Contributions to Management Science
For further volumes: http://www.springer.com/series/1505
.
Patrick Heinecke
S...
127 downloads
5092 Views
4MB 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
Contributions to Management Science
For further volumes: http://www.springer.com/series/1505
.
Patrick Heinecke
Success Factors of Regional Strategies for Multinational Corporations Appropriate Degrees of Management Autonomy and Product Adaptation
Patrick Heinecke
This work was originally published as doctoral thesis with the title Success factors of regional strategies for Multinational Corporations: The appropriate degree of regional management autonomy and regional product/service adaptation. This work was accepted as doctoral thesis in 2010 at the University of Erlangen-Nu¨rnberg.
ISSN 1431-1941 ISBN 978-3-7908-2639-5 e-ISBN 978-3-7908-2640-1 DOI 10.1007/978-3-7908-2640-1 Springer Heidelberg Dordrecht London New York # Springer-Verlag Berlin Heidelberg 2011 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Cover design: eStudio Calamar S.L. Printed on acid-free paper Physica-Verlag is a brand of Springer-Verlag Berlin Heidelberg Springer-Verlag is a part of Springer ScienceþBusiness Media (www.springer.com)
Abstract
To enhance understanding of success factors of regional strategies, we examine the effects of regional management autonomy and regional product/service adaptation on the regional success of Multinational Corporations (MNCs) as well as the interaction effects of regional orientation and inter‐regional distance. Based on conceptual insights from different International Business (IB) scholars and recent work in the field of regionalization theory, we develop a comprehensive regional success factor model, which is evaluated by means of the partial‐least‐ squares (PLS) method on the basis of a survey-based inquiry of the Fortune Global 500 firms with success indicator data for a period of 9 years (2000–2008). Our findings suggest that low degrees of regional management autonomy and high levels of regional product/service adaptation are appropriate for MNCs to be regionally successful. Possible adverse effects of high degrees of regional management autonomy on regional success are found to be mitigated by a MNC’s interregional distance. Furthermore, our results indicate that the regional performance associated with high levels of regional product/service adaptations is positively influenced by both a MNC’s regional orientation and its inter-regional distance. In an explorative analysis of the main variables that influence the success factors of regional strategies, we find that the regional strategies and the regional success of MNCs are restricted mainly to their home region. In addition, a trend to expand outside of the home region could be identified, where MNCs following bi-regional strategies were found to achieve the relatively highest corporate success according to our evidence for an S-curve development of the regional strategy-performance relationship. For corporate practice, these research findings imply that MNCs should optimize their regional success by varying their regional management autonomy and regional product/service adaptation in light of their organizational and environmental context, and taking into consideration their specific regional strategy as well as their regional expansion plans.
v
.
Preface
This work was submitted in March 2010 as a dissertation to the Institute of Economics of the University of Erlangen-Nu¨rnberg. Here, it was realized under the supervision of Prof. Dr. Dirk Ulrich Gilbert – mainly during my sabbatical from February to December 2009, which I was granted from the Volkswagen Group. The completion of this work was partly made possible by the support, cooperation and encouragement of several individuals to whom I would like to express my gratitude. First and foremost I would like to thank my doctoral advisor and academic teacher Prof. Dr. Dirk Ulrich Gilbert. He has supported my academic development ever since I first worked for him during my earlier studies in the field of International Management at the European Business School in Oestrich-Winkel. Throughout the entire dissertation process, as well as encouraging my conceptual ideas in an open-minded fashion, he provided thoughtful guidance for this work. I have learned much from his suggestions regarding structure and content that will remain valuable for years to come. I also thank Prof. Ju¨rgen Ka¨hler for being the second advisor of my dissertation and particularly for his helpful methodological advice throughout the entire process. In addition, I want to thank Hans Dieter Po¨tsch, member of the Board of Management of Volkswagen AG, responsible for “Finance and Controlling”. He enabled and encouraged me to take my sabbatical to realize my dissertation. Over more than 2 years as his personal assistant, I have learned much from him about how to design strategies for different regions and how to manage them appropriately. With respect to his tremendous analytical capabilities, as well as his highly motivating and inspiring management style, he will always be a very important role model for my future professional career. In the midst of realizing this work, I was given support and encouragement by many individuals to whom I want to express my special thanks: l
My friend and colleague Ingo Schedel for his support of my work. He has always been a very valuable advisor on content and style, in an exemplary and dedicated manner, even during some very difficult moments of my dissertation. Many thanks, Ingo!
vii
viii l
l
l
l
l
l
l
l l l
Preface
Prof. Dr. Michael Behnam for his conceptual advice and helpful comments on the content of this work. Many thanks for all that I have learned from you, Prof. Behnam, about the theoretical underpinnings of the fascinating field of International Business. Professors Alan M. Rugman and Alain Verbeke, as well as Dr. Andre´ Sammartino, for their helpful comments on the theoretical approach of this dissertation during the IB conference at the University of Reading in the United Kingdom (30–31 March, 2009) and concerning the content of the survey-based inquiry of this work. Dr. Arne Schneider (Associate Principal, McKinsey & Company) for his friendship and support, also regarding his participation in the pretest of the survey of this work. Concerning the latter, I want to express further special thanks to Dr. Detlef Mohr (Director, McKinsey & Company), Dr. Martin Reitz (Head of Investment Banking, Rothschild GmbH), Gerhard Mennecke (Head of Corporate Strategic Planning, Volkswagen Passenger Cars), Roberto Siotto (Head of Corporate Strategic Planning, Volkswagen Group), Dr. Hendric Hallay (Head of Brand Management and Strategy, Volkswagen Passenger Cars), Dr. Mattha¨us Sielecki (Vice President, Deutsche Bank AG), and Dr.-Ing. Julian Eckert (Head of Service-Oriented Computing Research Group, E-Finance Lab Frankfurt am Main e.V.). Dr. Uwe Elsner (Head of Macro and Competitive Analysis, Volkswagen Group) for providing me with macroeconomic data on regions. My friends Demian Nelli (Assistant of the Head of Sales Steering, Volkswagen Passenger Cars) and Markus Grothe from the Institute for Technology and Innovation Management at the Hamburg University of Technology for their methodological advice on a proper utilization of the partial-least-squares method. Stefan Windt for his efforts in setting up a very well-structured Internet platform that facilitated the collection and evaluation of the secondary data of my work. Johannes Friedl, Manuel Heidecke, and my wife Liz for their dedicated support in collecting the secondary data of this work. Dr. Jochen Eßmann for showing me how to properly evaluate my data via Excel. Helen Beckermann for her thorough linguistic review of this book. All of my friends for their support and enjoyable “distractions” during that time.
My deepest gratitude belongs to my wife, my parents and step-parents for their encouragement and unconditional support during this dissertation period, and more broadly, during every phase of my life. My wife Liz has lived through all of my ups and downs during the entire dissertation project. Her love and patience gave me the strength to make my thinking and writing more precise within this work. I dedicate this book to her. Erlangen, Germany
Patrick Heinecke
Contents
Part I
Theoretical Setting for Regional Strategies and Regional Success
1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Introduction and Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Problem Set and Aim of Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Course of Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 3 4 9
2
Theoretical Foundation and Literature Review . . . . . . . . . . . . . . . . . . . . . . . 2.1 Multinational Corporations and Regional Strategies . . . . . . . . . . . . . . . . . 2.1.1 Multinational Corporations and Their Regional Dimensions of Strategy and Organization . . . . . . . . . . . . . . . . . . . . . 2.1.2 Regional Strategies and Their Theoretical Foundation . . . . . . . 2.2 The Rugman Regional Strategy Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 The Development of the Regional Strategy Matrix . . . . . . . . . . . 2.2.2 Home-Regional Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3 Host-Regional Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.4 Bi-regional Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.5 Multi-regional (or Global) Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.6 Further Regional Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.7 Regional Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.8 Regional Products/Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Contingency Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Contingency Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2 Contingency Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Concept of Regional Corporate Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Success Factor Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.2 Regional Corporate Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13 13 15 18 33 33 37 38 39 40 40 43 45 50 50 53 57 57 60
ix
x
Contents
Part II
3
4
Development of a Regional Success Factor Model for the Analysis of the Regional Strategy–Success Relationship
Structural Equation Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Foundations of Structural Equation Modeling . . . . . . . . . . . . . . . . . . . . . . . 3.1.1 Alternative SEM Techniques: LISREL Versus PLS . . . . . . . . . . 3.2 The Partial Least Squares Approach to Structural Equation Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Structural Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Measurement Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Quality Criteria for Evaluating Structural Equation Models . . .
65 65 68 73 73 75 80
Regional Success Factor Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 4.1 Model of Success Factors of Regional Strategies . . . . . . . . . . . . . . . . . . . . 95 4.1.1 Regional Management Autonomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 4.1.2 Regional Product/Service Adaptation . . . . . . . . . . . . . . . . . . . . . . . . 105 4.1.3 Contingency Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 4.1.4 Regional Corporate Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 4.1.5 Control Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 4.2 Summary: Regional Success Factor Model . . . . . . . . . . . . . . . . . . . . . . . . 124
Part III
Success Factors of Regional Strategies: A New Perspective on the Geographic Competitiveness of Multinational Corporations
5
Research Design and Research Methodology . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Research Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.1 Sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.2 Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Research Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1 Methodology of Explorative Data Analysis . . . . . . . . . . . . . . . . . . 5.2.2 Survey Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.3 Modeling Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
129 129 129 131 133 133 138 144
6
Empirical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Explorative Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.1 Data Analysis over Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.2 Further Data Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Model Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
149 149 152 152 155 158
7
Discussion of Results and Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 7.1 Discussion of Theoretical Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 7.2 Discussion of Practical Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Contents
xi
7.2.1 Regional Management Autonomy as a Success Factor . . . . . . 184 7.2.2 Regional Product/Service Adaptation as a Success Factor . . . . 188 7.3 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 8
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 8.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 8.2 Recommendations for Further Research . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
.
List of Figures
Fig. 1.1 Fig. 1.2 Fig. 2.1 Fig. 2.2 Fig. 2.3 Fig. 2.4 Fig. 2.5 Fig. 2.6 Fig. 2.7 Fig. 3.1 Fig. 4.1 Fig. 4.2 Fig. 4.3 Fig. 4.4 Fig. 4.5 Fig. 4.6 Fig. 6.1 Fig. 6.2
Research approach of this work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Course of analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Strategy models of MNCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Structure-stadium model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Multi-regional division of labor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Global integration-national responsiveness framework . . . . . . . . . . . . 29 The regional strategy matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Geographic reach of the FSAs of regional products/services . . . . . 46 Categories of regional product/service adaptation . . . . . . . . . . . . . . . . 49 Structural equation model with two latent variables . . . . . . . . . . . . . . . 66 Structural model of success factors of regional strategies . . . . . . . . . 96 Measurement model of the multidimensional construct regional management autonomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Measurement model of the multidimensional construct regional product/service adaptation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Measurement model of the contingency variables regional orientation and inter-regional distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Measurement model of the endogenous variable regional success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Regional success factor model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Basic structural model of regional success factors . . . . . . . . . . . . . . . 163 Extended structural model of regional success factors . . . . . . . . . . . 165
xiii
.
List of Tables
Table 2.1 Table 2.2 Table 2.3 Table 2.4 Table 3.1 Table 3.2 Table 3.3 Table 3.4
Administrative heritage and management traditions . . . . . . . . . . . . . Forms of regional adaptation of products/services . . . . . . . . . . . . . . Dimensions of inter-regional distance . . . . . . . . . . . . . . . . . . . . . . . . . . . Requirements of “good” success factor research . . . . . . . . . . . . . . . . Comparison of PLS and LISREL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality criteria for the evaluation of reflective measurement models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality criteria for the evaluation of formative measurement models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality criteria for the evaluation of the structural model . . . . . .
31 48 55 58 69 82 86 91
xv
.
Abbreviations
AFTA ASEAN AVE BRIC CA CAGE Cf. CI Co. Corp. CSA CT DIN DUV E.g. EBIT EBITA EBITDA ECB Ed. EU Excl. FDI FL FSA GfK H HRA/TA HRE/TE HRI/TI
ASEAN Free Trade Area Association of Southeast Asian Nations Average variance extracted Brazil, Russia, India, China California Cultural, administrative, geographic, economic Confer (Latin word for: compare, consult) Condition index Company Corporation Country specific advantage Connecticut Deutsches Institut fu¨r Normung Deutscher Universita¨tsverlag Exempli gratia (Latin word for: for example) Earnings before interest and taxes Earnings before interests, taxes, and amortization Earnings before interest, taxes, depreciation, and amortization European Central Bank Editor European Union Exclusive Foreign direct investment Florida Firm-specific advantage Gesellschaft fu¨r Konsumforschung Hypothesis Home-regional to total assets Home-regional to total employees Home-regional to total investments
xvii
xviii
HRIQ HRP/E HRP/TP HRPG HRS/E HRS/TS HRSG I.e. IAS IB IFRS IL IMF Inc. Incl. IR ISO Japan GAAP K.K. Ltd. LVPLS MA MIMIC MIT MNC MNE NAFTA NAICS NJ No. NY PIMS PLS R&D SBU SEC SEM SFAS SIC SME T TIQ TPG TSG
Abbreviations
Home-regional investment quota Home-regional profit per employee Home-regional to total profits Home-regional profit growth Home-regional sales per employee Home-regional to total sales Home-regional sales revenue growth Id est (Latin word for: that is) International Accounting Standards International business International Financial Reporting Standards Illinois International Monetary Fund Incorporated Inclusive Integration-responsiveness International Standards Organization Generally Accepted Accounting Principles in Japan Kabushiki Kaisha (Japanese word for: business corporation) Limited Latent variable partial least square Massachusetts Multiple indicators and multiple causes Massachusetts Institute of Technology Multinational corporation Multinational enterprise North American Free Trade Agreement North American Industry Classification New Jersey Number New York Profit impact of marketing strategies Partial-least-squares Research and development Strategic business unit Securities and Exchange Commission Structural equation modeling Statements of Financial Accounting Standards Standard Industry Classification Small and medium enterprise Tolerance Total corporate investment quota Total profit growth Total sales revenue growth
Abbreviations
UK UN UNCTAD US US GAAP USA VIF Vol. Vs. Yrs.
xix
United Kingdom United Nations United Nations Conference on Trade and Development United States Generally Accepted Accounting Principles in the United States United States of America Variance inflation factor Volume Versus (Latin word for: against) Years
.
Part I
Theoretical Setting for Regional Strategies and Regional Success
.
Chapter 1
Introduction
1.1
Introduction and Motivation
In his seminal book, the Wealth of Nations, Adam Smith (1776) recognizes an important influence from the extent of the market on the competitiveness of firms (Tribe 1995: 24, 26). Nowadays Multinational Corporations (MNCs), the world’s largest companies, face an enormous extent of the market, as nearly all of them operate around the globe. To be competitive on a worldwide level, their design of successful strategies becomes a very challenging task. It involves addressing the differing needs of globally dispersed customers in multiple environments. In the past, the extant research in International Management has widely discussed two opposing approaches to deal with this complexity: the concepts of global integration and local responsiveness (Bartlett 1986; Bartlett and Ghoshal 1989; Doz 1980, 1986; Prahalad 1976; Prahalad and Doz 1987). The global integration approach focuses on similarities across countries. The standardization of products/services and the coordination of efficient operations on a worldwide level leads to economies beyond national borders (Bartlett and Ghoshal 1989: 5). In contrast, local responsiveness concentrates on national differences within a variety of host countries. Differentiated products/services and a high sensitivity to economic, geographical, and cultural differences aim to exploit localmarket needs and opportunities in the various countries in which the firm operates (Doz 1986: 16–17). Thus, these two approaches contrast the managerial perspectives “world” and “country”, each leading to different strategic implications for the product/service offerings and distinct organizational requirements of MNCs. While the “world” approach primarily focuses on the realization of cost advantages from global economies of scale, the “country” orientation aims to achieve revenue advantages from a deep penetration of single markets (Bartlett 1986: 369–371). However, recent research by Alan M. Rugman (2000, 2003a, b, c, 2005a, b, 2007, 2008, 2009) and Alain Verbeke (Rugman and Verbeke 2004, 2005, 2006, 2007, 2008a, b, c; Verbeke 2009) indicates that these perspectives for MNC strategies have to be reframed fundamentally. Rugman’s (2000) analysis of aggregate data on trade and foreign direct investment (FDI) for the triad markets of North America, the European Union (EU), and Asia provides evidence of primarily
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_1, # Springer-Verlag Berlin Heidelberg 2011
3
4
1 Introduction
regional economic activity. The non-existence of an integrated global market (Rugman 2003b: 409; Rugman and Hodgetts 2001: 341), and the overlapping efforts, duplications, and other inefficiencies of “country” approaches lead to the dominance of triad regions for MNC strategies (Rugman 2005b: 205; Rugman and Verbeke 2008b: 310). More specifically, based on an analysis of the Fortune Global 500 firms, Rugman (2003b, c, 2005b) shows that the majority of large MNCs are regionally based in their home region of the triad. Only a few are bi-regional, or host-regional firms, and even less are truly global, deriving their sales evenly from all three triad regions (Rugman 2005b: 4–5). The work of Rugman (2005b) implies that managers should have a more differentiated picture of geographical space when formulating corporate strategy. When looking at a world map, in contrast to prior recommendations, managers should devote special attention to regional, rather than global or national contours (Rugman 2005b: 202). This observation is fundamental, as it is no longer sufficient to polarize the concepts of global integration and national responsiveness, to design strategies for foreign operations. Rather the case in between these geographic extremes should be integrated into the agendas of MNC strategy meetings, where the regional elements of their strategies should receive utmost importance. The recent academic work on regional strategies provides a promising path for generating such conceptual insights – where regional strategies are characterized by an autonomous management at the regional level and by product/service adaptations to regional market requirements (Rugman 2005b: 48–49; Rugman and Verbeke 2006: 124–125).1 A transfer of these theoretical concepts into corporate practice leads to important strategic and managerial implications for MNCs from this new, regional geographic dimension. The main motivation of this work is to further explore these implications for MNCs, to provide conceptual and practical guidance of how to become regionally competitive – thus how to turn this regional market extent into corporate success. The practical relevance of our motivation is highlighted by Rugman (2005b: 2, 34) who strongly recommends top managers should now design such regional, rather than global, solutions to strategy.
1.2
Problem Set and Aim of Analysis
Despite recognizing this advice, a central question for managers still remains unanswered by researchers: “[. . .] why do MNCs succeed as regional organizations [. . .]” (Rugman 2005b: 240). What drives their regional success? What are the success factors of a regional strategy? Which degree of regional management autonomy and regional product/service adaptation is appropriate to be successful?
1
For a more detailed description cf. Sect. 2.2.
1.2 Problem Set and Aim of Analysis
5
This knowledge is essential for MNC managers to develop and implement a regional strategy in a successful manner. We want to enhance the understanding of success factors of regional strategies by building mainly on the work of Rugman (2005b) and related academic scholars. They have made substantial contributions to International Business (IB) research which can be classified into four categories: regions, regional strategies, regional success, and the regional strategy–performance relationship. Regions as a geographical space have already been well-delineated by two research contributions. First, the geographic frontiers of regions have been defined on the basis of important world economies, mostly by the triad markets of Japan/AsiaPacific, United States (US)/North America, and Europe/EU (Ohmae 1985: 9; Rugman 2000: 1; Rugman and Verbeke 2004: 3).2 As these triad markets constitute the main fields of worldwide MNC activities (Rugman 2000: 114–122), this is an adequate definition of regional frontiers. Second, the main differences between regions have already been well-classified into cultural, administrative/political/institutional, geographic, and economic dimensions (Ghemawat 2001: 140; Ghemawat 2008: 41; Ricart et al. 2004: 181). These differences between regions explain the interregional distance that MNCs have to consider in the formulation of their regional strategies. Regional strategies have been explored in three facets; first, regional strategies have been defined in a differentiated manner. Rugman (2005b: 4) distinguishes between home-regional, host-regional, bi-regional, and global strategies. This classification of regional strategies describes the varying degrees of regional penetration that MNCs have achieved by means of their internationalization strategies. Second, the characteristics of Asian, American, and European MNCs from different industries have already been thoroughly investigated (e.g., Collinson and Rugman 2008; Oh 2009; Rugman and Brain 2004; Rugman and Collinson 2004; Rugman and Girod 2003). These findings shed light on the nature and diversity of MNCs and their regional strategies. Third, the regional operations of MNCs, representing the organizational implications of regional strategies, have been categorized into regional management centers (regional headquarters, regional offices) and local offices (Enright 2005a: 66; Enright 2005b: 84–85). The observations about the main functions of these regional management centers provide important insights about their particular roles within the organizational structures of MNCs. Regional success has been studied by empirical evidence on the intra-regional sales revenue of MNCs, where the Fortune Global 500 firms have acted as the dominant sample (Rugman 2005b: 4; Rugman and Verbeke 2004: 6). According to sales revenue from customers’ purchases of the products/services of these firms in worldwide economies, they are the largest companies in the world (Rugman 2005b: 85; Westney 2006: 446). As regional strategies aim to address customer needs in all
2 Geographically these triad regions are distinguished from other types of “regions”, such as subnational regions like states in the USA or semi-autonomous regions like Scotland or Catalonia (Buckley and Ghauri 2004: 91).
6
1 Introduction
regions of the world, the Fortune Global 500 firms are a very good unit of analysis for regional success. The regional strategy–performance relationship has been examined along different aspects of the regional performance of MNCs by various academic scholars (e.g., Chen 2007; Li and Li 2007; Richter 2007; Rugman and Sukpanich 2006b). Their work has contributed to a much better understanding of the distinctive performance patterns of different MNCs’ regional strategies. These theoretical and empirical advancements in studying regions, regional strategies, regional success and the regional strategy–performance relationship reflect the considerable attention that has been devoted to these topics in the academic community in recent years. However, alongside these four categories, the existing approaches in the IB field to explain the success of regional strategies are limited in two aspects, which form the central problem set of this work. 1. The existing research provides a narrow conceptual and theoretical picture of regional geographical space, regional strategies, and regional success The geographic restriction of regions to the triad markets neglects the regions of South America, Africa, and the Middle East (Delios and Beamish 2005: 22). As several of the largest MNCs of the world sell their products/services in these regions, we agree with Proff (2002: 233–234) and Westney (2006: 447) that additional opportunities for non-triad expansion of MNCs exist. Extant literature has paid only little attention to a theoretically founded differentiation of regional strategies vis-a`-vis other strategies of worldwide operating MNCs (Rugman 2005b: 225). In addition, most authors apply the classification thresholds of 50% and 20% of MNC regional sales revenue (Rugman 2005b: 4; Rugman and Verbeke 2004: 7) for classifying regional strategies into homeregional, host-regional, bi-regional, and global strategies. The high sensitivity of this classification to these cut-off points has been criticized, particularly for being overly sensitive to minor shifts in a firm’s sales distribution and for relatively over-/ understating the different regional strategies (Asmussen 2009: 1193; Osegowitsch and Sammartino 2008: 186–188). Regional success has often been defined solely on the basis of regional sales revenue (Rugman 2005b: 4). Furthermore, many studies of the regional success of MNCs have been relatively static, covering only one or two financial reporting periods (e.g., Collinson and Rugman 2008; Delios and Beamish 2005; Rugman 2005b). 2. Existing studies of the regional strategy–performance relationship suffer theoretical and methodological limitations for a proper study of the success factors of regional strategies Much existing conceptual and empirical work focuses either on different phenomena of regional strategies (e.g., Cerrato 2009; Girod and Rugman 2005; Li 2005; Yin and Choi 2005), or on regional performance (e.g., Yip et al. 2006). However, in existing combinations of these two research fields, leading to a regional strategy–performance relationship, a foundation in organization-theory
1.2 Problem Set and Aim of Analysis
7
is largely missing in the IB literature. Only very few authors consider the organizational context of MNCs, thus the fact that the relationship between regional strategies and MNC performance is contingent on conditions and developments inside and/or outside of the firm (e.g., Bausch et al. 2007; Fortanier et al. 2007; Goerzen and Asmussen 2007). Furthermore, regarding the research methodology to study this relationship, we find that much empirical work about the performance of regional strategies focuses on relative figures of regional data – such as regional sales divided by total sales or return on sales (e.g., Li 2005; Li and Li 2007; Rugman and Oh 2007; Yip et al. 2006). This methodology may be eye-opening for regional performance phenomena, but it is not able to explore their causal influences. Thus the relationship between the multinationality of the firm and performance is still an emerging theme in IB research (Griffith et al. 2008: 1227). Consequently, we could not identify any work in the existing literature, which addresses these limitations for a proper study of the underlying factors of regional strategies that drive the success of MNCs. In this work, we aim to address this research gap by broadening the scope of analysis for regions, regional strategies, and regional success, to properly examine the success factors of regional strategies on solid theoretical and methodological grounds. An overview of our research approach is given in Fig. 1.1, and following this we will explore its two main facets in more detail. 1. Conceptual and theoretical extensions to regions, regional strategies, and regional success Subsuming the Middle East under the Asia-Pacific region, we add the regions South America and Africa to the triad markets, and thereby enlarge the relevant geographical space for regional strategies to five regions that encompass all continents of the world. This continental perspective is commonly applied to the definition of regions (Ghemawat 2005: 107). We aim to analyze the characteristics of regional strategies in light of other strategic alternatives according to well-established theoretical approaches in the IB field. Following the research avenue proposed by Osegowitsch and Sammartino (2008), we want to utilize sensitivity analyses of these classification thresholds for regional strategies to test their stability and robustness over time. For regional success, according to the external financial reporting of MNCs and their internal evaluation of alternative strategies (Behn et al. 2002: 33; Epstein and Jermakowicz 2009: 50; Langguth 2006: 629; Lee et al. 2008: 54; Prather-Kinsey and Meek 2004: 215; Schween 2006: 517), we apply a large spectrum of regional success indicators (e.g., relative shares of regional sales, regional profits, and regional assets).3 Furthermore, like Rugman and Oh (2007), we aim to study these regional success indicators in a longitudinal study of the Fortune Global 500 firms.
3
Cf. Sect. 2.4.2 for a more detailed description.
8
1 Introduction Regions
Europe
North America
Inter-regional distance Rest of the world
AsiaPacific
+ South America + Africa + Middle East Regional success factor model
Regional strategy
Homeregional
Hostregional
Regional strategy Biregional
Global
Regional strategy + Theoretical foundation + Variation of classification thresholds
Analysis of the regional strategyperformance relationship:
• Asian, American, European MNCs from different industries • Regional management centers
Regional success
FORTUNE GLOBAL 500
Regional success
• Underlying factors of regional strategies that drive their success • Contingency perspective • Structural equation modeling • Analysis of interaction effects
+ Additional success indicators + Longitudinal study of success
Regional success
• Mainly based on sales revenue • Relatively static
Fig. 1.1 Research approach of this work Source: own illustration
2. Elaboration of a regional success factor model and extensions in methodology to properly study regional success factors of MNCs Based on theoretical grounds, we want to develop a regional success factor model to explore the interactions of regional success factors with MNC performance. In this regional success factor model, we take a contingency perspective for the explanation of the success of the regional strategies of MNCs. In methodological terms, we follow Rugman’s (2005b: 228) suggestion to employ surveys of MNC managers, which we will analyze with modern structural equation models – on the basis of the partial-least-squares (PLS) method, which has not been utilized yet for studying regional strategies of MNCs and their performance. The development of such analytical methods for assessing the regional drivers of MNC success is highly encouraged (Rugman and Hodgetts 2001: 341). We believe that this statistical approach is particularly well suited for the study of
1.3 Course of Analysis
9
this research object, as it allows an analysis of interaction effects from moderating variables – in for example, the external context of MNCs – even for relatively small sample sizes (G€otz and Liehr-Gobbers 2004: 731). The study of such moderating influences on returns from multinationality is considered as a fruitful research topic in the IB field (Griffith et al. 2008: 1227). On this basis, we want to investigate whether regional success is driven by the above mentioned, main elements of regional strategies according to Rugman (2005b: 48–49) – regional management autonomy and regional product/service adaptations. Here, the main focus will be on identifying the degree of these two characterizing factors of regional strategies that is appropriate for a MNC’s superior regional performance – which thus leads to their classification as regional success factors of MNCs. In our opinion it is worthwhile pursuing this research for at least four reasons. First, by utilizing well-established theoretical models in the IB field, we aim to provide additional insights about the main elements of regional strategy in Rugman’s (2005b: 49) regional strategy matrix. For MNC managers, this knowledge of both managerial and product-/service-related success factors seems crucial, to properly translate regional strategies into corporate performance. Second, we hope that extensions of the triad regions, a broadened success definition, detailed sensitivity analyses of classification thresholds, and the application of a longitudinal approach are able to produce further knowledge about both regional strategies and their performance patterns. Third, a regional success factor model, which includes contingency variables, should offer advantages in reflecting the reality of regional MNC strategies and their performance. Recognizing the rigor-relevance debate in studies on success factors (e.g., Gulati 2007; Kieser and Leiner 2009; Kieser and Nicolai 2005; Rasche and Behnam 2009; Tushman and O’Reilly 2007), by improving such model specifications, this approach should both lead to a rigorous research and deliver relevant information for the managerial practice of MNCs. Fourth, the Fortune Global 500 firms include very large MNCs with an important role in the world economy, and a high impact on overall strategic and industry trends (Dunning et al. 2007: 188; Rugman 2000: 2; Rugman 2005b: 224). A study of these firms should thus produce relevant results for management practitioners. Therefore the findings of our study should make a valuable contribution to the existing knowledge in theory and practice about the framing and implementation of regional strategies in MNCs.
1.3
Course of Analysis
We aim to address the theoretical and practical research objectives of this work in three parts. This is illustrated in an overview of the course of analysis of this work in Fig. 1.2, which we will explain below in more detail. In the first part, after introducing the central themes and the research approach of this work (Chap. 1), we lay the theoretical foundations for regional strategies and
10
1 Introduction Success factors of regional strategies for Multinational Corporations: The appropriate degree of regional management autonomy and regional product/service adaptation I Theoretical setting for regional strategies and regional corporate success Chapter 1: Introduction Introduction and motivation Problem set and analysis
Chapter 2: Theoretical foundation and literature review Terminological foundation Theoretical foundation and review
II Development of a regional success factor model for the analysis of the regional strategy-success relationship Chapter 3: Structural equation methodology Methodological foundation Structural equation modeling
Chapter 4: Regional success factor model Regional strategy variables Contingency and success variables
III Success factors of regional strategies: A new perspective on the geographic competitiveness of Multinational Corporations Chapter 5: Research design and methodology Sample and data base Research methodology
Chapter 6: Empirical results Descriptive statistics Explorative data and model evaluation
Chapter 7: Discussion of results and implications Discussion of theoretical and practical results and implications Regional success factors and limitations Chapter 8: Conclusions
Fig. 1.2 Course of analysis Source: own illustration
regional corporate success (Chap. 2). Initially we establish the terminological basis for regional strategic and organizational dimensions of MNCs. The definition of the central terminology for regional phenomena is important, as this constitutes a relatively new research area in the IB field. In this context, we then theoretically substantiate regional strategies by a literature review of other strategic alternatives according to well-established theoretical approaches in the IB field. As this work cannot cover all possible strategic models for MNCs, only those strategic models which had an important theoretical influence on Rugman’s (2005b) concept of regional strategies are presented here. Subsequently we introduce his regional strategy matrix (Rugman 2005b: 49), to explain different regional strategies of MNCs and its constituting variables, regional management and regional products/ services. As an organization theory for MNCs, we present the contingency theory in the following, which adequately captures the fact that a MNC’s successful realization of regional strategies depends on its organizational context in the form of contingency variables. Based on this conceptual setting for regional strategies of
1.3 Course of Analysis
11
MNCs, we then focus on their regional corporate success in our analysis. Here, to ensure both the rigor and relevance of this work, we begin with a critical reflection of success factor theory, before subsequently presenting the conceptualization of regional corporate success. This theory-driven understanding of regional strategies and regional success is important for being able to properly derive hypotheses in the empirical part of this work. In the second part of this work, we develop our own model for the study of the interactions between regional strategies and regional corporate success of MNCs. In Chap. 3 we first provide a methodological foundation for the study of this relationship – by presenting structural equation methodology, and by reflecting on the adequacy of the PLS approach for this work – before we present the particular elements and quality criteria of structural equation models. By this sensitivity in the methodical – as well as in the earlier terminological, theoretical and organizational – foundation of this work, we aim to establish commensurable terminological and theoretical concepts and methods, to ensure a consistent research approach for the study of regional success factors of MNCs. Thereafter in Chap. 4 we apply this modeling methodology to compose our regional success factor model. In a first step, we add the characterizing variables of regional strategy – regional management autonomy and regional product/service adaptation – to the formerly developed success indicators. In a second step, we complement our model by moderating variables that satisfy our contingency view on the relationship under research. Finally, after operationalizing the measurement of regional success, we introduce relevant control variables for our model. Throughout all these steps, we formulate the central research hypotheses on the variable relationships in this regional success factor model. In the last part of this work, we empirically test the aforementioned hypotheses to derive the success factors of regional strategies. We set the stage for this research in Chap. 5 by presenting our research sample and the respective database, as well as the research methodology related to both our primary and secondary data sources. By a subsequent description of this data, its explorative analysis, and its evaluation in the regional success factor model along our research hypotheses in Chap. 6, we generate the empirical results of this work. In Chap. 7 we then discuss in detail the theoretical and practical implications of these findings. Here, using a differentiated analysis of the practical results and their respective implications for MNCs, we derive their characteristic success factors regarding regional management autonomy and regional products/services. However, this analytical separation of regional success factors does not correspond to the strategic reality of MNCs, who mostly have to combine these regional success factors in their regional strategies. Considering this, and the conceptual restrictions of this work, the methodological and theoretical limitations of this research are presented. Finally, this work ends with a short conclusion (Chap. 8).
.
Chapter 2
Theoretical Foundation and Literature Review
2.1
Multinational Corporations and Regional Strategies
In the extant literature different terms are used for MNCs. Apart from MNC, the mostly widely-used terminologies include multinational enterprise (MNE) (e.g., Rugman 2005b), transnational corporation (TNC) (e.g., Morrison and Roth 1992), and international corporation (e.g., Caves 1971). In this work we use the term MNC, as corporate multinationality – the degree of internationalization reflected in the activities of firms (Riahi-Belkaoui 1999: xiii) being organized in more than one national economy (Kreikebaum et al. 2002: 6) – effectively describes the varying geographic spread of corporate activities, which is important for the distinction of different regional strategies. Furthermore, we utilize “corporation” in this work, as this term is not used in a legal sense, like the term “enterprise”: “[. . .] clearly including a network of corporate and non-corporate entities in different countries joined together by ties of ownership [. . .] but rather according to common usage [. . .]” (United Nations 1973: 4). In the explanation of regional strategies of MNCs, we want to address the criticism that much IB research “[. . .] focuses on the ‘why, where and how’ of the MNC, with little attention to the ‘when’ [in time]” (Eden 2009: 535). Following this research recommendation, in establishing a proper definition and terminological basis of regional strategies of MNCs, we briefly outline the main historical developments which have formed them over time. Since the end of the Second World War, either on the basis of FDI or on trade, MNCs have increasingly penetrated international markets (Morrison and Roth 1992: 38). From the 1960s onwards, cross-border economic activity increased significantly, even breaking pre-war records (Ghemawat 2003: 150). This time was popularized as globalization, while still in the 1980s, the triad was limited to the US, Japan, and Western Europe, declaring the expansion across triad regions as the key route for MNC growth (Scholz 1998: 182–183; Westney 2006: 447). Regional strategies from the development of trading blocs did not emerge until
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_2, # Springer-Verlag Berlin Heidelberg 2011
13
14
2 Theoretical Foundation and Literature Review
the 1990s – such as the EU “single market” of 1992, the foundation of the ASEAN Free Trade Area (AFTA) in 1992 and the North American Free Trade Agreement (NAFTA) in 1994 – which enabled MNCs to benefit from growing FDI opportunities inside of their own regions (Enright 2005b: 84; Krugman 1993: 58; Morrison et al. 1991: 23–24; Proff 2002: 233; Rugman and Verbeke 1998b: 115; Rugman and Verbeke 2004: 4; Westney 2006: 447). While such free-trade zones have still benefited from positive effects by external FDI (Buckley et al. 2001: 267–268; Feils and Rahman 2008: 156), the economic data shows that, at the same time, intraregional trade and FDIs have increased significantly (Dunning 2001: 308; Rugman 2000: 114–122; Rugman 2005b: 216; Rugman and Hodgetts 2001: 334–335). By the beginning of the millennium, after “[. . .] the Asian financial crisis, episodes of instability in Russia and Latin America, a perceived “globalization backlash”, a global economic slowdown, and the war on global terrorism” (Ghemawat 2003: 150) – again the increasing trend towards regionalization, rather than globalization, was observed (Giddens 2001: 19; Rugman 2005b: 216). This was further supported by the fact that regional integration is an efficient substitute for alternative multilateral integration processes which in the case of the World Trade Organization (WTO) requires concession of all its member countries, by now more than 150 (Krugman 1993: 73; Rugman 2005b: 220–221; Rugman and Verbeke 2003b: 85; Rugman and Verbeke 2004: 5).1 Many other influences, such as: “[. . .] limits to global economies of scale, regional differences in markets and employees, physical and psychic distances between regions and the corporate headquarters, regional variations in business rules and social behavior, and the heterogeneity of the economies of some regions [. . .]” (Enright 2005b: 84) have led to a semi-globalized world today, where the region plays a pivotal role in the cross-border integration of markets (Ghemawat 2003: 150; Ghemawat 2007b: 57; Ghemawat 2008: 30–31). Along with these past developments, MNCs have continued to adjust their strategies and organizational designs, as well as relocate their subsidiaries, to align their product-/service-related and functional market interfaces with the requirements of their geographic markets (Doz and Prahalad 1991: 146). For the cross-border configuration of these corporate activities, regionalization does suggest MNCs should exploit their strengths and determine their competitive strategies separately for each region (Morrison et al. 1991: 26). Such regional strategies of MNCs, whether directed to their own region or to foreign regions, require a marketstrategy approach tailored to regional requirements and the creation of region-based components in the MNC’s: “[. . .] coordination and control structure to address the specific managerial challenges in those regions” (Rugman 2005b: 196). Recognizing both the macroeconomic environment of regionalism described above and its impact on the firm-level over the last years, we believe that two aspects of regional strategies of MNCs require further analysis.
1
For the political consequences of regionalization regarding the denationalization of countries cf. Walter and Z€urn (1998: 48–58).
2.1 Multinational Corporations and Regional Strategies l
l
15
The MNC itself should be further explored, in particular its strategic and organizational facets, as these have undergone significant changes in the course of regionalization over recent years. This is in line with Rugman’s (2005b: 221) claim that the MNC is the appropriate unit of analysis for IB research in regional integration. In addition to recent scholarly work on regional strategies, their theoretical foundation should also include historical approaches and frameworks. Such a procedure takes into account the effect that regionalization has had on the academic perspective of internationalization strategies. The theoretical implications of these earlier findings have contributed to our understanding of regional strategies today. This sensitivity to, and awareness of, the nature of historical evidence is crucial for a proper interpretation of more recent research results about regionalization phenomena like the regional strategies of MNCs (Jones and Khanna 2006: 465).
2.1.1
Multinational Corporations and Their Regional Dimensions of Strategy and Organization
Historically the MNC and its strategic and organizational characteristics became central to IB research in the 1970s (Buckley 2002: 367).2 Since then, MNCs have usually been described as large firms (Chandler 1990: 132; Vernon and Wells 1986: 1–2) with operations crossing national borders (Rugman 2005b: 224): “[. . .] that, at least to some extent, attempt to organize, coordinate and control these units in accordance with a common strategy” (Stopford and Wells 1972: 5). The extent of coordination and control of its operations is influenced by the essential difference between MNCs and simpler organizations: the combined consequences of multidimensionality and heterogeneity (Doz and Prahalad 1991: 146). The fact that MNCs are business organizations with integrated operations across more than one nation (Bartlett et al. 2003: 2; Caves 1971: 1; Caves 1996: 1; Verbeke 2009: 3; Verbeke and Kenworthy 2008: 941), leads to a high degree of strategic and structural as well as political multidimensionality (Doz and Prahalad 1991: 146). Heterogeneity is a result of: “[. . .] the differences between the optimal trade-offs for different businesses, countries, functions and tasks as a function of a whole range of economic and political characteristics which differ between countries and affect individual businesses and tasks in quite varied ways” (Doz and Prahalad 1991: 146). Thus the strategies and organizational configurations of MNCs have both a multidimensional and heterogeneous nature. 2
This impulse came from two directions – first, from Chandler’s (1962) resource based theory and second, from the internalization approach which derived from ideas of Coase (1937) and from concepts related to Williamson’s (1975) transaction cost economics approach (Buckley 2002: 367–368).
16
2 Theoretical Foundation and Literature Review
In a world of increasing regionalization, the multidimensionality and heterogeneity of MNC strategies is influenced by cross-border differences, depending on the firm’s geographic breadth – meaning the scope of the market that is served by the MNC (Roth and Morrison 1992: 478). Taking into consideration their underlying logic of growth and competition (Chandler 1990: 131), MNCs basically have three options for value creation when responding to these regional differences: adaptation, aggregation and arbitrage (Ghemawat 2008: 105). The strategic MNC dimension adaptation implies coping with cross-country differences by partially modifying a basic template by local changes (Ghemawat and Ghadar 2006: 619), which still leads to some economies of scale. Financed by firm-level investments, regional adaptations often imply production relocations to the most efficient subunits, resulting in regional scale economies and a review of the subsidiaries’ functions (Rugman 2005b: 222). Further examples of possible levers of adaptation include product variations, a focus on certain customer segments, externalization by franchising, product designs on the basis of platforms, or innovations from business model recombinations (Ghemawat 2008: 115–130). All these adaptations aim at competitive advantages that boost MNC revenues and market share (Ghemawat 2007a: 60). The strategic MNC dimension aggregation involves exploiting the similarities across several countries, by somehow overcoming their differences and as a consequence, reaping increasing returns to scale (Ghemawat 2003: 148). Regional aggregation occurs, if competitive advantages of MNCs can be deployed across national borders, but only in a limited geographic region (Rugman 2005b: 70). These intra-regional opportunities are supported by free-trade zones, tax regulations, inner-regional preferences, and currency unifications (Ghemawat 2008: 142). MNCs can realize aggregation advantages within the region by standardizing products/services and grouping processes of development and production (Ghemawat 2007a: 60). The strategic MNC dimension arbitrage involves exploiting differences among countries, such as different absolute cost levels or variations in willingness-to-pay (Ghemawat 2003: 148). The advantages of regional arbitrage are realized mostly across regions (Rugman 2005b: 70), and may include an inter-regional division of the supply chain along different locations, for example: “[. . .] call centers in India, factories in China, and retail shops in Western Europe” (Ghemawat 2007a: 60). Furthermore, regionalization affects the multidimensionality and heterogeneity of the MNC organization by adding new dimensions of product/service and/or geographical diversification (Pearce and Papanastassiou 2006: 152). Assuming hierarchical coordination methods to organize cross-national interdependencies (Hennart 2000: 72), MNCs establish increasingly complex administrative structures and enlarge their centralized planning facility for the coordination and organization of new geographical activities (Pearce and Papanastassiou 2006: 152). However, the fact that a complete central planning and management of these interdependencies in MNCs is not possible, leads to a decentralization and delegation of decisionmaking, which gives rise to opportunities for linkages between subunits at various locations and organizational levels (Doz and Prahalad 1991: 147). Consequently, in
2.1 Multinational Corporations and Regional Strategies
17
the course of regionalization, the regional organization is increasingly resuming responsibilities from both centralized and decentralized units (Enright 2005a: 61). Mainly, these regional management units are responsible for the coordination of product-/service-related, geographic, and functional activities (Williams 1967: 87). In the literature, two types of such regional management centers are distinguished: regional headquarters and regional offices (Daniels 1987: 410; Enright 2005a: 66; Enright 2005b: 84; Grosse 1981: 48; Heenan 1979: 410; Lasserre 1996: 33–35; Lasserre and Probert 1998: 49; Morrison and Roth 1992: 52; Rugman 2005b: 73; Rugman and Verbeke 2004: 14; Sch€utte 1997: 441; Williams 1967: 87–88; Yeung et al. 2001: 158). Both regional headquarters and regional offices are defined as offices that have: “[. . .] control over the operation of one or more other offices and subsidiaries in other economies or countries in the region [. . .]” (Enright 2005b: 84). The central difference between these two management centers is the broadened regional management autonomy of regional headquarters, including the supervision of corporate planning, budgets, and performance evaluations in the region (Morrison and Roth 1992: 52). As a consequence of their high autonomy for both coordination and control, regional headquarters do: “[. . .] not need to make frequent referrals to the overseas parent companies” (Enright 2005a: 66).3 These referrals to the parent headquarters are more common for regional offices, which are rather “coordination offices” with limited staff support, who maintain their contact with subsidiaries mainly by fax and telephone, as well as frequent 1-day meetings (Morrison and Roth 1992: 52). We agree with Ghemawat (2008: 150) that only on the basis of properly designed regional strategic dimensions, and a clear sense of how regional concepts are supposed to add value, can the coordination requirements at the regional level be specified. Therefore, in the following, we will usually use the term regional management center, and only if more detailed analyses are necessary for our research, we will differentiate between regional headquarters and regional offices. In such cases, given that we want to investigate regional management autonomy in this work, we follow the recommendation of Rugman and Verbeke (2001: 248) in focusing on regional headquarters, which are situated in regional continents such as America, Europe and Asia. Recognizing the explanations above, we extend Rugman’s (2005b: 224) definition of a regional MNC, as: “[. . .] a firm with the majority of its sales inside one of the triad regions [. . .]”, to the following: A regional MNC offers its products/services at the level of regional continents on the basis of its strategic dimensions of adaptation, aggregation, and/or arbitrage, each leading to different organizational configurations of coordination and control in regional management centers.
3
Cf. Lasserre’s (1996: 33) typology of regional headquarters regarding their different roles in MNCs.
18
2 Theoretical Foundation and Literature Review
2.1.2
Regional Strategies and Their Theoretical Foundation
2.1.2.1
Overview of Strategy Models for Multinational Corporations
Geographic level of analysis
In the course of scholarly research about internationalization strategies over recent years, various strategy models for MNCs have been developed by different IB scholars. Their conceptualizations of strategy vary from approaches that are formulated either for the MNC as a whole or only for certain parts of the firm, such as subsidiaries. In addition, their models can be differentiated along their geographic level of strategy analysis, focusing on the “world” and/or “country” level, or the “region”. A classification of MNC strategy models according to these levels of analysis is illustrated in Fig. 2.1.4 The horizontal axis shows the two main fields of analysis in MNC research (Birkinshaw and Pedersen 2009: 368–369). On the vertical axis, the main strategic options of MNCs are depicted – which arise from geographic spread, cross-national integration, and local adaptation (Tallmann and Yip 2009: 327–334) – reflecting the strategic MNC dimensions of adaptation, aggregation and arbitrage. The appropriateness of a particular MNC strategy model for our research depends on its coverage of the characteristics of regional MNCs which, according to the definition outlined above, are
Region
Triad power (Ohmae, 1985) Regional strategy matrix (Rugman, 2005b)
Structure-stadium model (Stopford & Wells, 1972) Multifocal strategy (Doz, 1986; Prahalad & Doz, 1987)
World/ country
International,multinational, global, and transnational strategy (Bartlett, 1986; Bartlett & Ghoshal, 1989) Diamond model (Porter, 1990), double diamond model (Rugman & Verbeke, 1993; Rugman & D’Cruz, 1993)
Subsidiary strategy paradigm (White & Poynter, 1984) Susidiary strategy typologies (Birkinshaw & Morrison, 1995)
Whole firm
Parts of the firm MNC level of analysis
Fig. 2.1 Strategy models of MNCs Source: own illustration
4 Some strategy models may, in certain aspects, cover more than one field in the matrix of Fig. 2.1. Therefore it should be noted that our classification is based on the main focus of the strategy models presented here.
2.1 Multinational Corporations and Regional Strategies l l l
19
Regional product/service offerings Regional adaptation, aggregation, and/or arbitrage Organizational implications for regional management centers
Regarding the level of our analysis, the strategy-related emphasis of our research is primarily on the MNC as a whole firm.5 The geographic level of our strategyrelated analysis is the region. This focus of our research, as also mentioned in the explanation of our aim of analysis,6 is best reflected in Rugman’s (2005b: 49) regional strategy matrix – where both regional products/services and regional management centers are integral parts of this matrix.7 Furthermore, Rugman (2005b: 70, 77–78, 197, 203, 222, 235) explicitly applies the strategy dimensions of regional adaptation, aggregation, and arbitrage to generate further knowledge about his regional strategies. In addition to the regional strategy matrix, as outlined above, we aim to consider historical developments in the IB field for our conceptualization of regional strategy. As we described previously, in the past, regional strategy was mainly affected by the regionalization of economies. This influence was effective at the level of the whole firm, introducing the “region” as a new geographic level of analysis for MNC strategies. Therefore, at the level of the MNC as a whole, we will consider three MNC strategy models in the theoretical foundation of regional strategy. First, we will consider Ohmae’s (1985) concept of a triad power in our analysis, as the triad regions constitute the relevant geographical space for Rugman and Verbeke’s (Rugman 2000: 1; Rugman 2005b: 202; Rugman and Verbeke 2004: 4) explanation of regional strategies. Furthermore Ohmae’s (1985: 206, 209) MNC strategies of “insiderization” and market commitment to become a triad power suitably describe both limitations and opportunities of regional strategies along the dimensions of adaptation, aggregation, and arbitrage (Rugman 2005b: 59, 66, 233–234; Rugman and Verbeke 2004: 4). Second, we utilize Bartlett and Ghoshal’s (Bartlett 1986; Bartlett and Ghoshal 1989) international, multinational, global, and transnational strategies, as these strategic alternatives are directly compared to Rugman’s (2005b: 206–212) conceptualization of regional strategies (Rugman and Verbeke 2008c: 330), offering important insights into their specific characteristics. Third, recognizing the importance of historical IB theory, we employ the structure-stadium model of Stopford and Wells (1972), which represents an important antecedent to Bartlett and Ghoshal’s (1989: 30–33) work (Pla-Barber 2002: 152; Rugman 2005b: 201). This theoretical approach offers insights for the
5
It should be noted here that – while it could be argued from an organizational perspective that regional management centers are a part of the firm – we view the degree of decision-making autonomy granted to regional management centers as a managerial implication of the overall MNC strategy, formulated at the level of the firm as a whole. 6 Cf. Sect. 1.2. 7 For a more detailed description cf. Sect. 2.2.
