTAD. (2003). World investment report, 2003: FDI policies for development; National and international perspectives. Geneva: United Nations Conference on Trade and Development. UNCTAD. (2005). World investment report, 2005: FDI policies for development; Transnational corporations and the internationalization of R&D. Geneva: United Nations Conference on Trade and Development. Wan, W. P., & Hoskisson, R. E. (2003). Home country environments, corporate diversification strategies, and firm performance. Academy of Management Journal, 46(1), 27–46. Wells, L. T. (1983). Third world multinationals: The rise of foreign investment from developing countries. Cambridge, MA: The MIT Press. Wells, L. T. (1998). Multinationals and the developing countries. Journal of International Business Studies, 29(1), 101–114. Yin, E., & Choi, C. J. (2005). The globalization myth: The case of China. Management International Review, 45(Special Issue 1), 103–120. Yiu, D., Bruton, G. D., & Lu, Y. (2005). Understanding business group performance in an emerging economy: Acquiring resources and capabilities in order to prosper. Journal of Management Studies, 42, 183–206.

THE LOCATION AND PERFORMANCE OF FOREIGN AFFILIATES IN GLOBAL CITIES Lilach Nachum and Clifford Wymbs ABSTRACT We suggest that the entire world may not always be the appropriate frame of reference in analyses of Multinational Enterprises (MNEs) location choices. In some industries and activities, more narrowly defined geographic areas, such as regions and cities, are more relevant level of analyses. Employing global cities as the geographic frame of reference, we extend the theory of the location choices of MNEs by challenging the assumption that location attributes have identical values for all MNEs. Rather, we explicitly acknowledge the relative value of such attributes for individual MNEs, and search for the firm-specific characteristics that affect this variation. The empirical testing is based on analysis of 673 financial and professional service MNEs that entered New York and London via mergers and acquisitions (M&As). The findings confirm that it is the interaction between location and firm-specific attributes, rather than each of these independently, which affects location choices.


221

222

LILACH NACHUM AND CLIFFORD WYMBS

INTRODUCTION The determinants of the location choices of Multinational Enterprises (MNEs) have been a central question in international business. Traditionally this question was studied with reference to the world as a whole, underlined by the (most typically implicit) assumption that the entire globe comprises the level-playing field that MNEs consider (Dunning, 1993). More recently, there has been growing recognition that the appropriate geographic level of analysis may in fact be much smaller than the entire globe. Scholars have acknowledged the importance of the region as a more meaningful unit of analysis (Rugman, 2005; Ghemawat, 2005; Ohmae, 1995). These arguments are underpinned by the observation that intra-regional homogeneity often drives MNEs to concentrate their activities in a single region. Rugman and Verbeke (2004) show that a single region accounts for most of the sales of the majority of the world’s largest firms. Here, we take a further step in this direction and employ global cities as the geographic unit of analysis. We argue that in the financial and professional service industries we study, global cities might be the most relevant geographic level of analysis for the understanding of the determinants of location choices. Research in economic geography provides support to this argument in illustrating the distinctive nature of location attributes of cities (Glaeser & Kohlhas, 2004; Gordon, 2002; Nahm & Semple, 1995). The concentration of financial service MNEs in the world’s global cities suggests that cities, rather than countries, might be the driving force of location choices in these activities (Sassen, 2001). By focusing on global cities, we are able to examine the specific location attributes that appear to be most pertinent in these industries. We fully recognize that the processes taking place at different geographic levels of analyses may differ in kind and hence require separate attention (Lomi, 1995). With reference to global cities, we seek to challenge a major assumption that has underlain Foreign Direct Investment (FDI) theory, namely that the location decisions of MNEs are determined by the relative location advantages of particular locations for certain activities. The sources of such advantages include resource abundance, liberal government policies and favorable institutional environment. Implicit in these formulations is the assumption that particular location advantages have the same value for all MNEs, that is, within an industry, firms value particular resources similarly and benefit from them to the same degree (Dunning, 1993). In this paper, we argue that the characteristics of the investing firms affect their evaluation, and ability to take advantage, of various location

The Location and Performance of Foreign Affiliates in Global Cities

223

advantages. As these characteristics vary across firms, so would vary the value of specific location advantages for them. Hence, the factors affecting location choices are not identical across MNEs and do not exist in isolation from the characteristics of the investing firms. For example, market size is often regarded as a major location advantage, enabling firms to reap the benefits of scale advantages. While this characteristic of markets is a highly valuable advantage for large firms, producing and selling products that enable standardization and mass production, it may have limited, if any, value for small, specialized firms, whose core competitive advantage lies in a highly specialized technology that may involve considerable adaptation to the specific needs of individual customers or small groups of customers. While market size is likely to have significant effect on the location choices of the former, it may have no explanatory power for the location choices of the latter. By acknowledging such differences between firms we introduce a notion of location advantages that are not absolute but rather vary across firms, in line with their firm-specific characteristics. We focus on one specific type of location advantages – the agglomeration externalities that may emerge as a result of the geographic proximity of firms engaged in similar activities. In line with the argument outlined above, we maintain that various firm-specific attributes determine the firm’s ability to benefit from collective dynamics, and hence the value of and benefits from taking part in agglomeration externalities resulting from cluster location. This explains why, while there is a tendency for economic activity within an industry to cluster in concentrated geographic areas, there is nonetheless considerable variation in the location choices of individual firms (Shaver & Flyer, 2000; Fischer & Harrington, 1996). This approach is consistent with the framework advanced by Rugman regarding the interaction between country-specific advantages (CSAs) and firm-specific advantages (FSAs) in determining location choices (Rugman & Verbeke, 2001). In the next section, we review the state of extant research on the factors affecting the location choices of MNEs and position our study vis-a`-vis the few attempts made to incorporate FSAs in location models. We then identify the firm-specific attributes that are likely to affect the location choices of MNEs and advance hypotheses on the nature and direction of their impact. These hypotheses are put forward for empirical test in the following section based on analysis of 673 financial and professional service MNEs who entered New York or London markets via mergers and acquisitions (M&As) during the last two decades. We show that location decisions are affected not only by the location advantages of countries, but also by the firm-specific attributes of individual firms and by the interaction between the two. We find

224


that the characteristics of MNEs, on their own and in interaction with cluster advantages, improve significantly the explanatory power of the location model. Most influential attributes were found to be firms’ length of operation in a market, the geographic scope of their activity and their organizational structure. The paper concludes by discussing the implications of the findings for the theory of the location of MNEs and by suggesting directions in which future research may make further progress. Finding out the extent to which firm attributes affect the value of certain location advantages, and identifying those attributes that are most influential, has important implications for theory and practice. For theory, the interest in the link between firm and location advantages signifies a departure from the traditional conceptualization of location advantages as existing in isolation from the advantages of firms. Examining location choices in light of the heterogeneity among firms may provide a useful starting point for refinement of the theory of MNE location choices. For MNEs, it implies a need to evaluate location advantages with explicit reference to their own unique resources and attributes, and to make their location choices individually, rather than following the ‘norm’ in their industry. It suggests that there is an idiosyncratic element in location choices, and location decisions should be undertaken based on a careful examination of the firm’s specific attributes in relation to specific location advantages rather than viewing the latter on their own.

THEORY AND HYPOTHESES The assumption of the heterogeneity among firms – in terms of the resources and assets they control (Barney, 1991, 2001) and their strategic action to take advantage of these resources (Porter, 1985) – is underlying the attempts of strategic management scholars to understand the determinants of the strategic behavior of firms and their outcomes. These scholars have sought to understand the nature and sources of firms’ heterogeneity and to draw their implications for competitive performance (see Holbrook, Cohen, & Hounshell, 2000) for a comprehensive survey). International business and management scholars, often drawing directly on these conceptualizations, notably the resource-based view of the firm, have only recently begun acknowledging the heterogeneity among MNEs as an important factor that provides a fuller understanding of their strategic behavior. Attempts were made to introduce systematically the idiosyncratic attributes of individual MNEs to models explaining choice of entry mode and


225

subsequent performance (Shaver, 1998). Heterogeneous investment motives were found to affect the impact of MNEs on productivity of the host economies (Chung, 2001b). Heterogeneity in entry mode, investment size and location were shown to explain variation in the impact of MNEs on host country competition (Chung, 2001a), and the tempo and location of sequential investment and reaction to environment changes (Song, 2002). Taken together, these studies show the limitations of the traditional approach whereby all MNEs within an industry were treated alike, and the merit of acknowledging systematically intra-industry variation in their behavior. There have been only a few attempts to apply a similar approach toward location decisions, and to examine whether there are some relationships between the location choices of MNEs and their firm-specific attributes (Kravis & Lipsey, 1982; Shaver & Flyer, 2000; Chung & Alcacer, 2002). Most previous analyses assume, implicitly or explicitly, that all MNEs within an industry have identical resource requirements and search for the same complementary resources in their external environment, that is, they have no individual preferences in terms of their location choices, beyond those related to industrial affiliation.1 The tendency to rely on industry-, rather than firm-level data, to test the determinants of the location choices of MNEs (see e.g., Yamori (1998) for a representative approach) has further obscured the differences among MNEs within an industry in this regard. In line with this industry approach, the models constructed to identify the factors affecting the location decisions of MNEs typically include a set of country (or geographic areas within them) characteristics that are hypothesized to affect the choice of MNEs between potential locations (Dunning, 1993). These models do not take account of possible influences of firm-specific attributes on this choice. The argument underlying this paper is that since firms differ in terms of their skills, the assets they control, their strategic objectives and the way they organize their international activity, identical location advantages have different values for them. Different capabilities imply different needs for complementary resources sought in the external environment, and different ability to use them as the basis for competitive advantages. We put the heterogeneity among firms at the center of the analysis and search for the firm-specific attributes that may affect the value of specific location advantages for them, and hence their location choices. We thus eliminate the traditional distinction between location and firm advantages that had underlain, explicitly or implicitly, the theories of international business and introduce a different notion, which emphasizes the interaction between the two as the factor determining the location choices of MNEs.

226


By so doing, we build on the work of Shaver and Flyer (2000), Chung and Alcacer (2002) and Kravis and Lipsey (1982), and advance their arguments in several directions. Theoretically, our conceptualization of the firms’ attributes that affect their location choices is broader than the one undertaken by these studies. Shaver and Flyer (2000) refer to the strength of firms’ technological competence as the major factor affecting the net benefits they receive from agglomeration externalities and hence their attraction to clusters. Chung and Alcacer (2002) put the differences in R&D intensity between the home and host countries as the driving factor of agglomeration tendency. Kravis and Lipsey (1982) assign the selection of country locations primarily to the size of firms and the composition of their production factors. We broaden this conceptualization and argue that a whole range of firm attributes affects their location choices. We incorporate the differences among firms directly and systematically into a model of location choices, and analyze their influence on the location choices of MNEs. By bringing these different determinants together into a single model we can assess their relative magnitude. Methodologically, Kravis and Lipsey (1982) used industries, rather than firms, as the unit of analysis. Shaver and Flyer (2000) and Chung and Alcacer (2002) use firm data as the dependent variable but their explanatory variables are, for the most part, at the level of industries and/or countries (or regions within them). This has fundamental implications for what is being measured and tested. The aggregation of firms within an industry or a country is a major deficit when attempting to uncover the firm-specific attributes that affect location choices, where the variation across firms within an industry is the major issue of interest. By using firms and their specific attributes as the unit of analysis, we base our analysis on firm characteristics that aggregated industry data naturally does not possess, and we are thus able to illuminate the variation across firms within an industry. We limit the analysis to one kind of location advantages – agglomeration externalities emerging from cluster participation.2 We argue that the view of location advantages as isolated from the characteristics of firms is particularly inadequate with reference to clusters. The advantages that such a context provides are created by processes taking place between firms located in geographic proximity to each other (Scott, 1998), rather than by the relative abundance of particular resources that are external to firms. Under such circumstances, the characteristics of the firms themselves create the advantages of a location and the two cannot be seen in isolation. In the rest of this section, we draw on the MNE literature, combined with insights from the literature on the behavior of firms in clusters, to identify


227

the firm-specific attributes that are likely to affect the value of cluster location for MNEs, and advance hypotheses as to the nature and direction of such impact. Size Size differences are hypothesized to affect the location choices of MNEs via their impact on the needs of firms for complementary resources. A major reason for cluster location is the search by firms for complementary resources that they do not possess (Scott, 1998). Such resources are often more abundant in and around clusters than elsewhere, and geographic proximity reduces the costs associated with accessing them. Large firms tend to internalize certain activities and are less dependent on external resources than their smaller counterparts, and hence might have less need for taking part in cluster dynamics for this reason. Several studies have shown that smaller firms are significantly more likely to agglomerate than larger ones (Rauch, 1993a; Shaver & Flyer, 2000). Considerable evidence, although often without specific spatial reference, shows that small firms tend to engage in networking relationships more than large ones, and that they use these relationships to compensate for the lack of resources that larger firms obtain internally (e.g., Chen & Chen, 1998). Formally: H1. The size of firms significantly affects their location choices, with the distance from the center of business activity positively related to the size of firms (i.e., the larger the firms, the more likely they are to be located away from the center), ceteris paribus. Geographic Scope The geographic scope of the activities of firms is likely to affect their need for, and benefit from cluster participation. The more firms have operations outside the cluster, the greater their need to tap into diverse knowledge network and the more constrained is the value of local knowledge for them. For example, a Japanese investment bank with operations in, e.g., Singapore, London, Hong Kong, would, on a relative basis, be less dependent on New York-specific knowledge than if the only foreign operation of this bank were in New York. Cluster participation may not be sufficient to the former, as it will not provide the knowledge needed for operations that take place beyond the boundaries of the cluster. Hence, firms whose operations are

228


geared toward markets external to the cluster may gain fewer benefits from taking part in local dynamics. This is consistent with studies that found that the more firms sell outside the cluster, the greater are their needs for sources of knowledge that are not available locally (e.g., Saxenian, 1994; Scott, 1998). Some support for this can be found in Shaver and Flyer’s (2000) analysis where they found that one possible reason for firms to cluster might be the different markets these firms serve. Firms serving national and international markets are more likely to locate away from the cluster than those that serve mostly local customers. Formally: H2. The geographic scope of firms significantly affects their location choices, with the distance from the center of business activity positively related to the scope of activity (i.e., the greater the scope of activity, the more likely are firms to locate away from the center), ceteris paribus. Innovation Recent conceptualizations of clusters have strongly emphasized the collective learning opportunities created via the interaction with other firms in the cluster as a major driver of cluster location. Such a location opens up possibilities for learning from competitors – via spillovers, through labor mobility, leakage of information and know-how or by imitation – and these are used by firms to develop and upgrade their own technological competencies (e.g., Scott, 1998; Rauch, 1993a, 1993b). Such benefits, however, are naturally uneven for firms with different technological capabilities. Firms possessing the best technologies are likely to benefit less from access to competitors’ technological knowledge, while firms with weaker technological capabilities will probably benefit more from such learning. This would enable weaker firms access to competitors’ sources of knowledge and learning which are superior to those they possess. Furthermore, firms possessing the most advanced technological capabilities may have incentives to locate away from clusters, to protect their core technologies from disseminating to other firms. This explains why, while there is a general tendency for innovative activities to cluster geographically (Audretsch & Feldman, 1996), the dominant firms in their industries sometime locate away from clusters, for example, Microsoft in Seattle and not in Silicon Valley; George Lucas near San Francisco and not in Hollywood (Star Wars, 1999). More systematic evidence for such location preferences, whereby technologically weaker firms exhibit greater tendency for cluster location than their competitors who possess stronger technological


229

capabilities, was shown by a number of studies (Rauch, 1993a; Suarez-Villa & Walrod, 1997; Gersbach & Schmutzler, 1999; Shaver & Flyer, 2000; Chung, 2001a; Chung & Alcacer, 2002). Formally: H3. The strength of the innovative capabilities of firms significantly affect their location choices, with the distance from the center of business activity positively related to innovative capabilities (i.e., the stronger the innovative capabilities of firms, the greater their tendency to locate away from the center), ceteris paribus. Experience in a Foreign Country We hypothesize that previous experience in a foreign country would affect MNEs location choices. Newly established operations are likely to confront higher information costs, a result of lack of knowledge of how to run business operations in an unfamiliar setting and of limited ability to forecast the economic events in a foreign country (Mariotti & Piscitello, 1995; Nachum & Keeble, 2001). A central cluster location is likely to reduce these costs as it tends to facilitate the interaction with other members of the cluster. As foreign affiliates become more established in the foreign environment, they may have less need for locating in proximity to other firms as a way to overcome their lack of local knowledge and unfamiliarity with the local environment. Therefore, we would expect that firms that had previously been operating in a given geographic area would exhibit lesser tendency for central cluster location. Formally: H4. Previous experience in a country significantly affects location choices, with the distance from the center of business activity positively related to previous experience (i.e., the greater a firm’s experience the more likely it is to locate away from the cluster), ceteris paribus. Nationality We expect that the national origin of firms, in particular the cultural and institutional distance between the home and the host countries concerned, would significantly affect the needs of firms for cluster participation and the benefits they may derive from it. However, the direction of this impact is hard to hypothesize a priori. There are both theoretical reasons and empirical evidence that suggest that the direction of the impact could go both ways. On the one hand, firms from culturally and institutionally remote countries have greater need to gain knowledge on the host country (Shenkar, 2001), and

230


greater difficulties in doing so, and they may use a more central cluster location to eliminate these difficulties. They may also have greater need for local complementary resources, as their own resources, or those that can be transferred internally within the MNEs, may be less adequate for the foreign environment. At the same time, however, greater cultural and institutional distance between the home and host countries may increase the difficulties of integrating in the cluster and taking active part in the local dynamics of collective learning and shared experiences that underlie the benefits of cluster location. The findings of studies designed to test for this link empirically are indeed inconclusive. In a survey of foreign firms investing in New York, Wymbs (2001) found that firms originating from countries that are culturally distant from the United States were more likely to locate in geographic proximity to an established knowledge center. Such location helps them to overcome difficulties in acquiring local knowledge and compensate for impediments associated with the transfer of tacit knowledge. In contrast, a number of studies have documented a tendency of Japanese firms to locate away from clusters of economic activity, and have attributed it to cultural differences between the United States and Japan that eliminate the ability of Japanese firms to successfully take part in local dynamics (Chung, 2001a; Friedman, Gerlowski, & Silberman, 1992). Formally: H5. The nationality of firms, as it relates to the cultural and institutional distance between the home and host countries, significantly affects location choices, but existing theory and empirical research provide inconclusive predictions for the direction of the impact. Differentiation Product differentiation is likely to influence the advantages associated with cluster participation, and hence affects the location choices of firms. The underlying advantages of cluster location are based on the assumption that firms share some resources and knowledge, and hence can benefit collectively from the availability of these in the immediate vicinity. Product differentiation enhances this logic, from both the demand and supply sides, implying that the more differentiated the firms’ products are, the greater the benefits of joining a cluster. From the supply side, greater differentiation enhances the ability of firms to appropriate the competitive benefits of their investments in innovation and new product development (De Bondt, Slaets, & Cassiman, 1992),


231

without it being mitigated by the dissemination to other cluster members. The leakage of their proprietary knowledge does not inhibit as much the attainment of competitive advantage as it would with undifferentiated outputs. They may also enjoy less fierce competition for localized resources, as they seek distinct functions in which they hold competitive advantage (Baum & Mezias, 1992; Baum & Haveman, 1997; Lomi, 1995) and the costs of factors of production would not necessarily rise, due to their specificity. The demand-based dynamics are such that with greater variation in the products being offered, consumers search more in order to find the best offer (Fischer & Harrington, 1996). Greater anticipated search makes going to a cluster more attractive, as a consumer can then search several firms at once. Greater product heterogeneity thus often means that firms have greater demand for their products. Greater differentiation also tends to reduce a main cost of joining a cluster, which is that proximity of firms intensifies price competition (Baum & Mezias, 1992). A number of empirical studies found support for these theoretical arguments. These studies show that greater differentiation is related to higher levels of geographic concentration (Swaminathan, 2001) and that inter-industry variation in the tendency of firms to agglomerate is directly associated with their level of product variety (Fischer & Harrington, 1996). Formally: H6. Differentiation significantly affects location choices, with the distance from the center of business activity negatively related to differentiation (i.e., the more differentiated MNEs are, the more likely they are to locate close to the center of activity), ceteris paribus. MNEs Organizational Structure The organizational structure of the MNEs, and the degree of control exercised by the parents over affiliates’ activities, are likely to affect the benefits of cluster location and hence the location choices of affiliates. The more an affiliate is integrated – in terms of its decision making and production – into the parent and the MNE as a whole, the less open it is likely to be toward its immediate environment, and the more limited will be its need for local cluster interaction (Nachum & Keeble, 2000). When affiliates enjoy more autonomy and implement the entire value-added function locally, they are likely to be more integrated in the locality that hosts them and to benefit more from cluster linkages (Birkinshaw & Hood, 2000). These theoretical arguments received strong empirical support by Nachum

232


and Keeble (2000). This line of argument suggests that there will be adverse relationships between the strength of control and internal MNE linkages and the need by foreign affiliates to locate in proximity to other firms. Formally: H7. The organizational structure of MNEs significantly affects location choices, with the distance from the center of business activity negatively related to the degree of control (i.e., the more centrally controlled MNEs are the more likely they are to locate away from the center), ceteris paribus.

THE DATA To control for industry effects, we limited the scope of the study to financial and professional service industries.3 Most previous studies of location choices have focused on manufacturing industries (e.g., Head, Ries, & Swenson, 1995; Shaver & Flyer, 2000; Kravis & Lipsey, 1982), and the validity of their findings for financial and professional service industries requires specific examination. The non-tradability of these service industries and the high tailor-made element in their production often implies that the location decisions of firms are affected by factors different from those affecting manufacturing firms. While the location of the latter is typically driven by the need for proximity to factors of production (Winsted & Patterson, 1998) using trade to access their customers, the location of the service firms studied here is determined by the location of their clients, in addition to the location of the factors of production (Pitt, Berthon, & Watson, 1999). In the context studied here, this difference implies that the agglomeration tendencies of these firms are driven by both demand and supply factors, while existing knowledge often only refers to supply-driven agglomeration (e.g., Shaver & Flyer, 2000). The few studies that directly address demand-driven agglomerations (Chung & Kalnins, 2001; Baum & Haveman, 1997) focus on service industries that differ considerably from those studied here, and as such the validity of their findings cannot be taken for granted. Determining the adequate industrial level of aggregation to capture the agglomeration effects arising from the interaction with other firms is a critical issue in a study of this kind (Moomaw, 1998). A narrow definition of an industry may underestimate the effect of the economies arising from the geographic concentration of related activities, and a broad definition may not be meaningful for analysis of agglomeration externalities, as it will group together firms that do not interact with each other. We have selected the 4-digits as the adequate level of aggregation. Several empirical studies of


233

agglomeration used this level, particularly with reference to MNEs (e.g., Head et al., 1995; Nachum, 2000; Shaver & Flyer, 2000) and found it to be a meaningful level of aggregation for the purpose of measurement of agglomeration processes. Geographically, the research was confined to MNEs who established operations in London and New York. These cities receive most of the total FDI flowing to their countries in financial and professional service industries (Sassen, 2001). So strong is the concentration of these activities in these cities that it has been argued that cities, rather than countries as a whole, are the real driving force of FDI in these activities (Sassen, 2001; Fainstein, Gordon, & Harlue, 1992). The selection of the geographic unit of analysis is critical, as different levels of spatial aggregation are associated with different processes of interaction between firms, and some processes only operate at lower levels of analysis (Lomi, 1995). By referring to such small geographic areas we are able to examine cluster processes at levels that were previously ignored by most studies, who used either whole countries (Wheeler & Mody, 1992; Braunerhjelm & Svensson, 1996) or states within them (Nachum, 2000; Shaver & Flyer, 2000). Such geographic levels may not capture fully the processes of interaction between firms that often occur at levels far smaller than a country, or even a state.4 Our selection of the specific boundaries of New York and London for the purpose of the study was guided by the search for boundaries which are meaningful in the sense that the firms residing within them can be judged to form a cluster, and do not take part in cluster interaction elsewhere. After experimenting with a number of possibilities, we adopted a definition that is broad enough to enable us to differentiate between firms at different geographic distances from the centers of activity, but at the same time, enables us to have a reasonable level of confidence that the firms residing within these boundaries have the potential for meaningful interaction with each other and do not belong to clusters elsewhere. When analyzing London, we refer to the area bounded by the Green Belt, known as ‘Greater London’, which covers 610 square miles. For New York City, we refer to the area known as ‘the core region’ or ‘New York region’, which covers about 600 square miles (Fainstein et al., 1992). By confining the study to firms residing within these boundaries, we regard London and New York as selfcontained world, whose business relations can be studied in isolation from the surrounding environment (Shefter, 1993). Within these boundaries we further divide the cities into districts, based on postal code areas (46 and 44 postal code areas for New York City and

234


Greater London, respectively). Previous studies found evidence that small districts within cities are meaningful spatial units for the analysis of the agglomeration benefits of firms in cities (Nachum & Keeble, 2000; Wymbs, 2001). The scope of the study was confined to entry via M&As that involved a complete transfer of ownership. The location decisions of firms entering via M&As naturally differ from those of firms entering via greenfield, since the latter require explicit location decisions, while the location of an acquisition is predetermined by the location of the acquired firms. For this reason, this entry mode is often excluded from analyses of location determinants of MNEs (e.g., Shaver & Flyer, 2000). Such a focus on greenfield investment, however, applies only to small shares of FDI. M&As have become the dominant mode of cross border expansion, accounting in 2000 for about 85% of total world FDI. In the service industries studied here it reached nearly 100% (UNCTAD, 2001). With such a proportion of FDI undertaken via M&As, analyses based on greenfield investment, although having obvious theoretical appeal, are only valid for a small share of FDI. There is an important need to examine the location determinants of investment via M&As, and to see the extent to which it conforms to the predictions of existing theories, which are typically based on greenfield investment. A number of studies have stressed the different nature of FDI associated with these two entry modes (e.g., Mata & Portugal, 2000) and have illustrated that the validity of knowledge generated based on one of them to the other is often limited (Blonigen, 1997; Ray, 1998). Furthermore, we argue that the choice of a particular target for acquisition is influenced, among other things, by its location. Hence the location of MNEs entering via M&As involves specific and active location decisions, and cannot be taken as if it is fully determined by the location of the target firms. This argument is consistent with Baum and Usher (2000) who found that acquiring firms seek to gain access to local information about demand, behavior of local competitors and feasibility of operation at particular locations, and these have strong influence on the selection of targets for acquisition. The focus on M&As that involve only a complete transfer of ownership enables us to control for differences associated with levels of ownership. It has been shown that the levels of ownership of MNEs over their foreign operations is associated with considerable differences in terms of affiliates behavior, including among other things their local embeddedness and the intensity of the linkages between them and their parents (Barbosa & Louri, 2002).


235

We take a long time span for the study, covering all M&As undertaken during the last two decades. Such a time span frees the findings from a possible bias that might have been introduced by focusing on a particular point in time. Data for the study were obtained from the Worldwide Merger & Acquisitions database of Thomson Financial Securities Data. This database is the largest of its kind, providing detailed information on more than 273,000 M&As worldwide over the last two decades. From this source, we obtained the exhaustive list of all foreign M&As in financial and professional services that involved a complete ownership transfer that took place in New York and London during the last two decades (about 2000 M&As), as well as information on the characteristics of the parties involved in the M&As. This source was supplemented by data from Disclosure Global Access, Hydra and Compustat that contain financial and various other firm-level data on corporations around the world. After excluding cases with large numbers of missing observations, we were left with 673 MNEs, of which 235 and 438 entered New York and London, respectively.5 t-tests show no significant differences between the omitted and included observations in terms of a number of economic and strategic attributes, enabling us to regard our sample as a reasonable representation of the entire population. About 40% of the firms in the sample are commercial and investment banks, 10% insurance and 50% professional services.

THE MODEL We start by estimating a model whereby the location decisions of MNEs are determined entirely by location advantages, which is essentially the traditional model of location decision. We use this model as a yardstick for comparison with the models we develop below. Formally: Y i ¼ f ðb0 þ b1 xj Þ þ i

(1)

Where Yi is the geographic distance of firm i from the center of business activity in its industry (i ¼ 1, y, m); xj signifies the cluster advantages of industry j ( j ¼ 1, y, n); ei is a firm-specific random disturbance that is attributable to errors associated with inadequate assessment by firms of the value of cluster interaction for them and possibly also distortions resulting from inertia of location patterns and the limited tendency by firms to

236


re-locate. It also takes account of unobservable firm characteristics that affect the distance of firms from the center of the cluster. We assume that the errors follow the standard normal distribution. Distance from the center of business activity is undertaken as an indication of the advantages that firms derive from cluster participation. The use of geographic distance as an indication of the strength of firms’ linkages with other firms has many precedents. It has been widely used in network literature (e.g., Blau, 1977) as well as in organization literature (e.g., Baum & Mezias, 1992) and more recently also in international business literature (Chung, 2001a). The center of business activity is defined for our purposes here as the city district that hosts the largest number of affiliates in an industry. The geographic distance from the center of business activity is measured by the distance of affiliate i from the city district that hosts the largest number of affiliates in affiliate i’s major industry. The number of affiliates was found to correspond closely with volume of business activity, making the number of affiliates an adequate indication of the distribution of activity. To construct this measure, we convert the location of each firm into longitudes and latitudes. We then calculate Euclidean distance between the location of each firm and the center of activity in its industry. This measure has the advantage that it treats the clustering tendency as a continuous variable, and thus overcomes the theoretical difficulties associated with the need to pre-define clusters, and to determine what level of geographic concentration should be regarded as a reasonable approximation for a ‘cluster’, a decision that without doubt involves a strong arbitrary element (Martin & Sunley, 2002). This avoids the limitations of selecting inappropriate boundaries, particularly as such boundaries are often determined according to administrative and political scales. To this traditional location model, which is based on location advantages alone, we add a set of firm-specific characteristics. We use this combination to test the basic argument underlying this study, namely that firm-specific characteristics together with location advantages would explain the location choices of MNEs. Formally: Y i ¼ f ðb0 þ b1 xj þ b2 xi Þ þ i

(2)

where xi 2 is a vector of firm-specific characteristics (i ¼ 1, y, m). Kolmogorov–Smirnov test, with a Lilliefors significance test for normality, shows that the dependent variable is not normally distributed and hence we take the natural logarithm. Since the dependent variable gets the value


237

zero for some cases, we construct it as (Weisberg, 1985): (20 )

logðY þ 1Þi ¼ f ðb0 þ b1 xj þ b2 xi Þ þ i

To this basic formulation we add a vector of control variables l, comprised of various characteristics that are expected to influence the location choices of firms, regardless of the firm-specific attributes hypothesized to affect them: logðY þ 1Þi ¼ f ðb0 þ b1 xj þ b2 xi þ lÞ þ i

(200 )

The theoretical discussion above suggests that the factors affecting location choices may also exist in interaction with the specific location advantages concerned, that is, the firm-specific attributes affect location choices in interaction with the cluster advantages. Formally: logðY þ 1Þi ¼ f ðb0 þ b1 xj þ b2 xi þ b3 ðxj n xi Þ þ lÞ þ i

(3)

By introducing these interaction variables, we seek to model explicitly the different value of identical location advantages for different firms, that is, the variation in the effect of the firm-specific characteristics on location choices as depending on the level of cluster advantages (Friedrich, 1982). Models 2 and 3 describe the relationships between the independent variables in different terms – while in model 2 the relationships are additive, in model 3 they are assumed to be both additional and conditional. That is, in model 2 we assume the effect of the firm characteristics on location choices to be constant, regardless of the level of the cluster advantages, while model 3 describes the effect of each firm-specific attribute on location choices as varying according to the level of cluster advantages.

OPERATION OF THE INDEPENDENT VARIABLES Cluster Advantages Geographic concentration of firms is taken as an indication of the existence of agglomeration economies. Measuring agglomeration by the extent of concentration of activity in a particular geographic area is well established in the literature (e.g., Shaver & Flyer, 2000; Nachum, 2000; Chung & Kalnins, 2001).

238


We apply a variation of a widely used measure of geographic concentration at the level of countries (e.g., Aw & Batra, 1998), to cities, as follows: X GDIj ¼ 1 S 2 gj where S is the share of firms based in district g in the total number of firms in industry j. GDI thus measures the overall level of concentration in an industry. It can get any value between 0 and 1. The higher the values of GDI the more dispersed is the industry. The GDI has the advantage that it takes account of both the number of geographic areas (districts in our case) in which an industry locates and the magnitude of activity in each of them. We use the number of affiliates to calculate the GDI, rather than alternatives like sales or employment, because this is consistent with data used in previous studies (e.g., Shaver & Flyer, 2000; Head et al., 1995), and this enhances the comparability of our findings with those studies. Firm-Specific Characteristics H1: Size. Size is typically operationalized either in terms of sales or employment. The correlation between the two measures in our dataset is very high (0.92, po0.05), and since the employment data is more complete we select employment. H2: Geographic scope. The share of activity in international markets (typically measured by sales) is commonly used to measure the geographic scope of MNEs operations (e.g., Dunning, 1993). Sales data that distinguish between foreign and domestic markets were only available for a small number of observations; therefore, we used the ratio of foreign acquisitions to total acquisitions undertaken over the last 20 years. The correlation between this measure and the sales ratio for the observations for which this data is available is 0.89 ( po0.01), enabling us to accept it as an adequate measure. H3: Innovative capabilities. The share of R&D investment to sales, which is the commonly used measure of the innovative capabilities of manufacturing firms, may not be a meaningful measure of the innovative capabilities of service firms. Unlike in manufacturing, innovation in services is not necessarily technological and can be, and often is, realized without any direct investment in R&D. R&D activities in service firms usually involve large amounts of ad hoc ‘innovations’ that are difficult to measure since no specific resources are assigned to these activities (Licht & Moch, 1999; Djellal & Gallouj, 1999). The limitations of the manufacturing-based measures


239

appear to be particularly severe in the service industries studied here, where the active participation of the clients in the production often implies that some of the inputs into the innovation process are incurred by the clients. Attempts to propose more adequate measures of innovation in service industries – often based on the judgment of service firms themselves – suggest that such measures should reflect changes in the inputs, methods or outputs which improve the commercial position of the firm. In line with these arguments, we operationalized innovation by two measures: (1) productivity, that is annual growth of turnover per employee. This operation reflects the realization of the need to take account of process and organizational innovation resulting in increased productivity of the resources employed in production (Licht & Moch, 1999; Djellal & Gallouj, 1999); (2) profitability (measured as share of revenues to adjust for size). Profitability is often used as an overall proxy for the possession of intangible advantages by MNEs and was shown to be a reasonable proxy for the possession of intangible assets (e.g., Shaver & Flyer, 2000), notably innovative capabilities. FDI theory states that firms possessing a substantial stock of intangible advantages are more likely to succeed in FDI. H4: Experience in a foreign country is operationalized by two measures. (1) Number of acquisitions made previously in a market. This is consistent with studies examining acquisition experience, in which the number of previous acquisitions is used as an indicator of accumulated experience (Ingram & Baum, 1997; Haleblian & Finkelstein, 1999). (2) Length of activity measured by the number of years since acquirers have made their initial entry into a market. We argue that these two operations together are needed to capture experience, since some firms may have smaller magnitude of activity lasting over long time span and vice versa. H5: Cultural distance between the home country of the acquirer and the host country. Following Kogut and Singh (1988), we calculated a composite index, based on Hofstede’s indices (Hofstede, 2001). The index is based on the deviation along each of the four cultural dimensions (power distance, uncertainty avoidance, masculinity and individualism) of each country from the host country (the US for New York FDI; the UK for London FDI). We correct for differences in the variances of each dimension and then calculate the arithmetic average between the four dimensions. H6: Differentiation. The common operational measure of differentiation is advertising expenditure, which captures two important aspects of product

240


differentiation: product branding and marketing skills (Dunning, 1993). Since such data are not available for us, we follow Owen (1982), and use an alternative marketing proxy – selling, general and administrative expenses (expressed as % of total costs to adjust for size). This is a broad measure, which captures a set of factors not directly related to promotion activities, but it provides indications of the strength of the marketing efforts undertaken by MNEs. In an analysis of inter-industry determinants of FDI, Owen (1982) found that this measure captures adequately the differentiation efforts undertaken by MNEs. H7: MNE organizational structure. This construct is operationalized by the industrial specialization of the affiliates (targets) vis-a`-vis the parent (acquirer). The assumption here is that the more remote are the affiliates from the parents in terms of their industrial affiliation, the less control would be exercised by the parents and the more limited will be the mutual transfer of resources. This variable was coded as a dummy variable that gets the value 1 if the affiliates and the parents share the same industrial affiliation at the 4-digit level; 2 for 3-digit level; 3 for 2-digit level; 4 for 1-digit level and 5 when the affiliates and the parents are engaged in completely different industries. This is consistent with a number of studies (e.g., Haleblian & Finkelstein, 1999), in which the industrial affiliation of targets and acquirers (at the 4-digit SIC level) is used to measure the relatedness between them. It is also consistent with Ray’s (1998) finding of a significant and positive link between the value of investment and the industrial association between the parents and affiliates. When the parent invests in product lines similar to its own, it transfers more of its capabilities to the affiliates and it is thus able to minimize the disadvantages that the affiliate faces as a result of operation in foreign market. Control Variables We add several control variables. First, we seek to control for the differences in the nationality of ownership of the acquired firm, particularly whether domestic or foreign. Since foreign and domestic firms often exhibit different location patterns (Shaver, 1998; Nachum & Keeble, 2001), it is important to control for differences in location that were pre-determined by the nationality of the acquired firm. This variable gets the value 1 if the firm acquired is foreign-owned, 0 otherwise. Second, we introduce control variables for industry affiliation to account for the possibility of unobserved differences arising due to industrial


241

influences that may affect both independent and dependent variables (as a common cause). To eliminate any spurious effects, we add fixed industry effects by entering industry dummy variables (at the 2-digit level). Third, we also measure the level of foreign activity in an industry, as research has shown that foreign firms often seek the proximity of other foreign firms (Florida & Kenney, 2000; Head et al., 1995; Nachum, 2000), and this may influence their location choices regardless of their specific attributes or location advantages. We measure variation in foreign activity by the industry’s share of foreign acquisitions in the total over the last two decades. Fourth, we control for the possibility that district differences (in terms of, e.g., property availability and prices, transport access) affect the choice of location, and hence the distance from the center of the cluster. We introduce district fixed-effect specification as dummy variables. This approach does not require identification and measurement of district attributes that may affect location choices in order to control for their effect. This overcomes the problem that the district characteristics that can attract (or deter) investment are manifold and often difficult to assess and measure. The fixed-effect approach is consistent with Head et al. (1995), Lomi (1995) and Shaver and Flyer (2000). Finally, we introduce a dummy variable to distinguish between London and New York to enable us to capture possibly significant differences between these two cities. Data refer to the characteristics of the acquirers, the MNEs entering foreign markets undertaking the location decisions in which we are interested. Table 1 summarizes the explanatory variables included in the model, their operational measures, descriptive statistics and correlation coefficients. Most coefficients are low (well under 0.5), enabling one not to be concerned about correlation. Tests of normality suggest that a number of independent variables are not normally distributed and hence we took the natural logarithm transformation in the analysis. Independent sample t-tests show that the missing value patterns are not random, and they were estimated from available observations, by testing a model based on all observations for which there were no missing values, and using it to estimate the missing values.

MODEL RESULTS AND DISCUSSION The models constructed above were estimated by a moderated multiple linear regression, a commonly used method to test hypotheses that involve interaction effects (Aiken & West, 1991) (Table 2). Overall, the analyses

242

Table 1.

The Explanatory Variables Included in the Model – Descriptive Statistics and Correlation Coefficients.

Constructs

Firm-specific characteristics H1: Size H2: Geographic scope

H3: Innovative capabilities

H4: Experience

GDI index ()

No. of employees (+) Share foreign acquisitions (+) Marginal profitability (+) Productivity (+) Years since establishment (+) No. of previous acquisitions (+)

Descriptive Statistics Means (S.D.)

.263 (.225)

Pearson Correlation Coefficients GDI

Size

Geog scope

Innov- Innovprofits product

Years Experience Nationality differnti

1.000

5,076 (8,824)

.059 (.505)

.758 (.132)

.017 (.692)

.216 .092 (1.812) (.416) 681,192 .033 (6,808,499) (.578) 6.872 (4.918)

.038 (.359)

14.282 (18.801)

.009 (.829)

1.000

.096 1.000 (.238) .052 .039 (.712) (.589) .069 .074 (.520) (.172) .032 (.694)

1.000 .037 (.705)

.105 .041 (.006) (.850)

.084 .068 (.313) (.090)

.088 (.231)

1.000

.048 1.000 (.380) .024 .020 (.667) (.618)

1.000

Af-par- nationality Foreign industry industry

City


Location advantages Cluster advantages

Operations (Causality)

Cultural distance index (?) H6: Differentiation Selling costs (shares total costs) () H7: Organizational structure Affiliate/parent industry () Control variablesa Nationality target Foreign activity

City

Dummy (1-foreign) Share foreign acquisitions (industry) Dummy (1-London)

14.090

.022

.065

.028

.104

(.609) .026

(.444) (.481) .071 .029

(3.216) 3.221

(.652) .019

(.498) .088

(.588) .089

(.862) .105

(.929) (.298) .022 .002

(1.577)

(.732)

(.318)

(.035)

(.172)

(.713)

.036 (.387) .757 .160 (3.922E02) (.000) b

.164 (.000)

(.527) .007

.047

(19.385) .790

b

(.153) .016

.035

(.244) .056

(.964)

.109

.085

1.000

(.959) .066

.083

(.188)

(.130)

(.160)

.285 (.000) .009 (.840)

.048 (.229) .070 (.103)

.051 .046 (.345) (.238) .006 .019 (.925) (.648)

.100 (.218)

.037 .117 .093 (.493) (.002) (.020)

.068 (.689)

.003

(.128) .058

.058 .054 .152 (.471) (.163) (.033) .074 .285 .047 (.397) (.000) (.539) .042 (.273)

1.000

(.008)

.066 (.099)

1.000

.096 .047 1.000 (.077) (.268) .068 .128 .006 (.244) (.002) (.878) .076 .062 (.160) (.139)

.177 (.000)

1.000

.008 (.854)

1.000

Correlation is significant at the 0.05 level (2-tailed). Correlation is significant at the 0.01 level (2-tailed). a

Statistics of two additional control variables are not reported: industry and district dummies (due to the large number of dummies in each of these categories). b Dummy variables. Means and S.D. not meaningful.