20
2 Theoretical Foundation and Literature Review
explanation of the organizational aspects of regional strategy, as it shows how elements of MNC strategies during their worldwide expansion – given by foreign sales and foreign product diversity – relate to organizational implications (Wolf and Egelhoff 2002: 182). At the level of the MNC as a whole, we will not address Doz and Prahalad’s (Doz 1986; Prahalad and Doz 1987) multifocal strategy, as it constantly seeks a tradeoff between the conflicting demands of worldwide integration and national responsiveness, for example by coordinating collective operations, while being responsive to each local context (Doz and Prahalad 1991: 146; Roth and Morrison 1990: 544). Thus the perspective of this strategy is both global and local, but not regional. However, Doz and Prahalad’s (Doz 1980, 1986; Prahalad 1976; Prahalad and Doz 1987) integration-responsiveness (IR) framework, which is grounded in the work of Lawrence and Lorsch (1967), will indirectly be considered as being an important foundation of Bartlett and Ghoshal’s (2002: 7, 10) work. Furthermore we will exclude the diamond model (Porter 1990) and the double diamond model (Rugman and D’Cruz, 1993; Rugman and Verbeke 1993) from the theoretical explanation of regional strategy. This is justified by the fact that these models explain the national competitive advantage of MNCs by country-specific attributes (Asmussen et al. 2009: 42; Dunning 1993: 7; Rugman and Verbeke 1993: 72), which are already well-captured in Rugman’s (2005b: 34–35, 202) concept of country-specific advantages (CSAs), which we will explore later. As a consequence of our strategy-related research focus on the MNC as a whole, instead of its parts, we will disregard the subsidiary strategy paradigms of White and Poynter (Poynter and White 1984; White and Poynter 1984), and the subsidiary strategy typologies of Birkinshaw and Morrison (1995). Regarding the empty field in Fig. 2.1, we could not identify an approach that explicitly develops a strategy model for certain parts of the firm at the regional level.8 An exclusion from the following theoretical foundation of regional strategy, however, does not mean that we will not sometimes draw on particular insights of these MNC strategy models, if beneficial to our research. In the subsequent presentation of the central strategy models for this work, we follow a chronological order along their historical occurrence, as we aim to consider the sequence of knowledge development in the IB community. This procedure allows the distinction between already existing and newly developed knowledge about MNC strategies.
8
Also in the work on regional management centers and/or regional offices (e.g., Enright 2005a: 66; Enright 2005b: 84; Grosse 1981: 48; Heenan 1979: 410; Lasserre 1996: 33–35; Yeung et al. 2001: 158), we could not identify any explicit strategy model for these parts of the firm. This might be due to the fact that the establishment of a regional organizational level is a result of the overall MNC strategy (e.g., to achieve intermediate solutions between local responsiveness and global integration) (Enright 2005a: 61). Consequently, the roles and particular functions of these regional management units are a function of the regional strategy of the MNC as a whole, which is addressed by the strategy models of Ohmae (1985) and Rugman (2005b).
2.1 Multinational Corporations and Regional Strategies
2.1.2.2
21
The Stopford and Wells Structure-Stadium Model
Building on the work of Chandler (1962), the structure-stadium model of Stopford and Wells (1972: 63) describes the development of the organizational structure of MNCs while they are becoming more involved in worldwide markets (Morschett 2007: 283; Stopford and Wells 1972: 10). Along different degrees of multinationality in their expansion across borders, MNCs can follow alternate paths of structural change with varying performance implications (Stopford and Wells 1972: 79–84). The degree of multinationality is measured by two elements of MNC strategy: foreign product diversity and foreign sales as a percentage of total sales (Stopford and Wells 1972: 63). The structure-stadium model results from the interaction of these two variables, specifying four different strategic domains, each of which are associated with a different type of structure (Wolf and Egelhoff 2002: 182). Figure 2.2 illustrates the structure-stadium model. Early in the development of their foreign activities, MNCs expand and diversify through one international division which is specialized for the coordination of a relatively small set of foreign operations (Stopford and Wells 1972: 21, 63; Wolf and Egelhoff 2002: 187). The MNCs’ increasing cross-border expansion and/or product diversification lead to intolerable organizational stress for the international division, requiring alternate paths of structural development: the establishment of area divisions or worldwide product divisions (Stopford and Wells 1972: 64–65). Area divisions are geographical region structures (e.g., in the form of regional
Worldwide product division
Alternate paths of development
Foreign product diversity (The number of products sold internationally)
Global matrix
International division
Area division
Foreign sales as percentage of total sales (Importance of international sales to the company)
Fig. 2.2 Structure-stadium model Source: own illustration on the basis of Stopford and Wells (1972: 63), Kreikebaum et al. (2002: 117), and Ghoshal and Nohria (1993: 23–24)
22
2 Theoretical Foundation and Literature Review
management centers) that – as an organizational response to the regional integration of production inside economic blocs – mainly coordinate and control intra-regional production, sourcing and product development (Stopford and Wells 1972: 59–60; Wolf and Egelhoff 2002: 187). However, if increases in product diversity imply higher external market diversity as well as internal technical diversity, a worldwide product division structure is needed, to coordinate the increasing product-related information-processing between the centers of product knowledge in the parent headquarters and the foreign subsidiaries (Stopford and Wells 1972: 41; Wolf and Egelhoff 2002: 183). At high levels of foreign sales and international product diversity, the coordination and control skills of both area divisions and product divisions are required. However, the unity of command in each of these divisions – implying isolated responsibilities and decision-making structures – leads to high inter-divisional communication barriers, which can only be overcome by the dual or multiple reporting relationships of global matrix structures with shared product and area responsibilities (Fouraker and Stopford 1968: 49–50; Pla-Barber 2002: 143; Stopford and Wells 1972: 27, 87). Therefore any increases in complexity, either driven by a rise in international sales or in product diversity, imply a structural challenge to MNCs (Stopford and Wells 1972: 6). Given that the structure-stadium model of Stopford and Wells (1972: 63) – as described in the previous chapter – serves to explain the organizational challenges that result from regional strategies about regional product/service offerings,9 the structural implications of area and product/service divisions at the regional level form the center of the forthcoming analysis. Area and product/service divisions represent structural alternatives, each allocating responsibility differently among senior managers (Stopford and Wells 1972: 10). Such managerial decision-making, at high levels of diversification along both product/service and geographical dimensions, is largely delegated within the MNC (Stopford and Wells 1972: 74–75). Therefore, a high regional decision-making autonomy is a consequence of a MNC’s product/service diversification and its commitment to regional markets. The administration of products/services by organizational structures at the regional level is encouraged to ensure a close intra-regional coordination of product/service and technical information flows (Stopford and Wells 1972: 43). The importance of such regional product/service coordination is supported by the fact that high levels of foreign R&D are associated with the geographical region structure (Wolf and Egelhoff 2002: 187). At mature stages of products/services and technologies, where marketing becomes the critical function, benefits from area coordination of products/services further increase (Stopford and Wells 1972: 43). In these cases, even though being best-suited for a single product line, area divisions may be efficiently adapted to foreign product diversification by the utilization of 9
While Stopford and Wells (1972) describe their concepts solely in the context of products – as they do not differentiate between products and services in their work – in our explanations here, we will add the service dimension where this is appropriate and/or useful for the theoretical foundation of this work.
2.1 Multinational Corporations and Regional Strategies
23
common functional facilities for all product lines, such as distribution channels (Stopford and Wells 1972: 80–81). Area structures in MNCs result from a shift in the balance of power towards regional units, reflecting the impact of growing foreign activities on the geographical balance of the business (Stopford and Wells 1972: 51). Instead of area diversity per se – such as the number of foreign countries in which a MNC has manufacturing sites – only increasing foreign sales change the geographical balance of the business and thus lead to the development of area divisions (Stopford and Wells 1972: 53). At the regional level, these area divisions mainly coordinate marketing and the rationalization of production (Stopford and Wells 1972: 62, 114). The regional coordination of marketing may include a regional standardization of brand names, image, advertising, and uniform terms of sale and service (Stopford and Wells 1972: 56–57). The regional rationalization of production, accompanied by a standardization of quality and product mix, leads to lower manufacturing costs (Stopford and Wells 1972: 60–61, 115). Besides these activities, regional management groups are responsible for planning the allocation of production and markets, thereby coordinating and controlling inter-regional transfer pricing (Stopford and Wells 1972: 59).10 The analysis above shows that regional MNC strategies are associated with regional product/service structures to administer technical and R&D related information, and with regional area structures to coordinate marketing and production. The establishment of such product/service divisions and area divisions at the regional level should always be in line with the overall regional MNC strategies of diversification and geographical expansion. Such a match between strategy and structure give MNCs a performance advantage over other firms, whose structural characteristics are not well-tailored to strategic and managerial needs (Stopford and Wells 1972: 82, 84). Regarding the direction of this fit between strategy and structure, Stopford and Wells (1972: 5–6) view limitations of their principal assumption that “structure follows strategy” (Chandler 1962: 383) – where organizational structure has to meet the administrative requirements of strategy – as they notice that reorganizing costs and resisting interest groups may inhibit the introduction of MNC strategies. The structure-stadium model has been largely confirmed and, in some cases, extended by subsequent research (Daniels et al. 1984, 1985; Davidson and Haspeslagh 1982; Egelhoff 1982, 1988a, b; Franko 1976; Pla-Barber 2002), while supporting its underlying principle of a fit between strategy and structure (Pla-Barber 2002: 154; Wolf and Egelhoff 2002: 182). However, the structurestadium model has been criticized regarding four aspects. First, even though being descriptive in nature, it was utilized prescriptively by consultants, academics, and
10
Area-based organizations mainly employ wholly-owned subsidiaries instead of joint ventures, while product divisions are associated with both types of ownership (Stopford and Wells 1972: 147–148).
24
2 Theoretical Foundation and Literature Review
managers, fostering the belief that every major strategic problem had a structural solution of simplistic choices between “centralization” and “decentralization”, or between product- and geography-based structures (Bartlett 1986: 368; Bartlett and Ghoshal 1989: 29–31; Kreikebaum et al. 2002: 121). Secondly, the global matrix in particular was criticized, as its dual chains of command and reporting relationships implied overlapping responsibilities, which routinely cause duplicated communications, costly and time-consuming approval processes, frequent meetings and regular intra-organizational travel (Bartlett and Ghoshal 1989: 31–32, 200). Third, Stopford and Wells (1972) fail to address a detailed and critical discussion of the direction of the fit between strategy and structure (Kreikebaum et al. 2002: 121–122), which would have contributed to a better understanding of the organizational dynamics of their four structural alternatives. Fourth, most importantly and compounding the previous point, the model focuses on only one element in the fit between structure and strategy, the formal structure of the MNC – representing solely the organizational anatomy of the firm (Bartlett and Ghoshal 1989: 201; Ghoshal and Nohria 1993: 24; Kreikebaum et al. 2002: 151) – that: “[. . .] could not capture the complexity of the strategic task facing the worldwide company” (Bartlett and Ghoshal 1989: 32). Consequently Stopford and Wells (1972) consider only static elements of the internal firm structure – such as roles, responsibilities, and relationships – in a dynamic and rapidly developing task environment of MNCs (Bartlett and Ghoshal 1989: 32), thus they fail to address the external context in the conceptualization of their organizational models (Kreikebaum et al. 2002: 122). Recognizing this criticism, this work uses the structure-stadium model mainly as an orientation for the configuration of an organization (Kreikebaum et al. 2002: 121), particularly concerning the main functions of product/service and geographical area structures at the regional level, as these represent organizational implications of a regional strategy.
2.1.2.3
The Ohmae Triad Power
Ohmae (1985: 9) observes a concentration of the most important industries worldwide in a triad, a geographical space consisting of Japan, the US, and Europe. These triad markets share many commonalities, such as low macroeconomic growth, similarities in customer demand and in technological infrastructures, large capital and knowledge intensive MNCs, and protectionist pressures (Ohmae 1985: 3, 23–25; Rugman and Verbeke 2004: 4). Furthermore, Ohmae (1985: 45) recognizes that while MNCs have a very solid competitive position within their own triad region, they experience difficulties in repeating their home triad performance in foreign triad markets, a phenomenon which he calls global impasse (Ohmae 1985: 36). This results from the dilemma that for MNCs, building market share in foreign triad markets involves high risks from enormous upfront investments and fixed costs – for setting up a distribution network and for developing strong products – while conversely, if successfully sold
2.1 Multinational Corporations and Regional Strategies
25
at home, such so-called “superproducts” could be easily copied and sold by competition in major markets (Ohmae 1985: 39; Rugman 2005b: 59). For noncapital intensive products/services, such upfront MNC efforts are rather related to the establishment of time-honored, trust-based relationships with customers (Ohmae 1985: 40). Global impasse therefore illustrates limitations of exploiting differences across regions by a strong product/service, the arbitrage dimension of regional MNC strategies. To overcome these limitations of a market dominance restricted to the home region, Ohmae (1985: 165) recommends MNCs to become a triad power that: “[. . .] has (1) equal penetration and exploitation capabilities, and (2) no blind spots in each of the triad regions [. . .]”. With regards to the first point, an equal and deep penetration into each triad region, is required for recovering the firm’s productrelated costs and investments (Ohmae 1985: 165), particularly for R&D and distribution (Ohmae 1985: 12–13). As for the second point, the non-existence of blind spots is important to avoid surprising moves by foreign or domestic competitors (Ohmae 1985: 165), and to benefit from a global-scale utilization of strong worldwide brands and distribution channels (Hamel and Prahalad 1985: 142; Rugman and Verbeke 2004: 4). MNCs can achieve this aggregation dimension of regional MNC strategies via a “sprinkler model”, meaning a rapid flooding of their key markets by selling their products/services simultaneously all over the globe (Ohmae 1985: 17, 32). A triad power combines the above described aggregation dimension with an adaptation dimension of successfully establishing an insider position in all triad markets (Ohmae 1985: 209). One reason behind such insiderization for MNCs is to become relatively independent of protectionist barriers (Ohmae 1985: 30), or even to influence governmental policies in foreign regions (Ohmae 1985: 84). Even nontariff barriers from structural and cultural differences between triad regions – such as distinct distribution systems, different business practices, value systems and preferences, and/or languages – can be overcome by insiders, especially in the long run (Ohmae 1985: 89, 92, 100–101; Ohmae 1989a: 158). MNCs can achieve such a long-term approach by their staying power as an insider, reflecting their commitment, creativity and competitiveness (Ohmae 1985: 110–111). Commitment is related to time, money, and effort of MNCs, while creativity stands for groundbreaking approaches to structural and cultural differences such as innovative distribution systems (Ohmae 1985: 111–113). Meanwhile competitiveness results from establishing a MNC’s overseas business like a new company with a novel business strategy (Ohmae 1985: 113–114). Their significant presence in all triad regions as a staying power with a well-established business model covering all corporate functions – being highly responsive to local and regional customers and conditions by active communication, and the autonomous decision-making of a continuous regional management – are necessary conditions for MNCs to reach a robust insider position in foreign regions (Ohmae 1985: 206; Ohmae 1989a: 155). It takes years to establish such a full business system with competent managers that are familiar with regional customer needs and the key linkages and relationships with the environment (Ohmae 1985: 51, 77).
26
2 Theoretical Foundation and Literature Review
From an organizational viewpoint, MNCs can become such an insider by consortia and joint ventures (Ohmae 1985: 171–180; Ohmae 1989b: 154; Rugman 2005b: 233) or through wholly owned units, while in the latter case they should establish regional headquarters – each with a geocentric mentality focusing on both worldwide and local objectives (Perlmutter 1969: 13) – in every respective triad region (Rugman and Verbeke 2004: 4–5).11 In addition to these triad markets, Ohmae (1985: 121) adds a fourth, developing region to the relevant geographical space of MNC strategies. The resulting fourheaded triad constitutes Ohmae’s (1985: 121–123) tetrahedron model, in which US MNCs should focus on South America, European MNCs on Africa and the Middle East, and Japanese MNCs on Asia in their selection of the fourth region.12 All four regions are covered by a multi-regional company with its regional headquarters in the critical tetrahedron regions (Ohmae 1985: 184). This structure involves aggregation advantages, as common functions can be shared in and across the regions, in: “[. . .] such a way as to gain synergistic advantage over competitors’ quality and cost position” (Ohmae 1985: 184). At the same time however, in the adaptation of products/services according to the requirements of foreign triad customers, MNCs should be sensitive regarding the amount of modification (Ohmae 1985: 104–105). These adaptations, given the fact that a differentiation of upstream functions – such as purchasing and production – is not possible, have shifted downstream to the customer, where the source of differentiating products lies in: “[. . .] image, logistics, distribution, service, and [the] degree of perfection in execution” (Ohmae 1985: 32). To realize the appropriate degree of aggregation and adaptation, MNCs should divide their functions – for each of their tetrahedron regions – among corporate headquarters, regional headquarters, and country management (Ohmae 1985: 185–186), as illustrated in Fig. 2.3. The light grey shaded area in Fig. 2.3 represents the functions of regional headquarters of multi-regional MNCs. Exemplary for the manufacturing function, the corporate headquarters focuses on the transfer of production technology, whereas country subsidiaries are responsible for productivity control and labor relations. The regional headquarters centrally coordinates the capital expenditure plan at the regional level, while synchronizing the sourcing of the production plan with inner-regional decentralized units. The white shaded areas show such benefits of centralization and decentralization for country subsidiaries and corporate headquarters, along the different value chain functions.13 Dividing the multi-regional MNC’s functions into support activities, as well as upstream and downstream functions (Ohmae 1985: 34; Porter 1985: 37; Porter 1986: 21, 24), shows that
11
Ohmae (1985: 123) describes this mentality as an “Anchorage perspective” of a corporate center being symbolically located in Anchorage (Alaska), and thus being equidistant from the economic and political centers of the triad regions in Japan, the US, and Europe (Rugman and Verbeke 2004: 4). 12 For general directions given to MNCs from developing regions in their triad-oriented expansion cf. Ohmae (1985: 160–161). 13 For a detailed description of different roles of world headquarters cf. Ohmae (1985: 194–204).
2.1 Multinational Corporations and Regional Strategies Role player National subsidiary
Benefits of centralization Local law
27 Benefits of decentralization
• Accounting • Working capital management
Budget of subsidiary
Local manager
• Productivity control • Labor relations
• Pricing • Advertising • Distribution channel policy
Multi-regional MNC
Product innovation
Regional headquarters
Cash Management Regional Public Relations Regional plan
Local executive training
Specific R&D
• Sourcing production plan
Corporate headquarters
Country analysis
Public Relations/ Legal
World strategy
Planning
• Basic R&D technology information
R&D
Regional headquarters of a multi-regional MNC
• Capital expenditure plan • Foreign exchange • Intercompany loan • Top management • Transfer pricing • Corporate values • Long-term funding
Finance
Support activities
Personnel
Transfer of production technology
Export
Manufacturing
Upstream functions
Brand policy
Sales
Downstream functions
Fig. 2.3 Multi-regional division of labor Source: own illustration on the basis of Ohmae (1985: 186)
moving down the value chain, an increasing level of management autonomy is granted to the regional headquarters and particularly to the national subsidiaries. To prevent regional headquarters from becoming: “[. . .] an extra layer of control and/or a bureaucratic nuisance” (Ohmae 1985: 188), its use should be determined by an analysis of the consumers’ product/service preferences and their degree of worldwide universality, thereby deriving the main implications for the MNCs’ key functions (Ohmae 1985: 189, 191). Such an analysis highlights opportunities of regional strategies along the dimensions of adaptation, aggregation, and arbitrage. For example, regional aggregation benefits may arise from commonalities in the triad customers’ basic product/service requirements – allowing the utilization of basic modules and concepts – leading to economies of scale and competitive cost advantages (Ohmae 1985: 189–190). By a careful evaluation of its strategy dimensions, a multi-regional MNC can realize a high payoff from an insider position in the triad markets (Ohmae 1985: 119). Even though Ohmae’s (1985) findings have been criticized for not considering the development of the broad triad of NAFTA, the expanded EU, and Asia in his tetrahedron model (Rugman and Verbeke 2004: 5), his insights and concepts lay important foundations for Rugman and Verbeke’s (Rugman 2005b: 59–62, 202; Rugman and Verbeke 2004: 4–6) analysis of regional strategies, and thus represent crucial research contributions for this work. Despite recognizing the importance of Ohmae’s (1985) approach for the conceptualization of regional strategies, he explains MNC strategies solely in the light of triad-based expansion patterns. Alternative strategies of worldwide operating
28
2 Theoretical Foundation and Literature Review
MNCs, however, receive little attention, or explicit theoretical foundation. This will be presented in the following approach for strategy models of MNCs.
2.1.2.4
The Bartlett and Ghoshal Integration-Responsiveness Framework
Building on Chandler (1962), Vernon (1966), Lawrence and Lorsch (1967), Stopford and Wells (1972), and on the IR model of Doz and Prahalad (Doz 1986; Prahalad and Doz 1987), Bartlett and Ghoshal (Bartlett 1986; Bartlett and Ghoshal 1987a, b, 1988, 1989; Ghoshal 1987) describe how external forces for global integration, national responsiveness, and worldwide learning shape both the internal strategic approaches, and organizational structures and processes of MNCs (Bartlett and Ghoshal 1987b: 7; Bartlett and Ghoshal 1989: 3; Bartlett and Ghoshal 2002: 10–11; Buckley 2002: 369; Kreikebaum et al. 2002: 148). The forces for global integration emerge from a worldwide homogenization (Levitt 1983: 102) and convergence of consumer tastes, falling transportation and communication costs, and new economic, social and technological developments which offer MNCs competitive aggregation advantages from considerable global economies of scale, experience curve effects, and/or economies of scope (Bartlett and Ghoshal 1987b: 8; Bartlett and Ghoshal 1989: 5–6; Ghemawat 2003: 149; Kreikebaum et al. 2002: 148). Examples of how MNCs can achieve these global efficiencies include standardization and rationalization of their products and processes, and specialization of their manufacturing operations (Bartlett and Ghoshal 1989: 5; Kreikebaum et al. 2002: 148). The forces for national responsiveness originate from different conditions in distinct national markets, where MNCs can achieve competitive adaptation advantages by offering locally differentiated products/services, thus being responsive to the needs and requirements of consumers and governments in national markets (Bartlett and Ghoshal 1987b: 9; Bartlett and Ghoshal 1989: 8–9; Ghemawat 2008: 113). This flexibility to adapt to multiple national markets and local conditions can be achieved by flexible manufacturing and a significant local presence, to critically observe national markets and to ensure a high sensitivity to local market differences (Bartlett and Ghoshal 1989: 8–10; Kreikebaum et al. 2002: 148). The forces for worldwide learning result from the increasing cost of R&D, shortened life cycles of technologies, and local content regulations, which together require MNCs to internationally develop and diffuse worldwide innovations, to quickly amortize their heavy R&D investments over global volume (Bartlett and Ghoshal 1987b: 10; Bartlett and Ghoshal 1989: 12). MNCs achieve worldwide learning by linking and leveraging knowledge in multiple national markets, thus realizing competitive arbitrage, adaptation and aggregation advantages (Ghemawat 2007a: 63): “[. . .] by sensing needs in one country, responding with capabilities located in a second, and diffusing the resulting innovation to markets around the globe” (Bartlett and Ghoshal 1989: 12). These external forces, both separately and in combination, lead to conflicting strategic demands that require different internationalization strategies of MNCs
Forces for global integration
2.1 Multinational Corporations and Regional Strategies
High
Low
Global strategy
International strategy
Low
29
Transnational strategy
Multinational strategy
High
Forces for national responsiveness
Fig. 2.4 Global integration-national responsiveness framework Source: own illustration on the basis of Bartlett (1986: 377), Bartlett and Ghoshal (1989: 65, 97), Ghoshal (1987: 429), Gilbert (1998: 47), and Kreikebaum (1997: 257)
(Bartlett and Ghoshal 1989: 14–16). This is illustrated in Bartlett and Ghoshal’s (Bartlett 1986: 377; Bartlett and Ghoshal 1989: 65, 97; Ghoshal 1987: 429) global integration-national responsiveness framework (Gilbert 1998: 47; Kreikebaum 1997: 257), as illustrated in Fig. 2.4. Responding to worldwide learning forces, MNCs with international strategies exploit the knowledge and capabilities developed at the parent headquarters by their transfer and adaptation to overseas markets (Bartlett and Ghoshal 1987b: 10; Bartlett and Ghoshal 1989: 15). In this process, instead of leaving adaptations to autonomous and independent national subsidiaries, international companies mainly build on a strong functional management which establishes worldwide standards and specifications for the rapid cross-border diffusion of new products/services (Bartlett and Ghoshal 1989: 13–15, 137). MNCs achieve a much higher sensitivity and responsiveness to national differences by multinational strategies, with strong national entities in multiple markets that largely manage their local businesses independently (Bartlett 1986: 371; Bartlett and Ghoshal 1987b: 9–10; Bartlett and Ghoshal 1989: 14). Therefore multinational companies employ a strong geographic management to identify and flexibly respond to local opportunities on a country-by-country basis (Bartlett and Ghoshal 1989: 137). MNCs with global strategies aim to exploit an integrated and unitary world market by the realization of cost advantages at highly centralized global scale operations (Bartlett 1986: 371; Bartlett and Ghoshal 1987b: 9; Bartlett and Ghoshal 1989: 14–15). The centralization of strategic and operational decisions in global companies results in a dominance of business management to rationalize manufacturing, to standardize products/services, and to achieve low-cost global sourcing (Bartlett and Ghoshal 1989: 15, 137).
30
2 Theoretical Foundation and Literature Review
In contrast to the three internationalization strategies previously outlined, which cope with a single dominant strategic MNC demand (Bartlett and Ghoshal 1989: 16), a transnational strategy aims to simultaneously achieve global efficiencies, multinational responsiveness and flexibility, and worldwide transfers of parent company knowledge and competencies (Bartlett 1986: 372; Bartlett and Ghoshal 1989: 25). This primarily implies an organizational challenge for transnational companies (Bartlett and Ghoshal 1989: 16), as multidimensional strategic capabilities have to be developed, to properly exploit and complementarily manage multiple sources of competitive advantages from adaptation, aggregation, and arbitrage (Bartlett 1983: 145–146; Bartlett and Ghoshal 1989: 33; Ghemawat 2008: 218). According to Bartlett and Ghoshal (1989: 33), the MNC’s ability to develop and manage the new strategic capabilities highly depends on its administrative heritage which is defined as the MNC’s: “[. . .] existing organizational attributes – its configuration of assets and capabilities, built over the decades; its distribution of managerial responsibilities and influence, which cannot be shifted quickly; and an ongoing set of relationships that endure long after any structural change”. The administrative heritage of MNCs is mainly shaped by leadership, home country culture, and organizational history (Bartlett and Ghoshal 1989: 41). The effect of the latter, organizational history, on administrative heritage and management traditions is illustrated in Table 2.1 and will be explained in more detail, as this work aims to address historical developments. During the inter-war period, many European MNCs applied a multinational organization model in their cross-border expansion, a decentralized federation with decentralized assets and responsibilities in nationally self-sufficient units, each exploiting local opportunities and generating knowledge independently (Bartlett 1986: 372–373; Bartlett and Ghoshal 1989: 49, 65). These firms typically applied socialization as a means of coordination, where negotiations between knowledgeable groups – who share common objectives, values, and perspectives – were preferred over authoritarian decision-making processes (Bartlett and Ghoshal 1989: 69, 163). During the post-war period, many American MNCs expanded by an international organization model, a coordinated federation with centralized core competencies at the parent company, which were the basis for transferring knowledge and expertise to overseas operations (Bartlett and Ghoshal 1989: 49–50, 65). To coordinate their units, these companies largely utilized formalization – establishing formal systems and procedures – to realize operating efficiencies by decisionmaking routines (Bartlett and Ghoshal 1989: 69, 161). From the mid 1960s to the late 1970s, Japanese MNCs utilized the global organizational model, a centralized hub with centralized assets, knowledge, and responsibilities to build global economies of scale at their corporate headquarters, based on the market reach of overseas operations (Bartlett 1986: 372–374; Bartlett and Ghoshal 1989: 51, 65). These firms coordinated their units mainly by centralization – substantial decision-making and interventions by the corporate headquarters’ management group – which limited the role of overseas firms to implementing the parent company’s strategies (Bartlett and Ghoshal 1989: 65, 69, 158).
Multinational strategy: Internationalize through rationally differentiated and responsive strategies
MNC strategies
Improved transport, communications Global strategy: Internationalize by capturing global scale advantages
Leverage management skills International strategy: Internationalize through a transfer of technological and managerial innovations
Intensified integration of world markets
Consolidation of industries Transnational strategy: Internationalize by simultaneously realizing global economies of scale and national responsiveness Japanese MNC Transnational expansion expansion by global by transnational organization model organization model (centralized hub) (integrated network)
Homogenizing markets
Transfer of new technologies
Post-war reconstruction
Mid 1960s to late 1970s 1980s to beginning of 1990s Falling trade barriers End of Cold War
Asian financial crisis Global terrorism Regional strategy: Internationalize by realizing regional adaptation, aggregation, and/or arbitrage advantages
Free trade areas and regional integration Inter-regional distance
Mid 1990s until today
MNC organization models
Regional expansion by American MNC European MNC regional expansion by expansion by organization model international multinational (regional organization model organization model management center) (coordinated (decentralized federation) federation) Source: own illustration on the basis of Bartlett and Ghoshal (1989: 45–55), Bartlett et al. (2003: 342–348), Rugman (2005b: 192, 197), Ghemawat (2003: 148–150), Ghemawat (2007b: 59), and Enright (2005b: 84)
National market differences Transport/ communication barriers Rising protectionism
Environmental trends
Table 2.1 Administrative heritage and management traditions Inter-war period Post-war period
2.1 Multinational Corporations and Regional Strategies 31
32
2 Theoretical Foundation and Literature Review
From the 1980s onwards, Bartlett and Ghoshal (Bartlett 1986: 380–381; Bartlett and Ghoshal 1987a: 44; Bartlett and Ghoshal 1989: 59–61, 65) view an increased necessity of MNCs to combine the above described multinational, international, and global approaches by expanding transnationally – based upon a transnational organizational model, an integrated network with centralized and decentralized, interdependent and specialized assets and responsibilities of both national units and integrated worldwide operations (Kreikebaum et al. 2002: 149). A transnational MNC selectively applies centralization, formalization, and/or socialization – along businesses, functions, areas, and tasks – as coordination mechanisms (Bartlett and Ghoshal 1989: 69–70; Ghoshal 1987: 438).14 This gives MNCs the flexibility to simultaneously optimize economies of scale, scope and factor costs, to provide responsive reactions to local environments, and to share and develop knowledge jointly worldwide (Bartlett and Ghoshal 1988: 73; Bartlett and Ghoshal 1989: 59–60, 65). Therefore any changes in competitive strategies or in formal organizational structures should be accompanied by matching changes in the MNC’s values and management processes (Bartlett and Ghoshal 1989: 54). In such decisions about organizational configurations, the opportunities and limitations of a particular MNC’s administrative heritage have to be considered on equal terms with economic, political, or any other environmental pressures (Bartlett and Ghoshal 1989: 81). Consequently, according to the environment-strategy-structure paradigm of Bartlett and Ghoshal (1989: 53), superior MNC performance primarily results from a: “[. . .] good fit between corporate strategy and environmental demands, and between organization structure and strategy” (Bartlett and Ghoshal 1989: 53), which has been confirmed by empirical research (Ghoshal and Nohria 1993: 32–33). Criticisms of Bartlett and Ghoshal’s (1989) work mainly focus on the high difficulties in building transnational capabilities – such as varying requirements of different organizational structures, costly and time-consuming socialization mechanisms, complications in decomposing interdependent MNCs into national units with different roles (Rugman 2005b: 195–196; Rugman and Verbeke 2008b: 306), duplications of competencies, and complex coordination processes (Kreikebaum et al. 2002: 152). This might be explained to a large degree by the fact that Bartlett and Ghoshal (1989) studied MNCs in the sense of business divisions (Bartlett and Ghoshal 2002: 20), where such transnational capabilities can be developed more easily than for the whole firm. Consequently, and particularly regarding the changes from 1980 to 2000, Bartlett and Ghoshal (2002: 9) state that: “[. . .] many of the old models, concepts, and frameworks that [they] developed to describe the workings of an earlier model of MNC no longer prove as powerful and robust as they once did”. 15 This work supports their insight with Table 2.1, reasoning that most MNCs today – driven by the increasing regionalization and inter-regional distance over recent 14
For a conceptualization of differentiated subsidiary roles and responsibilities in a transnational organization cf. Bartlett and Ghoshal (1989: 105–111) and Kreikebaum et al. (2002: 150). 15 Here it should be noted, that Bartlett and Ghoshal (1989: 142) recognized the existence of regional headquarters and regional offices, without however further conceptualizing their implications for MNCs and their strategies.
2.2 The Rugman Regional Strategy Matrix
33
years – tend to follow regional strategies to selectively exploit adaptation, aggregation, and/or arbitrage opportunities within regions (Ghemawat 2001: 140; Ghemawat 2003: 148–150; Ghemawat 2005: 107–108; Ghemawat 2007a: 60; Ghemawat 2007b: 59; Ghemawat 2008: 105; Ghemawat and Ghadar 2006: 619–620; Ricart et al. 2004: 181; Rugman 2005b: 192, 197). These MNCs develop regional organizational models in the form of regional management centers (Enright 2005a: 66; Enright 2005b: 61–63). Therefore in the following, we will further explore regional strategies, regional management centers and their coordination mechanisms, as well as environmental developments that affect strategic and managerial issues at the regional level. The inclusion of such external trends in our analysis – given their importance for analyzing MNC performance in a regional environment-strategy-structure relationship – is highly encouraged by the academic IB community (Czinkota and Ronkainen 2009: 263; Verbeke et al. 2009: 157).
2.2 2.2.1
The Rugman Regional Strategy Matrix The Development of the Regional Strategy Matrix
In the IB field, Rugman (2000: 1–2) has recognized the trend towards regionalization, based on an analysis of FDI and trade data. He shows that the economic activities of MNCs are primarily triad-based in the regions EU, NAFTA, and Asia (Rugman 2000: 114). The dominant paradigm in many MNCs are intermediate degrees of globalization (Asmussen 2009: 1192; Ghemawat 2003: 150; Rugman 2005b: 62) particularly, as the focus of their business activities is rather intraregional – mainly directed towards their own region – than across several triad regions (Rugman 2000: 114–122; Rugman 2003b: 409). As a result, MNCs are advised to design strategies on a regional rather than a global basis (Rugman 2003c: 3; Rugman and Hodgetts 2001: 341; Rugman and Moore 2001: 68). Following a resource-based view, Rugman (2005b: 46–48, 201–206) shows how this new geographical perspective affects firm-specific advantages (FSAs) of MNCs in the global integration-national responsiveness framework of Bartlett and Ghoshal (1989) in Fig. 2.4. FSAs are unique capabilities belonging to the firm, which are developed on the basis of either product or process technologies, marketing or distributional skills (Rugman 2000: 87; Rugman 2005b: 34; Rugman and Verbeke 2005: 12). Thus, reflecting the knowledge base for the competitiveness of MNCs, examples of FSAs include technological know-how, brand names, marketing abilities, or even firm size and industry type (Rugman 1981: 61; Rugman and Sukpanich 2006a: 357; Rugman and Verbeke 1991: 5; Rugman and Verbeke 2004: 6).16 16
A routine application and recombination of these capabilities may generate FSAs in the form of the MNC’s core competencies (Prahalad and Hamel 1990: 82; Verbeke 2009: 77; Zentes 1998: 179).
34
2 Theoretical Foundation and Literature Review
According to Rugman (2005b: 47), MNCs can deploy two different types of FSAs in the global integration-national responsiveness framework of Fig. 2.4, either location-bound FSAs and/or non-location-bound FSAs. Location-bound FSAs are firm-specific capabilities which can be exploited only in a limited geographic space (Rugman 2005b: 195). They may lead to benefits of national responsiveness along the horizontal axis of Fig. 2.4, if the MNC addresses local concerns in its countries served – such as national consumer tastes and governmental regulations (Rugman 2005b: 47). From this Rugman (2005b: 129) concludes that location-bound FSAs are influenced by changes in country-specific advantages (CSAs) of MNCs. CSAs are benefits from locating certain activities in particular countries (Rugman and Verbeke 1992: 762), as parameters exogenous to the MNC which influence its competitiveness along political, economic, cultural, and financial dimensions of countries (Rugman 2005b: 34–35; Rugman and Verbeke 2005: 12). Examples of CSAs are labor markets, country cultures, regulations, infrastructure, and natural resources such as minerals, energy, or forests (Rugman 2005b: 34–35). At the country-level, for instance, CSAs such as new patent legislations influence the location-bound FSAs of pharmaceutical firms – the technological know-how of their patented products within this national market (Rugman 2005b: 129). Non-location-bound FSAs are internationally transferable capabilities of MNCs (Rugman and Verbeke 2001: 238; Verbeke 2009: 14). MNCs can deploy and exploit non-location-bound FSAs across borders by following a global integration approach along the vertical axis of Fig. 2.4 (Rugman 2005b: 47). As this implies producing and distributing their products/services of a homogeneous type and quality worldwide, MNCs – by means of non-location-bound FSAs – can generate economies of scale and scope, and benefits from exploiting national differences (Rugman 2005b: 194–195). An example of a non-location-bound FSA, if valued highly by worldwide customers, is the outstanding production quality of a automobile manufacturer – which is internationally transferable, as it is embodied in its globally exported vehicles (Verbeke 2009: 14). By adding the regional dimension to location-bound FSAs, Rugman (2005b: 50) introduces region-bound FSAs as company strengths beyond the limited geographic scope of a single country. However, as these FSAs cannot be deployed globally, such a regional responsiveness approach of MNCs is still location-bound within the countries of a particular region (Morrison et al. 1991: 27; Morrison and Roth 1992: 45; Rugman 2005b: 50). By being regionally responsive to the requirements imposed by regional host environments, MNCs can improve their penetration of, and market success in, these host regions (Rugman 2005b: 197). Both the MNC’s location-bound and region-bound FSAs may be affected by CSAs like an EU or NAFTA regulation at the regional level – while WTO, IMF, or UN policies usually act at the global level (Kolk 2005: 160; Rugman 2005b: 38; Rugman and Kirton 1998: 449; Rugman et al. 1999: 225–228; Rugman and Verbeke 1991: 4–5; Rugman and Verbeke 1998a: 821–822; Rugman and Verbeke 1998b: 123; Rugman and Verbeke 2003b: 84). Regarding the regional exploitability of non-location-bound FSAs, Rugman (2005b: 47) shows that MNCs may achieve an integration approach not only at the
2.2 The Rugman Regional Strategy Matrix
35
MNC’s product/service characteristics
global, but also at the regional level. Within this regionally limited geographic space, MNCs can realize regional integration benefits similar to global integration benefits – such as regional economies of scale and scope (Morrison and Roth 1992: 45; Rugman 2005b: 47). The key to regionally exploiting the above described benefits from intermediate levels of national responsiveness and global integration may be best achieved by a regional adaptation approach (Morrison and Roth 1992: 46; Rugman 2005b: 49, 197). This should be accompanied by a high organizational responsiveness of MNCs by their organizational capabilities – which are location-bound FSAs, as different strategies for each country or regional market may be required (Rugman 2005b: 48). Such sub-global responsiveness implies the allocation of considerable resources, and the delegation of substantial decision-making to the regional level (Rugman 2005b: 50). Therefore, according to Rugman (2005b: 52), a proper analysis of the regional dimension in MNC strategies requires a focus on the: “[. . .] actual extent of triadbased decision-making power [. . .] and/or adapted products and services [. . .]” of MNCs. As such an analysis requires a decomposition of the MNCs’ strategic decision-making processes and their product/service offering into global, regional, and national components, a more complex analytical tool than the global integration-national responsiveness framework of Fig. 2.4 is required (Rugman 2005b: 52, 225). Consequently, Rugman (2005b: 49) introduces the regional strategy matrix, as illustrated in Fig. 2.5. The horizontal axis of Fig. 2.5 represents the management level for strategic decision-making in the MNC’s organizational structure, being concentrated either
1
Global product / service
4
2
Trans-regional (world mandate) strategy
Bi-regional strategy
Multi-regional (global) strategy
5
Directiveregional strategy
Regional product / service
7
Host-regional strategy
8
Cooperativeregional strategy
Regional strategy Home-regional strategy
3
National product / service
6
9
International strategy
Adaptiveregional strategy
Multinational strategy
Corporate headquarters
Regional management centers
National subsidiaries
Management level for strategic decision-making in the MNC’s organizational structure
Fig. 2.5 The regional strategy matrix Source: own illustration on the basis of Rugman (2005b: 49)
36
2 Theoretical Foundation and Literature Review
at corporate headquarters, regional management centers, or national subsidiaries (Rugman 2005b: 49, 203; Rugman and Verbeke 2006: 124–125). These different decision-making locations lead to appropriate levels of organizational responsiveness to national, regional, or global market concerns (Rugman 2005b: 49; Rugman and Verbeke 2006: 125). Thus, the decision-making locations reflect the breadth (Roth and Morrison 1992: 478) or geographic scope of MNCs, which is defined as: “[. . .] the distribution of the firm’s assets, activities, and sales across geographic space” (Rugman and Verbeke 2005: 10). The geographic scope of the MNC’s activities and operations should correspond to the locus of its competitive advantages (Rugman 2005b: 37, 182). These locational choices represent managerial reactions to locational advantages from CSAs of the firm (Rugman 2005b: 37). In contrast, the vertical axis reflects the actual product/service characteristics according to the market strategies of MNCs (Rugman 2005b: 48–49, 203). Here national and regional products/services – which are adapted to national or regional requirements – are differentiated from global products/services that may be standardized on a worldwide level (Rugman 2005b: 203). These different market strategies of MNCs, pursuing either responsiveness or integration objectives, vary along the geographic reach of their FSAs – which can be either location-bound, region-bound and/or non-location-bound (Rugman 2005b: 37, 49, 70, 221). Consequently in cell 5, MNCs follow a regional strategy which according to Fig. 2.5 is defined as the regional adaptation of products/services and a high decision-making autonomy of regional management centers (Rugman 2005b: 49). Within their regional environments, by means of these region-based adaptation strategies and organizational structures, MNCs can achieve both regional responsiveness and regional integration benefits (Rugman 2005b: 49–50, 201; Rugman and Verbeke 2006: 125; Rugman and Verbeke 2008b: 310). Thus the regional strategy matrix effectively illustrates a regional environment-strategy-structure relationship. In addition to Rugman’s (2005b: 51) positioning of the international strategy in cell 3 and of the multinational strategy in cell 9 of Fig. 2.5 (Rugman and Verbeke 2008b: 310–311), which have been previously presented,17 the regional strategy matrix sheds light on various new strategic alternatives for MNCs – particularly different types of regional strategies. Before going on to explain in more detail these alternative regional strategies, it should be noted that Rugman’s (2005b) approach for their conceptualization has been exposed to several important criticisms. The first criticism concerns the offering of a narrow theory of regionalization (Clark et al. 2004: 513). Related criticisms include the lack of a direct measurement of his main theoretical concepts like FSAs, or the missing details about the specific regional management forms that have worked well in certain industries and firms following regional strategies (Westney 2006: 447–448). Second, his study of regional phenomena has been criticized for being limited to triad regions, to sales-based data, and solely to the
17
Cf. Sect. 2.1.2.4.
2.2 The Rugman Regional Strategy Matrix
37
Fortune Global 500 firms (Delios and Beamish 2005: 20–23; Stevens and Bird 2004: 504–505; Westney 2006: 446). Third, methodological criticisms include his classification of regional strategies based only on 50% and 20% thresholds, and the missing differentiation of sales in the home country versus sales in the home region (Asmussen 2009: 1193; Dunning et al. 2007: 178; Osegowitsch and Sammartino 2008: 186–188; Stevens and Bird 2004: 503; Westney 2006: 446). At the same time, his critics highly acknowledge this work for pushing the IB field’s attention towards regional elements (Westney 2006: 447–448), which in many MNCs: “[. . .] are becoming increasingly important [. . .], both in terms of organizational structure and market approach” (Rugman 2005b: 206). Pursuing this research avenue – after a description of different regional strategies – we further explore regional management and regional products/services.
2.2.2
Home-Regional Strategy
According to Rugman (2005b: 38), home-regional MNCs are firms, whose geographic reach of their FSAs and locational scope is limited to their home region. Consequently, the strategies of these MNCs are characterized by region-bound FSAs and regional management structures, placing them in cell 5 of Fig. 2.5. Based on an analysis of regional sales data for the Fortune Global 500 firms,18 Rugman and Verbeke (2004: 7) categorize MNCs as following such a homeregional strategy if they: “[. . .] have at least 50% of their sales in their home region of the triad”. They show that the majority of MNCs, more than 80% (or 320 firms) according to their data, are oriented towards their home region (Rugman 2003a: 1–2; Rugman 2003b: 412; Rugman 2003c: 4; Rugman 2005a: 444; Rugman 2005b: 4–5; Rugman and Verbeke 2004: 7). Their findings of a strong home-regional orientation have been confirmed for MNCs from Europe (Beleska-Spasova and Glaister 2009: 301; Oh 2009: 342; Rugman and Collinson 2005: 258; Rugman et al. 2007: 300; Yip et al. 2006: 248), North America (Hejazi 2007: 24; Moore and Rugman 2003: 2), Asia (Collinson and Rugman 2008: 216; Oh and Rugman 2007: 39; Rugman and Collinson 2006: 168; Yin and Choi 2005: 110), the emerging markets (Banalieva and Santoro 2009; Sethi 2009), and for particular industries (Girod and Rugman 2005: 335–336; Oh and Rugman 2006: 165; Rugman and Collinson 2004: 472; Rugman and Girod 2003: 32; Rugman and Verbeke 2008a: 405). An example of an industry with many home-regional MNCs is the pharmaceutical sector, with its heavily regulated markets of North America and Europe, and its institutional frameworks for distribution and marketing (Rugman 2005b: 136). 18 In Rugman and Verbeke’s (Rugman 2005b: 3; Rugman and Verbeke 2004: 7) study of the Fortune Global 500 firms, no data was available for 120 firms, so their actual sample included 380 firms.
38
2 Theoretical Foundation and Literature Review
In general, such investigations of home-regional strategies are based upon Rugman and Verbeke’s (2004: 7) sales-based classification threshold of 50%, which we presented previously. Even though this classification threshold of 50% may indicate a substantial home-regional influence on the decisions and actions of MNCs (Rugman 2005b: 64; Rugman and Verbeke 2004: 7), the choice of this cutoff point turns the classification overly sensitive to the 50% value (Asmussen 2009: 1193; Osegowitsch and Sammartino 2008: 186). In addition, the assumption that a geographical sales concentration solely reflects priorities in decision-making and corporate actions, misses theoretical underpinnings (Osegowitsch and Sammartino 2008: 186). Home-regional sales also indicate a geographic focus in the deployment of FSAs (Osegowitsch and Sammartino 2008: 186–187), which implies a home region-specificity of the MNCs’ capabilities. Thus, possibly driven by the administrative heritage of MNCs (Bartlett and Ghoshal 1989: 33; Osegowitsch and Sammartino 2008: 186), FSAs are configured primarily to penetrate home region markets, to address home region-bound customer needs. As shown in Fig. 2.5, such a home-regional strategy of a MNC should be measured by the degrees of its homeregional product/service adaptation and its organizational responsiveness to homeregional demands. Both of these influences do not act on a domestic market, but at the level of the home region. Therefore, we believe that – even though a distinction of domestic from home-regional sales might lead to a better picture of the geographic reach of FSAs within the home region (Asmussen 2009: 1201; Osegowitsch and Sammartino 2008: 187) – the level of the home region is the appropriate level of analysis for home-regional strategies. This is supported by Ghemawat’s (2001: 140), Ghemawat’s (2005: 100, 102) and Ghemawat’s (2008: 34) analysis of different facets of cultural, administrative/political/institutional, geographic and economic distance that act on a regional level – e.g., regional trading blocs and currency unifications.19 As a result, specific home-regional FSAs and the emphasis of home-regional decision-making to address regional demands are the main influencing factors of home-regional sales. These elements of MNCs should be more explicitly emphasized in the conceptualization of their home-regional strategies. Consequently, we define a home-regional strategy as the exploitation of home region-bound FSAs and the focus on home-regional decision-making, to achieve the majority of sales in the home region.
2.2.3
Host-Regional Strategy
Rugman and Verbeke (2004: 7) show that almost 3% (or 11 firms) of the MNCs in their sample are host region oriented firms with “[. . .] more than 50% of their sales in a triad market other than their home region”. Host-regional MNCs were 19
In Sect. 2.3.2.2, these different dimensions of inter-regional distance will be presented in more detail.
2.2 The Rugman Regional Strategy Matrix
39
encountered for example in the automotive industry, where a foreign host region offered better growth opportunities than the home-regional market (Rugman 2005b: 147–148). As the same arguments apply principally here as in the previous home-regional strategy, we define a host-regional strategy as follows: a host-regional strategy is the exploitation of host region-bound FSAs and the focus on host-regional decision-making, to achieve the majority of sales in the host region. As the MNC’s host-regional strategies align its FSAs with host region-bound demands and requirements (Rugman 2005b: 43), they achieve a geographic reach outside of the firm’s home region. Therefore, even though being host-regionally limited, MNCs following host-regional strategies realize a higher geographic reach than firms pursuing home-regional strategy. This is indicated by placing the hostregional strategy above the home-regional strategy in cell 5 of Fig. 2.5.
2.2.4
Bi-regional Strategy
Bi-regional MNCs are characterized by a significant presence in two regions, thus having a global reach for their FSAs, but no global geographic scope (Rugman 2005b: 38). Approximately 7% (or 25 firms) in Rugman and Verbeke’s (2004: 7) sample are bi-regional firms with: “[. . .] at least 20% of their sales in each of two regions, but less than 50% in any one region”. These bi-regional MNCs usually derive between 20% and 50% of their sales from both their home region and another triad region (Rugman and Verbeke 2004: 7). The financial services industry is comprised of a high amount of such bi-regional MNCs, which generally focus on North America and Europe (Grosse 2005: 129, 139). While the threshold of 20% of sales was chosen to reflect considerable market success from FSA deployments in two regions (Rugman 2005b: 64; Rugman and Verbeke 2004: 7), the huge size of MNCs and the heterogeneity in product market sizes warrant lower cut-off points (Osegowitsch and Sammartino 2008: 188). Therefore, and again to avoid the high sensitivity to the operationalization of classification thresholds (Asmussen 2009: 1195),20 we focus on those MNC strategies that achieve a considerable part of their sales in each of the two regions. Such bi-regional strategies are positioned in cell 4 of Fig. 2.5, as the global, non-location bound nature of FSAs is reflected in their reach beyond one particular region (Girod and Rugman 2005: 351; Rugman 2005b: 40). At the same time, bi-regional organizational structures reflect the limited geographic scope of the MNC’s activities (Rugman 2005b: 40). Consequently, we define a bi-regional strategy as the exploitation of region-bound FSAs in two regions and the focus on bi-regional decision-making, to achieve a considerable part of sales in each of the two regions. 20
Cf. Sect. 2.2.2.
40
2.2.5
2 Theoretical Foundation and Literature Review
Multi-regional (or Global) Strategy
A global MNC is present in all three triad regions with a global reach of its FSAs and a global scope (Rugman 2005b: 38). According to the data of Rugman and Verbeke (2004: 7), only slightly over 2% (or 9 firms) are global firms: “[. . .] having sales of 20% or more in each of the three parts of the triad, but less than 50% in any one region of the triad”. This 20% classification threshold corresponds to Ohmae’s (1985: 165) description of a triad power with a balanced presence and sales distribution across the triad (Osegowitsch and Sammartino 2008: 185; Rugman 2005b: 64; Rugman and Verbeke 2004: 7).21 A global strategy achieves the first condition of a triad power – equal penetration and exploitation capabilities in each triad market (Ohmae 1985: 165) – by deploying non-location bound FSAs in each of the three triad regions (Rugman 2005b: 39, 64). The second condition – the nonexistence of blind spots (Ohmae 1985: 165) – is realized by a global strategy through the global geographic scope of the MNCs’ operations (Delios and Beamish 2005: 22; Rugman 2005b: 39; Rugman and Verbeke 2004: 4). As both key financial and many corporate strategy decisions are taken at the parent headquarters (Rugman 2005b: 50), global strategies are placed in cell 1 of Fig. 2.5. Many MNCs with such global strategies can be encountered, for example, in the computer, office, and electronics industry – particularly as they allow the transport of standardized components across the world (Rugman 2005b: 173). As in this work we focus on five regions in all continents in the world, a global strategy across more than two regions can be tri-regional (three regions), quadregional (four regions) (Delios and Beamish 2005: 22), or global (five regions). Consequently, we subsume these geographical variations of a global strategy in the term multi-regional strategy, while both terms are used interchangeably in this work. Recognizing the arguments before regarding classification thresholds,22 we define this strategy as the exploitation of non-location-bound FSAs in more than two regions and the focus on central decision-making at the corporate headquarters, to achieve a considerable part of sales in more than two regions.