H5: Nationality

243

244

Table 2.

A Model Linking Distance with Cluster Advantages and Firm-Specific Characteristics: Whole Sample.

Constructs

Operations

Multiple Regression

Location advantages Cluster advantages Firm-specific characteristics Size Geographic scope Innovative capabilities Experience Nationality Differentiation Organizational structure

GDI index

2.126 (5.789)

No. of employees Share foreign acquisitions Marginal profitability Productivity Years since establishment No. of previous acquisitions Cultural distance index Sale costs Affiliate/parent industry

Interaction variable (cluster advantages firm-specific characteristics) Size No. of employees Geographic scope Share foreign acquisitions Innovative capabilities Marginal profitability

Model 2 1.725 (2.802) 3.370 1.065 1.054 6.940 9.224 1.066 3.255 1.919 6.638

(.376) (5.069) (.865) (1.186) (3.213) (.506) (2.234) (1.561) (2.925)

Model 3 1.022 (1.775)w 3.370 4.282 4.557 4.423 4.581 3.245 1.260 5.533 .476

(.376) (2.572) (.354) (1.063) (.082) (1.308) (.510) (.893) (1.713)w

.177 (4.511) 1.327 (4.197) 4.557 (.354)


Model 1

Control variablesa Nationality target City

Dummy (1=foreign) Dummy (1=London)

Regression statistics Adj. R2 SE F stat. Sig. F N

3.906 .168 6.504 1.089 5.533 .203 .310 (1.897)w .478 (3.378) .377 .906 177.520 .000

3.804 (.091) .217 (3.761) .743 .691 64.827 .000

(1.063) (2.807) (1.198) (2.329) (.893) (2.275)

0.092 (2.733) .119 (2.007) .779 .741 88.455 .000

673

a The control variable ‘foreign activity’ at the industry level was excluded from the analyses due to high VIF values (we adopt the suggested cut-off point of 10 (Studenmund, 1992)). Statistics of the industry and district dummies are not presented (see text for discussion). po.001. po.01. po.05. w po.10.


Nationality Differentiation Organizational structure

Productivity Years since establishment No. of previous acquisitions Cultural distance index Sale costs Affiliate/parent industry

Experience

245

246


provide strong support to the basic assumption that underlay this study, namely that firm-specific characteristics affect location choices and the latter cannot be fully understood without taking explicit account of these attributes. The overall explanatory power of the model increases significantly with the addition of the firm-specific attributes while the cluster advantage variable loses much of its significance (model 2 in comparison with model 1). The findings show that most of the variation in location choices is explained by the firm-specific attributes. After taking account of these attributes, cluster advantages on their own possess only little explanatory power. Model 3, which is based on the firm-specific attributes in interaction with the cluster advantages, yielded the best fit, providing support to our argument that the same location advantages have different values for different firms. A few caveats might be in order when interpreting the results of the study. For one, the firm-specific characteristics that comprise the independent variables in the model cannot be said to capture all possible attributes that can potentially affect the location choices of firms. However, by including the major factors recognized in the international business and management literature to affect the strategic behavior of MNEs in foreign countries, combined with those identified in cluster theories to affect the behavior of firms in clusters, we expect the problem to be minimized. A second reservation to be made is that the expected link between location and firm attributes may sometimes be biased by the tendency for inertia in location decisions, whereby the initial location choices of firms determine their subsequent location (Hay, 1976). This tendency may weaken some of the associations between location and firm characteristics hypothesized here. With these reservations in mind we turn to the discussion of the findings. H1 is only supported by model 3, that is, size of firms is a significant determinant of cluster location only when taken in interaction with the cluster advantages. It possesses no explanatory power on its own (model 2). This is an interesting finding by itself, as it is the only variable that shows such a difference between the two models. It implies that the effect of size on location choices varies, depending on the level of cluster advantages, that is, size has a statistically significant effect on location choices only when conditional on the cluster advantages. Strong support is found for H2, which is the most significant variable in both analyses. That is, when firms’ operations are geared toward markets external to the cluster, they gain less from central cluster location and are more likely to locate away from clusters. This highlights the importance of demand forces on the location decisions of MNEs operating in the industries studied here. Support for this was found in a survey of foreign


247

MNEs in New York, where access to customers was cited by the majority of the respondents as the main reason for their investment in New York (Grein, 1996). The failure to support H3 implies that innovative capabilities do not affect the location choices of the firms studied here. The rationale for hypothesizing a link between innovative capabilities and location rested upon the risk of dissipation of knowledge to competitors, but it might be that this does not apply to financial and professional service industries. The idiosyncratic nature of the knowledge of these firms may eliminate the need for remote location as a way to protect their proprietary knowledge from disseminating to competitors. At the same time, it may also imply that there is less for them to learn from competitors. This finding signifies an important departure from previous studies that found close association between innovative capabilities and location choices, and appears to highlight the merits of analyzing narrowly defined industries in analyses of this type. H4 (experience) is supported by only one of the operation measures – years since establishment. The second operation – number of previous acquisitions – is not significant. The tendency for inertia in location decisions discussed above might be particularly important with reference to this attribute of firms, which by its very nature is changing over time, and may explain the weak support for this hypothesis. Another possible reason is that much of the foreign activity in the US and the UK is well established, with its origin going back decades and even centuries. The rationale for a link between previous acquisitions and location is more applicable to new investment and is likely to weaken or even disappear over time (Vermeulen & Barkema, 2001). The nationality variable (measured by the cultural distance index) is highly significant, in support of H5. The negative sign of this variable implies that the larger the cultural distance between the home and host countries, the greater would be the tendency for central location. MNEs thus use central cluster location to ease the difficulties arising from cultural distance, such as acquisition of local market information and access to the local cluster of knowledge. H6 (differentiation) receives only weak support – it has weak and negative explanatory power in model 2 and is not significant in model 3. These findings may suggest that there are no strictly linear relationships between differentiation and the tendency of firms to cluster; rather there are counteracting forces. While both theory and empirical evidence provide strong basis to hypothesize negative relationship between the extent of differentiation and distance from the cluster, opposite forces may also be in

248


place. Greater differentiation might be associated with lesser benefits of cluster location, as the specific needs of highly differentiated firms imply that there are fewer benefits of the complementary resources abundant in the cluster. High differentiation is also likely to eliminate the gains in received knowledge and spillovers from other firms. Furthermore, the argument for a significant link between differentiation and location choices lies in part on reduction of search costs by consumers and on diminishing price competition. Both mechanisms are less applicable to financial and professional service industries, where the customers are for the most part other businesses, whose demand tends to be price inelastic. The significant, negative sign of H7 supports the hypothesized relationships between organizational structure and level of control and cluster location. This aspect of international operation was somewhat neglected by previous research on location of MNEs, and our findings suggest that it should not be overlooked by analyses of this kind. These findings should be interpreted in light of the nature of organization of international activity in financial and professional service industries. Investment in these industries is typically horizontal, whereby the affiliates are small replicas of the parent, implementing the entire value-added chain locally. Organizations of this type typically exhibit a tendency for limited control and great autonomy for affiliates. If organizational structure is highly significant in this industrial context, it can certainly be argued that it is an important determinant of location choices. The highly significant sign of the ‘target city’ dummy variable implies that there are significant differences between London and New York. To analyze these in greater details we estimate the models for sub-samples of New York and London MNEs separately (Table 3). In both of the sub-sample models, geographic scope (firm-specific characteristics) and size and geographic scope (interaction effects) were significant and had the same sign as the combined model, lending a great deal of credibility to the importance of these firm-specific variables. For the London model, organizational structure and cultural distance were significant and had the same sign as in the combined model. A number of the explanatory variables have changed their signs and significance in these analyses, implying that in addition to the general characteristics identified above, there are also context-specific characteristics that modify the value of individual attributes in affecting the location choices of MNEs. Thus, while not significant in the analysis of the whole sample and in London, innovative capabilities, proxy by profitability is highly significant in the analysis of New York.


249

Some other variables lose their significance in the individual cities analyses, probably implying greater homogeneity within the sub-samples. Thus, H4 (experience) and H7 (organizational structure) are strongly supported in the analyses of the whole sample (Table 2) but are not significant in the analyses of both New York and London. Concerning H4, these findings maybe in line with the well-recognized tendency of FDI to flow in tandem to certain locations, illuminating variation in this respect across firms based in same location. The analyses reported in Table 3 provide additional support to the basic argument underlying this study, namely that location advantages on their own provide only partial explanation for the location choices of MNEs. Also in these analyses, the cluster advantages measured explains a smaller portion of the variation across firms and loses much of its significance with the addition of the firm-specific attributes. And also here, the model based on the interaction between the firm-specific and cluster advantages as independent variables yields the best fit, further supporting our argument that location advantages are relative whose value changes in line with specific attributes of firms. Some of the industry dummies were highly significant in most analyses, implying that in addition to firm characteristics, industry characteristics also affect the benefits of agglomeration. This is consistent with the argument that agglomeration benefits vary across industries (Audretsch & Feldman, 1996). None of the district dummies was significant in any of the analyses, perhaps suggesting that after taking account of cluster advantages, there are no influences of district characteristics on location choices.

Validation Tests We employ several analyses to validate the statistical significance of the findings. First, we check for a possible impact of dependence among observations. Some of the firms studied have made a number of acquisitions in New York and/or London, so there are multiple observations for them, which are probably not totally independent of each other. Although from a purely statistical point of view, such observations should be removed, conducting the analysis under the assumption that firms only invest once would significantly weaken its explanatory and predictive power. Hence, we leave these dependent observations and test for a potential bias that might be introduced by their inclusion. We examined the variance of the residuals of the observations with duplicate entries and those with a single entry and

250

Table 3. Constructs

A Model Linking Distance with Cluster Advantages and Firm-Specific Characteristics: London and New York.

Operations

Model 1 London

GDI index

NY

London

1.561 (5.991) 1.504 (6.759) 1.237 (3.761)

Firm-specific characteristics Size No. of employees Geographic scope Share foreign acquisitions Innovative capabilities Marginal profitability Productivity Experience Years since establishment No. of previous acquisitions Nationality Cultural distance index Differentiation Sale costs Organizational Affiliate/parent structure industry Interaction variable (cluster advantages firm-specific characteristics) Size No. of employees Geographic scope Share foreign acquisitions

Model 3 NY

London

NY

1.002 (2.986)

0.983 (2.067)

.863 (1.922)

.715 (.106) 1.362 (8.393)

6.110 (.402) 1.485 (1.273) 3.376 (.079) .661 (3.048) 4.646 (2.618) .516 (2.311)

4.341 (1.273) .690 (1.179) 1.939 (.495)

5.596 (1.816)w 3.016 (2.747) 2.374 (.042)

5.015 (1.348)

2.142 (.948)

5.704 (3.264) 1.604 (.534) 1.071 (.868) .117 (.871) 5.048 (.879) 5.280 (.169)

1.581 (.977) 6.789 (1.384) .104 (1.093) 6.778 (1.572)

.265 (.740) 5.770 (3.416) 6.123 (.567) 2.505 (.818)

4.674 (1.007) 4.522 (1.199) 3.378 (1.250) 2.546 (.157) 3.462 (.942) 1.345 (3.110)

.153 (3.183) .282 (3.021) 1.594 (4.081) 1.356 (1.894)


Location advantages Cluster advantages

Model 2

Nationality Differentiation Organizational structure

Marginal profitability Productivity Years since establishment No. of previous acquisitions Cultural distance index Sale costs Affiliate/parent industry

Control variablesa Nationality target

Dummy (1=foreign)

Experience

Regression statistics Adj. R2 SE F stat. Sig. F N

3.169 (.023) .434 (1.121) .140 (1.125)

.841 (2.388) 0.0494 (.687) .190 (.869)

1.059 (1.399) 1.339 (.156) 1.234 (2.187) .794 (.602) .578 (.877) .104 (.197) .190 (1.064) .281 (1.159)

.475 (5.521) .655 (5.894) .237 (.258)

.382 .868 75.527 .000

.359 .800 84.658 .000 438

.638 .717 43.914 .000 235

.124 (1.344)

.669 .626 27.343 .000 438

.932 (2.231)

.779 .652 38.568 .000 235

.188 (2.038)

.733 .659 23.511 .000 438

235

a The control variable ‘foreign activity’ at the industry level was excluded from the analyses due to high VIF (VIFW10 in all analyses). Statistics of the industry and district dummies are not presented (see text for discussion). po.001. po.01. po.05. w po.10.


Innovative capabilities

251

252


found no significant differences between them. This shows no discernable pattern for the duplicates. Second, we test whether our results are robust for narrower industrial classifications. Many of the control variables for industrial affiliation were significant in the previous analyses, implying an influence of industry affiliation, in addition to the firm attributes studied. We estimated the model separately for financial and professional service MNEs. The results of these two sub-samples are generally consistent with the results of the full sample but overall, the model of financial services is slightly less significant than the one of professional services, suggesting weaker link between location and firm-specific attributes for the former. A number of differences between the industries might explain this finding. First, in financial services there are greater possibilities to benefit from scale economies by the standardization of the production. Hence, there are greater internal advantages and lesser needs to take part in the external economies of cluster that may drive firms to locate in proximity to other firms. Second, unlike professional services that sell almost exclusively to other firms, financial services have also a large component of activity geared toward small private consumers. Geographic spread provides a considerable advantage for the latter type of activity and may eliminate some of the advantages associated with geographic concentration. Third, and somewhat related to the other two points, technological advances have, to a certain extent, eliminated the importance of location for some type of financial services, but to a lesser extent have affected professional services. Third, we test whether our findings hold if we take explicit account of a possible effect of external macroeconomic conditions on the relations we found. The study covers a period of about two decades, during which there naturally have been considerable changes in the macroeconomic environment facing foreign firms entering New York and London. It might be argued that these have affected the location choices of firms, regardless of their firm-specific attributes. For example, in fast growing economies firms may have a greater need for central cluster location, as a way to acquire knowledge on the rapidly changing conditions in their markets. Likewise, cyclical changes in property prices may affect the balance between the costs and benefits associated with central cluster location. To test for this, we add to the model dummy variables for each of the years analyzed (minus one). None of these dummy variables is significant. The inclusion of the dummy variables slightly diminishes the overall significance of the model (probably due to less degrees of freedom), but the findings are similar in sign and magnitude to those reported based on the estimation of the model without the year dummy variables.


253

CONCLUDING REMARKS In this paper, we have sought to extend the theory of the location decisions of MNEs, by incorporating the firm-specific attributes of these firms as factors affecting the value of specific location advantages to them and hence their location choices. By explicitly acknowledging the heterogeneity among firms in terms of the value of location advantages to them, we implied that location advantages are likely to vary for individual firms. We were further able to identify a set of specific characteristics that affect this variation and to illustrate their combined impact on the location choices of firms. Our approach signifies a departure from traditional conceptualizations of MNE location choices (Dunning, 1993), whereby the location choices of MNEs were attributed solely to the advantages of various locations. The findings suggest that rather than being absolutes whose value is identical for all firms within an industry, location advantages vary across firms, in line with certain firm-specific attributes. They demonstrate the promise of models of location determinants that take explicit account of firm-specific attributes for the development of a fuller understanding of the determinants of location choices that is sensitive to differences among individual MNEs. We addressed these firm idiosyncratic issues with reference to global cities a geographic level of analysis that has received limited attention within FDI research. In this, we contribute to the recent stream of research that has called for examination of the activities of MNEs at disaggregated geographic levels of analyses, such as regions (Rugman, 2005; Rugman & Verbeke, 2004). This approach perhaps suggests a need to adopt multiple geographic levels of analyses in international business research, with the appropriate one probably varies across industries and activities (e.g., computer chip design across whole regions, while security trading in one part of a city). The search for the suitable level of analysis might by itself be a challenge. The importance of this endeavor lies in that different geographic levels of analyses are associated with different location attributes and processes. The isolation of global cities from the national context and their linkages with other global cities (Sassen, 2001) is perhaps another distinctive attribute of global cities that is likely to bear implications for location choices. Our findings indeed illustrate the merit of this approach in that we were able to identify location attributes that are perhaps particularly noticeable in the context of cities and may have gone unnoticed in different geographic settings. The study makes several additional contributions, related to the data used and the specific setting in which the hypotheses were tested. First, by

254


focusing on M&As and identifying the location determinants of this, increasingly common, entry mode of FDI, the study has made an important contribution to the literature on location determinants, which is based almost entirely on greenfield investment or else does not control for entry mode. Given the dominance of M&As as an entry mode, this contribution is of considerable value. Second, by studying FDI to global cities, a geographic area that has received limited attention by previous research, we have made a contribution to closing a large gap in knowledge on the specific attributes of FDI concentrating in these geographic areas. Global cities often host the focus of decision-making within MNEs, and thus may represent a distinctive location (Semple & Phipps, 1982; Ward, 1994), which requires specific research attention. Finally, by studying the location decisions of service MNEs, which account for large and rapidly growing shares of FDI, we have contributed to the knowledge on the location of MNEs in these industries. Our findings support the need to pay more attention to these specific contexts. We have attributed the inconsistency of some of our findings with those of previous research to the specific nature of foreign activity in global cities and to certain characteristics of the service industries studied here. There is a need for further research in order to improve our ability to specify the conditions under which different types of associations between firm-specific attributes and location advantages affect the location choices of MNEs. Similar analysis is needed for different location advantages. The focus here was on advantages associated with cluster location, but there is a need to establish the effect of firm-specific characteristics on their evaluation of other location attributes, such as size and growth of the market, labor quality, costs of factors of production, etc. Further analysis is also needed for additional firm characteristics, notably those related to the underlying motivation for undertaking foreign investment and the strategic objectives of MNEs joining a cluster. With the sources of data used here these important characteristics could not be tested, but theory suggests that they are likely to exercise significant impact on the location decisions of MNEs.

NOTES 1. Some references to the influence of firm-specific attributes, along with industrial and country characteristics, on the Ownership, Location and Internalization (OLI) configuration was made by Dunning (e.g., Dunning, 1993, notably Table 4.3), who argued that the propensity of firms to engage in FDI will vary according to certain characteristics of the MNE concerned. However, explicit reference to firm-specific


255

attributes as affecting the location choices of MNEs has not been introduced systematically into the theoretical formulations of the OLI. 2. Agglomeration externalities arise from supply and/or demand reasons. From the supply side, such externalities arise from the concentration in relatively dense geographic area of firms and specialized resources, such as skilled labor and infrastructure, and it tends to be associated with lower unit costs of production than would be the case if typical businesses were located elsewhere. Demand driven agglomeration is driven by the location of the customers and arises from advantages such as shared customer bases of firms locating close to each other, informational externalities regarding the extent of demand, and reduction in consumer search costs, which is beneficial for total market demand. 3. Financial and professional services include banking, insurance, advertising, software and information services, accounting, management and engineering consulting, corresponding to the 60, 62, 63, 64, 67, 73 and 87 SIC codes. 4. Economists and economic geographers have been aware of the limitations of wide geographic areas for the study of agglomeration economies and have formulated their theoretical frameworks with reference to far smaller geographic areas. Krugman (1991) observed that states are not the right geographic units to analyze agglomeration processes because they are too large to capture the local processes that facilitate innovative activity. Florida (1998) found that regions, rather than countries as a whole, are the focal points of knowledge creation and learning. 5. The large differences between the number of acquisitions in New York and London can perhaps be attributed to differences in the general orientation of activity in these cities. Activity in New York is more locally focused than in London, with more domestic than foreign acquisitions.

REFERENCES Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. London: Sage. Audretsch, D. B., & Feldman, P. M. (1996). R&D spillovers and the geography of innovation and production. American Economic Review, 86(3), 630–640. Aw, B. Y., & Batra, G. (1998). Firm size and the patterns of diversification. International Journal of Industrial Organization, 16, 313–331. Barbosa, N., & Louri, H. (2002). On the determinants of multinationals’ ownership preferences: Evidence from Greece and Portugal. International Journal of Industrial Organization, 20, 493–515. Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120. Barney, J. (2001). Is the resource-based view a useful perspective for strategic management research? Yes. Academy of Management Review, 26(1), 41–56. Baum, J. A., & Haveman, H. A. (1997). Love thy neighbor? Differentiation and agglomeration in the Manhattan hotel industry, 1898–1990. Administrative Science Quarterly, 42(2), 304–338. Baum, J. A., & Mezias, S. J. (1992). Localized competition and organizational failure in the Manhattan hotel industry, 1898–1900. Administrative Science Quarterly, 37(4), 580–599.

256


Baum, J. A., & Usher, J. M. (2000). Making the next move: How experiential and vicarious learning shape the locations of chains’ acquisitions. Administrative Science Quarterly, 45, 766–801. Birkinshaw, J., & Hood, N. (2000). Characteristics of foreign subsidiaries in industry clusters. Journal of International Business Studies, 31(1), 141–154. Blau, P. M. (1977). Inequality and heterogeneity: A primitive theory of social structure. New York: The Free Press. Blonigen, B. A. (1997). Firm-specific assets and the link between exchange rates and foreign direct investment. American Economic Review, 87(3), 447–465. Braunerhjelm, P., & Svensson, R. (1996). Host country characteristics and agglomeration in foreign direct investment. Applied Economics, 28, 833–840. Chen, H., & Chen, T. J. (1998). Network linkage and location choice in foreign direct investment. Journal of International Business Studies, 29(3), 445–468. Chung, W. (2001a). Mode, size, and location of foreign direct investments and industry markups. Journal of Economic Behavior and Organization, 45, 185–211. Chung, W. (2001b). Identifying technology transfer in foreign direct investment: Influence of industry conditions and investing firm motives. Journal of International Business Studies, 32(2), 211–229. Chung, W., & Alcacer, J. (2002). Knowledge seeking and location choice of foreign direct investment in the United States. Management Science, 48(12), 1534–1554. Chung, W., & Kalnins, A. (2001). Agglomeration effects and performance: A test of the Texas lodging industry. Strategic Management Journal, 22, 969–988. De Bondt, R., Slaets, P., & Cassiman, B. (1992). The degree of spillovers and the number of rivals for maximum effective R&D. International Journal of Industrial Organization, 10, 35–54. Djellal, F., & Gallouj, F. (1999). Services and the research for relevant innovation indicators: A review of national and international surveys. Science and Public Policy, 26(4), 218–232. Dunning, J. H. (1993). Multinational enterprises and the global corporation. New York: AddisonWesley. Fainstein, S. S., Gordon, I., & Harlue, M. (Eds). (1992). Divided cities: New York and London in the contemporary world. Oxford, UK: Blackwell. Fischer, J. H., & Harrington, J. E. (1996). Product variety and firm agglomeration. RAND Journal of Economics, 27(2), 281–309. Florida, R. (1998). Calibrating the learning region. In: J. De la Mothe & G. Paquest. (Eds), Local and regional systems of innovation (pp. 19–28). Norwell, MA: Kluwer Academic Publishers. Florida, R., & Kenney, M. (2000). Transfer and replication of organizational capabilities: Japanese transplant organisations in the United States. In: G. Dosi, R. R. Nelson & S. G. Winter (Eds), The nature of dynamics of organizational capabilities (pp. 281–307). Oxford: Oxford University Press. Friedman, J., Gerlowski, D. A., & Silberman, J. (1992). What attracts foreign multinational corporations? Evidence from branch plant location in the United States. Journal of Regional Science, 32(4), 403–418. Friedrich, R. J. (1982). In defense of multiplicative terms in multiple regression equations. American Journal of Political Science, 26(4), 797–833. Gersbach, H., & Schmutzler, A. (1999). External spillovers, internal spillovers and the geography of production and innovation. Regional Science and Urban Economics, 29, 679–696.


257

Ghemawat, P. (2005). Regional strategies for global leadership. Harvard Business Review, 83(12), 98–108. Glaeser, E. I., & Kohlhas, J. (2004). Cities, regions, and the decline of transport costs. Papers in Regional Science, 83(1), 197–228. Gordon, I. (2002). Global cities, internationalization and urban systems. In: P. McCann (Ed.), Industrial location economics (pp. 187–206). Cheltenham, UK: Edward Elgar. Grein, A. (1996). Foreign firms’ attitudes, experiences, plans and expectations re New York City: survey results and trends. In: M. L. Chadwin (Ed.), International business: The New York experience (pp. 65–92). New York, NY: The Weissman Center for International Business, CUNY. Haleblian, J., & Finkelstein, S. (1999). The influence of organizational acquisition experience. Administrative Science Quarterly, 44, 29–56. Hay, D. A. (1976). Sequential entry and entry-deterring strategies in spatial competition. Oxford Economic Papers, 28(2), 240–257. Head, K., Ries, J., & Swenson, D. (1995). Agglomeration benefits and location choice: Evidence from Japanese manufacturing investments in the United States. Journal of International Economics, 38, 223–247. Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions and organizations across nations (2nd ed.). Thousands Oaks, CA: Sage. Holbrook, D., Cohen, W. M., & Hounshell, D. A. (2000). The nature sources and consequences of firm differences in the early history of the semiconductor industry. Strategic Management Journal, 21(11), 1017–1041. Ingram, P., & Baum, J. A. C. (1997). Chain affiliation and the failure of Manhattan hotels, 1896–1980. Administrative Science Quarterly, 42, 68–102. Kogut, B., & Singh, H. (1988). The effect of national culture on the choice of entry mode. Journal of International Business Studies, 19(3), 411–432. Kravis, I. B., & Lipsey, R. E. (1982). The location of overseas production and production for export by US multinational firms. Journal of International Economics, 12, 201–223. Krugman, P. (1991). Geography and trade. Cambridge, MA: MIT Press. Licht, G., & Moch, D. (1999). Innovation and information technology in services. Canadian Journal of Economics, 32(1), 363–382. Lomi, A. (1995). The population ecology of organizational founding: Location dependence and unobserved heterogeneity. Administrative Science Quarterly, 40, 111–144. Mariotti, S., & Piscitello, L. (1995). Information costs and location of FDI within host country: Empirical evidence from Italy. Journal of International Business Studies, 26(4), 815–842. Martin, R., & Sunley, P. (2002). Deconstructing clusters: Chaotic concept or policy panacea? Journal of Economic Geography, 3(1), 1–53. Mata, J., & Portugal, P. (2000). Closure and divestiture by foreign entrants: The impact of entry and post-entry strategies. Strategic Management Journal, 21, 549–562. Moomaw, R. L. (1998). Agglomeration economies: Are they exaggerated by industrial aggregation? Regional Science and Urban Economics, 28, 199–211. Nachum, L. (2000). Economic geography and the location of MNEs: Financial and professional service FDI to the US. Journal of International Business Studies, 31(3), 367–386. Nachum, L., & Keeble, D. (2000). Localised clusters and the eclectic paradigm of foreign investment: The clustering of film MNEs in central London. Transnational Corporations, 9(1), 1–37.

258


Nachum, L., & Keeble, D. (2001). TNC linkages in localised clusters: Foreign and indigenous firms in the media cluster of central London. In: Academy of Management Best Papers Proceedings, Washington, DC. Nahm, K. B., & Semple, K. (1995). Concentration and dispersion of global control links and changes in the multinational quaternary place system. In: M. B. Green & R. B. McNaughton (Eds), The location of direct investment: Geographic and business approaches (pp. 223–269). Aldershot, UK: Avebury. Ohmae, K. (1995). The end of the nation state: The rise of regional economies. New York: Free Press. Owen, R. F. (1982). Inter-industry determinants of foreign direct investment. In: A. M. Rugman (Ed.), New theories of multinational enterprise (pp. 135–149). London: Croom Helm. Pitt, L., Berthon, P., & Watson, R. (1999). Cyberservices: Taming service marketing with the WWW. Business Horizons, 42(1), 11–18. Porter, M. (1985). Competitive advantage: Creating and sustaining superior performance. New York: The Free Press. Rauch, J. E. (1993a). Does history matter only when it matters little? The case of city-industry location. The Quarterly Journal of Economics, 108(3), 843–867. Rauch, J. E. (1993b). Productivity gains from geographic concentration of human capital: Evidence from cities. Journal of Urban Economics, 34(4), 380–400. Ray, E. J. (1998). Takeovers and transplants: Reassessing FDIUS. In: D. Woodward & D. Nigh (Eds), Foreign ownership and the consequences of direct investment in the United States: Beyond us and them (pp. 69–96). Westport, CT: Quorum Books. Rugman, A. M. (2005). The regional multinationals: MNEs and ‘global’ strategic management. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Verbeke, A. (2001). Location, competitiveness, and the multinational enterprise. In: A. Rugman & T. Brewer (Eds), The Oxford handbook of international business (pp. 150–177). Oxford, UK: Oxford University Press. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Sassen, S. (2001). The global city: New York, London, Tokyo (2nd ed.). Princeton, NJ: Princeton University Press. Saxenian, A. (1994). Regional advantage: Culture and competition in Silicon Valley and Route 128. Cambridge, MA: Harvard University Press. Scott, A. J. (1998). Regions and the world economy: The coming shape of world production, competition, and political order. Oxford, UK: Oxford University Press. Semple, R. K., & Phipps, A. G. (1982). The spatial evolution of corporate headquarters within an urban system. Urban Geography, 3, 258–279. Shaver, J. M. (1998). Accounting for endogeniety when assessing strategy performance: Does entry mode choice affect FDI survival? Management Science, 44(4), 571–592. Shaver, J. M., & Flyer, F. (2000). Agglomeration economies, firm heterogeneity and foreign direct investment in the United States. Strategic Management Journal, 21, 1175–1193. Shefter, M. (Ed.), (1993). Capital of the American century: The national and international influence of New York City. New York: Russell Sage Foundations. Shenkar, O. (2001). Cultural distance revisited: Towards a more rigorous conceptualization and measurement of cultural differences. Journal of International Business Studies, 32(3), 519–535.


259

Song, J. (2002). Firm capabilities and technology ladders: Sequential Foreign Direct Investments of Japanese electronics firms in East Asia. Strategic Management Journal, 23, 191–210. Star Wars rule of clusters. (1999). The Economist, 76, May 8. Studenmund, A. H. (1992). Using econometrics: A practical guide. New York: Harper Collins. Suarez-Villa, L., & Walrod, W. (1997). Operational strategy, R&D and intra-metropolitan clustering in a polycentric structure: The advanced electronic industries of the Los Angeles Basin. Urban Studies, 34(9), 1343–1380. Swaminathan, A. (2001). Resource positioning and the evolution of specialist organizations: The role of location and identity in the US wine industry. Academy of Management Journal, 44(6), 1187–1210. UNCTAD. (2001). World investment report 2000: Promoting linkages. Geneva: United Nations. Vermeulen, F., & Barkema, H. (2001). Learning through acquisitions. Academy of Management Journal, 44(3), 457–476. Ward, S. (1994). Trends in the location of corporate headquarters, 1969–1989. Urban Affairs Quarterly, 29, 468–478. Weisberg, S. (1985). Applied linear regression (2nd ed.). New York: Wiley. Wheeler, D., & Mody, A. (1992). International investment location decisions. Journal of International Economics, 33, 57–76. Winsted, K., & Patterson, P. (1998). Internationalization of services: the service exporting decision. The Journal of Services Marketing, 12(4), 294–311. Wymbs, C. (2001). Key drivers of foreign firms’ locational choice: Knowledge nodes, cross-cultural differences and investment motivation. Working Paper, ISRN Annual Meeting, Toronto, Canada. Yamori, N. (1998). A note on the location choice of multinational banks: The case of Japanese financial institutions. Journal of Banking and Finance, 22, 109–120.

TESTING THE GLOBAL AND REGIONAL STRATEGIES OF MULTINATIONAL ENTERPRISES Lei Li and Dan Li ABSTRACT This study compares U.S. firm international strategies between two starkly different industries. We find that firms are more inclined to adopt global strategies in the integrated global industry than in the multidomestic industry. The global strategy does not seem to be effective unless a firm possesses substantial intangible assets. R&D-based intangible assets play a more significant role than marketing-based intangible assets in both the integrated global industry and (to a lesser extent) the multidomestic industry. Additionally, internationalization pace has a positive direct impact, and a negative interaction effect with the global strategy on firm performance in the integrated global industry.

INTRODUCTION Regional strategy as opposed to global strategy for multinational enterprises (MNEs) had attracted scholarly attention as early as the second half of 1980s when a tenet was emerging in international management which suggests that Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 263–296 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13011-4

263

264

LEI LI AND DAN LI

‘‘markets are fast becoming ‘borderless,’ and strategies that fail to recognize the [global] integration of markets are both shortsighted and misguided’’ (Birkinshaw, Morrison, & Hulland, 1995, p. 637; also Morrison, Ricks, & Roth, 1991). However, the academic debate on the tension between regional and global strategies did not seem to gain the momentum until the publication of The End of Globalization (Rugman, 2000). By analyzing the international trade and foreign direct investment (FDI) stocks as well as the cross-border activities of some of the world largest MNEs, Rugman illustrated that the international business activities were mostly undertaken intra-regionally in each of the triad-regions of North America, European Union, and Japan/Asia rather than configured and integrated globally as had been suggested (e.g., Levitt, 1983; Yip, 1992). More recently, Rugman and Verbeke (2004) examined the year 2001 geographical sales distribution of the global Fortune 500 firms. They identified that among the 365 firms (with sufficient data), 320 firms adopted a home-region-oriented strategy; 11 firms were host-region oriented; 25 were bi-regional firms; and only 9 firms had a global presence. Delios and Beamish (2005) also reported similar results based on an investigation of Japanese multinationals. While Rugman and other scholars have showed the prevalence of regional focus of MNEs (Delios & Beamish, 2005; Grosse, 2005; Rugman & Hodgetts, 2001; Rugman & Verbeke, 2004; Rugman, 2005), little is known as to how this regional orientation varies from industry to industry. This neglect is somewhat puzzling considering that the influential global integration– national responsiveness matrix for industry analysis was developed nearly two decades ago (Prahalad & Doz, 1987; Bartlett & Ghoshal, 1989), and our understanding of industry-level globalization was significantly enhanced by the important studies such as Kobrin (1991) and Makhija, Kim, and Williamson (1997). It appears that a substantial amount of effort has been made to investigate the global integration strategies and mechanisms in the contexts of global industries (Morrison & Roth, 1992; Birkinshaw et al., 1995; Kim, Park, & Prescott, 2003) whereas few studies compare the patterns of firm international strategies (e.g., regional versus global) in different industries. Moreover, our understanding of the performance implications of regional versus global orientation is limited. It is not clear under what circumstances a global strategy is superior to a regional strategy and vice versa. For example, Delios and Beamish (2005) found that Japanese firms with home-regional strategies appeared to under-perform those with global reach, whereas Li (2005) demonstrated that home-region orientation tended to enhance the performance of U.S. multinationals in the service industries. There is

Testing the Global and Regional Strategies of MNEs

265

also a lack of evidence as to how firm resources, in particular, firm-specific intangible assets contribute to the implementation of a global strategy visa`-vis a regional strategy despite a number of studies that report the role of intangible assets in enhancing the shareholder value and profitability of MNEs (e.g., Morck & Yeung, 1991; Kotabe, Srinivasan, & Aulakh, 2002; Li, 2003; Lu & Beamish, 2004). We contend that international strategies and performance implications need to be addressed in specific industry contexts for at least two major reasons. First, different industries tend to vary in terms of dominant internationalization motives and processes due to differences of economies of scale, technological intensity, marketing intensity, capital intensity, and industrial concentration etc. (Shan & Song, 1997; Sarkar, Cavusgil, & Aulakh, 1999). As a consequence, the potential of value creation by adopting a specific international strategy differs considerably. Second, different industries vary greatly in terms of existing global configuration and integration of assets and business activities (Kobrin, 1991; Makhija et al., 1997; Kim et al., 2003). In a more integrated global industry, firms tend to face fewer challenges to coordinate cross-border activities because there are more internationalization examples to follow, and the international linkages are easier to establish. In contrast, firms have to either operate various foreign subsidiaries separated from each other or take more burdens to establish and coordinate intra-firm relationships in a less integrated global industry such as a multidomestic industry. In this study, we seek to contribute to the emerging literature of regional multinationals by examining: (1) How do the patterns of international strategies vary in two entirely different industry contexts? More specifically, is global strategy more common in an integrated global industry while regional strategy more prevalent in a multidomestic industry? (2) How do firms perform with different international strategies in different industry contexts? (3) How is a specific international strategy’s effectiveness contingent upon the intensity and international deployment speed of firmspecific intangible assets?