2.2.6
Further Regional Strategies
The cells 2, 6, 7, and 8 of Fig. 2.5 represent additional regional strategies to the home-, host-, bi-, and multi-regional (or global) strategies outlined above. Even though Rugman and Verbeke (Rugman 2005b: 50–52, 204–206; Rugman and Verbeke 2008b: 310–311) described them in their work, they did not devise any wording or particular names to characterize them. Therefore, based on their
21
Cf. Sect. 2.1.2.3. Cf. Sect. 2.2.2.
22
2.2 The Rugman Regional Strategy Matrix
41
description, we will proceed to develop a terminology for these regional MNC strategies.23 Cell 2 of Fig. 2.5 combines centralized decision-making at the corporate headquarters with substantial regional responsiveness regarding the product/service offerings of MNCs (Rugman 2005b: 204–205). Here, the centralization of decisionmaking should predominantly focus on host-regional, and only partly on home-regional, products/services. This is due to our assumption that most frequently, the corporate headquarters should be identical to the regional management center in the home region. One reasoning behind centralizing decision-making for host-regional products/services might be a physical separation of a MNC’s operations between the headquarters and a foreign host region (Rugman 2005b: 204). Another reason might be the need for a central coordination of strategic investment decisions, if substantial resources are allocated to a foreign host region (or within the home region) (Stopford and Wells 1972: 43). We term this headquarters-based coordination of host-(home-) regional product/service adaptations a directiveregional strategy. We define this strategy as the exploitation of host-(or home-) region-bound FSAs in a host region (or within the home region), and the focus on central decision-making at the corporate headquarters, to achieve the majority of sales in the host (or home) region. To achieve decision-making synergies across markets (Rugman 2005b: 205), in cell 6 of Fig. 2.5, decisions about national product/service offerings are made by a regional management center. These decision-making synergies are realized in particular functional areas of regional management centers (Rugman 2005b: 51), for example in R&D or marketing at the pan-European level. In our view, geographically this can take two forms. First, the home-regional management center might use functional decision-making synergies for national products/services of the home region. Second, a host-regional management center might install functional decision-making synergies for national products/services of its own host region. The utilization of region-based functional synergies for a national product/ service should account for its location-bound FSAs. Therefore, as a home-regional or a host-regional management center has to ensure an adaptation to national requirements, we term this an adaptive-regional strategy. We define this strategy as the exploitation of location-bound FSAs and the focus on regional decisionmaking in either the home region or a host region, to achieve the majority of sales in the respective region. In cell 7 of Fig. 2.5, global products/services are delivered by powerful national subsidiaries (Rugman 2005b: 51–52). Here, the subsidiary has worldwide responsibility for a global product/service, and autonomously evaluates its market potential in national economies (Birkinshaw and Morrison 1995: 734; Ghemawat 2005: 104–105; Ghemawat 2008: 153–154; Rugman 2005b: 206). Based on this world mandate, a subsidiary usually controls all aspects of R&D, production, sales, and
23
We were able to personally accord our terminology with both Rugman and Verbeke during the IB conference at the University of Reading in the UK, 30–31 March 2009.
42
2 Theoretical Foundation and Literature Review
distribution for a limited range of global products/services (Birkinshaw and Morrison 1995; Morrison and Roth 1992: 51; Poynter and Rugman 1982: 54; Rugman and Bennett 1982: 59; White and Poynter 1984: 61). This represents a typical approach of firms in their evolution toward Bartlett and Ghoshal’s (1989: 25, 65) transnational MNC (Rugman 2005b: 206; Rugman and Verbeke 2008b: 311). To emphasize that a world mandate strategy is beyond not only the national, but also the regional level, we utilize the term trans-regional (world mandate) strategy. As it combines cells 1 and 9 of Fig. 2.5 (Rugman 2005b: 51), we define this strategy as the exploitation of non-location-bound FSAs and the focus on national decisionmaking for the products/services of more than two regions, to achieve a considerable part of sales in more than two regions. In cell 8 of Fig. 2.5 national subsidiaries jointly standardize their product/service offering at the regional level, thus benefiting from strong similarities in the demand of national markets within the home region or a host region (Rugman 2005b: 205; Rugman and Verbeke 2008b: 310). To realize these economies of scale and scope, subsidiary initiative is critical (Birkinshaw 2000: 8; Morrison et al. 1991: 27; Morrison and Roth 1992: 51–52; Rugman 2005b: 205; Rugman and Verbeke 2001: 238; Rugman and Verbeke 2008b: 310).24 Subsidiary initiatives aim to identify opportunities to use or expand the MNC’s resources (Birkinshaw 1997: 207; Birkinshaw 1998: 356; Birkinshaw 2000: 20). Thus, in cell 8 of Fig. 2.5, regional integration benefits are realized by the cooperation of national subsidiaries and their initiatives to extend the geographic reach of FSAs to the regional level (Rugman 2005b: 222). We term this strategy a cooperative-regional strategy, which we define as the exploitation of region-bound FSAs in either the home region or a host region and the focus on national decision-making, to achieve the majority of sales in the respective region. The majority of the diverse regional strategies outlined above, particularly those in cells 2, 4, 5, 6, and 8 of Fig. 2.5 are not addressed by Bartlett and Ghoshal’s (1989: 65, 97) IR framework of Fig. 2.4 (Rugman 2005b: 52). However, considerable empirical evidence has been provided by Rugman and Verbeke (Rugman 2005b: 4; Rugman and Verbeke 2004: 7) for the dominance of such regional strategies of MNCs – particularly of home-regional strategies, which are situated in cell 5 of Fig. 2.5. Consequently, the regional strategy matrix represents an important framework both in conceptual and empirical terms for the strategic alternatives of MNCs. In the following, we will further explore the regional components of structure and strategy – the former represented by regional management on the horizontal axis, the latter by regional products/services on the vertical axis of Fig. 2.5 (Rugman 2005b: 202; Rugman and Verbeke 2008b: 310).
24
Possibilities for a MNC to encourage its national subsidiaries to act at this regional level, beyond their national market, include judging them on their contribution to regional growth and profitability (Stopford and Wells 1972: 60), or giving the most enthusiastic subsidiary a lead for certain regional product attributes or service elements (Bartlett and Ghoshal 1989: 149).
2.2 The Rugman Regional Strategy Matrix
2.2.7
43
Regional Management
Manifold reasons exist for installing a regional management as an additional decision-making layer in the MNC. Some argue that it reduces the complexity at corporate headquarters of managing hundreds of national subsidiaries to less than 10 regional management units (Arregle et al. 2009: 103). Therefore regional management, following a transaction cost economics argumentation (Williamson 1981, 1996), represents an efficient intra-firm regional governance mechanism for the bounded rationality and bounded reliability of MNCs (Rugman and Verbeke 2005: 9, 15; Rugman and Verbeke 2007: 201; Verbeke and Kenworthy 2008: 944). Bounded rationality reflects the limitations of an intended rational human behavior (Simon 1957: xxiv) to absorb, process and purposefully act upon complex and often insufficient information (Verbeke 2009: 9, 52; Verbeke and Kenworthy 2008: 954). As bounded rationality affects managers of MNCs in the exploitation of FSAs (Verbeke and Kenworthy 2008: 954), it forces corporate headquarters to delegate decisions to the regional level (Rugman 2005b: 50), where national and regional market-related information can be better captured (Buckley and Ghauri 2004: 87). Therefore bounded rationality gives reasons for a regional management presence, if MNCs set up locally-oriented activities in a distant region (Enright 2005b: 100), or if the number of distinct economies within a particular region increases (Enright 2005a: 75; Rugman and Verbeke 2005: 15).25 Bounded reliability describes limitations of individuals regarding their realization of a promised outcome (Rugman and Verbeke 2005: 9; Verbeke 2009: 9; Verbeke and Greidanus 2009: 1482; Verbeke and Kenworthy 2008: 954). Bounded reliability of national subsidiaries is influenced by their distance to the incentive systems of corporate headquarters – including immediate sanction and reward systems, as well as monitoring and controlling mechanisms (Verbeke and Kenworthy 2008: 954; Yeung et al. 2001: 179). Therefore bounded reliability explains how the importance of regional management might be influenced by the geographical proximity to corporate headquarters (Yeung et al. 2001: 179). In addition, MNCs need to create a favorable organizational context internally, and in conformity with their external environment, for deploying non-locationbound and location-bound knowledge and innovations (Verbeke and Kenworthy 2008: 944, 955). Regarding a MNC’s internal organization, a regional management ensures the exploitation of synergies and the consistency of policies across the region, and it may pool functional resources (Daniels 1987: 412–413; Lasserre 1996: 31; Yeung et al. 2001: 166). With respect to the external environment of a MNC, regional management signals commitment to the region (Birkinshaw et al. 2006: 697; Ohmae 1985: 111; Yeung et al. 2001: 179),26 and detects both new
25
An example of increases in the number of distinct economies within a particular region is given by the rising number of member states within the EU (Berger and Steger 1998: 26). 26 Cf. Sect. 2.1.2.3.
44
2 Theoretical Foundation and Literature Review
regional business opportunities and changes in the regional environment (Lasserre 1996: 31; Lasserre and Probert 1998: 49). Furthermore, regional management is an organizational response of MNCs – such as the transnational type of organization by Bartlett and Ghoshal (1989: 59–61, 65) – to balance the pressures to stay local and to globalize (Buckley and Ghauri 2004: 86–87; Daniels 1987: 417; Enright 2005a: 78; Sch€utte 1997: 440–441; Williams 1967: 91; Yeung et al. 2001: 179). These two conflicting pressures affect the firm’s functional activities along its value chain, as they impose the need for downstream and upstream FSAs on MNCs (Rugman 2005b: 197, 173; Rugman and Verbeke 2008b: 311). Downstream FSAs, or customer-end FSAs: “[. . .] refer to knowledge strengths deployed in activities with a direct interface with the customers, and are required to achieve successful market penetration” (Rugman 2005b: 197). These MNC capabilities are contrasted to upstream FSAs or back-end FSAs, which: “[. . .] are deployed in activities that lack this direct interface, but are critical to creating an efficient internal production system” (Rugman 2005b: 197). Regional management functions realize adaptation advantages by a regional responsiveness to downstream, intra-regional differences at the country level (Arregle et al. 2009: 89–90; Ghemawat 2008: 199; Sch€utte 1997: 441). Also at the customer-end, regional management may deploy aggregation advantages from capitalizing on similarities across markets and from exploiting synergies of regionally integrated national subsidiaries (Arregle et al. 2009: 89–90; Ghemawat 2003: 148; Rugman 2005b: 77). At the same time, regional management achieves backend arbitrage advantages by benefiting from absolute economies across borders (Ghemawat 2008: 169; Rugman 2005b: 70); for example by reallocating investments across national boundaries (Arregle et al. 2009: 91). In addition, arbitrage leads to dispersed assets in the MNC, which may act as sources of back-end economies of scale and scope in specialized regional management centers (Ghemawat 2008: 199; Rugman 2005b: 77–78; Rugman and Verbeke 2008b: 311). Consequently, regional management may realize revenue advantages from downstream FSAs by the regional adaptation of products/services, and cost advantages both from exploiting aggregation economies of downstream FSAs and from realizing arbitrage effects regarding upstream FSAs (Buckley and Ghauri 2004: 86–87; Rugman and Verbeke 2008b: 311). To optimally respond to the dual requirements of economic integration and national responsiveness, different functional MNC activities require varying degrees of hierarchical coordination and control (Rugman and Verbeke 1992: 769). The coordination and control of customer-end activities in the MNC is achieved mainly by participative decentralization, leading to a high involvement of regional management in corporate strategic planning (Freeland 1996: 484; Rugman 2005b: 74). In contrast, back-end activities are coordinated and controlled mainly by administrative centralization, leading to a high involvement of the corporate headquarters in regional management affairs (Freeland 1996: 484; Rugman 2005b: 74; Williamson 1975: 148). Regional management exists and persists only if it creates value for the MNC (Enright 2005a: 62), on the basis of downstream and/or upstream FSAs.
2.2 The Rugman Regional Strategy Matrix
45
The deployment of customer-end and back-end FSAs, however, is irrespective of a particular value chain function (Rugman 2005b: 70, 72). As a result, the geographical division of labor – given by the most important functional activities at the regional level – represents the appropriate unit of analysis for regional management (Buckley and Ghauri 2004: 94). A recent application of this approach – which has also been proposed by Ohmae (1985: 186) according to Fig. 2.3 – can be found in Enright’s (2005a: 65; 2005b: 89) empirical results on the functional activities in the Asia-Pacific region of 1,100 North American, European, and Japanese MNCs. In our analysis of Enright’s (2005a: 66–68; Enright 2005b: 100) work – by focusing on those functional activities which were perceived as most important by his respondents, and those which are realized at the level of regional headquarters – we could identify four categories of regional management activities: regional strategy development, regional market and product/service development, regional market coordination, and regional operational administration.27 As outlined above, these regional management functions of MNCs require different balances of customer-end and back-end FSAs, to achieve regional responsiveness and regional integration benefits. Given the bounded rationality and bounded reliability of the management at corporate headquarters, and the need to create a favorable organizational context in the region, a regional management is established, which is defined as functional activities at the regional level of a MNC in the areas of regional strategy development, regional market and product/ service development, regional market coordination, and regional operational administration, which aim to achieve both downstream and upstream FSAs.
2.2.8
Regional Products/Services
A product represents a bundle of technical and functional attributes that might satisfy a need or a utility of a customer (De Bu´rca et al. 2004: 258; Kotler et al. 2007: 12; Meffert et al. 2008: 398–399). A service is a form of a product consisting of intangible activities, benefits, or satisfactions that are offered for sale to a customer (Kotler and Armstrong 2004: 276; Meffert et al. 2008: 29; Onkvisit and Shaw 2009: 365). The “regional” character of a product/service depends upon the geographic reach of its embodied FSAs (Rugman 2005b: 203; Verbeke 2009: 14). More specifically, a regional product/service results either from augmenting the geographic reach of location-bound FSAs to the regional level, or from the regionally limited deployability of non-location-bound FSAs (Rugman 2005b: 200). This is schematically illustrated in Fig. 2.6, where various combinations of triangles – each representing a possible regional product/service configuration in a MNC – show different levels of the geographic reach of FSAs.
27
In Sect. 4.1.1, these regional management activities will be presented in detail.
46
Location advantages host country (influenced by CSAs in host country) Non-transferable (or location-bound) FSAs host country Internationally transferable (or non-location-bound) FSAs
2 Theoretical Foundation and Literature Review Home region Host country 1 Host country 2
Host region Host country 3
Intra-regional border Internationally transferable (or non-location-bound) FSAs Non-transferable (or location-bound) FSAs home country Location advantages home country (influenced by CSAs in home country)
Home country Home-regional product /service 1 Home-regional product/service 2 Bi-regional product/service
Fig. 2.6 Geographic reach of the FSAs of regional products/services Source: own illustration on the basis of Verbeke (2009: 34)
In the case of the home-regional product/service 1, a MNC could augment its location-bound FSAs to non-location-bound FSAs – as indicated by the lower, grey shaded arrow in Fig. 2.6. These non-location-bound FSAs are then transferable to the foreign host country 1. There, the transferred (non-location-bound) FSAs of the home country are complemented by the development of new location-bound FSAs in host country 1, to access and benefit from the location advantages of host country 1 (Verbeke 2009: 34). These location advantages of the MNC in host country 1 are influenced by its CSAs in this host country – such as a skilled workforce, or good supplier networks (Rugman 2005b: 44). The home-regional product/service 1 is therefore characterized by region-bound FSAs (being located in the home country and in host country 1). In the case of the home-regional product/service 2 in Fig. 2.6 – as non-location-bound FSAs cannot be transferred beyond the home region – the same procedure leads again to region-bound FSAs in the home region (being situated in the home country, in host country 1, and in host country 2). In the case of a bi-regional product/service in Fig. 2.6, non-location-bound FSAs were successfully transferred to host country 3 in a foreign host region. Here again, to exploit the location advantages of host country 3, the MNC develops new locationbound FSAs in host country 3 as a complement of its (non-location-bound) FSAs of the home country. The resulting bi-regional product/service – even though its FSAs are region-bound within the home region and the host region – has extended the
2.2 The Rugman Regional Strategy Matrix
47
deployability of non-location-bound FSAs beyond one particular region. This increased geographic reach of FSAs corresponds to the characteristics of a global product/service, as illustrated on the vertical axis of Fig. 2.5. Over time – in the case of all three regional products/services presented before – MNCs might also develop new, non-location-bound FSAs in their respective host countries 1, 2, and/or 3 (Rugman 2005b: 229). Depending on the geographic reach of these new (nonlocation-bound) FSAs, MNCs can expand the relevant market space of their products/services within a particular region, or towards a more global level. In other words, the transformation of a regional to a more global product/service is realized by changing FSAs, rather than by expanding geographic scope (Rugman 2005b: 226).28 In all of the cases described above, the regional geographic reach of FSAs is a consequence of the adaptation of MNCs to national or regional market requirements (Rugman 2005b: 203, 229). The regional adaptation of products/services can take different forms, as illustrated in the overview in Table 2.2. First, products/services could be adapted vertically (e.g., by changing their objective quality) or horizontally (e.g., by changing different product features) to regional requirements (Beath and Katsoulacos 1991: 4–5; Onkvisit and Shaw 2009: 338). Second, the attributes of products could be regionally adapted according to either esthetic (e.g., colors), physical (e.g., norms and standards), functional (e.g., specifications), or symbolic (e.g., design, packaging) dimensions (Emrich 2007: 219–220). Third, services could be regionally adapted along physical good quality (e.g., diversity of services for a product offering), service quality (e.g., reliability), and/or servicescape (e.g., atmosphere) (Keillor et al. 2004: 10–12). Fourth, different degrees of regional product adaptation can be differentiated, ranging from low adaptations for universal products (e.g., language in user manuals), medium adaptations in the case of modified products (e.g., size), to high adaptations for countryspecific products (e.g., taste) (Takeuchi and Porter 1986: 133–134; Takeuchi and Porter 1989: 151–152). A combination of the different forms of regional product/service adaptations with the reasons for regional product/service adaptations (De Bu´rca et al. 2004: 267–270; Onkvisit and Shaw 2009: 349–359) – as illustrated by Fig. 2.7 – leads to four broad categories: design, functional scope, brands, and offering portfolio. These categories of regional product/service adaptations in Fig. 2.7 illustrate the core customer benefits and characteristics of an actual product/service offering – including features, design, packaging, quality level and brand name (Kotler and Armstrong 2004: 279) – which corresponds to the vertical axis of Fig. 2.5 (Rugman 2005b: 48). Their application by academic scholars to describe regional and crossborder product/service adaptations – along design (Albaum and Tse 2001: 71; Calantone et al. 2004: 186; Cavusgil et al. 1993: 487–488; Melewar and Saunders 1999: 584; 28
As these explanations aim to achieve a schematic understanding of the geographic reach of the FSAs of regional products/services, a MNC’s host-regional, multi-regional (or global), and further regional strategies are not explored in more detail. However, it should be noted, that here a likewise argumentation would apply.
Source: own illustration
Table 2.2 Forms of regional adaptation of products/services Adaptation of regional product Direction of regional product/ attributes (Emrich 2007) service adaptation (Onkvisit and Shaw 2009) Vertical adaptation Esthetic product attributes – Objective quality of – Different colors – Proportions of products (e.g., North product/service – For example: diamond American vs. Asian Barbie doll) necklace Physical product attributes instead of gemstone – Norms and standards (e.g., DIN, ISO) Horizontal adaptation Functional product attributes – Different product features – For example: strawberry – Specifications (e.g., size of beds) or chocolate ice cream Symbolic product attributes – Design, packaging – Quality and diversity of a product offering (e.g., fastfood or grocery services) Service quality – For example: attention, reliability Servicescape – Atmosphere – For example: ambient, social environment, design
Physical good quality
Adaptation of regional services (Keillor et al. 2004)
Degrees of regional product adaptation (Takeuchi and Porter 1986, 1989) Universal products – Low adaptation (e.g., language in user manuals) – Physically the same (e.g., steel, plastics, chemicals) Modified products – Medium adaptation (e.g., color, size, packaging accessories) – Changes to important attributes (e.g., cars, microwaves) Country-specific products – High adaptation (e.g., taste) – Consideration of countryspecific requirements (e.g., many food products)
48 2 Theoretical Foundation and Literature Review
2.2 The Rugman Regional Strategy Matrix (Onkvisit and Shaw 2009; Emrich 2007; Keillor, Hult and Kandemir 2004; Takeuchi and Porter 1986, 1989)
49
Categories of regional product/service adaptation
(Onkvisit and Shaw 2009; Búrca, Fletcher and Brown 2004)
• Own language and symbolism • Own style and design preferences
• Esthetic product attributes • Symbolic product attributes Design
• Local customs and culture • Taste requirements
• Horizontal adaptation
• • • •
Functional product attributes Physical product attributes Physical good quality Service quality
Functional scope
• Servicescape Brands
Environmental conditions Regulations and standards Physical appearance of society Room and space restrictions Regional usage conditions
• Performance standards • Usage habits • Income level
• Vertical adaptation • Universal products • Modified products • Country-specific products
• • • • •
Offering portfolio
• Level of innovation • Historically developed preferences
Fig. 2.7 Categories of regional product/service adaptation Source: own illustration
Rugman 2005b: 146, 161–162, 192, 205; Samiee 1994: 580), functional scope (Grosse 2005: 132; Kogut 1991: 42; Kogut 2002: 78), brands (De Bu´rca et al. 2004: 265; Kotler and Armstrong 2004: 285; Onkvisit and Shaw 2009: 390; Rugman 2005b: 34, 36, 141, 152, 162–163, 179, 188, 191, 225–226; Rugman and Collinson 2004: 475; Rugman and Verbeke 2004: 6; Schuiling and Kapferer 2004: 108), and offering portfolio (Ghoshal 1987: 426; Rugman 2000: 93; Rugman 2005b: 131, 134, 162, 204; Rugman and Verbeke 2008b: 310) – confirms the appropriateness of these categories. All four categories of regional product/service adaptations involve intra-regional resource commitments in the form of investments in foreign markets and subsidiaries (Rugman 2005b: 229; Ward 1973: 79–80). As argued elsewhere by Rugman and Verbeke (1992: 763), Rugman and Verbeke (2001: 241), Rugman and Verbeke (2003a: 128–129), Rugman and Verbeke (2004: 13), Rugman and Verbeke (2007: 201) and Rugman and Verbeke (2008b: 305–306), this particularly includes: “[. . .] (a) investments in the development of location-bound FSAs in foreign markets (leading to benefits of national responsiveness) to complement non-location-bound FSAs, and (b) investments in the development of new, non-location-bound FSAs in foreign subsidiaries” (Rugman 2005b: 229). The result of these region-specific adaptation investments is a linkage of the MNC’s existing (non-location-bound and location-bound) FSAs with its CSAs in the form of regional/locational advantages (Rugman 2005b: 221; Rugman and Verbeke 1992: 763; Rugman and Verbeke 2001: 241; Rugman and Verbeke 2003a: 127; Rugman and Verbeke 2004: 6; Rugman and Verbeke 2008b: 305–306). If these location-/region-specific investments are directed towards
50
2 Theoretical Foundation and Literature Review
downstream FSAs, to attract potential customers in foreign markets, MNCs face high risks related to two aspects (Rugman 2005b: 229–230). First, these marketoriented investments are fully one-sided from the part of the MNC, which is compounded by the second aspect, that no customer guarantees to purchase its products/services exist (Rugman 2005b: 229–230). In contrast, investments in upstream FSAs are not completely one-sided, as here all relevant parties (e.g., workers, foreign suppliers, and logistics providers) are themselves engaged in reciprocal commitments to generate value for the jointly established manufacturing or logistics chain apparatus (Rugman 2005b: 229–230). Given that production and delivery cannot be separated in service firms, service MNCs face high location-/ region-specific adaptation investments in both upstream and downstream FSAs to achieve regional responsiveness (Rugman 2005b: 233). Recognizing the explanations above, a regional product/service of a MNC is the result of the adaptation of downstream and upstream FSAs to regional requirements and conditions in the areas of design, functional scope, offering portfolio, and brands. This adaptation of its regional products/services and the previously described, functional activities of its regional management represent changes in a MNC’s regional strategy and structure, by which it aims to address requirements imposed by its regional environment (Rugman and Verbeke 2008b: 308). Therefore, the MNC as an organization faces opportunities and constraints in dependence with this external, as well as with its internal, organizational context. In the following we present an important theoretical framework for the conceptualization of these contextual factors that MNCs face at the regional level.
2.3 2.3.1
Contingency Approach Contingency Theory
Contingency theory aims to develop empirically founded statements about the relationships between external situational conditions, internal organizational structures, the behavior of organizational members, and the achievement of organizational objectives (Gerdin and Greve 2004: 307; Kreikebaum et al. 2002: 22). The theory was developed mainly in the 1960s (Burns and Stalker 1961; Lawrence and Lorsch 1967; Thompson 1967; Woodward 1965) and in the course of the “Aston Programme” that was guided by Derek S. Pugh (Pugh and Hickson 1976) at the University of Aston in Birmingham (UK) (Doz and Prahalad 1991: 151; Kreikebaum et al. 2002: 22). At a more abstract level, the contingency approach suggests that the relationship between an independent variable and a dependent outcome depends, or is contingent, upon a third variable (Donaldson 2001: 5; Drazin and Van de Ven 1985: 514; Gerdin and Greve 2004: 307). In other words: “[. . .] the effect of X on Y when W is low differs from the effect of X on Y when W
2.3 Contingency Approach
51
is high” (Donaldson 2001: 5–6). In the context of MNCs, the contingency approach – also referred to as the situational approach (Kieser 2006: 215; Kieser and Kubicek 1992: 45–46) – declares organizations as effective if their management realizes a structure and strategy which correspond to the requirements of their environmental and internal situation (Ebers 1981: 6–7). Thus, the organizational effectiveness of a MNC is explained as a function of the match, or fit, of its strategy and structure with its external and internal contingencies (Sillince 2005: 608; Venkatraman and Camillus 1984: 513–514; Zeithaml et al. 1988: 44). These external and internal situational characteristics of the MNC are termed contextual variables or contingency variables (Sousa and Voss 2008: 703). Examples of internal contingency variables include products, organizational size, operational technology, the age and development stage of the organization, and the degrees of diversification and internationalization of MNCs (Kreikebaum et al. 2002: 24–25; Roth and Morrison 1992: 484; Schrey€ ogg 2008: 331; Whittington 2006: 814). External contingency variables include dynamics and uncertainties in the environment, the intensity of competition, customer demand, a MNC’s customer and supplier structures, and its country culture (Ginsberg and Venkatraman 1985: 427; Hambrick and Lei 1985: 765; Hoque 2004: 488–489; Kreikebaum et al. 2002: 25; Lindsay and Rue 1980: 393; Whittington 2006: 814; Zheng Zhou et al. 2007: 308). For all possible constellations of these variables, neither a universally optimal organizational structure (Kieser 2006: 215; Zeithaml et al. 1988: 39), nor a universally superior strategy exist (Venkatraman 1989b: 424). Instead, each contextual constellation corresponds to only one efficient organizational structure, which constitutes an organizational equilibrium of the MNC (Kreikebaum et al. 2002: 23). In the case of an organizational disequilibrium, contingency theory assumes a gradual, steady, and continuous adaptation of the organization to its context (Kreikebaum et al. 2002: 23; Meyer et al. 1993: 1177–1178). In contrast to the configuration approach – which assumes that firms are either static or that their organizational change follows quantum jumps between discrete sets of ideal fit types (Doty et al. 1993: 1202; Mintzberg et al. 2007: 352–353, 387) – contingency theory describes a continuous development of MNC organizations over time, along typically linear variable relationships (Donaldson 2001: 122; Schoonhoven 1981: 353). Along those organizational continuums, the contingency approach allows for generalizations of the appropriateness of different strategies to internal or external contextual settings, which is why it is regarded as being well-suited for the research on MNC strategies (Doz and Prahalad 1991: 151; Hambrick and Lei 1985: 765; Hofer 1975: 807). The formerly described IB theories of Stopford and Wells (1972) and Bartlett and Ghoshal (1989) also draw on the contingency approach (Donaldson 2001: 81–83; Doz and Prahalad 1991: 151; Whittington 2006: 814). While Stopford and Wells (1972: 6) describe organizational change as a function of varying international sales and/or product diversity (Donaldson 2001: 81–82), Bartlett and Ghoshal (2002: 7) view the strategies and organizational designs of MNCs as contingent upon the contextual effects from diverse economic, political, social and technological forces that influence their need for integration and responsiveness (Donaldson
52
2 Theoretical Foundation and Literature Review
2001: 82–83). Ohmae (1985: 77) also describes successful insiderization strategies as being contingent on the establishment of linkages and relationships with the environment. This demonstrates the commensurability of the theoretical approaches applied in this work, and is reinforced by Rugman’s (2005b: 49) regional strategy matrix in Fig. 2.5 – as outlined earlier in Sect. 2.2.1 – which assumes that the simultaneous achievement of regional integration and regional responsiveness by a MNC is contingent upon its particular regional environmentstrategy-structure relationship. This relationship is affected by the MNC’s geographic reach of location-/ region-bound and non-location-bound FSAs, and by its CSAs, which are particularly influenced by two contingency variables: regional orientation and inter-regional distance. After the following critical review of the contingency approach, we will present them in more detail. A critical assessment of contingency theory is comprised of four facets along methodological and conceptual issues. The first criticism regards its limited explanatory power – as empirical results are only meaningful if the variables under research are independent of each other, which rejects the more realistic interdependencies between the influencing variables of an organization’s structure and efficiency (Kreikebaum et al. 2002: 25). Second, the contingency approach has been criticized for the lack of clarity and precision in its theoretical concepts, interactions, and underlying assumptions (Kieser and Kubicek 1992: 220, 412; Schoonhoven 1981: 349–354; Schrey€ ogg 2008: 355). Third, criticisms are directed towards the missing dynamics in its variable relationships and the encouragement of dichotomous thinking (Doz and Prahalad 1991: 151; Kieser 2006: 233–237; Kreikebaum et al. 2002: 25). Fourth, a generalization of the statements of contingency theory is not possible due to the shortcomings in the comparability, validity, and reliability of its applied measures (Kieser 2006: 231–232; Kreikebaum et al. 2002: 25). Recognizing these problems, we agree with Schoonhoven (1981: 349), that the application of contingency theory requires precise hypotheses on the expected empirical relationships, which this work aims to achieve. Besides this, we do not assume a strong determinism between an organization’s context and its structure and/or strategy (Kreikebaum et al. 2002: 25). Rather we take a moderate voluntaristic perspective for a MNC’s achievement of appropriate degrees of managerial decision-making autonomy and regional product/service adaptation (Gilbert 2005: 410–411; Kreikebaum et al. 2002: 171).29 This application of contingency theory highly recognizes that there is neither one universal “best way” for realizing regional strategies, nor that their achievement is solely situation-specific (Hambrick and Lei 1985: 765; Zheng Zhou et al. 2007: 305).
29
According to this perspective, the external and internal organizational context influence the managerial decision-making autonomy and regional product/service adaptation of MNCs – yet, at the same time, a firm can intervene in the processes related to the realization of its regional strategy – thus the relationship between regional strategy and its result can be described as emergent (Gilbert 2005: 411; von Hayek 1969).
2.3 Contingency Approach
2.3.2
Contingency Variables
2.3.2.1
Regional Orientation
53
The regional orientation is a form of geographic orientation of MNCs, in contrast to a local or global orientation (Banalieva and Santoro 2009: 344–345). It relates to the focus of senior management in the geographic expansion of MNCs (Delios and Beamish 2005: 20; Rugman 2005b: 202). It reflects a firm-internal, regiocentric orientation of MNC managers, who mainly view the region as a potential market (Proff 2002: 242, 245; Wind et al. 1973: 15),30 resulting in a narrow geographic market focus (Rugman 2005b: 212, 231; Rugman and Verbeke 2008b: 312) and a selectivity in the geographic scope of activities (Rugman 2005b: 77, 212; Rugman and Verbeke 2004: 16). The distribution of sales and assets of MNCs among economic regions indicates the type of a market-related, or respectively, an activityrelated regional orientation, each ranging from home-regional, host-regional, bi-regional, to multi-regional (global) (Goerzen and Asmussen 2007: 66–67; Oh 2009: 342; Rugman 2005b: 4; Rugman and Verbeke 2004: 5, 7). Therefore a MNC’s high orientation towards, or focus on, a particular region results from a low geographic distribution of its sales and assets. Consequently, the strongest form of regional orientation is a focus on the home region, indicating the lowest possible geographic dispersion of markets and activities. Thus, a MNC’s regional orientation decreases from home-regional to host-regional, from host-regional to bi-regional, and from bi-regional to multi-regional (global) – illustrating different extents of its regional multinationality (Contractor 2007b: 16; Verbeke et al. 2009: 150). The regional orientation of MNCs following further regional strategies – i.e., directiveregional, adaptive-regional, cooperative-regional, and trans-regional (world mandate) strategies – also depends on their particular geographic focus. These varying degrees of regional orientation are based on the geographic spread of markets and activities, which depends upon a MNC’s geographic reach and scope. We believe that the regional orientation of a MNC is an important internal contingency variable in the relationship between a MNC’s regional strategy and/or structure components and the achievement of its corporate goals and objectives. This is supported by Rugman (2005b: 212) who states that a strong regional orientation, given by a narrow geographic market and product/service focus: “[. . .] will help to develop the FSAs capable of providing maximum value added to customers”. For example, in the case of a home-regional orientation, the capitalization on such customer-end opportunities relates to an organizational and strategic efficiency, as no additional, one-sided, location-specific investments in other regions are needed (Rugman 2005b: 231). Therefore a MNC’s regional orientation – as the selection of a particular geographic focus – should be an important internal 30
This may relate to attitudinal characteristics of MNCs (Sullivan 1994: 331), regarding their: “[. . .] orientation toward ‘foreign people, ideas, resources’, in headquarters and subsidiaries, and in host and home environments [. . .]” (Perlmutter 1969: 11).
54
2 Theoretical Foundation and Literature Review
contextual variable for MNCs that influences the probability of achieving their corporate goals and objectives (Rugman and Verbeke 2008b: 312).
2.3.2.2
Inter-regional Distance
MNCs doing business in new markets face costs and risks from barriers created by distance (Ghemawat 2001: 138). This distance can either be intra-regional within a particular region (Verbeke 2009: 453), or inter-regional between two or more regions. According to Rugman (2005b: 77), MNCs have to consider the costs and risks of inter-regional distance in the economic evaluation of their international growth plans. Inter-regional distance is comprised of cultural, administrative/ political/institutional, geographic, and economic dimensions (Ghemawat 2001: 140; Ghemawat 2008: 34; Ricart et al. 2004: 181; Rugman 2005b: 230), as illustrated in the so-called “CAGE framework” (Ghemawat 2001: 140; Ghemawat 2008: 41; Ricart et al. 2004: 181) in Table 2.3.31 Inter-regional cultural distance refers to differences in the cultural attributes between the home country/region and host regions, which determine the interactions among people (Ghemawat 2001: 140; Ghemawat 2008: 40; Rugman 2005b: 223).32 These inter-regional cultural differences can sometimes be easily perceived (e.g., language), but they can also be nearly invisible (e.g., social norms) (Ghemawat 2008: 42). From one region to another, they cause differing, culturespecific demands on MNCs (Gilbert 1998: 66), such as color preferences of consumers in their choices between substitute products/services (Ghemawat 2001: 142). Inter-regional administrative, political, or institutional distance relates to differences in laws, policies, and institutions across regions, which affect crossborder economic activity (Ghemawat 2008: 42–43). These inter-regional differences may result, for example, from a lack of colonial ties between the home country/region and host region, from different currencies, or from not belonging to the same regional trading bloc (e.g., EU versus NAFTA membership) (Ghemawat 2001: 142–143; Ghemawat 2008: 43). Furthermore, the characteristics of a particular region (e.g., governmental policies, weak institutional infrastructure conditions) might impede MNCs from entering this region during their cross-border expansion (Ghemawat 2001: 144). 31
Inter-regional distance refers to bilateral differences between two particular regions, as well as to unilateral/multilateral differences between the home region and two or more other regions (Ghemawat 2008: 39). 32 Following Rugman (2005b: 223), we mainly focus on cultural differences at the regional level. This does not include the different levels of corporate culture differences within a particular MNC (Gilbert 1998: 67; Schein 1995: 30). Instead this subsumes country culture influences (Hofstede 1983: 75; Hofstede 1993: 22–23, 206–207) and more aggregate, regional cultural effects. However, it should be noted, that this represents a conceptual simplification, as a considerable cultural heterogeneity may prevail within both regions and countries (Gilbert 1998: 67; Hofstede 1997: 7–11; Macharzina et al. 1998: 160).
disease environments
Differences in time zones Differences in climates and
communication links
Differences in consumer
Physical distance, remoteness Lack of a common (land) border Lack of sea or river access Size of country/region Weak transportation or
incomes (rich-poor differences) Other differences in costs and quality of – Natural resources – Financial resources – Human resources – Infrastructure – Intermediate inputs – Information or knowledge
Economic distance
Geographic distance
Source: own illustration on the basis of Ghemawat (2001: 140), Ghemawat (2008: 40), and Ricart et al. (2004: 181)
Table 2.3 Dimensions of inter-regional distance Inter-regional distance Cultural distance Administrative/political/ institutional distance Different languages Lack of colonial ties Different ethnicities; lack of Lack of shared regional trading connective ethnic or social bloc (monetary or political networks association) Different religions Lack of common currency Level of trust Political hostility Different values, social norms, Government policies and dispositions Institutional weakness
2.3 Contingency Approach 55
56
2 Theoretical Foundation and Literature Review
In addition to the physical distance between regions, inter-regional geographic distance includes the presence or absence of a common (land) border, access to waterways and the ocean, size of the country, weak transportation or communication links, and different time zones and climates (Ghemawat 2001: 144–145; Ghemawat 2008: 44). These various aspects of geographic distance influence transportation and communication costs of MNCs (Ghemawat 2008: 44), which may explain reductions in inward FDI of a particular region (Feils and Rahman 2008: 160). Inter-regional economic distance results from differences in the wealth or income of consumers, and from any other variations in the cost and quality of resources, inputs, infrastructure and information (Ghemawat 2001: 145; Ghemawat 2008: 45). Macroeconomic effects from differences in consumer incomes can be observed in a higher cross-border economic activity of regions with many rich countries than of those regions with more poor countries (Ghemawat 2001: 145). At the microeconomic level, regional differences in per-capita incomes affect the labor costs of a MNC, while regional differences in skill levels may influence the quality of its products/services (Ghemawat 2008: 45). The inter-regional distance faced by a particular MNC may result in a competitive disadvantage from its liability of foreignness in the form of additional costs of doing business abroad (Ellis 2007: 377; Ghemawat 2008: 56; Hymer 1976: 34–36; Kindleberger 1969: 11–14; Rugman 2005b: 229; Rugman and Verbeke 2007: 201; Rugman and Verbeke 2008c: 331; Zaheer 1995: 342–343; Zaheer 2002: 351–352). Inter-regional distance, therefore, is an important external contingency variable in the achievement of a MNC’s corporate goals and objectives – particularly regarding adaptation, aggregation, and arbitrage benefits from regional strategy-structure combinations (Arregle et al. 2009: 88–92; Buckley and Ghauri 2004: 87; Rugman 2005b: 76). For example, the adaptation of a MNC’s products/services to a foreign host region requires higher investments to link its existing FSAs with foreign CSAs, if the inter-regional distance is high (Ghemawat 2008: 110–113; Rugman 2005b: 225). Aggregation economies of scale or scope are also contingent upon interregional distance, as they are easier to achieve, for example if the MNC can benefit from a higher economic integration or institutional proximity between regions (Arregle et al. 2009: 89; Ghemawat 2008: 158–159). A lower inter-regional distance facilitates earning arbitrage benefits for MNCs, as they may, for example, more flexibly relocate their foreign investments to exploit differences across regions (Arregle et al. 2009: 91; Ghemawat 2008: 174–180; Rugman 2005b: 76; Rugman and Verbeke 2004: 16). Furthermore, inter-regional cultural, economic and physical differences require regional decision-making by MNCs, to be responsive to their customers (Enright 2005b: 84; Yeung et al. 2001: 169). Consequently, the organizational and strategic effectiveness and overall competitiveness of MNCs is highly externally contingent upon the inter-regional distance they face (Ghemawat 2001: 147; Ghemawat 2008: 228; McLarney and Dastrala 2001: 349, 366; Rugman and Verbeke 2004: 5). Organizational and strategic effectiveness are usually measured by the corporate success of MNCs, for example by their return on investment, profit, or other
2.4 Concept of Regional Corporate Success
57
performance variables (Donaldson 2001: 10; Venkatraman and Ramanujam 1986: 803; Zeithaml et al. 1988: 49). A MNC’s effectiveness at the regional level depends on different factors – like its particular regional product/service offering, regional management configuration, and contingency variables – and is measured by its regional corporate success that we will present next.
2.4 2.4.1
Concept of Regional Corporate Success Success Factor Theory
Success factor research aims to identify the determinants that influence the longterm success of a MNC, where those measures that quantify the long-term success of a MNC are called success indicators – while those variables influencing these success indicators are termed success factors (Fritz 2004b: 24–25; Haenecke and Forsmann 2006: 45). The origins of success factor research date back to 1961, when the McKinsey consultant D. Ronald Daniel advocated a concept of success factors to improve internal management information systems (Daniel 1961: 116; Nicolai and Kieser 2002: 580). Around this time, the PIMS (Profit Impact of Marketing Strategies) program was developed, which systematically gathers firm data since the 1960s of more than 300 companies with approximately 3,000 strategic business units (SBUs) (Haenecke and Forsmann 2006: 45; Kreikebaum 1997: 113–117). While some of the following approaches such as Rockart’s (1979: 86) concept of critical success factors tried to identify such factors on the basis of extensive, case-by-case interviews of particular companies (Trommsdorff 1990: 2–3), other researchers tried to derive success factors for whole industries or even across all types of companies (Nicolai and Kieser 2002: 580). Since its very beginning in the 1960s, success factor research has been criticized either for not being relevant enough to practitioners, or for not being sufficiently rigorous according to theoretical quality requirements (Gulati 2007: 776; Kieser and Nicolai 2003: 592).33 For success factor studies, this persisting trade-off between rigor and relevance (Kieser and Nicolai 2003: 591; Kieser and Nicolai 2005: 275) causes the: “[. . .] problem of bridging the gap between the social system that produces scientific knowledge [. . .] and the social system where professional practice takes place [. . .]” (Simon 1967: 16). Some argue, based on Luhmann (1986, 1998, 2005), that this gap will remain, as science and practice are selfreferential (autopoietic) social systems that face major differences between each other (Kieser and Leiner 2009: 518–519; Kieser and Nicolai 2003: 591; Kieser and Nicolai 2005: 276; Nicolai and Kieser 2002: 592). Others argue that 33
From the 1960s onwards, the trend was first towards more relevance, while since the 1980s a development towards more rigor was observed (Gulati 2007: 776).
58
2 Theoretical Foundation and Literature Review
Table 2.4 Requirements of “good” success factor research “High” rigor “High” relevance Application of adequate theories Research based on practical issues Theory development based on different perspectives Translation function of researchers Theoretically grounded, causal relationships Consideration of the organizational context Methodological rigor Representative research sample Source: own illustration
this rigor-relevance gap can be bridged by theoretically grounded, problem-oriented research (Bettis 1991: 317; Gulati 2007: 775; Luther 1998: 704; Schendel 1995: 1), thus by: “[. . .] work that is rigorous both theoretically and methodologically and centered on issues of focal concern to a wide community of stakeholders (e.g. managers, government policy makers, trades unionists, and consumer groups) [. . .]” (Hodgkinson et al. 2001: S46). In this work, we accept the existence of a gap between theory and practice (Kieser and Nicolai 2003: 592), in the sense that insights from success factor studies cannot be used instrumentally (Ghemawat 1991: 11; Kieser and Nicolai 2005: 278; Nicolai and Kieser 2002: 584), like an immediate prescription or advice for exercising certain managerial actions that will increase corporate profits by X percent (Vermeulen 2007: 754). However, we agree with Fincham and Clark (2009: 514) that management studies as a social science is less of a closed system than natural science, for example. Therefore we assume that crossing this gap is possible, particularly by a certain level of practical relevance (Kieser 2004: 3), one that enables critical reflections on current practices (Kieser and Leiner 2009: 528).34 Such an approach offers alternatives to real-world issues, it questions preferred argumentation patterns of practice and points to relationships that have been neglected in practice so far (Kieser and Nicolai 2003: 592; Kieser and Nicolai 2005: 278; Rasche and Behnam 2009: 252). The previous arguments indicate that success factor theory has a small core (Bauer and Sauer 2004: 623); one that refers to “good” success factor research, a combination of “high” rigor – concerning the specification and the theoretical foundation of research models, data inquiry and data analysis (Homburg and Krohmer 2004: 630) – and “high” relevance for the reflection of real-world issues. Therefore in the following and as illustrated in Table 2.4, we formulate requirements for our work of “good” success factor research, based on the recommendations of academic scholars and researchers (Markides 2007: 766) related to both “high” rigor and “high” relevance. With respect to “high” rigor, Ghoshal (2005: 86) argues that: “[. . .] nothing is as dangerous as a bad theory”, meaning that a theory can only reveal true insight if it 34
This is in line with the argumentation of Markides (2007: 766) who claims for a higher differentiation of the different types of managerially relevant research, where one might be the development of grand new theories without any empirical evidence, and another the development of new theoretical rationales for current business phenomena.
2.4 Concept of Regional Corporate Success
59
has reality in mind (Tushman and O’Reilly 2007: 769, 772; Vermeulen 2007: 755). Examples of such adequate theories to explain real-world issues include Porter’s (1985) influential work on strategy or Hofstede’s (1983) study of cultural influences on values in the workplace (Starkey et al. 2009: 553).35 Therefore theory development should be based on different perspectives – e.g., implying a regular, intensive dialogue and interaction with practitioners (Markides 2007: 762; Rynes 2007: 745; Starkey et al. 2009: 551; Tushman and O’Reilly 2007: 769; Vermeulen 2007: 757; Worrell 2009: 129), to get to the heart of practical matters (Gulati 2007: 780; Rynes 2007: 746). Even though researchers should trade some rigor for more relevance (Gulati 2007: 777; Starkey et al. 2009: 550), the ideas of researchers about causal relationships should still be grounded in existing or new theory (Gulati 2007: 780) and/or should integrate different theories (McGahan 2007: 750).36 Furthermore, methodological rigor should ensure the proper operationalization of variables and the utilization of appropriate statistical methods (Fritz 2004a: 625; Nicolai and Kieser 2002: 584).37 Methodologically sound research should properly address the key informant bias (Fritz 2004b: 26–30; Haenecke and Forsmann 2006: 50; Hurrle and Kieser 2005: 598; Kieser 2004: 3–9; Nicolai and Kieser 2002: 584; Nicolai and Kieser 2004: 633), as well as problems from endogeneity (influences of non-collected variables on independent variables), simultaneity (effects from prior success on managerial actions), unobserved heterogeneity (differences in the long-term success potential of firms), regression-to-the-mean (effects from randomly occurring successes and failures of firms), and the survival bias (research focusing solely on firms that have survived) (Nicolai and Kieser 2002: 584–585). “High” relevance relates to generating insights for practitioners to better understand their own organizations and situations (Markides 2007: 764–765; Vermeulen 2007: 755). Therefore, practitioners are viewed as stakeholders, the central audience of research, to realize a research based on practical issues that addresses important managerially relevant topics and problems (Bennis and O’Toole 2005: 103; Markides 2007: 763–764). Here, researchers have a translation function in the communication between science and practice (Gulati 2007: 780–781; Hodgkinson and Rousseau 2009: 543; McGahan 2007: 748; Vermeulen 2007: 759). More specifically, they should establish a common frame of reference (Kieser and Nicolai 2005: 276) that ensures speaking the same language, and that communicates
35
For further examples cf. Hodgkinson and Rousseau (2009: 539–541). Here it should be noted that – as researchers can only derive knowledge from falsifying hypotheses about causal relationships, instead of measuring them directly – success factor research can only shed light on possible reasons of success (Haenecke and Forsmann 2006: 48). 37 In the operationalization of variables, no average values across firms should be utilized, as success is firm-specific and unique (Nicolai and Kieser 2002: 585–586). Furthermore it should be considered that one success indicator cannot adequately address the diversity of the real-world success (Nicolai and Kieser 2002: 586–587), and that success factor research is always based on past data, which limits the transferability of its findings to the future (Nicolai and Kieser 2002: 587). 36
60
2 Theoretical Foundation and Literature Review
meaning by the transfer of schemas between the two contexts (Hodgkinson and Rousseau 2009: 537–538; Kieser and Leiner 2009: 528; Starkey et al. 2009: 556).38 In addition, not necessarily by introducing many control variables (Kieser and Nicolai 2005: 276–277; Nicolai and Kieser 2002: 582),39 the organizational context should be considered (Eggert et al. 2005: 103; Starkey et al. 2009: 556), as more informed theory building can be achieved by research that is well-grounded in organizational contexts (Hodgkinson and Rousseau 2009: 542). To achieve relevance, some claim for large-scale empirical analyses (McGahan 2007: 749), while others question the usefulness of existing big research projects and meta-analyses (Fritz 2004a: 624; Homburg and Krohmer 2004: 627; Nicolai and Kieser 2002: 582–584; Nicolai and Kieser 2004: 632). Therefore, rather than solely size, the quality reflected in a representative research sample (Nicolai and Kieser 2002: 584) seems to be more important for practical relevance. By applying the requirements of “good” success factor research in our work, we aim to realize academic rigor in the sense of Kieser and Nicolai (2005: 278), one that achieves practical relevance by enhancing the dialogue about alternatives to existing practices of MNCs in their regional strategies, either regarding their regional product/service adaptation or their regional management autonomy.
2.4.2
Regional Corporate Success
The establishment of a solid theoretical and practical basis in this dialogue requires a concept of regional corporate success that recognizes both the requirements of academic rigor and practical relevance. To achieve such a “good” success factor research, we aim to focus on five facets of regional corporate success. First, regarding practical relevance, the generation of regional success indicators should be based on the external regional financial reporting of the Fortune Global 500 firms, as this leads to a common frame of reference in the translation between our research and the language and terminology used by practitioners. A synthesis of the requirements from accounting standards for geographical segment reporting (Behn et al. 2002: 33; Epstein and Jermakowicz 2009: 50; Langguth 2006: 629; Lee et al. 2008: 54; Prather-Kinsey and Meek 2004: 215; Schween 2006: 517) and an analysis of the annual reports of the Fortune Global 500 firms – for the period
38
This form of dialogue still regards science and practice as separated social systems, as it does not assume the rather uncommon case of collaborative research projects of scientists and researchers to jointly produce research results (Kieser and Leiner 2009: 529; Rasche 2007: 297). 39 Kieser and Nicolai (Kieser and Nicolai 2005: 277; Nicolai and Kieser 2002: 582) argue that the introduction of more and more control variables leads to a reduced clarity of the simple causal relationship which was originally identified by research.
2.4 Concept of Regional Corporate Success
61
2000–2008 – leads to five main categories in their external regional financial reporting: sales revenues, profits, investments, assets, and employees.40 Therefore, this regional information reported by the Fortune Global 500 firms will serve as guidance in generating regional success indicators. Second – with respect to the methodological rigor of this work – we want to analyze the development of regional corporate success indicators over time – to properly analyze the problems of simultaneity, unobserved heterogeneity, regression-to-the-mean, and of the survival bias (Nicolai and Kieser 2002: 584–585). This will be achieved by studying various success indicators of the Fortune Global 500 firms in a longitudinal study over nine financial reporting periods from 2000 to 2008. Here, based on the financial reporting information mentioned above, we aim to investigate absolute (e.g., regional sales revenue, regional profit, regional assets), relative (e.g., regional sales to total sales, regional profit to total profit, regional assets to total assets, regional sales per regional employee, regional profit per regional employee), and dynamic regional success indicators (e.g., sales growth, profit growth) of MNCs over time. These are very commonly utilized success indicators in success factor studies (Nicolai and Kieser 2002: 581). Such a study of regional performance over long time frames with large samples is highly encouraged (Glaum and Oesterle 2007: 315; Hult et al. 2008a: 1073; Seno-Alday 2009: 376). Third – as no consensus exists on which indicator appropriately constitutes firm performance (Nicolai and Kieser 2002: 586–587) – in our modeling of regional corporate success,41 we aim to define regional corporate success as a construct of various indicators. Here, following academic rigor, we apply adequate theories – particularly the dominant conceptual work on regional strategies (e.g., Rugman 2005b; Rugman and Verbeke 2008b) – to identify the most important indicators of regional corporate success. Fourth, regarding these most important regional corporate success indicators, the differences between their objective presentation by means of financial reporting information and its subjective estimations by MNC managers should be analyzed (Chen 1999: 161; Dawes 1999: 71; Dess and Robinson 1984: 270; Fritz 2004b: 29; Geringer and Hebert 1991: 256; Hult et al. 2008a: 1070–1071; Schmidt 2004: 151). Using the analysis of these differences, we aim to ensure that meaning is communicated between our research and its theoretical and practical implications by the transfer of schemas between the two contexts – which represents an important condition to bridge the gap between rigor and relevance (Hodgkinson and Rousseau 2009: 537–538; Kieser and Leiner 2009: 528; Starkey et al. 2009: 556). Fifth, following the moderate voluntaristic approach of this work, firm-specific regional corporate success indicators are understood as being contingent upon the internal and external organizational context of MNCs. Consequently, regional
40
This wording excellently represents the different terminology applied by the Fortune Global 500 firms in their geographical segment reporting, as illustrated in Table A2. Cf. Sects. 5.1.2 and 6.1. 41 Cf. Sect. 4.1.4.
62
2 Theoretical Foundation and Literature Review
corporate success is regarded as the result of a “fit” of appropriate degrees of firmspecific regional management autonomy (as regards managerial decisions) and regional product/service adaptation (as regards product attributes/service elements) with interaction effects from the internal regional orientation of the MNC and its external inter-regional distance. To capture and to properly examine these influences in the relationship between regional strategies and regional corporate success, in the following – based on methodological and conceptual grounds – we will develop a regional success factor model.
Part II
Development of a Regional Success Factor Model for the Analysis of the Regional Strategy–Success Relationship
.
Chapter 3
Structural Equation Methodology
3.1
Foundations of Structural Equation Modeling
Since the 1970s, structural equation modeling (SEM) has been applied for the analysis of causal relationships in economic and social sciences (G€otz and LiehrGobbers 2004: 714; Henseler 2005: 70; Herrmann et al. 2006: 35). SEM is a multivariate statistical technique that – by integrating different regression- and factor-analytic methods – allows the testing and estimation of theoretically derived, casual relationships between variables (Bortz 1993: 436; Rigdon 1998: 251). Its high popularity in economic and social sciences is mainly attributed to two factors. First, it is the only multivariate statistical method that allows one to simultaneously assess the quality of construct measurement in terms of reliability and validity on the one hand, while on the other hand estimating the strength of a relationship between constructs (Backhaus et al. 2008: 511; Henseler 2005: 70). Second, SEM enables scientists to measure not only observable (manifest) variables, but also unobservable (latent) variables (Chin 1998b: 296; Chin and Newsted 1999: 307; Herrmann et al. 2006: 35; Rigdon 1998: 251). As latent variables are very common in economic and social sciences, the ability to model latent variables is of particular importance (Chin 1998b: 296). Due to the fact that latent variables cannot be observed directly, they are assigned manifest variables – which can be derived empirically and measured on the basis of metric scales – in a measurement model (or outer model) (Backhaus et al. 2008: 513; Henseler 2005: 70), as illustrated in Fig. 3.1. Measured by its respective manifest (indicator) variables, the independent (exogenous) and dependent (endogenous) latent variables are then related with each other in a structural model (or inner model) (Henseler 2005: 71).1 This ability to model latent variables differentiates SEM from the so-called firstgeneration approaches, such as principal components analyses, factor analyses,
1
The independent variables are termed exogenous, as their causes lie outside the structural model, while the endogenous variables are determined by variables within the structural model (Bollen 1989: 12). The characteristics of structural models and measurement models, as well as the applied nomenclature in Fig. 3.1, will be explained in more detail in Sects. 3.2.1 and 3.2.2.
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_3, # Springer-Verlag Berlin Heidelberg 2011
65
66
3 Structural Equation Methodology Latent variables
Manifest (indicator) variables
x11
x12 .. .
Manifest (indicator) variables
π11 π12 .. . π1H1
λ11
υ2 ξ1
β21
ξ2
x1H1
λ12 .. . λ1H 2
x21
ε21
x22
ε22
.. .
.. .
x2H 2
Measurement model of the exogenous latent variable (here: formative mode)
ε2H 2
Measurement model of the endogenous latent variable (here: reflective mode)
Structural model
Fig. 3.1 Structural equation model with two latent variables Source: own illustration on the basis of Henseler (2005: 71)
discriminant analyses, or multiple regression methods (Chin and Newsted 1999: 307–308). SEM has several advantages over these statistical techniques, which have led to its classification as a second-generation approach (Chin 1998b: 296; Chin and Newsted 1999: 307–308; Fornell 1987: 408).2 Its main advantages in comparison to first-generation techniques are the following (Chin and Newsted 1999: 308): l
l l l
l
The possibility to model complex relationships between multiple independent and dependent variables The possibility to model non-observable, latent variables The possibility to model measurement errors for observable, manifest variables The possibility to: “[. . .] statistically test a priori substantive/theoretical and measurement assumptions against empirical data (i.e. confirmatory analysis)” (Chin 1998b: 297) The possibility to process correlated measures of latent variables (Rigdon 1998: 254)3
These advantages are important for a proper study of regional success factors of MNCs, mainly for four reasons. First, regional management autonomy and regional
2 As SEM is regarded as an extension of first-generation approaches, any constraints or assumptions that limit its application would set it back to a first-generation procedure (Chin and Newsted 1999: 308). 3 While using highly correlated measures as independent variables in regression analyses leads to multicollinearity (Homburg 1992: 499), in SEM – as correlated items are modeled as measures of one common variable, and as only this one variable enters the causal relationship – high correlations may even improve the robustness of the applied measurement and/or structural model (Rigdon 1998: 254–255).
3.1 Foundations of Structural Equation Modeling
67
product/service adaptation are constructs of the IB field that represent non-observable latent variables. Second, considering these latent variables and the additional, latent contingency variables regional orientation and inter-regional distance, the research approach of this work involves multiple independent variables in a complex regional strategyperformance relationship. Third, SEM represents a well-suited methodological technique to address the rigor-relevance distinction, as it explicitly distinguishes between a structural model (substance theory) and a measurement model (measurement theory). This separation reflects the two-languages-theory of Hempel (1951) and Carnap (1966), which differentiates between a theoretical language (level of the latent variable constructs) and an empirical language (level of the observed indicator variable) (Eggert and Fassott 2005: 34–35; Fornell 1989: 160; Homburg and Hildebrandt 1998: 18–19; Sarkar and Pfeifer 2005: 137). Until recently, these language differences have received considerable attention in the rigor-relevance debate (Fincham and Clark 2009: 512; Hodgkinson and Rousseau 2009: 543; Kieser and Leiner 2009: 517, 525, 528; Kieser and Nicolai 2005: 276; Starkey et al. 2009: 550, 556). In SEM, these two different language levels are methodologically connected by specifying rules of correspondence between manifest and latent variables. These rules of correspondence are statements that link abstract theoretical dimensions with their empirical indicators (Bagozzi 1977: 214; Bagozzi 1998: 50; Costner 1969: 245–246; Jarvis et al. 2003: 199), and that – depending on their direction – imply either a reflective measurement model or a formative measurement model (Eggert and Fassott 2005: 36).4 Fourth, a general problem of any of such measurement models is their – at least to a certain degree – incorrect measurement of latent variables. Therefore, regarding the trade-off between rigor and relevance, it could be argued that an exact measure of a latent variable in a strict sense is not possible. Assuming that this is true, particularly those analytical methods of examination should be applied which are able to reduce the proportion of measurement errors in the indicator variables by statistical controls. Here, the importance of SEM’s advantage of being the most powerful analytical method for controlling the proportion of measurement errors in manifest variables – while at the same time estimating the (measurement error adjusted) relationship between the latent variables – becomes evident (Albers and Hildebrandt 2006: 3; Chin et al. 2003: 197; Eggert and Fassott 2005: 36–37; Rigdon 1998: 253–254).5 Therefore SEM allows a theoretical reflection of practical phenomena, even though the observations of the latter involve measurement errors. Consequently, SEM represents an appropriate approach for this work, particularly 4 The distinctive characteristics of reflective and formative measurement models will be presented in Sect. 3.2.2. 5 The estimates of imperfections in the parameters of the measurement model give important information about the reliability and validity of the measurement model, which is an important pre-condition to derive reliable and valid insights from the theoretical substance of the relationship between the latent variables (Hildebrandt 2004: 545; Rigdon 1998: 254).
68
3 Structural Equation Methodology
due to its flexibility for integrating theory with empirically derived data (Chin and Newsted 1999: 308).
3.1.1
Alternative SEM Techniques: LISREL Versus PLS
In recent years, in the application of SEM, two second-generation approaches in particular have gained importance: covariance structure analysis and PLS (Herrmann et al. 2006: 35; Rigdon 1998: 252–253; Ringle et al. 2006: 81). Developed mainly by J€oreskog and S€ orbom (J€ oreskog 1973; J€ oreskog and S€orbom 1988), the covariance structure analysis is a causal analytical approach, which is often termed LISREL (linear structural relations), according to the so-called computer program for conducting SEM analysis (Backhaus and Ebers 2006: 607; Homburg and Hildebrandt 1998: 17).6 Even though LISREL still represents the dominant method for SEM, in recent times the alternative approach offered by PLS path modeling – which has been developed by Wold (1966) and Lohm€oller (1984, 1989) – has increasingly received attention for the analysis of latent variables in structural equation models (Chin 1998b: 297–298; G€ otz and Liehr-Gobbers 2004: 714; Scholderer and Balderjahn 2005: 88; Tenenhaus et al. 2005: 159–160).7 Based on a comparison of PLS and LISREL (Bliemel et al. 2005: 11; Chin and Newsted 1999: 314), as illustrated in Table 3.1, we give reasons for applying the PLS approach as a SEM method in this work. The first and most important argument for utilizing PLS is driven by the research aim of this work, which according to the following analysis is methodologically better captured by PLS than LISREL. The covariance-based8 analytical technique LISREL aims to minimize the difference between the covariances of an empirical sample and those predicted by a theoretical model (Chin 1998b: 297). Therefore, the parameter estimation process intends to most closely reproduce the covariance matrix of the empirically observed measures (Chin and Newsted 1999: 309). The objective of the variance-based9 method PLS, instead of trying to explain the covariation of all indicators, is prediction (Chin and Newsted 1999: 312). Based on an analysis of the raw data matrix (Henseler 2005: 70), the parameter estimation process of PLS aims to obtain the best weight estimates for each block of indicators that correspond to each latent variable (Chin 1998b: 301; Chin and Newsted 1999: 312; Herrmann et al. 2006: 39). This process employs the least squares method, which estimates parameters so that the sum of the squared residuals has its least 6 Further statistical programs for covariance structure analysis are, for example, AMOS, EQS, CALIS, SEPATH, and Mplus (Chin 1998b: 295; Grace 2006: 325). 7 The growing popularity of PLS path modeling is reflected by its increased utilization in recent journal publications (Henseler et al. 2009: 282). 8 The term “covariance” is used here in a broader sense, including both variance and correlation. 9 PLS is described as a variance analytic approach (Backhaus et al. 2008: 515).
3.1 Foundations of Structural Equation Modeling Table 3.1 Comparison of PLS and LISREL Criteria PLS Methodological Variance-based approach Main objective Prediction oriented: Explanation of latent and/or indicator variables
69
LISREL Covariance-based
Parameter oriented: Explanation of empirical data structures Implication Optimal for prediction accuracy Optimal for parameter accuracy Theory requirements Flexible High Assumptions Predictor specification Multivariate normal distribution and independent observations Measurement model Reflective and/or formative Reflective Large (minimum Sample size Also appropriate for small sample recommendations range sizes (minimum recommendations from 200 to 800 cases) range from 30 to 100 cases) Source: own illustration on the basis of Bliemel et al. (2005: 11) and Chin and Newsted (1999: 314)
value (Herrmann et al. 2006: 37; Schulze 1998: 137). The resulting parameter estimates based on the estimated indicator weights minimize the variance of all dependent (both latent and observed) variables (Backhaus et al. 2008: 515; Chin 1998b: 301; Chin and Newsted 1999: 312–313; Scholderer and Balderjahn 2005: 92). This process in PLS leads to: “[. . .] parameter estimates (i.e., weights for each individual indicator) in order to create latent variable scores that can be used to predict its own indicators or other latent variables” (Chin and Newsted 1999: 311).10 These parameter estimates, which aim at the best possible reproduction of the raw data matrix, have better predictive abilities than those estimated by LISREL – despite being less consistent (Bliemel et al. 2005: 11; Scholderer and Balderjahn 2005: 92).11 This explorative character of PLS is in contrast to the confirmative approach of LISREL, which tests the validity of a priori defined models (G€otz and Liehr-Gobbers 2004: 721). Thus, LISREL is more adequate for causal model/ theory testing, while PLS is more appropriate for component-based predictive modeling (Chin and Newsted 1999: 312; Henseler et al. 2009: 296–297; Homburg and Klarmann 2006: 734–735). However, it should be noted, that the application of PLS is not solely restricted to explorative research, but may also be utilized for confirmative research (Albers and
10
For a more detailed description of the PLS path algorithm and its estimation procedure cf. Chin (1998b: 302–303), Chin and Newsted (1999: 315–321), and G€otz and Liehr-Gobbers (2004: 722–724). 11 The limited consistency of the parameter estimates in PLS is due to the fact that a consistent approximation of the correct parameter value requires a large sample and numerous indicators for each latent construct (consistency at large) (Albers and Hildebrandt 2006: 15; Bliemel et al. 2005: 11; Chin 1998b: 329–330). Therefore, while PLS is coherent in a predictive sense (Chin 1998b: 303), its parameter estimation process – in contrast to LISREL – is not consistent (Scholderer and Balderjahn 2005: 92).
70
3 Structural Equation Methodology
Hildebrandt 2006: 29). This flexibility is due to the fact that PLS has much lower requirements – in comparison to LISREL – on prior theoretical knowledge (Chin and Newsted 1999: 311). Consequently, PLS also allows the confirmation of unknown relationships that have been defined solely on the basis of plausible hypotheses (Bliemel et al. 2005: 10–11; Henseler 2005: 70). This possibility to: “[. . .] suggest where relationships might or might not exist [. . .]” (Chin and Newsted 1999: 313), that have been neglected in practice so far, is important to overcome the gap between rigor and relevance (Kieser and Nicolai 2003: 592; Kieser and Nicolai 2005: 278).12 Here, due to its ability to test the prediction accuracy of a model, PLS is applied particularly in early stages of research and theory development (G€ otz and Liehr-Gobbers 2004: 721). As the theoretical basis for the relationships between our latent variables – regional management autonomy, regional product/service adaptation, regional orientation, inter-regional distance, and regional success – is still at an early stage in the IB field, PLS is the appropriate modeling technique for this work. Its predictive ability is particularly important for being able to offer practically relevant alternatives to real-world issues (Kieser and Nicolai 2003: 592; Kieser and Nicolai 2005: 278). Besides the focus on assessing the prediction quality of a certain model, PLS also analyzes its explanatory power based on the explained variance given by the R2 value (Chin 1998b: 316–318; Fassott 2005: 28). As this is not crucial for the fit values of covariance-based procedures, LISREL has shortcomings – in comparison to PLS – regarding both the explanatory and predictive power of SEM techniques (Fassott 2005: 28). This was confirmed by Weiber and Adler (2002: 12) who – despite a good model fit – could not realize a satisfactory explanatory power of their model (Fassott 2005: 28). The possibility to assess both the prediction accuracy of a model and its explanatory power – i.e., whether the impact of a certain independent latent variable on a dependent latent variable has a substantive impact (Chin 1998b: 316) – is important for identifying the central driving factors of a MNC’s regional success in this work. In addition, LISREL has much more restrictive assumptions than PLS (Henseler et al. 2009: 280; Homburg 1992: 502; Lee 2000: 196). Covariance-based approaches like LISREL – to achieve unbiased and asymptotically consistent estimations of model parameters based on the maximum likelihood method – require a multivariate normal distribution of the data, that the observations are independent of each other, and a large sample size (Chin and Newsted 1999: 309; Henseler et al. 2009: 281; Scholderer and Balderjahn 2005: 91; Venaik et al. 2005: 665). In contrast, PLS does not require a normal distribution of manifest variables, as its parameter estimations are based solely on the regression principle (G€otz and Liehr-Gobbers 2004: 721). The assumptions in PLS are related to the following specification of latent variables xj in dependence of other latent variables xi within the structural model (Chin and Newsted 1999: 321–322; Henseler 2005: 71):
12
Cf. Sect. 2.4.1.
3.1 Foundations of Structural Equation Modeling
xj ¼
X
71
bji xi þ uj ;
(3.1)
i
X bji xi : E xj ?jxi ¼
(3.2)
i
In (3.1), bji is the path coefficient linking the endogenous and exogenous latent variables xi and xj over the range specified by the index i, and uj is the inner residual variable (Chin 1998b: 312). This shows the clearly defined causal direction between the latent variables in the structural model, as illustrated by the one-way arrows in Fig. 3.1 (Chin and Newsted 1999: 321). Equation (3.2) shows that the expected value of a latent variable corresponds to a linear function of its predictors (Chin 1998b: 313). These assumptions are much less restrictive than those of normally distributed data. They imply that there are no linear relationships between the predictors and the residual, which means that – as illustrated in (3.3) and (3.4) – the expected value of the error term uj equals zero, as well as the covariance of uj and xi (Chin and Newsted 1999: 322): Eðuj jxi Þ ¼ 0
(3.3)
Covðuj ; xi Þ ¼ 0:
(3.4)
Besides its less restrictive assumptions, PLS allows the application of both reflective and formative measurement models, while covariance-based SEM analyses by LISREL are typically based on reflective measurement models (Chin and Newsted 1999: 310; Diller 2006: 614; Eberl 2006: 661; G€otz and Liehr-Gobbers 2004: 715; Henseler et al. 2009: 283; Homburg and Giering 1996: 6).13 This is due to the fact that the estimation of formative relationships in covariance-based approaches is complicated, as it depends upon certain conditions (G€otz and Liehr-Gobbers 2004: 715; Homburg and Klarmann 2006: 735; Jarvis et al. 2003: 213),14 and it may involve: “[. . .] identification problems, implied covariances of zero among some indicators, and/or the existence of equivalent models” (Chin and Newsted 1999: 310). Due to the fact that our regional success factor model, as we will show below,15 involves formative measurement models – PLS is better suited for the SEM of this work. Furthermore, minimal recommendations in LISREL for the size of the research sample range from 200 to 800 cases, while in PLS a range from 30 to 100 cases is
13
The distinctive characteristics of reflective and formative measurement models will be presented in Sect. 3.2.2. 14 For more details about these conditions cf. G€ otz and Liehr-Gobbers (2004: 715, 732), and Jarvis et al. (2003: 213). 15 Cf. Sect. 4.1.1.
72
3 Structural Equation Methodology
sufficient (Backhaus et al. 2006: 717; Chin and Newsted 1999: 314; Diller 2006: 615; Henseler et al. 2009: 283, 291–292; Homburg and Klarmann 2006: 733–734). The latter range of survey responses seems more realistic for the research sample of this work, as we do not expect to achieve the minimum sample size in LISREL of 200 cases for the Fortune Global 500 firms.16 Despite these advantages of PLS in comparison to LISREL for this work, PLS also faces several limitations – which are briefly presented in the form of its main criticisms. First, in comparison to LISREL, PLS is restricted in its understanding of underlying theoretical relationships – due to its lack of goodness-of-fit criteria (Henseler et al. 2009: 296–297). A related, second criticism refers to the fact that PLS does not provide quality criteria that relate to the structural model in its entirety (Herrmann et al. 2006: 59–60; Homburg and Klarmann 2006: 734–735). Third, PLS does not allow a separation of the “true” variance of a latent construct from the variance of measurement errors (Scholderer and Balderjahn 2005: 97). These advantages of LISREL to measure the “true” construct variance, and its benefits of being an inference-statistical procedure, are particularly important in the comparison of competing theoretical models (Scholderer and Balderjahn 2005: 98), which thus cannot be achieved by PLS. In recognizing these criticisms, we agree with Scholderer and Balderjahn (2005: 98) that neither PLS nor LISREL is generally superior over the other, so that the research objective should be the main influence in selecting an appropriate SEM technique. Here, as outlined previously, the explorative nature of PLS, in particular, is important for the study for the relationships in our structural model, which have not been analyzed in the IB field in this constellation before. Therefore, also considering the conceptual and methodological reasons stated above, PLS represents the appropriate SEM technique of this work. For its application, in a similar manner as the prevailing PLS implementations of structural equation models, we will use the SEM software SmartPLS (Henseler et al. 2009: 309).17 To ensure a proper methodological application of PLS in this work, we will present the distinguishing characteristics of structural models and measurement models, and particularly the quality criteria for their evaluation.
16
This represents a response rate of 40%, which we do not expect to realize for the Fortune Global 500 firms, as cross-national mail surveys usually achieve response rates between 6% and 16% (Harzing 1997: 643; Harzing 2000: 244). 17 We apply the most recent version SmartPLS 2.0 M3 (Ringle et al. 2005). Alternative software applications are, for example, LVPLS (latent variable partial least square), PLS-graph, PLS-GUI, SPAD-PLS, and ParLeS (Albers and Hildebrandt 2006: 27; G€otz and Liehr-Gobbers 2004: 715).
3.2 The Partial Least Squares Approach to Structural Equation Modeling
3.2
3.2.1
73
The Partial Least Squares Approach to Structural Equation Modeling Structural Model
The structural model in Fig. 3.1 describes the theoretically derived relationships among the exogenous (independent) and endogenous (dependent) latent variables. These relationships of the inner model are defined by a system of linear structural equations (Chin 1998b): xj ¼ Bxj þ Gxi þ u:
(3.5)
In (3.5), xj represents the vector of the endogenous latent variables, xi the vector of the exogenous latent variables, u the vector of residual variables (i.e., unexplained variance), where the parameter matrices (or coefficient matrices) B and G depict the relationships (or path coefficients) in the structural model (Betzin and Henseler 2005: 53; Chin and Newsted 1999: 321; G€otz and Liehr-Gobbers 2004: 717). While B represents the relationship between the endogenous latent variables xj, the matrix G depicts the effects between the exogenous and endogenous latent variables. The vector u describes the measurement errors of the endogenous latent variables, which – as illustrated in formulas (3.3) and (3.4) above – are assumed to have an expected value of zero, and no correlations with the latent variables (G€otz and Liehr-Gobbers 2004: 717). A necessary condition for a valid structural model is its so-called recursivity (Henseler 2005: 71; Tenenhaus et al. 2005: 166). A structural model can be assumed to be recursive, if there is no latent variable within the model that is either directly or indirectly (via another latent variable) related with itself (Henseler 2005: 71; Ringle et al. 2006: 81). The existence of such non-recursive effects would be highly problematic, as this would not allow to unambiguously depict the causal relationships in the structural model (Reinecke 2005: 59), which form the basis for testing the underlying research hypotheses in SEM. Structural models may take various forms and can be much more complex than the one depicted in Fig. 3.1. Here, two possible extensions of such simple structural models are important for this work: interaction effects from moderating variables and second-order constructs. In the following these extensions of the structural model will be briefly presented. First, the relationship or interaction between an exogenous and an endogenous variable in a structural model may be influenced by moderating variables (Eggert et al. 2005: 102).18 According to Baron and Kenny (1986: 1174) a moderator
18
Despite not being relevant for this work, it should be noted that – apart from moderating variables – mediating variables may also influence this relationship (Shrout and Bolger 2002: 422). In the case of such mediation variables, the impact of an exogenous variable on a dependent,
74
3 Structural Equation Methodology
variable: “[. . .] is a qualitative (e.g., sex, race, class) or quantitative (e.g., level of reward) variable that affects the direction and/or strength of the relation between an independent or predictor variable and a dependent or criterion variable” (Eggert et al. 2005: 104; Huber et al. 2006: 697). Therefore, in the analysis of moderating effects, besides the relationship between an exogenous and an endogenous variable, the direct influence of the moderator variable on the endogenous variable and the influence of an interaction variable – the Cartesian product of the exogenous variable and the moderator variable – is also assessed (Eggert et al. 2005: 107; G€ otz and Liehr-Gobbers 2004: 725; Hartmann and Moers 1999: 293; Homburg and Klarmann 2006: 730). The estimation of such interaction effects poses a problem for covariance-based approaches like LISREL due to its previously mentioned, restrictive assumptions – e.g., that the error terms of the indicator variable are uncorrelated (Chin et al. 2003: 197; Eggert et al. 2005: 107; Scholderer et al. 2006: 646).19 Consequently, even though highly encouraged, only very few analyses of moderating effects in complex relationships exist (Chin et al. 2003: 191–193; Eggert et al. 2005: 107–108; Homburg and Giering 2001: 47). Given that the PLS algorithm does not assume uncorrelated error terms, in its estimation of the interaction effect, these correlations may actually help to improve the accuracy of estimates for moderating influences (Chin et al. 2003: 198; Eggert et al. 2005: 108; Huber et al. 2006: 701). Secondly, another complexity that may be added to structural equation models is given by second-order or multidimensional constructs (Albers and G€otz 2006: 670; Chin 1998a: x; Giere et al. 2006: 678; Jarvis et al. 2003: 204). A second-order construct consists of multiple, uncorrelated dimensions (first-order constructs) that are measured by their respective indicators (Albers and G€otz 2006: 670). Consequently, in contrast to unidimensional latent constructs – which are directly measured by its manifest indicators, as illustrated in Fig. 3.1 – second-order constructs are not directly connected to any empirically measured items (Chin 1998a: x; Homburg and Giering 1996: 6; Maloney 2007: 258). Instead, they represent a multidimensional construct which integrates different, but related (first-order) dimensions – that are measured in either a formative or reflective manner – to one (second-order) latent construct (Giere et al. 2006: 678).20 This allows analyzing (non-observable) latent constructs on a comparable (first-order) endogenous variable is partly or completely mediated by a so-called mediator (Eggert et al. 2005: 103). The existence of mediation effects thus leads to a partial or complete disconnection of the relationship between the exogenous constructs and the endogenous variables (Eggert et al. 2005: 103–104). Instead of the disconnection of this relationship, only effects on its strength are assumed in this work, which are better conceptualized by moderating variables (Baron and Kenny 1986: 1174; Eggert et al. 2005: 103–104). In Sect. 4.1.3, we will explain the moderating effects on the strength of this relationship in more detail. 19 Latent interaction variables do not conform to this assumption, as their indicators result from multiplication with an exogenous variable, and thus inevitably share a part of their variance with the indicators of the exogenous construct (Eggert et al. 2005). 20 In Sect. 3.2.2., the distinctive characteristics of reflective and formative measurement models will be presented.
3.2 The Partial Least Squares Approach to Structural Equation Modeling
75
level of abstraction within the same (second-order) nomological network (Chin 1998a: x; Giere et al. 2006: 679; Jarvis et al. 2003: 204). In the empirical research of social science, many latent constructs are more properly conceptualized as multidimensional (e.g., business strategy), being divided into separate latent constructs that represent different dimensions (e.g., aspects of business strategy like strategic decision-making) (Hulland 1999: 196–197; Jarvis et al. 2003: 204; Maloney 2007: 257; Venkatraman 1989a: 944, 947–949). As neither unidimensional nor multidimensional latent variables of a structural model can be measured directly, as they are not observable, they have to be operationalized in a measurement model. Here, different forms of measurement models can be distinguished, which will be presented in the following.
3.2.2
Measurement Model
As illustrated in Fig. 3.1, the operationalization of measurement models is achieved by defining how each block of manifest indicator variables xjh in the outer model relates to its respective latent variable xj in the inner model (Chin and Newsted 1999: 322; Henseler 2005: 70). The latent variables of the measurement model in Fig. 3.1 are depicted as circles, and its indicators as rectangles. In this specification of the measurement models for the endogenous and exogenous variables, besides assigning appropriate manifest (indicator) variables to the latent constructs, the form of measurement model has to be defined as reflective or formative (G€otz and Liehr-Gobbers 2004: 717; Henseler 2005: 70). Depending on the particular construct, a measurement model may consist exclusively of reflective or formative indicators, or may include both reflective and formative indicators (G€otz and LiehrGobbers 2004: 717). The decision for a reflective or a formative measurement model depends on the causal direction (the rules of correspondence) in the relationship between the latent variable and its indicators (Eggert and Fassott 2005: 36). In Fig. 3.1, this causal direction of the measurement model is illustrated by the arrows that point from the latent variable to the indicators (reflective), or that point from the indicators to the latent variable (formative).
3.2.2.1
Reflective Measurement Model
A reflective measurement model assumes that the theoretically derived, latent variable causes its related manifest variables (Eggert and Fassott 2005: 36; Henseler 2005: 70). Consequently, a change in the latent variable always leads to a change of all its reflective (effect) indicators (Bollen and Ting 2000: 3; G€otz and Liehr-Gobbers 2004: 718; Herrmann et al. 2006: 36). This logic corresponds to the confirmative factor analytic approach, where each indicator is a linear function of its factor (its corresponding latent variable) and of possible measurement errors (Henseler et al. 2009: 285; Homburg and Giering 1996: 9). Accordingly, the relationship between a
76
3 Structural Equation Methodology
latent variable and its indicators in a reflective measurement model is depicted as follows (Betzin and Henseler 2005: 54; G€ otz and Liehr-Gobbers 2004: 718; Henseler 2005: 71): xih ¼ lih xi þ eih for the exogenous reflective measurement model, and
(3.6)
xjh ¼ ljh xj þ ejh for the endogenous reflective measurement model:
(3.7)
In the (3.6) and (3.7), eih and ejh are the measurement errors of the indicator variables xih and xjh, while the loading coefficients lih and ljh describe the relationship between xi and xih, or xj and xjh (Chin 1998b: 313; G€otz and Liehr-Gobbers 2004: 718). The measurement errors of the exogenous and endogenous indicator variables eih and ejh are not correlated and neither do they correlate with the latent variables nor the measurement errors of the structural model (Homburg and Hildebrandt 1998: 21; Tenenhaus et al. 2005: 163). As a result, each indicator represents an erroneous measurement of its latent variable (G€otz and Liehr-Gobbers 2004: 718). As a reflectively measured latent variable is related to several indicators, such measurement error-related biases of particular indicators can be absorbed (Eggert and Fassott 2005: 36–37; Homburg and Dobratz 1998: 450). The indicators in reflective measurement models should be highly correlated, as within every block of indicators, each of these indicators measures the same latent variable (G€otz and Liehr-Gobbers 2004: 718). This implies that the measurement error of reflective indicators should be as marginal as possible, because the lesser the measurement error of one indicator, the higher its correlation with the other indicators (Eggert and Fassott 2005: 37). Therefore, a high indicator correlation implies a high reliability and validity of the measurement model for a latent variable (Eggert and Fassott 2005: 37; Homburg and Giering 1996: 8–10). Furthermore, even though the indicators of a reflective measurement model should ideally illustrate the different facets of the latent variable, they are exchangeable measures of this latent construct (Eggert and Fassott 2005: 38; Herrmann et al. 2006: 47). For example, an elimination of the indicator x21 in Fig. 3.1 would neither change the inter-correlations between the remaining indicators x22 and x2H2 , nor between x22, x2H2 , and x2 (Bollen and Lennox 1991: 308; Eggert and Fassott 2005: 37). This is due to the fact that reflective indicators are solely some possible representatives of a latent construct (Herrmann et al. 2006: 47).
3.2.2.2
Formative Measurement Model
The opposite is true for formative measurement models, where the manifest (indicator) variables cause the parameter values of their respective latent variable (Henseler 2005: 71). Here, a change in a formative (or causal) indicator necessarily leads to a change in the latent variable (Eggert and Fassott 2005: 38; Herrmann et al. 2006: 36). However, the values of the other indicator variables are not affected
3.2 The Partial Least Squares Approach to Structural Equation Modeling
77
by this, meaning that a change of the latent variable may be related to the change of exclusively one indicator (Eggert and Fassott 2005). Therefore, besides its belonging to the same latent construct, formative indicators are relatively independent of each other and thus are not necessarily correlated (G€otz and Liehr-Gobbers 2004: 718; Herrmann et al. 2006: 48).21 Grounded on the logic of a (multiple) regression analysis – which analyzes the effect of (multiple) independent variables on a dependent variable – the latent variable in a formative measurement model represents a linear combination of its indicators plus a residual term (Chin 1998b: 314; Diamantopoulos and Winklhofer 2001: 270; G€ otz and Liehr-Gobbers 2004: 718–719; Henseler 2005: 71; Henseler et al. 2009: 286; Krafft et al. 2005: 78; Tenenhaus et al. 2005: 164–165): xi ¼
X
pih xih þ di for the exogenous reflective measurement model, and (3.8)
h
xj ¼
X
pjh xjh þ dj for the endogenous reflective measurement model:
(3.9)
h
In addition to the variables xi, xj, xih, and xjh – which have been described before – equations (3.8) and (3.9) include (multiple) regression coefficients in the form of matrices pih and pjh, and error terms represented by di and dj (G€otz and LiehrGobbers 2004: 719; Henseler 2005: 71), where the latter are not correlated with the formative indicators (Bollen and Lennox 1991: 306; Tenenhaus et al. 2005: 165). The (multiple) regression coefficients of formative indicators represent their loading (p) on their respective latent variable (x) (Chin and Newsted 1999: 323; Eggert and Fassott 2005: 38). These formative indicator weights are different from the loading coefficients in a reflective measurement model, as their variation cannot be attributed to changes of the latent construct (Herrmann et al. 2006: 48). Instead, their variation explains changes in their corresponding latent variable. Here, the formative measurement model usually assumes an error-free measurement of all facets of the latent variable (with di and dj ¼ 0) (G€otz and Liehr-Gobbers 2004: 719; Hildebrandt and Temme 2006: 620). As, however, a complete coverage of the latent construct is not always possible (G€ otz and Liehr-Gobbers 2004: 719), measurement errors in the form of di and dj may occur. In contrast to a reflective measurement model, formative measurement models do not assume that these error terms arise at the level of the indicators, but at the level of the latent variables (Diamantopoulos and Winklhofer 2001: 271). As formative indicators determine a latent variable – in contrast to reflective indicators – they are not interchangeable (Diamantopoulos and Winklhofer 2001; Eggert and Fassott 2005: 39). This is due to the fact that the elimination of one item 21
Even though formative indicators are usually slightly positively or negatively correlated, it is possible that formative indicators are highly correlated with each other, which leads to the problem of multicollinearity (Krafft et al. 2005: 78). Appropriate quality criteria to control for this problem will be presented in Sect. 3.2.3.
78
3 Structural Equation Methodology
from the block of formative indicators may omit a unique part of its corresponding latent variable, and thus may alter the empirical meaning of the composite latent construct (Jarvis et al. 2003: 202). Given their importance for the adequate measurement of latent variables, the differentiating characteristics of reflective and formative measurement models will be elaborated in more detail.
3.2.2.3
Differentiation Between Reflective and Formative Measurement Models
Since the mid 1960s, a more stringent differentiation between reflective and formative measurement models has been postulated in the literature (Eggert and Fassott 2005: 32), which is particularly due to substantial misspecifications of measurement models in the past. This has been illustrated by Jarvis et al. (2003: 206–207), who find that over 20 years from 1977 to 2000, almost one third of academic publications in important marketing journals22 utilize incorrectly specified measurement models – where almost all misspecifications are related to the use of a reflective measurement model for latent variable constructs that should have been formatively modeled (Eggert and Fassott 2005: 32; Henseler et al. 2009: 290–291). This dominance of reflective measures has been attributed to the fact that covariance-based evaluation techniques, such as LISREL or AMOS, by default assume reflective measurement models (Diamantopoulos and Winklhofer 2001: 274; Eberl 2006: 661; Eggert and Fassott 2005: 32). Independent of the SEM technique, such misspecifications can lead to considerable distortions of the parameter estimates in the structural model, which may seriously affect the theoretical conclusions drawn from the model (Diamantopoulos and Papadopoulos 2010: 368; Eggert and Fassott 2005: 32–33; Jarvis et al. 2003: 207–212). More specifically, the theoretical relationships between latent variables in the structural model may be either over- or underestimated by such misspecifications (Jarvis et al. 2003: 212). Furthermore – and even more critical than incorrectly specified measurement models – while indicators in a reflective measurement model may be omitted due to their low correlation with the other indicators, in a formative measurement model this elimination changes the meaning of the latent constructs (Albers and Hildebrandt 2006: 11, 24–25). Consequently, the differentiation between reflective and formative measurement models is important. To properly differentiate reflective and formative measurement models, besides considering their conceptual differences, explicit decision rules should be applied to determine whether a construct requires a reflective or formative measurement (Fassott 2006: 71; Henseler et al. 2009: 291). Such decision rules have been proposed by Jarvis et al. (2003: 203), who propose the differentiation of reflective 22
The academic journals of their study included the Journal of Consumer Research, the Journal of Marketing, the Journal of Marketing Research, and Marketing Science (Jarvis et al. 2003: 206).
3.2 The Partial Least Squares Approach to Structural Equation Modeling
79
and formative measurement models according to four criteria: the direction of causality from the latent construct to its measure; the interchangeability of the indicators; the correlation between the indicators; and the nomological aspects of indicators (i.e., if they have the same antecedents and consequences).23 Despite properly applying these decision rules, Jarvis et al. (2003: 206) could not tell in 14% of the cases in their above mentioned analysis of academic journals, whether a construct should have been modeled in a reflective or formative measurement model. The same has been observed by Eggert and Fassot (2005: 44) in their analysis of these decision rules for 180 latent variables, in which they could not differentiate in 18 cases (10%) between a reflective and a formative measurement model. In his analysis of important German academic journals,24 Fassot (2006: 75) could not make this distinction for 16% of the published construct measurement models, particularly due to missing statements about the reasons for either a reflective or formative operationalization of the respective measurement model. Furthermore, it has been argued that out of the four decision rules of Jarvis et al. (2003: 203), only the first decision criterion – the direction of causality – has to be considered, as the other three decision criteria represent its respective consequences (Eberl 2006: 657; Herrmann et al. 2006: 48). Following these insights, this work will mainly apply the direction of causality between latent variables and their indicators, to properly differentiate between reflective and formative measurement models. If, however, this decision criterion is not sufficient for an adequate differentiation, the other three decision rules of Jarvis et al. (2003: 203) will be supplemented. In general, to avoid any ambiguities, we follow the recommendation of Eggert and Fassot (2005: 47) to always explicitly outline the rationale for selecting either a reflective or a formative measurement model.25 Besides these preliminary considerations in the development of measurement models, additional criteria have to be introduced to evaluate the quality of SEM results. Therefore in the following, we will present quality criteria for evaluating both measurement models and structural models.
23
The nomological aspects of indicators may differ in the case of formative measurement models (i.e., they are not required to have the same antecedents and consequences) (Jarvis et al. 2003: 203). Therefore, in contrast to reflective measurement models (i.e., where indicators have the same antecedents and consequences), the elimination of an indicator in a formative measurement model requires a check, if the new model version functions in predictable ways (Diamantopoulos and Winklhofer 2001: 273; Jarvis et al. 2003: 203). 24 His analysis covered the following German academic publications: Zeitschrift f€ ur Betriebswirtschaft, Die Betriebswirtschaft, and Zeitschrift f€ ur betriebswirtschaftliche Forschung (Fassott 2006: 74). 25 Cf. Sect. 4.1.
80
3.2.3
3 Structural Equation Methodology
Quality Criteria for Evaluating Structural Equation Models
The ability of a specified model to describe the relationships between the observed variables is the main focus in the quality evaluation of structural equation models (Krafft et al. 2005: 72). This evaluation process usually involves two steps: first, the quality of the measurement models is assessed – before evaluating in a second step, if the structural model reproduces the indicator values (G€otz and Liehr-Gobbers 2004: 727). In the first step, the differences between reflective and formative measurement models – as outlined above – require different quality criteria for their evaluation (G€ otz and Liehr-Gobbers 2004: 728). While the quality assessment of reflective measurement models is primarily based on the internal consistency of their indicators (reliability and validity),26 the quality of formative measurement models is evaluated primarily by the existence of multicollinearity (G€otz and LiehrGobbers 2004: 729). The second step mainly focuses on evaluating the explanatory power and predictive ability of the structural model (Krafft et al. 2005: 83–85). In the following, we follow this stepwise procedure in the presentation of the most important quality criteria for the SEM of this work.
3.2.3.1
Quality Criteria for Evaluating Reflective Measurement Models
Each indicator in reflective measurement models, as outlined above, represents an erroneous measurement of its latent variable (G€ otz and Liehr-Gobbers 2004: 727). Here, random (non-systematic) measurement errors, which influence construct measurement without any noticeable systematics, are differentiated from systematic measurement errors, which occur (independently of random measurement errors) in the same size for each measurement repetition (Churchill 1987: 73; Krafft et al. 2005: 73). Given these different types of measurement errors, the quality of reflective measurement models is evaluated mainly on the basis of their reliability and validity. A reflective measurement is termed reliable if it – under identical measurement conditions – repeatedly leads to exactly the same measures, thus where the random measurement error equals zero (Bollen 1989: 207; G€otz and Liehr-Gobbers 2004: 727; Homburg and Giering 1996: 6). A reflective measurement is termed valid if it really measures what it intends to measure, thus where both the random and the systematic measurement errors equal zero (Bollen 1989: 184; Homburg and Giering 1996: 7; Krafft et al. 2005: 73). Thus – as the reliability of measurement is included in its validity – the reliability is a necessary, but not a sufficient, condition for the validity of reflective measurement models (Homburg and Giering 1996: 7). In this work, different quality criteria are applied to evaluate the reliability and validity of reflective measurement models. This evaluation involves the application of both first-generation and more powerful, second-generation quality criteria 26
The terms reliability and validity will be explained in more detail in the next chapter.
3.2 The Partial Least Squares Approach to Structural Equation Modeling
81
(Homburg and Giering 1996: 8; Kirstein 2009: 175).27 As illustrated in Table 3.2, four forms of quality evaluation for reflective measurement models are distinguished (Homburg and Giering 1996: 8, 13; Kirstein 2009: 182; Krafft et al. 2005: 75). Content validity describes the degree by which the variables of a measurement model correspond to the content-semantic field of the latent construct (Bohrnstedt 1970: 92; G€otz and Liehr-Gobbers 2004: 727). Content validity is usually assessed by two quality criteria. First, expert validity refers to a quality evaluation procedure during the pretest, where test persons (e.g., experts or probands of the pretest sample) are asked to allocate indicators to constructs (Krafft et al. 2005: 76). To evaluate the unambiguousness of the allocation of the indicators to the respective construct and their content relevance, Anderson and Gerbing (1991: 734) suggest two indices (Eggert and Fassott 2005: 41): psa ¼
nc for the indicator’s proportion of substantive agreement, and N
(3.10)
nc n0 as substantive-validity coefficient for the content N relevance of the indicators:
(3.11)
csv ¼
Equation (3.10) depicts the degree of correspondence between the a priori intended and the factual indicator allocation by the expert to a latent construct (Eggert and Fassott 2005: 41). Here, nc refers to the number of experts that have correctly allocated the indicator to its construct – corresponding to the a priori intended allocation – and N depicts the total number of all probands (Krafft et al. 2005: 77). The psa index produces values from 0 to þ1, where higher values indicate a high degree of correspondence (Krafft et al. 2005: 77). Equation (3.11) reflects the substantive validity of the indicators, calculated by the difference between the number of “correct” (nc) and the most frequently “wrong” (no) allocation of the indicator to the latent construct, divided by the total number N of all probands (Eggert and Fassott 2005: 42). The csv index produces values from 1 to þ1, where high positive values reflect a high substantive validity (Anderson and Gerbing 1991: 734; Krafft et al. 2005: 77). If the values are close to 1, the indicator has a high substantive validity for another latent variable, different from the a priori intended allocation (Eggert and Fassott 2005: 42).
27
First-generation quality criteria (e.g., explorative factor analysis, Cronbach’s alpha) were originally developed in the 1950s in the area of psychology/psychometrics and have been promoted mainly by the work of Churchill (1979) in marketing research (Homburg and Giering 1996: 8). With the introduction and spread of the confirmatory factor analysis in marketing research, increasingly new and more powerful quality criteria (e.g., indicator reliability, construct reliability) were developed – which were termed second-generation quality criteria (Homburg and Giering 1996: 8, 13).
Degree of measurement quality for a latent variable by its assigned indicator variables
r0:4
used scales)
x
Internal consistency
P 2 ð lij Þ Internal consistency ¼ P 2 i P r0:7 l þ varðeij Þ ij i i P 2 li Average variance extracted r0:5 AVE ¼ P 2 iP
Cronbach’s alpha
jh Þ
Test of significance of factor loadings by (onesided) t-values, requiring a 5% level with a minimum value r 1.66 (based on bootstrapping method) P 2 s n a ¼ n1 1 s2 i r0:7 (or r 0.5 for first-time
jh
jh Indicator reliability ¼ l2 þvarðe
l2
Proportion of substantive agreement psa ¼ nNc (high value within range from 0 to 1) 0 (high Substantive-validity coefficient csv ¼ nc n N value within range from 1 to 1) Factor loading 0.4 with lower cross-loadings
Methods/criteria
Degree by which measurements of distinct constructs differ from each other, utilizing the l þ varðei Þ i i i same measurement instrument Source: own illustration on the basis of Homburg and Giering (1996: 8, 13), Kirstein (2009: 182), and Krafft et al. (2005: 75)
Discriminant validity
Construct reliability
Table 3.2 Quality criteria for the evaluation of reflective measurement models Quality criteria for reflective measurement models Form of quality Definition Quality criteria evaluation Expert validity Content validity Degree of correspondence between a priori intended and factual indicator allocation, or degree by which construct meaning is captured by the indicators Explorative factor analysis Indicator reliability Degree of explanation of indicator variance by Interpretation of loadings the construct Reliability of parameter estimations
82 3 Structural Equation Methodology
3.2 The Partial Least Squares Approach to Structural Equation Modeling
83
A second procedure to evaluate content validity is the explorative factor analysis (Krafft et al. 2005: 73). This technique analyzes the indicators regarding its underlying factor structure, where it aims to depict all indicators by the least number of factors (Homburg and Giering 1996: 8). In this empirical evaluation of the hypothetically derived factor structure, content validity is given if all indicators have sufficiently high factor loadings to its corresponding latent construct (e.g., r 0.4 with lower crossloadings to other latent variables) (Homburg and Giering 1996: 8).28 Besides their content validity, reflective measurement models are evaluated on the basis of indicator reliability, which assesses the degree by which the variance of an indicator is explained by its underlying latent variable (G€otz and Liehr-Gobbers 2004: 727). It is calculated as follows (Hildebrandt and Temme 2006: 625; Nacif 2003: 182): Indicator reliability ¼
l2jh l2jh þ varðejh Þ
:
(3.12)
In (3.12), ljh represents the estimated factor loading of the indicator xjh on its corresponding latent construct xj – as illustrated in Fig. 3.1 – while var (ejh) corresponds to the estimated variance of the measurement error ejh (Nacif 2003: 182). This quality criterion thus describes how well a single indicator is explained by the latent variable and not by its measurement error, where researchers usually suggest a cut-off point of 0.5 within its value range from 0 to þ1 (Nacif 2003: 182). This cut-off point has been chosen to reflect that more than 50% of indicator variance can be related to the latent variable (G€ otz and Liehr-Gobbers 2004: 727; Hulland 1999: 198). A value above 0.7 is perceived as acceptable (Carmines and Zeller 1979: 27; Herrmann et al. 2006: 56; Krafft et al. 2005: 73), while newly developed scales may have lower values that, however, should not be less than 0.4 (Huber et al. 2007: 25; Hulland 1999: 198). As this work utilizes newly developed scales,29 a minimum value of 0.4 appears reasonable. Furthermore, indicator reliability is assessed by the reliability of parameter estimations, which is measured by the significance of factor loadings in terms of their t-values (Chin 1998b: 319). To calculate t-values, the resampling procedures bootstrapping and jackknifing are applied in PLS (Chin 1998b: 318–320). Here, bootstrapping is the preferred method, given its lower standard error and the fact that it is more efficient than jackknifing, as the latter represents an approximation of bootstrapping (Chin 1998b: 320; Efron and Tibshirani 1993: 145–146, 291–293; Herrmann et al. 2006: 56). The underlying null hypothesis of bootstrapping assumes that the factor loading equals zero, which can be rejected by t-values in a one-sided test of 5% significance – corresponding to a minimum value of 1.66 (Huber et al. 2007: 35). 28
For a detailed description of explorative factor analysis cf. H€uttner and Schwarting (1999). The scales of the questionnaire applied in our survey have been developed for the purpose of this work and have not been utilized before (cf. Sect. 5.2.2).