THEORIES AND HYPOTHESES The choice of firm international strategies depends upon the exogenous structural forces and the patterns of international competition, or broadly speaking, the international industry contexts, which vary markedly from one industry to another (Porter, 1986; Birkinshaw et al., 1995). Although the

266

LEI LI AND DAN LI

international industry contexts can be classified in various ways, there appears to be a consensus that industry-specific pressures for global integration and/or local market responsiveness vary along a broad spectrum with endpoints that can be labeled as ‘‘integrated global’’ and ‘‘multidomestic’’ (Porter, 1986; Morrison et al., 1991; Birkinshaw et al., 1995; Makhija et al., 1997). An integrated global industry is defined as the one in which ‘‘competitive advantage is derived by firms from high coordination of geographically dispersed value-added activities’’ (Makhija et al., 1997, p. 690) whereas a multidomestic industry is an industry in which ‘‘international firms derive competitive advantage through largely separate and unlinked FDIs in multiple countries’’ (Makhija et al., 1997, p. 689). The emergence of ‘‘integrated global industries’’ is driven by (1) the potential for economies of scale from the global configuration of value-adding activities, (2) differences in comparative advantages across countries, and (3) standardized market demand across countries (Levitt, 1983; Ghoshal, 1987; Kim, Hwang, & Burgers, 1989; Birkinshaw et al., 1995). The persistence of ‘‘multidomestic industries’’ may be attributed to the relative absence of the above three drivers and/or the significance of cultural and institutional (e.g., legal and political) barriers. In light of the contingency perspective (Lawrence & Lorsch, 1967; Thompson, 1967), a firm tends to adopt an international strategy that fits with the industry context (Nohria & Ghoshal, 1997; Child, Chung, & Davies, 2003). A global strategy (i.e., a strategy with a worldwide geographic reach) seems to be desirable in an industry dominated by global integration drivers (Hout, Porter, & Rudden, 1982; Yip, 1992; Birkinshaw et al., 1995). By contrast, a firm is likely to embrace a regionally oriented strategy (i.e., a strategy with a limited geographic reach) in an industry lacking a need for global integration, and/or confronting significant institutional and cultural barriers. Therefore: Hypothesis 1. Firms tend to adopt a global strategy in an integrated global industry, and a regionally oriented strategy in a multidomestic industry. Firm performance is contingent upon the fit between environment (industry) and strategy (Porter, 1986; cited in Birkinshaw et al., 1995). A firm can gain competitive advantages through accurately assessing the underlying industry structural potential for globalization and by adopting an appropriate international strategy accordingly (Nohria & Ghoshal, 1997). Dunning (1993, pp. 427–428) noted that MNEs helped refashion the international division of labor both between developed and developing


267

countries and within regionally integrated developed countries. To the extent that such a specialization has taken place in industries in which economies of scale or scope are the most prevalent, and the comparative advantages (e.g., low-cost resource endowments, high national investment incentives) and the standardization for product offerings and distribution systems are most critical (i.e., integrated global industries), this has often given a notable competitive edge to MNEs that adopt a global strategy. By contrast, in an industry characterized by limited-scale benefits beyond national or regional boundaries and/or significant barriers across nations (i.e., a multidomestic industry), a regionally oriented strategy seems to be more economically rational. Certainly, the implementation of an international strategy, whether global or regional, entails a varying degree of worldwide geographic reach and corresponding vertical/horizontal cross-border integration of operations (Kobrin, 1991). The external transaction costs and internal administrative costs for dealing with national/regional differences and cross-border coordination need to be weighed against the scale and integration benefits (Contractor, Kundu, & Hsu, 2003; Li, 2005; Lu & Beamish, 2004). As Ghemawat (2003) points out, a distinctive feature of international strategy is that regional specificity matters. In an integrated global industry where similar products and services are valued by the customers worldwide, and suppliers or partners (inside or outside of the firm) in different geographic regions are more easily accessible due to the intensive cross-border industry network established over time, both the external transaction and internal administrative costs would be less significant than in a multidomestic industry. In summary, a global strategy tends to generate more benefits and face fewer challenges in an integrated global industry as opposed to a multidomestic industry. Therefore, we expect: Hypothesis 2a. Global strategy is more effective than regionally oriented strategies in an integrated global industry. Hypothesis 2b. Regionally oriented strategies are more effective than global strategy in a multidomestic industry. The effectiveness of an international strategy is contingent upon firmspecific intangible assets (e.g., Severn & Laurence, 1974; Kotabe et al., 2002; Li, 2003; Lu & Beamish, 2004). Firm-specific intangible assets refer to the unique organizational resources and managerial capabilities that a firm has accumulated over time. ‘‘Given the resources and time costs of developing such assets, the efficiency of and returns to their exploitation is greater when

268

LEI LI AND DAN LI

their scope of use is greater’’ (Teece, 1986, cited in Lu & Beamish, 2004, p. 602). The core FDI theories as well as the resource-based view have made compelling arguments associating firm-specific intangible assets with MNE performance (Vernon, 1966; Caves, 1971; Buckley & Casson, 1976; Rugman, 1981; Tallman & Li, 1996; Hitt, Hoskisson, & Kim, 1997) because international presence provides an opportunity for MNEs to leverage their intangible assets (which ‘‘have some characteristics of public goods’’) in a broad market scope (Morck & Yeung, 1991). However, a major problem with deriving benefits from leveraging globally firm-specific intangible assets is that certain intangible assets (e.g., experiential knowledge) are sticky and do not move easily across countries, even when transferred willingly (Kogut & Zander, 1993; Szulanski, 1996; Rugman & Verbeke, 2001). Thus, at least a portion of intangible assets are characterized by mobility barriers or isolating mechanisms that make full exploitation across countries very difficult. This argument is consistent with Vermeulen and Barkema (2002) and Goerzen and Beamish (2003) who found that when an MNE has operations in very dissimilar countries, its absorptive capacity to benefit from the dispersed information is strongly taxed. The difficulties of diffusing some firm-specific intangible assets across national/regional borders are less serious in an integrated global industry than in a multidomestic industry due to the abundance of international linkages of trade, and other value-added activities in the former (Makhija et al., 1997). We expect that firm-specific intangible assets enhance the effectiveness of the global strategy in an integrated global industry. Though the intangible assets can also improve the effectiveness of a regionally oriented strategy, the economies of scope would not be as fully realized when a firm’s operation is restricted to a limited number of markets. Therefore, we argue that firm-specific intangible assets enhance the effectiveness of the global strategy relative to regionally oriented strategies in an integrated global industry. In accordance with Rugman and Verbeke’s (1992, 2001, 2004) arguments, firm-specific intangible assets are harder to transfer in a multidomestic industry as opposed to an integrated global industry because they tend to be more location or region bound. Although the intangible assets are valuable in a multidomestic industry, the value exploitation may be only effective on a regional basis. Moreover, the significant costs associated with the diffusion of the intangible assets tend to discount the benefits of international deployment when the geographic reach becomes too broad as in the case of the global strategy. In some instances (e.g., chewing tobacco), such diffusion


269

of firm-specific intangible assets across national boundaries may not be likely. Therefore, we expect that firm intangible assets enhance the effectiveness of regionally oriented strategies relative to the global strategy in a multidomestic industry Hypothesis 3a. Firm intangible assets enhance the effectiveness of the global strategy relative to regionally oriented strategies in an integrated global industry. Hypothesis 3b. Firm intangible assets enhance the effectiveness of regionally oriented strategies relative to the global strategy in a multidomestic industry. Although firm intangible assets may include a wide range of firm resources and capabilities such as brand name, technological know-how, organizational routines, corporate culture, and social relations, the literature identifies two major types of intangible assets: R&D-based and marketing-based intangible assets (Morck & Yeung, 1991; Hitt et al., 1997; Pantzalis, 2001; Kotabe et al., 2002; Goerzen & Beamish, 2003; Lu & Beamish, 2001, 2004). As noted by some scholars, the majority of the industries are characterized by semi-globalization whereby firms have the tendency to standardize some links of value-added chain (typically upstream activities) by global integration and differentiate other links (typically downstream activities) by local adaptation (Kogut, 1985; Ghemawat, 2003). R&D-based intangible assets seem to reflect a firm’s central capability of performing upstream activities (e.g., product development and manufacturing), while marketing-based intangible assets indicate the firm’s capability of managing downstream activities (e.g., branding, packaging, labeling, distributing). Indeed, as Kotabe et al. (2002) argued, a firm’s R&D commitment would enable it to achieve greater return by differentiating its product from its competitors and by improving its manufacturing processes, while a firm’s marketing commitment could help enhance its revenues by better targeting the customer needs and by developing standardized marketing programs across foreign markets. Rugman and Verbeke (2003) elaborated that for an MNE to leverage downstream intangible assets, the resource commitments made to attract potential foreign customers are one-sided; the MNE invests its resources up-front whereas, on the other side of the equation, the customers do not. Thus, the challenge of a successful exploitation of downstream intangibles

270

LEI LI AND DAN LI

becomes greater, demanding greater resource commitments. In contrast, while foreign markets may also require location-specific linking investments in the case of the MNE’s leveraging upstream intangible assets, the firm is likely to enjoy greater local support thereby increasing the chances of success in this area of the value chain. The above arguments are consistent with Porter’s (1986) insight that ‘‘downstream activities create competitive advantages that are largely country-specific: a firm’s reputation, brand name, and service network in a country grow largely out of a firm’s activities in that country and create entry/mobility barriers in that country alone. Competitive advantage in upstream and support activities often grows more out of the entire system of countries in which a firm competes than from its position in any one country.’’ Put differently, an MNE’s intangible assets resulting from downstream activities tend to be more location bound than those from upstream activities. In an integrated global industry where competitive advantages can be derived from leveraging firm-specific intangible assets, those from upstream activities (i.e., R&D-based intangible assets) are expected to have a larger impact on the effectiveness of a global strategy than those from downstream activities (i.e., marketing-based intangible assets). In a multidomestic industry, however, where firm competitive advantages tend to come from location-bound firm-specific advantages, the intangible assets employed to manage downstream activities (i.e., marketing-based intangible assets) are likely more significant. Indeed, Kobrin (1991) reveals at the industry-level that the R&D intensity is positively associated with the degree of industry globalization whereas the marketing intensity tends to be negatively associated with the degree of industry globalization. Hypothesis 4a. In an integrated global industry, R&D-based intangible assets enhance the effectiveness of a global strategy relative to regionally oriented strategies more than marketing-based intangible assets. Hypothesis 4b. In a multidomestic industry, marketing-based intangible assets enhance the effectiveness of regionally oriented strategies relative to global strategy more than R&D-based intangible assets. The effectiveness of a firm’s international strategy is contingent not only upon the intensity of firm-specific intangible assets, but also upon the deployment and development process of intangible assets in the course of


271

internationalization. A central feature of this process is captured by firm internationalization pace (Vermeulen & Barkema, 2002). The assumption here is that the speed of establishing foreign subsidiaries is equivalent to the speed of deploying and acquiring intangible assets. This is consistent with the insight that tangible-asset flows (e.g., monetary investment, products) often embody intangible-asset flows such as technology transfers (Kobrin, 1991; Makhija et al., 1997). The international deployment and development of intangible assets is often path dependent and subject to time compression diseconomies (Dierickx & Cool, 1989). A fast internationalization pace tends to compromise managers’ abilities to deploy and acquire intangible assets effectively in the international markets for at least two reasons. First, the ability to evaluate foreign institutional environments and design appropriate market entry programs takes time to develop. Managers’ learning capacity and information processing capabilities have constraints. Trying to learn about foreign business environments rapidly may result in managers having less time to digest, absorb, accumulate, and apply the knowledge. Second, even after the market entries are materialized, the management of multiple foreign subsidiaries established in a short time period can be challenging. For instance, Tsai (2000) reported that it took considerable amount of time for an MNE to establish linkages between different units in order to function effectively. Under the time pressure imposed by a fast internationalization pace, it is less likely for the management to carefully evaluate the environments, build new subsidiaries, and then integrate them into the existing organizational system (Birkinshaw, 1997; Birkinshaw & Hood, 1998). Evidently, the wider the geographic reach an international strategy entails, the more challenges a fast internationalization pace will impose on firm managers. Thus, a global strategy in conjunction with a fast internationalization pace is expected to hamper firm performance relative to regionally oriented strategies. As noted previously, institutional and cultural barriers tend to be more significant in a multidomestic industry than in an integrated global industry. Therefore, the negative moderating effects of internationalization pace on global strategy and performance relationship are expected to be more serious in a multidomestic industry. Hypothesis 5. The faster the firm internationalization pace, the less effective a global strategy will be, relative to regionally oriented strategies; the effects are more salient in a multidomestic industry than in an integrated global industry.

272

LEI LI AND DAN LI

METHODOLOGY Empirical Setting We have chosen two 3-digit SIC manufacturing industries: the computer and office equipment industry (hereinafter COE industry, SIC 357), and the soap, cleanser, and toilet goods industry (hereinafter SCT industry, SIC 284). Although both are knowledge-intensive industries, the COE industry is characterized by high R&D intensity and high global integration, whereas the SCT industry is particularly marketing intensive and country-centered without much interdependence across the national borders. In fact, the two industries are categorized as integrated global industry and multidomestic industry respectively by Makhija et al. (1997), which is consistent with Kobrin (1991), Morrison and Roth (1992), and Kim et al. (2003). Thus, they serve as excellent settings for examining the patterns of international strategies and performance implications comparatively.

Data and Sampling We relied on the Directory of Corporate Affiliations (DCA) compiled by Lexis–Nexis to identify the geographical distribution of a firm’s foreign subsidiaries (i.e., wholly owned foreign subsidiaries or joint ventures with equity ownership greater than 50%). We also searched the Securities Data Corporation Platinum (SDC) to obtain information regarding the firm’s foreign alliances (i.e., international joint ventures with equity ownership less than 50% or non-equity-based contractual alliances). We then checked the annual reports filed to the Securities and Exchange Commission (SEC) to corroborate the data from the above sources and fill the missing values. The accounting data items were extracted from the Compustat database, and were corroborated and complemented by searching through the annual reports of the sample firms. Our sample is obtained through the following steps. First, the initial list of U.S. firms in the two industries (SIC 357 and 284) was extracted from the Compustat database. Second, the online DCA was searched to obtain the data on geographic distribution of subsidiaries from 1993 to 2004. After dropping the firms lacking information about their subsidiaries during our study period, our final sample includes 200 U.S. firms and 1,265 firm-year observations from both industries. It should be noted that our sample includes the firms that did not survive the observation period, and thus the


Table 1.

273

Sample Distribution by Industry and Year.

Year

Computer and Office Equipment (COE)

Soap, Cleanser and Toilet Products (SCT)

Total

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

73 78 81 79 76 71 73 81 86 87 87 84

28 30 30 27 27 24 27 26 23 23 23 21

101 108 111 106 103 95 100 107 109 110 110 105

Total

956

309

1265

current study is able to mitigate the problem of survival bias. Table 1 reports the distribution of our sample by industry from 1993 to 2004. Variables and Measures Dependent Variable Firm Performance. We use three indicators to measure performance. Return on sales (ROS ) measures the profitability associated with market performance. Return on assets (ROA) takes into account both ROS and the utilization of firm assets in accordance with Dupont formula, and thus, is arguably the best conventional accounting-based performance indicator (e.g., Grant, 2005). Tobin’s q is perhaps the most frequently used long-term performance indicator in the literature (e.g., Morck & Yeung, 1991; Lu & Beamish, 2004). We use Chung and Pruitt’s (1994) approximation for Tobin’s q because it has a very high correlation with the more rigorous calculation (Berry, 2006). Independent Variables Firm International Strategy. We adopt Rugman and Verbeke’s (2004) criteria to define the four types of international strategies, namely homeregion-oriented strategy, host-region-oriented strategy, bi-regional strategy, and global strategy. Rugman and Verbeke use the geographic sales distribution (i.e., market output) to classify these international strategies. We focus on the geographic distribution of firm subsidiaries (i.e., firm operational activities)

274

LEI LI AND DAN LI

instead. We argue that the configuration of firm subsidiaries (or operations) may be a more reasonable yardstick to gauge a firm international strategy according to the definition of MNEs – the firms that have established two or more country business enterprises in which they exercise some minimum level of control (Dunning, 1993; Caves, 1996). For the purpose of comparison, we also use sales-based measures to replicate a part of the analyses. The definitions of the four international strategies are as follows: (1) Home-region oriented: At least 50% of the subsidiaries are located in the home region of the Triad (i.e., Americas including both North and South America).1 (2) Host-region oriented: At least 50% of the subsidiaries are located in the host region of the Triad (i.e., Europe including Mideast and Africa, or Asia-Pacific). (3) Bi-regional: At least 20% of the subsidiaries are located in two Triad regions each, and none of the three Triad regions has more than 50% of the subsidiaries. (4) Global: At least 20% of the subsidiaries are located in each of the three Triad regions, but less than 50% in any one region of the Triad. R&D-Based Intangible Assets. We assess a firm’s R&D-based intangible assets by adopting the proxy of R&D intensity, which is computed as the ratio of R&D expenses to total sales. This measure largely reflects a firm’s internal R&D capability. Marketing-Based Intangible Assets. We measure a firm’s marketing-based intangible assets by its marketing intensity, which is calculated as the ratio of sales and marketing expenses to total sales. This measure largely reflects a firm’s capability to manage marketing and distribution activities. Internationalization Pace. We measure ‘‘internationalization pace’’ using the average number of new foreign subsidiaries established during the past three years, weighted by the total number of foreign subsidiaries in the current year. t n P P

Pace ¼

ðM i;T CDi Þ

T¼t2 i¼1

Nt 3

where Pace stands for internationalization pace in year t; Mi,T is number of newly established foreign subsidiaries in host country i in year


275

T (T = t2 to t); CDi is cultural distance between host country i and the U.S. (Hofstede, 1980; Kogut & Singh, 1988); Nt is total number of foreign subsidiaries in year t; n is the total number of the firm’s host countries. We incorporate both the growth rate of the number of foreign subsidiaries and the speed at which the firm expands into culturally dissimilar foreign countries. For example, if two firms have established the same number of subsidiaries during a certain time period, the one that has entered more dissimilar foreign countries has a higher internationalization pace. Control Variables We control for the variables that are widely acknowledged as important factors in prior research on the relationship between firm strategy and performance. Firm size is measured as the number of employees in logarithmic form. Firm age is the number of years since the founding, which reflects the firm’s overall business experience. Consistent with the literature (e.g., Palepu, 1985; Baysinger & Hoskisson, 1989; Hoskisson, Johnson, & Moesel, 1994; Hitt et al., 1997), we adopt the entropy index to measure business diversity. The P entropy index is based on business segment sales reported by firms, i.e., i ½BSPi ln ð1=BSPi Þ, where BSPi is the sales attributed to business segment i and ln(1/BSPi) is the salesbased weight given to business segment i. The firm’s financial leverage is computed as the ratio of long-term debt to total assets. This measure of firm capital structure is particularly relevant to long-term performance indicators such as Tobin’s q. We also control for the firm’s entry mode preference. Firms enter foreign markets by either founding majority-owned subsidiaries or forming alliances (including minority-owned joint ventures). We use the percentage of the cumulative international market entries accounted for by establishing majority-owned foreign subsidiaries to measure a firm’s entry mode (or internalization) preference. This variable reflects a firm’s corporate control of foreign operations. Finally, we use 11-year dummy variables to control the temporal trend (1993–2004). Analytical Models We first compare the patterns of international strategies between the COE and SCT industry using subsidiary-based measures (as well as the salesbased measures adopted by Rugman and Verbeke). We apply the general

276

LEI LI AND DAN LI

linear model (GLM) to examine whether the differences of the patterns are significant. The GLM model takes into account the unbalanced number of observations between the two groups. To sense whether the differences are inter-temporally stable, we divide the study period into two time periods (1993–1998 and 1999–2004). We then conduct cross-sectional time-series regressions on the two industries separately to test the effects of international strategies on firm performance. Hausman tests indicate that feasible generalized least square (FGLS) regression is needed to control for the first-order autocorrelation and heteroskedasticity. To test Hypotheses 2–5, we need to compare the performance implications between three categories of international strategies: global strategy, bi-regional strategy, and single-region-oriented strategy (home and host). Technically, to compare the global strategy with the bi-regional strategy, we drop the observations of the single-region-oriented strategy; to compare the global strategy with the single-region-oriented strategy, we drop the observations of the bi-regional strategy; to examine the bi-regional strategy versus the single-region-oriented strategy, the observations of the global strategy are removed.

RESULTS The descriptive statistics of all variables are summarized in Table 2a (COE industry) and Table 2b (SCT industry). The descriptive statistics indicate that the COE industry are more R&D intensive while the SCT industry are more marketing intensive, though both can be classified as knowledgeintensive industries because the R&D spending is more than 5% of sales according to the OECD criterion (www.oecd.org). Table 2b also shows that no firms in the SCT industry adopt a global strategy (see details below). Table 3 compares the patterns of international strategies between the COE and SCT industries to test Hypothesis 1. Tables 4 and 5 report the impact on firm performance of the global strategy vis-a`-vis the bi-regional strategy and the single-region-oriented (home and host) strategy, respectively, in the COE industry. These two tables also show the interaction effects on firm performance between firm-specific intangible assets (R&D and marketingbased) and global strategy in the COE industry. Since Table 3 shows that no firms in the SCT industry adopted the global strategy during our observation period, the hypotheses such as Hypotheses 2b, 3b, and 4b are, strictly speaking, not testable. Thus, we treat the bi-regional strategy as a proxy for

Table 2a. Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

ROA Tobin’s q ROS Firm size Firm age Business diversity Financial leverage Entry mode preference Home region Host region Bi-regional Global R&D intensity Marketing intensity Internationalization pace

po0.05.

N 956 956 956 956 956 956 956 956 956 956 956 956 956 956 956

Mean

S.D.

0.066 0.342 2.154 2.153 0.080 0.483 6.882 1.855 16.238 11.050 0.226 0.349 0.120 0.186 0.946 0.159 0.482 0.206 0.215 0.096 0.135 0.217 0.037

0.500 0.405 0.411 0.295 0.154 0.086 0.317

Minimum Maximum 5.460 0.001 8.074 0 0 0 0 0.083 0 0 0 0 0 0.004 6.508

1.742 32.577 1.709 11.925 79 2.099 2.171 1 1 1 1 1 1.755 0.928 0.950

Descriptive Statistics (COE Industry). 1

2

3

4

5

6

7

1.000 0.164 1.000 0.782 0.036 1.000 0.210 0.025 0.164 1.000 0.054 0.120 0.074 0.272 1.000 0.023 0.095 0.011 0.307 0.266 1.000 0.000 0.137 0.003 0.118 0.188 0.072 1.000 0.085 0.074 0.048 0.308 0.230 0.009 0.015 0.023 0.028 0.070 0.009 0.014 0.049 0.038 0.008 0.347 0.048 0.125 0.012 0.023 0.058

0.001 0.160

8

9

10

11

12

13

14

15

1.000

0.141 0.061 0.090 0.000 1.000 0.018 0.051 0.107 0.131 0.061 0.003 0.492 1.000 0.011 0.129 0.051 0.015 0.043 0.020 0.506 0.267 1.000 0.038 0.161 0.020 0.097 0.010 0.024 0.315 0.166 0.171 1.000 0.520 0.258 0.184 0.085 0.058 0.075 0.025 0.001 0.047 0.022 1.000 0.226 0.256 0.056 0.072 0.093 0.026 0.009 0.038 0.099 0.100 0.381 1.000 0.003 0.027 0.035 0.022 0.018 0.012 0.153 0.045 0.091 0.071 0.049 0.036 1.000

Table 2b. Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

ROA Tobin’s q ROS Firm size Firm age Business diversity Financial leverage Entry mode preference Home region Host region Bi-regional Global R&D intensity Marketing intensity Internationalization pace

po0.05.

N 309 309 309 309 309 309 309 309 309 309 309 309 309 309 309

Mean

S.D.

0.005 0.196 2.314 6.417 0.005 0.196 7.080 2.098 24.761 15.258 0.375 0.420 0.178 0.163 0.986 0.058 0.796 0.016 0.188 0 0.059 0.358 0.015

0.404 0.126 0.391 0 0.049 0.298 0.072

Minimum

Maximum

2.550 0.045 2.169 0.000 0 0 0 0.616

0.247 97.884 0.380 11.608 75 1.570 0.726 1

0 0 0 0 0 0.094 0.355

1 1 1 0 0.117 3.048 0.306

Descriptive Statistics (SCT Industry). 1

2

3

4

5

6

7

1.000 0.315 1.000 0.971 0.299 1.000 0.321 0.084 0.282 1.000 0.141 0.028 0.153 0.522 1.000 0.058 0.010 0.075 0.394 0.415 1.000 0.027 0.082 0.036 0.017 0.094 0.002 1.000 0.089 0.024 0.084 0.438 0.567 0.489 0.036

8

9

10

11

12

13

14

15

1.000

0.139 0.004 0.124 0.305 0.386 0.117 0.058 0.367 1.000 0.030 0.002 0.030 0.119 0.060 0.115 0.128 0.031 0.253 1.000 0.134 0.004 0.118 0.353 0.418 0.158 0.019 0.389 0.950 0.062

1.000

0.150 0.011 0.142 0.408 0.115 0.186 0.234 0.114 0.012 0.095 0.043 0.019 0.325 0.009 0.184 0.198 0.093 0.162 0.091 0.078 0.042 0.067 0.140 0.020 0.132 0.169 0.037 0.099 0.024 0.029 0.011 0.276 0.078

1.000 0.060 1.000 0.114 0.072 1.000

Period

1993–1998

1999–2004

w

Patterns of International Strategy: Comparison between COE and SCT Industry.

Classification of International Strategy

Subsidiary-Based Computer and office equipment (COE)

Soap, cleansers and toilet products (SCT)

Sales-Based F-value (PrWF )

Computer and office equipment (COE)

Soap, cleansers and toilet products (SCT)

F-value (PrWF )

Home region Host region Bi-regional Global

220 124 78 36

(48.03%) (27.07%) (17.03%) (7.86%)

134 1 31 0

(80.72%) (0.6%) (18.67%) (0%)

57.78 (o0.0001) 58.09 (o0.0001) 0.23 (0.6333) 14.12 (0.0002)

308 16 12 7

(89.80%) (4.66%) (3.50%) (2.04%)

125 2 3 0

(96.15%) (1.54%) (2.31%) (0%)

4.95 2.52 0.43 2.70

(0.0265) (0.1131) (0.5104) (0.1012)

Total Home region Host region Bi-regional Global

458 241 73 128 56

(100%) (48.39%) (14.66%) (25.70%) (11.24%)

166 112 4 27 0

(100%) (78.32%) (2.80%) (18.88%) (0%)

– 42.78 (o0.0001) 15.09 (o0.0001) 2.82w (0.0934) 18.06 (o0.0001)

343 344 31 39 16

(100%) (80%) (7.21%) (9.07%) (3.72%)

130 91 6 6 0

(100%) (88.35%) (5.83%) (5.83%) (0%)

3.87 0.25 1.13 3.97

– (0.0496) (0.6204) (0.2883) (0.0469)

Total

498 (100%)

143 (100%)

–

430 (100%)

103 (100%)


Table 3.

–

po 0.10,

po 0.05, po 0.01, po 0.001.

279

280

Table 4. Global versus Bi-regional Strategy (COE industry only). ROA

Control Firm size Firm age Business diversity Financial leverage Entry mode preference

Model 4.7 (global vs. bi-region)









0.002 0.002 0.005 0.122 0.086

0.011w 0.002w 0.008 0.150 0.106

0.006 0.003 0.008 0.113 0.133w

0.006 0.003w 0.005 0.105 0.277

0.003 0.002 0.015 0.132 0.084

0.003 0.002 0.015 0.107w 0.170

0.147 0.058 0.064 1.834 1.732

0.124 0.055 0.049 1.620 1.910

0.129 0.054 0.089 1.931 1.660

0.073 0.902 0.102 0.699

0.069 0.837 0.191

0.008 1.249 0.121

0.080 1.261 0.007 0.782

0.005 1.096 0.095

0.058 1.751 1.607w

0.040 1.686 1.866 0.453

0.245 1.967 1.523w

Independent variables Global strategy 0.004 R&D intensity 0.677 Marketing intensity 0.071 Global*R&D intensity Global*marketing intensity

0.315w

0.154 286

0.130 286

0.178w 286

244.48 135.68

246.93 161.84

242.22 169.69

0.037 0.462 286 186.67 96.14

0.232w 286

0.317w 286

179.79 92.48

187.61 82.77

1.779 3.155 286

3.290 286

3.260 286

439.52 235.81

432.02 176.08

425.69 243.07

Notes: Because of space limitations, we did not include individual coefficients on year dummies in this table. w po 0.10, po 0.05, po 0.01, po 0.001.

LEI LI AND DAN LI

Intercept Number of observations Log likelihood Wald w2

Tobin’s q

ROS

Global versus Single-Region-Oriented Strategy (Home and Host) (COE Industry only). ROA

Control Firm size Firm age Business diversity Financial leverage Entry mode preference

Model 5.7 (global vs. home/host)









0.016 0.001 0.005 0.168 0.032

0.020 0.001 0.026w 0.158 0.006

0.017 0.001 0.012 0.159 0.022

0.015 0.001w 0.026w 0.182 0.001

0.016 0.002 0.011 0.223 0.005

0.014 0.003 0.008 0.177 0.018

0.088 0.006 0.053 1.058 0.939

0.080 0.005 0.134 1.098 0.833

0.089 0.005 0.025 1.027 1.132

0.041 0.340 0.111 0.363

0.032 0.342 0.085

0.008 0.817 0.235

0.100 0.981 0.349 0.809

0.070w 0.874 0.402

0.131 0.116 0.070

0.245 0.298 0.506 1.441

0.443 0.225 0.449

Independent variables Global strategy 0.016 R&D intensity 0.337 Marketing intensity 0.006 Global*R&D intensity Global*Marketing intensity Intercept Number of observations Log likelihood Wald w2

Tobin’s q

ROS

0.039 730

0.111 730

0.071 730

620.24 133.57

612.95 115.96

627.33 119.95

3.184

0.312w

0.129

0.024 730

0.100 730

0.122 730

598.50 282.89

589.94 213.69

648.93 281.28

3.519 730 948.67 131.62

954.80 107.38

3.823 730 938.03 262.42

281


3.472 730


Table 5.

282

LEI LI AND DAN LI

the global strategy in the SCT industry when necessary. Tables 6 and 7 exhibit the performance implications of the bi-regional strategy vis-a`-vis the single-region-oriented strategy, and the corresponding moderating effects of firm-specific intangible assets in the COE and SCT industries, respectively. In addition, Table 8 reports the impact of internationalization pace on firm performance in both the COE and SCT industries. Hypothesis 1 states that firms tend to adopt a global strategy in an integrated global industry, and a regionally oriented strategy in a multidomestic industry. Table 3 compares the patterns of international strategies adopted by the firms in the two industries. It is clear that from 1993 to 2004, no firms in the SCT industry (a multidomestic industry) adopted a global strategy in contrast to the COE industry (an integrated global industry). Moreover, there were a larger percentage of firms that were home-region oriented in the SCT industry than in the COE industry. The results are similar when the sales-based data are used, though to a lesser significance level. In addition, there appear to be higher percentages of host-oriented firms and bi-regional firms in the COE industry than in the SCT industry. Overall, the differences tend to be more significant in more recent years (1999–2004). Evidently, Hypothesis 1 is supported. Hypothesis 2a states that global strategy is more effective than regionally oriented strategies in an integrated global industry. Hypothesis 2b states that regionally oriented strategies are more effective than global strategy in a multidomestic industry. From Tables 4 and 5, we see that the global strategy per se does not make a significant impact on the firm performance in the COE industry. When the interaction effects are included, however, the direct impact on short-term performance (ROA and ROS ) of the global strategy as opposed to the bi-regional strategy or the single-region-oriented strategy (home and host) is either mostly negative (when R&D intensity is incorporated) or largely insignificant (when market intensity is entered). Also, the direct impact on long-term performance (Tobin’s q) is always insignificant. Similar results are found when the bi-regional strategy is compared with the single-region-oriented strategy, except Model 6.2 in Table 6, which we discuss in detail later. These findings indicate that Hypothesis 2a is not supported. Regarding Hypothesis 2b, the fact that no firms in the SCT industry adopted the global strategy implies that the global strategy is at least not perceived effective or too challenging by top executives. Thus, Hypothesis 2b seems to be supported albeit indirectly. On the other hand, Table 7 shows that the bi-regional strategy tends to be more effective than the single-region-oriented strategy (see Models 7.1, 7.4, and 7.8 in Table 7). This implies that an international strategy with a moderate

Bi-Regional versus Single-Region-Oriented Strategy (Home and Host) (COE Industry). ROA

Control Firm size Firm age Business diversity Financial leverage Entry mode preference Independent variables Bi-region strategy R&D intensity Marketing intensity Bi-region*R&D intensity Bi-region*marketing intensity Intercept Number of observations Log Likelihood Wald w2

Tobin’s q

ROS Model 6.7 (bi-region vs. home/host)

Model 6.8 (bi-region vs. home/host)


0.010 0.001 0.022 0.109 0.029

0.075 0.013 0.021 0.837 0.744

0.070 0.012 0.033 0.866 0.474

0.056 0.011 0.092 0.830 0.776

0.060w 0.863 0.398

0.071 0.205 0.057

0.029 0.002 0.124 0.859

0.438 0.091 0.615

Model 6.1 (bi-region vs. home/ host)

Model 6.2 (bi-region vs. home/ host)





0.012 0.001 0.003 0.120 0.080

0.012 0.001 0.005 0.107 0.072

0.021 0.002 0.005 0.312 0.047

0.011 0.001 0.018 0.098 0.018

0.010 0.001 0.015 0.123 0.022

0.004 0.468 0.009

0.050 0.373 0.000 0.382

0.040 0.650 0.405

0.002 0.776 0.114

0.021 0.780 0.276 0.222

0.311

0.121

0.053 850

0.020 850

0.115 850

0.026 850

0.082 850

0.125 850

646.03 185.34

642.32 202.60

602.46 341.99

460.10 199.73

550.18 219.03

574.35 248.24


Table 6.

1.513w 3.429 850 1132.59 135.06

1125.65 133.66

3.391 850 1147.64 138.14

283


3.084 850

284

Table 7. Bi-regional versus Single-Region-Oriented Strategy (Home and Host) (SCT Industry). ROA

ROS

Tobin’s q

Model 7.9 Model 7.8 Model 7.7 Model 7.6 Model 7.5 Model 7.4 Model 7.3 Model 7.2 Model 7.1 (bi-region vs. (bi-region vs. (bi-region vs. (bi-region vs. (bi-region vs. (bi-region vs. (bi-region vs. (bi-region vs. (bi-region vs. home/host) home/host) home/host) home/host) home/host) home/host) home/host) home/host) home/host) Control Firm size Firm age Business diversity Financial leverage Entry mode preference Independent variables Bi-region strategy R&D intensity Marketing intensity Bi-region*R&D intensity Bi-region* marketing intensity

0.019 0.001 0.029 0.014 0.109

0.015 0.001w 0.026 0.022 0.184

0.012 0.000 0.011 0.018 0.070

0.014 0.000 0.015 0.019 0.082

0.013 0.000 0.010 0.011 0.034

0.210w 0.010 0.210 1.265 3.182

0.200 0.016 0.152 1.057 3.928

0.197 0.015 0.204 1.279 4.217w

0.023w 0.423 0.026

0.004 0.404 0.006 0.276

0.026 0.426 0.030

0.020w 0.433 0.013

0.022 0.472 0.014 0.549

0.016 0.431 0.016

0.373 2.084 0.371

1.030 0.337 1.410 10.559w

0.829 1.807 1.112

0.017

0.013

0.112 307

0.036 307

0.160 307

0.030 307

0.030 307

477.35 95.83

481.56 150.52

471.67 87.65

510.24 113.32

505.73 128.43

0.020 307 509.48 98.46

1.553

3.422 307 517.43 16


4.002w 307 500.54 29.28w

4.432 307 501.25 26.10

LEI LI AND DAN LI

Intercept Number of observations Log likelihood Wald w2

0.018 0.001 0.027 0.013 0.163

The Impact of Internationalization Pace.

ROA

Tobin’s q

ROS

Model Model 8.1 Model 8.2 Model 8.3 Model 8.4 Model 8.5 Model 8.6 Model 8.7 Model 8.8 Model 8.9 8.10 (global vs. (global vs. (bi-region (bi-region (global vs. (global vs. (bi-region (bi-region (global vs. vs. home/ vs. home/ bi-region) (global vs. home/ vs. home/ vs. home/ bi-region) bi-region) home/ home/ host) host) host) host) host) host) host) Control Firm size 0.017 Firm age 0.002 Business diversity 0.011 Financial leverage 0.003 Entry mode preference 0.091 Independent variables Global strategy Bi-region strategy R&D intensity Marketing intensity Internationalization pace Global*pace Bi-region*pace Intercept Number of observations Log likelihood Wald w2

0.027

0.013 0.013 0.000 0.001 0.009 0.001 0.112 0.075 0.016 0.086

0.021 0.000 0.020w 0.025 0.063

0.025w

0.005 0.002 0.015 0.031 0.215w

0.001 0.001 0.007 0.066 0.052

0.034

0.013

0.004 0.001w 0.036w 0.062 0.080w

0.010 0.000 0.005 0.001 0.009

0.015 0.020 0.702 0.203 0.436 0.119 0.091 0.152 0.157 0.018 0.199 0.013 0.025w 0.064

0.004 0.011 1.243 0.661 0.916 0.399 0.087 0.511 0.604 0.013 0.159 0.028 0.022 0.070

0.208

0.146

0.058 244

0.036

0.047 512

0.075 0.086 630

0.049

0.084 0.196 218

0.400

0.185

244

512

0.114

630

0.211 0.062 0.108 1.570 1.365w

0.063 0.019 0.367 0.742 0.314

0.212

0.129

1.804 1.763w 0.185

0.084 0.964 0.552 3.339 0.424 0.311 0.255

0.787

218

0.012 0.016 0.209c 1.033 0.805

Model 8.12 (biregion vs. home/ host)

0.233 0.005 0.322 0.872 3.325

0.408 5.545 2.734 0.596

0.080

0.060

0.271 0.061

Model 8.11 (biregion vs. home/ host)


Table 8.

0.383

1.020

2.776

2.991

3.074

244

512

630

2.752 218

191.28 476.73 496.56 331.71 142.27 385.00 378.44 371.77 388.31 1637.45 851.37 388.06 219.30 100.64 188.43 129.71 99.37 269.08 254.69 64.51 298.86 940.64 143.75 30.78

285


286

LEI LI AND DAN LI

geographic reach appears to be more beneficial relative to the one with a limited geographic reach for firms in the SCT industry. We will further explore this finding in the discussion section. Hypotheses 3a and 3b state that firm intangible assets enhance the effectiveness of the global strategy in an integrated global industry; and the effectiveness of regionally oriented strategies in a multidomestic industry, respectively. Tables 4 and 5 show that high R&D intensity helps enhance the effectiveness of the global strategy as opposed to the bi-regional and the single-region-oriented strategies (i.e., the impact on short-term performance such as ROA and ROS is positive and highly significant). Moreover, marketing intensity, though to a lesser extent, tends to enhance the effectiveness of the global strategy versus the single-region-oriented strategy (see Models 5.6 and 5.9 in Table 5). Thus, Hypothesis 3a is supported. Regarding Hypothesis 3b, Table 7 reports mixed findings. R&D intensity tends to positively moderate the impact of a bi-regional strategy on ROS and negatively moderate the impact of the bi-regional strategy on Tobin’s q (as opposed to the single-region-oriented strategy). On the other hand, marketing intensity does not appear to have any significant effect. Thus, Hypothesis 3b is only partially supported. Hypothesis 4a states that in an integrated global industry, R&D-based intangible assets enhance the effectiveness of the global strategy relative to regionally oriented strategies more than marketing-based intangible assets. Hypothesis 4b states that in a multidomestic industry, marketing-based intangible assets enhance the effectiveness of regionally oriented strategies relative to the global strategy more than R&D-based intangible assets. Tables 4 and 5 show that the interaction effects on ROA and ROS between R&D intensity and global strategy are mostly positive and significant whereas the interaction between marketing intensity and global strategy is only marginally significant and positive in Model 5.6, Table 5. If we use Models 5.5 and 5.6 in Table 5 as examples, an increase of 0.01 of R&D intensity will increase the moderating effect by 0.00809 whereas an increase of 0.01 of marketing intensity will enhance the moderating effect by 0.00312. Clearly, R&D-based intangible assets enhance the effectiveness of global strategy more than marketing-based intangible assets. The only exception is that the moderating effects of marketing-based intangible assets appear to be more significant than that of R&D-based intangible assets in increasing Tobin’s q of global strategy firms versus single-region-oriented strategy firms (see Model 5.9 in Table 5). We will address this point further in the discussion. Overall, Hypothesis 4a is supported. Since the interaction between marketing intensity and bi-regional strategy is not significant in Table 7, Hypothesis 4b is not supported.


287

Hypothesis 5 states that the faster the firm internationalization pace, the less effective the global strategy will be, relative to regionally oriented strategies; the effects are more salient in a multidomestic industry than in an integrated global industry. Evidently, our focus is on the interaction effect of internationalization pace with global strategy. Table 8 shows that in the COE industry, this interaction effect is mostly negative but only significant in Model 8.1 (global versus bi-regional strategy). For the SCT industry, the interaction between internationalization pace and bi-regional strategy is not significant (see Models 8.4, 8.8, and 8.12 in Table 8). In contrast with the positive and significant interaction effect between internationalization pace and bi-regional strategy in the COE industry (Models 8.3 and 8.7 in Table 8), however, the insignificant effect implies that internationalization pace compromises the effectiveness of an international strategy with relatively wide geographic reach in the SCT industry more than that in the COE industry. Thus, Hypothesis 5 is supported to a certain extent.

DISCUSSION Rugman and Verbeke (2004) show that most of the Global Fortune 500 firms are home-region oriented and only 9 MNEs are truly global firms based on the sales data. Our investigation of the U.S. MNEs in the COE industry and the SCT industry has confirmed such findings. We have further demonstrated that more MNEs can be labeled as ‘‘global’’ based on the geographical dispersion of the operations (the ‘‘inputs’’) than being based on the sales (the ‘‘outputs’’). This is also consistent with Rugman and Verbeke’s statement that ‘‘[many] large MNEs do have a strong geographical dispersion of their sourcing and production, both in resource industries and in manufacturing, but appear incapable (or unwilling) of capitalizing on this position to achieve global sales penetration’’ (2004, p. 13). Our unique contribution in this respect lies in the comparative analysis between the two vastly different industries over time. Our results indicate that firms face substantially different imperatives and/or challenges in international expansion in different industry contexts. While some global firms exist in the COE industry, there is not even a single global firm in the SCT industry (including P&G, Colgate-Palmolive, and AVON). This suggests that top managers might have perceived it to be either unnecessary or too difficult to implement a global strategy in a multidomestic industry as opposed to an integrated global industry. This difference appears to be persistent over time (see Table 3).

288

LEI LI AND DAN LI

Our further analysis (not reported here) shows that even the few biregional U.S. firms in the SCT industry operate mainly in Western European countries (besides the home region, North America), which is a regional block perceived to be economically and institutionally most similar to the United States. Even the majority (over 70%) of the U.S. bi-regional firms in the COE industry have a major presence in North America and Western Europe. The evidence suggests that there is perhaps indeed a need to pay more attention to the development and implementation of regional strategies versus global strategies in both academia and business world (Rugman, 2005; Westney, 2006). Effectiveness of Global versus Regional Strategies Our analysis shows that the global strategy per se is not effective in the sense that it neither increases a firm’s short-term performance (ROS and ROA) nor helps with the firm’s long-term performance (Tobin’s q). This finding begs the question as to whether the economies of scale are really critical and whether the standardized worldwide demand is an illusion even in an integrated global industry in which a global strategy may be perceived as a ‘‘no-brainer’’ (Kobrin, 1991; Douglas & Wind, 1987). It also makes one wonder if an industry such as the COE is over-globalized and fraught with too intense worldwide competition (Birkinshaw et al., 1995) so that even the potential advantages from global arbitrage have become less significant. Indeed, an HP supply-chain manager once commented: A new belief in supply chain theory, especially in the commoditized industry of computer and office equipment, is that multinational corporations will no longer compete against each other via marketing and sales-based promotional efforts, but rather supply chains will compete against supply chains to get the product to the customer faster, cheaper and with the most ease y . To be brutally honest, we HP supply chain guys are always ‘‘chasing’’ Dell’s supply chain practices. 2

Our findings also suggest that bi-regional strategies, with a narrower geographical reach than global strategies, might not be effective either. Although we have generated some marginal evidence that a bi-regional strategy increases firm performance (see Model 6.2 in Table 6 and Models 7.1, 7.4, and 7.8 in Table 7), we realize that this is likely a special case rather than a general application. Indeed, for the sample firms adopting a bi-regional strategy, the second region (other than the home region) is Western Europe in over 70% of the cases in the COE industry and 100% of the cases in the SCT industry. The overall economic and institutional


289

similarity between Western Europe and United States may play an important role here.

International Strategies and Firm-Specific Intangible Assets Our results confirm that firm-intangible assets help enhance the effectiveness of a global strategy in the COE industry, consistent with the internalization theory in international business literature. This suggests that the economies of scope may be still a major benefit for MNEs in integrated global industries. A somewhat puzzling finding is that the interaction between intangible assets and bi-regional strategy is negative in some models with ROA and ROS as dependent variables (see Tables 6 and 7). As noted earlier, over 70% of bi-regional strategies involve North America and Western Europe. Considering that the COE industry has been commoditizing, and the North American and Western European markets have been maturing, a bi-regional strategy involving these two regions may not be the best way to exploit the product-oriented intangible assets. Moreover, the regional similarity between North America and Western Europe makes it less likely to effectively explore and develop process-oriented intangible assets, which is critical for a maturing industry. Indeed, both HP and Dell accelerated the investment in Asia-Pacific (rather than in Western Europe) around late 1990s to improve the efficiency of their supply chains. In the case of the SCT industry, the negative impact on Tobin’s q suggests that leveraging intangible assets across North America and Western Europe are not well perceived by the stock market, though R&D-based intangible assets may be exploited profitably (in terms of short-term performance) across these two Triad regions.

R&D versus Marketing-Based Intangible Assets At a very general level, Porter (1986) argued that upstream value chain activities (compared to downstream ones) could be more easily organized at the corporate level to produce competitive advantages. Most recently, Rugman and Verbeke (2004) suggest that firm-specific advantages (FSAs) in upstream activities to achieve global sourcing (of R&D outputs, intermediate inputs, labor, and capital) and production may be more readily available for MNEs than the FSAs required in downstream activities to achieve a global distribution of sales.

290

LEI LI AND DAN LI

Our findings support these arguments to a certain extent. Regarding the globally integrated COE industry, we show that R&D-based intangible assets play a significant role in enhancing the short-term performance (ROA and ROS) of the global strategy. In fact, even if we factor in the direct negative effect of the global strategy, the overall impact on ROA and ROS are positive if the R&D intensity is above 0.104 and 0.102, respectively, in Models 4.2 and 4.5, for example. The marketing-based intangible assets tend to be effective in enhancing Tobin’s q in conjunction with the global and biregional strategies as opposed to the single-region-oriented strategy, though their impact on ROA and ROS seems to be very limited, and somewhat inconclusive. We suggest that the effect is likely related to the impact of global branding, which tends to be much valued by the shareholders. As for the multidomestic SCT industry, R&D-based intangible assets are found to have mixed impact on the effectiveness of the bi-regional strategy focusing on North America and Western Europe, while the marketing-based intangible assets appear to be not important. The above discussion extends Rugman and Verbeke’s (2004) insights about the distinction between location-bound and non-location-bound bundles of FSAs. We have found that non-location-bound intangible assets are not only available in upstream value chain activities, but also in downstream value chain activities in the integrated global industries such as the COE. However, it seems that intangible assets are mostly location (home country or home region) bound in the multidomestic industries such as the SCT. This point is even more valuable, considering that the two industries we have selected are all knowledge-intensive industries. If indeed global (or cross-regional) knowledge transferability is problematic for at least some industries, a research niche may become necessary to address firm regional strategies as a category qualitatively different than global strategies and national strategies (Westney, 2006). It is worth noting that R&D and marketing efforts are measured at the corporate level in this study; it would be promising for the future research to direct more attention to the R&D and marketing-based intangible assets that are developed at the subsidiary level and leveraged across different regions or worldwide (Katsikeas, Samiee, & Theodosiou, 2006).

International Strategies and Internationalization Pace Internationalization pace has not been widely investigated. Vermeulen and Barkema (2002) report that internationalization pace has a negative


291

moderating impact on firm international operations. The current study shows that the moderating effects vary by the industry contexts as well as the degree of geographical reach entailed by the specific international strategy. Although we confirm that internationalization pace tends to have a negative moderating impact on the effectiveness of global strategy, we find that it appears to help enhance the effectiveness of certain bi-regional strategies (as opposed to the single-region-oriented strategy) in the globally integrated COE industry. Moreover, internationalization pace has a positive direct impact on both short-term and long-term performances. This suggests that internationalization pace needs to fit with the industry context and the geographical reach of an international strategy. For a dynamic and highvelocity industry such as the COE, the product life cycle is short. Thus, fast pace in international expansion is likely to increase firm performance, especially when the firm chooses an international strategy of limited geographical reach. As for the multidomestic SCT industry, internationalization pace seems to be not significant according to Table 8. We have conducted further analyses (not reported here) by comparing the bi-regional strategy with two variants of the home-region-oriented strategy as well as the host-region-oriented strategy, respectively. The results show that internationalization pace has an overall negative impact on ROA and ROS when the home-region-oriented firms have less than 50% of their Americas-based subsidiaries in the United States.3 This has boosted our confidence to claim that the negative impact of internationalization pace is more salient in multidomestic industries, further supporting Hypothesis 5. Limitations We believe that the current study makes some unique contributions to the emerging research niche of regional strategies and operations of MNEs. Owing to several limitations, however, we fall short of producing adequate results to fully address all our research questions. First, there is not a single global firm in the SCT industry. As a consequence, we only have the indirect evidence to support our hypothesis that regional strategies are more effective than the global strategy in multidomestic industries. Second, the bi-regional firms in both industries are largely involved in North America and Western Europe. The relevant findings regarding biregional strategies may not be applicable to the bi-regional firms involving North America and Asia Pacific, or Europe and Asia Pacific, for example.

292

LEI LI AND DAN LI

Third, we do not have the data to incorporate the geographic-area-based management structure of our sample firms. The effectiveness of international strategies is closely related to implementation (e.g., centralization, coordination) (Kim et al., 2003). If the trend for MNEs is toward ‘‘think globally, act locally, and manage regionally’’ (Enright, 2005), it is critical to take into account the management structure (e.g., regionally based versus globally centralized) in future research.

CONCLUSION To the best of our knowledge, this is the first study that compares systematically the patterns of international strategies and the corresponding performance implications between two vastly different industry contexts. Consistent with and extending the insights and evidence provided by Rugman and other scholars (Rugman, 2000; Rugman & Verbeke, 2004; Delios & Beamish, 2005), we find that the choice of the global strategy is rare even in an integrated global industry and completely non-existent in a multidomestic industry; bi-regional strategies (involving two Triad regions) are not common either. On the one hand, this result suggests that researchers need to make more endeavors to investigate firm regional strategies, which are clearly more prevalent than the global strategy. On the other hand, it also indicates the importance to study how MNEs may cope with the challenging international business environments (the geovalent elements) in order to realize the potential of the global strategy (Guisinger, 2001). The analyses reveal that firm-specific intangible assets are largely nonlocation bound in the integrated global industry whereas they appear to be much home-region bound in the multidomestic industry. Moreover, R&D-based intangible assets required for upstream value chains tend to be more significant in enhancing firm short-term performance while marketingbased intangible assets (perhaps an important indicator for global branding) appears to be more important for the shareholder valuation (Tobin’s q) in the integrated global industry. Unlike some existing studies, the current one shows that internationalization pace has a direct positive impact on firm performance and rapid internationalization pace seems to work best for firms adopting bi-regional strategies (i.e., international strategies with moderate geographical reach) in the globally integrated COE industry, which happens to be a dynamic and high-velocity industry. It also reveals to a certain extent that internationalization pace has a negative impact on the performance of the bi-regional


293

firms as opposed to single-region-oriented firms in the multidomestic SCT industry.

NOTES 1. According to the original definition, the Triad regions are the U.S., Western Europe, and Japan. Rugman & Verbeke (2004) redefines the Triad as North America, Europe, and Asia-Pacific to reflect the dramatic change of world economy in the past 20 years. We extend the Triad region of North America to include Central and South American countries, and Europe to include African and Middle Eastern countries. This extension is largely consistent with the worldwide organizational structure of many large MNEs. 2. Personal communication with an HP supply-chain manager in April 2006. 3. We have repeated all the regressions by separating the single-region-oriented firms into (1) home-region-oriented with equal to or more than 50% of Americasbased subsidiaries in the United States, (2) home-region oriented with less than 50% of Americas-based subsidiaries in the United States, and (3) host-region-oriented firms (mostly Europe oriented). The results are largely consistent with those in Tables 4–8.

ACKNOWLEDGMENTS This study is financially supported by an Arthur Butine Faculty Development Grant at the University of Portland and a CIBER research award at Indiana University. We would like to acknowledge the able assistance in data collection from Aaron M. Kendrick, Dara Khowkachaporn, Amit Suri, and Cynthia Santoso.

REFERENCES Bartlett, C. A., & Ghoshal, S. (1989). Managing across borders: The transnational solution. Boston: Harvard Business School Press. Baysinger, B. D., & Hoskisson, R. E. (1989). Diversification strategy and R&D intensity in multiproduct firms. Academy of Management Journal, 32, 310–332. Berry, H. (2006). Shareholder valuation of foreign investment and expansion. Strategic Management Journal, 27, 1123–1134. Birkinshaw, J. (1997). Entrepreneurship in multinational corporations: The characteristics of subsidiary initiatives. Strategic Management Journal, 18, 207–229. Birkinshaw, J., & Hood, N. (1998). Multinational subsidiary evolution: Capacity and character change in foreign-owned subsidiary companies. Academy of Management Review, 23, 273–285.

294

LEI LI AND DAN LI

Birkinshaw, J., Morrison, A., & Hulland, J. (1995). Structural and competitive determinants of a global integration strategy. Strategic Management Journal, 16, 637–655. Buckley, P., & Casson, M. (1976). The future of the multinational enterprises. New York: Holmes & Meier Publishers, Inc. Caves, R. E. (1971). International corporations: The industrial economics of foreign investment. Econometrica, 38, 1–27. Caves, R. E. (1996). Multinational enterprise and economic analysis (2nd ed.). Cambridge, UK: Cambridge University Press. Child, J., Chung, L., & Davies, H. (2003). The performance of cross-border units in China: A test of natural selection, strategic choice and contingency theories. Journal of International Business Studies, 34, 242–254. Chung, K. H., & Pruitt, S. W. (1994). A simple approximation of Tobin’s Q. Financial Management, 23(3), 70–74. Contractor, F. J., Kundu, S. K., & Hsu, C-C. (2003). A three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34(1), 5–18. Delios, A., & Beamish, P. W. (2005). Regional and global strategies of Japanese firms. Management International Review, 45(1), 19–36. Dierickx, I., & Cool, K. (1989). Asset stock accumulation and sustainability of competitive advantage. Management Science, 35, 1504–1511. Douglas, S. P., & Wind, Y. (1987). The myth of globalization. Columbia Journal of World Business, 22(4), 19–29. Dunning, J. H. (1993). Multinational enterprises and the global economy. Wokingham, UK: Addison-Wesley Publishers, Ltd. Enright, M. J. (2005). Regional management centers in the Asia-Pacific. Management International Review, 45(1), 59–82. Ghemawat, P. (2003). Semiglobalization and international business strategy. Journal of International Business Studies, 34, 138–152. Ghoshal, S. (1987). Global strategy: An organizing framework. Strategic Management Journal, 8, 425–440. Goerzen, A., & Beamish, P. (2003). Geographic scope and multinational enterprise performance. Strategic Management Journal, 24, 1289–1306. Grant, R. M. (2005). Contemporary strategy analysis (5th ed.). Malden, MA: Blackwell Publishing. Grosse, R. (2005). Are the largest financial institutions really ‘global’? Management International Review, 45(1), 129–144. Guisinger, S. (2001). From OLI to OLMA: Incorporating higher levels of environmental and structural complexity into eclectic paradigm. International Journal of the Economics of Business, 8(2), 257–272. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management Journal, 40, 767–798. Hofstede, G. (1980). Culture’s consequences. Beverly Hills, CA: Sage Publications. Hoskisson, R. E., Johnson, R. A., & Moesel, D. D. (1994). Corporate divestiture intensity in restructuring firms: Effects of governance, strategy, and performance. Academy of Management Journal, 37, 1207–1251. Hout, T., Porter, M. E., & Rudden, E. (1982). How global companies win out. Harvard Business Review, 60, 98–108.


295

Katsikeas, C. S., Samiee, S., & Theodosiou, M. (2006). Strategic fit and performance consequences of international marketing standardization. Strategic Management Journal, 27(9), 867–890. Kim, W. C., Hwang, P., & Burgers, W. P. (1989). Global diversification strategy and corporate profit performance. Strategic Management Journal, 10, 45–57. Kim, K., Park, J. H., & Prescott, J. E. (2003). The global integration of business functions: A study of multinational businesses in integrated global industries. Journal of International Business Studies, 34, 327–344. Kobrin, S. (1991). An empirical analysis of the determinants of global integration. Strategic Management Journal, 12, 17–31. Kogut, B. (1985). Designing global strategies: Profiting from operational flexibility. Sloan Management Review, 27(1), 27–38. Kogut, B., & Singh, H. (1988). The effect of national culture on the choice of entry mode. Journal of International Business Studies, 19(3), 411–432. Kogut, B., & Zander, U. (1993). Knowledge of the firm and the evolutionary theory of the multinational corporation. Journal of International Business Studies, 24(4), 625–645. Kotabe, M., Srinivasan, S. S., & Aulakh, P. S. (2002). Multinationality and firm performance: The moderating role of R&D and marketing capabilities. Journal of International Business Studies, 33(1), 79–97. Lawrence, P. R., & Lorsch, J. W. (1967). Differentiation and integration in complex organizations. Administrative Science Quarterly, 12, 1–47. Levitt, T. (1983). The globalization of markets. Harvard Business Review, 61, 92–102. Li, L. (2003). Liability of internationalization: Technical efficiency-based evidence. The 63rd Academy of Management Best Paper Proceedings, L1–L6. Li, L. (2005). Is regional strategy more effective than global strategy in the US service industries? Management International Review, 45(1), 37–57. Lu, J. W., & Beamish, P. W. (2001). The internationalization and performance of SMEs. Strategic Management Journal, 22, 565–586. Lu, J. W., & Beamish, P. W. (2004). International diversification and firm performance: The S-curve hypothesis. Academy of Management Journal, 47(4), 598–609. Makhija, M. V., Kim, K., & Williamson, S. D. (1997). Measuring globalization of industries using a national industry approach: Empirical evidence across five countries and over time. Journal of International Business Studies, 28(4), 679–710. Morck, R., & Yeung, B. (1991). Why investors value multinationality? Journal of Business, 64, 165–187. Morrison, A. J., & Roth, K. (1992). A taxonomy of business-level strategies in global industries. Strategic Management Journal, 13, 399–418. Morrison, A. J., Ricks, D. A., & Roth, K. (1991). Globalization versus regionalization: Which way for the multinational? Organizational Dynamics, 19(3), 17–29. Nohria, N., & Ghoshal, S. (1997). The differentiated network: Organizing multinational corporations for value creation. San Francisco, CA: Jossey-Bass. Palepu, K. (1985). Diversification strategy, profit performance and the entropy measure. Strategic Management Journal, 6, 239–255. Pantzalis, C. (2001). Does location matter? An empirical analysis of geographic scope and MNC market valuation. Journal of International Business Studies, 32(1), 133–155. Porter, M. (1986). Competition in global industries: A conceptual framework. In: M. Porter (Ed.), Competition in global industries (pp. 5–60). Boston, MA: Harvard Business School Press.

296

LEI LI AND DAN LI

Prahalad, C. K., & Doz, Y. L. (1987). The multinational mission: Balancing local demands and global vision. New York: Free Press. Rugman, A. M. (1981). Inside the multinationals: The economics of internal markets. London: Croom Helm. Rugman, A. M. (2000). The end of globalization. New York: AMACOM. Rugman, A. M. (2005). The regional Multinationals: MNEs and ‘global’ strategic management. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Hodgetts, R. (2001). The end of global strategy. European Management Journal, 19(4), 333–343. Rugman, A. M., & Verbeke, A. (1992). A note on the transnational solution and the transaction cost theory of multinational strategic management. Journal of International Business Studies, 23(4), 761–772. Rugman, A. M., & Verbeke, A. (2001). Subsidiary-specific advantages in multinational enterprises. Strategic Management Journal, 22, 237–250. Rugman, A. M., & Verbeke, A. (2003). Extending the theory of the multinational enterprises: Internalization theory and strategic management perspectives. Journal of International Business Studies, 34(2), 125–137. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35, 3–18. Sarkar, M. B., Cavusgil, T. S., & Aulakh, P. S. (1999). International expansion of telecommunication carriers: The influence of market structure, network characteristics, and entry imperfections. Journal of International Business Studies, 30, 361–381. Severn, A. K., & Laurence, M. M. (1974). Direct investment, research intensity, and profitability. Journal of Financial and Quantitative Analysis, 9(2), 181–190. Shan, W., & Song, J. (1997). Foreign direct investment and the sourcing of technological advantage: Evidence from the biotechnology industry. Journal of International Business Studies, 28, 267–283. Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practice within the firm. Strategic Management Journal, 17, 27–43. Tallman, S., & Li, J. T. (1996). Effects of international diversity and product diversity on the performance of multinational firms. Academy of Management Journal, 39, 179–196. Teece, D. J. (1986). Profiting from technological innovation. Research Policy, 15, 285–306. Thompson, J. D. (1967). Organizations in action. New York: McGraw-Hill. Tsai, W. (2000). Social capital, strategic relatedness, and the formation of intra-organizational strategic linkages. Strategic Management Journal, 21, 925–939. Vermeulen, F., & Barkema, H. (2002). Pace, rhythm, and scope: Process dependence in building a profitable multinational corporation. Strategic Management Journal, 23, 637–654. Vernon, R. (1966). International investment and international trade in the product cycle. Quarterly Journal of Economics, 80, 190–207. Westney, D. E. (2006). Book review on ‘‘The Regional Multinationals: MNEs and ‘‘Global’’ Strategic Management’’. Journal of International Business Studies, 37, 445–449. Yip, G. (1992). Total global strategy: Managing for worldwide competitive advantage. Englewood Cliffs, NJ: Prentice-Hall.

THE REGIONAL DIMENSION OF UK MULTINATIONALS Alan M. Rugman, Alina Kudina and George S. Yip ABSTRACT As multinational enterprises (MNEs) expand internationally (for example, as the ratio of foreign (F) to total (T) sales increases), there is a positive effect on firm performance (usually measured by return on total assets (ROTA). We advance this literature in three ways: (i) we focus on the recent performance of UK MNEs, in terms of ROTA, but also in terms of their return on foreign assets (ROFA); (ii) to supplement (F/T), we examine the ratio of European (E) to total (T) sales of these UK MNEs; and (iii) we test the relationship between (E/T) and both ROFA and ROTA, and find a significant non-linear fit.

INTRODUCTION The locational (or geographic) decisions of multinational enterprises (MNEs) is one of the most extensively researched areas of international business. Many scholars have investigated factors which make an MNE choose one investment location over another (Aharoni, 1966; Green & Smith, 1972; Dunning, 1980; Nigh, 1985; Loree & Guisinger, 1995). Locational strategy has been traditionally studied by economists in terms of

Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 297– 315 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13012-6

297

298

ALAN M. RUGMAN ET AL.

trade theory at the level of individual countries. In the field of international business, a number of scholars have emphasized that firm-level analysis of international business activity by MNEs can be undertaken (Rugman, 1981; Dunning, 1981; Porter, 1990; Bartlett, Doz, & Hedlund, 1990; Yip, 2003). Indeed, the geographic locational decisions of MNEs can be considered in a broader system (network) context (Nohria & Ghoshal, 1997), rather than as a collection of individual (country-level) decisions. Furthermore, the recent evidence on regional activities of MNEs provided by Rugman (2000, 2005) suggests that the home-triad region (as opposed to a national or global view) is the most appropriate level of analysis for the activities of MNEs. The great majority of the Fortune Global 500 firms average over 72% of their sales in their home region, with only a very few of them qualifying as global firms. Hence, examination of the relative merits of regional strategy (as opposed to country-level or global strategy) is of great importance to both managers and academics. Owing to data limitations there are some untested aspects of the regionalization hypothesis. The work by the advocates of regional strategy provides evidence on the geographic location of revenues (Rugman & Girod, 2003; Rugman & Verbeke, 2004; Rugman, 2005). Data on the production side (foreign assets) and value added are generally not available for most of the 500 largest MNEs. At the same time, business strategy researchers now define global strategy less narrowly than Levitt (1983) who argued that the same product could be sold in the same way across the world, a situation of perfect homogeneity. Today, most strategy scholars accept that there is heterogeneity across geographic space and that the world lacks commonality in its markets for goods and services. For example, Yip (2003, p. 1) posits: ‘‘A strategy is global to the extent that it is integrated across countries. Global strategy should not be equated with any one element – standardized products, worldwide market coverage, or a global manufacturing network. Global strategy should, instead, be a flexible combination of many elements.’’ Using this broad definition any international operations of the MNE can be thought of as either ‘‘global’’ or ‘‘regional’’. Indeed, much of the strategy research in international business has examined ‘‘global’’ strategy when only ‘‘international’’ (i.e., cross-border) strategy is really being examined. As a result, the analysis of the global (not regional) strategies of MNEs has been dominant in the business literature to date. In particular, there is a thick body of research analyzing the impact of an MNE’s international activities on its financial performance (Tallman & Li, 1996; Hitt, Hoskisson, & Kim, 1997; Contractor, Kundu, & Hsu, 2003). These studies use data on the geographic source of revenues as a measure of

The Regional Dimension of UK Multinationals

299

the multinationality of a business (Geringer, Beamish, & Da Costa, 1989; Sullivan, 1994; Tallman & Li, 1996; Reeb, Kwok, & Bayek, 1998). Yet, these and related studies yield inconsistent results on the relationship between multinationality and performance (Grant, 1987; Ruigrok & Wagner, 2002; Gomes & Ramaswamy, 1999; Contractor et al., 2003). While there are many potential explanations for this inconsistency (Grant, 1987; Hitt et al., 1997; Goerzen & Beamish, 2003; Contractor et al., 2003), one important aspect is the extent to which scholars focus on country, global, or regional levels of analysis. As a result, we specifically address in this chapter this new issue of the regional level of analysis. More precisely, this is the first chapter that explicitly takes into account and measures the underlying geography of the international business phenomenon of the MNE and the impact of regional location on firm-level performance. By putting geographic lenses on first, we are able to accurately connect the empirical evidence and existing theoretical frameworks to conduct an analysis of the international (in this case regional) strategy of multinational companies and their performances. As a result, this chapter addresses two highly acute international business issues: (1) the locational (geographic) positioning of MNEs and (2) the performance implications of their regional strategies. We conduct our analysis for a set of the largest UK multinationals. The chapter starts with an analysis of the evidence related to the international activities of UK firms in the Fortune Global 500. Then, we relate the evidence on the regional nature of the UK companies to the regional scope of the world’s largest MNEs to show that the described pattern is not unique to the UK. We further proceed by testing the effect of regional strategy on the performance of the UK multinationals. A discussion of our findings and some concluding remarks complete this chapter.

THE REGIONAL DIMENSION OF UK MULTINATIONALS One method of assessing the international competitiveness of British business is to study the performance and international geographic presence of the larger UK MNEs. As a preliminary insight into this, let us first consider the 27 British MNEs listed in the Fortune Global 500 ranking of the world’s largest firms (by revenues in 2001), as analyzed by Rugman (2005). Of these 27 MNEs, only 8 are in the manufacturing sector whereas 19 are in services.

300


It is important to position the 27 largest British MNEs in relation to rival firms in the rest of the Europe, North America, and Asia. These three regions represent the ‘‘triad’’ of economic activity which dominates international business (Rugman, 2000). The world’s largest 500 firms account for over 90% of the world’s stock of foreign direct investment and about half of the world’s trade. Despite evidence that most of them operate largely within their home regions (the 380 firms average 71% home-region sales), many of these MNEs state that they have a global strategy (Yip, 2003; Govindarajan & Gupta, 2001). Before examining this issue, we examine the basic empirical evidence on the international business activities of the world’s largest MNEs. In an influential study, Rugman and Verbeke (2004) report that of the world’s 500 largest firms, only 9 can be classified as global whereas 320 of the 380 firms reporting regional sales have an average of 80% of their sales in their home region. There are also some 36 bi-regional firms. These data are reclassified in Table 1. This lists the 380 firms by their home country. There are 169 from the United States, 66 from Japan, and 119 from Europe, of which 27 are from Britain. The average intra-regional sales of the U.S. firms is 77.3%; for the Japanese it is 74.7%; and for the European it is 62.8%. Within the European group, the British firms have intra-regional sales of 64.5%. This suggests that the British (and European) firms may be somewhat less regionally based than U.S. and Japanese firms. The British firms are in the mid range of European firms, with the most ‘‘global’’ firms being the five from the Netherlands. Table 2 reports the classifications of the 27 British MNEs, based on the Rugman and Verbeke (2004) and Rugman (2005) methodology. As with most of the world’s other large firms, 19 of them are home-region oriented. However, 4 are bi-regional, and 2 are host-region oriented (AstraZeneca & Wolseley) where these latter 2 firms have over 50% of their sales in North America. The four bi-regionals are also big players in North America, in particular, BP, GlaxoSmithKline, and Diageo. Yet many of the service-based firms are very home-region based: Tesco has 94% of its sales in Europe; Vodafone has 93%; Centrica has 94%; Abbey National has 99.5%; Kingfisher has 95%; Safeway has 100%, as does Alliance Unichem. These data suggest that most large British service MNEs do most of their business within Britain and the EU. They are poor prospects for globalization. Why is this? What are the strategies, structures, processes, and internal managerial factors that explain this lack of global sales? Are there any public policy factors such as government regulations that have an impact on these location-bound firms? But perhaps it does not really matter


Table 1. Country

United States Japan Germany France Britain Canada Switzerland Italy Australia Sweden Netherlands European bi-nationala Norway South Korea Belgium Finland Spain Taiwan Luxembourg Denmark Brazil Singapore Total

301

The World’s Largest 500 Firms by Country. Number of Firms

Average Revenues (US$ billion)

Average IntraRegional Sales (%)a

169 66 29 27 27 16 8 5 5 5 5 3 2 2 2 2 2 1 1 1 1 1 380

30.3 28.9 37.3 27.2 25.3 13.5 34.7 38.7 13.6 16.4 42.1 73.9 21.6 26.3 18.8 20.0 29.1 11.6 13.0 10.9 24.5 13.1 29.2

77.3 74.7 68.1 64.8 64.5 74.1 49.6 83.4 71.4 54.3 39.1 47.9 83.0 71.2 58.4 55.1 50.3 100.0 95.0 94.3 88.0 22.4 71.9

Notes: Data are for 2001. Numbers might not add up due to rounding. There are 120 firms in the world’s largest 500 which report no data in regional sales. For further information on the data and definitions used, please see the Chapter 2 in Alan M. Rugman’s The Regional Multinationals (Cambridge University Press, 2005). a Average intra-regional sales are by the firm’s size according to weighted revenues.

as other large firms from Asia and North America are also very insular and home-region based. Are there any industry effects that matter in explaining the competitiveness of British business? The preliminary answer is yes – there is a major difference between the manufacturing and service sectors, as can be seen in Table 3. In Table 3, the 27 British MNEs fall into two classifications, with only 8 in manufacturing but 19 in services. The eight British manufacturing firms have lower intra-regional sales than for other European or world firms. This indicates that these eight British manufacturing MNEs are potentially more

302


Table 2. The Regional Nature of UK Multinational Enterprises. 500 Rank

Bi-regional 4 140 262 390

Company

BP GlaxoSmithKline Diageo BAE Systems

Host-region oriented 301 AstraZeneca 487 Wolseley Home-region oriented 114 Tesco 115 Royal Bank of Scotland 123 Vodafone 139 BT (q) 150 HBOS 154 Barclays 184 J. Sainsbury 206 Lloyds TSB Group 222 Royal & Sun Alliance 270 Centrica 280 Abbey National 314 Kingfisher (q) 409 Compass Group 418 Safeway 428 British Airways 439 Marks & Spencer 452 Corus Group 453 Old Mutual (q) 478 Alliance Unichem

Revenues (US$ billion)

Percentage of Total North America

Europe

Asia Pacific

174.2 29.5 18.6 13.0

48.1 49.2 49.9 32.3

36.3 28.6 31.8 38.1

NA NA 7.7 2.7

16.5 10.4

52.8 66.3

32.0 28.7

5.2 NA

33.9 33.8

12.0

93.6 81.0

6.4 NA

32.7 30.0 27.8 27.6 24.6 22.8 21.5

0.1 8.3 NA 6.0 16.7 NA 27.1

93.1 87.0 92.1 88.0 83.3 81.2 64.8

4.8 4.7 NA NA – NA NA

18.2 17.8 16.1 12.6 12.3 11.9 11.6 11.1 11.1 10.5

6.2 0.5 0.8 32.4 NA 18.6 NA 11.5 NA –

93.8 99.5 98.3 67.6 100.0 64.8 85.1 82.7 NA 100.0

NA – 0.6 – NA NA NA 5.8 NA –

Insufficient information 271 British American Tobacco (q) 341 Anglo American

18.1

NA

31.3

9.9

14.8

18.9

46.1

17.8

Weighted average

25.3

Notes: Data are for 2001. For further information on the data and definitions used please see the appendix and company notes in Alan M. Rugman’s The Regional Multinationals (Cambridge University Press, 2005).


303

Table 3. The Regional Nature of British, European, and World Firms. Industry

Manufacturing Aerospace and defense Chemicals and pharmaceuticals Computer, office, and electronics Construction, building materials, and glass Energy, petroleum, and refining Food, drug, and tobacco Motor vehicle and parts Natural resource manufacturing Other manufacturing Services Banks Entertainment, printing, and publishing Merchandisers Telecommunications and utilities Transportation services Other financial services Other services Total

Britain

Europe

Number of firms

Average intraregional sales

1

38.1

2

Number of firms

World


Number of firms


2

42.7

11

66.3

29.8

7

37.6

18

56.5

0

NA

6

49.4

36

56.2

0

NA

5

60.6

11

73.5

1

36.3

7

53.8

31

66.0

2

31.6

5

36.4

14

55.0

0

NA

8

54.4

29

60.6

2

61.8

6

71.8

17

77.6

0

NA

5

54.0

13

57.8

5 0

87.4 NA

23 3

75.4 67.2

40 9

78.3 73.1

6 3

92.4 91.0

15 11

75.5 82.8

63 27

87.9 87.6

1

64.8

4

73.9

13

83.7

2

74.5

8

62.9

27

71.9

2

50.0

4

51.5

21

75.8

27

64.5

119

62.8

380

71.9

Notes: Data are for 2001. Numbers might not add up due to rounding. Average intra-regional sales are by the firm’s size according to weighted revenues. For further information on the data and definitions used please see the appendix and company notes in Alan M. Rugman’s The Regional Multinationals (Cambridge University Press, 2005).

304


‘‘global’’ than other large firms. In contrast, in all but two of the service sectors, the British firms have higher intra-regional sales than others. This indicates that these firms are much less global than others. Are there any industry-specific effects that have led to such location-based activity? In summary, this preliminary analysis suggests a lack of globalization for all these large British firms in Table 3. As these firms are often ‘‘flagships,’’ at the hubs of business clusters, it is likely that other businesses, especially small- and medium-size businesses, are even less global and more location bound (Rugman & D’Cruz, 2000), than frequently thought. Before moving onto report a much more detailed econometric analysis of the performance of a larger set of UK companies, we shall further relate the regional nature of UK companies to the regional scope of all the world’s largest MNEs. While there is nothing special or unique about the regional nature of UK business, it is necessary to show the relevance of the empirical work reported later by first establishing the general nature of intra-regional activity for the world’s largest firms.

THE WORLD OF REGIONAL MULTINATIONALS: THEORY AND HYPOTHESES The definition of an MNE is that it produces and/or distributes products and/or services across national borders. These MNEs have repeatedly been identified as the drivers of globalization, pursuing geographic integration and coordination through aspects of global strategy, such as global market participation, global products, global activity location, global marketing, and global competitive moves (Yip, 2003). Advantages of global strategy include reduced costs from product standardization, reduced duplication of activities (and relocation of activities to lower cost sites), enhanced customer preference from global uniformity and consistency, and greater global competitive advantage. But to fully achieve these benefits of global strategy, an MNE needs both globally dispersed sales and globally dispersed production (and other activities). Nevertheless, an MNE can achieve some of the advantages of global strategy, whatever its geographic scope. So an MNE can pursue a global strategy even within a single region. We define a regional strategy as one that seeks the benefits of geographic integration and coordination within a single region or within a collection of regions. Indeed, such an integration and coordination strategy is easier to pursue within a single region, especially if the home-region government


305

pursues policies of an internal market such as social, cultural, and political harmonization (as in the EU) or economic integration (as in NAFTA and Asia). As firms expand outside their home regions, they then face a greater liability of foreignness and other additional risks by this global expansion, to offset the additional benefits of greater geographic scope. Furthermore, inter-block business is likely to be restricted by government-imposed barriers to entry. The EU and the United States are now fighting trade wars and are responsive to domestic business lobbies seeking shelter in the form of subsidies and/or protection, as in the case of the steel and agricultural sectors. There will remain cultural and political differences between members of the triad, but there will be fewer of these within each triad block. Rugman (2000) argues for a trend, over the last quarter century, toward regionalization and increased intra-regional economic activity. Only in a few sectors, such as consumer electronics, is a global strategy of economic integration viable. For most other manufacturing (such as automobiles, chemicals, energy, etc.), and for all services (such as retail, banking, etc.), regional strategies are required. On the other hand, there is counter-evidence that companies who do implement global strategies perform better than those who focus on regional strategies, at least in the case of the automobile industry (Schlie & Yip, 2000). Rugman and Verbeke (2004) examined the triad/regional economic activity of the world’s largest firms in the core triad of the United States, EU, and Japan. In 2001, of the world’s largest 500 firms, 428 were in these core triad regions, whereas back in 1981 it was 445. Data were found for 380 firms reporting geographic sales distributions, based on information in the annual reports and web pages. The 380 firms were then classified according to those that are global, bi-regional, and domestic. There were no data for 120 firms (most of which are entirely domestic) and insufficient data for another 15 firms. The main results show that the intra-regional average sales for each group was

80.3% 42.0% 30.9% 38.3%

for for for for

the the the the

320 home region oriented firms; 25 bi-regional MNEs; 11 host-region ranked firms; and 9 global MNEs.

These data confirmed the study of the 49 retail MNEs in the 500, in Rugman and Girod (2003). In that study, only one retail MNE was found to be global, namely Moet Hennessey Louis Vuitton (LVMH). This result is evident across all industry sectors except for electronics, which includes 7 of the 9 global firms in the set.

306


It is possible that the upstream ‘‘back end’’ production of the value chain is more globalized than the downstream ‘‘front end’’ of sales. However, even there it was found that regionally based production clusters and networks, similar to the automobile sector are the norm (Rugman, 2000, 2005). Based on the above analysis, regional strategy is hypothesized to have a direct impact on the performance of MNEs. However, we suggest that this relationship is not linear. Regional strategy contributes negatively to MNE performance at early stages of regionalization due to the costs associated with any international expansion. However, once an MNE achieves a certain level of familiarity with the whole region, further expansion contributes positively to its performance due to the learning benefits of multinationality, most of which can be achieved within the home region. We shall use data on the ROTA to measure the overall performance of the MNE, as explained later in the data section. Hypothesis 1. The degree of regionalization has a negative impact on the overall performance (ROTA) of an MNE at early stages of regionalization, but this changes to a positive one once an MNE has familiarized itself with the method of conducting business within the region, leading to learning benefits. However, a different picture emerges if we consider not only the overall performance of MNEs, but also the more specific performance of its foreign subsidiaries. This is probably a more accurate way to measure the success of an MNE’s internationalization efforts as a significant home-country effect may distort the ROTA data on overall performance. In this case, an MNE will be facing a trade-off between benefits/costs of regionalization and globalization, or, in other words potential performance differentials among its subsidiaries in the home region and other parts of the world. We can address this issue by obtaining data on the ROFA. Such data have become available only in recent years for most MNEs. Here, we suggest that early to medium stages of regionalization will have a positive effect on the performance of the foreign subsidiaries (ROFA) of the MNE. This occurs as familiarity in conducting business within the region will outweigh the added costs of expansion beyond the region, assuming that an MNE first expands within its home region and then goes to other regions in line with the thinking of Rugman (2005) and the Uppsala School. However, once the home-region market is relatively saturated and the MNE starts expanding into other regions, additional sales within the home region are no longer contributing as much to the performance of the MNE’s foreign


307

operations. Then ROFA starts to decrease as the home region’s subsidiaries become relatively less profitable than the extra-regional ones. Hypothesis 2. The international expansion of an MNE in its home region has a positive impact on the performance of the foreign operations (ROFA) of the MNE at early to medium stages of regionalization, but ROFA changes to a negative one at higher stages of regionalization.

EMPIRICAL TESTING Data and Sampling In this section of the chapter, we examine whether the benefits of regional strategy are applicable to the case of UK based MNEs. In particular, we are investigating a link between regional strategy and financial performance for a set of over 210 UK MNEs. We find their sales in the ‘‘European’’ (E) region and compare that to their total (T) sales, yielding (E/T). Then, we test their performance against this degree of multinationality (E/T). This is the first study to test the regional (E/T) type of measure for any set of firms. As performance measures, we use return on foreign assets (ROFA). This is also one of the first studies to test for ROFA, which is calculated as a ratio of profits earned overseas to the amount of company’s assets located overseas. We use ROFA as it measures the performance of the subsidiaries of the MNE, which is our primary concern, and set this against a ‘regional’ measure of (E/T), for the first time. Usually, ROTA is used, or another firm-level metric, such as return on sales, as in Yip, Rugman, and Kudina (2007). We also regress ROTA against (E/T), and compare the difference between ROFA and ROTA as dependent variables. In line with previous research, ROTA is measured as the ratio of a company’s net income to the amount of its total assets. We use the OSIRIS database, which contains the annual reports of over 30,000 public and private companies from all over the world. From this database, we select the UK companies present in the top 100 across 89 industry sectors (an exhaustive list) as classified by Dow Jones (DJ). The DJ system has the advantage, relative to the U.S. Standard International Classification (SIC) system, of defining industries to fit the actual distribution of major MNEs from around the world. Hence this DJ system minimizes the problem that arises from industry diversification. Our selection method yields a total of 587 UK companies from the European total of 8,900 in the OSIRIS database. However, not all of these UK companies report the

308


necessary regional segment data, so we found only 210 UK companies for which we are able to calculate both the regionalization measure (E/T), and the return on foreign assets (ROFA) over time. These companies constitute our sample (those companies that reported the data for only one of the years under consideration were excluded from the sample). As the internationalization process is a longitudinal phenomenon, we are looking into its evolution over a 10-year period. Accordingly, we collected the data for four years: 1993, 1998, 2001, and 2003, thus creating an unbalanced panel of 495 observations, as data for some years were missing for some companies.

VARIABLES As was mentioned above, we use two measures of performance as dependent variables in this study: (1) return on foreign assets (ROFA) that reflects the performance of an MNE’s foreign operations (that is, its subsidiaries) and (2) ROTA – a traditional measure of overall performance used in studies, for example, by Hitt et al. (1997), Gomes and Ramaswamy (1999), and Contractor et al. (2003), amongst others. The central independent variable in this analysis is a measure of regionalization (E/T), which is a ratio of sales in a home region (Europe) to the total sales of a multinational enterprise (Rugman, 2005). Hence, this measure includes home-country (UK) sales, subsidiaries’ sales in the region (Europe), and the UK firm’s exports to the rest of Europe. Inclusion of such export data is partly governed by data availability, but it is also appropriate as exporting obviously gives foreign sales. Both exports and foreign subsidiary sales are components of foreign sales. According to internalization theory, an internationalizing firm chooses the most cost-efficient way among all possible modes of foreign sales (Buckley & Casson, 1976; Rugman, 1981). In this study we use a set of control variables which is traditionally used in the literature to moderate the effects of other factors which have been shown to have a significant impact on a firm’s performance according to earlier research. First, we control for a firm’s size effect by including a logarithm of company’s total revenues into the regressors (Tallman & Li, 1996; Hitt et al., 1997; Gomes & Ramaswamy, 1999). Second, recognizing the importance of industry effects in explaining firm performance in a multi-industry study (Schmalensee, 1985; Grant, Jammine, & Thomas, 1988; Montgomery & Porter, 1991; Tallman & Li, 1996), we use the corresponding DJ industry’s average return on assets


309

measured for all public companies in the world in a particular industry (as in, for example, Goerzen & Beamish, 2003). Third, we include a time trend to mediate a year or temporal effect in profitability data, as the former was shown to have a significant effect in previous research (Cowley, 1988; Mascarenhas & Aaker, 1989; Haskel & Martin, 1992; Li, 2005).

METHODOLOGY To test the two hypotheses outlined in the earlier section we estimate the following two major specifications: S1 : ROFAit ¼ b0 þ b1 ðE=Tit Þ þ b2 ðE=Tit Þ2 þ b3 ðE=Tit Þ3 X þ bj ControlVariableijt þ it j

S2 : ROTAit ¼ b0 þ b1 ðE=Tit Þ þ b2 ðE=Tit Þ2 þ b3 ðE=Tit Þ3 X þ bj Control Variableijt þ it j

where ROFAit is percentage return on foreign assets of company i in year t; ROTAit is percentage return on total assets of company i in year t; E/Tit is a ratio of European sales to total sales of company i in year t; Control Variableijt is a vector of control variable j for a company i (or industry k) in year t (which consists of log(TRit) is a natural logarithm of total revenues (in U.S.$billion) of company i in year t, ROTAWkt – an average percentage return on assets of all public companies in the world of industry k in year t; YEARt – a time trend); and eit is a corresponding error term. Previous researchers have found various functional forms for the degree of multinationality and performance relationship (for example, linear (Grant, 1987; Jung, 1991), quadratic (Geringer et al., 1989; Sullivan, 1994; Ramaswamy, 1995) and cubic (Contractor et al., 2003; Lu & Beamish, 2004). Consequently, although the two hypotheses developed earlier explicitly refer to U-shaped and inverted U-shaped relationships, we have decided to test for various functional fits between the regionalization variable and MNE performance (starting with linear, we add quadratic and cubic terms at later stages). Since the time series is short in our dataset, estimation of either fixed or random effect models does not seem feasible. Therefore, in this chapter we use a Feasible Generalized Least Squares (FGLS) estimator, which is more

310


efficient than a pooled OLS estimator when the series exhibit heteroskedasticity (that is a concern in the analyzed dataset). By downweighting estimated coefficients by an estimate of the cross-sectional residual standard deviation, FGLS allows assigning a smaller weight to observations coming from populations with greater variance and a larger weight to the ones coming from populations with smaller variance. In this way cross-sectional heteroskedasticity is addressed. Additionally, we use White heteroskedasticity consisted covariances to obtain estimates which are robust to general heteroskedasticity; that is, we allow for inter-temporal differences in variances along with cross-sectional heteroskedasticity, which is traditionally addressed by FGLS. A similar estimation technique was used in studies by Gomes and Ramaswamy (1999), Contractor et al. (2003), and Li (2005).