29
84
3 Structural Equation Methodology
Another form of quality evaluation – assessed at the level of constructs, instead of indicators – is construct reliability which describes the degree of how well a latent variable is measured by its assigned indicator variables (Krafft et al. 2005: 74–75). It requires that indicators corresponding to the same construct have a strong inter-relation (G€otz and Liehr-Gobbers 2004: 727). To assess how well a construct is measured by its assigned indicator variables, a coefficient developed by Cronbach (1951: 297–299) called Cronbach’s alpha, as well as Fornell and Larcker’s (1981: 45) internal consistency can be applied (Kirstein 2009: 178; Krafft et al. 2005: 74). Cronbach’s alpha measures the reliability of a group of indicators that can be aggregated in one factor, thus reflecting the internal consistency of the indicators of a latent construct (Homburg and Giering 1996: 8). It is calculated as follows (Cronbach 1951: 299; Zinnbauer and Eberl 2005: 568):
a¼
P 2 s n 1 2i : sx n1
(3.13)
In (3.13), n represents the number of indicators, s2i the variance of the indicator i, and s2x the variance of the total test (Cronbach 1951: 299; Zinnbauer and Eberl 2005: 568). Here, the higher the covariances or correlations between the indicator variables, the more it converges towards +1 within its values range from 0 to þ1 (Homburg and Giering 1996: 8; Zinnbauer and Eberl 2005: 568). While some argue for a minimum value of Cronbach’s alpha of 0.4 (Peter 1997: 180) or 0.6 (Malhotra 1993: 308), usually the minimum value proposed by Nunally (1978: 245) of 0.7 (or 0.5 for first-time used scales) is applied to evaluate construct reliability (Homburg and Giering 1996: 8, 12; McNaughton 2002: 370). However, it should be noted that Cronbach’s alpha has been criticized for its increase with the number of indicators and for unrealistically assigning equal weights to the indicators in its calculation (Homburg and Giering 1996: 8; Kirstein 2009: 179–180; Krafft et al. 2005: 74). A solution to this criticism is provided by internal consistency,30 another criterion of construct reliability originally developed by Werts et al. (1974: 27–29), which recognizes the actual factor loadings in the quality evaluation (Kirstein 2009: 180). According to Fornell and Larcker (1981: 45), internal consistency in reflective measurement models is defined as follows (Chin 1998b: 320; Huber et al. 2007: 35; Krafft et al. 2005: 74; Zinnbauer and Eberl 2005: 568): P ð i lij Þ2 : Internal consistency ¼ P P ð i lij Þ2 þ i varðeij Þ
30
(3.14)
In literature, other frequently used terms for internal consistency include convergent validity, factor reliability, J€ oreskog’s (1971) rho, and composite reliability (Homburg and Giering 1996: 74).
3.2 The Partial Least Squares Approach to Structural Equation Modeling
85
The variables of (3.14) have been described above in light of equation (3.12). Within its value range from 0 to +1, Nunally (1978: 245) suggests a minimum value of 0.7 for internal consistency (G€ otz and Liehr-Gobbers 2004: 728). Even though others argue that lower minimum values of 0.6 would be acceptable (Bagozzi and Yi 1988: 82; Homburg and Baumgartner 1998: 361), given the importance of this quality criterion for construct reliability, we will utilize the higher value of 0.7 in this work. Those indicators that have a low correlation with the other indicators of the reflective measurement model, have to be eliminated (Krafft et al. 2005: 74). In addition to content validity, indicator and construct reliability, a complete validation process also requires the evaluation of discriminant validity (G€otz and Liehr-Gobbers 2004: 728). This refers to the degree by which measurements of distinct constructs differ from each other, utilizing the same measurement instrument (Krafft et al. 2005). Discriminant validity is assessed by the average variance extracted (AVE), which analyzes the amount of indicator variance explained by a factor in comparison to the amount of non-explained variance (Chin 1998b: 321; Fornell and Cha 1994: 69; Huber et al. 2007: 35–36). This is defined as follows (Fornell and Larcker 1981: 45–46; Krafft et al. 2005: 74; Ringle 2004: 24): P
l2i : AVE ¼ P 2 i P ð i li Þ þ i varðei Þ
(3.15)
The variables of (3.15) correspond to those described above in (3.12). Within its value range from 0 to þ1, the AVE should have a minimum value of 0.5, as otherwise the overwhelming part of variance would be attributed to the error term (Hildebrandt and Temme 2006: 625; Homburg and Giering 1996: 12; Krafft et al. 2005). In our previous explanations, we have presented the most important quality criteria for the evaluation of reflective measurement models, mainly according to Chin (1998b), Homburg and Giering (1996), Kirstein (2009), and Krafft et al. (2005). Here, it should be noted that – despite being methodologically intended by this work – a simultaneous fulfillment of all of those criteria is not necessary. Rather, the overall picture of the quality criteria is important, where only multiple and substantial violations of minimum quality levels should lead to modifications of reflective measurement models (Giering 2000: 90).
3.2.3.2
Quality Criteria for Evaluating Formative Measurement Models
The quality criteria of formative measurement models are distinguished from those of reflective measurement models particularly by the change of the causal direction between the latent variable and its manifest indicators (G€otz and Liehr-Gobbers 2004: 728). Here, the fact that formative indicator variables, in comparison to reflective ones, do not have to correlate substantially, leads to an exclusion of several quality criteria applying to reflective measurement models (e.g., indicator
Test of significance of weight loadings by (two-sided) t-values, requiring a 5% level with a minimum value 1.98 (based on bootstrapping method) Correlation matrix (pairwise) Correlation assumed if correlation coefficients 0.7 Tolerance Ti ¼ 1 R2i Multicollinearity expected if Ti < 0.1 Multicollinearity confirmed if Ti < 0.01 1 Variance Inflation factor VIFi ¼ 1R 2 i Multicollinearity if VIFq 10.0 i ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Proportion of substantive agreement psa ¼ nNc (high value within range from 0 to 1) 0 Substantive-validity coefficient csv ¼ nc n (high value within range N from 1 to 1) No elimination of indicators based on size of weights
Methods/criteria
max Condition index CIi ¼ Eigenvalue Eigenvaluei Moderate multicollinearity if Cli ranges from 10 to 30 High multicollinearity if Cli > 30 External/nomological Quality of construct Nomological Direction and size, as well as significance of relationship (i.e. validity measurement evaluation confirmation of already theoretically and/or empirically grounded relationships) Source: own illustration on the basis of Kirstein (2009: 188) and Krafft et al. (2005: 82)
Evaluation of multicollinearity
Table 3.3 Quality criteria for the evaluation of formative measurement models Quality criteria for formative measurement models Form of quality Definition Quality criteria evaluation Content/expert validity Degree of correspondence Expert validity between a priori intended and factual indicator allocation Interpretation of Indicator relevance Contribution of indicators weights to latent construct building Significance of weight loadings
86 3 Structural Equation Methodology
3.2 The Partial Least Squares Approach to Structural Equation Modeling
87
reliability, internal consistency) (Fornell and Larcker 1981: 46; Kirstein 2009: 183; Krafft 1999: 124–125; Krafft et al. 2005: 76; Rossiter 2002: 307–308). As illustrated in Table 3.3, for the evaluation of formative measurement models, three categories of quality criteria are distinguished (Kirstein 2009: 188; Krafft et al. 2005: 82). In conformity with reflective measurement models, content or expert validity is assessed in the pretest by interviewing test persons (Krafft et al. 2005: 76). This is particularly important for formative measurement models, as content validity has to be ensured in advance of data collection (G€ otz and Liehr-Gobbers 2004: 728). An elimination of indicators is still possible in this preliminary research phase (Helm 2005: 247–248), while modifications at a later stage are not recommended (Kirstein 2009: 183).31 As in the case of reflective measurement models, content or expert validity is measured by an indicator’s proportion of substantive agreement and by its substantive-validity coefficient (Krafft et al. 2005: 77). In addition to content or expert validity, indicator relevance is assessed by evaluating the contribution of each indicator to latent construct building (Krafft et al. 2005: 77). This interpretation of indicator weights is different from the evaluation of factor loadings, as the values of the former are usually below those of the latter (Krafft et al. 2005: 78). Consequently, relatively low values of indicator weights should not be misinterpreted as a poor formative measurement model (Chin 1998b: 307; G€otz and Liehr-Gobbers 2004: 729). However, some argue that weights below 0.1 contribute only marginally to the explanation of latent variables (Lohm€oller 1989: 60), and thus should be eliminated (Seltin and Keeves 1994: 4356), as they may produce “[. . .] more noise than information [. . .]” (J€oreskog and Wold 1982: 270). Others refuse such an elimination of indicators to improve the reliability of formative measurement models, as this represents a trial-and-error process which – despite not being based on theory – would change the conceptual content of latent variables (Bollen and Ting 2000: 4; Huber et al. 2007: 38; Jarvis et al. 2003: 202; Rossiter 2002: 315). Following this argumentation, we will not eliminate formative indicators based on the size of their weights, as even marginally weighted indicators may contribute highly to the meaning of latent constructs (Helm 2005: 251–252). Therefore, in the course of assessing indicator relevance, rather than the size of weights, their significance is important (Huber et al. 2007: 38). Here, corresponding to reflective measurement models, the bootstrap method is applied as the preferred resampling procedure for formative measurement models (Efron and Gong 1983: 39–40; Herrmann et al. 2006: 57).32 The t-values should be significant in a two-sided test at the 5% level, which corresponds to a minimum value of 1.98 utilizing the bootstrap method (Huber et al. 2007: 45). Although this minimum value might serve in this work as an orientation to assess the relevance of
31
The reasoning for avoiding the elimination of indicators in formative measurement models has been explained in Sect. 3.2.2.2. 32 Cf. Sect. 3.2.3.1.
88
3 Structural Equation Methodology
formative indicators, their elimination should not be divorced from conceptual considerations (Diamantopoulos and Winklhofer 2001: 273; G€otz and LiehrGobbers 2004: 730). The elimination of formative indicators from the measurement model, however, is encouraged in the case of multicollinearity, meaning the degree of their linear dependence of each other (Krafft et al. 2005: 78). This is due to the fact that a substantial multicollinearity would lead to enormous biases in parameter estimations, which do not allow the isolation of the singular influence of an indicator in the measurement model (Diamantopoulos and Winklhofer 2001: 272; G€otz and LiehrGobbers 2004: 729). As formative measurement models ground on the principle of multiple regression analysis, the standard errors of beta coefficients increase with higher degrees of multicollinearity – which results in an increasing inefficiency of parameter estimations (Backhaus et al. 2008: 88; Krafft et al. 2005: 78). Given the importance of avoiding multicollinearity in formative measurement models, various quality criteria for its evaluation will be utilized in this work. First, it can be assessed by analyzing the correlation matrix of formative indicators, where substantial pairwise linear dependencies – expressed by high correlation coefficients close to the maximum value of 1 – reflect a high multicollinearity (Backhaus et al. 2008: 89; Eggert and Fassott 2005: 40).33 Regarding an appropriate cut-off point for assuming multicollinearity, simulations have shown that an indicator correlation of 0.7 still delivers robust results close to the “true” values in the PLS path algorithm, while some biases already exist (Cassel et al. 1999: 443–444). Therefore, correlation coefficients should be below the value of 0.7 in this work, to avoid biases from multicollinearity. As correlation coefficients, however, only measure pairwise dependencies, despite low values for the correlation coefficients of independent indicator variables, substantial multicollinearity may still exist (Backhaus et al. 2008: 89). Consequently, a regression of each independent indicator variable xi with the other independent variables is recommended to uncover multicollinearity, based on the respective multiple correlation coefficient or the coefficient of determination R2i (Backhaus et al. 2008: 89). Only if this coefficient of determination R2i is close to the value of zero, can a linear independence of indicator variables be assumed (Krafft et al. 2005: 79). The complementary logic of this procedure corresponds to the tolerance of indicator variables, a second approach to evaluate their multicollinearity (Backhaus et al. 2008: 89): Ti ¼ 1 R2i
(3.16)
Multicollinearity is rejected if Ti has values close to 1 (Krafft et al. 2005: 79), while multicollinearity may be expected at values <0.1 and definitely confirmed at values <0.01 (Hair et al. 2006: 230; Kleinbaum et al. 1988: 214).
33
The correlation coefficient is calculated by dividing the covariance of two metric formative indicators by the product of their standard deviations (Schulze 1998: 128).
3.2 The Partial Least Squares Approach to Structural Equation Modeling
89
A third measure of multicollinearity is the reciprocal value of tolerance, the socalled variance inflation factor (VIF) (Backhaus et al. 2008: 89): VIFi ¼
1 : 1 R2i
(3.17)
The minimum value of the VIFi is 1, which corresponds to a tolerance value of 1 – thus reflecting a complete linear independence of the formative indicators (Krafft et al. 2005: 79). VIFi values above 1 show by which factor the variances of the respective indicators increase, where values of 10 should not be exceeded (Huber et al. 2007: 39; Krafft et al. 2005: 79). A fourth quality criterion to evaluate multicollinearity is Belsey et al. (1980: 117–118) and Belsey et al. (2004: 104 –105) condition index (CI) (G€otz and LiehrGobbers 2004: 729; Krafft et al. 2005: 79): sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Eigenvaluemax : CIi ¼ Eigenvaluei
(3.18)
In (3.18), Eigenvaluemax represents the maximum eigenvalue in the estimation procedure, while Eigenvaluei is the eigenvalue of the variance-covariance matrix of unstandardized regression coefficients between the indicators and the latent variable (Krafft et al. 2005: 79). Multicollinearity increases with higher CIi values, where values between 10 and 30 reflect moderate multicollinearity and values above 30 a substantial multicollinearity (Krafft et al. 2005: 79–80). In addition to multicollinearity, the formative measurement model should be evaluated regarding its external or nomological validity (G€otz and Liehr-Gobbers 2004: 729; Reinartz et al. 2004: 298–299).34 An external or nomological validation of formative measurement models is necessary, as their underlying assumption – that an error-free measurement of all facets of the latent variable is possible – may not always hold (G€ otz and Liehr-Gobbers 2004: 719). It is assessed either by Hauser and Goldberger’s (1971: 81–82) so-called MIMIC (multiple indicators and multiple causes) model, a two-construct model, and/or by nomological validity (Krafft et al. 2005: 80–82), where the latter will be applied in this work.35 Nomological validity is 34
The external or nomological validity has been introduced to evaluate formative measurement models, as individual item reliability and convergent validity of reflective measurement models are irrelevant here, given that the latent variable is modeled as an effect instead of a cause of the manifest indicator variables (G€ otz and Liehr-Gobbers 2004: 729; Hulland 1999: 201). 35 The MIMIC model allows a single latent variable to be measured by both formative and reflective indicators (Krafft et al. 2005: 80). As the MIMIC model is currently not supported by all PLS techniques (e.g., SmartPLS), a two-construct model is often applied to assess the measurement errors (Diamantopoulos and Winklhofer 2001: 272–273; G€otz and Liehr-Gobbers 2004: 720). Here, for the external validation of the formative measurement model, an additional reflective measurement model (phantom model) is designed to specify a phantom variable (G€otz and Liehr-Gobbers 2004: 720; Rindskopf 1984: 38). External validity is confirmed if a strong and
90
3 Structural Equation Methodology
assessed by analyzing the relationships between the formatively operationalized latent constructs and other latent variables in the structural model (Eggert and Fassott 2005: 41; Krafft et al. 2005: 82). Nomological validity is given if the direction and size, as well as the significance of these relationships correspond to their respective theoretically and/or empirically derived values (Diamantopoulos and Winklhofer 2001: 273; Eggert and Fassott 2005: 41; Krafft et al. 2005: 82). For the operationalization of formative indicators, this work will follow a threestep procedure. First, the content of the formative measurement model will be specified for its respective latent variable, including a complete description and definition of the construct (G€ otz and Liehr-Gobbers 2004: 719). Second, formative indicators will be derived that are assumed – on the basis of theoretical considerations – to have an important influence on the latent constructs (G€otz and LiehrGobbers 2004: 719). A part of these two steps is already included in the theoretical foundation in Sect. 2.2.7 and will be completed in Sect. 4.1.1 of this work. In a third step, the formative measurement models will be evaluated according to their quality criteria in Table 3.3, mainly in Sect. 6.3 of this work. 3.2.3.3
Quality Criteria for Evaluating a Structural Model
For the evaluation of the structural model – instead of inference-statistical tests of covariance-based techniques – the PLS approach applies only non-parametric procedures (i.e., resampling procedures, such as bootstrapping), due to its less restrictive assumptions (Chin 1998b: 314–316, 318–320; Krafft et al. 2005: 83).36 Following this methodological direction, as illustrated in Table 3.4, three categories of quality criteria are distinguished for the evaluation of the basic structural model (Kirstein 2009: 192; Krafft et al. 2005: 85). A first orientation in the evaluation of the basic inner model is given by the coefficient of determination (R2) of the endogenous variables, based on the regressions in the structural model (Krafft et al. 2005: 83). The coefficient of determination measures the degree of alignment of the regression function to the empirical data by relating the portion of explained variance to total variance (Backhaus et al. 2008: 67–69). Even though higher values within the range from 0 to þ1 might significant relationship exists between the formatively operationalized latent construct and the reflectively measured, latent phantom variable (Krafft et al. 2005: 82; Rindskopf 1984: 38). If, however, no reflective indicators are available to specify the latent phantom variable, nomological validity may be assessed alternatively (Diamantopoulos and Winklhofer 2001: 273; Eggert and Fassott 2005: 41; G€otz and Liehr-Gobbers 2004: 730). No reflective indicator variables were added in the survey of this work – that would correspond to the formatively measured, latent constructs – to achieve a sufficient response rate based on a minimum amount of questions (cf. Sect. 5.2.2). 36 The less restrictive assumptions of PLS versus the covariance-based approach LISREL have been explained in Sect. 3.1. It should be noted here that even though a parametric approach – which recognizes the classical assumptions, particularly regarding the distribution of residuals – would allow an estimation of the standard errors of the parameters, this does not correspond to the PLS philosophy of soft modeling (Balderjahn 1986: 147; Chin 1998b: 315; Chin and Newsted 1999: 324; Lohm€oller 1989: 64), and thus will not be considered in this work.
Second order latent constructs
Interaction variables: Substantial explanatory contribution of interaction variable
Interaction variables: Interaction relevance
Source: own illustration on the basis of Kirstein (2009: 192) and Krafft et al. (2005: 85)
Additional quality criteria for more complex, extended structural models
R2included R2excluded 1R2included
>0
Test of significance of path coefficients by (twosided) t-values, requiring a 5% level with a minimum value 1.98 (based on bootstrapping method) P
Small effect if 0.02 f2 < 0.15 Medium effect if 0.15 f2 < 0.35 Large effect if f2 0.35 Signs of path coefficients Size of path coefficients >0.1
f2 ¼
Methods/criteria R2 0.4
Ejk Stone-GeisserQ2j ¼ 1 P k O >0 jk test k Relevance of interaction Path coefficients Analogous to the basic structural model variables for the basic and structural model significance R2withinteractionvariable R2maineffectsmodel Substantial influence of the Interaction effect f 2 >0 interaction ¼ 1R2maineffectsmodel interaction variable on the size Small effect if 0.02 f2 < 0.15 endogenous variable Medium effect if 0.15 f2 < 0.35 Large effect if f2 0.35 [Values should be interpreted jointly with beta changes] Evaluation of second order latent Interpretation Application of the same criteria as above for the constructs in the structural like first basic structural model, and of the quality model order criteria of Tables 3.2 and 3.3 for the constructs measurement models of first order dimensions
Structural model alignment to the Interpretation empirical data of path coefficients Reliability of parameter estimations
Prediction relevance
Quality criteria Interpretation like multiple regression Effect size
Substantial influence of the exogenous on the endogenous variable
Definition Portion of explained construct variance
Substantial explanatory contribution
Quality criteria for the structural model Form of quality evaluation Quality criteria for the basic Coefficient of determination structural model
Table 3.4 Quality criteria for the evaluation of the structural model
3.2 The Partial Least Squares Approach to Structural Equation Modeling 91
92
3 Structural Equation Methodology
indicate a better model fit, no universally valid cut-off point for a sufficient R2 exists (Backhaus et al. 2008: 70, 93; Krafft et al. 2005: 83). This depends rather on the underlying problem set, where for example a R2 value of 0.1 may be acceptable in highly random processes (e.g., weather, stock market) (Backhaus et al. 2008: 94). According to Chin (1998b: 323), R2 values of 0.19, 0.33, and 0.67 are perceived as weak, moderate, and substantial, respectively. In the current PLS research, values greater than 0.3 (Herrmann et al. 2006: 61) and 0.4 are considered as acceptable (Kirstein 2009: 189; Ringle 2004: 19) – where the latter, more conservative cut-off point will be applied in this work. In addition, the basic structural model may be evaluated by its effect size that – corresponding to the partial f-test – allows the assessment of whether the impact of a particular independent latent construct on a dependent latent variable has a substantive effect (Chin 1998b: 316; Cohen 1988: 410–413; Huber et al. 2007: 46; Krafft et al. 2005: 84): f2 ¼
R2included R2excluded : 1 R2included
(3.19)
The effect size f 2 is calculated by the change of the coefficient of determination R for the dependent latent variable, if the exogenous latent construct is used (R2included ) or omitted (R2excluded ) in the structural equation (Chin 1998b: 317; Huber et al. 2007: 46). Here, f 2 should be above zero, where values of 0.02, 0.15, and 0.35 indicate whether an exogenous latent variable has a small, medium, or large effect on the endogenous latent variable at the structural level (Chin 1998b: 317; Cohen 1988: 413; Huber et al. 2007: 46; Krafft et al. 2005: 84–85). Besides looking differently at the coefficient of determination, the prediction relevance of the basic structural equation model can be assessed by the signs, size, and significance of path coefficients (Chin 1998b: 316; Krafft et al. 2005: 83). Here, those paths that are not significant, or that have contrary signs to the formulated hypotheses, disprove these hypothesized relationships – while significant paths with the a priori postulated signs empirically prove the hypotheses on these relationships (G€ otz and Liehr-Gobbers 2004: 730).37 The interpretation of path coefficients thus includes their signs, in dependence with the hypothesized relationships, and their size – which should be at least above 0.1 (Lohm€oller 1989: 60–61; Ringle 2004: 19). Furthermore, corresponding to reflective and formative measurement models, the reliability of parameter estimations is assessed by their significance based on the bootstrap method as the preferred resampling procedure (Efron and Gong 1983: 39–40; Efron and Tibshirani 1993: 145–146; Herrmann et al. 2006: 58; Huber et al. 2007: 35; Krafft et al. 2005: 83).38 Here, in a two-sided test the t-values should be significant at the 5% level, corresponding to a minimum value of 1.98 utilizing the bootstrap method (Huber et al. 2007: 45). 2
37
In Sect. 4.1, the hypotheses of this work will be presented and explained in detail. Cf. Sect. 3.2.3.1.
38
3.2 The Partial Least Squares Approach to Structural Equation Modeling
93
In addition to the above described analysis of path coefficients, prediction relevance may be evaluated by the non-parametric Stone-Geisser-test (Chin 1998b: 317; Fornell and Cha 1994: 71–73; Geisser 1975: 320–322; Huber et al. 2007: 46; Krafft et al. 2005: 84; Stone 1974). This test follows the so-called blindfolding procedure, which omits a part of the data for a particular block of indicators during the parameter estimation process, and then tries to estimate the omitted data on the basis of the estimated parameters (Chin 1998b: 317; G€otz and Liehr-Gobbers 2004: 731). This procedure of omitting and reconstructing is repeated until every data point is omitted and estimated once, which results in a generalized cross-validation measure and jackknife standard deviations of parameter estimates (Chin 1998b: 317; Fornell and Cha 1994: 71; Krafft et al. 2005: 84). This is assessed by the StoneGeisser-test criterion Q2j which measures the alignment of the structural model to the empirical data, thus indicating how well the empirically derived data can be reconstructed by the model and its parameter estimates (Chin 1998b: 318; Fornell and Cha 1994: 72; G€ otz and Liehr-Gobbers 2004: 731): Q2j
P Ejk P ¼1 k : O k jk
(3.20)
In (3.20), Ejk represents the sum of squares of prediction errors, and Ojk the sum of squares of the difference between the estimated parameter value and the mean value of the remaining data of the blindfolding procedure (Chin 1998b: 317; Krafft et al. 2005: 85). The index j represents the examined, endogenous measurement model, while k corresponds to the index for all indicators of the measurement model (G€otz and Liehr-Gobbers 2004: 731). A value of Q2j > 0 confirms the predictive relevance of the model, whereas a value of Q2j < 0 represents a lack of predictive relevance (Chin 1998b: 318; Fornell and Cha 1994: 73; Herrmann et al. 2006: 58; Krafft et al. 2005: 85). For the evaluation of more complex structural models with interaction variables and/or second-order constructs, additional quality criteria apply. In the case of an extension of the basic structural model by interaction variables, their interaction relevance should be evaluated by their path coefficients and their significance (G€otz and Liehr-Gobbers 2004: 727), representing analogous quality criteria to the basic structural model evaluation. In addition, their substantial explanatory contribution is assessed by the interaction effect size39 which measures the change of the coefficient of determination (R2) if – in addition to the exogenous variable and the moderator variable (main effects model) – the interaction variable is utilized to explain the variance of the endogenous variable (Chin et al. 2003: 195–196; Eggert et al. 2005: 109; G€ otz and Liehr-Gobbers 2004: 727):
39
In the literature, this measure of the effect size of the interaction term is usually termed solely “effect size” (Chin et al. 2003: 195–196; Eggert et al. 2005: 109). To distinguish this measure of the effect size of interaction variables from the effect size of the exogenous variables in the basic structural model, we utilize the term “interaction effect size”.
94
3 Structural Equation Methodology
2 finteraction ¼
R2with interaction variable R2main effects model 1 R2main effects model
:
(3.21)
2 is calculated by the change of the In (3.21), the interaction effect size finteraction 2 coefficient of determination R for the endogenous latent variable, if the interaction variable is used ðR2with interaction variable Þ or omitted ðR2main effects model Þ in the structural equation (Chin 1998b: 317; Huber et al. 2007: 46). Here, analogous to the effect 2 should be above zero, while values of size of the basic structural model, finteraction 0.02, 0.15, and 0.35 indicate that an interaction variable has a small, medium, or large effect on the endogenous latent variable at the structural level (Eggert et al. 2005: 110). Here, if the resulting beta changes are meaningful, even a small 2 can be meaningful (Chin et al. 2003; Eggert et al. interaction effect size finteraction 2005: 110). The betas or path coefficients of interaction variables describe, to what extent the influence of the exogenous variable on the endogenous construct changes, if a change in the moderating variable occurs (Eggert et al. 2005: 109; G€otz and Liehr-Gobbers 2004: 724). Another complexity that might be added to the basic structural model, are second-order constructs that – by treating their different dimensions as separate constructs – are evaluated by following the same process that is applied to examine the validity of latent first-order constructs (Chin 1998a: x; Hulland 1999: 196–197; Maloney 2007: 259–260).40 Here, the formative/reflective question is reapplied at this higher order level, where a formative specification between first-order constructs and its respective second-order construct is most meaningful, as its dimensions clearly describe separable facets of the construct (Albers and G€otz 2006: 672–673; Chin 1998a: x; Maloney 2007: 259).41 Following this formative specification between first-order and second-order constructs in this work, the relationship of latent second-order variables with other latent variables in the structural model will be analyzed according to the quality criteria in Table 3.4, while the measurement models of the first-order dimensions will be evaluated based on the quality criteria of Tables 3.2 and 3.3. After having presented the structural equation methodology and the PLS approach as an adequate SEM technique, our explanations above have laid the methodological groundwork for a proper conceptualization and evaluation of both structural models and measurement models. In the following, we will apply this methodological basis to the establishment of a regional success factor model, to empirically test the success factors of regional strategies
40
For a detailed description of the sequences of this process cf. Chin (1998a: x), Giere et al. (2006: 683–689), and G€otz and Liehr-Gobbers (2004: 725–727). 41 A reflective relationship between the dimensions and the second-order construct would imply that the dimensions are interchangeable measures of the second-order construct, and thus are not inseparable – calling into question the necessity of a second-order construct (Albers and G€otz 2006: 672–673; Maloney 2007: 258–259).
Chapter 4
Regional Success Factor Model
4.1
Model of Success Factors of Regional Strategies
As we have previously illustrated, in the theoretical foundation of this work,1 the regional competitiveness of MNCs is influenced by the regional environmentstrategy-structure relationship. As illustrated in Fig. 2.5, this consists of the strategic decision-making of regional management and of the regional products/services resulting from the development of FSAs according to the demands of the regional environment. These region-bound company strengths are factors of regional management and regional product/service configurations that: “[. . .] can contribute to survival, profitability, and growth beyond the geographic scope of a single nation [. . .]” (Rugman 2005b: 50). This relationship of the organizational and strategic effectiveness of MNCs with their regional corporate success has been described to be contingent on contextual influences. More specifically, the regional success factors of MNCs – given by appropriate degrees of managerial decision-making autonomy and regional product/service adaptation – are highly internally contingent upon their regional orientation and highly externally contingent upon the interregional distance they face. By transferring this theoretically derived understanding of regional competitiveness to the SEM methodology, a structural model of success factors of regional strategies results, as illustrated in Fig. 4.1. Here, two types of relationships between the latent variables in Fig. 4.1 can be distinguished. First, as illustrated by the hypotheses H1 and H2, direct effects exist of the exogenous variables, regional management autonomy and regional product/ service adaptation, on the endogenous construct regional success (basic structural model). Second, moderating effects are hypothesized to emanate from the latent contingency variables regional orientation (H3a and H3b) and inter-regional distance (H4a and H4b) on the relationships of the basic structural model (extended structural model). Therefore in the following, we will first elaborate a measurement model for the exogenous latent constructs of the basic inner model, regional management
1
Cf. Chap. 2.
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_4, # Springer-Verlag Berlin Heidelberg 2011
95
96
4 Regional Success Factor Model
Regional orientation
H3a
Regional management autonomy
H3b
H1
Regional success H2
Regional product /service adaptation
H4a
H4b
Inter-regional distance
Fig. 4.1 Structural model of success factors of regional strategies Source: own illustration
autonomy and regional product/service adaptation. Second, we will develop an outer model for the exogenous latent variables of the extended structural model, regional orientation and inter-regional distance. Third, we will derive the measurement model for the endogenous variable regional corporate success. Fourth, we will introduce appropriate control variables, to control for additional influences on regional corporate success besides those of the latent constructs of the basic and the extended structural model. Along these steps, we will theoretically derive the central research hypotheses H1–H4b for the latent variable relationships in this regional success factor model.
4.1.1
Regional Management Autonomy
The total degree of regional management autonomy within a MNC, applying the definition of regional management in Sect. 2.2.7, consists of different dimensions of managerial decision-making autonomy of regional management centers, which may be granted in the areas of regional strategy development, regional market and product/service development, regional market coordination, and regional operational administration. Conversely, the degree of decision-making autonomy in each of these dimensions of regional managerial activities influences the overall extent to which regional management can autonomously decide at the regional level (Ohmae 1985: 185–188). Drawing on this rationale, a formative direction of
4.1 Model of Success Factors of Regional Strategies
97
Strategy formulation Competitor intelligence Business model development Budgeting, capital allocation, control
Regional strategy development
Regional orientation
Marketing planning and execution Sales planning and execution Customer service and support
Regional market development
R&D of products or service development Coordination of distribution Coordination of production/logistics
Regional market coordination
H3a
Regional management autonomy
H3b
H1
Regional success
Coordination of legal entities/operations Monitoring of legal entities/operations Liaison center for the parent company Reporting to the parent headquarters
Regional operational administration
Senior personnel resource management
H2
Regional product /service adaptation
H4a
H4b
Support of operations
Inter-regional distance
Fig. 4.2 Measurement model of the multidimensional construct regional management autonomy Source: own illustration
causality exists between the four dimensions and regional management autonomy.2 Given that these four dimensions are formulated at a similar level of abstraction and cannot be observed directly (Giere et al. 2006: 679), they represent latent first-order variables in the nomological network of the latent second-order construct regional management autonomy. Consequently, as illustrated in Fig. 4.2, the multidimensional latent construct regional management autonomy is measured at the level of the latent first-order variables. In the following, according to the phases recommended by Giere et al. (2006: 683) for the conceptualization of multidimensional constructs, we will first describe the separate dimensions of regional management autonomy and their indicators, before – in a second step – specifying the rules of correspondence between each dimension and its respective manifest indicator variables. In a third step, we will present our research hypothesis H1 and our reasoning for its hypothesized relationship between regional management autonomy and regional success.
4.1.1.1
Regional Strategy Development
One dimension of the latent construct regional management autonomy is regional strategy development, the extent to which regional management is responsible for the development of strategies at the regional level (Sch€utte 1997: 441; Yeung et al. 2001: 165). This responsibility relates particularly to managerial decision-making 2
As explained in Sect. 3.2.3.3, a formative specification between first-order constructs and secondorder variables is generally more meaningful.
98
4 Regional Success Factor Model
(Bartlett and Ghoshal 1989: 200; Rugman 2005b: 49; Rugman and Verbeke 2006: 124–125), where higher regional management autonomy for regional strategy development leads to lower, respective corporate and/or national decision-making autonomy, and vice versa. Thus, while some MNCs might have a strong home base for strategic development at the corporate headquarters (Rugman 2005b: 161), others might strongly rely on their national subsidiaries for strategy development (Enright and Subramanian 2007: 919; Hamel and Prahalad 1983: 343–344; Manea and Pearce 2006: 242–245), where the intermediate case – in the form of regional strategy development – leads to increases of regional management autonomy.3 This type of area coordination at the regional level leads to aggregation benefits by grouping strategic development processes, and thereby reducing the complexity of national decision-making in numerous countries (Arregle et al. 2009: 103; Ghislanzoni et al. 2008: 76; Vernon and Wells 1986: 31). Consequently, regional strategy development can be understood as the attempt of a MNC to build and manage the most appropriate decision process for geographically distinct issues of strategy – one that senses and responds to multiple changing environmental demands within each particular region (Bartlett and Ghoshal 1989: 201; Yeung et al. 2001: 165). Thus, regional strategy development refers to the establishment of a strategic architecture to guide the MNC in developing regionally relevant core competencies (Prahalad and Hamel 1990: 89; Sch€ utte 1997: 441; Verbeke 2009: 79). The development of these regional core competencies is dependent upon the degree of managerial decision-making autonomy with respect to four indicators of regional strategy development that – as illustrated in Fig. 4.2 – include regional strategy formulation, regional competitor intelligence, regional business model development, as well as regional budgeting, capital allocation, and control over profit/loss (Enright 2005a: 68; Enright 2005b: 100; Sch€utte 1997: 441; Williams 1967: 89). Regional strategy formulation refers to the degree of involvement of regional management in the phrasing of the regionally relevant strategy, thus the extent to which regional management can shape and stimulate the strategy formulated for a certain region (Lasserre and Probert 1998: 49; Sch€utte 1997: 441). Regional competitor intelligence relates to activities of regional management to gather information about competitors, to develop regionally unique core competencies that increase the MNC’s competitiveness (Enright 2005a: 62; Lasserre and Probert 1998: 49; Sch€ utte 1997: 441; Yeung et al. 2001: 160). Consequently, the familiarity of MNCs with local and regional competitors is one important condition to become a performing insider (Ohmae 1985: 206; Rugman 2005b: 66). The term regional business model development is derived from Ohmae’s (1985: 114) claim to set up an overseas business like a new company with a novel business strategy.4 This strategic commitment to the region leads to an insider position, and thus
3
These relationships illustrate the sphere of the multidimensional construct (Giere et al. 2006: 683), which is comprised of causal (formative) effects between the dimension regional strategy development and its second-order construct regional management autonomy. 4 Cf. Sect. 2.1.2.3.
4.1 Model of Success Factors of Regional Strategies
99
involves the delegation of important decisions to the region (Ohmae 1985: 114, 206; Sch€utte 1997: 440) – for example regarding regional products/services, site locations, sourcing, and/or sales channels. Thus, in contrast to regional strategy formulation, which solely relates to the degree of participation of regional management in the development of regional strategies, a regional business model development leads to a much higher decision-making autonomy of regional management covering many different strategy-related elements along the MNC’s value chain. The necessity of firms to adapt their business model when entering new regions has been highlighted by Rugman (2005b: 186, 189, 191) for different MNCs from various industries.5 Furthermore, for regional management to be effective in strategy development, a delegation of decision-making should be coupled with the responsibility for regional budgeting, capital allocation, and control over profit/ loss for the group of operations/entities (e.g., national subsidiaries) within the region (Morrison and Roth 1992: 52; Sch€ utte 1997: 441; Stopford and Wells 1972: 50). The responsibility of regional management for a region-wide capital budget allocation implies a regular performance evaluation of regional business operations (Hinrichs 2008: 109; Rugman 2005b: 198).
4.1.1.2
Regional Market and Product/Service Development
Another dimension of the latent variable regional management autonomy is regional market and product/service development, referring to the degree of decision-making autonomy of regional management in the development of regional markets and products/services (Rugman and Verbeke 2006: 125; Rugman and Verbeke 2008a: 310).6 Regional market development, or regional strategic market planning, relates to the development and implementation of a regional market approach by regional management – that systematically and periodically examines the regional environment, competitive opportunities and threats, the corporate goals and objectives to be achieved, and the products/services offered – to profitably exploit the MNC’s capabilities and resources in a competitive advantage that addresses the threats and opportunities arising from a changing regional environment (De Bu´rca et al. 2004: 464). Regional product/service development is closely intertwined with regional market development, as the development of regional products/services should be market driven, thus should be based upon new developments and opportunities in the regional environment (Rugman 2003a: 8–9; Rugman and Verbeke 2008a: 402–403; Verbeke 2009: 177, 182–183). The extent to which regional management can autonomously decide on regional market and product/service developments is indicated by four manifest variables – as 5
The firms include Wal-Mart from the retail industry, LaFarge of the cement industry, L’Ore´al from the cosmetics industry, and McDonald’s of the food services industry (Rugman 2005b: 186, 189, 191). 6 In Fig. 4.2, we used “regional market development” as an abbreviation for “regional market and product/service development”.
100
4 Regional Success Factor Model
illustrated in Fig. 4.2 – regional marketing planning and execution, regional sales planning and execution, regional customer service and support, and regional R&D of products, or respectively, the regional development of services (Enright 2005a: 67–68; Enright 2005b: 100; Rugman 2005b: 162; Williams 1967: 89). Regional marketing planning and execution relates to the analysis of consumer needs and market segments, to properly develop and offer those regional products/ services that lead to a competitive advantage in regional markets (De Bu´rca et al. 2004: 488–489). Rugman (2005b: 43, 45, 106, 117, 120, 122, 124, 134, 136, 162, 168, 185, 188) shows for several MNCs that,7 to achieve regional responsiveness, decisions about regional marketing planning and execution should be delegated to regional management. This involves location-driven adaptation costs to build these regionally-responsive, managerial competencies in the form of regional marketing capabilities, which are important FSAs of the MNC (Rugman 2005b: 34, 226). Regional sales planning and execution refers to the analysis of the current sales situation at the regional level, the setting of regional sales-related objectives, and the forecasting on how these business objectives may be executed (Sheldon 2006: 65–67). MNCs may assign substantial decision-making autonomy to regional management for regional sales planning and execution, to increase regional sales – e.g., by tailoring products/services to the requirements of regional markets (Rugman 2005b: 206; Rugman and Verbeke 2008b: 310). Another regional management activity that contributes to regional responsiveness is regional customer service and support, relating to the regional service and support activities provided by MNCs to their customers (De Bu´rca et al. 2004: 276, 306; Onkvisit and Shaw 2009: 362; Roth and Morrison 1990: 557). The regional R&D of products or the regional development of services refers to varying degrees of decision-making autonomy of regional management for product-related R&D, or respectively, for the development of services (Daniels 1987: 413; Grosse 2005: 142; Onkvisit and Shaw 2009: 366; Rugman and Verbeke 2006: 125; Verbeke 2009: 186–187).
4.1.1.3
Regional Market Coordination
A further, latent first-order variable of the second-order construct regional management autonomy is regional market coordination, referring to the delegation of decision-making to regional management for the coordination of market-related regional activities across borders and business divisions (Enright 2005a: 62; Sch€utte 1997: 441). In the course of this regional market coordination, regional management realizes aggregation advantages by ensuring that regional businesses exploit synergies and follow consistent policies and mutually coherent market approaches across the region (Lasserre 1996: 31–32; Yeung et al. 2001: 166). 7 The MNCs include Toyota, Citibank, Eli Lilly, Procter and Gamble, L’Ore´al, Canon, Coca-Cola, and several firms from the chemical and the pharmaceutical industry (Rugman 2005b: 43, 45, 106, 117, 120, 122, 124, 134, 136, 162, 168, 185, 188).
4.1 Model of Success Factors of Regional Strategies
101
The degree of decision-making autonomy assigned to regional management for regional market coordination is dependent on four indicators – as illustrated in Fig. 4.2 – the coordination of regional distribution, the coordination of regional production/ logistics, and the coordination and monitoring of regional legal entities and operations – e.g., national subsidiaries (Enright 2005a: 68; Enright 2005b: 100; Lasserre 1996: 36; Rugman and Verbeke 2006: 125; Sch€utte 1997: 441; Williams 1967: 89). The coordination of regional distribution refers to the organization of effective distribution channels by regional management, which link the MNC to its customers, where the effectiveness of distribution depends on the channels’ contribution to a better availability, processing, and information content of products/services (De Bu´rca et al. 2004: 390). Regional distribution is a fundamental capability, or FSA, of MNCs (Rugman 2005b: 59, 34; Rugman and Verbeke 2004: 4) to improve their downstream competitiveness, as it represents an important source of differentiating their products/services (Ohmae 1985: 32; Verbeke 2009: 298). Consequently, decision-making autonomy for the coordination of regional distribution is a crucial function of regional management, as it reflects its critical responsibility over realizing a competitive insider position in the region (Ohmae 1985: 112; Roth and Morrison 1992: 483; Rugman 2005b: 59; Rugman and Verbeke 2004: 4; Rugman and Verbeke 2006: 125; Stopford and Wells 1972: 81). Even though Enright’s (2005a: 67, 2005b: 96) respondents have not perceived the coordination of regional production/logistics as being highly important activities at regional headquarters, the frequent use of regional production and logistics hubs and manufacturing-related techniques8 – to coordinate reductions in production, logistics and transport costs (Buckley and Ghauri 2004: 87–88; Verbeke 2009: 205–206) – qualifies for adding this management function to the indicators of regional market coordination. The decision-making autonomy of regional management for the coordination of regional production and/or logistics implies the responsibility for achieving significant regional aggregation and arbitrage advantages, as these back-end functions offer considerable economies of scale and scope to the MNC (Ghemawat 2007a: 60; Rugman 2005b: 77–78). The coordination of regional legal entities and operations – including for example national subsidiaries or affiliates, sales offices, agencies, dealerships, franchises, joint ventures, consortia, or partnerships – particularly regarding their achievement of a deep regional market penetration, is important to become an effective insider (Ohmae 1985: 171–180; Rugman 2005b: 233). Here, making a joint venture or any of the other myriad forms of regional operations a success, may take as much effort as establishing a fully controlled national subsidiary or building a new greenfield plant (Ohmae 1985: 177). Thus, the progress of regional operations towards an effective insider position should be regularly evaluated and controlled by a continuous monitoring of regional legal entities and operations (Enright 2005a: 66; Enright 2005b: 100). The importance of these monitoring activities of regional management are further
8
Examples of production-related techniques include flexible manufacturing systems, just-in-time manufacturing, and total quality management (Verbeke 2009: 205–206).
102
4 Regional Success Factor Model
enhanced, as regional legal entities and operations should comply with regional laws and policies – which, increasingly, are enforced by regional agreements, such as EU or NAFTA, that have shifted regulatory authority to the regional level (Rugman 2005b: 53; Rugman and Brain 2004: 19; Rugman and Kirton 1998: 439; Rugman et al. 1999: 235–237). Consequently, senior executives of many MNCs state the need for environmental surveillance activities, relating to an active monitoring of business and market environments at the regional level (Bouquet et al. 2009: 110).
4.1.1.4
Regional Operational Administration
An additional dimension of the latent construct regional management autonomy is regional operational administration, which relates to the responsibility of regional management to support regional operations in the organization of their interfaces – either between them or vis-a`-vis their regional environment – as well as to pool resources, to provide shared services or commonly required expertise to national subsidiaries and other regional operations (Lasserre 1996: 33; Ohmae 1985: 206). Any forms of this pooling of resources lead to aggregation advantages that, by avoiding the duplication of efforts, raise the efficiency and effectiveness of a MNC’s regional operations (Sch€ utte 1997: 441). The latent second-order variable regional operational administration – as illustrated in Fig. 4.2 – is comprised of four indicators, a regional liaison center for the parent company, the reporting on regional activities to the parent headquarters, regional senior personnel resource management, and the support of regional operations (Enright 2005a: 68; Enright 2005b: 100; Lasserre 1996: 34; Williams 1967: 89). A regional liaison center for the parent company acts as a translator between the strategies of corporate headquarters and national markets (Lasserre 1996: 31–32), thus involving, for example, communication, public relations and/or governmental relations in accordance with the parent headquarters. These activities of a regional liaison center relate to conditions of a successful insiderization, such as achieving an active communication within the MNC as well as realizing a significant presence and weight felt throughout the regional environment and community where operations are located (Ohmae 1985: 206). The reporting on regional activities to the parent headquarters fosters the visibility of national subsidiaries and local operations vis-a`-vis the central corporate headquarters (Lasserre and Probert 1998: 49). Assigning this responsibility to the regional management center – rather than directly soliciting reports from national subsidiaries or consolidating regional data at central functions of the MNC – leads to considerable information flows and linkages between regional management and the corporate headquarters (Enright 2005a: 66). Regional senior personnel resource management might include, for example, employment, remuneration, and career development of regional senior personnel (Ohmae 1985: 187–188). In its strongest form, this implies that regional management in isolation decides on the career development plans of regional senior managers, while a weaker form would
4.1 Model of Success Factors of Regional Strategies
103
allow suggestions of national subsidiaries and/or corporate headquarters in these decisions (Ghoshal and Nohria 1989: 336). The support of regional operations may include different support activities at the regional level, such as: accounting, IT, and business process development (Enright 2005a: 67; Enright 2005b: 100; Rugman 2005b: 162). These support activities are important elements of regional operational administration as – even though they might only be weakly linked to operating activities – they represent very typically encountered functions of regional management centers (Ghemawat 2005: 104). By means of the previous explanations, we could theoretically derive four indicators for the dimension regional operational administration, as well as for each other dimension of the second-order construct regional management autonomy. In addition to the conceptual grounds of these indicators, their correspondence to each latent first-order construct has to be specified – as either reflective or formative – to generate the complete measurement model of the second-order variable regional management autonomy (Bagozzi 1998: 52–53; Eggert and Fassott 2005: 35–36).
4.1.1.5
Rules of Correspondence and Research Hypothesis 1
As outlined previously,9 this work will mainly apply the direction of causality for the specification of the rules of correspondence between the four dimensions of regional management autonomy and their respective four manifest variables. Here, for the relationship between each set of indicators and its respective latent firstorder construct – regional strategy development, regional market and product/ service development, regional market coordination, and regional operational administration – we could identify a formative direction of causality. This is due to the fact that a change of any one indicator would always cause a change in its respective first-order construct – by either increasing or reducing the overall degree of the dimension’s regional decision-making autonomy (Eberl 2006: 658; Fornell 1989: 163–164; Jarvis et al. 2003: 203). Given that each of these four indicators plays an important role in the overall extent of regional management autonomy for each of their respective dimensions (Arregle et al. 2009: 104; Enright 2005a: 66; Enright 2005b: 99–100), dropping an indicator would alter the conceptual domain of the respective first-order variable – as well as of the second-order construct – which again confirms the formative specification throughout the complete measurement model (Eggert and Fassott 2005: 43; Jarvis et al. 2003: 203). Consequently, the multidimensional construct regional management autonomy is specified by a first-order, formative measurement model. In the following, at the level of the structural model, we will present the rationale behind the hypothesis H1 on the relationship between the exogenous variable regional management autonomy and the endogenous variable regional success. 9
Cf. Sect. 3.2.2.3.
104
4 Regional Success Factor Model
Based on their analysis of the sales performance of 118 European MNCs of the Fortune Global 500 firms, Rugman and Collinson (2005: 271) observe that the majority of these firms have developed regional structures with specific resources and capabilities, primarily to meet the needs of regional customers. For another set of European MNCs, represented by a sample of 356 UK exporters, BeleskaSpasova and Glaister (2009: 303) find that the sales revenue generation in the home region or a host region requires a supporting regional organizational structure. The value of regional activities has also been confirmed by Venzin et al. (2008: 473), who find that a regional cash management of retail banks creates value for the affiliated banks within this region. In addition, by conducting case studies for seven of 29 automotive MNCs of the Fortune Global 500 firms, Rugman and Collinson (2004: 475–481) show that all of these firms rely on a regional, instead of a global, production for their sales-related performance. Furthermore, Rugman (2005b: 201) observes that a lack of market success can be explained by a headquarters-based, centralized decision-making, which: “[. . .] may not always be appropriate to address region-specific challenges [that] may be better handled through region-based [. . .] organizational structures”. Related to this argument, Richter (2007: 377) provides evidence on the fact that such regional activities prevent MNCs from suffering these performance decreases – or in other words, that regional management activities positively influence regional market success – particularly during their initial foreign expansion. Contrary to these arguments, an excessive decentralization of decision-making autonomy to regional units could also lead to inefficiencies of coordination and control within the MNC (Bartlett and Ghoshal 1989: 14), which are detrimental to regional performance (Stopford and Wells 1972: 70). This points to the fact that situations might exist, where the costs of regional management autonomy outweigh their benefits. In such cases, MNCs may prefer more centralized structures for some of their management activities (e.g., for R&D) (Rugman and Brain 2004: 12). Nevertheless, following a regional strategy implies that at least a substantial portion of decision-making autonomy should be delegated to regional management (Rugman and Verbeke 2008b: 310). More specifically, the strive of MNCs to achieve a market-oriented, regional success reflects their efforts to achieve an effective insider position, that should be associated with a high decision-making autonomy for regional management (Ohmae 1985: 206; Rugman 2005b: 66, 233). The hierarchical decentralization in these strategic decision-making processes within the MNC should be positively related to its performance (Ghemawat 2005: 100; Papadakis et al. 1998: 132). Overall, there exist more findings and arguments advocating a positive relationship between regional management autonomy and the market- or sales-related performance of MNCs. Consequently, it seems reasonable to hypothesize that increases in the decision-making autonomy of regional management should positively influence the regional performance of the MNC. Hypothesis 1 Regional management autonomy is positively associated with regional success
4.1 Model of Success Factors of Regional Strategies
105
Regional orientation Appearance Packaging Labeling
Regional design
Innovativeness of the design
H3a
Regional management autonomy
H3b
H1
Specifications Quality standards Innovativeness of features Number of features Number of products/services Diversity of products/services
Regional success
Regional functional scope
Regional offering portfolio
H2
Regional product /service adaptation
H4a
H4b
Newness of products/services Life cycle of products/services Factors that influence the brand image
Regional brands
Inter-regional distance
Brand positioning Brand development Brand name
Fig. 4.3 Measurement model of the multidimensional construct regional product/service adaptation Source: own illustration
4.1.2
Regional Product/Service Adaptation
As described in Sect. 2.2.8, a MNC may regionally adapt its products/services in the areas of design, functional scope, brands, and offering portfolio. The extent to which a MNC realizes an adaptation of downstream and upstream FSAs along these four dimensions, determines the MNC’s overall regional product/service adaptation to regional requirements and conditions. Thus a change in one of these four dimensions should cause a change in the resulting regional product/service adaptation, illustrating the formative direction of causality between the four dimensions and regional product/service adaptation.10 This is confirmed by Albers and G€otz (2006: 671) who show that the market orientation, or responsiveness, of a firm should be modeled in a formative manner, if it is achieved by a set of different strategic measures. Here, the regional market responsiveness of the MNC is expressed by the degree of its regional product/service adaptation that – as a latent second-order construct – is realized by four different strategic measures, or dimensions, at a similar level of abstraction (Chin 1998a: x). The resulting nomological network between the four dimensions and the latent second-order variable – as illustrated in Fig. 4.3 – requires a measurement model at the level of the four not directly observable, first-order constructs (Chin 1998a: x; Giere et al. 2006: 679), regional design, regional functional scope, regional brands, and regional offering portfolio.