RESULTS Table 4 shows the results of our analysis. Specifications 2 and 5 are directly related to the two hypotheses we developed earlier. Let us mention at the outset that both of them are confirmed, and we now continue with a more detailed discussion of the results. First, let us analyze the relationship between ROFA and the regionalization variables. We can see that both linear and quadratic terms are significant in the dataset under consideration. The overall fit (as shown by the F-statistics) is significantly better for a quadratic specification, which consequently becomes our preferred specification. We also find significant size, industry, and time effects. This analysis demonstrates that regional sales have a significant association with the financial performance of the subsidiaries of the UK MNEs. However, the relationship is not linear. In line with Hypothesis 2, smaller regionalization levels have a positive effect on the performance of foreign operations. In contrast, a high level of regional sales has a negative effect on the subsidiaries’ performance. Overall, this analysis shows that regional sales have a significant impact on performance of the international operations of UK MNEs. Having analyzed the impact of regional strategy on the foreign performance of an MNE (ROFA), let us contrast it with the relationship between ROTA and regional sales, a test of Hypothesis 1. As we can see from Table 4, in the case of ROTA, a cubic fit (or S-curve) is also significant as well as linear and quadratic. Yet, interestingly, the overall fit for linear regression is significantly better than that for other specifications. Judging by


Table 4. Dependent Variable Independent Variables

311

Regional Sales and the Performance of British Companies. Return on Foreign Assets (ROFA)

S1

S2

European/total 0.03*** 0.095*** sales (0.000) (0.000) European/total 0.001*** sales2 (0.000) European/total sales3 Total revenues 0.682*** 0.560*** (log) (0.000) (0.000) Industry return 0.332*** 0.326*** on total assets (0.000) (0.000) (world) Time trend 1.293*** 1.149*** (0.000) (0.000) Constant term 20.126*** 17.002*** (0.000) (0.000) Adjusted R2 0.682 0.689 F-statistics 264*** 5333*** Number of 495 495 observations

Return on Total Assets (ROTA)

S3

S4

0.113 (0.134) 0.001 (0.361) 0.000002 (0.834) 0.491*** (0.000) 0.288*** (0.000)

0.012*** (0.000)

1.124*** (0.000) 16.193*** (0.000) 0.643 147*** 495

0.716*** (0.000) 0.320*** (0.000)

S5

S6

0.028*** 0.147*** (0.000) (0.000) 0.0003*** 0.003*** (0.000) (0.000) 0.00002*** (0.000) 0.729*** 0.633*** (0.000) (0.000) 0.347*** 0.302*** (0.000) (0.000)

0.950*** 1.058*** (0.000) (0.000) 3.804*** 4.785*** (0.000) (0.000) 0.995 0.926 23061*** 1247*** 495 495

0.929*** (0.000) 15.607*** (0.000) 0.754 249*** 495

Note: p-values in parentheses. Significant at 1% level.

F-statistics, the linear specification has more explanatory power than the quadratic specification. The cubic specification’s F-statistic is lagging behind the former two, yet the overall fit for Hypothesis 1 is still significant. Hence, although the linear relationship’s fit is considerably better, we cannot neglect the other specifications, which are also significant. An interesting finding is that despite exhibiting an inverted U-shaped relationship in the ROFA model, the regionalization variable reveals a significant U-shaped relationship for the corresponding ROTA model. This difference in fit, with an inverted U-shape for ROFA, and a U-shape for ROTA, can perhaps be better understood if we consider the nature of two of the key variables: ROTA and (E/T). ROTA is the overall profitability of an MNE (it includes domestic profitability), while (E/T) also includes a significant proportion of home sales (where home sales are profitable). Hence, low levels of (E/T) indicate a highly (non-European) internationalized company, whereas a high (E/T) ratio is a sign of a more regionalized

312


company, where there are also large home sales. Consequently, our results show that highly regionalized and highly internationalized companies are performing considerably better overall than firms which operate at medium levels of regionalization. It is not surprising that at high levels of regionalization (that in practice translates into mostly domestic operations) companies are doing fairly well. However, when they start investing abroad (at medium levels of regionalization that also implies medium levels of internationalization), their overall performance deteriorates due to a variety of factors. These include: a strong liability of foreignness, since a company is investing not only into its home region, but also into other less familiar regions (Zaheer & Mosakowski, 1997); and a large minimum administrative burden (Contractor et al., 2003). However, when an average regionalization threshold is surpassed, and a UK company is moving toward becoming a truly global business, the pattern changes. From that moment on, a lower regionalization scope (or higher internationalization scope) has a positive impact on the overall performance, as the company has already learned how to effectively conduct business in other regions, and its global network of subsidiaries becomes an asset, not a liability. The signs and coefficients for control variables do not vary a lot across different specifications and dependent variables. As expected, there is a strong positive effect of a company’s size (measured by total revenues); and a strong positive association between industry profitability and that of individual companies. The time trend is significant with a negative coefficient suggesting that average performance is deteriorating over the years.

CONCLUSIONS In this study we have applied the location variable of geography to the international activities of large UK MNEs, within the context of the literature on multinationality and performance. We find that, in general, large UK MNEs conduct the majority of their sales in the European region and that the (E/T) ratio is a significant explanatory variable affecting firm performance in a positive, but non-linear manner, allowing for standard moderating control variables. We find that data on the return on foreign assets (ROFA) enrich the previous research on multinationality and performance which has used data on ROTA. We find significant non-linear effects, with a quadratic fit for ROFA and an S-curve for ROTA, where the latter is consistent with recent literature.


313

Overall, we suggest that it is necessary to move beyond analysis of the country level (used in traditional multinationality and performance studies) toward a regional level of analysis. Above all, it is necessary to recognize the value of regional metrics, such as (E/T) in the analysis of multinationality and performance. The MNE is much more than a trading firm, for, while exports are important to UK MNEs, so are the sales of their foreign subsidiaries. This chapter’s analysis of the relationship between the return on these foreign assets (ROFA) and intra-regional sales (E/T) presents new information and challenges to studies of the activities of MNEs.

REFERENCES Aharoni, Y. (1966). The foreign investment decision process. Boston: Division of Research, Graduate School of Business Administration, Harvard University. Bartlett, C. A., Doz, Y. L., & Hedlund, G. (Eds). (1990). Managing the global firm. London: Routledge. Buckley, P. J., & Casson, M. (1976). The future of the multinational enterprise. London: Macmillan. Contractor, F. J., Kundu, S. K., & Hsu, C. C. (2003). Three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34, 5–18. Cowley, P. R. (1988). Market structure and business performance: An evaluation of buyer/seller power in PIMS database. Strategic Management Journal, 9(3), 271–279. Dunning, J. (1980). Toward an eclectic theory of international production: Some empirical tests. Journal of International Business Studies, 11(1), 9–31. Dunning, J. (1981). International production and the multinational enterprise. London: Allen & Unwin. Geringer, J. M., Beamish, P., & Da Costa, R. (1989). Diversification strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10(1), 109–119. Goerzen, A., & Beamish, P. (2003). Geographic scope and multinational enterprise performance. Strategic Management Journal, 24(13), 1289–1306. Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between multinationality and performance. Journal of International Business Studies, 30(1), 173–188. Govindarajan, V., & Gupta, A. (2001). The quest for global dominance. San Francisco: Jossey-Bass. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies. Journal of International Business Studies, 18(3), 79–89. Grant, R. M., Jammine, A. P., & Thomas, H. (1988). Diversity, diversification, and profitability among British manufacturing companies, 1972–1984. Academy of Management Journal, 31, 771–801. Green, R., & Smith, C. (1972). Multinational profitability as a function of political instability. Management International Review, 6, 23–29.

314


Haskel, J., & Martin, C. (1992). Margins, concentration, unions and business cycle. International Journal of Industrial Organisation, 10(4), 611–631. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management Journal, 40(4), 767–798. Jung, Y. (1991). Multinationality and profitability. Journal of Business Research, 23, 179–187. Levitt, T. (1983). The globalization of markets. Harvard Business Review, 61(3), 92–102. Li, L. (2005). Is regional strategy more effective than global strategy in the US service industries? Management International Review, 45(1), 19–37. Loree, D., & Guisinger, S. (1995). Policy and non-policy determinants of U.S. equity foreign direct investment. Journal of International Business Studies, 26(2), 281–299. Lu, J. W., & Beamish, P. (2004). International diversification and firm performance: The S-curve hypothesis. Academy of Management Journal, 47, 598–609. Mascarenhas, B., & Aaker, D. (1989). Strategy over the business cycle. Strategic Management Journal, 10(3), 199–211. Montgomery, C. A., & Porter, M. E. (1991). Strategy: Seeking and securing competitive advantage. Boston: Harvard Business School Publishing. Nigh, D. (1985). The effect of political events on United States direct foreign investment: A pooled time-series cross-sectional analysis. Journal of International Business Studies, 16(1), 1–17. Nohria, N., & Ghoshal, S. (1997). The differentiated network. San Francisco: Jossey-Bass. Porter, M. E. (1990). The competitive advantage of nations. New York: Macmillan. Ramaswamy, K. (1995). Multinationality, configuration, and performance: A study of MNEs in the US drug and pharmaceutical industry. Journal of International Management, 1, 231–253. Reeb, D. M., Kwok, C. C., & Bayek, Y. (1998). Systematic risk of the multinational corporation. Journal of International Business Studies, 29, 263–279. Rugman, A. M. (1981). Inside the multinationals: The economics of internal markets. London: Croom Helm. Rugman, A. M. (2000). The end of globalization. London: Random House Business Books. Rugman, A. M. (2005). The regional multinationals: MNEs and global strategic management. Cambridge: Cambridge University Press. Rugman, A. M., & D’Cruz, J. (2000). Multinationals as flagship firms: Regional business networks. Oxford: Oxford University Press. Rugman, A. M., & Girod, S. (2003). Retail multinationals and globalization: The evidence is regional. European Management Journal, 21(1), 24–37. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Ruigrok, W., & Wagner, H. (2002). Internationalization and performance: An organizational learning perspective. Management International Review, 43, 63–83. Schlie, E., & Yip, G. S. (2000). Regional follows global: Strategy mixes in the world automotive industry. European Management Journal, 18(4), 343–354. Schmalensee, T. (1985). Do markets differ much? American Economic Review, 75, 341–351. Sullivan, D. (1994). Measuring the degree of internationalization of a firm. Journal of International Business Studies, 25, 325–342.


315

Tallman, S., & Li, J. T. (1996). Effects of international diversity and product diversity on the performance of multinational firms. Academy of Management Journal, 39, 179–196. Yip, G. S. (2003). Total global strategy II. Upper Saddle River, NJ: Prentice-Hall. Yip, G. S., Rugman, A. M., & Kudina, A. (2007). International success of British companies. Long Range Planning, 39(3), 241–264. Zaheer, S., & Mosakowski, E. (1997). The dynamics of the liability of foreignness: A global study of survival in financial services. Strategic Management Journal, 18(6), 439–463.

INTRA-REGIONAL SALES AND PERFORMANCE Nessara Sukpanich ABSTRACT This study examines the effect of a firm’s level of intra-regional sales in the triad markets of North American, Europe, and Asia on its performance. The form of the relationship is explored. The results show that there exists a strong positive relationship between a firm’s level of intra-regional sales and its performance (measured by return on equity (ROE) and return on assets, (ROA)). A firm tends to perform better when it has its sales in the home region of the triad. The hypothesis that there exists a non-linear relationship (second- and third-order curvilinear relationship) between performance and intra-regional sales is not supported.

INTRODUCTION Rugman and Verbeke (2004) present data on the distribution of sales of the world’s 500 largest firms across the broad triad regions of North America, the European Union (EU), and Asia. They find that of the 500 largest firms, 380 firms (for which data on intra-regional sales are available) have average


317

318

NESSARA SUKPANICH

level of intra-regional sales as high as 72%. That is, the majority of these firms have their sales regionally based in the home region of the triad. These findings are the starting point for introducing a regional component in international business research. The study in this chapter further explores this regional presence in the context of the effect of a firm’s intra-regional sales on its performance. The existing theory of multinational enterprises (MNEs) suggests that international expansion can generate both benefits and costs to the firm. Many empirical studies explore the form of the relationship between a firm’s degree of multinationality and performance (see, for example, Buhner, 1987; Sullivan, 1994b; Gomes & Ramaswamy, 1999; Ruigrok & Wagner, 2003). However, most of these studies neglect to capture the regional nature of a firm that might also affect a firm’s performance. Accordingly, this chapter applies the existing theory of MNE to explain the benefits and costs of interregional expansion, and then explores how a firm’s proportion of intraregional sales could affect its performance. The form of the relationship between a firm’s proportion of intra-regional sales and its performance is also examined. This chapter is organized into seven main sections. The first section is the introduction to the chapter. The second section is the literature review regarding the benefits and costs of international expansion. The third section explains the econometric models. The fourth section describes the data and variable measures used in the study. The fifth section shows and analyzes the results. The sixth section describes the limitations of the data used in the analysis. The seventh section provides the conclusions of the study.

THE LITERATURE ON THE BENEFITS AND COSTS OF INTERNATIONAL EXPANSION The existing literature concerning the relationship between a firm’s degree of multinationality and its performance suggests that a firm’s international expansion can generate both advantages and disadvantages for a firm. That is, the effect of a firm’s degree of multinationality on a firm’s performance can be positive or negative. The following is a review of literature regarding the benefits of international expansion, the costs of international expansion, and the interplay between the benefits and costs of international expansion (see Table 1 for a summary of theoretical literature related to benefits and costs of international expansion).

Intra-Regional Sales and Performance

Table 1. Frameworks

Benefits of international expansion Costs of international expansion

Interaction between benefits and costs of international expansion

319

Summary of Theoretical Literature Review Related to Benefits and Costs of International Expansion. Basic Arguments

Implication on Some Relationships

Economies of scale Economies of scope Exploitation of national differences

Higher degree of multinationality causes higher performance Higher degree of Liability of foreignness including: multinationality Internal costs, such as higher causes lower coordination costs performance External costs, such as financial risk and political risk Location choice studies: a firm will There is a non-linear relationship first expand to regions or locations between degree of that are geographically and multinationality culturally close to the home and performance country (net benefits increase), then expand to locations that it is less familiar (net benefits decline). Therefore, there exists ‘‘threshold of international expansion’’ or ‘‘optimal level of international expansion’’ Organization learning perspective: a firm may experience liability of foreignness when it first expands abroad (net benefits decline); however, learning opportunities along the international expansion process provide a firm with accumulated knowledge, leading to a firm’s success in the future (net benefits increase) A three-stage theory of international expansion: at the initial stage of international expansion a firm will face liability of foreignness (net benefits decrease); in the mid-stage further geographic expansion can reduce costs and enhance ability to arbitrage national differences (net benefits increase); further international expansion beyond the optimal level will generate higher coordination costs than benefits (net benefits decrease)

Representative Scholars Rugman (1976) Grant (1987) Buhner (1987 Hymer (1976) Michael and Shaked (1986) Zaheer and Mosakowski (1997) Sullivan (1994b) Gomes and Ramaswamy (1999) Lu and Beamish (2001) Contractor, Kundu, and Hsu (2003) Ruigrok and Wagner (2003)

320

NESSARA SUKPANICH

Benefits of International Expansion The international expansion of MNEs can generate several types of benefits. In his early work, Rugman (1976) suggests that MNEs might be able to reduce their risk of earnings through diversification of their international operations. Hamel and Prahalad (1985) indicate that the multinational presence of MNEs allows firms to exploit the advantages of crosssubsidization to compete with firms in host markets. Grant (1987) suggests that the geographic dispersion of MNEs can generate a wider range of investment opportunities than domestic firms. In summary, a firm’s international expansion can lead to advantages of scale, scope, and the exploitation of national differences (Rugman & Verbeke, 1992). Many MNE theories emphasize the benefits or advantages of international expansion. According to Hymer (1976), the creation of MNEs can internalize pecuniary externalities occurring when firms compete with each other (structural market imperfection). Further, Dunning and Rugman (1985) suggest that internalization can also overcome the natural market failure problems which can occur from the public good characteristic of some assets, such as knowledge and information. The resource-based view also emphasizes the importance of firm-specific assets. This theory suggests that a firm’s unique resources and heterogeneous capabilities can generate competitive advantages, which in turn contribute to sustainable superior returns (Barney, 1991; Rugman & Verbeke, 2002).

Costs of International Expansion Although international expansion generates several kinds of benefits, some scholars indicate that the multinational presence also comes with some costs or disadvantages. Zaheer and Mosakowski (1997) propose that a firm operating abroad may encounter the liability of foreignness, a comparative disadvantage compared to a local firm. Hymer (1976) also suggests that firms may face unavoidable costs when operating in foreign countries. The costs of firms doing business abroad can be related to internal costs or external costs of internationalization (Ruigrok & Wagner, 2003). The internal costs may arise from the fact that international expansion generates more organizational complexity and cultural diversity within the MNEs, resulting in higher coordination costs, and these costs may eventually exceed the benefits from foreign operation (Geringer, Beamish, & Da Costa, 1989; Zaheer & Mosakowski, 1997). According to Ruigrok and Wagner (2003),


321

the external costs may relate to financial risk (such as the fluctuation of exchange rates) and political risk (such as the discrimination of the host government and changes in government regulations).

Interplay between the Benefits and Costs of International Expansion Many studies have attempted to explain the interplay between the benefits and costs of international expansion. Gomes and Ramaswamy (1999), for example, suggest that during one period the benefits of internalization may exceed the costs, and during another period the costs may exceed benefits. That is, the relationship between the degree of multinationality and performance may exhibit a non-linear form, although there is still no agreement on the exact form of the relationship (Ruigrok & Wagner, 2003). According to Gomes and Ramaswamy (1999), several location choice studies suggest that in the initial stage of international expansion, a firm tends to expand its operation in the locations or regions in which it is most familiar. Johanson and Valhne (1977) suggest that a firm will first expand to the locations that are geographically and culturally close to its home region, and then move on to regions that are less familiar. Owing to the fact that these familiar locations tend to have similar administrative systems and consumer tastes, the initial expansion to these locations allows a firm to realize the benefits of internationalization with relatively small incremental costs (Davidson, 1983; Gomes & Ramaswamy, 1999). This is because during the initial phase of international expansion, a firm can use existing skills, resources, and organization to achieve the benefits of internationalization (Kogut, 1985). However, at a higher level of foreign operation a firm tends to engage in more complex organization and structure, and tends to exhibit higher cultural diversity (Geringer et al., 1989). As a result, at a very high level of internationalization, the costs might escalate over the benefits so that a firm’s net benefits might decline. This leads to the concept of optimal international expansion or critical ‘‘internationalization threshold’’ (Geringer et al., 1989; Gomes & Ramaswamy, 1999). That is, during the initial stage of international expansion the marginal benefits could exceed the marginal costs so that a firm could experience higher performance from internationalization; however, when the expansion goes beyond the optimal level the marginal costs might escalate to surpass the marginal benefits, leading to a firm’s performance decline (Gomes & Ramaswamy, 1999). On the other hand, Lu and Beamish (2001) suggest that during the initial stage of international expansion, a firm may inevitably face the liability of

322

NESSARA SUKPANICH

foreignness. These liabilities include organization reconfiguration, cultural diversity, high-coordination costs, and political influence in the host country (Zaheer & Mosakowski, 1997). The disadvantages from these liabilities may exceed the benefits of international expansion – leading to a performance decline in the initial stage of internationalization (Lu & Beamish, 2001). However, according to Ruigrok and Wagner (2003), the organizationlearning perspective suggests that although a firm may face the liability of foreignness when it first begins international expansion, learning opportunities along the internationalization process provide a firm with accumulated knowledge, leading to a firm’s success in the future. Accordingly, Lu and Beamish (2001) suggest that a firm’s performance may decline in the initial stage of internationalization, but greater levels of international expansion are associated with higher performance. Contractor, et al. (2003) propose a three-stage theory of international expansion. They suggest that a firm may face a liability of foreignness, including the costs of acquiring knowledge about the foreign markets and a large administrative overhead burden, at the initial stage of international expansion. In the mid-stage of international expansion, further geographic expansion may reduce the overheads per nation and create a firm’s ability to arbitrage national differences and to have better access to low-cost inputs (Daniels & Bracker, 1989; Kogut, 1985; Rugman, 1981). However, the benefits of this international expansion cannot go on forever. The international expansion over the optimum level may generate greater coordination and governance costs than benefits. Accordingly, the relationship between performance and degree of multinationality may be summarized with the horizontal S-curve form (Riahi-Belkaoui, 1998; Contractor et al., 2003). That is, a firm’s performance will be negatively related to lower and higher ranges of degree of multinationality, and positively related to a middle range of degree of multinationality (Riahi-Belkaoui, 1998).

The Empirical Literature Over the past 30 years, a number of researchers have investigated the relationship between a firm’s degree of multinationality and its performance. The findings are inconclusive. Some studies find a positive relationship (see, for example, Grant, 1987; Buhner, 1987), and some find a negative relationship (see, for example, Michael & Shaked, 1984; Collins, 1990). Moreover, most recent findings indicate that the relationship between degree


323

of multinationality and performance may exhibit a non-linear form (Ruigrok & Wagner, 2003). Some studies find that the relationship is in the inverted U-curve or inverted J-curve form (see for example, Daniels & Bracker, 1989; Gomes & Ramaswamy, 1999). Some studies find a U-curved or J-curved relationship (Lu & Beamish, 2001; Ruigrok & Wagner, 2003). Recent studies find that the relationship could exhibit a horizontal S-curve form, or a triagonometric wave relationship (Sullivan, 1994b; Riahi-Belkaoui, 1998; Contractor et al., 2003). Various empirical studies use different methods to measure degree of multinationality. Most studies use value of foreign sales as a percentage of total sales as a proxy for degree of multinationality (see, for example, Buhner, 1987; Geringer et al., 1989; Daniels & Bracker, 1989; Ruigrok & Wagner, 2003). Some studies measure degree of multinationality by the value of foreign assets as a percentage of total assets (Daniels & Bracker, 1989). Sullivan (1994a) uses a composite index of five variables: foreign sales as a percentage of total sales (FSTS), foreign assets as a percentage of total assets (FATA), overseas subsidiaries as a percentage of total subsidiaries (OSTS), psychic dispersion of international operations (PDIO), and top managers’ international experience (TMIE), to measure degree of multinationality. Moreover, different studies use different proxies to measure a firm’s performance. Most studies measure the performance of firm by accounting financial indicators including return on equity (ROE) (see, for example, Rugman, Lecraw, & Booth, 1985; Buhner, 1987; Luo & Peng, 1999), return on assets (ROA) (Buhner, 1987; Daniels & Bracker, 1989; Ruigrok & Wagner, 2003), and return on sales (ROS) (Grant, 1987; Tallman & Li, 1996; Luo & Peng, 1999). Some studies measure a firm’s performance by market value variables, such as Tobin’s q-value (see, for example, Morck & Yeung, 1991; Mishra & Gobeli, 1998) and risk-adjusted returns (see, for example, Michael & Shaked, 1986; Buhner, 1987; Collins, 1990). Some studies measure operational performance by the value of operating costs as a percentage of total sales (OCTS) (Ruigrok & Wagner, 2003; Gomes & Ramaswamy, 1999).

THE ECONOMETRIC MODEL According to the review of literature in the second section, international expansion can generate both benefits and costs for a firm. The same logic for factors generating these advantages and disadvantages can be applied to a

324

NESSARA SUKPANICH

firm’s inter-regional expansion. That is, a firm’s inter-regional sales across the broad-triad regions can generate advantages of scale, scope, and the exploitation of regional and national differences. However, Rugman (2005) suggests that firms that have sales in the home region of the triad and try to expand their sales into other regions will face the liability of foreignness. These liabilities might include greater internal costs from increasing coordination and organization costs, and higher external costs from financial and political risks across regions. For these reasons, inter-regional expansion leads to both advantages and disadvantages. Therefore, a firm’s intra-regional sales, a reflection of a firm’s inter-regional expansion, can also generate both benefits and costs for a firm. Accordingly, this chapter examines the effect of a firm’s proportion of intra-regional sales on a firm’s performance. Due to the fact that a firm’s intra-regional sales may generate both benefits and opportunity costs in expanding sales to other regions of the triad, this chapter also explores whether the relationship between a firm’s proportion of intra-regional sales and its performance is in a positive linear form, a negative linear form, or a non-linear form (U-curve or inverted U-curve form, possibly J-curve or inverted J-curve form, and triagonometric wave form). The model also controls for other factors that might affect a firm’s performance. These factors include firm size, knowledge, marketing ability, industry effect, and financial leverage. The model can be specified as follows: Linear model performance ¼ b0 þ b1 firm_size þ b2 knowledge þ b3 marketing_ability þ b4 industry_effect þ b5 financial_leverage þ b6 INTRA þ

ð1Þ

Non-linear model performance ¼ b0 þ b1 firm_size þ b2 knowledge þ b3 marketing_ability þ b4 industry_effect þ b5 financial_leverage þ b6 INTRA þ b7 INTRA2 þ

ð2Þ

performance ¼ b0 þ b1 firm_size þ b2 knowledge þ b3 marketing_ability þ b4 industry_effect þ b5 financial_leverage þ b6 INTRA þ b7 INTRA2 þ b8 INTRA3 þ

ð3Þ


325

where e stands for the error term, INTRA stands for a firm’s proportion of intra-regional sales (in the triad region of North America, Europe, and Asia), and performance is measured by a firm’s ROE, ROA, and ROS. Eq. (1) examines the linear relationship between a firm’s proportion of intra-regional sales and its performance, while Eqs. (2) and (3) explore the non-linear relationship. To test whether there exists a non-linear relationship (second- and third-order curvilinear relationship) between a firm’s intra-regional sales and its performance, each of the models in Eqs. (2) and (3) needs to be tested against the model in Eq. (1). The second-order curvilinear relationship (the model in Eq. (2)) rather than the linear model (the model in Eq. (1)) will be supported if (i) the Akaike Information Criteria (AIC) associated with the second-order curvilinear model is lower than that of the linear model and (ii) the coefficient on the squared term (INTRA2) in Eq. (2) has to be significant.1 The same rules can be applied for testing the model in Eq. (3) against the model in Eq. (1). That is, the third-order curvilinear model (model in Eq. (3)) rather than the linear model (model in Eq. (1)) will be supported if (i) the AIC associated with the third-order curvilinear model is lower than that of the linear model and (ii) the coefficients on the squared term (INTRA2) and the cubic term (INTRA3) are significant (and jointly significant). If both non-linear models (Eqs. (2) and (3)) fail to achieve the two conditions, it would be useful to turn the focus back to the traditional linear model (model in Eq. (1)). In Eq. (1), if the coefficient on INTRA is significant and positive, the relationship will exhibit a positive linear form. In contrast, if the coefficient on INTRA is significant and negative, the relationship will exhibit a negative linear form. On the other hand, if both non-linear models (Eqs. (2) and (3)) meet the two conditions specified earlier, it would be useful to further explore whether the relationship between a firm’s proportion of intra-regional sales and its performance exhibit a second- or third-order curvilinear relationship. Similar conditions suggested previously could be applied in the comparison between models in Eqs. (2) and (3). That is, the third-order curvilinear relationship (model in Eq. (3)) rather than the second-order curvilinear relationship (model in Eq. (2)) will be supported if (i) the AIC associated with the third-order curvilinear model is lower than that of the second-order curvilinear model and (ii) the coefficient on the cubic term (INTRA3) is significant. The models in Eqs. (1), (2), and (3) control for various factors that are likely to affect a firm’s performance. The resource-based view suggests that a firm’s unique resources and heterogeneous capabilities, such as its

326

NESSARA SUKPANICH

technological knowledge and marketing skills, can generate competitive advantages, which can lead to sustainable superior returns (Barney, 1991; Rugman & Verbeke, 2002). Firm size may determine a firm’s performance through potential economies of scale of large firms (Nayyar, 1992). Industry type may also have important effects on a firm’s performance (Schmalensee, 1985). Financial leverage reflecting observable risk factor and capital structure (especially debt) has been argued to affect a firm’s performance (Buhner, 1987; Hitt & Smart, 1994). Accordingly, in this study the models exploring the relationship between a firm’s proportion of intra-regional sales and its performance control for various variables including firm size, knowledge, marketing ability, industry type, and financial leverage.

VARIABLE MEASURES AND DATA SAMPLE In this study, the model in Eqs. (1), (2), and (3) can be estimated using the ordinary least squares (OLS) method. The dependent variable, a firm’s performance, is measured by three variables including a firm’s return on equity (ROE=net income/common equity), a firm’s return on assets (ROA=profits before interests and taxes/total assets), and a firm’s return on sales (ROS=profits before interests and taxes/total sales). The measures of performance (ROE, ROA, and ROS) are similar to those in Grant (1987). The major independent variable INTRA, is measured by a firm’s proportion of intra-regional sales in the triad of North America, Europe, and Asia (INTRA=intra-regional sales/total sales). Other control variables are measured by various proxies. Firm size is measured by a firm’s log of total assets (logasset). The value of a firm’s research and development (R&D) expenses as a proportion of total sales (RDpsale) is used as a proxy for knowledge. Marketing ability is measured by a firm’s selling and general administrative expenses as a proportion of total sales (selladminpsale). Industry type is identified by a dummy of whether a firm is in the manufacturing industry or service industry (servicedummy). Financial leverage (FL), the last controlled variable, is measured by a firm’s total debt (debt in current liabilities plus total long-term debt) as a proportion of total assets (see Table 2 for the descriptions of all the variables used in the analysis). The analysis in this study is based on the dataset of 91 large firms listed in the ‘‘Fortune Global 500 (2002)’’ for year 2001 data. Out of 91 observations, 67 firms are North American and 24 are European companies. In terms of industry type, there are 66 manufacturing and 25 service firms included in the dataset. The data for the variables used in the analysis of this study are


Table 2. Variables

327

List of Variables Used in the Estimation. Explanation

ROE ROA

Return on equity=(net income – preferred dividend)/common equity Return on assets=profits before interests and taxes/total assets, where profits before interest and taxes=net income+interest expenses+total income taxes ROS Return on sales=profits before interests and taxes/total sales logasset Log of total assets (millions of dollars) selladminpsale Selling and general administrative expenses as a proportion of total sales=selling and general administrative expenses/total sales RDpsale Research and Development expenses as a proportion of total sales=research and development expenses/total sales servicedummy Dummy, 1 if the firm is a service firm FL Financial Leverage=(debt in current liabilities+total long term debt)/total assets INTRA Intra-regional sales as a proportion of total sales=intra-regional sales/total sales INTRA2 The squared term of INTRA INTRA3 The cubic term of INTRA

Note: The original selling and general administrative expenses data item from COMPUSTAT database includes R&D expenses. The selling and general administrative expenses data used here are obtained by subtracting the R&D expenses from the original one.

derived from two data sources: (1) the annual industrial section of the COMPUSTAT North America database; (2) the ‘‘Regional Nature of Global Multinational Activity’’ database (the RNGMA database), the same database used in Rugman (2005).2 The first database provides year 2001 data on a firm’s performance and other accounting variables (including a firm’s consolidated net income, common equity, total sales, total assets, selling and general administrative expenses, and R&D expenditures). These data can be used to calculate the following variables: ROE, ROA, ROS, logasset, RDpsale selladminpsale, and FL. The second database provides year 2001 data on a firm’s proportion of intra-regional sales in the triad of North America, Europe, and Asia (INTRA) and the data on a firm’s industry type (servicedummy).

RESULT Tables 3–5 report the results of the estimation of the effects of a firm’s proportion of intra-regional sales on a firm’s ROE, ROA, and ROS, respectively. Column 1 of each table shows the estimation of the linear

328

NESSARA SUKPANICH

Table 3.

The OLS Estimation of the Effect of Intra-Regional Sales on ROE.

Dependent Variable

Independent variables logasset selladminpsale RDpsale servicedummy FL INTRA

ROE Linear model

Quadratic model

Cubic model

0.0231 (1.29) 0.6003 (3.07) 0.3627 (0.70) 0.2555 (3.80) 0.2941 (2.11) 0.2905 (2.44)

0.0240 (1.36) 0.6040 (3.06) 0.3527 (0.67) 0.2720 (3.78) 0.2752 (2.05) 0.2020 (0.30) 0.3782 (0.72)

0.1756 (0.87)

0.3266 (1.09)

0.0184 (0.97) 0.6155 (3.04) 0.3425 (0.66) 0.2807 (3.86) 0.2671 (2.04) 2.9062 (0.86) 4.6332 (0.86) 2.5332 (0.93) 0.3325 (0.43)

INTRA2 INTRA3 Constant

Number of observations 91 R2 0.229 Adjusted R2 0.174 AIC 0.474 Jointly significant test of INTRA2 and INTRA3 F-statistic p-value

91 0.234 0.169 0.458

91 0.242 0.168 0.446 0.64 0.5298

Note: Values in the parenthesis are the Huber–White robust t-statistic value. p-valuer0.05.

model (model in Eq. (1)). The magnitude of the correlations (illustrated in Table 6) and the results of the regression diagnostics of variance inflation factor (VIF) statistics (shown in Table 7) suggest that multicollinearity is not a serious problem. The results from column 1 of Table 3 (using ROE as a performance measure) and Table 4 (using ROA as a performance measure) show similar findings that the coefficient on a firm’s proportion of intraregional sales (INTRA) is significant at the 5% significance level with positive value, while in column 1 of Table 5 (using ROS as a performance measure), the coefficient on INTRA is not significant at the 5% significant


Table 4.

The OLS Estimation of the Effect of Intra-Regional Sales on ROA.

Dependent Variable


329

ROA Linear model

Quadratic model

Cubic model

0.0103 (1.08) 0.3086 (3.85) 0.2409 (0.97) 0.1045 (3.79) 0.1789 (3.01) 0.1294 (2.35)

0.0107 (1.16) 0.3104 (3.83) 0.2360 (0.94) 0.1127 (3.88) 0.1695 (2.89) 0.1152 (0.42) 0.1878 (0.91)

0.1211 (1.19)

0.1961 (1.44)

0.0080 (0.83) 0.3160 (3.83) 0.2310 (0.93) 0.1169 (4.08) 0.1655 (2.88) 1.3872 (0.92) 2.2346 (0.94) 1.2245 (1.02) 0.1226 (0.35)


Number of observations 91 R2 0.249 Adjuste R2 0.196 AIC 2.144 Jointly significant test of INTRA2 and INTRA3 F-statistic p-value

91 0.256 0.193 2.131

91 0.265 0.194 2.121 0.99 0.3756

Note: Values in the parenthesis are the Huber–White robust t-statistic value. p-value r0.05.

level. That is, for a linear model, a firm’s proportion of intra-regional sales significantly and positively affects a firm’s ROE and ROA, but insignificantly affects a firm’s ROS. Column 2 in Tables 3–5 show the results of the estimation of the quadratic model (model in Eq (2)). To compare the quadratic model in column 2 with the linear model in column 1, two criteria specified in the econometric model section need to be examined. The results from Tables 3 and 4 (using ROE and ROA as performance measures) show that the AIC of the quadratic model is greater than that of the linear model, while the results

330

NESSARA SUKPANICH

Table 5.

The OLS Estimation of the Effects of Intra-Regional Sales on ROS.

Dependent Variable


ROS Linear model

Quadratic model

Cubic model

0.0309 (1.43) 0.4367 (4.08) 0.0356 (0.11) 0.1709 (4.68) 0.0391 (0.41) 0.1092 (1.59)

0.0318 (1.48) 0.4404 (4.12) 0.0255 (0.08) 0.1875 (4.80) 0.0201 (0.21) 0.3843 (1.08) 0.3789 (1.47)

0.2826 (1.28)

0.4340 (1.65)

0.0299 (1.37) 0.4443 (4.10) 0.0220 (0.07) 0.1905 (4.79) 0.0174 (0.18) 0.6762 (0.39) 1.3309 (0.49) 0.8643 (0.66) 0.2091 (0.46)


Number of observations 91 R2 0.339 Adjusted R2 0.292 AIC 1.646 Jointly significant test of INTRA2 and INTRA3 F-statistic p-value

91 0.353 0.298 1.646

91 0.356 0.293 1.627 1.68 0.1923

Note: Values in the parenthesis are the Huber–White robust t-statistic value. p-value r0.05.

from Table 5 (using ROS as a performance measure) show that the AIC of the quadratic model is equal to that of the linear model. Accordingly, the first criterion in favor of the quadratic model for all the three tables is not satisfied. Considering the second criterion, according to all the three tables, the coefficient on INTRA2 is not significant at the 5% significant level. Therefore, the second criterion in favor of the quadratic model is not satisfied. Accordingly, the quadratic model, in contrast to the linear model, for all the three tables (three measures of a firm’s performance) cannot be supported.


Table 6.

331

Summary Statistics and Correlations.

Variables

Mean Standard Minimum Maximum Deviation

ROE ROA ROS logasset selladminpsale RDpsale FL INTRA

0.07 0.07 0.06 10.07 0.17 0.04 0.26 0.66

0.20 0.09 0.12 0.93 0.11 0.05 0.13 0.21

0.70 0.16 0.58 8.15 0.02 0.00 0.00 0.30

1

2

3

4

5

6

7

0.56 0.30 0.91 0.33 0.76 0.83 12.69 0.05 0.06 0.11 0.49 0.16 0.21 0.23 0.16 0.21 0.12 0.09 0.24 0.15 0.33 0.56 0.17 0.24 0.06 0.02 0.09 0.24 1.00 0.02 0.02 0.18 0.28 0.17 0.43 0.03

p-value r0.05.

Table 7. Variance Inflation Factor (VIF) of Linear Regression Model. Variable

VIF

1/VIF

logasset selladminpsale RDpsale servicedummy FL INTRA

1.11 1.32 1.56 1.89 1.08 1.83

0.9020 0.7560 0.6419 0.5300 0.9252 0.5458

Mean VIF

1.46

Column 3 in Tables 3–5 shows the results of the estimation of the cubic model (model in Eq. (3)). To compare the cubic model in column 3 with the linear model in column 1, two criteria specified in the econometric model section need to be examined. The results from all the three tables show that the AIC of the cubic model is greater than that of the linear model, and the coefficients on INTRA2 and INTRA3 are jointly insignificant at the 5% significant level, and each term is not significant at the 5% significant level. As a result, the two criteria in favor of the cubic model are not satisfied. Therefore, the cubic model, in contrast to the linear model, for all the three tables (three measures of a firm’s performance) cannot be supported. According to the results from Tables 3–5, it is clear that the non-linear relationship (second- and third-order curvilinear relationship) between a firm’s proportion of intra-regional sales and its performance (as measured by ROE, ROA, and ROS) hypothesis is not supported. For all the three tables, the results show that adding the squared term (INTRA2) or both the

332

NESSARA SUKPANICH

squared term and the cubic term (INTRA2 and INTRA3) of the proportion of intra-regional sales do not improve the model. Accordingly, further comparison between the second- and the third-order curvilinear model is not necessary. Considering the linear model, the results vary depending on the measure of a firm’s performance. When using ROS as a performance measure, the results show that intra-regional sales insignificantly affect a firm’s ROS. However, when using ROE and ROA as performance measures, the results show that all else being equal, a firm with higher level of intra-regional sales performs better (has greater level of ROE and ROA) than a firm with a lower level of intra-regional sales. It is possible that indeed the cost of interregional expansion might go beyond its benefits so that the firm can perform better in the home region rather than in other regions of the triad. Rugman and Verbeke (2004) suggest that a firm trying to expand its sales from the home region of the triad to other regions may face liabilities of interregional foreignness. These liabilities might be related to internal costs to the firm (such as higher coordination and control costs) and external costs (such as regional regulations, powerful foreign rivals in other regions, and local product preference). The findings of this study do not oppose the tradition theory related to the benefit of international expansion. However, the results can be interpreted that although there may exist some benefits of interregional expansion, its costs are non-negligible. That is, to design interregional strategy, the MNE managers need to take into account the liability of inter-regional foreignness.

LIMITATIONS This study has some limitations. First, due to the fact that the RNGMA database has available data on intra-regional sales only for year 2001, the data used for the analysis are cross-sectional. This generates limitations to analyze the relationship between a firm’s level of intra-regional sales and its performance across time. Accordingly, a timewise analysis would be a logical next step for future research. Secondly, all firms included in the analysis are North American and European firms. Lacking a sample of firms from all parts of the triad (North America, Europe, and Asia) could restrict the generalizability of the study. To be able to generalize the results, future research should try to incorporate firms from all parts of the triad and may cover firms from other regions. Moreover, firms from banking and other financial services industries are not


333

included in the data sample of this study. Since the financial firms may have very distinct characteristics from firms from other service industries, and, of course, from the manufacturing firms, a separate analysis for these type of firms would be an interesting case for future research. Thirdly, this study does not incorporate any analysis of a firm’s structure and its managerial capability that might affect a firm’s performance. This is a distinct limitation of studies using secondary data. Accordingly, future research may try to collect data regarding a firm’s structure and its managerial method from primary sources and incorporate these variables in the analysis. Finally, this study measures a firm’s performance by accounting financial indicator, including ROE, ROA, and ROS. The performance measures in terms of a market values variable (such as Tobin’s q-value and risk-adjusted returns) and operational performance (such as operating costs as a percentage of total sales) are not incorporated in the estimation. This study does not aim to expand on the existing performance measures. However, for more robustness of the results, future studies may test other performance measures. The new performance measures in terms of return on foreign assets (ROFA) and return on regional assets (RORA) could be used to measure a firm’s international and regional performance.

CONCLUSIONS The main contribution of this study is to redefine multinationality to capture the regional presence that can affect a firm’s performance. More specifically, this study explores the effect of a firm’s proportion of intra-regional sales on its performance. The form of the relationship between a firm’s performance and intra-regional sales is also examined. The results show that the hypothesis regarding the non-linear relationship (second- and third-order curvilinear relationship) between a firm’s performance and its intra-regional sales is not supported. For the linear relationship, the results are sensitive to the choice of performance measures. A firm’s proportion of intra-regional sales insignificantly affects a firm’s ROS. However, there exists a positive linear relationship between a firm’s performance (ROE and ROA) and its proportion of intra-regional sales. That is, a firm with a higher level of intraregional sales tends to perform better (has greater level of ROE and ROA) than a firm with a lower level of intra-regional sales. In other words, the costs of inter-regional expansion exceed its benefits. This is because a firm trying to expand its sales into other regions may face costly liability of

334

NESSARA SUKPANICH

inter-regional foreignness (which could be stronger than the liability of international expansion). Moreover, the possible benefits of inter-regional expansion (such as economies of scales and scopes) may first be realized within the home region of the triad itself. These findings have important implications for MNE strategies. That is, a firm does not need to expand its sales into other regions of the triad to generate higher profits. Indeed, it can perform better if it sells the products in the home region.

NOTES 1. According to Long and Freese (2001), the AIC is defined as

AIC ¼

^ k Þ þ 2P 2 ln LðM N

^ k Þ is the likelihood of the model with k regressors, and P is the number where LðM parameters in the model. They state that ‘‘all else being equal, the model with smaller AIC is considered the better fitting model’’ (Long & Freese, 2001, p. 86). 2. (a) The COMPUSTAT North America database provides financial statistics and market information of publicly traded companies in the United States and Canada. (b) The firms listed in the RNGMA database are the world’s 500 largest companies according to the ‘‘Fortune Global 500’’ (2002). The database contains year 2001 data on firms’ regional sales in the triad region of North America, Europe, and Asia.

REFERENCES Barney, J. (1991). Firm resources and sustained competitive advantages. Journal of Management, 17(1), 99–120. Buhner, R. (1987). Assessing international diversification of West German corporations. Strategic Management Journal, 8(1), 25–37. Collins, J. M. (1990). A market performance comparison of U.S. firms active in domestic, developed and developing countries. Journal of International Business Studies, 21(2), 271–287. Contractor, F. J., Kundu, S. K., & Hsu, C-C. (2003). A three-stage theory of international expansion: The link between multinationalilty and performance in the service sector. Journal of International Business Studies, 34, 5–18. Daniels, J. D., & Bracker, J. (1989). Profit performance: Do foreign operations make a difference? Management International Review, 29(1), 46–56. Davidson, W. H. (1983). Market similarity and market selection: Implications for international market strategy. Journal of Business Research, 11, 439–456.


335

Dunning, J. H., & Rugman, A. M. (1985). The influence of Hymer’s dissertation on the theory of foreign direct investment. The American Economic Review, 75(2), 228–232. Geringer, J. M., Beamish, P. W., & Da Costa, R. C. (1989). Diversification strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10(2), 109–119. Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between multinationality and performance. Journal of International Business Studies, 30(1), 173–187. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies. Journal of International Business Studies, 18(3), 79–89. Hamel, G., & Prahalad, C. K. (1985). Do you really have a global strategy? Harvard Business Review, 63, 139–148. Hitt, M. A., & Smart, D. (1994). Debt: A disciplining force for managers or a debilitating force for organizations. Journal of Management Inquiry, 3, 144–152. Hymer, S. H. (1976). The international operations of national firms: A study of direct foreign investment. Doctoral dissertation, M.I.T. Press, Massachusetts. Johanson, J., & Valhne, J. (1977). The internationalization process of the firm: A model of knowledge development and increasing foreign commitments. Journal of International Business Studies, 8, 23–32. Kogut, B. (1985). Designing global strategies: Profiting from operational flexibility. Sloan Management Review, 26, 27–38. Long, J. S., & Freese, J. (2001). Regression models for categorical dependent variables using STATA. College Station, TX: Stata Press. Lu, J. W., & Beamish, P. W. (2001). The international and performance of SMEs. Strategic Management Journal, 22, 565–586. Luo, Y., & Peng, M. W. (1999). Learning to compete in a transition economy: Experience, environment, and performance. Journal of International Business Studies, 30(2), 269–295. Michael, A., & Shaked, I. (1984). Does business diversification affect performance? Financial Management, 13(4), 18–24. Michael, A., & Shaked, I. (1986). Multinational corporations vs domestic corporations: Financial performance and characteristics.. Journal of International Business Studies, 17(3), 89–100. Mishra, C., & Gobeli, D. H. (1998). Managerial incentives, internalization, and market valuation of multinational firms. Journal of International Business Studies, 29(3), 583–597. Morck, R., & Yeung, B. (1991). Why investors value multinationality. The Journal of Business, 64(2), 165–187. Nayyar, P. (1992). Performance effects of three foci in service firms. The Academy of Management Journal, 35(5), 985–1009. Riahi-Belkaoui, A. (1998). The effects of the degree of internationalization on firm performance. International Business Review, 7, 315–321. Rugman, A. M. (1976). Risk reduction by international diversification. Journal of International Business Studies, 7(2), 75–80. Rugman, A. M. (1981). Inside the multinationals: The economics of internal markets. New York: Columbia University Press. Rugman, A. M. (2005). The regional multinationals. Cambridge, UK: Cambridge University Press. Rugman, A. M., Lecraw, D. J., & Booth, L. D. (1985). International business: Firm and environment. New York: McGraw-Hill.

336

NESSARA SUKPANICH

Rugman, A. M., & Verbeke, A. (1992). A note on the transnational solution and the transaction cost theory of multinational strategic management. Journal of International Business Studies, 23(4), 761–771. Rugman, A. M., & Verbeke, A. (2002). Edith Penrose’s contribution to the resource-based view of strategic management. Strategic Management Journal, 23, 769–780. Rugman, A. M., & Verbeke, A. (2004). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organization learning perspective. Management International Review, 43(1), 63–83. Schmalensee, T. (1985). Do markets differ much? American Economic Review, 75, 341–351. Sullivan, D. (1994a). Measuring the degree of internationalization of a firm. Journal of International Business Studies, 29(3), 325–342. Sullivan, D. (1994b). The threshold of internalization: Replication, extension, and reinterpretation. Management International Review, 34(2), 165–186. Tallman, S., & Li, J. (1996). Effects of international diversity and product diversity on the performance of multinational firms. The Academy of Management Journal, 39(1), 179–196. Zaheer, S., & Mosakowski, E. (1997). The dynamics of the liability of foreignness: A global study of survival of financial services. Strategic Management Journal, 18(6), 439–464.

TESTING REGIONAL EFFECTS IN THE INTERNATIONALIZATION– PERFORMANCE RELATIONSHIP IN ASIAN SERVICE FIRMS Stephen Chen ABSTRACT Despite many years of research, empirical studies of the relationship between internationalization and performance of firms have given conflicting results. Contractor, Kundu, and Hsu (2003) and Lu and Beamish (2004) have recently proposed a three-stage theory of international expansion that attempts to reconcile the conflicting findings. However, other studies suggest that there are good reasons to believe that firms from less-developed countries (LDCs) differ in their internationalization from firms in developed countries. Furthermore, little research has been conducted on service firms in LDCs. This paper aims to fill some of this gap by testing the internationalization–performance relationship in a sample of service firms in the Asia-Pacific region. The study confirms the three-stage model but adds two new dimensions. First, the results show that the internationalization–performance relationship varies significantly depending on whether internationalization is intra-regional or extra-regional. Extra-regional sales/total sales showed significant and positive relationships with return on assets (ROA) in all Regional Aspects of Multinationality and Performance Research in Global Strategic Management, Volume 13, 337–358 Copyright r 2007 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1064-4857/doi:10.1016/S1064-4857(07)13014-X

337

338

STEPHEN CHEN

three models but combined foreign sales/total sales showed no significant relationships with ROA and surprisingly intra-regional sales showed a significant relationship only in the quadratic model.

INTRODUCTION It has been said with some justification that the internationalization– performance relationship is the key question in international business and significant research has been conducted over the last 30 or so years on this question. However, most research has been on firms in the manufacturing sector. Although service firms have contributed to the majority of the job growth in the industrialized nations, with a few notable exceptions (e.g. Capar & Kotabe, 2003; Contractor, Kundu, & Hsu, 2003) research on internationalization in the service sector has only been explored to a limited extent. Even less research has been conducted on service firms in lessdeveloped countries (LDCs). This paper aims to fill some of these gaps by testing the internationalization–performance relationship in a sample of service firms in the Asia-Pacific region. Reasons for expansion of service firms are much the same as those for manufacturing firms such as labor costs, market access and resources (Murray & Kotabe, 1999). However, despite these similarities, there are also some differences between manufacturing and service firms. For instance, the nature of service businesses is mostly intangible. Second, the production and consumption of many services occur simultaneously owing to the impossibility of inventory in services (Habib & Victor, 1991). Therefore, there are good reasons to believe that the internationalization–performance relationship may be different for service firms (Capar & Kotabe, 2003; Contractor et al., 2003). There may also be good reasons to believe that the internationalization– performance relationship is different for firms in LDCs. First, firms in LDC are generally less advanced in managerial and technological knowledge in comparison with developed country firms (Tolentino, 1993). Second, they may face different macroeconomic conditions and the institutional environments from developed country firms (Khanna & Rivkin, 2001). Third, developed countries and LDCs differ in international competitiveness by sector (Nachum, 2004). For instance, while the manufacturing and commodity sectors have seen a decline in many industrialized countries, the manufacturing and commodity sectors have been growing in many developing countries.

Testing Regional Effects in the Internationalization–Performance Relationship

339

The remainder of the paper is organized as follows. First, we provide a review of the theoretical background and literature on the internationalization– performance relationship. Second, we describe the data collection, analysis and results of the study. Third, we discuss the implications of the results along with limitations of the study and possible future directions for research.

THEORETICAL BACKGROUND AND HYPOTHESES DEVELOPMENT The Internationalization–Performance Relationship The relationship between internationalization and firm performance has long been a topic of interest to international business researchers (e.g. Hymer, 1976; Rugman, 1979; Caves, 1982). However, despite years of research, there seems to be no clear consensus about the relationship between internationalization and performance from previous studies. The overall shape of the relationship between the performance of multinational enterprises (MNEs) and their degree of internationalization is a crucial question. Early researchers share a strong focus on the benefits of international expansion and agree that the internationalization– performance relationship is positive (Vernon, 1971; Errunza & Senbet, 1984; Bu¨hner, 1987; Grant, 1987) but later studies have found a negative relationship (Michel & Shaked, 1986; Collins, 1990; Shaked, 1986; Kumar, 1984). Addressing these contradictory findings, researchers in the 1990s elaborated both the benefits and costs and suggested a curvilinear form of the relationship (e.g. Geringer, Beamish, & Da Costa, 1989; Hitt, Hoskisson, & Kim, 1997; Gomes & Ramaswamy, 1999; Mauri & Sambharya, 2001; Capar & Kotabe, 2003). For instance, Geringer et al. (1989) found that when sales from foreign subsidiaries rise to the 80–100 percent quintile, the relation to performance turns negative. Reasons given for an inverted U-shaped relationship are that geographic expansion leads to enhanced corporate performance up to a second threshold, beyond which the organizational costs and complexity associated with managing widely scattered operations begin to outweigh the advantages of even further international expansion. More recently, Ruigrok and Wagner (2003) have found a standard U-shaped relationship between internationalization and firm performance.

340

STEPHEN CHEN

Building on Johanson and Vahlne’s (1977) classic work, they argue that the form of the internationalization–performance relationship is likely to be determined by organizational learning processes. However, they argue that by building on their domestic business, firms are initially able to exploit economies of scale and/or scope and thus achieve high-performance outcomes. In the course of further international expansion firms encounter an increasing imbalance between external environments and internal competencies. This triggers organizational learning and firms begin to reconfigure internal systems, mechanisms and processes to match their new global environment. Corporations that successfully pass through the readjustment phase then experience a reversal of fortunes and restore positive performance development. Contractor et al. (2003) and Lu and Beamish (2004) have independently suggested that all these contradictory findings might be reconciled in a new three-stage theory of international expansion in which positive, negative and U-shaped relationships can be found at different stages. In Stage 1, the relationship is a negative slope as costs and barriers to initial international expansion outweigh the benefits. In addition to economic and legal barriers, barriers to internationalization include cultural distance (Johanson & Vahlne, 1977) and establishing the firm’s legitimacy abroad (Zaheer & Mosakowski, 1997). For early internationalizers, the small scale of global operations is insufficient to recoup costs of creating an international organization (Hitt et al., 1997; Gongming, 1998). In Stage 2, the relationship becomes positive as benefits of international expansion are now realized. Further geographical expansion makes possible efficiencies that improves resource utilization (Kogut, 1985; Porter, 1985; Kim, Hwang, & Burgers, 1989) while market-seeking firms are better able to scan for market opportunities. Other benefits of Stage 2 international expansion are the ability of some companies to exercise global market power (Grant, 1987) and to extend the product cycle (Vernon, 1966). In Stage 3, the relationship becomes negative again as some firms expand beyond an optimal threshold. Contractor et al. (2003) identify some reasons why this occurs. First, in this stage having expanded into the most lucrative markets, the firm is left with minor or peripheral countries with a lower profit potential. Second, beyond the optimum point, the growth of coordination and governance costs may exceed the benefits of further expansion (Galbraith & Kazanjian, 1986; Hitt et al., 1997). This is especially true where the firm has to deal with different legal, cultural and linguistic environments (Gomes & Ramaswamy, 1999; Siddharthan & Lall, 1982). Table 1 summarizes the findings of previous empirical studies.


341

Table 1. Previous Literature on the Link between Degree of Internationalization and Performance. Results

Author(s) and Year

Positive (linear)

Errunza and Senbet (1984) Vernon (1971) Grant (1987)

Negative (linear)

Michel and Shaked (1986) Collins (1990)

U-Shape (nonlinear) Inverted U-shape (nonlinear)

Excess return ROI and ROS ROA, ROE, ROS

Risk-adjusted return Total risk, debt to equity, beta Siddharthan and Lall (1982) Sales growth Kumar (1984) ROA and ROS Capar and Kotabe (2003) Ruigrok and Wagner (2003) Gomes and Ramaswamy (1999) Sullivan (1994)

ROS ROA ROA and OCTS

Al-Obaidan and Scully (1995) Hitt et al. (1997)

Frontier production function ROA and R&D intensity ROS and ROA After-tax ROA and ROS ROA, ROS

Daniels and Bracker (1989) Geringer et al. (1989) S-shape

Performance Indicators

Contractor et al. (2003)

After-tax ROS and ROA

DOI

OSTS OSTS Change in overseas production ratio FSTS FSTS FSTS Overseas profitability ratio FSTS FSTS Multi-item index Composite index of five items; FSTS, FATA, TMIE, PDIO, OSTS FATA FSTS FSTS and FATA FSTS FSTS, FETE (foreign employees/total employees), FOTO (foreign offices/total offices)

Note: FATA, foreign assests/total assests; OCTS, operating costs/total sales; OSTS, overseas subsidiaries/total subsidiaries; PDIO, psychic dispersion of international operations; TMIE, top management international experience.

INTERNATIONALIZATION IN SERVICE FIRMS Some researchers (e.g. Boddewyn, Halbrich, & Perry, 1986) have argued that theories developed to explain the behavior of multinational manufacturing firms can be equally well applied to the internationalization of service firms for the same reasons (labor costs, market access, resources, etc.).

342

STEPHEN CHEN

For instance, Dunning (1989) argued that multinational service firms could benefit from global economies of scale in various aspects of the value chain such as personnel specialization, financial management and governance while Campbell and Verbeke (1994) note that service firms can achieve global economies of scale in marketing. However, other researchers have argued that despite the similar motivations of service firms to expand internationally, the unique characteristics of service firms are likely to lead to a different pattern with respect to performance. Lovelock and Yip (1996) noted that although some service firms can achieve economies of scale by providing global customers with standardized products and by centralizing some activities in the value chain, service firms generally require more local adaptation of their offering compared with manufacturing firms and many service firms also require a local presence. The need for immediate delivery in many cases also makes it difficult to transfer services from one country to another. Service firms are likely to face declining performance in international expansion for some reasons: First, many countries still have a strict control over the extent of foreign involvements in service industries, preventing service multinationals from operating efficiently (Feketekuty, 1988), such as ownership restrictions, domestic preference policies, unfavorable tax treatments and unbalanced employment rules. Second, the inherently higher need for adaptation of services to local requirements will likely raise the cost of operations for service firms higher than for manufacturing firms. Knight (1999) indicates that service firms require intensive customer contact, extensive customization and cultural adaptation. Third, some service industries require foreign direct investment from the very beginning, because many services require simultaneous production and consumption. The inseparability of many services makes it necessary for a parent firm to have a local facility (Boddewyn et al., 1986). Therefore, service firms must undertake considerably higher investments than manufacturing firms that begin foreign expansion. Katrishen and Scordis (1998) found that service firms suffered from diseconomies of scale, because of the initial investments required and the additional governance costs. They also observed that the diseconomy of scale increased even further at higher levels of internationalization. As Boddewyn et al. (1986) pointed out, even within the services sector there can be substantial differences in attributes such as intangibility, perishability and simultaneity. Recent work examining the relationship between performance and internationalization has introduced ‘moderating’ variables that distinguish between types of firm. Contractor et al. (2003) found that firms in knowledge-based sectors reap the positive (Stage 2) benefits of internationalization earlier (that is at a lower degree of


343

internationalization (DOI)) than firms in capital-intensive sectors. Second, knowledge-based sectors ‘over-internationalize,’ reaching the sub-optimal Stage 3, whereas capital-intensive sectors do not. Contractor et al. (2003) suggest three reasons why knowledge-based firms are able to reap the benefits of internationalization faster than capitalintensive firms: (1) They have a lower burden of tangible asset investment. (2) They have clients already established abroad. (3) They possibly have a greater global standardization, which lowers the cost per foreign entry, and enables the net benefits of foreign expansion to be reaped sooner. For capital-intensive service sectors, the fixed asset risk and capital cost is much higher. Consequently, such sectors are likely to be more cautious in their international expansion and less likely to stray into Stage 3 (sub-optimal for performance). Therefore, they hypothesize: Hypothesis 1. The form of the relationship between internationalization and performance for multinational service firms is a horizontal S-shape. In other words, there will be a negative relationship at a lower or higher range, and a positive relationship for a middle range.

REGIONAL DIFFERENCES There are also good reasons to believe that the internationalization– performance relationship may vary by country or geographic region. Many studies have identified significant differences between firms in developed and less-developed countries. First, LDCs typically have lower labor costs, less efficient production and less internationally competitive knowledge-based industries (Leff, 1978, 1979; Amsden & Hikino, 1994; Guillen, 2000). Second, LDCs often suffer from weak or non-existing institutions (Khanna & Palepu, 2000; Khanna & Rivkin, 2001), as well as underdeveloped capital markets (Singh, 1995). Third, LDCs are generally less economically and politically stable compared with developed countries. Industry diversification is often used to secure the supply of intermediates by internalizing these functions, thus reducing the systematic stock market risk that is typically high in most LDCs (Nachum, 2004). The same may be true for international diversification. A major reason for expecting strong association between internationalization and performance is that internationalization enables firms to benefit from economies of scale (Dunning, 1993). Scale benefits are likely to be particularly important for developing country firms, whose own markets may be too small to provide significant scale advantages. The ability to reap the benefits of scale is vital when competing on the basis of low cost, which is

344

STEPHEN CHEN

often the case of developing country firms. Furthermore, internationalization has an added value in high-risk economies that are subject to the uncertainties of instability and rapid structural change by enabling firms to alleviate these factors and by spreading market risk (Rugman, 1979). Internationalization may also allow developing country firms to access resources overseas, and thus make them less reliant on home country resources and to overcome limitations of the home market. These arguments suggest that: Hypothesis 2. In general the performance of firms from less-developed countries should vary positively with the degree of internationalization. There are also good reasons to believe that the shape of the curve may also vary with the foreign markets that it enters. The Uppsala model of internationalization (Johanson & Vahlne, 1977) proposed that firms internationalize incrementally from ‘psychically close’ countries to ‘psychically distant’ countries. This would predict a pattern of internationalization in which one would find internationalization in familiar countries in the first stage and internationalization in less familiar countries in the latter stages of the three-stage model. Other factors may also favor a regionalization strategy. By definition, regional markets are geographically closer. This reduces transportation costs. As they are in the same time zone, it is also easier to coordinate activities. By concentrating on markets in the same region, firms may also benefit from diseconomies of time compression (Vermeulen & Barkema, 2002). Finally, markets in the same region are also likely to be part of the same trade blocs and so benefit from reduced market entry barriers. These effects seem to be confirmed by the preponderance of regionalization strategies among leading MNEs. As shown by Rugman (2000), even among the most international companies, most still derive the bulk of their sales in the home region. Institutional theorists also point to significant differences between business systems internationally that may affect the competitive advantage of firms in particular markets (Griffiths & Zammuto, 2005). For example, there are many differences in the institutional systems that affect business in the US, Europe and Asia such as differences in legal and political institutions (Whitley, 1999). Since business systems can to a large extent be divided regionally, this too would tend to favor regionalization. Another factor that could encourage regionalization strategies is the need to localize some products. As described by Prahalad and Doz (1987) in their well-known Integration–Responsiveness framework, MNEs must balance the twin demands of integration and responsiveness. A regional strategy may allow MNEs to achieve a better balance between the two demands, gaining some of the benefits of globalization while remaining responsive to


345

local market needs (Ghoshal, 1987; Ghemawat, 2003). Evidence of this can be seen in the automobile industry (Schlie & Yip, 2000) or the white goods industry (Baden-Fuller & Stopford, 1990). In short, there are good reasons to believe that the internationalization– performance relationship may differ both according to the country of origin of the firm and the markets into which it internationalizes. Arguments based on minimizing costs arising from geographic and psychic distance would suggest that: Hypothesis 3. Internationalization within the region should lead to greater performance compared with internationalization outside the region.

METHODOLOGY Sample and Data Collection This study aims to shed some light on these hypothesized regional effects by examining how the performance of Asian service firms varies with internationalization into different regions. The hypotheses were tested in a sample of service firms from Asia. First, a sample of multinational service companies from the Asian region was drawn from the OSIRIS database, which provides details of listed companies in over 125 countries. To minimize distortion resulting differences in the political-economic system, this study focused on firms from seven countries (Singapore, Malaysia, Indonesia, Philippines, Taiwan, Thailand and India), which have a fairly wellestablished stock market. Companies from China and Hong Kong were excluded owing to the special nature of the economic regime, as were companies from other countries where sufficient data on companies could not be obtained. For each country, the top 200 companies (ranked by sales) in each of the following sectors were selected: Consumer, Construction, Health, Industrial transport, Software, Business Support, Telecom, Utilities and Real estate and finance. Foreign-owned companies and companies where more than 50 percent was owned by a single shareholder were excluded in order to minimize the effects of foreign and domestic parent companies on performance in the sample. Second, data on international operations (foreign sales, foreign assets, etc.) over the period 2001–2005 were obtained for each company. Firms with less than 10 percent of their sales in 2005 originating overseas were excluded, an implicit criterion used in many of the previous studies (e.g. Daniels & Bracker, 1989; Sambharya, 1995; Geringer et al., 1989), as well as those that

346

STEPHEN CHEN

Table 2.

Distribution of Sample by Country and Industry Sector.

Country

India Indonesia Malaysia Philippines Singapore Taiwan Thailand Total

Industry Code

Total

1

2

3

4

5

6

7

8

9

2 0 2 0 3 3 1

2 3 8 0 1 0 2

1 0 1 0 2 2 0

1 0 2 1 8 1 0

19 0 5 0 4 4 0

0 0 0 0 2 2 0

1 1 1 0 1 0 0

0 0 2 0 0 0 0

0 0 3 0 0 1 0

26 4 24 1 21 13 3

11

16

6

13

32

4

4

2

4

92

1, Consumer; 2, Construction; 3, Health; 4, Industrial transport; 5, Software; 6, Support; 7, Telecom; 8, Utilities; 9, Real estate and finance.

did not provide the figures necessary for calculating the ratio. The final sample of firms that had the required data on internationalization and performance amounted to 92 firms. Table 2 shows the distribution by country and industry sector.

VARIABLES AND MEASURES Dependent Variable Performance as measured by ROA was used as the dependent variable in all models.

Independent Variables Degree of internationalization (DOI) was measured using three measures. First, total foreign sales as a percentage of total sales, FSTS (e.g. Grant, 1987; Tallman & Li, 1996), was used as in previous studies. To examine regional effects two other measures were also used in this study: intraregional sales/total sales (ISTS) and extra-regional sales/total sales (XSTS), calculated as the percentage of total sales arising from within the Asia region and outside the region, respectively.


347

Control Variables SIZE: This was measured by the natural logarithm of number of employees (Gomes & Ramaswamy, 1999), and was used to control for the potential effect of scale economy differences. Logarithmic transformation not only makes the results easy to interpret, because the changes in the logarithm domain represent relative (percentage) changes in the original metric and also makes the distribution of data closer to normality. YEAR: Year for which figures were reported. INDUSTRY: Industry sector, 9 dummy variables, I1–I9. COUNTRY: Country in which company in headquartered, 7 dummy variables, C1–C7. Analysis Following Lu and Beamish (2004) and Contractor et al. (2003), the data were analyzed using a cross-sectional time series model, which makes it Table 3.

Results of Linear Regression of Percentage of Total Foreign Sales on ROA.

Independent Variable India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE FSTS cons

Note: R2=0.1349. Significant at 0.05 level.

Coefficient 0.0298017 0.0174274 0.0421335 (reference category) 0.109764 0.032063 0.2372076 0.0638276 0.0624123 0.0062817 0.007827 0.0485891 0.1532917 (reference category) 0.0574022 0.0316079 0.0114835 0.0156616 0.264067

348

STEPHEN CHEN

possible to reveal individual exploratory variables as well as the dynamics over time. However, following the advice of Beck and Katz (1995), as the number of time periods was relatively small, instead of using feasible generalized least-squares regression (FGLS) as did the previously cited researchers, OLS regression with panel-corrected standard errors was used, implemented using the xtpcse command in Stata. Unlike FGLS, disturbances are assumed to be heteroskedastic and contemporaneously correlated across panels (Stata, 2003). To test the three-stage theory model, linear, quadratic and cubic regression models were tested as follows: Model 1: Performance ¼ a þ b1 ðSIZEÞ þ b2 ðDOIit Þ þ

m ¼9 X m¼1

Table 4.

b2þm I m þ

n¼7 X

b2þmaxðmÞþn C n þ it

n¼1

Results of Quadratic Regression of Percentage of Total Foreign Sales on ROA.

Independent Variable India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE FSTS FSTS2 cons

Note: R2=0.1349. Significant at 0.05 level. Significant at 0.01 level.

Coefficient 0.0293364 0.0174927 0.0428058 (reference category) 0.1107 0.0322671 0.2361732 0.0632194 0.0633019 0.0076555 0.0085813 0.0497815 0.1536954 (reference category) 0.0566679 0.032381 0.011563 0.0014021 0.016019 0.2687619


349

Model 2: Performance ¼ a þ b1 ðSIZEÞ þ b2 ðDOIit Þ þ b3 ðDOIit Þ2 þ

m¼9 X

b3þm I m þ

m¼1

n¼7 X


n¼1

Model 3: Performance ¼ a þ b1 ðSIZEÞ þ b2 ðDOIit Þ þ b3 ðDOIit Þ2 þ b4 ðDOIit Þ3 þ

m ¼9 X m¼1

Table 5.

b4þm I m þ

n¼7 X


n¼1

Results of Cubic Regression of Percentage of Total Foreign Sales on ROA.

Independent Variable India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE FSTS FSTS2 FSTS3 cons


Coefficient 0.0355662 0.0107547 0.0356736 (reference category) 0.1071969 0.0275222 0.2420165 0.0664224 0.0677674 0.0120935 0.0100773 0.052371 0.1600323 (reference category) 0.0531066 0.02473 0.0120728 0.2534309 0.6313905 0.4190186 0.2520036

350

STEPHEN CHEN

Table 6. Results of Linear Regression of Percentage of Intra-Regional Sales on ROA. Independent Variable

Coefficient

India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE ISTS cons

0.1766468 0.1129244 0.2411716 0.1806833 0.3163055 0.224096 (reference category) 0.093174 0.0997992 0.1437872 0.1465051 0.1231496 (reference category) 0.151754 0.0959175 0.120217 0.0129928 0.0043344 0.334538


RESULTS The results show clearly that the internationalization–performance relationship varies significantly depending on whether internationalization is intraregional or extra-regional. Surprisingly XSTS showed the most significant effect and showed positive relationships with ROA in all three models. FSTS showed no significant relationships with ROA and intra-regional sales showed only a slightly significant relationship in one model. Tables 3–5 show the results using FSTS as a measure of internationalization. As can be seen, no significant relationships were found between FSTS and performance in any of the models and the R2 figures of all models were relatively low at around 0.13. Tables 6–8 show the results using ISTS as a measure of internationalization. As with FSTS no significant relationships between ISTS and ROA were found, except in the case of the quadratic model where ISTS and ISTS2 showed slightly significant effects. However, the R2 figures for all models at around 0.12 were lower than those for the models using FSTS.


351

Table 7. Results of Quadratic Regression of Percentage of Intra-Regional Sales on ROA. Independent Variable

Coefficient

India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE ISTS ISTS2 cons

0.1835298 0.0997215 0.2421984 0.1948098 0.3070729 0.2285517 (reference category) 0.0921166 0.0902188 0.1332039 0.1438601 0.1166491 (reference category) 0.151896 0.0943322 0.1135314 0.0129315 0.2179841 0.2561392 0.3505462

Note: R2=0.1253. Significant at 0.05 level. Significant at 0.01 level. Significant at 0.005 level.

The best fit was obtained with models using XSTS as a measure of internationalization. Tables 9–11 show the results for each of the three models. As can be seen the R2 figures for all models are significantly higher than the R2 figures for the models using FSTS or ISTS, ranging from 0.42 in the linear regression model to 0.50 in the quadratic regression model and 0.52 in the cubic regression model. There are also highly significant relationships with XSTS, XSTS2 and XSTS3.

DISCUSSION The results strongly suggest that may well be country or regional differences in the nature of the relationship between internationalization and performance. Contrary to Hypothesis 1 and previous studies by Contractor et al.

352

Table 8.

STEPHEN CHEN

Results of Cubic Regression of Percentage of Intra-Regional Sales on ROA.

Independent Variable India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE ISTS ISTS2 ISTS3 cons

Coefficient 0.1825161 0.1033272 0.2448808 0.1946388 0.3102459 0.2285174 (reference category) 0.0921646 0.0908686 0.128252 0.1452905 0.1144984 (reference category) 0.1545412 0.0959101 0.1140822 0.0135509 0.0623853 0.3493251 0.4970987 0.35666


(2003) and Lu and Beamish (2004), this study did not find support for a significant effect of internationalization on firm performance when internationalization was measured using combined FSTS. However, our results are consistent with the results of Nachum (2004), who also found no significant relationship between internationalization and performance in her sample of firms Southeast Asia. As she noted, one reason could be the considerable heterogeneity of the economies in the region or the fast changing economic conditions in the region which may mask any firm-level effects. This suggests a need to control more for macro-economic factors in future studies. The results also show that there are significant differences depending on whether internationalization takes place within the region or outside the region. Contrary to Hypothesis 2 the study found no significant relationship between internationalization as measured by FSTS and ROA but differences when internationalization was measured by ISTS and XSTS.


Table 9.

353

Results of Linear Regression of Percentage of Extra-Regional Sales on ROA.

Independent Variable India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE XSTS cons

Coefficient 0.0713951 0.0692256 0.1096293 0.085398 0.1050168 0.0868313 (reference category) 0.0217002 0.0268114 0.0212786 0.0370065 0.0192328 0.0971415 0.0460469 (reference category) 0.0150484 0.0025859 0.1260763 0.0904998

Note: R2=0.4184. Significant at 0.005 level.

Even more unexpectedly, and contrary to Hypothesis 3, there was little or no significant relationship with percentage of sales within the region but there was a highly significant relationship between the percentage of extraregional sales and performance. One possible reason could be that the sample examined in this study was drawn from the largest firms in each country and industry sector. Since such firms are likely to be more profitable they may be more likely to expand outside the region. In other words, performance could affect the degree and scope of internationalization rather than the other way round. Given the relatively small size of the sample and the small number of time periods, it was not possible to determine statistically the direction of the relationship from the present data but an examination of some firms in the sample suggests that this is plausible. For instance, the largest market for many of the software firms in the sample not surprisingly is the US where many successful firms such as WIPRO and Infosys have established operations.

354

STEPHEN CHEN

Table 10. Results of Quadratic Regression of Percentage of Extra-Regional Sales on ROA. Independent Variable

Coefficient

India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE XSTS XSTS2 cons

0.0521404 0.0146544 0.038288 0.0184602 0.03851 0.0215184 (reference category) 0.0242433 0.0446378 0.0243408 0.0278032 0.0162101 0.0948185 0.0448431 (reference category) 0.0091175 0.0033812 0.0534882 0.0188986 0.044094


CONCLUSION To summarize this adds two new dimensions to existing research on the relationship between internationalization and performance. Firstly and most importantly, it shows that there are significant differences in the internationalization–performance relationship depending on whether internationalization takes place within the region or outside the region. This suggests a need to more comprehensively consider the macroeconomic conditions under which internationalization takes place. Secondly, it highlights a need to more closely examine the direction of the internationalization–performance relationship in future studies, particularly in firms from developing economies such as many in Asia. High performance may trigger or drive internationalization rather than the other


Table 11.

355

Results of Cubic Regression of Percentage of Extra-Regional Sales on ROA.

Independent Variable India Indonesia Malaysia Philippines Singapore Taiwan Thailand i1 i2 i3 i4 i5 i6 i7 i8 i9 SIZE XSTS XSTS2 XSTS3 cons

Coefficient 0.0442035 0.0270125 0.0193829 0.0152039 0.0261179 0.0127973 (reference category) 0.0147556 0.036538 0.0548859 0.0406479 0.0172025 0.0840591 0.0509683 (reference category) 0.0037042 0.00377 0.0321039 0.0203437 0.0028078 0.0060042


way round. Therefore, more attention to firm-level factors such as motives for internationalizing may also be warranted. Clearly this study has limitations and there are several opportunities for further research. The sample of firms is relatively small and there are few examples of firms from some countries such as the Philippines or some sectors such as utilities. Secondly, although we controlled for firm size, for reasons of data availability, our sample was drawn from the largest firms in the world. Results may be different for small and medium-sized enterprises (SMEs) or firms that are not among the world’s largest firms. Thirdly, it should be noted that our sample of LDC firms only includes firms from Asia. While this controls for home country regional differences, it is possible that different results may be obtained with firms from other regions such as South America or Africa.

356

STEPHEN CHEN

REFERENCES Al-Obaidan, A. M., & Scully, G. W. (1995). The theory and measurement of the net benefits of internationalization: The case of the international petroleum sector. Applied Economics, 27, 231–239. Amsden, A., & Hikino, T. (1994). Project execution capability, organisational know-how and conglomerate corporate growth in late industrialisation. Industrial and Corporate Change, 3, 111–147. Baden-Fuller, C. W., & Stopford, J. (1990). Globalization frustrated: The case of white goods. Strategic Management Journal, 12, 493–507. Beck, N., & Katz, J. N. (1995). What to do (and not to do) with time-series cross-section data. American Political Science Review, 89, 634–647. Boddewyn, J. J., Halbrich, M. B., & Perry, A. C. (1986). Service multinationals: Conceptualisation, measurement and theory. Journal of International Business Studies, 17(3), 41–57. Bu¨hner, R. (1987). Assessing international diversification of West German corporations. Strategic Management Journal, 8, 25–37. Campbell, A. J., & Verbeke, A. (1994). The globalization of service multinationals. Long Range Planning, 27(2), 95–102. Capar, N., & Kotabe, M. (2003). The relationship between international diversification and performance of service firms. Journal of International Business Studies, 34(4), 345–355. Caves, R. E. (1982). Multinational enterprise and economic analysis. Cambridge, UK: Cambridge University Press. Collins, J. M. (1990). A market performance comparison of U.S. firms active in domestic, developed and developing countries. Journal of International Business Studies, 21, 271–287. Contractor, F. J., Kundu, S. K., & Hsu, C.-C. (2003). A three-stage theory of expansion of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34(1), 5–18. Daniels, J. D., & Bracker, J. (1989). Profit performance: Do foreign operations make a difference? Management International Review, 29, 46–56. Dunning, J. H. (1989). Transnational corporations and the growth of service: Some conceptual and theoretical issues. UNCTC Current Studies, Series A, No. 9. New York: United Nations Publications. Dunning, J. H. (1993). Multinational enterprises and the global economy. Wokingham, UK: Addison-Wesley. Errunza, V. R., & Senbet, L. W. (1984). International corporate diversification, market valuation and size-adjusted evidence. Journal of Finance, 39, 1325–1344. Feketekuty, G. (1988). International trade in services: An overview and blueprint for negotiations. Cambridge, MA: Ballinger. Galbraith, J. R., & Kazanjian, R. (1986). Strategy implementation. St Paul, MN: West Publishing. Geringer, J. M., Beamish, P. W., & Da Costa, R. C. (1989). Diversification strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10, 109–119. Ghemawat, P. (2003). Semiglobalization and international business strategy. Journal of International Business Studies, 34, 138–152. Ghoshal, S. (1987). Global strategy: An organizing framework. Strategic Management Journal, 8, 425–440.


357

Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between internationalization and performance. Journal of International Business Studies, 30, 173–188. Gongming, Q. (1998). Determinants of profit performance for the largest US firms, 1981–92. Multinational Business Review, 6(2), 44–51. Grant, R. M. (1987). Internationalization and performance among British manufacturing companies. Journal of International Business Studies, 18, 79–89. Griffiths, A., & Zammuto, R. F. (2005). Institutional governance systems and variations in national competitive advantage: An integrative framework. Academy of Management Review, 30(4), 823–842. Guillen, M. F. (2000). Business groups in emerging economies: A resource-based view. Academy of Management Journal, 43(3), 362–380. Habib, M. M., & Victor, B. (1991). Strategy, structure, and performance of US manufacturing and service MNCs: A comparative analysis. Strategic Management Journal, 12(8), 589–606. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product diversified firms. Academy of Management Journal, 40, 767–798. Hymer, S. H. (1976). A study of direct foreign investment. Cambridge, MA: MIT Press. Johanson, J., & Vahlne, J. (1977). The internationalization process of the firm: A model of knowledge development and increasing foreign commitments. Journal of International Business Studies, 8, 23–32. Katrishen, F. A., & Scordis, N. A. (1998). Economies of scale in services: A study of multinational insurers. Journal of International Business Studies, 29(2), 305–324. Khanna, T., & Palepu, K. (2000). The future of business groups in emerging markets: Long-run evidence from Chile. Academy of Management Journal, 43(3), 268–285. Khanna, T., & Rivkin, J. W. (2001). Estimating the performance effects of business groups in emerging markets. Strategic Management Journal, 22(1), 45–74. Kim, W. C., Hwang, P., & Burgers, W. P. (1989). Global diversification strategy and corporate profit performance. Strategic Management Journal, 10, 45–57. Knight, G. (1999). International services marketing: Review of research, 1980–1998. Journal of Services Marketing, 13(4/5), 347–361. Kogut, B. (1985). Designing global strategies: Profiting from operational flexibility. Sloan Management Review, 26, 27–38. Kumar, M. S. (1984). Growth acquisition and investment: An analysis of the growth of industrial firms and their overseas activities. Cambridge, UK: Cambridge University Press. Leff, H. N. (1978). Industrial organisation and entrepreneurship in the developing countries: The economic groups. Economic Development and Cultural Change, 4, 661–675. Leff, H. N. (1979). Entrepreneurship and economic development: The problem revisited. Journal of Economic Literature, 17(1), 46–64. Lovelock, C. H., & Yip, G. S. (1996). Developing global strategies for service business. California Management Review, 38(2), 64–86. Lu, J. W., & Beamish, P. W. (2004). International diversification and firm performance: The S-curve hypothesis. Academy of Management Journal, 47, 598–609. Mauri, A., & Sambharya, R. (2001). The impact of global integration on MNC performance: Evidence from global industries. Management International Review, 10, 441–454.

358

STEPHEN CHEN

Michel, A., & Shaked, I. (1986). Multinational corporations vs. domestic corporations: Financial performance and characteristics. Journal of International Business Studies, 17(3), 89–100. Murray, J. Y., & Kotabe, M. (1999). Sourcing strategies of US service companies: A modified transaction-cost analysis. Strategic Management Journal, 20(9), 791–809. Nachum, L. (2004). Geographic and industrial diversification of developing country firms. Journal of Management Studies, 41(2), 274–294. Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance. New York: Free Press. Prahalad, C. K., & Doz, Y. (1987). The multinational mission: Balancing local demands and global vision. New York: Free Press. Rugman, A. M. (1979). International diversification and the multinational enterprise. Lexington, MA: Heath. Rugman, A. M. (2000). The end of globalization: Why global strategy is a myth & how to profit from the realities of regional markets. New York: Amacom. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43(1), 61–84. Sambharya, R. B. (1995). The combined effect of international and product diversification strategies on the performance of U.S.-based multinational corporations. Management International Review, 35(3), 197–218. Schlie, E., & Yip, G. (2000). Regional follows global: Strategy mixes in the world automotive industry. European Management Journal, 18(4), 343–354. Shaked, I. (1986). Are multinational corporations safer? Journal of International Business Studies, 17, 75–80. Siddharthan, N. S., & Lall, S. (1982). Recent growth of the largest US multinationals. Oxford Bulletin of Economics and Statistics, 44, 1–13. Singh, A. (1995). Corporate financing patterns in industrialising economies. Technical Paper No. 2. International Finance Corporation, Washington, DC. Stata. (2003). Stata cross-sectional time series: Reference manual, release 8. College Station, TX: Stata Press. Sullivan, D. (1994). Measuring the degree of internationalization of a firm. Journal of International Business Studies, 25, 325–342. Tallman, S., & Li, J. (1996). Effects of international diversity and product diversity on the performance of multinational firms. Academy of Management Journal, 39(1), 179–196. Tolentino, P. E. E. (1993). Technological innovation and Third World multinationals. New York: Routledge. Vermeulen, F., & Barkema, H. (2002). Pace, rhythm, and scope: Process dependence in building a profitable multinational corporation. Strategic Management Journal, 23, 637–653. Vernon, R. (1966). International investment and international trade in the product cycle. Quarterly Journal of Economics, 80, 190–207. Vernon, R. (1971). Sovereignty at bay: The multinational spread of U.S. enterprises. New York: Basic Books. Whitley, R. (1999). Divergent capitalisms: The social structuring and change of business systems. Oxford, UK: Oxford University Press. Zaheer, S., & Mosakowski, E. (1997). The dynamics of the liability of foreignness. Strategic Management Journal, 18(6), 439–464.