10
In conformity with the previous chapter, this confirms the more meaningful, formative specification between first-order constructs and second-order variables (cf. Sect. 3.2.3.3).
106
4 Regional Success Factor Model
In the following – as in the previous Sect. 4.1.1 – we again adhere to the phases recommended by Giere et al. (2006: 683) for the conceptualization of second-order constructs, by first describing how the four dimensions are reflected by their indicators, before specifying the rules of correspondence between the latent firstorder variables and their indicators in a second step.11 Thirdly, we will outline our rationale for the hypothesized relationship between regional product/service adaptation and regional success, to derive its corresponding research hypothesis H2.
4.1.2.1
Regional Design
In the multidimensional construct regional product/service adaptation, one latent first-order variable is regional design, describing both the distinctive appearance and usefulness of a product/service in conformity with regional customer demands (Kotler and Armstrong 2004: 284; Meffert et al. 2008: 461; Onkvisit and Shaw 2009: 349). Here, the extent to which MNCs from different industries regionally adapt the design of their products/services may be very different (Van Pham 2006: 61–64). Consequently, the distinctiveness in the design of products/services may diverge considerably from one region to another, while also within a particular region – depending, for example, on different consumer tastes or governmentimposed tariff or non-tariff barriers, such as the requirement of design-relevant local inputs – varying degrees of adaptation may be realized (De Bu´rca et al. 2004: 269). Conversely this implies that MNCs – in the absence of strong forces for design adaptation within a region – might realize a certain level of standardization in the intra-regional design of their products/services, while they may still achieve a high level of overall regional product/service adaptation concerning interregional design differences (Lehrer and Behnam 2009: 282; Lim et al. 2006: 512).12 In this work, we predominantly focus on these inter-regional design variations, which result from tailoring the design of products/services to the needs of customers differing from one region to the other. Thus, customer preferences highly influence the overall degree of design-related product/service adaptations which are reflected in four indicators of regional design – as illustrated in Fig. 4.3 – ranging from regional products’/services’ appearance, packaging, labeling, to the innovativeness of their design (De Bu´rca et al. 2004: 268–269; Emrich 2007: 220; Kotler and Armstrong 2004: 284–288; Lehrer and Behnam 2009: 282; Onkvisit and Shaw 2009: 356). The regional adaptation of the appearance of products/services aims to attract regional customers by pleasing esthetics or by their usefulness (Kotler and Armstrong 2004: 284; Meffert et al. 11
According to our explanations in the previous chapter and in Sect. 3.2.3.3, we do generally assume the more meaningful, formative specification between first-order constructs and secondorder variables. 12 These relationships show the causal (formative) effects between the latent first-order variable regional design and its second-order construct regional product/service adaptation, where the dimension regional design forms its higher order variable within the inner model (Chin 1998a: x).
4.1 Model of Success Factors of Regional Strategies
107
2008: 461; Onkvisit and Shaw 2009: 356). The physical appearance of products/ services is useful, if their design suits, for example, the customs and size requirements of regional customers (e.g., size of cars, razors, tumblers, credit cards, servicescape) (Keillor et al. 2004: 18; Kotler and Armstrong 2004: 285; Onkvisit and Shaw 2009: 349, 356). Variations in packaging from one region to another refer to differences in the container or wrapper of products, or in the total service package consisting of tangible and intangible features (e.g., combination of staffing, customer throughput, length of wait for a particular service) (Chase and Erikson 1988: 194; Kotler and Armstrong 2004: 286; Meffert et al. 2008: 443, 459). Here, the main function of packaging relates to getting the product/service to its intended recipient in a serviceable shape and pleasing form (De Bu´rca et al. 2004: 280; Onkvisit and Shaw 2009: 410–412). Labeling may range from simple tags attached to products to complex graphics on their package, or may refer to service labels (e.g., eco service label ISO 14024 for ecological and sustainable car repair services) (Kotler and Armstrong 2004: 281; Rugman 2005b: 38). MNCs regionally adapt the labeling of their products/services, to respond to different languages (e.g., subtitles of movies), to varying regulations (e.g., required information on product composition and nutritional data, governmental permissions for the labels of patented drugs of pharmaceutical firms), or to any other information requirements of regional customers (e.g., instructions and user manuals) (De Bu´rca et al. 2004: 281; Onkvisit and Shaw 2009: 366; Rugman 2005b: 120). The innovativeness of the design of products aims to make a product easier to use, smaller, lighter, or sturdier (Athey and Schmutzler 1995: 558). In the case of services, the innovativeness of their design may be realized by innovations in the service process (e.g., improved customer integration by more client contacts during the service process) (Benkenstein and Stenglin 2006: 278–279, 287). For products, as well as services, the degree of their regionally differentiated design innovations is dependent upon the particular demands of regional customers and governments (Rugman 2005b: 141–142, 192). Depending on the extent of these demands, as described above, a MNC’s design innovations may reap both intra-regional aggregation and inter-regional adaptation advantages.13
4.1.2.2
Regional Functional Scope
Another dimension of the multidimensional construct regional product/service adaptation is regional functional scope, which relates to the regionally differentiated scope, or spectrum, of product/service functions offered to regional customers (Emrich 13
Lehrer and Behnam (2009: 282, 284, 291) show that this may be achieved by specific design principles, such as modularity or programmability – where the former decomposes a product into separable components (e.g., standardization of selected car components and regional/local adaptation of others), while the latter reconciles standardization and adaptation for programmable products (e.g., use of standardized software for computers or industrial robots, and regional/local adaptation of their user-specific programming).
108
4 Regional Success Factor Model
2007: 220; Grosse 2005: 132; Kogut 1991: 42; Kogut 2002: 78). Regional adaptations in the functional scope of products/services may be due to regulations and standards (e.g., voltage conversion functions of electrical equipment, miles-based speed information on North American car tachometers) or regional usage conditions (e.g., the frequent use of pencils in Asian countries requires copiers that can copy light pencil lines) (De Bu´rca et al. 2004: 267; Emrich 2007: 220; Onkvisit and Shaw 2009: 351, 356). All of these regional functional adaptations of products/services are reflected in four indicators of regional functional scope – as illustrated in Fig. 4.3 – ranging from the regional products’/services’ specifications, quality standards, innovativeness of their features, and number of their features (Kotler and Armstrong 2004: 283–284, 302; Meffert et al. 2008: 301). In conformity with regional customer demands, a product/service can be offered with varying specifications in different regions (Emrich 2007: 220; Johnson and Arunthanes 1995: 32–33). This may include, for example, regionally differentiated specifications in the taste, color, sound, or engine power of products, or contractual specifications for the scope of services (B€ohmann and Krcmar 2004: 166; Emrich 2007: 220; Ghemawat 2005: 105; Onkvisit and Shaw 2009: 349, 355–356). Furthermore, the functional scope of products/services may be regionally adapted by their quality standards, which describe the ability of products/services to perform their functions (Kotler and Armstrong 2004: 283, 302–303). Examples of regionally differentiated quality standards include the varying durability or reliability of products/services (e.g., as concerns regionally offered banking services), reflecting differences in their attributes valued by regional customers (Ghemawat 2005: 104, 107; Kotler and Armstrong 2004: 283–284, 302; Malhotra et al. 2005: 272). In dependence of these customer demands, some MNCs highly adapt their quality standards from one region to another, while others realize high levels of quality standardizations for their products/services across regions (Grosse 2005: 142; Rugman 2005b: 89, 164; Rugman and Girod 2003: 32). In addition, the functions of products/services may be regionally adapted by the innovativeness of their features, which is referred to as the degree of innovation, or newness, of features needed and valued by customers (Kotler and Armstrong 2004: 284, 302). Innovative features of regional products/ services may include for example new car security options, or innovative online banking functions (Claessens et al. 2003: 113–114; Grosse 2005: 142; Heitmann 2006: 236), which usually do not offer a long-lasting competitive advantage, as they are quickly imitated by regional competitors (De Bu´rca et al. 2004: 260). The fact that the pharmaceutical and chemical industries are considered the most innovative industries of the Fortune Global 500 firms (Rugman 2005b: 114), can thus be largely explained by their patent protection against competitive imitations (e.g., medicine product innovations are protected by patents of up to 10 years) (Meffert et al. 2008: 410). Irrespective of their innovativeness, MNCs can also change the number of features of their products/services, to regionally adapt their functional scope (Kotler and Armstrong 2004: 284, 302). By means of an increased number of features, MNCs aim to better serve regional customer demands with their product (e.g., adding right-hand steering functions to cars) or service (e.g., extending airline
4.1 Model of Success Factors of Regional Strategies
109
services to include beds, hot showers, and cooked-to-order breakfasts for international travelers) (Kotler and Armstrong 2004: 302; Onkvisit and Shaw 2009: 349).
4.1.2.3
Regional Offering Portfolio
An additional latent first-order variable in the multidimensional construct regional product/service adaptation is the regional offering portfolio of MNCs, which describes the portfolio, or mix, of their products/services offered at the regional level (Ghoshal 1987: 426; Kotler and Armstrong 2004: 290; Rugman 2000: 93; Rugman 2005b: 131, 134, 162, 204; Rugman and Verbeke 2008b: 310). In conformity with regional customer demands, MNCs regionally adapt their offering portfolio of products (e.g., pharmaceuticals’ portfolio of current drugs in the regional market) or services (e.g., financial services’ mix of regionally offered commercial banking, investment banking, and/or insurance services) (Grosse 2005: 132, 142; Meffert et al. 2008: 239; Rugman 2005b: 131). This regional differentiation in the products/services offered by the MNC influence the overall extent of its regional product/service adaptation, which is reflected in four indicators of its regional offering portfolio – as illustrated in Fig. 4.3 – the number, diversity, newness, and the life cycle of its products/services (De Bu´rca et al. 2004: 11–13, 263; Kotabe and Jiang 2009: 465; Kotler and Armstrong 2004: 290, 330; Meffert et al. 2008: 463; Pycraft et al. 2000: 84). By means of variations in the number of products/services in their regional offering portfolios – for example by offering and/or producing a limited number of car models exclusively in one region – MNCs can signal their commitment to building complete organizations in each of their regions (Ghemawat 2005: 103). Furthermore, MNCs can regionally adapt the diversity of their products/services, to realize regional differences in the variety, or number of versions, of products/services offered (Kotler and Armstrong 2004: 290). For example, MNCs of the oil and gas industry may regionally vary their alternative energy products – for example by utilizing different renewable energy sources, such as wind, solar, nuclear, wave, and biomass technologies – to adapt the regional diversity of their offering portfolio (Rugman 2005b: 181–182). In addition, MNCs can adapt their regional offering portfolio by the newness, or the degree of novelty, of their products/services (Kotler and Armstrong 2004: 290–291). For example, in addition to its worldwide offered products/services, a MNC could develop a completely new product/service offering for a certain region. Here MNCs, however, have to consider that such a new regional product/service entails high, irreversible, fixed costs – particularly at the downstream end, related to its market-oriented development and distribution (Ohmae 1985: 39; Rugman 2005b: 59). Moreover, MNCs can regionally adapt their regional offering portfolio according to the life cycle of their products/ services. (De Bu´rca et al. 2004: 11–13, 263; Kotabe and Jiang 2009: 465; Kotler and Armstrong 2004: 330; Meffert et al. 2008: 463; Pycraft et al. 2000: 84). The stages of a product/service life cycle are usually labeled “introduction”, “growth”, “maturity”, and “decline” (De Bu´rca et al. 2004: 11; Pycraft et al. 2000: 84).
110
4 Regional Success Factor Model
Variations in the regional offering portfolio according to the product/service life cycle implies, for example, that MNCs may extend the life of a product/service in their home region – by offering it in another region, where the “decline” stage has not yet been reached (De Bu´rca et al. 2004: 12).
4.1.2.4
Regional Brands
A further dimension of the multidimensional construct regional product/service adaptation are regional brands, which refers to regionally differentiated names, terms, signs, symbols, or designs, or a combination of these – that identify the regional products/services of a MNC (De Bu´rca et al. 2004: 265; Kotler and Armstrong 2004: 285; Onkvisit and Shaw 2009: 390; Rugman 2005b: 162; Rugman and Collinson 2004: 475; Rugman and Verbeke 2004: 6; Schuiling and Kapferer 2004: 108). Given that brands are generally associated with perceptions and feelings of consumers about a product/service and its benefits, the extent to which MNCs should employ regional, instead of global or local, brands is highly dependent upon regional customer and market requirements (De Bu´rca et al. 2004: 265; Kotler and Armstrong 2004: 291). This overall degree of a MNC’s brand-related product/service adaptation is reflected in four indicators of regional brands – as illustrated in Fig. 4.3 – ranging from factors that influence the brand image, brand positioning, brand development, and brand name of regional products/services (Kotler and Armstrong 2004: 292–297, 302; Rugman 2005b: 45, 89, 185; Rugman and Verbeke 2004: 6, 10). MNCs can regionally adapt factors that influence the brand image of their products/services – for example by regionally employing different symbols or messages that express characteristics and benefits of a regional product/service (Kotler and Armstrong 2004: 302; Meffert et al. 2008: 305; Schuiling and Kapferer 2004: 108). These image-related factors of regional products/services even affect the “region-of-origin” preferences of consumers – as a complement to “countryof-origin” effects – which influence their confidence, attitudes and purchase intentions (Agarwal et al. 2002: 464–465; Batra et al. 2000: 84; Pappu et al. 2007: 729; Rugman 2005b: 67). Consequently, MNCs can realize intra-regional aggregation advantages by a regionally standardized brand image, while still differentiating brand images from one region to another (Onkvisit and Shaw 2009: 390; Rugman and Girod 2003: 26; Stopford and Wells 1972: 56). Furthermore, MNCs can regionally adapt the brand positioning of their products/services, referring to regional variations in how MNCs position their brands in the minds of regional target customers (Kotler and Armstrong 2004: 292). This brand positioning leads to regionally different expectations of regional customers about particular attributes, benefits, and values of the products/services offered by MNCs (Kotler and Armstrong 2004: 292–293; Onkvisit and Shaw 2009: 390–391). MNCs may achieve such a regional variation in the positioning of their brands, for example, by regionally differentiating their television advertising – e.g., for Asia, North America, and/or Europe (Alden et al. 1999: 78). In addition, MNCs can regionally
4.1 Model of Success Factors of Regional Strategies
111
adapt the brand development of their products/services – by means of line extensions (introducing new forms, sizes, flavors or ingredients under the same regional brand name), brand extensions (use of an existing regional brand name to launch a new product/service), multibrands (introducing additional brands for the same product/service category), and new brands (creating a new brand for a new product/service category) (Kotler and Armstrong 2004: 296). An example of brand extensions is provided by several Asian MNCs who manufacture and offer high quality products under well-known brands (Chao et al. 2003: 486). Moreover, MNCs can realize a high regional adaptation of their brands by changing the brand name of their products/services from one region to another (Kotler and Armstrong 2004: 293; Rugman 2005b: 226–227; Rugman and Verbeke 2004: 6). For example, some MNCs from the financial services industry issue their credit cards under different brand names, while several MNCs from the automotive industry regionally adapt the branding of their products by varying the names of their cars (Nebenzahl and Jaffe 1996: 5; Onkvisit and Shaw 2009: 391; Rugman and Collinson 2004: 475–476; Schlie and Yip 2000: 350).14 According to our conceptual explanations before, the outer model for the latent first-order variable regional brands, as well as for each other dimension of the latent second-order construct regional product/service adaptation, consist of four indicators. Besides the conceptual grounds of these manifest variables, a complete definition of the measurement model for the multidimensional construct regional product/service adaptation also requires the specification of the rules of correspondence between the theoretical dimensions and their indicators (Jarvis et al. 2003: 199).
4.1.2.5
Rules of Correspondence and Research Hypothesis 2
In conformity with the arguments stated above,15 the specification of the rules of correspondence between the four dimensions of regional product/service adaptation and their respective four indicators is mainly based upon the direction of causality between these latent and manifest variables. Here, for all relationships of the latent first-order variables – regional design, regional functional scope, regional offering portfolio, and regional brands – with each of their respective sets of manifest variables, we could identify a reflective direction of causality. This is confirmed by Albers and G€otz (2006: 671) who argue for a reflective specified measurement model of strategic measures, if these aim to increase a firm’s market orientation, or responsiveness. In a similar vein, the four latent first-order variables above represent different strategic measures of a MNC, to realize regional market responsiveness by a certain degree of regional product/service adaptation. More specifically, 14
For example, Volkswagen’s “Fox” in Europe and South America is named “Lupo” in North America (Kulic 2009: 90). 15 Cf. Sect. 3.2.2.3.
112
4 Regional Success Factor Model
the latent first-order variables represent region-bound FSAs, the building blocks of a MNC’s regionally responsive product/service strategies (Rugman 2005b: 50). For example, when entering into a new host region, MNCs usually have to build up new strengths – e.g., competencies in how to adapt the design of their products/services to the customer demands of this foreign host region. If MNCs have created such host region-bound FSAs, they are able, or competent enough, to properly address host region-specific demands – for example by adapting the appearance or the labeling of their host-regional products/services. This leads to a reflective direction of causality, given that changes in each of these regional company strengths would always cause changes in their sets of indicators, as the latter are respective manifestations of these four latent first-order constructs (Eberl 2006: 658; Hulland 1999: 201; Jarvis et al. 2003: 203; Rugman 2005b: 50). In other words, the MNC’s ability to link its region-specific FSAs to CSAs of the respective region is engrained in the four latent first-order constructs, which decide on the extent to which a MNC is able to effectively implement their indicators – for example a regional packaging, regional quality standards, a regional diversity of their products/services, and/or regional brand names (Arregle et al. 2009; Rugman 2005b: 229; Rugman and Verbeke 2005: 13). Furthermore, a regional FSA (e.g., the regional design of products/services) should usually lead to an overall impression of the MNC’s products/services within a region. Conversely, we do not expect that regional customers analytically separate each of their indicators in their purchasing decisions (e.g., by distinguishing the products’/services’ appearance from their packaging, labeling, and innovativeness of their design). Consequently, we expect that the indicators within each set of manifest variables highly covary with each other, which again confirms the reflective specification of the complete outer model (Diamantopoulos and Winklhofer 2001: 271; Jarvis et al. 2003: 203). As a result, the measurement model of the multidimensional construct regional product/service adaptation is specified in a reflective manner. At the level of the structural model however, the relationship between the exogenous variable regional product/service adaptation and the endogenous construct regional success has not been explored yet. Therefore, in the following, we will outline the hypothesis H2 on this relationship and its underlying reasoning. For purely export-based strategies, Cavusgil and Kirpalani (1993: 2, 5, 11–13) find, in their analysis of 130 cases of international foreign market entries, that subsequent to market entry, especially for large firms, a local and/or regional product adaptation is important for sustained export success of MNCs. By means of 589 responses of firms from the US, Korea, and Japan, this has been confirmed by Calantone et al. (2006: 178–180, 182–183) who found a positive relationship between a MNC’s product adaptation strategy and its sales- and profit-related export performance (Kotabe and Jiang 2009: 467). A reason behind this is provided by Baden-Fuller and Stopford (1991: 497–505), who show in their study of the washing machine industry, that the creation of a global product suitable for many countries is very difficult and involves many pitfalls (Elango 2004: 433). A related explanation is offered by Malhotra et al. (2005: 272–273), who find high variations in how service quality of banking services is perceived by customers from North
4.1 Model of Success Factors of Regional Strategies
113
America and Asia – suggesting that the performance of MNCs is positively associated with regional adaptations of products/services (Kotabe and Jiang 2009: 465–466). The dependence of these performance effects on regional customer demands implies for MNCs that – in addition to back-end advantages from upstream FSAs – particularly: “[. . .] strong downstream FSAs are necessary to achieve market success in the market considered, [. . .] at the national, regional, or inter-regional level” (Rugman and Verbeke 2008b: 308). A high degree of regional product/service adaptation, however, also involves high costs, and thus might also result in negative performance effects. To avoid such negative performance outcomes, the regional adaptation of products/services should add sufficient incremental revenue, or economies of scale, to recapture the costs of its related investments (Calantone et al. 2004: 185, 189). Therefore, the success of a MNC is not only dependent upon the responsiveness of products/ services to regional demands and requirements, but also on its ability to regionally integrate their products/services across national markets – thus by realizing aggregation advantages from exploiting similarities within the region. Here, according to Rugman (2005b: 197), regional product/service adaptation represents the most promising path for the market success of MNCs, as this might be the key to achieve an optimal mix, or moderate levels, of both global integration and national responsiveness. Empirical evidence for this argument has been provided by Mauri and Sambharya (2003: 35–37, 40) who show – based on their analysis of 69 MNCs from the computer, office and electronics, telecommunications, and automotive industry for the period from 1992 to 1997 – that moderate levels of global integration are positively associated with firm performance,16 measured here by return on sales. Consequently, it seems reasonable to hypothesize that increases in the regional adaptation of products/services should positively influence the regional performance of the MNC. Hypothesis 2 Regional product/service adaptation is positively associated with regional success The hypothesized relationships between regional management autonomy and regional success (H1), as well as between regional product/service adaptation and regional success (H2), describe the main relationships in the structural model of Fig. 4.1. However, following the moderate voluntaristic approach of this work – as argued earlier in Sect. 2.4.2 – additional influences from contingency variables have to be considered in these relationships of the structural model, which we will present in the next chapter.
16
In conformity with Kobrin (1991: 19) and Mauri and Phatak (2001: 240), Mauri and Sambharya (2003: 35) measure global integration as the inter-subsidiary product/service, or resource, flows – which are estimated by calculating the percentage of total sales that is generated within the firm and across geographic regions. Geographic restrictions to inter-subsidiary product/service flows can be suitably explained by regional product/service adaptations of MNCs, which lead to high intra-regional trade and product/service flows – typical for the regionalization of markets (Rugman 2000: 114; Rugman and Hodgetts 2001: 334) – resulting in moderate levels of global integration.
114
4.1.3
4 Regional Success Factor Model
Contingency Variables
The conditions and contexts, under which the relationships of H1 and H2 may vary, are given by contingency variables, the MNC’s regional orientation and the interregional distance it faces. In methodological terms, each of these internal and external contextual variables interacts with every exogenous variable – regional management autonomy and regional product/service autonomy – to affect the endogenous variable regional success (Hartmann and Moers 1999: 203). In other words, the hypothesized relationships of H1 and H2 in the structural model are moderated by the contingency, or moderating, variables regional orientation and inter-regional distance (Chin et al. 2003: 191; Eggert et al. 2005: 102; Hartmann and Moers 1999: 203). Given that these moderators act at the level of the structural model – as not directly observable, latent constructs – measurement models for both regional orientation and inter-regional distance are required, as illustrated in Fig. 4.4. In the following – given that the conceptual foundations of the contingency variables regional orientation and inter-regional distance have been laid in Sect. 2.3.2 – we will mainly concentrate on two steps in the modeling of each of these moderators. Building on how these latent moderating variables are reflected by their indicators, we will first specify the rules of correspondence between them. Here, as argued earlier,17 the direction of causality will be applied as the dominant principle for specifying the rules of correspondence between the latent and manifest variables. Secondly, we will present our research hypotheses on the moderating effects of regional orientation (H3a and H3b) and inter-regional distance (H4a and H4b) on the hypothesized relationships H1 and H2.
4.1.3.1
Regional Orientation
As illustrated in Fig. 4.4, a MNC’s regional orientation is reflected by three indicators of this latent construct. Following our explanations in Sect. 2.3.2.1, this includes a market-related and an activity-related regional orientation. Furthermore, due to the fact that the regional orientation reflects a firm-internal, regiocentric orientation of MNC managers towards the region as a potential market (Wind et al. 1973: 15), a subjective estimation of MNC managers of this marketrelated regional orientation represents another important indicator of this latent variable. While this subjective estimate is derived from the survey of this work,18 the objective measurement of the MNCs’ market-related regional orientation is measured by the distribution of sales across economic regions – whereas its activity-related regional orientation is also assessed objectively by the regional dispersion of its assets (Goerzen and Asmussen 2007: 66–67; Oh 2009: 342; Rugman 2005b: 4; Rugman and Verbeke 2004: 5, 7). All three manifest variables indicate 17
Cf. Sect. 3.2.2.3. Cf. Sect. 5.2.2.
18
4.1 Model of Success Factors of Regional Strategies
Objective market-related regional orientation
Objective activities-related regional orientation
115
Subjective market-related regional orientation
Regional orientation
H3a
Regional management autonomy
H3b
H1
Regional success H2
Regional product/service adaptation
H4a
H4b
Inter-regional distance
Cultural distance
Administrative distance
Geographic distance
Economic distance
Fig. 4.4 Measurement model of the contingency variables regional orientation and inter-regional distance Source: own illustration
different degrees of a MNC’s regional orientation, which is assumed to decrease from home-regional to host-regional, from host-regional to bi-regional, and from biregional to multi-regional (global).19 These conceptual characteristics of the manifest variables are important for the subsequent specification of the rules of correspondence between the latent moderating variable and its indicators.
4.1.3.2
Rules of Correspondence and Research Hypotheses 3a/3b
The regional orientation of MNCs causes a focus in their geographic expansion – related to markets or to activities – on their own region (home-regional orientation), 19
Cf. Sect. 2.3.2.1.
116
4 Regional Success Factor Model
one foreign region (host-regional orientation), two regions (bi-regional orientation), or several regions (multi-regional, or global, orientation). Thus, any changes of the observed measures reflect variations in the latent moderating variable, indicating a reflective direction of causality between the contingency variable regional orientation and its indicators (Henseler et al. 2009: 289; Jarvis et al. 2003: 203). Furthermore, both the market-related and the activities-related geographic spread of a MNC result from a careful cost-benefit analysis in its regional orientation (Elango 2004: 433) – a common theme shared by the indicators – again confirming the reflective mode of the outer model’s correspondence rules (Eggert and Fassott 2005: 43; Jarvis et al. 2003: 203). Consequently, the measurement model of the latent moderator regional orientation is specified in a reflective form. In the following, we will outline our reasoning for our research hypotheses H3a and H3b that refer to the moderating effects of regional orientation on the hypothesized relationships of the basic, or main, structural model – either between regional management autonomy and regional success (H1), or between regional product/service adaptation and regional success (H2). As described in Sect. 2.3.2.1, a MNC’s regional orientation – its regional spread of markets and activities – depicts its regional geographic reach and its regional geographic scope. In their study of 22 MNC’s over 26 years from 1967–1992, Vermeulen and Barkema (2002: 637, 643, 648–649) found that a broader geographic scope of these firms – measured by their spread of geographical and product markets entered – negatively moderates increases in profitability resulting from their international expansion (Rugman 2005b: 75). They argue that this is due to the fact that a broader geographic scope strains the firm’s absorptive capacity (Cohen and Levinthal 1990: 131–132), indicating that foreign expansion is easier to absorb by the MNC if it occurs in countries that are “related” – e.g., according to Ronen and Shenkar’s (1985: 452) classification of country clusters (Rugman 2005b: 75; Vermeulen and Barkema 2002: 648–649).20 This explains why MNCs expand by regions – as each of them consists of such “related” countries that are relatively close to one another (Ghemawat 2005: 100) – while allocating substantial decisionmaking autonomy to their regional management to better absorb, or exploit, the growing diversity of regional environmental developments and organizational resources (Enright 2005a: 60–61; Lehrer and Asakawa 1999: 284). Here, a limited geographic scope, represented by a strong regional orientation, should facilitate this organizational learning of regional management, aiming at realizing both adaptation and aggregation advantages in handling region-specific challenges (Rugman
20
The MNC’s absorption of its international expansion relates to cross-border organizational learning, the processing of geographically dispersed knowledge by a firm’s management (Cohen and Levinthal 1990: 131–132, 134–135; Rugman 2005b: 75; Vermeulen and Barkema 2002: 641). The extent to which the need for management structures rises, due to an increasing international expansion of MNCs, is shown by the fact that the headquarters of companies with a wider geographical scope tend to be significantly larger (Collis et al. 2007: 394).
4.1 Model of Success Factors of Regional Strategies
117
2005b: 76; Ruigrok and Wagner 2003: 77–79).21 Conversely, for MNCs with a high geographic scope, absorptive capacity constraints are salient that – even though the firm builds on an extensive internationalization experience and highly autonomous regional management in all regions – lead to an organizational and environmental complexity, which is extremely difficult to handle successfully (Bausch and Krist 2007: 339; Elango 2004: 433, 436; Ruigrok et al. 2007: 363–364; Verbeke et al. 2009: 156–157). As a result, the extent to which a high degree of regional management autonomy is effective in the exploitation of adaptation and aggregation economies at the regional level – meaning that it considerably contributes to the regional success of a MNC – is contingent upon the internal contextual variable regional orientation. Here, increases in a MNC’s regional orientation should positively interact with the degree to which increases in regional management autonomy leads to regional success. This brings us to our research hypothesis that the regional orientation of a MNC should positively moderate the hypothesized relationship H1 between regional management autonomy and regional success. Hypothesis 3a A firm’s regional management autonomy will have a stronger impact on regional success given a higher regional orientation In addition to this moderating effect related to the regional success of regional management, the regional orientation of a MNC also affects its successful regional product/service adaptation – as its widespread geographic diversification may involve: “[. . .] pitfalls similar to the conventional drawbacks of product diversification” (Rugman 2005b: 77). Given that a firm’s regional products/services, by definition, are not attractive to consumers all around the world (Rugman and Verbeke 2004: 5), advances in a MNC’s geographical expansion – corresponding to a lower regional orientation – increase its product/service diversification. As illustrated in Sect. 2.2.8, MNCs can achieve this by augmenting, region-by-region, the geographic reach of their FSAs. In his study of 130 firms of the Directory of Multinationals for the year 2001, Elango (2004: 433–434, 437) finds that these MNCs are able to handle moderate levels of product diversification effectively, while a high diversity in products and markets constrains their effective performance. Therefore, the degree to which a MNC can successfully adapt its products/ services to each region – referring to its regional success from properly addressing regional demands by its FSAs – is dependent upon the internal contingency variable regional orientation. More specifically, a MNC’s increasing regional orientation should positively interact with the influence that regionally adapted products/ services have on its regional success. As a result, it appears reasonable to hypothesize that the regional orientation of a MNC should positively moderate the
21
The fact that a firm’s geographic concentration has a positive influence on its performance – e.g., by means of aggregation advantages from shared resources and capabilities – has also been confirmed for a more restricted geographical space of MNCs at the sub-national level (MolinaMorales 2001: 291).
118
4 Regional Success Factor Model
relationship between regional product/service adaptation and regional success, which we postulated in hypothesis H2. Hypothesis 3b A firm’s regional product/service adaptation will have a stronger impact on regional success given a higher regional orientation By means of the previous explanations, we have derived the research hypotheses H3a and H3b about the moderating effects of the latent construct regional orientation. In addition to these internal contextual variables – in the successful realization of its regional strategies – the MNC is exposed to external contingencies from the inter-regional distance it faces, which we will explore in the following.
4.1.3.3
Inter-Regional Distance
As outlined in Sect. 2.3.2.2 and depicted in Fig. 4.4, the inter-regional distance faced by a MNC is a latent moderating variable, which is observed by its manifest variables indicating an inter-regional cultural, administrative/political/institutional, geographic, and economic distance.22 The inter-regional distance – based on all of the different firm-external, environmental elements represented by its indicators – can substantially increase the liability of foreignness of MNCs, which results in costs and investments to complement their existing FSAs with new, location-bound FSAs and CSAs (Rugman 2005b: 225). Before further exploring how these conceptual underpinnings of inter-regional distance, and its related economics, affect the hypothesized relationships H1 and H2, we will specify the rules of correspondence between the latent moderating variable and its manifest variables.
4.1.3.4
Rules of Correspondence and Research Hypotheses 4a/4b
According to Ghemawat (2001: 138), inter-regional distance is a construct that: “[. . .] can manifest itself along four basic dimensions: cultural, administrative, geographic, and economic”. Consequently, these four facets are indicative manifestations of inter-regional distance, implying a reflective direction of causality from the latent moderating variable to its indicators (Eggert and Fassott 2005: 43; Jarvis et al. 2003: 203). This categorization of the four indicators as the manifestations of inter-regional distance is supported by Rugman and Verbeke (2008a: 406), who observe that inter-regional distance has: “[. . .] regulatory/institutional, cultural and economic components [. . .]”. Furthermore, the indicators are expected to covary with each other, as: “[. . .] these four components often intertwine [. . .]” (Ghemawat 2008: 40), again confirming the reflective character of the relationship 22
In Fig. 4.4, we utilized “administrative distance” as an abbreviation for “administrative/political/ institutional distance”.
4.1 Model of Success Factors of Regional Strategies
119
between the latent moderating variable inter-regional distance and its manifest variables (Diamantopoulos and Winklhofer 2001: 271; Jarvis et al. 2003: 203). As a result, the measurement model of the latent moderating variable interregional distance is specified in a reflective form. In the following, at the level of the structural model, we will present our rationale behind the hypotheses H4a and H4b that describe the moderating effects of inter-regional distance on the hypothesized relationships H1 and H2 of the main structural model. Various academic scholars have investigated the relationships between different indicators of inter-regional distance and the effectiveness of a regional management in achieving regional market success. Yeung et al. (2001: 169) observe that the higher the geographic distance, the more important are increased regional decisionmaking responsibilities, to effectively explore regional markets. This increase of regional management autonomy is related to regional performance effects, if it leads to a better linkage of a MNC’s institutional distance to its resources and structures (Venzin et al. 2008: 474; Xu and Shenkar 2002: 615), to overcome its liability of foreignness in host regions (Miller and Richards 2002). Another argument for a high regional management autonomy in light of an increased interregional distance is provided by Collinson and Rugman (2008: 225–228), who show for 75 Asian MNCs of the Fortune Global 500, that the effectiveness of their regional management activities is highly contingent on their societal, cultural, and economic environment.23 Due to these contextual influences of inter-regional distance – according to the analysis of 18 Fortune Global 500 firms of the chemicals and pharmaceuticals sector by Rugman and Brain (2004: 8, 23–24) – a high regional management autonomy is required, as marketing activities cannot be transferred to host regions.24 Moreover, as shown in the study of 148 US manufacturing firms during 1985–1996 by Tong and Reuer (2007: 219, 223, 226), negative performance effects from risks and coordination costs of the MNC increase with the cultural heterogeneity in the firm’s portfolio of international operations – mainly due to complexities in transferring management practices abroad, and in adapting the firm’s value chain activities to the respective market requirements.25 In particular the latter argument indicates how a high inter-regional distance translates into the need for a highly autonomous regional management to 23
The regional management activities in their study included regional operational administration and regional market coordination activities – the former represented by regional human resource management (e.g., recruitment practices, career paths), and the latter by the coordination of intraregional networks (e.g., input-related supply chains or R&D alliances, and output-related distribution keiretsu or retailers) (Collinson and Rugman 2008: 225–226). The fact that these 75 Asian MNCs are highly tied to the political, economic, social context and business infrastructures in their activities, has also been highlighted elsewhere by Rugman and Collinson (2006: 195). 24 This explains why political and cultural sensitivity are considered as important traits of regional managers (Lasserre and Probert 1998: 49). 25 In Tong and Reuer’s (2007: 215–216) findings, the risk-related negative effect on MNC performance is given by the downside risk of international investments, which is found to be an increasing function of the average cultural distance between a MNC’s home base and the host countries of its foreign subsidiary operations.
120
4 Regional Success Factor Model
achieve regional success. Conversely, in the absence of a highly autonomous regional management (e.g., MNCs that focus on exporting), detrimental effects of inter-regional distance on performance may result. This has been validated in the analysis of 345 Chinese exporters by Ellis (2007: 378, 383), who shows that the increased distance of these exporting MNCs to their customers and markets is negatively associated with realizing the required market orientation through their marketing activities. Therefore, inter-regional distance can hamper the marketseeking expansion of MNCs in host regions – as it may lead to a high liability of foreignness, particularly in downstream activities (Rugman 2005b: 212) – if it is not associated with a high regional management autonomy to link the MNC’s: “[. . .] existing knowledge base with host-region location advantages [. . .] which do not simply meld together without managerial intervention” (Rugman 2005b: 223). Therefore, an increasing inter-regional distance faced by the MNC should positively interact with the degree to which increases in regional management autonomy lead to regional success. More specifically, we expect that the external contingency variable inter-regional distance has a positive moderating effect on the hypothesized relationship H1 between a MNC’s regional management autonomy and its regional success. Hypothesis 4a A firm’s regional management autonomy will have a stronger impact on regional success given a higher inter-regional distance This moderating effect of inter-regional distance not only acts on the regional success of MNCs that is related to regional management autonomy – but also influences the successful regional adaptation of their products/services, as the size of these regional product/service adaptation investments is driven by: “[. . .] home/ host region differences in the cultural, administrative, geographic, and economic sphere [. . .]” (Rugman 2005b: 230). Consequently, increases in the inter-regional distance faced by the MNC lead to rising costs from its liability of foreignness, which can be reduced by means of a successful insiderization (Zaheer 2002: 353) – via regionally adapted products/services, particularly by FSAs at the downstream end (Ohmae 1985: 32, 206; Rugman 2005b: 66, 229–230; Rugman and Verbeke 2008b: 307). Given that these adaptations imposed by inter-regional distance are mainly related to the downstream end – as explained in Sect. 2.2.8 – this involves high risks for the MNC from one-sided, upfront investments for adapting its products/services to host region demands. Due to the inseparability of upstream and downstream FSAs in services, a: “[. . .] higher inter-regional distance will impose a significant additional burden of adaptation on services firms” (Rugman and Verbeke 2008a: 409). These high risks, however, have to be incurred by MNCs, to achieve regional success, as cultural and economic differences between regions influence consumers’ product/service choices as well as the diffusion and acceptance of new products/services (Kotabe and Jiang 2009: 466–468; Malhotra et al. 2005: 271–272; Yaveroglu and Donthu 2002: 61–62; Yeniyurt and Townsend 2003: 390–391). Furthermore, Chakrabarti et al. (2007: 107–108, 117) show in their study of 3,177 Asian companies in the period 1988–2003, that the
4.1 Model of Success Factors of Regional Strategies
121
performance of regional product/service adaptations – measured by changes in their return on assets from increases in regional product/service diversity through diversification – is highly contingent on the stability of the institutional and economic environment. In addition, different elements of an increased geographic interregional distance (e.g., physical distance, climate and temperature differences) influence the successful regional product/service adaptation of MNCs (e.g., with respect to packaging requirements) (De Bu´rca et al. 2004: 280). Therefore, the extent to which a MNC can successfully address regional demands by its regionally adapted products/services – which lead to regional success by the successful deployment of its FSAs – is contingent on the external contextual variable interregional distance. Here, increases in inter-regional distance should positively interact with the effect that a regional product/service adaptation has on the regional success of a MNC. Consequently, we expect that the inter-regional distance faced by a MNC should positively moderate the hypothesized relationship H2 between regional product/service adaptation and regional success. Hypothesis 4b A firm’s regional product/service adaptation will have a stronger impact on regional success given a higher inter‐regional distance The above described exogenous constructs and the moderating variables of the structural model – as well as the interaction variables, calculated as the Cartesian product of the exogenous variables and the moderators – all influence the extent to which a MNC is regionally successful and thus represent all independent variables of the endogenous construct regional success (Eggert et al. 2005: 107; G€otz and Liehr-Gobbers 2004: 725; Hartmann and Moers 1999: 293; Homburg and Klarmann 2006: 730), which will be presented in the following chapter.
4.1.4
Regional Corporate Success
An adequate operationalization of regional success according to the requirements for its conceptualization – as outlined in Sect. 2.4.2. – implies its measurement by various indicators. As the indicator of the regional success of MNCs, Rugman (2005b: 3–4) utilizes the percentage of intra-regional sales, meaning the relationship between sales revenue within a particular region and total sales revenue of the MNC. Even though this geographical dispersion of sales indicates a MNC’s ability of a successful regional market penetration (Rugman and Verbeke 2004: 6; Rugman and Verbeke 2008b: 308), the focus on only one success indicator has stimulated critique and extensions of Rugman’s (2003b, 2005b) way of measuring the regional success of MNCs (Clark et al. 2004: 513–514; Delios and Beamish 2005: 23; Osegowitsch and Sammartino 2007; Rugman and Oh 2007). Rugman (2005b: 231) himself recognizes that an important additional condition besides salesbased success should be fulfilled, when he states that: “[. . .] ultimately it is market
122
4 Regional Success Factor Model
penetration (if achieved in a profitable way) that provides the best, in fact the only, indicator of global corporate success”. Consequently, the profitability that a MNC realizes in a particular region is a further important indicator of regional corporate success. Here, also a relative figure should be utilized – given by the ratio of regional profits to total profits – to assess the degree of relative importance of this region’s profitability in comparison to total firm profits. For such a successful market penetration, customer-end FSAs in particular are relevant, or more specifically, the MNC’s: “[. . .] ability to adapt successfully the deployment of its existing FSAs to the specific circumstances of foreign markets, i.e., by better aligning FSAs and CSAs” (Rugman 2005b: 229). According to Rugman and Verbeke (2008b: 311), besides a success indicator of these downstream FSAs – given by the ratio of regional sales to total sales – another regional success indicator has to be considered: the ratio of regional assets to total assets. Measures of this geographical dispersion of both the sales and the assets of MNCs reflect the outcome of their regional strategies and structures in the form of downstream and upstream FSAs (Rugman and Verbeke 2008b: 311). As a result, regional corporate success will be measured by three indicators: the ratio of regional sales to total sales, the ratio of regional profits to total profits, and the ratio of regional assets to total assets. The direction of causality in performance measures is usually reflective – which also applies in this work, as these three
Regional orientation
H3a
Regional management autonomy
H3b Regional sales to total sales
H1
Regional success H2
Regional product/service adaptation
Regional profits to total profits Regional assets to total assets
H4a
H4b
Inter-regional distance
Fig. 4.5 Measurement model of the endogenous variable regional success Source: own illustration
4.1 Model of Success Factors of Regional Strategies
123
indicators are manifestations of the latent construct regional success (Hulland 1999: 201; Jarvis et al. 2003: 203). The resulting measurement model of regional success is illustrated in Fig. 4.5. With the specification and operationalization of the measurement model for regional corporate success, all outer models have been defined for observing the latent variables in the structural model of success factors of regional strategies. To assure that the relationships postulated in the research hypotheses H1–H4b between the exogenous variables of the structural model and regional corporate success are not affected by additional influences, control variables are introduced to control for the latter.
4.1.5
Control Variables
For this work, two control variables are particularly relevant: firm size and regional competition. The strategic decision-making of MNCs is influenced by firm size (Stopford and Wells 1972: 72), in particular the comprehensiveness and rationality of their strategic decision-making (Papadakis et al. 1998: 135). As illustrated in Sect. 2.2.7, the bounded rationality of MNCs highly influences their delegation of decisions to the regional level. Therefore, firm size influences the establishment of regional management centers (Enright 2005a: 63; Stopford and Wells 1972: 72) – which constitute an important organizational influence in the realization of regionally successful MNC strategies.26 In addition, Beleska-Spasova and Glaister (2009: 296, 302) found that firm size plays an important role in the degree of geographic spread of a MNC’s operations across triad regions that – by influencing geographic scope – may also affect regional performance outcomes. Therefore, firm size should be controlled for, and will be measured by total company sales (Arregle et al. 2009: 94; Hutzschenreuter and Gr€ one 2009: 1157; Rugman 2005b: 227). This should be an adequate measure of firm size, given that – even for the Fortune Global 500 firms – the difference between the lowest and the highest value of their sales data from 2000–2008 varies by a factor of 20 to 24 (Fortune 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009). In addition, MNCs may face geographically varying degrees of competition. These differences in the competitive forces acting at the regional level are given by the degree of regional competition – which is influenced both by intra-regional and foreign competitive attacks that affect a MNC’s regional performance by triggering different strategic responses (Hutzschenreuter and Gr€one 2009: 1164). Due to high variations in regional competition, the performance of a firm in one region might not be comparable to the performance of another firm in a different region – e.g., if its underlying regional product/service adaptations are highly dependent upon these
26
Cf. Sect. 4.1.1.
124
4 Regional Success Factor Model
regional differences in competition.27 This has been shown by Miller and Richards (2002: 332), who found that the liability of foreignness faced by a MNC increases in highly competitive host regions – which restricts their ability to successfully exploit its FSAs in foreign host regions. Therefore, regional success should be controlled for these effects of regional competition. By means of the explanations above, we have derived the latent variables of the structural model and its respective measurement models, as well as control variables. In the following, we will briefly describe the resulting, complete regional success factor model.
4.2
Summary: Regional Success Factor Model
As illustrated in Fig. 4.6,28 the regional success factor model describes the relationship between the exogenous variables regional management autonomy and regional product/service adaptation and the endogenous variable regional success (H1 and H2) – considering the interaction effects from the moderating variables regional orientation and inter-regional distance on this relationship (H3a/3b and H4a/4b), as well as control variables.29 The regional success factor model of Fig. 4.6 illustrates the relationships between the main elements of regional strategies according to Rugman’s (2005b: 48–49) regional strategy matrix – regional management autonomy and regional product/service adaptation – and regional success, formulated by the hypotheses H1 and H2. These relationships are described as contingent upon the interaction with each of the moderators, regional orientation (H3a and H3b) and inter-regional distance (H4a and H4b). Those factors within this regional success factor model, or latent variables, of the multidimensional constructs – regional management autonomy and regional product/service adaptation – that qualify as important for
27
It should be noted that these geographical differences in the extent of competition faced by a MNC may be very different from its degree of its industry competition (Porter 1986: 17). For example, a MNC might be substantially exposed to a region-specific competition of a foreign host region (for example given by regional, or even local competitors), while these regional competitors of the foreign host region might not form part of the core competitors of its industry. Due to the fact that this work focuses on regional success factors of MNCs related to their regional competitiveness, we perceive a MNC’s regional competition as a more relevant control variable than its industry competition. However, while we will not control for industry competition, we will present industry-specific differences for the MNCs of our study (cf. Chap. 6). 28 Due to the complexity of the structural model of this work, the control variables are not explicitly depicted in Fig. 4.6. It should be noted, however, that they are modeled as two additional, exogenous control variables on regional success, each being measured by only one reflective manifest variable (Reinartz et al. 2004: 295) – firm size, or respectively, regional competition. 29 The respective latent constructs and their measurement models have been described in the previous chapter.
4.2 Summary: Regional Success Factor Model
125
Objective market-related regional orientation
Strategy formulation
Objective activities-related regional orientation
Subjective market-related regional orientation
Competitor intelligence Business model development Budgeting, capital allocation, control Marketing planning and execution
Regional strategy development
Regional orientation
Regional market development
H3a
Sales planning and execution Customer service and support R&D of products or service development
H3b
Coordination of distribution Coordination of production/logistics Coordination of legal entities/operations
Regional market coordination
Monitoring of legal entities/operations Liaison center for the parent company Reporting to the parent headquarters Senior personnel resource management
Regional management autonomy H1
Regional operational administration
Regional sales to total sales
Support of operations Appearance Packaging
Regional success Regional design
Regional assets to total assets
H2
Labeling
Regional profits to total profits
Innovativeness of the design Specifications Quality standards
Regional functional scope
Innovativeness of features Number of features Number of products/services
Regional product/service adaptation H4a
Regional offering portfolio
H4b
Diversity of products/services Newness of products/services Life cycle of products/services
Regional brands
Inter-regional distance
Factors that influence the brand image Brand positioning Brand development Brand name
Cultural distance
Administrative distance
Geographic distance
Economic distance
Fig. 4.6 Regional success factor model Source: own illustration
realizing a superior regional success of the firm, are classified as regional success factors of MNCs. Here, the latent second-order constructs are direct regional success factors, as they directly affect regional success, while the latent first-order variables can be classified as indirect regional success factors, if they highly influence the values of the second-order variables and thus, their impact on regional success. Such indirect influences on regional success are explained by the formative specification of the rules of correspondence between the latent first-order variables and the direct regional success factors, as changes in the former should always lead to changes in the construct (Jarvis et al. 2003: 203). The evaluation of the extent to which a first-order and latent second-order variable is important, will be based upon a survey-based inquiry about current practices regarding these multidimensional constructs, methodological considerations – mainly the quality criteria of Sect. 3.2.3 – and ultimately, theoretical reflections. This shows how the approach of this work – represented in the regional success factor model of Fig. 4.6 – aims to bridge the gap between rigor and relevance by enabling critical reflections on current practices (Kieser and Leiner 2009: 528).
.
Part III
Success Factors of Regional Strategies: A New Perspective on the Geographic Competitiveness of Multinational Corporations
.
Chapter 5
Research Design and Research Methodology
The research design and research methodology of this work is oriented on the theoretically derived latent variables of the regional success factor model, and its respective measurement models, which we presented in the previous chapters. For the measurement of these latent constructs, in addition to utilizing secondary data, the use of primary data is highly encouraged to improve methodological rigor (Bergh et al. 2006: 91; Yang et al. 2006: 603) – as this contributes to avoiding a common method bias (Chang et al. 2010: 179; Homburg and Klarmann 2009: 149; Podsakoff et al. 2003: 882).1 Due to the fact that methodological rigor – according to our explanations in Sect. 2.4.1 – represents a necessary condition for achieving “high” rigor, we will apply both data sources in this work. Here, secondary data is related to the more “objective” firm-level data, whereas primary data focuses on the “subjective” experimental, cultural, and knowledge (or information) related variables (Buckley et al. 2007: 1071; Hult et al. 2008a: 1066; Venkatraman and Ramanujam 1986: 804). In this work, the former is given by a database with firm-level information about MNCs, whereas the latter corresponds to a survey-based inquiry of these firms. In the following chapter, we will present this database of our research sample – before we will outline in the subsequent chapter the research methodology for our survey-based research, as well as for the explorative analysis and modeling of our data.
5.1 5.1.1
Research Design Sample
To test the theoretically derived hypotheses of Sect. 4.1, this work aims to use a representative sample of large MNCs, to ensure the practical relevance of the empirical research results for such companies (Nicolai and Kieser 2002: 584). 1 A common method bias occurs if the same data source is applied for measuring both the independent and dependent variables of a dependence-analytical model (Homburg and Klarmann 2009: 149; Podsakoff et al. 2003: 882).
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_5, # Springer-Verlag Berlin Heidelberg 2011
129
130
5 Research Design and Research Methodology
The research sample of this work consists of the Fortune Global 500 firms, whose regional strategies have a large impact on the world’s economy (Buckley and Ghauri 2004: 81) – given that they account for a great part of the world’s stock of FDI and more than half of world trade (Dunning et al. 2007: 188; Rugman 2000: 2; Rugman 2005b: 224) – and thus are practically relevant for competitors, governments, suppliers, employees, and/or any other stakeholders of these MNCs. The sample period for these firms covers nine reporting years, 2000–2008. For three reasons we believe that this timeframe is adequate to analyze the success factors of regional strategies. First, it is consistent with the long-term strategic perspectives that MNCs apply in building their competitive insider positions at the regional level (Ohmae 1985: 51, 77), either by adapting their products/services to regional markets, and/or by establishing their regional management structures. Second, the period 2000–2008 recognizes the fact that a lag in time might exist between the mid 1990s, where the trend of an increasing regionalization began,2 and the possibility to empirically observe regional strategies of MNCs. Third, the International Accounting Standards (IAS) were introduced at the beginning of the millennium – being later termed International Financial Reporting Standards (IFRS) – which aim to enhance the international comparability of financial MNC data (Rugman and Oh 2007: 32). More importantly, the IFRS as well as other important financial accounting standards for MNCs – such as the Generally Accepted Accounting Principles for the United States (US GAAP) or the Japan GAAP – require the reporting of geographic segments within our sample period. As from 1 January 2009, the earlier segment reporting according to IAS 14 is replaced by IFRS 8 “Operating Segments”, so that firms have to focus on their internal control structure instead of geographical segments in their financial reporting (Epstein and Jermakowicz 2009: 50; Schween 2006: 516). Therefore the 9 years 2000–2008 appear well-suited for the research purpose of this work. Our analysis of companies during this period – which were ranked as Fortune Global 500 firms by the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009) – resulted in a total of 663 MNCs. This number is net of any type of industry consolidation for these firms, such as mergers and acquisitions (M&A), as well as of those MNCs that filed for bankruptcy in this period. On the basis of the countries, where the headquarters of our sample companies are located, we could identify the home regions of these MNCs – utilizing the UNCTAD (2008: xii–xiv) methodology for the regional classification of countries. The application these UNCTAD or UN classifications for grouping countries into regions is common in the IB field (Arregle et al. 2009: 94–95; Dunning et al. 2007: 178–179). This leads to a continental perspective for the regional presence of these MNCs (Ghemawat 2005: 107),3 covering the five regions: Africa, Asia-Pacific, Europe, North America, and South America. Following Rugman and Oh’s (Rugman 2005b: 79–182; Rugman and Oh 2007: 42–43) classification of industries and by utilizing the
2
Cf. Sect. 2.1.2.4. Cf. Sect. 1.2.
3
5.1 Research Design
131
industry information of Hoover’s (2009) firm groupings – according to Standard Industry Classification (SIC) and North American Industry Classification (NAICS) codes – we could classify our sample firms in 16 industries including both manufacturing and service sectors, as illustrated in Table A.1. An overview of the resulting complete research sample of all 663 firms – including their geographic and industry affiliation – is provided in Table A.14. The MNCs of our sample constitute our firm database, which we will present in the next chapter.
5.1.2
Database
Following the methodology applied by Rugman (2005b: 10) – and due to the fact that geographical region data is not supplied by the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009) – we compiled our database from the geographical segment reporting in the annual reports and/or SEC filings of our sample firms. Our research sample of 663 companies, during the 9-year period, 2000–2008, leads to 5,967 firm-year observations in our firm database. For each of these years, we collected the geographically provided information of these firms – according to the five regions mentioned above, supplemented by a “rest” category, if this classification was not possible from the available data – which produced five sorts of regional data: sales revenues, profits, investments, assets, and employees.4 We found that the resulting 179,010 data points were very heterogeneous in seven aspects, for which we collected additional information to improve their comparability. First, the financial information was stated in 24 different currencies for our research sample. Based on the foreign exchange rate fixing by the European Central Bank (ECB) (2009), we translated these currencies into US dollars – which was applied by more than one-third, or 263 MNCs, of our sample firms. Here, according to the accounting rules for currency translation, the assets were translated at the spot rates in effect on each related balance sheet date – whereas the sales, profits, and investments were translated at the average rate for the reporting period of the respective MNC (Schildbach 2008: 123–124). Second, the geographical segment information that we encountered for our sample MNCs was based on different closing dates of their financial statements (e.g., 31.01., 31.03., 30.06., 30.09., or 31.12.). Here, the reporting period from January to December was most frequently applied by more than two-thirds, or 458 companies, of our research sample. Consequently, we used this accounting period to harmonize the different timeframes underlying the published data of our sample firms. For example to harmonize the financial reporting year 2008 of our sample firms, we allocated all their information published for accounting periods up to the end of June (e.g., 31 March, 30 June) to the previous year 2007, while all data reported after June (e.g., 4
Cf. Sect. 2.4.2.
132
5 Research Design and Research Methodology
31 July, 30 September) was recognized in 2008.5 Besides improving the timely comparability of the reported data, the different closing dates were also used to assess the appropriate timeframes for calculating the above mentioned, average currency translation rates. Third, the financial data was reported in different units (e.g., in thousands, millions, billions), and we converted all into millions.6 Fourth, the financial data reported by our sample firms also varied according to its definition (e.g., operating profit, profit before taxes).7 We collected these definitional differences in the reported data of each company, to take them into account for data comparisons.8 Fifth, 194 firms of our research sample did not report geographical segment information, mostly due to the accounting rule that allows them to exclude the reportable data (e.g., sales revenue) for a particular foreign country or region, if it does not exceed 10% of the firm’s total amount or value for this data (e.g., consolidated sales revenue) (Stanko et al. 2002: 101). For these firms with an insignificant presence in foreign countries or regions, we collected only their consolidated number for each of the 5 yearly data sets described above, as a proxy of the data reported on their home region. Given that this approximation sets their reported data as being derived by 100% from the home region, these firms will be termed 100%-HR MNCs in the following. Sixth, the published information on the foreign activities of the firms of our research sample spans different countries and regions, which we allocated to the five regions or to the “rest” category outlined before, to achieve the comparability of their regionally reported data. Seventh, several MNCs applied different accounting standards during the period 2000–2008, mostly due to their application of local accounting rules (e.g., Korean GAAP) in the first years of this period, and their later transition to a more internationally relevant accounting standard (e.g., IFRS or US GAAP).9 By means of collecting this 5 Alternatively, each MNC’s financial year could be portioned to the base period of the research sample (January–December). This would imply, for example, splitting an annual report ending on 31 March into two periods from 1 January–31 March and 1 April–31 December, allocating the former to the previous year and the latter to the actual reporting period. This would not be reasonable, as the annually reported financial statements of a MNC are a cohesive presentation of their situation according to the financial accounting rules relevant at the respective closing date, and is thus not necessarily comparable with its annual report of the previous or next financial year. This is illustrated by the frequent restatements of financial data from one reporting period to another. We considered such effects from restatements by collecting the reported data for each firm in a reverse order – starting with 2008, then 2007, and so on, to improve the longitudinal comparability of each MNC’s reported information. 6 This is due to the fact that most MNCs, or 587 firms, of our research sample published their financial information in millions. 7 Differences in the data reported by these MNCs on their employees included distinctions of yearend and average numbers – each on a full-time equivalent basis – as well as different terminologies (e.g., headcount, personnel, or staff). As most of the MNCs in our research sample reported yearend figures, these differences were not material and thus were not further differentiated in our database. 8 Cf. Sect. 6.1. 9 If a MNC published its financial statements according to both a local accounting standard (e.g., Australian GAAP) and a more international standard (e.g., US GAAP), then the information based
5.2 Research Methodology
133
information on their applied accounting principles, we aim to ensure the comparability of the reporting periods across our sample firms.10 By applying the above described procedures to harmonize the reported geographical segment information of our sample firms – and with respect to our conceptual explanations in Sects. 2.4.2, 2.3.2.1 and 4.1.5 – we could derive different measures of regional success, the regional orientation of these MNCs, and the control variable firm size. The use of such secondary data sources to measure the latent construct regional success is appropriate, given that it is related to firm-level – as opposed to SBU-level – strategies of MNCs (Hult et al. 2008a: 1070–1071; Venkatraman and Ramanujam 1986: 807). Such objective data for the performance measurement of regional strategies is commonly applied in the IB field (Bausch et al. 2007: 155; Chen 2007: 346; Richter 2007: 367–368; Rugman and Oh 2007: 35; Sukpanich 2007: 136–137), as well as for assessing a MNC’s regional orientation (Banalieva and Santoro 2009: 348; Rugman 2005b: 4; Rugman and Verbeke 2004: 7), and its firm size (Arregle et al. 2009: 94; Hutzschenreuter and Gr€one 2009: 1157; Rugman 2005b: 227). Before further analyzing this information of our database, and its utilization in our regional success factor model, we will present our research methodology in the following chapter. Here, besides the methodological foundations for the subsequent exploration and modeling of our data, we will introduce the methodology for our survey-based research.
5.2 5.2.1
Research Methodology Methodology of Explorative Data Analysis
To enhance our understanding of the patterns of latent variables in our regional success factor model, we aim to explore their measurement models and the underlying nature of our hypothesized success factors in more detail. On the one hand, the indicators of regional success will be investigated by means of a data analysis over time. In this longitudinal study of regional corporate success, we will devote special attention to the development of home-regional figures. This selection grounds on the fact that – according to our explanations in Sect. 2.2.2 – a dominance of homeregional success should be expected for our research sample. To analyze if this is true, we will investigate in a longitudinal study from 2000 to 2008 the degree to which such home-regional success dominates over hostregional, bi-regional, or any type of multi-regional (e.g., tri-regional) success. For a differentiation of these forms of regional success, we will utilize the 50% home solely upon the more international reporting standard (here US GAAP) was utilized for comparability reasons. 10 Cf. Chap. 6.
134
5 Research Design and Research Methodology
region and 20% host region thresholds, following the methodology proposed by Rugman (2005b: 4), as illustrated in Sect. 2.2. We will apply these cut-off points both to the sales spread and to the asset dispersion of MNCs, as brought forward by Rugman and Verbeke (2008b: 311) for the analysis of regional success. For the subsequent analysis of particularly home-regional data,11 several methodological aspects should be highlighted here. First, following the methodology applied by Rugman and Oh (2007: 35), we will calculate for each MNC the ratio of home-regional to total sales (HRS/TS) and the ratio of home-regional to total assets (HRA/TA) for the 2000–2008, 9-year time period.12 To assess industry-related differences of these regional success measures (Rugman and Oh 2007: 39), the HRS/TS and HRA/TA will also be analyzed for the manufacturing sector and the service sector as a whole, as well as by single industries, over the sample period. In addition to these two success indicators, we will analyze the ratio of home-regional to total profits (HRP/TP), as this represents another manifest variable of regional corporate success in our regional success factor model.13 Second, to further substantiate our analysis of these three success indicators – as outlined in Sect. 2.4.2 – we will explore additional home-regional figures. This includes the ratio of homeregional investments to total investments (HRI/TI), which represents an important basis for HRS/TS and HRA/TA – as these investments are employed in the home region to link a MNC’s upstream and downstream FSAs with its CSAs in the form
11
It should be noted that in this analysis – by drawing on Rugman and Verbeke’s (Rugman 2005b: 3–4; Rugman and Verbeke 2008b: 311) approach for assessing regional success, as outlined in Sect. 4.1.4 – we investigate different measures of home-regional success. More specifically, we calculate for all of our 663 sample firms the average values of their relative shares of homeregional sales and assets – leading to nine figures for these two regional success metrics over our sample period. This methodology for assessing home-regional success of MNCs is different from the classification of home-regionally successful firms, which we presented before – where the regional success of each company is evaluated separately by Rugman’s (2005b: 4) classification scheme according to the 50% home region and 20% host region thresholds. In the latter approach, the number of those MNCs that are classified as being home-regionally successful is counted – which also leads to nine figures for both the relative amount of sales-based and assets-based homeregionally successful firms. This shows that the respective nine figures derived by the former and the latter methodology have a different meaning. The latter approach takes a broader view on the relative importance of firms being classified as home-regionally successful vis-a`-vis alternative regional success classifications – which corresponds to the methodological basis of the very early influential work of those academic scholars that discovered the phenomenon of the dominance of home-regional success (Rugman 2003b: 412; Rugman 2005b: 4; Rugman and Verbeke 2004: 7). The former approach concentrates specifically on the home-regional success patterns of MNCs – which was applied more recently, to further exclusively explore the development of this dominant form of regional success over time (Oh 2009: 341; Rugman and Oh 2007: 36–37; Rugman and Verbeke 2008c: 328). In this work, we apply both approaches in the same sequence as these IB scholars – following its underlying logic from broad to specific – as both methodological techniques look at the same phenomenon in different ways, and thus contribute to our understanding of the multi-facetted nature of the (home-)regional success of our sample firms. 12 A subtraction of the values of HRS/TS and HRA/TA from a 100% leads to rest of world sales and assets, respectively (Rugman and Oh 2007: 36–37). 13 Cf. Sect. 4.1.4.
5.2 Research Methodology
135
of regional/locational advantages.14 The relative size and importance of HRI/TI will be further analyzed by the MNC’s home-regional investment quota (HRIQ),15 and the difference of this HRIQ to the total corporate investment quota (TIQ). Such investment-based forms of a MNC’s home-regional commitment are often accompanied by other resource allocations to the home region – such as personnel resource commitments that can be well observed by the ratio of home-regional to total employees (HRE/TE) (Ohmae 1985: 111). Furthermore, both HRS/TS and HRP/TP will be explored more deeply, given their importance as success indicators of a MNC’s intra-regional market penetration within the home region,16 by means of analyzing their productivity given in the home-regional sales revenue per employee (HRS/E) and the home-regional profit per employee (HRP/E). Finally, the dynamics of HRS/TS and HRP/TP will be assessed by comparing the firms’ home-regional sales revenue growth (HRSG) with their total sales revenue growth (TSG) – and respectively, the MNCs’ home-regional profit growth (HRPG) with their total profit growth (TPG). In our longitudinal studies, we will calculate all of these values including and excluding the 100%-HR MNCs, to avoid any homeregional bias of MNCs. On the other hand, pursuing the aim of analysis of this work,17 a proper examination of the underlying nature of the hypothesized regional success factors in our structural model – besides regional corporate success – also involves the exploration of regions, regional strategies, and the regional strategy–performance relationship. Consequently, we will present further data observations for these three aspects of MNCs’ approaches to regionalization, by means of applying the following methodology. Regarding the geographical distribution of MNC activities across regions, the number of MNCs operating in each of the five regions mentioned previously will be counted on the basis of the information reported in their annual reports. This sheds light on their geographical expansion patterns within and beyond triad regions, which we aim to analyze more profoundly by examining Ohmae’s (1985: 121–123) tetrahedron model. Following our explanations in Sect. 2.1.2.3, this will be explored by analyzing if MNCs that are active in South America primarily stem from the US or North America, and if those operating in Africa are MNCs from Europe, or if those with activities in the Asia-Pacific are mainly Japanese or Asian MNCs. Concerning regional strategies, we aim to examine the robustness of their classification by Rugman (2005b: 4) into home-regional, host-regional, bi-regional, or any form of multi-regional strategies according to the sales spread of MNCs
14
Cf. Sect. 2.2.8. The home-regional investment quota is calculated by dividing the MNC’s home-regional investments by its sales revenue in the home region. This illustrates the portion of sales that MNCs invest in their home region. 16 Cf. Sect. 4.1.4. 17 Cf. Sect. 1.2. 15
136
5 Research Design and Research Methodology
across regions. Here, we apply the methodology proposed by Osegowitsch and Sammartino (2008: 188–189) to perform a sensitivity analysis of the 50% home region and 20% host region cut-off points. The minimum value for the home region threshold will be increased by 5% steps from 50% up to the maximum value of 95%, whereas the minimum value of the 20% host region cut-off point – in accordance with the procedure applied by Osegowitsch and Sammartino (2008: 188–189) – will be varied by 15% and 10%.18 To ensure that the classifications of the regional strategies of MNCs are not distorted by any fluctuations in their sales revenue in certain years, these sensitivity analyses are based on the classifications resulting from the average sales revenue values for the 9 years from 2000 to 2008. Furthermore, in addition to the stability of their classification, the particular characteristics of regional strategies will also be examined by means of our surveybased inquiry of the Fortune Global 500 firms.19 In this analysis, we will focus on how important the survey respondents consider certain characteristics of regional strategies,20 for which we will also assess industry-related differences.21 This should produce additional insights about how these characteristics vary from one regional strategy to another, or in other words, about the extent to which MNCs place a different emphasis on certain characteristics during their regional expansion. Here, we also aim to investigate if variations in this emphasis of regional strategy characteristics can be explained by a MNC’s administrative heritage according to Bartlett and Ghoshal (1989: 45–53). Given that the influence of administrative heritage becomes evident in a MNC’s overall disposition (centralized versus decentralized) in foreign host regions (Osegowitsch and Sammartino 2008: 186), we will concentrate on non-home-regional MNCs and their preferences for certain coordination mechanisms in their host regions. In this analysis, the coordination mechanisms of further regional strategies – as outlined in Sect. 5.2.2.1 – will serve as proxies for differences in the administrative heritage of MNCs. Following our explanations in Sects. 2.1.2.4 and 2.2.6, we will utilize a MNC’s tendency to implement adaptive-regional coordination forms – which reflect the centralization of functional decision-making synergies at the level of 18
The highest value of 95% for the 50% home region threshold seems reasonable, to include MNCs at a very high level of home-regionalization as well as the 100%-HR firms. To distinguish these 100%-HR firms from the other home-regional MNCs, the 10% cut-off point appears adequate – given the accounting rule outlined above, according to which MNCs have to report their geographical activities in foreign host regions that exceed 10% of the firm’s total amount, or value, of the respective geographical segment reporting data. 19 The methodology for our survey-based inquiry of the Fortune Global 500 firms will be introduced in the next chapter. 20 To differentiate regional strategies form one another, we utilize Rugman’s (2005b: 4) salesbased classification of regional strategies according to the 50% home region and 20% host region cut-off points. The selection of this classification scheme is mainly due to its prevailing application in the IB literature, which ensures the comparability of our results with other studies in this field. 21 Given their importance for our regional success factor model of Fig. 4.6, we will devote particular attention to regional management autonomy, regional product/service adaptation, regional orientation and inter-regional distance in this analysis of industry-specific differences.
5.2 Research Methodology
137
the host region to transfer knowledge and expertise from the regional management center to intra-regional operations – as an approximation of the formalization approaches of North American firms (Bartlett and Ghoshal 1989: 49–50, 161). In addition, again based on our explanations in these previous chapters, a preference for cooperative-regional coordination forms will be used as a proxy for the socialization mechanisms typically encountered in European MNCs, while a tendency for directive-regional coordination of regional activities in foreign host regions will serve as an approximation of centralization approaches of Asian MNCs (Bartlett and Ghoshal 1989: 49, 51, 158–161, 163–165). Consequently, we will investigate the degree to which North American MNCs coordinate their activities in host regions by adaptive-regional approaches, European MNCs by cooperativeregional coordination forms, and Asian MNCs by directive-regional coordination mechanisms. As regards the regional strategy–performance relationship, regional strategies will be examined concerning their effect on total corporate performance. This analysis investigates the degree to which the regional geographic spread expressed in the regional strategies of MNCs – corresponding to the latent variable regional orientation – leads to different corporate performance outcomes. Given the importance of this moderator regional orientation in our structural model, which reflects the regional multinationality of MNCs (Contractor 2007b: 16; Verbeke et al. 2009: 150),22 we aim to profoundly analyze this regional strategy–performance relationship. Consequently, based on the 50% home region and 20% host region cut-off points of Rugman (2005b: 4), various geographic segment information reported by the MNCs will be applied for classifying their regional strategies – including their regional sales revenue, profits, assets, and employees.23 As measures of the total corporate performance, the average values of a MNC’s consolidated sales revenue published by the magazine Fortune (2005, 2006, 2007, 2008, 2009) for the 5 years, 2004–2008, will be calculated – to ensure that this performance data is based upon largely comparable international accounting standards.24 By means of the explanations above, we have outlined the methodology of the explorative data analyses of this work. These explorative studies largely draw on the objective information of our database, and to a lesser extent on the subjective data of our survey-based inquiry of the Fortune Global 500 firms – which will be analyzed mainly by means of the SEM approach PLS.25 Before outlining the
22
Cf. Sect. 2.3.2.1. These four elements of the MNCs’ geographical segment information were chosen, as they are also published by the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009) on a consolidated basis. This allows an examination of their overall plausibility – which we perceived as being necessary here, given the importance of this analysis. 24 We will show in Sect. 6.1 that – from 2004 to 2008 – most MNCs prepare their financial statements according to IFRS, US GAAP, or Japan GAAP. The focus on corporate sales revenue and profits is based on Rugman’s (2005b: 231) definition of corporate success, as a combination of sales-based success and the profitability of a MNC. 25 Cf. Sect. 6.3. 23
138
5 Research Design and Research Methodology
methodology for modeling this survey-based data in PLS, we will present the main methodological aspects of developing and implementing this survey in the next chapter.
5.2.2
Survey Methodology
5.2.2.1
Survey Design
According to Rugman (2005b: 228) and in conformity with our aim of analysis,26 an in-depth study of the success factors of regional strategies – in addition to the data analyses outlined previously – should rest upon surveys of MNC managers. Consequently, we focus on survey-based, quantitative inquiry techniques in this work.27 Here, we target those MNC managers of our 663 Fortune Global 500 sample firms that are corporate experts for the regional strategies of their companies. Given that we concentrate on the MNC as a whole in our research,28 this particularly includes managers at the corporate headquarters who are responsible for the worldwide corporate strategy and the international business development of their firm.29 As we expect considerable time constraints of the managers in these departments, and due to the fact that – as illustrated in Table A.14 – the headquarters of our sample firms are situated all over the globe, a web-based questionnaire was chosen as being most adequate for conducting our survey (Meffert et al. 2008: 159).30 Such inquiries in written form leave sufficient time for well thought-through responses, and thus should improve the quality of the answers to our survey (Schnell et al. 2005: 359). After these preliminary considerations of our survey design, we divided the outline of the questionnaire into three parts: contact and introduction, company details and main content, and respondent information (Meffert et al. 2008: 158–159). The first part of contact and introduction consists of three sub-sections. First, a short introduction in the opening of the questionnaire briefly presents the objective and academic background of the survey – and offers a firm-specific
26
Cf. Sect. 1.2. For a comparison of quantitative and qualitative research methods cf. Creswell (2009) and De Vaus (2002: 5–7). 28 Cf. Sect. 2.1.2.1. 29 The resulting key informant bias will be controlled by various measures, which we will describe below in the presentation of the outline and the pretest of our survey. 30 For creating our online questionnaire, we utilized the online survey software provided by SurveyMonkey (http://www.surveymonkey.com). The selection of this web-based software was mainly due to its advantages for contacting the MNC managers of our research sample, given for example by design aspects (such as optical appeal, seriousness) and its functionality (such as progress bars). 27
5.2 Research Methodology
139
company report to MNC managers,31 to encourage their participation. Another important condition for their participation is included in the general remarks of the second section, which highlights that the purpose of this survey is solely academic, and not commercial – and which briefly gives some directions for answering the questionnaire. In the third section, a preliminary note and general definitions introduce the main topics and most relevant definitions of the survey. The second part of company details and main content is composed of two subsections. First, general information about the company is requested – including the name of the MNC and details on its regional presence and experience, as well as on its competitors at the regional level. The second section of this part represents the main content of our survey, which largely results from operationalizing the measurement models of those latent variables – as presented in Sect. 4.1 – that should be assessed by primary data. This includes regional management autonomy, following Enright’s (2005a: 65) and Enright’s (2005b: 89) approach for measuring this latent construct. The same applies to the extent of regional product/service adaptations, which are difficult to assess by other means than by subjective knowledge (Rugman and Verbeke 2008b: 311). In addition, the inter-regional distance faced by the MNC: “[. . .] is necessarily subjective” (Ghemawat 2001: 147), and thus is as well observed by means of this part of our survey. Following our explanations in Sect. 4.1.3.1, subjective estimates of the market-related regional orientation are also assessed here. Furthermore, to control and measure regional success by both secondary and primary data (Hult et al. 2008a: 1072),32 regional competition and regional success are also assessed by subjective estimations in this part of the questionnaire.33 Additional information requested here is related to the explorative analyses – outlined in Sect. 5.2.1 – and to details on the insiderization of MNCs. Primary data corresponding to the former includes subjective preferences for the different coordination forms of further regional strategies,34 while the latter is 31
In this company report, the specific results of the responding MNC are compared with the average results of the survey as a whole. 32 As explained in Sect. 5.1.2, the control variable firm size will be measured by secondary data. 33 In our regional success factor model, as outlined in Sect. 4.1.4, regional success will be measured solely by means of objective data. However, to draw proper conclusions on the quality of regional performance measurement (Hult et al. 2008a: 1072), subjective estimations of regional success are also assessed to validate the secondary data of our database (Meffert et al. 2008: 159). In assessing this criterion validity of our success measurement (Fritz 2004b: 29), we focus on the subjective estimations of regional sales revenue and profit generation, which – according to Rugman (2005b: 231) – are the most important indicators of a MNC’s performance. 34 These different coordination mechanisms are assessed by the decision-making structures of MNCs. Here, directive-regional decision-making usually implies that the parent headquarters takes most decisions about regional products/services. In adaptive-regional decision-making, generally the regional management center(s) develop(s) many products/services for the national subsidiaries within the region. A manifestation of a trans-regional (world mandate) coordination mechanism is that the regional management center(s) develop(s) many products/services for the national subsidiaries outside the region. Furthermore, the establishment of cooperative-regional decision-making structures typically leads to the fact that the national subsidiaries in the region participate strongly in the development of regional products/services within the region.
140
5 Research Design and Research Methodology
assessed by the direction of communication in the MNC’s know-how transfer and by different measures of its regional commitment (Ohmae 1985: 206).35 In the third part of respondent information, extensive information on the respondent is requested – representing a possibility to control for the key informant bias (Fritz 2004b: 26–30; Haenecke and Forsmann 2006: 50; Hurrle and Kieser 2005: 598; Kieser 2004: 3–9; Nicolai and Kieser 2002: 584; Nicolai and Kieser 2004: 633). Besides the age of the informants, this includes their management level,36 management experience, functional area, and functional experience. These measures aim to assess the competency and experience of the key informants of our research sample, and the degree to which their responsibilities are closely associated with the regional strategies of their company (Kumar et al. 1993: 1635). By means of this respondent information, the objectivity as well as the reliability and validity of our primary data is assessed – which constitute important conditions for conducting success factor research based on key informants (Fritz 2004b: 27–30; Haenecke and Forsmann 2006: 50; Hurrle and Kieser 2005: 598; Kieser 2004: 8).37 After demanding this information related to the key informant bias – in this last part of the questionnaire – the email address of those respondents who would like to receive a firm-specific company report is requested, before the survey ends by acknowledging their participation. Regarding the questioning techniques in our survey, we utilized closed questions for the company and respondent information, as this allows a standardized evaluation of alternative answer choices (Meffert et al. 2008: 160; Schnell et al. 2005: 330). For the main content in the second part of the survey however, we utilized statements on five-point Likert scales,38 as this part focuses on the subjective estimates of the responding MNC managers – which are better captured by the extent to which they agree or disagree with such statements (Meffert
35
Measures of a MNC’s regional commitment include its emphasis placed upon management continuity (e.g., continuity of the senior management team), market-oriented investments (i.e., resource commitments to attract potential foreign customers such as sales offices, plants, R&D centers), and its staying power (i.e., staying in a region while it faces periods of difficulty, such as economic downturns). Cf. Sect. 2.1.2.3. 36 A MNC’s management levels were differentiated by the reporting structures within the firm (Stopford and Wells 1972: 10), which led to the distinction of senior management – those MNC managers that form part of the Board of Management or directly report to the Board of Management – and middle management with no direct reporting to the Board of Management. 37 It should be noted that multi-informant designs can also lead to measurement problems (Homburg and Klarmann 2009: 148). Due to the fact that usually only very few respondents in the MNCs of our research sample – apart from the Board of Management and the corporate strategy department – are deeply involved in the conceptualization of the firm’s regional strategies, a key informants approach appears to be more reasonable for the research of this work (Fritz 2004b: 26–27, 30). 38 Latent constructs that are measured by means of Likert scales for multiple indicators, can be interpreted as metric scales – which are required to apply the PLS approach (Chin et al. 2003: 199; G€ otz and Liehr-Gobbers 2004: 721, 733).
5.2 Research Methodology
141
et al. 2008: 86).39 Furthermore, in this section – given its importance for measuring the latent variables of our regional success factor model – we utilized two additional techniques to control for potential biases. First, by clearly separating the information requested in our survey, we aim to avoid a common method bias in the presentation of our statements (Homburg and Klarmann 2009: 149; Podsakoff et al. 2003: 882; Temme and Paulssen 2009: 125–126). Second, we controlled the respondents’ attention paid to these statements by reversely coding some of the survey statements (De Vaus 2002: 168). The methodology described previously in the design of our survey aims to ensure that the primary data provided by our respondents leads to an adequate measurement of the latent variables in our regional success factor model – which represents an important condition for testing their hypothesized relationships (Schnell et al. 2005: 211). In the course of implementing our survey, additional techniques were applied to methodologically ensure a proper measurement of our latent constructs by their outer models.
5.2.2.2
Survey Implementation
In the implementation of our survey – following our explanations in Sects. 3.2.3.1 and 3.2.3.2 – we tested the content and expert validity of our reflective and formative measurement models by means of an indicator’s proportion of substantive agreement and by its substantive-validity coefficient (Krafft et al. 2005: 77), as well as by two subsequent in-depth pretests of our survey. In assessing the indicator’s proportion of substantive agreement and its substantive-validity coefficient, we concentrated on those latent variables that are exclusively measured by means of primary data – including regional management autonomy, regional product/service adaptation, and inter-regional distance. Three research assistants of our university participated in this preliminary test of content and expert validity. While we could realize good to high values in this test of both the indicator’s proportion of substantive agreement and its substantive-validity coefficient,40 we were able to identify 6 of 48 manifest variables that were wrongly allocated by more than half of the participants. Consequently, we devoted particular attention to these indicators in the subsequent pretests.
39
In the case of the five-point Likert scale, a “don’t know” field – whereas for closed questions an “other” field – was also added. To better understand the reasoning behind selecting these answer choices, for both questioning techniques, we asked the respondents to specify reasons or to provide additional details for choosing these answers. These scales of our survey have been developed for the purpose of this work and have not been utilized before. 40 The respective test results – relating to the psa index of the indicator’s proportion of substantive agreement, and to the csu index of its substantive-validity coefficient – will be presented in more detail in Sect. 6.3.
142
5 Research Design and Research Methodology
We conducted two in-depth pretests of our survey,41 with valuable input from experts that are highly recognized and experienced either in the area of regional strategies and/or with the methodological issues of this work.42 The main changes resulting from these two pretests can be divided into two categories. The first category included adjustments in the survey’s content – such as wording changes or additional examples – and in the sequence of its sections, statements, and questions.43 The second category relates to methodological changes of the questionnaire – such as a separation of the survey for respondents from home-regional and from non-home-regional MNCs – and inquiry techniques.44 Two adjustments of the second category are particularly important for the research of this work, and thus should be highlighted here. First, Verbeke proposed to make the respondents of non-home-regional MNCs choose one foreign region outside of their home region for the survey completion – where their firm has been the most successful (e.g., in terms of sales, profitability, and/or sales growth). This change leads to the application of our regional success factor model only to the most successful region of our sample firms – given either by their home region or by one of their foreign regions in the case of non-home-regional MNCs – which constitutes a very meaningful
41
We conducted our pretest over 6 weeks from the end of February 2009 to the beginning of April 2009. 42 The six academic experts included Alan M. Rugman of the University of Reading (UK), Alain Verbeke from the University of Calgary (Canada), Andre´ Sammartino from the University of Melbourne (Australia), Michael Behnam from Suffolk University in Boston (MA, USA), and Dirk Ulrich Gilbert and J€ urgen K€ahler from the University of Erlangen-N€urnberg (Germany). The cultural background of these academic experts relates to Asia, Europe, and North America – which represent the most frequently encountered regions of our research (cf. Sect. 6.1). This is an important condition for the cross-cultural application of our survey, for which we aim to ensure that the numbers on the response scales, or the items that the respondents are responding to, have the same meaning across cultures (Hult et al. 2008b: 1036). In addition, the six practitioners were comprised of two heads of corporate strategy in each the manufacturing and the service sector, as well as one director and one associate principal of McKinsey & Company. The high crosssectional experience and competence of these practitioners regarding regional strategies of MNCs appeared necessary to ensure that our survey can be applied to the diverse industries of our sample firms. Furthermore, the six methodological experts included German professionals in survey-based research from the Gesellschaft f€ ur Konsumforschung (GfK) and TNS Infratest GmbH, as well as other questionnaire specialists and PLS modeling experts. 43 For example, our pretest participants advocated a change in the order of regional management autonomy and regional product/service adaptation, as a beginning with the latter area of interest was perceived by our pretest participants as facilitating the respondents’ involvement with the theoretical concepts of the survey. 44 Concerning the inquiry techniques, some pretest participants observed that respondents could also strongly agree to some of the reversely coded survey statements – even though here we would expect a disagreement. Consequently, to avoid such wrong implications of our control questions, we decided to directly ask the respondents, if we would encounter such ambiguous cases – where they answered to the reversely coded statements in the opposite direction of our expectation – in the returned responses to our final survey.
5.2 Research Methodology
143
approach for the study of regional success factors.45 Second, the pretest participants of McKinsey & Company proposed adding the phrase “in relation to competition” to those statements that can be externally observed – such as a MNC’s regional product/service adaptation and the subjective estimates regarding its regional orientation and its success.46 These changes led to the final survey of this work, which is illustrated in Fig. A.1. To ensure that the survey is completed by respondents with profound knowledge about the regional strategies of their firm, we directly contacted all 663 MNCs of our sample via telephone. For the identification of appropriate key informants, we utilized the listings of the Hoover’s (2009) database on the names, functions, and departments of possible respondents in our sample firms. Besides a verification of the adequacy of our respondents and their email addresses, we were able to shortly describe the background of our research and the benefits of their participation in our survey – mainly given by the firm-specific company report. Based on these verified contact details of our respondents,47 we sent our survey to all our 663 sample firms.48 After sending a total of five remainders to each contact at our sample MNCs, we received a total of 114 responses – corresponding to a response rate of 17.2%.49 This is beyond the usual expected response rate of cross-national mail surveys of between 6% and 16% (Harzing 1997: 643; Harzing 2000: 244), and thus was perceived as satisfactory. The responses of 18 firms had to be eliminated due to
45
Due to the fact that also non-home-regional MNCs (e.g., a bi-regional or tri-regional MNC) may be most successful in their home region, for these MNCs we included the possibility to complete the questionnaire for their home region. It only had to be assured that the respondents of our sample firms complete the survey for only one, or more specifically, their most successful region. 46 Throughout all these adjustments, particular attention was paid to the total response time of the survey, which should not exceed a maximum of 15 min. Furthermore, by means of a professional proofreading service, the linguistic accuracy of the English survey was confirmed. 47 The contact details of the respondents mostly consisted of their direct email addresses (in 633 cases). Mainly due to their company policy, 30 respondents did not want to give out their email address and preferred either the contact form on their firm’s webpage or a fax for the transmission of the survey. 48 The transmission of our survey was always accompanied by a brief email-based cover letter and the link to our online questionnaire. The content of this cover letter was largely identical to the first page of our survey – illustrated in Fig. A.1 – including a deadline for completing the questionnaire within 4 weeks after initial transmission. 49 From 24 April 2009 to 8 June 2009, we sent our survey to 621 MNCs, and from 4 August 2009 to 4 September 2009 to another 42 firms of our sample. our questionnaire was due to the fact that the Fortune Global 500 firms for the year 2008 were published by the magazine Fortune (2009) on 20 July 2009 – thus after our initial survey transmission phase. We followed the same procedure as in the earlier phase – including the direct contacting of our respondents and the number of remainders. The response rate of 23.8% for the later responding 42 MNCs was even above the response rate of 16.7% for the earlier 621 responding firms. This might be due to the fact that during our first transmission period, many MNCs were preparing their annual financial statements – leading to time constraints for the participation in our survey. However, regarding the content of their survey responses, we could not find any systematic differences between early and late respondents.
144
5 Research Design and Research Methodology
incomplete or implausible answers,50 leading to a total of 96 useable responses for our 663 sample MNCs, or 14.5%. Throughout all these steps of our survey methodology, we aimed to ensure a proper observation of those latent variables of our regional success factor model that are based upon such primary data. In the following, we will briefly describe the methodology applied for modeling both this survey-based data and the secondary data of our database in PLS.
5.2.3
Modeling Methodology
To properly apply the PLS approach to the regional success factor model of Fig. 4.6, three methodological aspects for modeling its latent variables should be highlighted, before evaluating the respective empirical results. This includes the methodology utilized for modeling the latent constructs’ measurement models, latent second-order variables, and interaction effects. For a proper modeling of the measurement models of our latent variables in PLS, several methodological adjustments to their respective indicator data were necessary. These changes related both to our survey and to our database, as the indicators of the latent variables in our regional success factor model are derived from both primary and secondary data sources. A first set of adjustments included a transfer of the reversely coded statements in the survey into the scale logic of the other responses. Second, the scales of all indicators in our regional success factor model that are derived from our database – including the manifest variables of regional success, firm size, and regional orientation – were transformed into scales similar to our survey-based primary data. For regional success and firm size, in a first step, this involved calculating the average values of their indicators for the 5 years, 2004–2008 – to ensure that this secondary data is based upon largely comparable international accounting standards, as mentioned in Sect. 5.2.1. In a second step, all indicators of regional success – including regional sales to total sales, regional assets to total assets, and regional profits to total profits – and of firm size, given by the total MNC revenue in US dollars, were coded from one to five. Appropriate intervals for this coding were derived by centering the indicators – thus by assuming that the value of three corresponds to the arithmetic mean of each of these indicators (Chin et al. 2003: 198–199). The addition of one or two standard deviations led to values of four and five, while a respective subtraction led to values of two and one. Furthermore, we assigned values of one to seven for a 50
The implausibility of responses was assessed by reversely poled items and by an overall judgment of the answers. To understand the reasons for not completing the survey and for the encountered implausible answers, each of the respondents for these 18 sample firms was contacted, before we eliminated their responses. The main reasons for not completing the questionnaire were time constraints or company policy matters (e.g., no participation in strategy-related surveys).
5.2 Research Methodology
145
MNC’s regional orientation – ranging from home-regional (7), host-regional (6), bi-regional (5), tri-regional (4), quad-regional (3), global (2), to “rest-regional” (1) – as these values represent different degrees of this latent construct (Rossiter 2002: 323).51 Third, those answers that were not completed by the respondents, or for which they selected the “don’t know” field, were coded as missing values (Huber et al. 2007: 79). The existence of missing values, mainly due to unreported information in the annual reports of our survey sample MNCs,52 led to a reduction from 96 to 36 modeled cases in PLS – as here, only those latent constructs are recognized which are completely observed by their manifest variables.53 Even though 36 cases represent a small number of observations, they still fall into the minimum range of 30–100 cases, as mentioned in Sect. 3.1. Due to the fact that PLS is based on the estimation of a series of ordinary least squares analyses, it allows the modeling of large models with such very small sample sizes (Herrmann et al. 2006: 39). However – particularly with respect to the more complex, latent second-order variables of our structural model of Fig. 4.6 – we aim to provide a more detailed analysis of their modeling in light of such a small sample size. As these sample size restrictions particularly concern latent variables that are specified in a formative manner (Chin and Newsted 1999: 327), we will concentrate on the formative measured latent second-order construct regional management autonomy in this analysis. Here, according to Giere et al. (2006: 687–688), two approaches in particular have received considerable attention in literature for modeling formative second-order constructs: the hierarchical component model and the factor scoresbased approach. In the hierarchical components model by Wold (1980), the indicators of the first-order dimensions are defined as direct manifest variables of the latent second-order construct – also known as the repeated indicators approach (Giere et al. 2006: 688; Lohm€ oller 1989: 128–133; Wold 1982: 39). For this modeling approach however, we found that the bootstrapping method did not produce reasonable results – given by our observation that those path coefficients with a larger distance from zero had lower t-values. Even after further controls for
51
A “rest-regional” orientation of a MNC results for example, if a substantial part of the firm’s revenue is derived from the “rest” category in its geographical segment reporting. Such “restregional” orientations of MNCs can be explained by their orientation to emerging markets – such as the BRIC countries, Brazil, Russia, India, and China – if the corresponding financial data was allocated to the “rest” category in their geographic segment reporting, instead of one of the five regions that we outlined before in Sect. 5.1.1 (Banalieva and Santoro 2009: 347). 52 The number of missing values increased considerably by the lack of annual reporting information on the regional profitability of our survey sample firms – which was published by only 54 of these MNCs. 53 One of these 36 survey respondents completed the questionnaire for a foreign region that was not declared in an isolated form in the respective firm’s annual reports. Based on an analysis of the financial statements of this MNC, we utilized the figures of the reported “other” figure in its geographical segment reporting as the best possible approximation of its regional success in this particular foreign region.
146
5 Research Design and Research Methodology
multicollinearity and of the underlying indicator data (Giere et al. 2006: 687),54 these implausible results persisted for the formative measured latent second-order construct regional management autonomy. By taking a closer look at the formula for calculating t-values (Smith et al. 2008: 186), we found that these results can be explained by the small sample’s high standard deviation of the large number of noncollinear formative indicators of this multidimensional construct. As their standard deviation enters in the denominator of this formula, low t-values result – which means, in other words, that the significance of the relationships in our structural model could not be confirmed on the basis of this small sample size. Consequently, we applied the factor scores-based approach that utilizes the factor scores of the four first-order dimensions as direct indicators of the latent second-order construct (Giere et al. 2006: 688; Wetzels et al. 2009: 180).55 After controlling for effects of multicollinearity in these first-order factor scores,56 this still led to 36 cases – which is above the required sample size of at least 20 observations, following a rule of thumb for the minimum number of cases for a proper modeling in PLS.57 Apart from the fact that they fulfill the sample size requirements for modeling the regional success factor model of Fig. 4.6 – according to Giere et al. (2006: 688, 691) – factor scores should be preferred over alternative methods of indexation for formative manifest variables, such as the mean values of the formative indicators (Albers and
54
Regarding multicollinearity, we evaluated the VIF values of all sets of formative indicators in SPSS (2009) – where all VIF values were lower than their limit value of 10 (Giere et al. 2006: 687), as outlined in Sect. 3.2.3.2. By applying an even more conservative approach – assuming that any VIF substantially greater than one indicates multicollinearity (Henseler et al. 2009: 302) – we were able to ensure that no multicollinearity exists in this modeling alternative (given for example by the fact that no indicator weight loadings were greater than one). The underlying indicator data was checked mainly regarding the comparability of the applied objective performance data – as explained in Sect. 5.1.2 – and the appropriateness of the subjective responses to reversely coded questions. Here, no discernible implausibilities in the underlying indicator data could be found. 55 An indexation of the formative indicators for each of these four first-order dimensions is possible, as they fully capture the respective constructs’ domain of content – and as they share a common meaning, given by changes in the degree of regional management autonomy (Diamantopoulos et al. 2008: 1212; Diamantopoulos and Winklhofer 2001: 271–272). 56 We calculated the factor scores after having ensured that no multicollinearity exists between the formative indicators. Following our explanations before, this was achieved by analyzing VIF values on the basis of SPSS (2009) – assuming that any VIF substantially greater than one indicates multicollinearity (Henseler et al. 2009: 302). By means of these analyses, we could identify two formative indicators (the coordination of regional legal entities and operations, and the regional liaison center for the parent company) which – to avoid effects from multicollinearity, such as for example indicator weight loadings greater than one – we did not use in the calculation of the factor scores. 57 As a rule of thumb, sample size in PLS should at least be equal to the larger of the following: (1) five times the scale with the largest number of formative indicators (the scales for constructs specified in a reflective manner can be ignored); or (2) five times the largest number of structural paths directed at a particular latent variable in the structural model (Chin et al. 1996: 39; Sosik et al. 2009: 15; Tabachnick and Fidell 1989: 129). Here, the first criterion of the rule of thumb applies to the case at hand, resulting in the minimum number of 20 observations for a proper modeling the regional success factor model of Fig. 4.6.
5.2 Research Methodology
147
Hildebrandt 2006: 25–26; Diamantopoulos et al. 2008: 1212; Diamantopoulos and Winklhofer 2001: 271).58 This might explain why the factor scores-based approach has been applied in recent times by several other researchers for modeling formative measured latent second-order constructs (Bruhn et al. 2008: 1298; Reinartz et al. 2004: 298; Venaik et al. 2005: 663, 675). Another implication from formative measured latent second-order constructs relates to the modeling of interaction effects. These effects are usually modeled by means of interaction variables – which are calculated as the Cartesian product of the exogenous variable and the moderator variable, as explained in Sect. 3.2.1. More specifically, this implies a pairwise multiplication of the (standardized or centered) indicators of the exogenous variable with the moderator variable (Eggert et al. 2005: 108). However: “[. . .] since formative indicators are not assumed to reflect the same underlying construct (i.e. can be independent of one another and measuring different factors), the product indicators between two sets of formative indicators will not necessarily tap into the same underlying interaction effect” (Chin et al. 2003: Appendix D). Therefore, in the case of formative indicators, the pairwise multiplication of manifest variables is not possible (Eggert et al. 2005: 108). Consequently, following the modeling recommendation of Eggert et al. (2005: 108, 112), in a first step, we derived the factor scores of the exogenous variable and the moderator variable from PLS – and, in a second step, calculated the interaction variable by multiplying the standardized values of their factor scores.59 On this basis, we were able to properly model the main effects and the interaction effects in PLS. By means of the previous explanations, we have outlined the most important preliminary considerations for the modeling of the regional success factor model of Fig. 4.6 in PLS. Before evaluating the respective results of this structural model in more detail, we will describe and explore the patterns of its latent variables on the basis of our primary and secondary data sources in the following.
58
The disadvantage of utilizing mean values is given by the fact that the influence of the nonexplained variance, or the measurement error, is not explicitly included in the estimation of model parameters, which may lead to biases in the estimates (Albers and G€otz 2006: 674; Homburg and Baumgartner 1995: 1092, 1102–1103). 59 This two-step procedure has also been described by Chin et al. (2003: Appendix D).
.
Chapter 6
Empirical Results
6.1
Descriptive Statistics
The basic features of our data will be described for our research sample of 663 MNCs and our survey sample of 96 MNCs – as the empirical results of this work draw upon both of these secondary and primary data sources. To gain insights into the geographical and sectoral origin of our research sample, we will describe these 663 MNCs along their home country and their home region according to the location of their corporate headquarters (Rugman 2000: 9; Rugman and Verbeke 2004: 4) and their industry membership (Rugman 2005b: 79–182; Rugman and Oh 2007: 39, 42–43). Furthermore, based on company data, we will illustrate how our sample firms develop on average over the sample period 2000–2008 – utilizing Fortune Global 500 data on a consolidated basis issued by the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009).1 For a better understanding of the comparability of this company data over time, we will describe the financial accounting standards applied by these MNCs during the sample period. In addition, regarding the regional data provided by our sample firms, we will give an overview on the terminology applied in their annual reports for this geographical segment information. The descriptive statistics on the home country and the home region of our sample MNCs are depicted in Fig. A.2. This shows that the largest part, or 210 firms, of our research sample are US companies, followed by 108 Japanese, 45 French, 45 German, and 43 British MNCs. Besides these five most frequent home countries of our sample firms – nearly all of their, in total, 36 home countries form part of the broad triad of North America, Europe, and Asia-Pacific. More specifically, in terms of their home region, 234 firms (or 35.3%) are from North America, 230 (or 34.7%) from Europe, and 192 (or 29.0%) from the Asia-Pacific – and only 7 companies (or 1.1%) are from South America, whereas no MNC is from Africa.
1
This data for our 663 sample firms on a consolidated basis will be described by means of the information provided by the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009), as our database of Sect. 5.1.2 is based upon regional data in an unconsolidated form (i.e., which does not take into consideration any consolidation-related reconciliations or eliminations).
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_6, # Springer-Verlag Berlin Heidelberg 2011
149
150
6 Empirical Results
Consequently, the home regions of 656 firms of our research sample (or 98.9%) form part of the broad triad. The industry membership of our sample MNCs is illustrated in Fig. A.3. Here, we first observe that, with 359 firms (or 54.1%), the number of manufacturing firms exceeds the number of 304 (or 45.9%) service MNCs in our research sample. Second, taking a closer look at those industries which dominate these two categories, we notice that the most frequently encountered industries in the manufacturing sector are from the energy, petroleum and refining industry (122 firms or 18.4%) – while most MNCs of the service sector are banks (72 firms or 10.9%). With respect to the company data of our research sample, which is expressed in average values of all sample companies in Fig. A.4, we observe an upward trend of the MNCs’ sales revenue, assets, and their number of employees from 2000 to 2008 – where the development of assets more than doubles during this time. However, the profitability of these Fortune Global 500 firms rather fluctuates – given by its decline from 2000 to 2002, then its continuous rise until 2007, before its latest drop by almost 50% in 2008.2 The financial accounting standards applied by the 663 MNCs from 2000 to 2008 are displayed in Fig. A.5. Here, we notice that US GAAP is applied over this sample period by 36–37% of our sample firms. Quite strikingly, we observe that the portion of our sample MNCs applying IFRS increases substantially – at the expense of local accounting standards, for example Japan GAAP, Korean GAAP, or Australian GAAP. This leads to the fact that from 2004 to 2008, the overwhelming part of our sample firms, or more than 60%, apply either US GAAP or IFRS – while the portion of those MNCs utilizing local accounting standards further decreases.3 For the geographical segment information presented in the annual reports of our sample firms – as illustrated in Table A.2 – five categories of regional data can be distinguished: sales revenues, profits, investments, assets, and employees. With respect to each category, different definitions and terminologies are applied by our sample firms. However, we observe that these differences can be largely integrated within each category – as concerns 88–100% of our sample firms for sales revenue, investments, assets, and employees, and almost 60% for profits. Consequently, we generally will not differentiate between these definitional and terminological variations in the following.4 Besides these secondary data-based descriptions of our research sample, we have more detailed information on a subset of 96 of these firms – given by our 2
The regional performance patterns of these MNCs will be presented in the explorative analysis of this data in the following chapter. 3 From 2004 to 2008, the most frequently encountered accounting standard was Japanese GAAP – which was applied by 35–45% of those MNCs utilizing local accounting standards during this period. 4 In other words, all the five categories are adequate approximations of the geographical segment information provided by our sample firms. Therefore, we will draw on these different definitions and terminologies, only if such more detailed analyses are required.