INTRA-REGIONAL SALES AND THE INTERNATIONALIZATION AND PERFORMANCE RELATIONSHIP Nicole Richter ABSTRACT This study provides a deeper insight into the performance effects of internationalization of the most international multinational enterprises (MNEs). Most MNEs perform their business activities within their homeregional block of the world – North America, Europe or the Asia-Pacific block. Whether these regional strategies pay off is explored by means of two analyses: first, the impact of internationalization in terms of the transnationality index, and second, the impact of foreign intra-regional sales on performance is examined. Results indicate that regional strategies smooth performance declines in the early stages of internationalization but also smooth performance increases during a phase of high-foreign expansion.

INTRODUCTION Over the past three decades, the relationship between internationalization and firm performance has been subject to extensive research. Researchers


359

360

NICOLE RICHTER

from international business or management who examined this relationship have produced mixed arguments and results. Early research focused on the theoretical benefits of internationalization, and thus, attempted to prove empirically a monotonic positive effect on performance (Caves, 1971; Vernon, 1971; Errunza & Senbet, 1984; Bu¨hner, 1987; Benvignati, 1987; Kim & Lyn, 1987). Today, sound theoretical hypotheses which take the interaction of benefits and costs into account argue in favor of a nonlinear relationship. Empirical analyses, however, yielded both evidence for U-shaped as well as inverted U/J-shaped relationships. Moreover, recent findings support a horizontal S-curve (Al-Obaidan & Scully, 1995; Hitt, Hoskisson, & Kim, 1997; Gomes & Ramaswamy, 1999; Ruigrok & Wagner, 2003; Contractor, Kundu, & Hsu, 2003). One possible cause of these conflicting findings which was shortly emphasized is a lack of investigation into the regions targeted in the internationalization process. In fact, most of today’s multinational enterprises (MNEs) do not follow a global but a regional strategy that concentrates on the countries in their home leg of the triad North America, Europe and the Asia-Pacific (AP) region (Rugman & Verbeke, 2004a; Rugman & Verbeke, 2004b; Rugman, 2005). Taking this locus of destination into consideration, this study aims at further illuminating and resolving the controversy surrounding the different shapes of the internationalization and performance relationship. Following this objective the paper first offers a brief review of theoretical frameworks and past empirical findings in order to derive hypotheses of curvilinearity in the next section. This is followed by an explanation of the research methodology and presentation of results. A summary of the main conclusions will close the article.

PREVIOUS FINDINGS AND THEIR THEORETICAL FOUNDATION The Benefits and Costs of Internationalization Theoretical frameworks describing the internationalization and performance relationship differ with regard to the emphasis on benefits or costs of internationalization, their interplay and development over time. Two main theoretical perspectives that specify the benefits of internationalization can be identified: the first theory stream concentrates on the motives for foreign direct investment of enterprises, such as resource-, market-, efficiency- and strategic-asset-seeking. Economies of scale and scope as well as an improved

Intra-Regional Sales and the Internationalization and Performance Relationship 361

risk-return performance through portfolio diversification are the major benefits claimed by scholars in this field. The second strand of research adopts a more managerial perspective and stresses internationalization benefits in the company itself and not from external opportunities. Here, internationalization driven company-specific competencies are promoted and benefits arise from organizational learning, knowledge development and more operational flexibility (Rugman, 2002; Buckley & Casson, 1976; Fayerweather, 1978; Wernerfelt, 1984; Kogut, 1989). The costs of internationalization can also be looked at from external and internal perspectives: focusing on the company’s environment, costs can accrue due to political uncertainty such as unanticipated changes to the business environment, and financial risks such as exchange-rate fluctuations or inflation. Referring to the intra-company situation, increased costs stem from higher information-processing demands, and transaction costs which are highly influenced by a corporation’s geographic spread and cultural diversity (Boddewyn, 1988; Reeb, Kwok, & Beak, 1998; Teece, 2002; Hofstede, 1980; Graham, 2001). Companies face both the benefits and the costs of internationalization – the interpretation of their interaction offers theoretical support for various empirical findings, as will be seen in the next section.

Past Empirical Research Empirical studies on the internationalization and performance relationship have produced contradicting results. In the past, research especially focused on a monotonic linear relationship between these two variables: most researchers found a positive (Vernon, 1971; Errunza & Senbet, 1984; Benvignati, 1987; Bu¨hner, 1987; Grant, 1987; Kim & Lyn, 1987; Grant, Jammine, & Thomas, 1988; Kim, Hwang, & Burgers, 1989; Tallman & Li, 1996), fewer a negative (Siddharthan & Lall, 1982; Michel & Shaked, 1986; Chang & Thomas, 1989; Collins, 1990) or no internationalization and performance relationship (Haar, 1989; Morck & Yeung, 1991). Early investigators concentrated on possible benefits and paid less attention to the costs of internationalization, and therefore tested a rather positive linear relationship. Among these authors is Grant (1987) who realizes that the success of internationalization is influenced by the cultural and geographical distance or the business situation in the host country, but stresses the following benefits of international activity: returns to intangible assets, market power conferred by international scope, the capacity to undertake risky investments and the broadening of investment opportunities.

362

NICOLE RICHTER

Since the late 1980s, researchers started to take the interplay between benefits and costs of internationalization more into consideration and explored curvilinear relationships. Results which support a nonlinear relation are, however, likewise inconsistent: the types of nonlinear relationships found range from the most prevalent inverted U/J-shaped curve (Daniels & Bracker, 1989; Geringer, Beamish, & Da Costa, 1989; Sullivan, 1994b; Ramaswamy, 1995; Al-Obaidan & Scully, 1995; Gomes & Ramaswamy, 1999; Hitt et al., 1997), to a U-curve (Lu & Beamish, 2001; Ruigrok & Wagner, 2003; Ruigrok & Wagner, 2006), to multiple waves or a horizontal S-curve (Hitt, Hoskisson & Ireland, 1994; Riahi-Belkaoui, 1998; Contractor et al., 2003). Daniels and Bracker (1989) who found an inverted U-shaped relationship agree with Grant (1987) as regards the benefits of internationalization. However, they argue that there is an optimal point after which the marginal costs of internationalization will exceed its marginal returns. The rationale behind it is that the costs of managing information and resource flows in the company escalate at a faster rate than returns with growing corporate complexity. Their reasoning is supported by the assumption of an internationalization path of corporations from culturally familiar to unfamiliar and geographically close to far countries. The entry into culturally unfamiliar or geographically far regions aggravates the information and resource management, while at the same time diminishes internationalization benefits stemming from similarities such as similar customer needs. Riahi-Belkaoui (1998), in contrast, found a horizontal S-curve. While he accepts the theoretical explanations above, i.e. an inverted U-shaped curve, he argues that a corporation’s entry into international business is costly up to a first level of internationalization, leading to a first downturn before passing into the inverted U-shape. Contractor et al. (2003) as well as Lu and Beamish (2004) offer a more thorough explanation of the S-curve relationship by differentiating three stages of internationalization as illustrated in Fig. 1. Early internationalizers face large so-called liabilities of foreignness, because of unfamiliarity with foreign markets, cultures and environments. These result in costs which can outweigh the benefits realized in the first stages of internationalization. The costs attributed to foreignness are mitigated in the following stages of internationalization due to organizational learning effects. Moreover, a growing geographic diversification increases the opportunities of exploiting internationalization benefits such as better access to low-cost inputs and a greater variety of market opportunities. Although, the coordination costs begin to rise during the mid-stage of internationalization, the incremental benefits of further

Performance


Early Internationalizers

Mid-Stage Internationalizers

High Internationalizers

Liability of foreignness Initial learning costs Insufficient economies of scale

Ressource augmentation/ exploitation Internalization of transaction costs Economies of scale and scope Extension of product life cycle Access to lower cost resources

Cultural distance Coordination costs due to very dispersed markets Expansion into peripheral markets

Internationalization

Fig. 1. A Three-Phase Model of the Internationalization and Performance Relationship. Source: Own illustration, based on Contractor et al. (2003, p. 7).

international expansion exceed the incremental costs. This relationship of benefits and costs turns around in a stage of high internationalization, as also stressed by Daniels and Bracker (1989). MNE’s profit potential declines because they already have expanded into the most lucrative markets, and coordination costs escalate due to the growing complexity of global operations (Contractor et al., 2003; Lu & Beamish, 2004).

RESEARCH HYPOTHESES The theoretical foundation of an inverted U-shaped internationalization and performance relationship is most convincing, although there might be a first downswing in the early stages of internationalization. Whether there are arguments in favor of a downswing and how benefits and costs interplay is highly influenced by the internationalization pattern of MNEs. What is crucial is the question, whether corporations follow a regionalization or globalization strategy. MNEs pursuing a regional strategy primarily concentrate on international activities within a geographically close or familiar region and do not seek the benefits of worldwide expansion. The work of Rugman (2005) on the ‘‘Regional Multinationals’’ offers strong

364

NICOLE RICHTER

support for prevalent regional strategies of MNEs. He found that 320 out of 380 of the world’s top MNEs, i.e. 84.2%, are the so-called home-region oriented companies. Following the differentiation of the world into the triad, a home-region oriented corporation generates more than 50% of its sales in its home triad. In contrast to the literature on international marketing (e.g. Meffert & Bolz, 1998; Backhaus, Bu¨schken, & Voeth, 2003), Rugman and Verbeke (2004a) stress that globalization in terms of a balanced geographic distribution of sales across the triad is a rather unusual phenomenon. Since there is still an essential distance separating North America, Europe and Asia as regards culture, administrative and economic roots, MNEs which pursue a regional strategy are confronted with less liabilities of foreignness in the early stages of internationalization compared to their globalizing competitors. By choosing familiar settings, corporations can exploit internationalization benefits without huge cost increases. The concentration on familiar markets reduces a possible information lack regarding the political or business situation. In addition, informationprocessing, distribution or administration costs are kept low due to similar administrative mechanisms, consumer tastes, attitudes to work or languages. In the mid-stages of internationalization, however, global corporations have a greater potential to exploit internationalization benefits in comparison to regional MNEs. Regional MNEs face a faster declining profit potential as they already have expanded into the most lucrative markets in the region. In the last stages of internationalization, the costs of managing information and resource flows in the company escalate with growing corporate complexity at a faster rate than returns. This is aggravated in the case of global corporations. Raised management demands due to growing internationalization are intensified by the difficulties of managing a multi-cultural workforce and serving distinctly different customer markets. Moreover, highly differing regional peculiarities with respect to the business environment have to be taken into consideration (Rugman & Verbeke, 2007, p. 201; Erikkson, Johanson, Majkgard, & Sharma, 2000, p. 29; Johanson & Vahlne, 1990, p. 13; Gomes & Ramaswamy, 1999, pp. 174–178; Geringer et al., 1989, p. 112; Papadopoulos & Denis, 1988, p. 44). Thus, the relationship between internationalization and performance is more likely to show a first downturn in the case of global internationalization. Following the argument that internationalization into geographically or culturally distant locations increases costs, concentrating on close areas mitigates possible internationalization driven cost increases.


Internationalization benefits can, however, possibly not be exploited to the full extent. Thus, the following research hypotheses are outlined: Hypothesis 1. The relationship between internationalization and performance is curvilinear. After a first downswing of performance in the first stages of internationalization, performance is increased up to an optimal level beyond which higher degrees of internationalization again lead to a decrease in performance. Hypothesis 2. The relationship between regionalization and performance does not show a first downswing of performance. It follows an inverted U shape. However, performance increases at a lower pace up to an optimal level beyond which performance declines less steep compared to the case of international expansion throughout the world. To verify the hypothesized relationships, linear, quadratic as well as cubic functions between internationalization and performance will be tested using the following research methodology.

RESEARCH METHODOLOGY The next section discusses the operationalization of performance as the dependent variable, the internationalization measures used as independent variables and the three control measures which are included.

Variables and Measures Performance is measured from an accounting-based view via the return on assets (ROA). It is based on after tax income, in order to take possible tax advantages of international operations into account. Due to its frequent application in previous studies, the ROA allows for cross-study comparisons (Bu¨hner, 1987; Haar, 1989; Daniels & Bracker, 1989; Geringer et al., 1989; Hitt et al., 1997; Gomes & Ramaswamy, 1999; Wan & Hoskisson, 2003). Researchers have identified a considerable variety of measures in order to specify the independent variable of this study, a firm’s degree of internationalization. The most adequate indicators are the foreign sales to total sales ratio (FSTS), the foreign shares of assets (FATA) and employees (FETE). The most commonly used measure, the FSTS, can be described as a proxy of a company’s foreign market dependence, whereas the FATA and

366

NICOLE RICHTER

FETE cover a company’s foreign production presence (Annavarjula & Beldona, 2000; Hassel, Ho¨pner, Kurdelbusch, Rehder, & Zugeho¨r, 2003; Sullivan, 1994a). A useful measure comprising all three ratios is the index of transnationality (TNI) introduced by the United Nations Conference on Trade and Development (UNCTAD). It is derived as the average of the three ratios (UNCTAD, 1995, pp. 19–26; Holden, 2005, pp. 203–204; Be´langer, Giles, & Grenier, 2003, p. 472; Fisch & Oesterle, 2003, pp. 12–13; Pheng, Hongbin, & Leong, 2004, pp. 719, 722). The suitability of this index was confirmed via the following analyses: after a verification that all ratios show close mean values and standard deviations during the three years of the study (2001–2003), the TNI’s reliability was tested using Cronbach’s alpha. All alpha coefficients exceed the recommended threshold of 0.7 (0.9 in 2001, 0.8 in 2002 and 0.9 in 2003), and thus reinforce the reliability of the TNI (Nunnally, 1978, p. 245; Santos, 1999, p. 2; Streiner, 2003, p. 103). Finally, an exploratory factor analysis was carried out. It supports the assumption of one underlying construct ‘‘internationalization’’ behind the three ratios, as one single factor with an eigenvalue between 1.918 and 2.319 during the 2001–2003 period was extracted. The second indicator used in the analysis captures the locus of destination. When analyzing the impact of regionalization, a company’s sales in its home-triad region are related to its total sales. This variable is called the home-region to total sales ratio (R/T). The R/T can further be divided into the domestic sales and foreign sales in the rest of the home-region (ROR). Moreover, Rugman’s classification of companies into home-region and nonhome-region oriented (i.e. host-region oriented, bi-regional or global) companies is pursued. Host-region oriented companies derive – in contrast to home-region oriented corporations – at least 50% of their sales from a foreign triad region (Rugman, 2005, p. 10). In addition to examining the impact of the ROR on performance the subgroup of home-region oriented companies is investigated in detail. The analysis is supplemented by three control variables: the first two control variables which have a theoretical and empirically proven impact on firm performance are organizational size and industry membership (Buckley, Dunning, & Pearce, 1978; Kumar, 1984; Hsu & Boggs, 2003). Moreover, the region of origin is controlled for in order to take the importance of the locus of origin which was lately stressed by Wan and Hoskisson (2003) into account. Size is measured as the natural logarithm of total employees in order to meet the requirement of a normal distribution of variables in the following regression analysis. Industry effects are represented by the industry’s average ROA, and the region of origin is operationalized via dummy variables.


Sample The sample is drawn from the UNCTAD, which annually publishes the 100 most internationalized non-financial corporations ranked by foreign assets. Thus, this sample is biased toward corporations for which it can be assumed that they possess the necessary knowledge to successfully work in a foreign market. To smooth annual fluctuations in the accounting data, a three-year average from 2001 to 2003 is used. Data on FSTS, FATA and FETE are taken from the corresponding World Investment Reports (UNCTAD, 2005; UNCTAD, 2004; UNCTAD, 2003). The performance numbers needed as well as data on regional sales is retrieved from the Hoover’s and the Thomson Research databases. The corporations included in the initial UNCTAD ranking come from 18 different home economies. The home economies accounting for the largest numbers of companies are the USA (30), France (17), the UK (16), Germany (14) and Japan (8). These 85 corporations form the final sample, in order to keep the variance resulting from different home countries low. Moreover, 29 industries are found in the sample, which can – in order to give a review – further be assigned to 12 industry groups: telecommunications and utilities (13), chemicals and pharmaceuticals (13), energy (8), construction and building materials (6), machinery and equipment (4), motor vehicle and parts (9), computer, office and electronics (7), food, tobacco and textiles (9), diversified (2), media (4), merchandise (6) and other services (4). Table 1 gives an overview of the internationalization patterns and performance ratios found for different subgroups in the sample. As can be seen, European corporations are most internationalized in terms of the average TNI and statistical testing found that the European averages differ significantly from the American and AP values. In addition to the different size of the home markets, this could be due to the fact that the major expansion of European companies already started in the 1920s and 1930s, while US corporations enjoyed their main overseas thrust after World War II, and Japanese corporations even later in the 1970s (Bartlett & Ghoshal, 2000, pp. 508–510). As regards regionalization, US firms report the highest, whereas European firms show the lowest R/T. In further analysis, it was found that US firms show the highest and European firms the lowest percentage of home-region oriented firms – which corresponds to Rugman’s findings (Rugman, 2005). Thus, a regionalization strategy is more prevalent among US corporations. In addition to the intra-regional sales, R/T, the shares of sales in the ROR, i.e. the foreign sales within the home-region, are given in parentheses. US corporations generate for instance only 10% of

368

Table 1.

NICOLE RICHTER

Internationalization and Performance Ratios of the Sample.

2001–2003

Number of Firms

Average TNI

Average R/T (% Foreign=ROR)

Average ROA

Triad region North America (here: US)

30

0.442

0.033

Europe

47

0.614

France

17

0.633

Germany

14

0.465

UK

16

0.725

8

0.432

0.634 (10%) 0.578 (48%) 0.643 (43%) 0.693 (39%) 0.407 (53%) 0.605 (19%)

Company orientationa Home-oriented Non-Home-oriented Host-oriented Bi-regional Global

56 28 5 19 3

0.468 0.656 0.658 0.676 0.571

0.710 0.379 0.346 0.388 0.371

0.019 0.053 0.065 0.047 0.066

Total

85

0.536

0.600

0.026

AP (here: Japan)

0.024 0.006 0.019 0.046 0.019

Source: Data on TNI: UNCTAD (2005, pp. 267–269), UNCTAD (2004, pp. 276–280) and UNCTAD (2003, pp. 187–188); data on R/T, ROA: Hoover’s and Thomson Research databases. a For one corporation was no data on the regional spread of sales available.

their intra-regional sales from foreign countries within the home triad whereas the major part (90%) of the R/T is yielded at home, which brings to bear the importance of the home economy size. Concerning the average performance, US firms are the best performers followed by European and Japanese corporations. Among the Europeans, corporations from the UK outperform German as well as the worst-performing French corporations. Drawing the attention to the company orientation in the second part of Table 1, 66% of the corporations of this sample are home-region oriented with respect to intra-regional sales. Rugman and Verbeke (2004b) who have analyzed the same sample found that even those companies which are nonhome-region oriented as regards their sales exhibit some regional features in corporate strategy or structure. The non-home-region oriented corporations in the sample reach significantly better performance ratios than home-region


oriented corporations. As will be seen later in the analysis, this can be attributed to the internationalization phase they are in: a strategy focusing on intra-regional sales does not pay off for these corporations in the later stages of internationalization. Analysis Technique Hypothesis testing is done using a classical regression analysis on the threeyear averaged data of the 2001–2003 period. In order to test the assumed cubic relationship between internationalization and performance (Hypothesis 1), regression equation (Eq. (1)) is used. ROA ¼ b0 þ b1 TNI þ b2 TNI2 þ b4 TNI3 þ b4 SIZE þ b5 IND þ

X

bi REGi þ e (1)

where ROA, return on assets; TNI, transnationality index; SIZE, natural logarithm of total employees; IND, average industry return on assets; REG, two dummy variables (US, AP) for three regions of origin (US, EU and AP); b1–b7, regression coefficients; e, residual. Performance is measured by the ROA. The independent research variable is the TNI. The three control variables are company size measured by the natural logarithm of total employees, industry which is included via the average industry return on assets and country of origin which is operationalized via two dummy variables. The cubic relationship is tested versus a possible linear as well as a quadratic relationship. While the above regression is applied in order to examine the cubic model, a quadratic relationship is explored with the same regression, except for omitting the cubic term (b3TNI3). As a consequence, for exploring linearity the cubic and quadratic terms (b3TNI3 and b2TNI2) are left out. The hypothesis will be considered confirmed if the following requirements are met: firstly, the cubic model has a significant adjusted coefficient of determination (R2), secondly, the regression coefficient of the cubic term of the internationalization variable is significant, and finally, the cubic model shows a higher R2 compared to the linear and quadratic model. An equivalent methodology is applied in order to investigate the impact of regionalization on performance. The relationship between ROR and performance is supposed to be quadratic rather than cubic (Hypothesis 2). An analogous regression equation is used. ROA ¼ b0 þ b1 ROR þ b2 ROR2 þ b3 ROR3 þ b4 SIZE þ b5 IND þ

where ROR, foreign sales within the home-triad region.

X

bi REGi þ e (2)

370

NICOLE RICHTER

While the ROA is again the performance proxy, the independent variable now is the ROR. The same control variables as given in regression equation (Eq. (1)) are used. The quadratic model is also tested versus linearity and a possible cubic relationship, resulting in another three regression analysis. Results will validate the hypothesis of a quadratic relationship, if the aboveformulated requirements are met with respect to the quadratic model. In Hypothesis 2, not only a quadratic relationship between internationalization within the home triad and performance was assumed, but also the relationship was supposed to be less steep: performance increases before reaching the optimum are less steep and the optimum is followed by a slower downturn of performance. This assumption is analyzed, too. For this purpose, a detailed exploration of the internationalization and performance relationship of home-region oriented corporations is implemented.

RESULTS Before an in-depth discussion of results, Table 2 reports means, standard deviations and correlations of all variables. Both the correlations as well as other diagnostic tests did not indicate problems of multi-collinearity between the independent variables (TNI, ROR, SIZE and IND). However, there was strong multi-collinearity between the linear, squared and cubic terms of the TNI and ROR. Therefore, the TNI and ROR values have been centered via subtracting their respective mean values before building their squared and cubic terms. After this procedure, all variance inflation factors showed values below 4 indicating no further problems of multi-collinearity Table 2.

Descriptive Statistics and Correlations.

Variables

Mean

S.D.

TNI

ROR

SIZE

IND

ROA TNI ROR SIZE IND

0.026 0.536 0.174 5.032 0.031

0.071 0.188 0.156 0.379 0.051

0.058

0.102 0.440

0.091 0.306 0.058

0.301 0.110 0.181 0.088

Number of observations: 255, Number of corporations: 85. SIZE, logarithm of total employees; IND, industry average return on assets. *po0.10. po0.05. po0.01.


(Urban & Mayerl, 2006, pp. 232, 238–241). A test for normal distribution and autocorrelation as well as analyzing the residual plots did not reveal any further violations of regression premises. Internationalization and Performance An examination of the assumption of a cubic relationship between the TNI and ROA (Hypothesis 1) yields the following key results which are discussed in detail below: The cubic model is clearly superior to the linear and quadratic model. The cubic model is able to explain 51% of total variance. Thus, the internationalization and performance relationship is best explained by the assumed horizontal S-curve. The optimal level of internationalization is given at a TNI of 60%. Table 3 summarizes the values of the standardized regression coefficients, the adjusted R2 values as well as the statistical levels of significance of the three models tested. Table 3.

The Internationalization and Performance Relationship. ROA Linear

Quadratic

Cubic

0.116

0.176 0.237

0.293 0.399 0.608

Control variables SIZE INDUSTRY REGION: US REGION: AP

0.113 0.639 0.037 0.047

0.081 0.597 0.017 0.046

0.035 0.614 0.051 0.003

Adjusted R2 F-statistics

0.371 10.929

0.417 10.999

0.514 13.701

Research variables TNI TNI2 TNI3

Number of observations: 255, Number of corporations: 85. SIZE, logarithm of total employees; IND, industry average return on assets. po0.10. po0.05. po0.01.

372

NICOLE RICHTER

The quadratic model yields an adjusted R2 of 0.417 which already is a high value compared to other studies in this field (Grant, 1987; RiahiBelkaoui, 1998; Ruigrok & Wagner, 2003; Thomas & Eden, 2004). Not only the total quadratic model, but also the quadratic term (TNI2) is highly significant with a probability of error below 0.01. Thus, the quadratic model better explains the internationalization and performance relationship compared to the linear model, which reports a R2 of only 0.371 and a non-significant linear term (TNI). However, the results of the exploration against the most recently discussed cubic model confirm Hypothesis 1 in which a horizontal S-curve relationship was assumed. The cubic model best explains the general internationalization and performance relationship with a highly significant adjusted R2 of 0.514. Among the control variables, only significant industry effects are found. Both size and region of origin are nonsignificant. The bold line in Fig. 2 depicts the relationship found. Performance decreases during the initial internationalization process. The decline in performance can be attributed to initial learning costs due to a strong liability of foreignness, as internationalization includes also less familiar regions. This initial phase is marked by a TNI of 26% or less. Performance increases thereafter during the following internationalization stage. Organizational learning reduces cost increases and enables corporations to fully profit from their international activities up to an optimal level. This optimal level of internationalization is given at a TNI of 60% which is within the range of values (50–82%) suggested by several 0.02 0.00 Return on assets

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

-0.02 -0.04 -0.06 -0.08

all corporations home-region oriented corporations

-0.10 Transnationality index -0.12

Fig. 2.

The Internationalization and Performance Relationship.


authors (Daniels & Bracker, 1989; Ramaswamy, 1995; Gomes & Ramaswamy, 1999) who assume a rather deterministic internationalization and performance optimum. In the last stages of internationalization, especially as corporations expand into geographically and culturally distant regions, the costs of managing information and resource flows escalate. This leads to a decline in performance at higher degrees of internationalization. The dashed line in Fig. 2 describes the internationalization and performance relationship of home-region oriented corporations which will be referred to in the next section.

Foreign Intra-Regional Sales and Performance Following the proposal of Rugman and Verbeke (2004a), a deeper investigation into the impact of intra-regional sales on performance is sought and the special characteristics of the internationalization and performance relationship of home-region oriented corporations are analyzed. Rugman and Verbeke (2004b) identify several factors which strongly approve a regional strategy. Among these are different ‘‘rules of engagement’’ in different regions such as a different industry structure, a different regulatory system and a different competitive position of the firm. Moreover, dissimilar customer requirements and regional cooperation agreements such as the North American Free Trade Agreement or the European Union are stressed. In which way a regional strategy contributes to a corporation’s performance will be tested by analyzing, first, the relationship between the ROR and performance, and second, by analyzing the subgroup of home-region oriented corporations. The following key aspects of the analysis are initially highlighted: The curvilinearity hypothesis is neither clearly confirmed nor rejected. Results support a rather quadratic or cubic model compared to a linear relationship. A first downswing in the cubic function is hardly visible, and the gradient of the cubic and quadratic function is very similar. Costs initially weakening performance when entering international markets in the home triad are negligible. Concentration on the home triad offers less potential to benefit from international activities. Costs that reduce performance beyond an optimal internationalization level will have less impact on performance, if a company focuses on the home triad.

374

NICOLE RICHTER

Turning to the in-depth regression analysis, which explores the ROR effects on performance, Table 4 summarizes the results of the linear, quadratic and cubic regression model. All regression models are significant on a 0.01 level, but neither model yields significant regression coefficients for any of the ROR variables. Thus, only comparing the adjusted R2 values provides indication on the best model. Here, the quadratic model is slightly superior to both, the linear as well as the cubic model. However, the differences in the adjusted R2 values are that low that a well-founded decision in favor of either model is hardly possible: a cubic or quadratic relationship between foreign sales within the region and performance is conceivable. A distinct confirmation or rejection of Hypothesis 2 is impossible. Since results of the ROR and performance relationship are dissatisfying, further investigation is needed to gain deeper insights into the performance effects of foreign sales in the home-region. In order to provide some insight into the different performance effects of ROR, Fig. 3 depicts the ROR and ROA performance relationship for both the cubic (the bold line) as well as the quadratic model. As can be seen when comparing both models, there is no visible downswing in performance during the initial process of foreign expansion Table 4.

The Effect of Foreign Intra-Regional Sales on Performance. ROA

Research variables ROR ROR2 ROR3 Control variables SIZE INDUSTRY REGION: US REGION: AP Adjusted R2 F-Statistics

Linear

Quadratic

0.096

0.217 0.144

Cubic

0.201 0.006 0.140

0.039 0.679 0.140 0.006

0.034 0.664 0.197 0.016

0.022 0.661 0.186 0.020

0.401 10.712

0.429 10.025

0.423 8.552

Number of observations: 255, Number of corporations: 85. SIZE, logarithm of total employees; IND, industry average return on assets. po0.10. po0.05. po0.01.

Intra-Regional Sales and the Internationalization and Performance Relationship 375 0.04 0.02

Return on assets

0.00 0

0.1

0.2

0.3

0.4

0.5

0.6

-0.02 -0.04 -0.06

cubic model quadratic model

-0.08 -0.10 ROR - Foreign sales within the home triad -0.12

Fig. 3.

The Effect of Foreign Intra-Regional Sales on Performance.

within the home-region. Fewer foreign liabilities were identified in geographic and culturally close countries. These paired with efficient regional strategies prevent corporations from experiencing a performance decline during the first stages of intra-regional expansion. An optimal level of intra-regional expansion is given at a ROR of 36–40%. Thereafter, a further engagement in the home-region worsens performance. This could be attributed to saturated markets or exploited resources within the homeregion from a revenue perspective and rising needs for administration and coordination from a cost perspective. These conclusions are supported by an analysis of the subgroup of homeregion oriented corporations. Regression analysis also revealed a quadratic or cubic relationship between internationalization and performance for this subgroup. Drawing the attention back to Fig. 2 which shows the TNI and performance relationship of all corporations compared to the cubic fit of the home-region oriented corporations, it can be concluded: The first U-shaped relation within the cubic model is less incisive in the case of home orientation. This confirms the idea of lower initial costs during the first international expansion of home-region oriented corporations. However, the profit potential during the mid-stages of foreign expansion is much lower compared to non-home-region oriented corporations. These findings point to the fact that home-region oriented corporations are unable to fully profit

376

NICOLE RICHTER

from all possible internationalization benefits. However, the performance downturn after an optimal level of internationalization is much weaker within this subgroup of corporations. It can be concluded: A regional strategy in terms of intra-regional sales clearly is preferable during the first and mid-stages of internationalization. After a certain level of intra-regional activity, in this sample at a ROR and a TNI of about 40%, non-home foreign expansion is advocated in order to fully profit from further international engagement. However, as given in Rugman and Verbeke (2004b) even the corporations in this sample which can be classified as non-home-region oriented still show elements of regional strategy or structure. This mixture of inter-regional sales and regional strategy or structure seems to be adequate in the context of performance enhancement. This successful international sales and strategy mix is also the reasoning, why the non-home-region oriented MNEs in this sample achieve on average a better performance than their home-region oriented competitors. Although the home-region oriented MNEs might not have suffered from a heavy decrease in performance during their initial foreign expansion, they do not exploit all possible benefits of non-regional foreign activity. Obviously, international coordination mechanisms have to be improved in order to mitigate the strong decline in performance after a certain level of internationalization, here given at a TNI of 60%. Against the background of the sample looked at, which consists of the world’s most international corporations, two aspects are stressed: Firstly, the great success of inter-regional sales might be due to the rich experience of these corporations. Secondly, the heavy downturn of performance after an optimal level of internationalization points to the lack of efficient coordination mechanism even at the most experienced corporations in the field of overseas business.

CONCLUSIONS The actual relationship between internationalization and performance is subject to a controversial discussion in the literature. Theoretically sound arguments can be given for a curvilinear relationship and recent empirical research identifies a cubic relation. A better understanding of this relationship will not only provide insights into the interactions of underlying factors, but is also essential for improving internationalization strategies in business practice. An important factor, which has lately been introduced to the discussion, is the influence of the locus of destination.


In accordance with the more recent theories it was found that an optimal level of internationalization beyond which profitability gradually diminishes seems to exist. In line with the findings of Daniels and Bracker (1989) and Gomes and Ramaswamy (1999), it was confirmed that this optimum lies in between 50% and 82%. The results further strongly indicate that whether the progression follows a horizontal S-curve instead of a squared curvilinear relationship depends on the business strategy pursued. The regression analysis based on foreign sales within the home triad did not provide unambiguous results; however, the succeeding subgroup analysis yielded evidence for the following presumptions: Regional activities prevent corporations from suffering heavy performance decreases during their initial foreign expansion. However, pure regionalization also seems to impede corporations from fully exploiting all the benefits of non-home internationalization. Focusing on regional strategies in terms of intra-regional sales ensures a smoother performance progression and therefore less risk, but impedes stronger profit potential after a certain level of internationalization – here identified at TNI and a ratio of foreign sales in the home-region of 40%. The results indicate that the locus of destination plays a key role in resolving the controversy regarding the different shapes of the internationalization and performance relationship. Further research in this direction seems very promising. Especially industry-specific studies might be able to further gain deeper insights into the question of whether and at what point in international development a regional or non-regional strategy should be applied.

REFERENCES Al-Obaidan, A. M., & Scully, G. W. (1995). The theory and measurement of the net benefits of multinationality: The case of the international petroleum industry. Applied Economics, 27(2), 231–238. Annavarjula, M., & Beldona, S. (2000). Multinationality–performance relationship: A review and reconceptualization. The International Journal of Organizational Analysis, 8(1), 48–67. Backhaus, K., Bu¨schken, J., & Voeth, M. (2003). Internationales marketing (5th ed.). Stuttgart: Scha¨ffer-Peoschel. Bartlett, C. A., & Ghoshal, S. (2000). Transnational management. Boston: Irwin/McGraw-Hill. Be´langer, J., Giles, A., & Grenier, J.-N. (2003). Patterns of corporate influence in the host country: A study of ABB in Canada. International Journal of Human Resource Management, 14(3), 469–485. Benvignati, A. M. (1987). Domestic profit advantages of multinational firms. Journal of Business, 60(3), 449–461.

378

NICOLE RICHTER

Boddewyn, J. J. (1988). Political aspects of MNE theory. Journal of International Business Studies, 19(3), 341–363. Buckley, P. J., & Casson, M. (1976). The future of the multinational enterprise. London: Macmillan Press. Buckley, P. J., Dunning, J. H., & Pearce, R. D. (1978). The influence of firm size, industry, nationality and degree of multinationality on the growth and profitability of the world’s largest firms, 1962–1972. Weltwirtschaftliches Archiv, 114(2), 243–257. Bu¨hner, R. (1987). Assessing international diversification of West German corporations. Strategic Management Journal, 8(1), 25–37. Caves, R. E. (1971). Industrial corporations: The industrial economics of foreign investment. Economica, New series, 38(149), 1–27. Chang, Y., & Thomas, H. (1989). The impact of diversification strategy on risk-return performance. Strategic Management Journal, 10(3), 271–284. Collins, J. M. (1990). A market performance comparison of U.S. firms active in domestic, developed and developing countries. Journal of International Business Studies, 21(2), 271–287. Contractor, F. J., Kundu, S. K., & Hsu, C.-C. (2003). A three-stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34(1), 5–18. Daniels, J. D., & Bracker, J. (1989). Profit performance: Do foreign operations make a difference? Management International Review, 29(1), 46–56. Erikkson, K., Johanson, J., Majkgard, A., & Sharma, D. D. (2000). Effect of variation on knowledge accumulation in the internationalization process. International Studies of Management and Organization, 30(1), 26–44. Errunza, V. R., & Senbet, L. W. (1984). International corporate diversification, market valuation, and size-adjusted evidence. Journal of Finance, 39(3), 727–743. Fayerweather, J. (1978). International business strategy and administration. Cambridge, MA: Ballinger. Fisch, J. H., & Oesterle, M.-J. (2003). Exploring the globalisation of German MNCs with the complex spread and diversity measure. Schmalenbach Business Review, 55(1), 2–21. Geringer, M. J., Beamish, P. W., & Da Costa, R. C. (1989). Diversification strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10(2), 109–119. Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between multinationality and performance. Journal of International Business Studies, 30(1), 173–187. Graham, J. L. (2001). Culture and human resources management. In: A. M. Rugman & T. L. Brewer (Eds), The Oxford handbook of international business (pp. 503–536). Oxford: Oxford University Press. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies. Journal of International Business Studies, 18(3), 79–89. Grant, R. M., Jammine, A. P., & Thomas, H. (1988). Diversity, diversification, and profitability among British manufacturing companies. The Academy of Management Journal, 31(4), 771–801. Haar, J. (1989). A comparative analysis of the profitability performance of the largest U.S., European and Japanese multinational enterprises. Management International Review, 29(3), 5–18.


Hassel, A., Ho¨pner, M., Kurdelbusch, A., Rehder, B., & Zugeho¨r, R. (2003). Two dimensions of the internationalization of firms. Journal of Management Studies, 40(3), 701–719. Hitt, M. A., Hoskisson, R. E., & Ireland, R. D. (1994). A mid-range theory of the interactive effects of international and product diversification on innovation and performance. Journal of Management, 20(2), 297–326. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance in product-diversified firms. Academy of Management Journal, 40(4), 767–798. Hofstede, G. (1980). Culture’s consequences: International differences in work-related values. Newbury Park, CA: Sage. Holden, C. (2005). The internationalization of corporate healthcare: Extent and emerging trends. Competition and Change, 9(2), 201–219. Hsu, C.-C., & Boggs, D. J. (2003). Internationalization and performance: Traditional measures and their decomposition. The Multinational Business Review, 11(3), 23–49. Johanson, J., & Vahlne, J.-E. (1990). The mechanism of internationalism. International Marketing Review, 7(4), 11–24. Kim, W. C., Hwang, P., & Burgers, W. P. (1989). Global diversification strategy and corporate profit performance. Strategic Management Journal, 10(1), 45–57. Kim, W. S., & Lyn, E. O. (1987). Foreign direct investment theories, entry barriers, and reverse investments in U.S. manufacturing industries. Journal of International Business Studies, 18(2), 53–66. Kogut, B. (1989). Research notes and communications: A note on global strategies. Strategic Management Journal, 10(4), 383–389. Kumar, M. S. (1984). Comparative analysis of UK domestic and international firms. Journal of Economic Studies, 11(3), 26–42. Lu, J. W., & Beamish, P. W. (2001). The internationalization and performance of SMEs. Strategic Management Journal, 22(6/7), 565–586. Lu, J. W., & Beamish, P. W. (2004). International diversification and firm performance: The S-curve hypothesis. Academy of Management Journal, 47(4), 598–609. Meffert, H., & Bolz, J. (1998). Internationales Marketing–Management (3rd ed.). Stuttgart: Kohlhammer. Michel, A., & Shaked, I. (1986). Multinational corporations vs. domestic corporations: Financial performance and characteristics. Journal of International Business Studies, 17(3), 89–100. Morck, R., & Yeung, B. (1991). Why investors value multinationality. Journal of Business, 64(2), 165–187. Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill. Papadopoulos, N., & Denis, J.-E. (1988). Inventory, taxonomy and assessment of methods for international market selection. International Marketing Review, 5(3), 38–51. Pheng, L. S., Hongbin, J., & Leong, C. H. Y. (2004). A comparative study of top British and Chinese international contractors in the global market. Construction Management and Economics, 22(7), 717–731. Ramaswamy, K. (1995). Multinationality, configuration and performance: A study of MNEs in the US drug and pharmaceutical industry. Journal of International Management, 1(2), 231–253. Reeb, D. M., Kwok, C. C. Y., & Beak, H. Y. (1998). Systematic risk of the multinational corporation. Journal of International Business Studies, 29(2), 263–279.

380

NICOLE RICHTER

Riahi-Belkaoui, A. (1998). The effects of the degree of internationalization on firm performance. International Business Review, 7(3), 315–321. Rugman, A. M. (2002). Internalization as a general theory of foreign direct investment: A re-appraisal of the literature. In: A. M. Rugman (Ed.), International business: Critical perspectives on business and management (Vol. I, pp. 73–87). London: Routledge. Rugman, A. M. (2005). The regional multinationals: MNEs and ‘‘global’’ strategic management. Cambridge, UK: Cambridge University Press. Rugman, A. M., & Verbeke, A. (2004a). A perspective on regional and global strategies of multinational enterprises. Journal of International Business Studies, 35(1), 3–18. Rugman, A. M., & Verbeke, A. (2004b). Regional transnational and triad strategy. Transnational Corporations, 13(3), 1–20. Rugman, A. M., & Verbeke, A. (2007). Liabilities of regional foreignness and the use of firmlevel versus country-level data: A response to Dunning et al. (2007). Journal of International Business Studies, 38(1), 200–205. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43(1), 63–83. Ruigrok, W., & Wagner, H. (2006). Internationale Standortverlagerung und Mitarbeiterproduktivita¨t: Eine empirische Untersuchung deutscher Industrieunternehmen. Zeitschrift fu¨r Betriebswirtschaftliche Forschung, 57(4), 310–324. Santos, J. R. A. (1999). Cronbach’s alpha: A tool for assessing the reliability of scales. Journal of Extension, 37(2). Retrieved on February 20, 2007, from http://www.joe.org/joe/ 1999april/tt3.html Siddharthan, N. S., & Lall, S. (1982). The recent growth of the largest US multinationals. Oxford Bulletin of Economics and Statistics, 44(1), 1–13. Streiner, D. L. (2003). Starting at the beginning: An introduction to coefficient alpha and internal consistency. Journal of Personality Assessment, 80(1), 99–103. Sullivan, D. (1994a). Measuring the degree of internationalization of a firm. Journal of International Business Studies, 25(2), 325–342. Sullivan, D. (1994b). The ‘‘threshold of internationalization’’: Replication, extension, and reinterpretation. Management International Review, 34(2), 165–186. Tallman, S., & Li, J. (1996). Effects of international diversity and product diversity on the performance of multinational firms. Academy of Management Journal, 39(1), 179–196. Teece, D. J. (2002). Transactions cost economics and the multinational enterprise: An assessment. In: A. M. Rugman (Ed.), International business: Critical perspectives on business and management (Vol. I, pp. 151–176). London: Routledge. Thomas, D. E., & Eden, L. (2004). What is the shape of the multinationality–performance relationship? The Multinational Business Review, 12(1), 89–110. UNCTAD. (1995). World investment report, 1995. Transnational corporations and competitiveness. New York/Geneva: United Nations. UNCTAD. (2003). World investment report, 2003. FDI policies for development: National and international perspectives. New York/Geneva: United Nations. UNCTAD. (2004). World investment report, 2004. The shift towards services. New York/ Geneva: United Nations. UNCTAD. (2005). World investment report 2005. Transnational corporations and the internationalization of R&D. New York/Geneva: United Nations.


Urban, D., & Mayerl, J. (2006). Regressionsanalyse: Theorie, Technik und Anwendung (2nd ed.). Wiesbaden, Germany: VS Verlag fu¨r Sozialwissenschaften. Vernon, R. (1971). Sovereignty at bay. New York: Basic Books. Wan, W. P., & Hoskisson, R. E. (2003). Home country environments, corporate diversification strategies, and firm performance. Academy of Management Journal, 46(1), 27–45. Wernerfelt, B. (1984). A resource-based view of the firm. Strategic Management Journal, 5(2), 171–180.

AUTHOR INDEX Aaker, D. 309 Agarwal, S. 160 Aggarwal, R. 18 Agmon, T. 147, 160 Aharoni, Y. 69, 72, 297 Aiken, L.S. 241 Alcacer, J. 225–226, 229 Al-Obaidan, A.M. 341, 360, 362 Amman, W. 24 Amsden, A. 343 Anand, J. 76 Anderson, J.E. 101 Annavarjula, M. 366 Ansoff, H.J. 151 Arregle, J.L. 208 Asiedu, B.J. 160, 163 Asmussen, C.G. 47, 65 Audretsch, D.B. 228, 249 Aulakh, P.S. 23, 265, 267, 269 Aw, B.Y. 238

Baysinger, B.D. 275 Beak, H.Y. 361 Beamish, P.W. 13, 17–19, 20, 24, 27, 66–67, 69, 73, 116–118, 120, 133, 140, 144, 149–150, 152, 161, 203, 208, 264–265, 267–269, 273, 299, 292, 309, 319–323, 339–341, 345, 347, 352, 362–365 Beck, N. 348 Be´langer, J. 366 Beldona, S. 366 Benvignati, A.M. 360–361 Bergstrand, J.H. 101 Berry, H. 21, 273 Berthon, P. 232 Bhaumik, S.K. 149, 161 Bierman, L. 203 Birkinshaw, J. 231, 264–266, 271, 288 Black, J.S. 17, 148 Blau, P.M. 236 Blonigen, B.A. 86, 234 Boddewyn, J.J. 148, 160, 341–342, 361 Boesecke, K. 143 Boggs, D.J. 366 Bolz, J. 364 Booth, L.D. 323 Borza, A. 208 Bowen, H.P. 113, 128 Bracker, J. 322–323, 341, 345, 362–363, 365, 373, 377 Brain, C. 66, 68, 70, 86 Brainard, S.L. 101 Braunerhjelm, P. 233 Brewer, H.L. 148, 160 Bruton, G.D. 210

Backhaus, K. 364 Baden-Fuller, C.W. 345 Baek, Y.H. 148, 160 Bain, J.S. 152 Baltagi, B. 211 Barbosa, N. 234 Barkema, H.G. 73, 147, 247, 268, 271, 290, 344 Barney, J. 224, 320, 326 Bartlett, C.A. 17, 264, 367 Batra, G. 238 Baum, J.A.C. 231–232, 234, 236, 239 Bausch, A. 143, 160 Bayek, Y. 299 383

384 Buckley, P.J. 18, 70, 73, 160, 202, 268, 308, 361, 366 Bu¨hner, R. 318–319, 322–323, 326, 339, 360–361, 365 Bu¨schken, J. 364 Burgers, W.P. 33, 266, 340, 361 Calvet, A.L. 148 Campbell, A.J. 342 Capar, N. 338–339, 341 Capron, L. 150 Cassiman, B. 230 Casson, M.C. 18, 70, 73, 160, 268, 308, 361 Casti, J.L. 146 Caves, R.E. 18, 73, 115, 148, 160, 268, 274, 339, 360 Cavusgil, T.S. 265 Certo, S.T. 155 Chang, S.J. 73, 146, 210 Chang, Y. 361 Chen, H. 227 Chen, S. 337 Chen, T.J. 227 Child, J. 266 Choi, C.J. 204 Choi, U. 210 Christophe, S.E. 33–34 Chung, K.H. 273 Chung, L. 266 Chung, W. 225–226, 229–230, 232, 236–237 Coase, R.H. 146, 160 Cohen, W.M. 224 Collins, J.M. 322–323, 339, 341, 361 Collinson, S.C. 45, 162–163 Combs, J.G. 155, 164 Connelly, C. 115, 117, 128–129, 133–134 Contractor, F.J. 11, 13, 15, 20, 22, 24, 26–27, 34, 117–118, 120, 130, 132, 203, 267, 298–299, 308–310, 312, 319,

AUTHOR INDEX 322–323, 338, 340–343, 347, 352, 360, 362–363 Cool, K. 271 Covin, J.G. 162 Cowley, P.R. 309 Crook, T.R. 155, 164 Csiszar, E.N. 149 Da Costa, R.C. 19, 152, 299, 309, 320–321, 323, 339, 341, 345, 362, 364–365 Dacin, M.T. 208 Daily, C.M. 156, 162 Dalton, D.R. 155–156, 162 Daniel, F. 156 Daniels, J.D. 322–323, 341, 345, 362–363, 365, 373, 377 Das, T.K. 149 Davidson, W.H. 321 Davies, H. 266 D’Cruz, J. 304 De Bondt, R. 230 Deardorff, A.V. 101 Delios, A. 17, 66, 69, 73, 76, 152, 208, 264, 292 Denis, D.J. 33–34 Denis, D.K. 33–34 Denis, J.-E. 364 Dierickx, I. 271 Djellal, F. 238–239 Douglas, S.P. 288 Doukas, J. 147, 151, 160 Douma, S. 73 Doz, Y.L. 14, 264, 344 Duade, C. 101 Dunning, J.H. 114, 147, 151, 160, 222, 225, 238, 240, 253–254, 266, 274, 297–298, 320, 342–343, 366 Eden, L. 13, 20, 25, 205, 372 Ellstrand, A.E. 156 Ennew, C. 149, 161 Enright, M.J. 292

Author Index Erikkson, K. 364 Errunza, V.R. 33–34, 147, 339, 341, 360–361 Esfahani, H.S. 160, 163 Ethier, W. 101 Farjoun, M. 72 Fayerweather, J. 361 Feenstra, R.C. 101 Feketekuty, G. 342 Feldman, P.M. 228, 249 Feldstein, M. 86 Finkelstein, S. 239–240 Fisch, J.H. 366 Fischer, J.H. 223, 231 Florida, R. 241, 255 Floyd, D. 209 Flyer, F. 223, 225–229, 232–234, 237–239, 241 Fornaciari, C.J. 156 Frankel, J.A. 101 Freese, J. 334 French, K.R. 86 Friedman, J. 230 Friedrich, R.J. 237 Fritz, T. 143 Galbraith, J.R. 340 Gallouj, F. 238–239 Gaur, A.S. 201 Gelb, S. 149, 161 Geringer, J.M. 19, 118, 144, 152, 299, 309, 320–321, 323, 339, 341, 345 Geringer, M.J. 362, 364–365 Gerlowski, D.A. 230 Gersbach, H. 229 Ghauri, P. 202 Ghemawat, P. 60, 68, 222, 267, 269, 345 Ghoshal, S. 17, 264, 266, 298, 345, 367 Giles, A. 366 Girod, S. 68, 86, 298, 305 Glaeser, E.I. 222

385 Glass, G.V. 155, 157 Gobeli, D.H. 323 Goerzen, A. 47, 65, 67, 74, 268–269, 299, 309 Gomes, L.K. 19, 33, 117–119, 299, 308, 310, 318–319, 321, 323, 339–341, 347, 360, 362, 364–365, 373, 377 Gomez-Mejia, L.R. 152, 163 Gongming, Q. 18, 340 Gooding, R.Z. 156 Gordon, I. 222 Govindarajan, V. 66, 300 Graham, J.L. 361 Grant, R.M. 18, 33, 147, 160, 273, 299, 308–309, 319–320, 322–323, 326, 339–341, 346, 361–362, 372 Greene, W.H. 129–131, 136–137, 140 Grein, A. 247 Grenier, J.-N. 366 Griffith, D. 113, 136 Griffiths, A. 344 Grosse, R. 101, 264 Grubel, H.G. 147 Grubert, H. 101 Guillen, M.F. 343 Guisinger, S. 292, 297 Gupta, A. 66, 300 Haar, J. 361, 365 Habib, M.M. 338 Halbrich, M.B. 341–342 Haleblian, J. 239–240 Hamel, G. 147, 320 Hamilton, B. 140 Harrington, J.E. 223, 231 Haskel, J. 309 Haspeslagh, P.C. 146 Hassel, A. 366 Haveman, H.A. 231–232 Hay, D.A. 246 Head, K., 101, 232–233, 238, 241 Hejazi, W. 85–87, 89–90, 98, 100–101, 106

386 Helpman, E. 101 Henisz, W. 17, 152 Hennart, J.-F. 70, 148, 160 Higgins, R.C. 162 Hikino, T. 343 Hill, C.W.L. 147, 160 Hitt, M.A. 17–19, 67, 115, 117, 128–129, 133–134, 203, 208, 268–269, 275, 298–299, 308, 326, 339–341, 360, 362, 365 Hodgetts, R. 66, 264 Hofstede, G. 148, 160, 239, 275, 361 Holbrook, D. 224 Holden, C. 366 Hongbin, J. 366 Hood, N. 147, 160, 231, 271 Ho¨pner, M. 366 Horioka, C. 86 Hoskisson, R.E. 17–19, 67, 204–205, 268–269, 275, 298–299, 308, 339–341, 360, 362, 365–366 Hounshell, D.A. 224 Hout, T. 266 Hsu, C.-C. 13, 20, 22, 24, 26–27, 34, 117–118, 120, 130, 132, 203, 267, 298–299, 308–310, 312, 319, 322–323, 338, 340–343, 347, 352, 360, 362–363, 366 Hulland, J. 150, 264–266, 288 Hunter, J.E. 155, 157–158 Hwang, P. 33, 266, 340, 361 Hymer, S. H. 12, 18, 70, 79, 115, 147, 150, 160, 319–320, 339 Ingram, P. 239 Ireland, R.D. 362 Jackson, G.B. 155, 157 Jammine, A.P. 308, 361 Jeannet, J. 66 Jemison, D.B. 146

AUTHOR INDEX Jensen, M.C. 73, 146, 148–149, 159, 161 Johanson, J. 14, 22, 71, 80, 146–147, 152, 160, 163, 321, 340, 344, 364 Johnson, J.L. 156 Johnson, R.A. 275 Jones, G.R. 147, 160 Jung, Y. 309 Kachra, A. 144 Kalnins, A. 232, 237 Kang, N.-H., 144 Kapur, D. 209 Katrishen, F.A. 342 Katz, J.N. 348 Kazanjian, R. 340 Keats, B.W. 164 Keeble, D. 229, 231–232, 234, 240 Kenney, M. 241 Khanna, T. 23, 204–205, 210, 338, 343 Kihlstrom, R. 77 Kim, H. 17–19, 67, 268–269, 275, 298–299, 308, 339–341, 360, 362, 365 Kim, J.B. 204 Kim, K. 264–266, 268, 271–272, 292 Kim, W.C. 33, 266, 340, 361 Kim, W.S. 33–34, 360–361 Kindleberger, C. 150, 160 King, D.R. 162 Knight, G. 342 Kobrin, S. 264–265, 267, 270–272, 288 Kogut, B. 15, 73, 114, 146, 152, 239, 268–269, 275, 321–322, 340, 361 Kohlhas, J. 222 Koslowky, M. 157 Kostova, T. 14, 24, 152, 208 Kotabe, M. 23, 265, 267, 269, 338–339, 341 Koza, M.P. 150 Kravis, I.B. 225–226, 232 Krist, M. 160 Krugman, P. 255

Author Index Kudina, A. 297, 307 Kumar, M.S. 339, 341, 366 Kumar, V. 201 Kundu, S.K. 13, 20, 22, 24, 26–27, 34, 117–118, 120, 130, 132, 203, 267, 298–299, 308–310, 312, 319, 322–323, 338, 340–343, 347, 352, 360, 362–363 Kurdelbusch, A. 366 Kwok, C.C.Y. 148, 160, 299, 361 Lall, S. 18, 207, 340–341, 361 Lau, C.M. 205 Laurence, M.M. 18, 267 Lawrence, P.R. 266 Leamer, E.E. 101 Lecraw, D.J. 323 Lee, S.H. 33, 204 Leff, H.N. 210, 343 Leong, C.H.Y. 366 Lessard, D. 147, 160 Levitas, E. 208 Levitt, T. 264, 266, 298 Li, D. 263 Li, J.T. 33, 151, 268, 298–299, 308, 323, 346, 361 Li, L. 209–210, 263–265, 267, 309–310 Licht, G. 238–239 Lipsey, R.E. 101, 225–226, 232 Lohrke, F.T. 156 Lomi, A. 222, 231, 233, 241 Long, J.S. 334 Loree, D. 297 Lorsch, J.W. 266 Louis, H. 144 Louri, H. 234 Lovelock, C.H. 342 Lu, H. 116–118, 120, 133, 140 Lu, J.W. 13, 17–18, 20, 24, 27, 203, 265, 267–269, 273, 309, 319, 321–323, 340, 347, 352, 362–363 Lu, Y. 201, 210 Lubatkin, M. 151

387 Luo, Y. 217, 323 Lyn, E.O. 33–34, 360–361 Mahoney, J.T. 63 Majkgard, A. 364 Makhija, M.V. 264–266, 268, 271–272 Makino, S. 149, 161 Mariotti, S. 229 Markowitz, H.M. 146 Markusen, J.R. 101–102 Martin, C. 309 Martin, R. 236 Mascarenhas, B. 309 Mata, J. 234 Mathews, J. A., 202, 208 Mauri, A. 339 Mayerl, J. 371 McCallum, J. 103 McDougall, P. 15 McGraw, B. 157 McManus, J. 209 McNamara, G. 205 McWilliams, A. 32 Meffert, H. 364 Meyer, M. 72 Mezias, S.J. 231, 236 Michael, A. 319, 322–323 Michel, A. 339, 341, 361 Miller, S. 25 Miller, T. 115, 117, 128–129, 133–134 Mishra, C. 323 Mitchell, W. 146 Moch, D. 238–239 Mody, A. 233 Moesel, D.D. 275 Montgomery, C.A. 308 Moomaw, R.L. 232 Moore, K. 86 Morck, R. 18, 33, 70, 73, 265, 268–269, 273, 323, 361 Morrison, A.J. 264–266, 272, 288 Mosakowski, E. 11–12, 14, 24, 312, 317, 319–320, 322, 337, 340

388 Murray, J.Y. 338 Mutti, J. 101 Nachum, L. 203, 207, 221, 229, 231–234, 237, 240–241, 338, 343, 352 Nahm, K.B. 222 Nayyar, P. 73, 326 Nickerson, J. 140 Nigh, D. 297 Nohria, N. 266, 298 Nunnally, J.C. 366 O’Donnell, S. 147, 160 Oesterle, M.-J. 366 Oh, C.H. 31, 48–49, 59–61 Ohmae, K. 222 Olsen, M. 118 Osegowitsch, T., 12, 45–46, 48, 59, 61–63 Oviatt, B. 15 Owen, R.F. 240 Palepu, K. 23, 205, 210, 275, 343 Palich, L.E. 152, 163 Pantzalis, C. 136, 269 Papadopoulos, N. 364 Park, J.H. 264–265, 272, 292 Patterson, P. 232 Pearce, R.D. 366 Peng, M.W. 17, 114, 117, 202, 217, 323 Pennings, J. 73 Penrose, E. 63 Perry, A.C. 341–342 Pheng, L.S. 366 Phipps, A.G. 254 Pisano, G. 149 Piscitello, L. 229 Pitt, L. 232 Porter, M.E. 135, 224, 265–266, 270, 289, 298, 308, 340 Porterba, J.M. 86 Portugal, P. 234

AUTHOR INDEX Prahalad, C.K. 147, 264, 320, 344 Prescott, J.E. 264–265, 272, 292 Pruitt, S.W. 273 Ramamurti, R. 205, 209 Ramanujam, V. 164 Ramaswami, S.N. 160 Ramaswamy, K. 19, 33, 117–119, 299, 308–310, 318–319, 321, 323, 339–341, 347, 360, 362, 364–365, 373, 377 Rauch, J.E. 227–229 Ray, E.J. 234, 240 Ray, S. 207 Reeb, D.M. 148, 160, 299, 361 Rehder, B. 366 Reuer, J.J. 150 Riahi-Belkaoui, A. 18, 20, 27, 322–323, 362, 372 Richter, N. 359 Ricks, D.A. 264, 266 Ries, J. 101, 232–233, 238, 241 Riordan, M. 77 Rivkin, J.W. 204, 338, 343 Roengpitya, R. 155 Rolfe, R.J. 152 Roll, R. 150 Rose, A.K. 101 Rosenthal, R. 158 Ross, D. 146 Ross, S.A. 146, 159 Roth, K. 147, 160, 264, 266, 272 Ruback, R.S. 148 Rudden, E. 266 Rugman, A.M. 12, 14–16, 31, 34–36, 45–49, 53, 59–63, 66–68, 70–73, 76–77, 79, 86, 114–115, 125, 147, 162–163, 202–204, 208, 222–223, 253, 264, 268–269, 273, 287–290, 292–293, 297–300, 304–308, 317, 319–320, 322–324, 326–327, 332, 339, 344, 360–361, 363–364, 366–368, 373, 376 Ruigrok, W. 18–19, 23–24, 135, 160, 203, 299, 318–323, 339, 360, 362, 372

Author Index Rumelt, R.P. 38, 146, 150 Russell, C. 113, 136 Safarian, A.E. 101 Sagie, A. 157 Sakai, K. 144 Salter, M.S. 151 Sambharya, R. 339 Sammartino, A. 12, 45–46, 48, 59, 61–63 Santos, J. R. A. 14, 366 Sarkar, M.B. 265 Sassen, S. 222, 233, 253 Saxenian, A. 228 Schall, L.D. 162 Scherer, F.M. 146 Schlie, E. 305, 345 Schmalensee, T. 308, 326 Schmidt, F.L. 155, 157–158 Schmutzler, A. 229 Scholes, M.S. 147, 160 Schweiger, D.M. 149 Scordis, N.A. 342 Scott, A.J. 226–228 Scully, G.W. 341, 360, 362 Semple, R.K. 222, 254 Senbet, L.W. 33–34, 339, 341, 360–361 Senebet, L.T. 147 Severn, A.K. 18, 267 Shaked, I. 319, 322–323, 339, 341, 361 Shan, W. 265 Sharma, D.D. 364 Sharpe, W. 73 Shaver, J.M. 223, 225–229, 232–234, 237–241 Shenkar, O. 17, 229 Shimizu, K. 203 Shook, C.L. 155, 164 Siddharthan, N.S. 18, 340–341, 361 Siegel, D. 32 Silberman, J. 230 Singh, A. 343 Singh, H. 239, 275

389 Sinha, J. 205, 208 Slaets, P. 230 Smart, D. 326 Smith, M.L. 157 Song, J. 225, 265 Srinivasan, S.S. 23, 265, 267, 269 Stein, E. 101 Steiner, P.O. 162 Stopford, J. 71, 345 Streiner, D.L. 366 Studenmund, A.H. 245 Suarez-Villa, L. 229 Sukpanich, N. 66, 70–72, 79, 317 Sullivan, D. 22, 116, 299, 309, 318–319, 323, 341, 362, 366 Sundaram, A.K. 148 Sunderam, A. 17 Sunley, P. 236 Svensson, R. 233 Swaminathan, A. 231 Swenson, D. 232–233, 238, 241 Szulanski, G. 268 Tallman, S.B. 33, 118, 148, 151, 268, 298–299, 308, 323, 346, 361 Tan, D. 63 Teece, D.J. 149, 160, 163, 268, 361 Teng, B.-S. 149 Thomas, D.E. 13, 20, 203, 372 Thomas, H. 308, 361 Thompson, J.D. 266 Tihanyi, L. 115, 117, 128–129, 133–134, 136 Tolentino, P.E.E. 338 Travlos, N. 147, 151, 160 Trefler, D. 101 Trevino, L. 101 Tsai, W. 271 Turner, R.A. 156 Uhlenbruck, K. 203 Urban, D. 371 Usher, J.M. 234

390 Vaaler, P.M. 205 Vahlne, J.-E. 14, 22, 71, 80, 146–147, 152, 160, 163, 321, 340, 344, 364 Varadarajan, N. 164 Venkatraman, N. 164 Verbeke, A. 15, 31, 34–36, 45–47, 59, 62–63, 66–68, 70–72, 76–77, 85–86, 114, 125, 202–204, 222–223, 253, 264, 268–269, 273, 287, 289–290, 292–293, 298, 300, 305, 317, 320, 326, 332, 342, 360, 364, 368, 373, 376 Vermeulen, F. 247, 268, 271, 290, 344 Vermeulen, W.K. 147 Vernon, R. 268, 339–341, 360–361 Victor, B. 338 Voeth, M. 364 Wagner, H. 18–19, 23–24, 135, 160, 203, 299, 318–323, 339, 360, 362, 372 Wagner J.A., III. 156 Walrod, W. 229 Wan, W.P. 204, 365–366 Ward, S. 254 Watson, R. 232 Wei, S.J. 101 Weinhold, W.A. 151 Weisberg, S. 237 Weiss, M.Y. 101 Wells, L.T. 71, 207, 217 Wernerfelt, B. 146, 361 West, S.G. 241 Westney, D.E. 288, 290 Wheeler, D. 233 Whitley, R. 344 Wiedersheim-Paul, F. 71, 80, 152, 163

AUTHOR INDEX Wiersema, M.F. 128 Williamson, O.E. 146–147, 160 Williamson, P. 14 Williamson, S.D. 264–266, 268, 271–272 Wilson, W.W. 86 Wind, Y. 288 Winsted, K. 232 Wolfson, M.A. 147, 160 Wong, P. 149, 161 Woodcock, C.P. 149, 161 Woodward, D.P. 152 Wright, M. 149, 161, 205 Wymbs, C. 221, 230, 234 Yamori, N. 225 Yeung, B. 18, 33, 70, 73, 265, 268–269, 273, 323, 361 Yeung, H. 23 Yin, E. 204 Yip, G.S. 66, 148, 264, 266, 297–298, 300, 304–305, 307, 342, 345 Yiu, D. 210 Yost, K. 33–34 Young, S. 147, 160 Zacharakis, A.L. 150 Zaheer, S. 12, 14, 24, 70, 150, 152, 160, 208, 312, 319–320, 322, 340 Zammuto, R.F. 344 Zander, I. 202 Zander, U.B. 114, 146, 268 Zugeho¨r, R. 366

SUBJECT INDEX abnormal return (AR), 32 accumulated knowledge, 138, 319, 322 advertising intensity, 23, 73–77, 179, 185–186 alliances, 4, 144–146, 148, 150–151, 154–155, 159, 161, 165, 272, 275 antecedents, 129, 133 Armington bias, 86 Asia, 34, 38, 48–49, 51, 53–55, 57–59, 62–63, 68–69, 74, 81, 86, 88–90, 92, 94–99, 145, 152, 159, 162–163, 190, 193, 204, 206–207, 217, 264, 274, 289, 291, 293, 300–302, 305, 317, 325–327, 332, 334, 337–338, 344–346, 352, 354–355, 359–360, 364 Asia-Pacific, 48–49, 51, 53–55, 57–59, 62–63, 69, 74, 81, 274, 289, 293, 337–338, 359–360 average cost, 121

competitive advantages, 114, 143, 146–147, 152, 160, 225, 266, 270, 289, 320, 326 control variable, 5–6, 125, 129, 132, 134, 211, 237, 240, 243, 245, 251–252, 275, 308–309, 312, 326, 347, 366, 369–372, 374 cost, 12, 15, 20, 22, 47, 51, 77, 85–91, 100, 105, 107, 115, 118–121, 125–126, 128, 131, 137, 139–140, 146–147, 152, 178, 181, 184, 195, 204, 209, 214, 231, 267, 304, 308, 322, 332, 342–343, 362–364, 372, 375 costs of internationalization, 25, 180, 183, 192, 320, 360–362 cross-sectional time series, 347 cubic model, 122, 127, 328–331, 369, 371–375 cubic relationship, 120, 122, 126–127, 369–371, 375

benefits of internationalization, 67, 178–180, 184, 191, 195, 321, 342–343, 360, 362 bi-regional, 2, 45–46, 48, 61–62, 71–72, 204, 264, 273–274, 276–280, 282–284, 286–292, 300, 302, 305, 366, 368 business combinations, 4, 144–148, 150–154, 158–165 business groups, 210

decentralization, 183, 188–190, 197 degree of regionalization, 207, 306 degree of internationalization (DOI), 15, 17–19, 26, 67, 69, 71, 78, 80–81, 152, 160, 163, 181, 184, 190, 193–194, 1201, 203, 209, 211–216, 339, 341, 344, 346, 365 degree of multinationality, 3, 32, 70, 72, 113, 115–116, 118–121, 126, 132, 134, 137–140, 307, 309, 318–319, 321–323 degrees of freedom, 130, 252 developed country, 136, 338 developing country, 136, 207, 343–344

capital invested, 67, 71–72, 74, 78–81 centralization, 182, 187, 189, 292 country environments, 67 391

392 downstream activity/production, 77, 269, 270, 289 dummy variables, 101–103, 130, 140, 190, 192, 241, 243, 252, 275, 347, 366, 369 economies of scale, 18, 56, 121, 146, 151, 180–182, 265–267, 288, 319, 326, 334, 340, 342–343, 360, 363 economies of scope, 140, 268, 289, 319 emerging markets, 25, 202–203, 210, 217 endogenous, 132–134, 138–140 entropy, 33, 80, 116, 139, 275 European Union (EU), 3, 6, 15, 17, 37–38, 55–56, 86–88, 91–100, 102, 104–107, 300, 305, 317, 369 external growth strategies, 143, 145, 165 extraregional sales, 353 factor analysis/input, 73, 115, 117, 189, 366 Feldstein-Horioka puzzle, 86 financial and professional service, 5, 221–223, 232–233, 235, 247–248, 252, 255 firm performance, 3–4, 6, 18, 23, 66, 113, 115, 125, 127–128, 132–133, 135–138, 141, 143–145, 148, 150–151, 153, 161, 164, 178, 182, 195, 201, 203, 210, 212–214, 216–217, 263, 266, 271, 273, 276, 282, 288, 291–292, 297, 308, 312, 339, 352, 359, 366 firm specific advantage (FSA), 2, 32, 46, 68, 115, 160, 163, 180, 183, 223, 270, 289 FIT, 190–192 fixed effects, 129, 211, 214 foreign assets, 5–7, 16–17, 21, 32–33, 37, 116–117, 125, 137, 297–298, 307–309, 311–313, 323, 333, 345, 367 foreign direct investment (FDI), 3–5, 22, 26, 66–67, 69–70, 72, 74, 79–81, 85, 87–88, 98, 100–102,

SUBJECT INDEX 106–107, 135, 143–144, 147, 204–207, 209, 214–215, 222, 233–234, 239–240, 249, 253–254, 264, 268 foreign intra-regional sales and performance, 373 foreign sales ratio, 53, 115–116, 121–125, 137–138 foreign subsidiaries, 7, 22, 69, 71–72, 74, 77, 182, 187, 189–190, 265, 271–272, 274–275, 306, 313, 339 fully-pooled model, 130 geographic orientation, 65, 67–79 global cities, 222, 254 global integration, 178, 182–183, 185, 264, 266, 269, 272 global orientation, 65–66, 75, 77, 264 global strategy, 40, 65–66, 69–70, 78–80, 204, 210, 215, 263–271, 273, 276, 280–282, 285–292, 298, 300, 304–305 global supply chain, 34 globalization, 46 gravity model, 3, 85–91, 93, 95, 97, 99–106 Heckman, 136 Herfindahl index, 22, 116 Herfindahl, 22, 116, 139 heterogeneity, 8, 23, 69, 80, 128–132, 135–138, 140–141, 147, 152, 154–155, 157, 182, 211, 224–225, 231, 253, 298, 352 hierarchy, 188 holding of domestic equities, 86 home-oriented corporation, 364, 366 368, 369, 370, 372, 373, 375 home region, 2, 6–7, 12, 34, 36–37, 39–40, 45–48, 51–55, 57, 59–63, 66–70, 72, 74, 77–81, 87–88, 90, 93, 101, 105–106, 162–163, 204, 207–208,

Subject Index 274, 277–279, 288, 290, 298, 300, 305–308, 312, 317–318, 321, 324, 332, 334, 344 homogeneous, 130 host-region oriented, 264, 274, 300, 302, 366 index, entropy, 275 index, transnationality, 7, 16–17, 22, 117, 359, 366, 369, 372 India, 5, 15, 21, 25, 202–203, 205–210, 214–216, 345–355 industry differences, 38–39 chemicals and pharmaceutical industry, 39 construction, building materials and glass industry, 40 labor intensive industries, 40 manufacturing industry, 39 merchandising firms, 40 natural resources oriented industries, 40 service industry, 39–40, 326 institutional distance, 208, 229, 230 intangible assets, 20, 67, 74, 77, 79–80, 115, 129, 178, 180–181, 188, 239, 263, 265, 267–271, 274, 276, 282, 286, 289–290, 292, 361 integration, 4, 6, 47, 55–56, 59, 91, 145, 147–149, 161, 177–179, 182–187, 192, 195–196, 264–267, 269, 272, 304–305, 344 interaction between location and firm, 221 attributes, 221–227, 235, 237, 241, 246, 248–249, 252–255, 342 internalization, 69, 72, 73, 114, 146, 147, 166, 181, 211, 254, 275, 289, 308, 320, 321, 363 international experience, 22, 71, 73–74, 77–78, 80, 323, 341 international trade, 3, 101, 264

393 internationalization, 45–46, 48–51, 53–55, 57, 61–62 internationalization pace, 263, 271, 274, 275, 277, 278, 282, 285, 287, 290, 291, 292 inter-regional, 47, 51–52, 56, 58–59, 61, 68, 324, 332–334, 376 intra-firm, 87, 92–93, 96–99, 104–105, 265 intra-regional, 35–40 assets, 37 foreign region, 40 rest of the home region, 36–37 rest of the world, 36–37 sales, 38–39 instrumental variables, 133 intangible assets, 20, 67, 74, 77, 79–80, 115, 129, 178, 180–181, 188, 239, 263, 265, 267–271, 274, 276, 282, 286, 289–290, 292, 361 inverted U-curve, 178, 180, 323–324 inverted J-curved, 178, 323, 324, 362 J-curve, 178, 323–324 Japanese, 2, 20, 34, 49, 57, 67–69, 71–73, 135, 227, 230, 264, 300, 367–368 Jensen’s Alpha, 73 knowledge, 15, 22, 40, 73, 114–115, 129, 132, 134, 138–139, 146–147, 149–151, 160–161, 163–164, 180, 188, 197, 209, 216–217, 227–232, 234, 247–248, 252, 254–255, 268, 271–272, 290, 292, 319–320, 322, 324, 326, 338, 342–343, 361, 367 less-developed countries (LDCs), 337 liability of foreignness (LOF), 2, 24, 25, 68, 70, 78, 79, 115, 119,121, 125, 135, 136, 160, 181, 183, 192, 305, 312, 319, 320, 321, 322, 324, 333, 363, 372 liability of regional foreignness, 2

394 linear, 6, 18, 20–21, 23, 32, 34, 114, 117, 122, 124, 126, 191–192, 196, 211, 214, 241, 247, 276, 297, 306, 309–312, 317, 319, 321, 323–325, 327–333, 341, 347–348, 350–351, 353, 361, 365, 369–374 linkages, 179, 188, 231–232, 234, 236, 253, 265, 268, 271 local adaptation, 269, 342 location choices, 221–232, 236–237, 241, 246–249, 252–255 locus of destination, 360, 366, 376–377 longitudinal data, 128, 139, 210 marginal cost, 22, 120–121, 321, 362 market imperfections, 115, 146, 160, 180 marketing assets, 70, 73, 76–77 mergers and acquisitions (M&A), 144 meta-analysis, 3–4, 19, 23, 143, 145, 153–155, 158–159, 162 moderator, 80, 128, 140, 160–163, 183 modes of entry, 22 modes of multinationality (MoMs), 3, 114, 125, 132, 135 multidomestic industry, 6, 263, 265–272, 282, 286–287, 292 multinational activity, 1, 327 multinational enterprises (MNEs), 2, 32, 65–66, 202, 221–222, 263, 297, 318, 339, 359–360 multinationality, 31–40 multi-stage theory of internationalization, 11 number of foreign affiliates (NOFA), 32 ratio of foreign to total assets (F/TA), 35 ratio of foreign to total sales (F/TS), 135 ratio of regional to total assets (R/TA), 35 ratio of regional to total sales (R/TS), 35

SUBJECT INDEX return on foreign assets (ROFA), 5–6, 32, 297, 307–308, 311–312, 333 return on total assets (ROTA), 6, 32, 297, 311 return on total sales (ROTS), 32 nonlinear, 6, 114, 118–119, 128, 132, 138, 141, 341, 360, 362 non-location bound firm specific advantage, 32, 208, 270 off-shore production, 125 OLS, 133, 179, 191–192, 310, 326, 328–330, 348 omitted variable, 129, 137–138 optimal level of internationalization, 371–372, 376–377 organizational structure, 5, 177–179, 181–183, 185, 187–192, 195, 197, 224, 231–232, 240, 243–245, 248–251, 293 outward FDI, 5, 98, 100, 106–107, 205 panel, 128–130, 133, 137, 308, 348 performance of MNEs, 306 price discrimination, 115 principle components, 117 product diversification, 134, 140–141, 181 profit, 115, 118–122, 125–126, 140, 162, 196, 201, 203, 340, 363–364, 372, 375–377 proprietary assets/ resources and capabilities, 68, 73–76, 79, 114, 146, 180, 204, 208, 216, 269 quadratic, 6, 18, 20–21, 23, 32, 118, 122, 124–127, 138, 309–312, 328–330, 338, 348, 350–351, 354, 365, 369–375 R&D intensity, 73–77, 131, 179, 185–186, 226, 270, 272, 274, 277–278, 280–286, 290, 341 random effects, 129–131, 138, 211

Subject Index regional clusters, 34 regional economics, 305 region of origin, 63, 143, 152–153, 162, 366, 372 regional bias in international trade, 3, 264 regional bias in MNE activity, 86 regional firms, 2, 61, 66, 69, 74, 76–81, 264, 282, 288, 291, 300 regional hypothesis, 298 regional strategy, 68–69, 204, 210, 215–216, 263–265, 273, 276, 280, 282, 286–291, 298–299, 304, 306–307, 310, 332, 344, 360, 363–364, 373, 376–377 regionalization, 45–49, 51, 53, 55, 57, 59, 61, 63 resource-based view, 146, 224, 268, 320, 325 return on assets (ROA), 121, 273, 323, 337, 365 return on foreign assets, 5–6, 32, 125, 297, 307–309, 311–312, 333 return on sales (ROS), 190–192, 194 revenue, 55, 57–58, 71, 115, 118–119, 125–126, 128, 131, 137, 139–140, 192, 375 S-curve, 18–20, 23, 27, 32, 122, 127, 178, 181, 196, 310, 312, 322–323, 360, 362, 371–372, 377 Sharpe’s Measure, 73 sample selection, 136 segmentation, 115 selection equation, 137 semi-globalization, 68, 269 sequencing, 12

395 service firms, 39, 201, 203, 206–207, 210, 215, 232, 238–239, 326, 337–338, 341–343, 345 spatial, 12, 67, 74, 114, 125, 135, 227, 233–234 specific factors, 125, 139 strategic fit, 3–4, 177, 179, 182–185, 187, 191, 194–196 structural equations, 134 technological assets, 70, 73, 75–76 three-stage theory, 23, 319, 322, 337, 340, 348 time-series, 128, 137, 276 Tobin’s q, 21, 32–34, 273, 275, 277–278, 280–286, 288–290, 292, 323, 333 triad, 5–6, 12, 15, 34, 37–39, 45, 47–49, 53, 62, 66, 68, 184, 202, 204, 210, 264, 274, 289, 292–293, 298, 300, 305, 317–318, 324–327, 332, 334, 360, 364, 366, 368–370, 373, 375, 377 turnover by country of origin, 92, 95, 99 two-stage least squares, 133 UK firms, 299 upstream activity/production, 34, 77, 269, 270, 289 U-shape, 19–21, 23, 25, 34, 118–120, 125–127, 135, 141, 192, 214, 309, 311, 339–341, 360, 362–363, 375 uniform internationalization, 31 value chain, 15, 209, 270, 289–290, 292, 306, 342 world’s 500 largest firms, 34, 300, 317

Regional Aspects of Multinationality and Performance, Volume 13 (Research in Global Strategic Management)

Regional Economic Integration, Volume 12 (Research in Global Strategic Management)

The Regional Multinationals: MNEs and 'Global' Strategic Management

Global Strategic Management

Handbooks in Operations Research and Management Science, Volume 13: Simulation

Handbooks in Operations Research and Management Science, Volume 13: Simulation

Measuring Organizational Performance: Metrics for Entrepreneurship And Strategic Management Research

International Business Scholarship (Research in Global Strategic Management)

Internalization, International Diversification and the Multinational Enterprise, Volume 11: Essays in Honor of Alan M Rugman (Research in Global Strategic ... (Research in Global Strategic Management)

North American Economic and Financial Integration, Volume 10 (Research in Global Strategic Management)

Performance Measurement and Management: A Strategic Approach to Management Accounting

Advances in Environmental Research, Volume 13

Advances in Management Accounting, Volume 13 (Advances in Management Accounting)

Public Management and Performance: Research Directions

Essentials of Strategic Management

Neuromuscular Aspects of Sport Performance

Performance Management in Schools

Research Methodology in Strategy and Management, Volume 3 (Research Methodology in Strategy and Management)

Research Methodology in Strategy and Management, Volume 1 (Research Methodology in Strategy and Management)

Regional and Global Regulation of International Trade

Genetic and molecular aspects of sports performance

Research in Personnel and Human Resources Management, Volume 25 (Research in Personnel and Human Resources Management)

Internationalization of Banks: Strategic Patterns and Performance

Strategic Management

Research on Professional Responsibility and Ethics in Accounting, Volume 13

Progress in Brain Research Volume 13 Mechanism of Neural Regeneration

Performance and implementation aspects of nonlinear filtering

Performance and Fault Management

People Management and Performance

Global Aspects of Complex Geometry

Global Aspects of Complex Geometry

Regional Aspects of Multinationality and Performance, Volume 13 (Research in Global Strategic Management)

Regional Economic Integration, Volume 12 (Research in Global Strategic Management)

The Regional Multinationals: MNEs and 'Global' Strategic Management

Global Strategic Management

Handbooks in Operations Research and Management Science, Volume 13: Simulation

Handbooks in Operations Research and Management Science, Volume 13: Simulation

Measuring Organizational Performance: Metrics for Entrepreneurship And Strategic Management Research

International Business Scholarship (Research in Global Strategic Management)

Internalization, International Diversification and the Multinational Enterprise, Volume 11: Essays in Honor of Alan M Rugman (Research in Global Strategic ... (Research in Global Strategic Management)

North American Economic and Financial Integration, Volume 10 (Research in Global Strategic Management)

Performance Measurement and Management: A Strategic Approach to Management Accounting

Advances in Environmental Research, Volume 13

Advances in Management Accounting, Volume 13 (Advances in Management Accounting)

Public Management and Performance: Research Directions

Essentials of Strategic Management

Neuromuscular Aspects of Sport Performance

Performance Management in Schools

Research Methodology in Strategy and Management, Volume 3 (Research Methodology in Strategy and Management)

Research Methodology in Strategy and Management, Volume 1 (Research Methodology in Strategy and Management)

Regional and Global Regulation of International Trade

Genetic and molecular aspects of sports performance

Research in Personnel and Human Resources Management, Volume 25 (Research in Personnel and Human Resources Management)

Internationalization of Banks: Strategic Patterns and Performance

Strategic Management

Research on Professional Responsibility and Ethics in Accounting, Volume 13

Progress in Brain Research Volume 13 Mechanism of Neural Regeneration

Performance and implementation aspects of nonlinear filtering

Performance and Fault Management

People Management and Performance

Global Aspects of Complex Geometry

Global Aspects of Complex Geometry

Recommend Documents