6.1 Descriptive Statistics
151
survey sample, which we will describe shortly. Here, we will first outline the differences of this subset compared to the larger research sample, with respect to the home regions of these 96 MNCs and their industry membership. In addition, using the questionnaire-based company information for this survey sample, we will describe the regional presence and experience of these firms in more detail – including their regional operating experience, regional management center experience, and regional subsidiary density. Furthermore, the objectivity, reliability, and validity of this primary data provided by key informants of our survey sample will be assessed – as outlined in Sect. 5.2.2.1 – by describing respondent information on the key informants’ age, management level, management experience, functional area, and their functional experience. As depicted in Fig. A.6, differences in the home regions of the 96 MNCs in comparison to the research sample are mainly given by a higher portion of European MNCs (49.0%) and a reduced share of North American firms (20.8%) in the survey sample. Concurrently, the percentage of firms from the Asia-Pacific and South America is relatively stable around 29%, or 1%, respectively. Overall, the portion of MNCs from the broad triad regions is above one-fifth in both samples. Regarding the industry membership of the companies in our survey sample visa`-vis those of our total research sample – illustrated in Fig. A.7 – we notice an increase of more than 5% points in the share of manufacturing firms (59.4%) and a simultaneous decrease in the number of service MNCs (40.6%). Furthermore, we observe that our survey sample – apart from aerospace and defense, and entertainment, printing and publishing – is composed of all industries of our total research sample. Here, similar to the 663 sample firms, the most frequently encountered industries within the manufacturing and service sectors are energy, petroleum and refining (17.7%), and banks (12.5%), respectively. More detailed company information on the survey sample firms concerning their regional presence and experience is illustrated in Table A.3. Here, we recognize that the overwhelming part of these 96 MNCs has a substantial regional operating experience – as nearly half of these companies (46%) have been operating for more than 30 years in either their most successful foreign region or in their home region.5 Even though to a lesser extent, their regional management experience is also considerable – given by the fact that most of these MNCs (45%) have been utilizing a regional management center for more than 20 years, in either their most successful foreign region or in their home region. Interestingly, most survey sample firms (50%) have a low subsidiary density of one to six national units in either their most successful foreign region or in their home region – while also a notable number of MNCs (19%) have more than 30 national units at the regional level.6 5 It should be noted here, that 63 firms of our survey sample have chosen to answer the questionnaire for their most successful foreign region, whereas 33 companies responded the survey on behalf of their home region. 6 Even though the open-ended responses on the competitors of our survey sample firms could not be categorized into distinctive descriptive statistics, two observations should be noted in this regard. First, the respondents of our 96 sample MNCs largely stated regional or even local
152
6 Empirical Results
Based upon the respondent information in Table A.4, we provide details on our key informants at the 96 MNCs that participated in the survey of this work. Concerning their age, we notice that most of our respondents (39%) are between 41 and 50 years old. With respect to their management level, most key informants (61%) are from middle management – whereas a substantial part (39%) of our survey sample firms are senior managers, thus forming part of the Board of Management or directly reporting to the Board of Management. Furthermore, most respondents (31%) have considerable management experience, having worked for 6–10 years at these management levels. The largest portion of our key informants is from the functional area business development or strategy (68%) and from general management (10%) – while the share of all other functional areas is below 10%. Within each functional area, our key informants have notable functional experience, ranging from 6 to 10 years for most of our survey respondents (27%). Using the previous descriptive statistics, we were able to describe the basic features of our primary and secondary data sources. Furthermore, we showed that the survey sample – with respect to the industry membership of these MNCs – is largely a representative subset of the total research sample (von der Lippe and Kladroba 2002: 140). Concerning their home regions however, our survey sample firms are less representative for the total research sample – while at least one-fifth of the MNCs of both samples have their home regions within the broad triad regions of North America, Europe, and the Asia-Pacific. Based on additional data descriptions, we provided several important insights about the MNCs of the different samples, and about the key informants of those firms participating in our survey. In the following, to gain a more profound understanding of both samples, we will explore their respective data in more detail.
6.2 6.2.1
Explorative Data Analysis Data Analysis over Time
The secondary data of this work will be analyzed over time – following the methodology presented in Sect. 5.2.1 – regarding different forms of regional success and the development of home-regional figures. The analysis of homeregional, host-regional, bi-regional, and multi-regional success over time – based upon the regional sales spread of the total research sample, as depicted in Fig. A.8 – sheds light on the relative proportionalities of these different forms of regional
competitors for either their most successful foreign region or their home region. Second, more than two-thirds of their competitors (71.9%) were from the region selected by our key informants for completing the questionnaire – thus from either their most successful foreign region, or their home region.
6.2 Explorative Data Analysis
153
success. Here, we observe that home-regional success represents the most dominant form of regional success for more than 80% of the 663 sample firms from 2000 to 2008. However, we find that the portion of MNCs with home-regional success decreases from 86 to 80% during this period – while the portion of bi-regionally successful companies increases from 5 to 10%. All other forms of regional success, including host-regional, tri-regional, quad-regional,7 and rest-regional success,8 remain below 4% during our sample period – while no MNC could be identified that is globally successful in all five regions from 2000 to 2008. A similar analysis – based on the regional asset dispersion of the 663 sample companies, as illustrated in Fig. A.9 – shows even more unambiguously the dominance of home-regional success, as more than 87% of our sample firms are home-regionally successful from 2000 to 2008. Due to this high percentage of assets-based home-regional success, we observe only a low and relatively stable portion of MNCs with biregional success (between 3 and 6%), host-regional-success (between 3 and 4%), and tri-regional success (around 1%) during this period – whereas no MNC being either quad-regionally or globally successful could be identified within this timeframe. Given this high portion of MNCs that are successful within their home region, we will explore the development of home-regional figures in more detail – according to the methodology outlined in Sect. 5.2.1. By focusing exclusively on the home regions of all our 663 sample firms, we find that – as depicted in Fig. A.10 – the HRS/TS of these MNCs range on average between 76% and 78% from 2000 to 2008. Even after controlling for a home region bias of the 100%-HR MNCs, the HRS/TS remain on a high level between 67 and 68%. Similarly, as portrayed in Fig. A.11, the HRA/TA of these companies range between 79 and 80% over the 9 years of our sample period and – after excluding the 100%-HR MNCs – only decrease marginally to still high values between 69 and 72%. An analysis of industry-specific differences for the HRS/TS and HRA/TA in Figs. A.12 and A.13 shows that MNCs of the service sector are substantially more home-regionally successful than manufacturing firms – with differences of 15 up to 19% points from 2000 to 2008, or of 7 up to 15% points after the exclusion of the 100%-HR MNCs. By means of a more detailed analysis of the single industries’ average values for HRS/TS and HRA/TA over this sample period – as illustrated in Table A.5 – we notice that the three least home-regionally successful, or most internationalized, industries are chemicals and pharmaceuticals, computer, office and electronics, and
7 It should be noted, however, that the portion of tri-regional MNCs rises from 2.7% in 2000 to 3.7% in 2008 – while only two quad-regionally successful MNCs (SAB Miller in 2007 and Citigroup in 2008) could be identified, generating more than 20% of their sales revenue in four of the five regions Europe, North America, Asia-Pacific, South America, and Africa. 8 According to our explanations in Sect. 5.2.3, a “rest-regional” success of a MNC may result from the fact that a substantial part of the firm’s revenue is derived from the “rest” category in its geographical segment reporting (e.g., if the firm generates considerable revenue in emerging markets like the BRIC countries, which form part of this “rest” category).
154
6 Empirical Results
food, drug and tobacco. The values of these three industries range between 54 and 65%, or between 54 and 64% after excluding the 100%-HR MNCs. Based upon a longitudinal study of HRP/TP, as displayed in Fig. A.14, we find high fluctuations for this home-regional success measure from 2000 to 2008 – whereas generally, an upward trend can be observed from 2000 to 2006, with more recent decreases in 2007–2008. Over the whole sample period, the HRP/TP ranges between 56 and 96%, or between 25 and 93% if controlling for the 100%-HR MNCs. By means of exploring the HRI/TI over the sample period – as displayed in Fig. A.15 – we find that our 663 sample MNCs highly invest within their home region. The percentages of their investments within the home region range from between 78 and 81% for these 9 years, or between 63 and 71% if excluding the 100%-HR MNCs. However, if relating these investments to sales revenue, we observe that the respective HRIQ – as depicted in Fig. A.16 – decreases by approximately 2.5 percentage points from 2000 to 2008, also after an exclusion of the 100%-HR MNCs. Nevertheless, a more detailed analysis of these reductions – by means of the difference between the HRIQ and the TIQ, as portrayed in Fig. A.17 – shows that from 2000 to 2006 and in 2008, the investments spent for the home region are above those of all other regions of the sample MNCs. Only in 2007, no considerable variations between the HRIQ and the TIQ were found – whereas in all other years, the HRIQ exceeds the TIQ by 2–8% points, or by 4–14% points if controlling for the 100%-HR MNCs. Another form of home-regional resource commitments is given by the HRE/TE that – as illustrated in Fig. A.18 – remains relatively constant around 76–80% from 2000 to 2008, or around 59–64%, after excluding the 100%-HR MNCs. If assessing the productivity of these home-regional employees by the HRS/E – as depicted in Fig. A.19 – we recognize considerable increases from 0.3 to 2.2 million US dollars over the sample period. An exclusion of the 100%-HR MNCs leads to lower, but still increasing, values of the HRS/E from 0.2 to 0.7 million US dollars during the first 8 years of this period – while in 2008 a drop to 0.5 million US dollars can be observed. This drop in the last year of our sample period is even more severe in the case of the HRP/E. Here, even after excluding the 100%-HR MNCs – as shown in Fig. A.20 – we find an upward trend from the year 2000 of approximately 20,000 US dollars to values of around 60,000 US dollars per home-regional employee in 2007. In 2008 however, this trend for HRP/E inverts tremendously – reaching levels of 60,000 US dollars, or 5,000 US dollars if controlling for the 100%-HR MNCs. When exploring the dynamics of home-regional figures, given by a comparison of the HRSG with the TSG, and of the HRPG with the TPG – as illustrated in Figs. A.21 and A.22 – we notice high fluctuations from 2001 to 2008.9 However we find that, even after an exclusion of the 100%-HR MNCs, the home-regional sales
9
Given that our database includes the reporting periods of 2000–2008, the growth figures can only be calculated for the 8 years beginning in 2001.
6.2 Explorative Data Analysis
155
growth surpasses the corporate revenue growth in 6 years of the sample period – whereas the home-regional profit growth solely exceeds the corporate profit growth in the years 2005, 2006, and 2008. By means of the longitudinal studies above, we were able to explore the patterns of different forms of regional success, and particularly, of home-regional success and its related figures for the 663 MNCs of our research sample from 2000 to 2008. In the next chapter, these examinations of regional success will be complemented by further data observations, to gain additional insight into the underlying nature of the hypothesized regional success factors in this work.
6.2.2
Further Data Observations
Following the methodology outlined in Sect. 5.2.1, further data observations will be provided for regions, regional strategies, and the regional strategy–performance relationship. The analysis of the geographical distribution of MNC activities across regions – as illustrated in Table A.6 – shows that most MNCs (at least 399 firms or 60%) of our total research sample are operating in the broad triad regions of North America (70.9%), Europe (60.3%), and Asia-Pacific (60.2%). However, we also observe that several of our sample firms operate in South America (141 firms or 21.3%) and in Africa (65 firms or 9.8%). By means of a more profound examination of these geographical expansion patterns of our 663 sample companies, we find evidence for Ohmae’s (1985: 121–123) tetrahedron model – as depicted in Table A.7. Here, we observe that European firms tend to operate more in Africa than the other MNCs of our research sample. Furthermore, US or North American companies are more active in South America than MNCs from Europe or from the Asia-Pacific.10 In addition, Japanese or Asian MNCs have more activities in the Asia-Pacific than companies from Europe or from North America.11 Apart from these regional expansion patterns, we investigated the relative amount of sales revenue generated in these regions and its impact on the classifications of regional strategies. First, we analyzed the sensitivity of these classifications by holding constant the 20% host region cut-off point, while increasing the minimum value for the home region threshold by steps of 5% points from 50% up to the maximum value of 95% – which is displayed in Fig. A.23. Here, we notice that the portion of MNCs, which are classified as following home-regional strategies, decreases from 85 to 43% throughout these steps. Simultaneously, the share of firms being classified as pursuing bi-regional strategies increases from 9 to 26%. The portion of MNCs classified as pursuing host-regional and tri-regional strategies 10
The 8% of US or North American companies in Table A.7 relate to 38 firms, of which most are US MNCs (35 firms or 92%). 11 In Table A.7, the 42% of Japanese or Asian MNCs with activities in the Asia-Pacific correspond to 116 MNCs. Most of these MNCs (89 firms or 77%) are home-regional companies – mainly from Japan (76 firms or 66%).
156
6 Empirical Results
remains below 5% in these sensitivity analyses – whereas no MNCs with quadregional or global strategies could be identified. Furthermore, due to the fact that stepwise increases of the minimum value for the home region threshold reduce the number of MNCs with home-regional strategies, an increasing number of MNCs is classified as following rest-regional strategies, if they generated any sales revenue within the “rest” category in their geographical segment reporting. Second, by following the same procedure after changing the constant host region threshold to 15% – as illustrated in Fig. A.24 – we find that the MNCs with home-regional strategies comparably decrease from 84 to 41%. However, the portion of MNCs with bi-regional strategies (5–28%) and tri-regional strategies (6–10%) notably increases. Again, the percentage of MNCs with host-regional strategies remains below 5% – while a small number of firms with quad-regional strategies (below 1%) and no MNCs with global strategies could be found. Third, by performing these sensitivity analyses based upon a constant host region threshold of 10% – as depicted in Fig. A.25 – we observe that the portion of MNCs with home-regional strategies again decreases from 84 to 41%. Concurrently, the share of firms classified as pursuing bi-regional strategies (from 3 to 24%) and the percentage of those with tri-regional strategies (from 7 to 18%) increases considerably. Once more, the portion of firms with host-regional strategies does not exceed 5% – while more firms with quad-regional strategies (around 2%) and also a few MNCs following global strategies (below 1%) could be identified. Besides these sensitivities in their classifications, we examined the particular characteristics of regional strategies based upon the information provided by the 96 survey sample MNCs. By means of analyzing the survey responses provided by MNCs following different regional strategies,12 as shown in Fig. A.26, we observe that regional management autonomy increasingly gains importance from homeregional, over host-regional and bi-regional, to tri-regional strategies of MNCs.13 Concurrently, the emphasis placed on regional product/service adaptations appears to be closely related to the inter-regional distance perceived by the MNC. This leads to an inverted U-curve from home-regional, over host-regional and bi-regional, to tri-regional strategies regarding the importance assigned to regional product/service adaptations by the respective MNCs. Interestingly however, while the inter-regional distance slightly increases for MNCs following tri-regional strategies, their respective emphasis placed upon the regional adaptation of their products/services further decreases along the inverted U-shape. The market-related regional orientation of our survey sample firms is perceived as moderate in the case of MNCs following home-regional and tri-regional strategies – while its lowest and highest values 12
The survey responses were analyzed regarding the respondents’ rating along the five-point Likert scales, corresponding to their perceived importance of different characteristics of regional strategies. 13 These were the only regional strategies that could be identified for our 96 survey sample firms, by applying Rugman’s (2005b: 4) sales-based classification of regional strategies according to the 50% home region and 20% host region cut-off points – which were chosen mainly due to comparability reasons, as explained in Sect. 5.2.1.
6.2 Explorative Data Analysis
157
are assigned by firms with host-regional, and respectively, bi-regional strategies. This pattern indicates an inverse relationship of the MNCs’ subjective perceptions of their market-related regional orientation and the inter-regional distance they face. In addition, industry-related differences concerning the respondents’ ratings for regional management autonomy, regional product/service adaptation, regional orientation, and inter-regional distance are illustrated in Figs. A.27–A.30. Here, for these four items, we find notable industry-specific differences within both the manufacturing sector and the service sector. The emphasis placed upon regional management autonomy and regional product/service adaptation was highest for firms from the food, drug and tobacco industry in the manufacturing sector, and respectively, for MNCs from the telecommunications and utilities industry as well as from the other services industry in the service sector. The subjective estimation of market-related regional orientation is particularly high for companies from the computer, office and electronics industry in the manufacturing sector – and for firms from the telecommunications and utilities industry in the service sector. The inter-regional distance is perceived as highest by MNCs from the food, drug and tobacco industry in the manufacturing sector, and by companies from the transportation services industry in the service sector. Furthermore, the values assigned to the different coordination forms of further regional strategies are illustrated in Fig. A.31. Here, we find that directive-regional coordination mechanisms received the highest rating by MNCs following triregional strategies. Adaptive-regional coordination forms were most important to firms with host-regional strategies, while home-regional and bi-regional companies attached the highest priorities to cooperative-regional coordination. Trans-regional (world mandate) coordination mechanisms were not perceived as most important within any particular regional strategy pursued by our MNCs – while it received the highest rating by firms with host-regional strategies. Moreover, the importance of inter-regional know-how transfer and intraregional knowledge sharing for MNCs with different strategies is depicted in Fig. A.32. This shows an upwards trend in the values assigned to the interregional know-how transfer from home-regional, over host-regional and biregional, to tri-regional strategies of MNCs. Here, the highest rating for the inter-regional know-how transfer was given by firms with tri-regional strategies, which perceived both directions as equally important (i.e., from the home region to foreign regions, and vice versa). High weights for the intra-regional knowledge sharing were placed by MNCs with bi-regional strategies, whereas no considerable differences in its importance could be found for the firms pursuing any other regional strategies. Finally, the importance that MNCs with different strategies attach to various aspects of their regional commitment is displayed in Fig. A.33. Here, we find that staying in regions while they face periods of difficulty was rated relatively highly by all MNCs of our survey sample – while firms pursuing host-regional strategies placed the highest weight on this form of regional commitment. Management continuity was perceived as important by MNCs following home-regional, bi-regional, and tri-regional
158
6 Empirical Results
strategies – whereas market-oriented investments were perceived as most important by companies with home-regional and bi-regional strategies. Such variations in the emphasis placed by our survey MNCs on different characteristics of their regional strategies were further analyzed with respect to the administrative heritage of these sample firms – as depicted in Table A.8. Here, we find that for non-home-regional MNCs from Europe, cooperative-regional decision-making was perceived as most important in the coordination of their regional operations. On the contrary, North American firms assigned the highest values to adaptive-regional coordination mechanisms, while MNCs from the AsiaPacific placed the highest weights on directive-regional coordination forms. In addition to these observations about different aspects of regional MNC strategies, we examined the regional strategy–performance relationship – given by their effect on total corporate performance, as illustrated in Fig. A.34. Here, irrespective of the classification logic applied for regional strategies – including sales revenue, profits, assets, and employees – we observe an S-curve development of the MNCs’ total corporate performance from home-regional, over host-regional and bi-regional, to tri-regional strategies.14 By means of the further data observations presented earlier, we were able to produce additional insights about the MNCs of our total research sample and the firms of our survey sample – with respect to their regions, regional strategies, and the regional strategy–performance relationship. Such an in-depth analysis of these main concepts, which represent the substance of regionalization theory and practice, is an important precondition for an effective research of regional success factors (Verbeke et al. 2009: 157–158). In the following, the results of this research – based on their modeling in the regional success factor model of Fig. 4.6 – will be presented.
6.3
Model Evaluation
Following our explanations in Sect. 3.2.3, we will evaluate the regional success factor model of Fig. 4.6 in two steps. First, we will assess the quality of the reflective and formative measurement models, before evaluating the structural model in a second step. The reflective measurement models of the regional success factor model of Fig. 4.6 are evaluated by their respective quality criteria – content validity, indicator reliability, construct reliability, and discriminant validity – which we outlined 14
The plausibility of the consolidated figures of the MNCs’ sales revenue, profits, assets, and employees published by the magazine Fortune (2005, 2006, 2007, 2008, 2009) was confirmed by their comparison with the respective information in our database. Taking into account the usual consolidation procedures applied by the MNCs of our total research sample, no considerable differences could be identified between the data published by Fortune (2005, 2006, 2007, 2008, 2009) and our database.
6.3 Model Evaluation
159
in Sect. 3.2.3.1. The content validity of the reflective specified latent variables, which are measured by primary data, was confirmed with respect to their expert validity, as mentioned in Sect. 5.2.2.2. More specifically, the psa index of the indicators’ proportion of substantive agreement for regional product/service adaptation was 0.71, and 0.90 for inter-regional distance. The csv index of the manifest variables’ substantive-validity coefficient was 0.67 for the indicators of regional product/service adaptation, and 0.77 for the outer model of inter-regional distance. Furthermore, we analyzed the indicators of the reflective measurement models of our latent variables – including regional orientation, inter-regional distance, regional success, and regional product/service adaptation and its first-order dimensions – concerning their underlying factor structure by means of an explorative factor analysis. Here, the appropriate allocation of the indicators to their latent variables regional orientation, inter-regional distance, and regional success was confirmed.15 The same applies to 13 of the 16 indicators of the multidimensional construct regional product/service adaptation. More specifically, two of its indicators were extracted as a separate component,16 whereas one indicator was extracted as forming part of another latent first-order variable.17 These explorative results do not reflect the results of the previous expert validity tests for these three manifest variables,18 and thus should be complemented by additional quality criteria for these indicators – prior to adjusting the respective reflective measurement models (Homburg and Giering 1996: 8–11). For the indicator reliability of these three indicators, we found that the loadings of those two indicators, which were extracted as a separate component, did not exceed the required minimum value of 0.4.19 Subsequently to the elimination of these two indicators, the reliability of all
15
For the indicators of inter-regional distance and regional success, only one component was extracted, each with factor loadings above 0.4 for the respective manifest variables. In the case of regional orientation, two components were extracted – with one separate component for the subjective market-related regional orientation. However, the measurement of regional orientation as a latent second-order construct with two first-order dimensions – where a single latent first-order variable contains only one indicator – would not be meaningful, as this would violate the identification of both the first-order dimensions and of the latent second-order construct (Baumgartner and Homburg 1996: 144; Bollen 1989: 249; Giere et al. 2006: 687). These considerations confirm the appropriateness of measuring regional orientation by its three conceptually derived indicators (cf. Sect. 4.1.3.1). 16 These indicators were the labeling and the diversity of products/services, representing manifest variables of the first-order dimensions regional design and regional offering portfolio. 17 After controlling for the two indicators that were extracted as a separate component, the innovativeness of regional design – an indicator of this first-order dimension – was extracted as forming part of the latent first-order variable regional functional scope (with respective factor loadings above 0.4). 18 Concerning their proportion of substantive agreement, only labeling had an psa index of 0.33, while the other two indicators reached values of 0.67 for the innovativeness of regional design, and respectively, 1.00 for the diversity of products/services. 19 The loading of the manifest variable labeling was 0.014, while the diversity of products/ services had a loading of 0.081.
160
6 Empirical Results
remaining reflective manifest variables was affirmed, as illustrated in Table A.9.20 This table also depicts the results for the quality criteria of construct reliability and discriminant validity. Concerning construct reliability, we find that the Cronbach’s alphas for the reflective outer models of all latent variables – apart from regional orientation – exceed the minimum value of 0.5. However, according to internal consistency, a more appropriate quality criterion than Cronbach’s alpha for construct reliability,21 we find that all reflective measurement models are beyond its threshold value of 0.7. As regards discriminant validity, we notice that the reflective measurement models of all latent constructs, besides regional orientation, have an AVE above its minimum value of 0.5. Even though the latent variable regional orientation does not achieve the respective thresholds of 0.5 for two quality criteria of reflective measurement models, in both cases the violations are not substantial – with a Cronbach’s alpha of 0.48 and an AVE of 0.47. Consequently, this overall picture of the quality criteria depicted in Table A.9 – following our explanations in Sect. 3.2.3.1 – does not require any further modifications of these reflective measurement models for regional orientation, inter-regional distance, regional success, and the multidimensional construct regional product/service adaptation.22 The quality criteria for the formative measurement models of the regional success factor model of Fig. 4.6 include content or expert validity, indicator relevance, and external or nomological validity.23 As mentioned in Sect. 5.2.2.2, we could confirm the content or expert validity of all indicators of the formative specified multidimensional construct regional management autonomy. More specifically, the psa index for regional management autonomy, reflecting its manifest variables’ proportion of substantive agreement, was 0.81 – whereas the csv index of its indicators’ substantive-validity coefficient was 0.71. The remaining quality criteria are applied to the four indicators of regional management autonomy that – as outlined in Sect. 5.2.3 – are given by the factor scores of its former first-order 20
In Table A.9, we observe that the loadings of all indicators are above the required value of 0.4, and that each manifest variable achieves the minimum significance level of at least 5% in the onesided t-test. 21 Cf. Sect. 3.2.3.1. 22 For the constructs regional orientation and regional success, we also analyzed the relationship between their values based upon secondary data and their subjective estimations by the survey respondents. For the exogenous variable regional orientation, we found a positive relationship of 0.068 between the objective and subjective market-related regional orientation. This correspondence between the primary and secondary data sources for these indicators of a MNC’s marketrelated regional orientation shows that they are well suited to reflect this aspect of regional orientation in the PLS modeling of this variable. For the endogenous construct regional success, we found a positive relationship of 0.136 between the objective and subjective sales-based measures – whereas the relationship between the primary and secondary data sources for the profit-based regional success indicators showed a slightly negative relationship of 0.040. This difference is at least marginal, which confirms the appropriateness of using the objective figures of our database in the PLS modeling. Furthermore, the utilization of objective figures in PLS ensures the comparability of the modeling results with the explorative data analyses related to regional success, as these also draw exclusively upon the secondary data sources of this work. 23 Cf. Sect. 3.2.3.2.
6.3 Model Evaluation
161
dimensions regional strategy development, regional market and product/service development, regional market coordination, and regional operational administration.24 An evaluation of the weights of these four indicators, as illustrated in Table A.10, shows a notable indicator relevance of these four manifest variables – whereas the bootstrap method did not generate t-values that exceed the required minimum values.25 Even though the latter result implies little indicator relevance, an elimination of any of these formative indicators should only be considered in the case of multicollinearity that – according to our explanations in Sect. 3.2.3.2 – constitutes another important quality criterion of indicator relevance. However no, or only very moderate, effects of multicollinearity could be identified – considering different quality criteria, such as the correlation matrix,26 the tolerance, VIF, and CI values of these formative indicators.27 Furthermore, with respect to their external or nomological validity, we find that the direction and size of the formative indicators,28 apart from regional market and product/service development, correspond both to their respective theoretically derived values in Sects. 2.2.7 and 4.1.1 and to their empirical values (Enright 2005a: 66–67; Enright 2005b: 99–100).29 The negative sign of regional market and product/service development implies that the respective decision-making of regional management negatively contributes to the degree of its management autonomy. This effect may be explained by the fact that most MNCs of 24
This shows how the elimination of the outer model by the factor scores-based approach leads to a transfer of the first-order latent variables in indicators of the formative construct regional management autonomy (Giere et al. 2006: 689). 25 As depicted in Table A.10, the weights of all formative manifest variables are above 0.3 – while no indicator exceeds the minimum significance level of at least 5% in the two-sided t-test. Due to the fact that those weights with a larger distance from zero have higher, and thus plausible, t-values – following our explanations in Sect. 5.2.3 – the low absolute figures for these t-values can be explained by the standard deviation of these non-collinear formative indicators. 26 In the correlation matrix, only regional market coordination correlates moderately with regional market and product/service development – given by their correlation of 0.7, the threshold value for multicollinearity. All other formative indicators have correlations below this maximum value, ranging from 0.3 to 0.6. 27 As illustrated in Table A.10, the tolerance values of these four indicators range from 0.4 to 0.7, and thus are considerably above their critical value of 0.1. Their VIF values range between 1.3 and 2.3, which are also notably below their maximum permissible value of 10. Even if assuming a more conservative threshold – i.e., that any VIF values substantially greater than one indicate multicollinearity (Henseler et al. 2009: 302) – this VIF value range only implies a moderate multicollinearity. The same applies to their CI, as the CI values for the formative indicators are between 12.1 and 14.7 – representing the lower end of its range from 10 to 30 for moderate multicollinearity. Cf. Sect. 3.2.3.2. 28 It should be noted that the significance of the indicator relationships could not be utilized for evaluating the external or nomological validity, due to the fact that no indicator exceeds the minimum significance level of at least 5% in the two-sided t-test, as outlined previously. 29 By means of our conceptual explanations in Sects. 2.2.7 and 4.1.1, we showed that these four indicators should be important for the degree of decision-making autonomy granted to regional management – which is reflected in the size of all formative indicator weights. This also conforms to Enright’s (2005a: 66–67) and Enright’s (2005b: 99–100) empirical results on the importance of these activities for the roles of regional management.
162
6 Empirical Results
our survey sample have an extensive regional operating experience and regional management center experience.30 At this advanced stage of regional management experience, such entrepreneurial roles of regional management centers like regional market and product/service development lose importance in comparison to integrative coordination roles (Lasserre 1996: 36; Yeung et al. 2001: 167).31 Due to the reduced importance of these entrepreneurial roles of regional management centers, they: “[. . .] are eventually eliminated and replaced by direct control and co-ordination by parent companies” (Yeung et al. 2001: 167). This reasonably explains how the responsibility for regional market and product/service development might turn into highly negative effects on the decision-making autonomy of regional management. Consequently, considering the extensive regional management experience of our survey sample companies, the external or nomological validity of the formative measurement model of regional management autonomy can also be confirmed. After having ensured the soundness of our reflective and formative measurement models, we evaluate the relationships of their respective latent variables in the structural model. For the evaluation of the structural model, as outlined in Sect. 3.2.3.3, we distinguish between the basic structural model and the extended structural model.32 The quality criteria for the evaluation of the basic structural model are depicted in Table A.11. Here, we observe that the portion of explained construct variance for regional success – given by its coefficient of determination – reaches a moderate level of 0.46 above its threshold value of 0.4. Furthermore, with respect to their effect size, we find that regional management autonomy, regional product/service adaptation, and inter-regional distance have medium effects between 0.16 and 0.22 on regional success – while regional orientation has only a small effect of 0.07. Concerning the prediction relevance of the structural model, we notice that the sizes of the path coefficients of all latent variables – including latent first-order and latent second-order constructs – range from 0.2 to 0.4, and thus exceed the critical value of 0.1. Furthermore, the required significance level of at least 5% in the twosided t-test is achieved by all latent constructs – apart from regional management autonomy and regional orientation that are significant at the 10% level. These results also imply that all latent first-order variables of regional product/service adaptation are important defining characteristics of this multidimensional construct.33 With
30
Cf. Table A.3 and Sect. 6.1. This also explains why – as depicted in Table A.10 – regional market coordination has the highest weights of all formative indicators. 32 Due to the fact that the latent second-order constructs are evaluated like latent first-order variables – as described in Sect. 3.2.3.3 and illustrated in Table 3.4 – their respective quality criteria will be presented together with those of the other latent constructs of both the basic structural model and of the extended structural model. 33 Given the formative specification of the relationships between the first-order dimensions and the latent second-order construct regional product/service adaptation, we analyzed the latent first-order variables concerning their multicollinearity. The respective tolerance values in the 31
6.3 Model Evaluation
163
Regional orientation
Regional management autonomy
0.200*
– 0.308* H1
R2 = 0.46 Regional success H2
0.315** Regional product/service adaptation
– 0.364***
Inter-regional distance
Fig. 6.1 Basic structural model of regional success factors Source: Own illustration. Note: Asterisks denote significance levels of two-sided t-test: * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level
respect to our research hypotheses, the sign of the path coefficient for regional product/service adaptation, recognizing its significance level of 5%, provides strong support for H2 – whereas H1 is rejected at the 10% significance level by the respective sign for regional management autonomy. Moreover, we observe the direct effects of the contextual variables regional orientation and inter-regional distance on regional success – which in the case of the former are positive and significant at the 10% level, and strongly negative at the 1% significance level for the latter. Finally, the value of 0.29 for the Stone-Geisser-test is above its threshold value of zero – implying overall a notable predictive relevance of the basic structural model. Recognizing the explanations above concerning its evaluation, an overview of the resulting basic structural model is given below in Fig. 6.1.34 range from 0.4 to 0.6, VIF values from 1.6 to 2.8, and CI values from 18.0 to 21.1 imply low, at most moderate, levels of multicollinearity. Therefore, the first-order dimensions can be assumed to be linear independent of each other. 34 Taking into account the theoretically derived importance of the contextual variables regional orientation and inter-regional distance in Sects. 2.3.2 and 4.1.3, we also illustrated the direct effects of this internal and external context of the MNC in Fig. 6.1.
164
6 Empirical Results
After the introduction of the control variables firm size and regional competition, only minor changes to the hypothesized relationships are observed.35 The research hypothesis H2 is still supported – even though only at the 10% significance level – whereas the hypothesized relationship in H1 is rejected now at the 5% significance level. The Stone-Geisser-test value decreases only marginally to 0.28,36 which shows that – even after introducing the control variables firm size and regional competition – the overall prediction relevance of the basic structural model remains relatively stable. The evaluation of the extended structural model according to the respective quality criteria is illustrated in Table A.12.37 This shows that the consideration of interaction effects leads to a substantial increase in the endogenous variable’s coefficient of determination to a value of 0.63. Furthermore, the effect size of regional management autonomy and inter-regional distance also increases – where values of 0.4 in both cases show a large effect size. However, the effect size of regional product/service adaptation decreases considerably to 0.02 – while regional orientation has only a very marginal, barely noticeable, effect size. The interaction variables, given by the Cartesian product of the exogenous variables and the moderator variables,38 have a small and medium interaction effect size. The interaction effect size for two interaction variables is small, with a marginal value for the interaction variable of regional orientation and regional management autonomy – and a higher, but still small, value of 0.1 for the interaction variable of interregional distance and regional management autonomy. A medium interaction effect size is observed for the other two interaction variables of regional orientation and regional product/service adaptation, and of inter-regional distance and regional product/service adaptation – with values of 0.31 and 0.27, respectively. With respect to the prediction relevance of the structural model, we find that the size of the path coefficients of nearly all its latent variables – apart from regional orientation, and the interaction variable of regional orientation and regional management autonomy – exceeds the threshold value of 0.1. Furthermore – again apart from these two latent constructs, as well as from regional product/service adaptation – almost all latent variables reach the required significance level of at
35
The respective quality criteria are depicted in Table A.11. Here, we observe that the coefficient of determination slightly increases to 0.49. At the same time, the effect size of the formerly described latent variables decreases to a value range from 0.08 to 0.13. In addition, we also calculated the effect size of the control variables, where we found a small effect of 0.05 for firm size – while regional competition with an effect size of 0.01 has almost no influence. 36 Cf. Table A.11. 37 In a comparison of the basic structural model with the extended structural model, we could not identify any major changes in the formative and reflective measurement models underlying their respective latent variables. Therefore, here we exclusively focus on the level of the inner model – concerning changes from the basic structural model to the extended structural model. 38 Following our explanations in Sect. 5.2.3, this involved the multiplication of the exogenous constructs regional management autonomy and regional product/service adaptation with the moderating variables regional orientation and inter-regional distance, respectively.
6.3 Model Evaluation
165
Regional orientation – 0.076 Regional management autonomy
0.438**
H3a
H3b 0.017
– 0.472*** H1
R2 = 0.63 Regional success H2
0.100 Regional product/service adaptation
H4a
– 0.268*
H4b
– 0.419**
0.400**
Inter-regional distance
Fig. 6.2 Extended structural model of regional success factors Source: Own illustration. Note: Asterisks denote significance levels of two-sided t-test: * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level
least 5% in the two-sided t-test. Once more, this also implies the importance of all latent first-order variables of regional product/service adaptation as defining characteristics of this multidimensional construct. Concerning our research hypotheses, H1 is rejected again – now at the 1% significance level – whereas H2 is not supported. This is due to the fact that, even though the sign of the path coefficient of regional product/service adaptation confirms the research hypothesis H2, the respective t-value is not significant. Furthermore, the signs of the path coefficients of the interaction variables of regional orientation and regional product/service adaptation, and of inter-regional distance and regional product/service adaptation – both at significance levels of 5% – provide strong support for H3b and H4b. This means that the positive relationship between a MNC’s regional product/service adaptation and its regional success is strongly positively affected by its regional orientation and its inter-regional distance. At the same time however, the signs of the path coefficients for the other interaction variables point in the opposite direction of the research hypotheses, implying that H3a and H4a are rejected. Here, the rejection of H4a – which relates to the interaction variable of interregional distance and regional management autonomy – is significant at the 10%
166
6 Empirical Results
level. In addition, we notice the direct effects of the moderating variables on regional success – which are marginally positive and not significant in the case of regional orientation, while strongly negative for inter-regional distance at the 5% significance level. Finally, the value of the Stone-Geisser-test increases to 0.41 – which overall implies a high predictive relevance of the extended structural model. Considering the quality criteria outlined before, an overview of the extended structural model is depicted in Fig. 6.2. Also in the case of the extended model, the introduction of the control variables firm size and regional competition leads only to minor changes in the hypothesized relationships.39 Again, the research hypothesis H1 is rejected – now with a significance at the 5% level – and H2 is also not supported, once again due to its missing significance. In addition, the hypothesized relationships in H3b and H4b are confirmed – with a significance of the former at the 5% level and the latter at the 10% level. Furthermore, H3a and H4a are rejected once again – while these rejections were not significant. Moreover, the value of the Stone-Geisser-test only slightly decreases to 0.40,40 which demonstrates the overall stability of the prediction relevance of the extended structural model. By means of evaluating the measurement models and the structural model of the regional success factor model of Fig. 4.6, we were able to uncover the relationships underlying its latent variables, particularly with respect to our research hypotheses. In the following, mainly drawing upon these insights – as well as upon the descriptive statistics and explorative analyses of this chapter – we will discuss our research findings concerning their implications for theory and practice.
39
Referring to Table A.12, we notice a marginal change in the coefficient of determination to 0.64. The effects of all endogenous latent variables are almost the same – with small to medium values for their effect size, or respectively, their interaction effect size. Furthermore, in an analysis of the control variables’ effect size, we find that both firm size and regional competition have a small effect of 0.02 on regional success. 40 Cf. Table A.12.
Chapter 7
Discussion of Results and Implications
The empirical results of the previous chapter shed light on the success factors of regional MNC strategies, or more specifically, on the extent to which regional management autonomy and regional product/service adaptation lead to regional success – taking into consideration the influence of contextual variables on these relationships. We substantiated this novel perspective on the regional competitiveness of MNCs with further research findings about regions, regional strategies, regional success, and the regional strategy–performance relationship. Collectively, these empirical results have both theoretical and practical implications that we will present in the following.
7.1
Discussion of Theoretical Implications
The theoretical implications of our empirical results, referring to the aim of analysis of this work,1 are comprised of two issues. First, our research results broaden the existing conceptual and theoretical knowledge about regions, regional strategies, regional success, and the regional strategy–performance relationship. Second, by building on these conceptual extensions in a regional success factor model, we advance regionalization theory by insights into the factors that drive the regional success of MNCs. Concerning the first issue, we found corroborating evidence for the concentration of regional MNC activities in the broad triad regions of North America, Europe, and Asia-Pacific. These geographical terrains represent not only the home regions of almost all firms of our total research sample, but also reflect the main areas of their regional expansion. This evidence is consistent with earlier findings on the triad-oriented market penetration of MNCs (Rugman 2000: 114–122; Rugman and Verbeke 2004: 5) – which is largely explained as a result of historical developments of regionalization, such as the liberalization of trade and investments within NAFTA, the EU, or ASEAN (Ghemawat 2005: 100; Rugman 1
Cf. Sect. 1.2.
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_7, # Springer-Verlag Berlin Heidelberg 2011
167
168
7 Discussion of Results and Implications
2005b: 215–216; Rugman and Verbeke 2004: 5). These developments lead to a semi-globalized world of incomplete cross-border integration for different types of region-based markets (i.e., for products, capital, labor, and knowledge) – where both the barriers and the links among these markets give reasons to the intra-triad expansion of MNCs (Ghemawat 2003: 139; Ghemawat 2007b: 57). However, our research adds a certain nuance to this geographically restricted space of MNC activities, that is, the extent and patterns of their expansion beyond the broad triad regions. Complementary to their expansion within the broad triad, nearly one-third of our 663 sample MNCs also expand into the regions South America and Africa – which are considered by only a few IB scholars in their studies of firmlevel regionalization phenomena (Arregle et al. 2009: 94–95; Delios and Beamish 2005: 22).2 These non-triad regions, however, constitute typical regional expansion patterns of MNCs according to Ohmae’s (1985: 121–123) tetrahedron model, for which we found supporting evidence. The respective research findings – showing that triad-based European firms expand particularly to Africa, that US or North American companies expand particularly to South America, and that Japanese or Asian MNCs expand particularly within the Asia-Pacific – imply that regionalization theory should take into account South America and Africa, to properly explain the diverse nature of regional expansion patterns of MNCs. Based upon this enlarged geographical space, we notice different regional strategies of MNCs that – following Rugman’s (2005b: 4) approach for their classification, according to the firms’ sales spread across regions – include homeregional, host-regional, bi-regional, and different multi-regional (or global) strategies. By scrutinizing Rugman’s (2005b: 4) classification scheme in a sensitivity analysis, we find that his chosen thresholds of 50% for home-regional sales and 20% for host-regional sales overstate home-regional vis-a`-vis in particular biregional and tri-regional MNC strategies. These results coincide with those of Osegowitsch and Sammartino (2008: 190) who show that this overemphasis of home-regional strategies can be avoided by abandoning the 50% cut-off point for home-regional sales and by relaxing the 20% host region thresholds to 15% or 10% of sales. In our extension of their sensitivity analyses – by employing stepwise increases of the minimum value for the home region threshold – we provide new insights into the extent to which home-regional sales of a MNC may influence the classification of its regional strategy. These results uncover a notable portion of biregional and tri-regional firm strategies, while interestingly – even at the very extremes of minimum home-regional and host-regional sales (e.g., 95% and 10%, respectively) – only very few quad-regional and global strategies of MNCs, the latter covering all five regions, could be identified. This empirical finding provides empirical evidence for the contention that truly global strategies are far from common for the world’s largest firms (Rugman 2005b: 4–5; Rugman and Hodgetts 2001: 341; Rugman and Verbeke 2004: 6–7; Rugman and Verbeke 2008c: 327),
2
For further evidence at the macro-level regarding the importance of such non-triad-regions in the processes of regionalization, provided by both trade and FDI patterns, cf. Poon et al. (2000).
7.1 Discussion of Theoretical Implications
169
which constitutes an important assumption of regionalization theory. Furthermore, the results of our sensitivity analyses suggest that more thorough analyses are necessary for comparing MNCs – e.g., from particular industries and/or from different home regions – with respect to their regional strategies. For example, supposing that the minimum value of the home region threshold in a specific industry is 85% on average, a MNC with 65% of home-regional sales has reached a much higher level of internationalization than its peers.3 This differentiation would be obscured both by applying and by abandoning the 50% cut-off point for home-regional sales. Consequently, even though this differentiation may not be relevant for the more diversified firms of our sample MNCs that span various industries (Rugman and Verbeke 2008c: 327), we call for more caution in the utilization of cut-off points for the classification of regional strategies – particularly as no threshold can actually claim to be the one and only “true” threshold (Osegowitsch and Sammartino 2008: 190). Such more cautious applications of threshold values could address criticisms related to the narrow thesis of regionalization (Clark et al. 2004: 513) – and thus could paradoxically redirect the attention from these quantitative data thresholds to the more qualitative nature of regional strategies, leading to further advances in the substance of this research field. This more qualitative information was provided by our survey respondents – where the relative importance assigned in their rating of different characteristics of the regional strategies of their MNCs produced several valuable insights.4 First, the increasing importance of regional management autonomy from home-regional, over host-regional and bi-regional, to tri-regional strategies of MNCs supports Stopford and Wells’ (1972: 53) contention that a rising portion of foreign sales – which are generated in foreign host regions – should turn into a higher perceived need for regional area structures. This shows that the delegation of decision-making autonomy to regional units appears to be considerably influenced by such changes in the geographic balance of the business (Ghemawat 2005: 104; Rugman and Verbeke 2007: 201; Rugman and Verbeke 2008c: 329; Stopford and Wells 1972: 51). The observed importance of a MNC’s geographic balance for its organizational configuration implies that, as suggested by regionalization theory, such spatial issues are strongly incorporated in the strategy of MNCs (Buckley and Ghauri 2004: 95). Second, our empirical results support the widespread conjecture that the degree of regional product/service adaptation of MNCs is closely related to the interregional distance they face (Ghemawat 2001: 147; Ghemawat 2005: 102; Ghemawat 2008: 113; Rugman 2005b: 230). The fact, however, that this relationship appears to persist along different regional strategies – and that it tends to follow an inverted U-shape – represents a new research finding. The observed shape of this 3 A comparison of MNCs from the same home region may lead to similar results – uncovering firms with regional strategies that are considerably less home-regional than those of other companies of the same home region. 4 Here, even though it overstates home-regional strategies – mainly for comparability reasons, as outlined in Sect. 5.2.1 – we utilize Rugman’s (2005b: 4) sales-based classification of regional strategies according to the 50% home region and 20% host region thresholds.
170
7 Discussion of Results and Implications
relationship may be explained by the relatively lower perceived inter-regional distance faced by MNCs with home-regional strategies, which requires to a lesser extent the regional adaptation of products/services than for those firms pursuing host-regional or bi-regional strategies (Rugman 2005b: 225). Those MNCs with a tri-regional strategy, however, may exploit multi-regional (or global) integration economies across borders – leading to competitive aggregation advantages for their products/services at the expense of regional adaptation (Bartlett and Ghoshal 1987b: 8; Bartlett and Ghoshal 1989: 5–6; Ghemawat 2003: 149; Kreikebaum et al. 2002: 148). This also explains why the perception of a slightly increased inter-regional distance by MNCs with tri-regional strategies does not translate into the regional adaptation of their products/services. Consequently, the underlying logic and benefits of “global”, or multi-regional, integration across borders appear to apply for MNCs following regional strategies – which supports the argument of Rugman and Verbeke (2008c: 330) that the academic work on regional strategies can be considered as an extension of Bartlett and Ghoshal’s (1989) work. Third, our observations shed light on the perceived difficulties of MNCs concerning their proper orientation towards, or focus on, particular regional customer demands in foreign host regions in light of an increased level of interregional distance. This provides support for the effects of a MNC’s liability of foreignness imposed by a high inter-regional distance that,5 particularly at the customer-end, requires considerable market knowledge – which, if not already existing, has to be built up by time-consuming learning processes (Barkema and Drogendijk 2007: 1143–1144; Rugman 2005b: 233; Zaheer 2002: 353). Given a high inter-regional distance, experimental learning thus appears to be critical, particularly in regards to the growing knowledge about regional markets – which lends support to the view that such insights of the internationalization model of the Scandinavian school (e.g., Barkema et al. 1996; Johanson and Vahlne 1977, 1990) should be considered in advances of the field of regionalization theory (Rugman 2005b: 65; Rugman and Verbeke 2004: 11–12). Fourth, the survey responses indicate that MNCs with distinct regional strategies coordinate their regional operations differently – where their preferences for any particular coordination mechanism in foreign host regions were found as being influenced by their administrative heritage (Bartlett and Ghoshal 1989: 49–51, 158–165). This provides support for the notion that the administrative heritage of MNCs may constitute an administrative rigidity in their penetration of host regions (Rugman 2005b: 197; Rugman and Verbeke 2003a: 128). More broadly, this implies that a firm’s administrative heritage represents an important element in the theoretical explanation of the regional strategies of MNCs, especially concerning their respective cross-border expansion to foreign regions. Fifth, the increasing importance of inter-regional knowledge transfers from homeregional, over host-regional and bi-regional, to tri-regional strategies of MNCs
5
The liability of foreignness is given here by the MNC’s unfamiliarity with the environment of a foreign host region (Zaheer 1995: 343).
7.1 Discussion of Theoretical Implications
171
suggests that the FSAs, or knowledge bundles, required in foreign host regions are quite distinct from the knowledge combinations effective in the home region (Rugman 2005b: 71). Conversely, the relatively high level of inter-regional knowledge transfer, which we found for MNCs with tri-regional strategies, points to the utilization of more complex internal knowledge diffusion systems at advanced levels of internationalization (Birkinshaw 2000: 112; Rugman and Verbeke 2001: 239–243; Rugman and Verbeke 2003a: 133; Rugman and Verbeke 2007: 202–203). The precise configuration of these firm-internal knowledge diffusion systems, however, may also depend upon the particular coordination mechanism applied by the MNC at the regional level. This is supported by the relatively high emphasis placed on intra-regional knowledge sharing by MNCs with bi-regional strategies – which prefer cooperative-regional coordination forms that involve extensive inter-unit communication and knowledge exchange within the region – i.e., related to subsidiary initiatives (Birkinshaw 1997: 225). As a result, both inter-regional and intra-regional knowledge flows appear to be dependent upon the MNC’s particular regional strategy and its respective regional coordination mechanism. Sixth, our primary data suggests that Ohmae’s (1985: 110–111, 206) concept of a MNC’s staying power to achieve a regional insider position constitutes an important element of regional strategies, largely irrespective of their particular type. Other forms of such a regional commitment of MNCs, however, including their management continuity and market-oriented investments, appear to be utilized differently along various regional strategies. Here, our observation that MNCs with host-regional strategies attach relatively low values to these two forms of regional commitment, may imply that these firms have decided to concentrate exclusively on their staying power to become a regional insider. Such a concentration on only one form of insiderization seems reasonable, considering our explanations above regarding the relatively high inter-regional distance faced by these companies and the respective high additional costs of further forms of regional commitment. The little importance attached to market-oriented investments by MNCs with triregional strategies again supports the existence of multi-regional (or global) integration advantages that – by economies of scale and scope and the increased transferability of non-location-bound FSAs – reduce the required amount of market-oriented expenditures (Rugman 2005b: 194–195). More generally, achieving an insider position by means of different forms of its regional commitment, particularly by its staying power, seems to be an integral part of a MNC’s regional strategy. Apart from these diverse conceptual insights into the regional strategies of MNCs, our empirical results contribute to the existing knowledge in the IB field about regional success. Over a considerable time period that has not been previously studied – covering nine reporting years from 2000 to 2008 – we find that the overwhelming part of MNCs, based on the relative amounts of their regional sales and assets, are classified as home-regionally successful companies,6 which provides
6
The classification of different types of regional success (e.g., home-regional, host-regional, bi-regional, etc.) – mainly for comparability reasons, as outlined before in this chapter and in
172
7 Discussion of Results and Implications
support for the finding of numerous regionalization scholars (e.g., Delios and Beamish 2005: 32; Rugman 2003b: 412; Rugman and Verbeke 2004: 7).7 With respect to the sales-based regional performance, however, our results indicate that the relative dominance of home-regionally successful firms declines, whereas the share of MNCs with bi-regional success doubles during the 9-year period of our research. While an increased percentage of bi-regionally successful firms has been observed for certain industries – e.g., for the financial services sector (Grosse 2005: 140) – this is the first study, to the best of our knowledge, which shows such a development for all the Fortune Global 500 firms during our sample period. The rising portion of MNCs with bi-regional success indicates that, a dominance of success in the home region: “[. . .] does not mean that the process of globalization is entirely non-existent at the macro level, nor that attempts at the micro level to achieve a more balanced distribution of sales should be dismissed” (Rugman and Verbeke 2008c: 331). Rather this suggests that firms, over longer periods of time, appear to be able to overcome their liability of foreignness in host region environments – both by exogenous changes in the inter-regional distance they face and by their own firmlevel learning in host region environments (Rugman 2005b: 65–66; Rugman and Verbeke 2008c: 331). While both this argumentation and our data indicate an increasing share of MNCs that are bi-regionally successful across regions,8 their relative share is still low – corresponding to approximately only one-eighth of those MNCs with home-regional success in 2008. Considering this in combination with our finding that only very few MNCs could realize a tri-regional or quad-regional success – while we could not identify any company being globally successful across all five regions – this implies that firms currently face a world of regionalization, instead of globalization, in which they are particularly home-regionally successful. Taking a closer look at home-regional success patterns, even after controlling for home region biases,9 we find that the percentage of the MNCs’ sales and assets within their home region on average is quite stable from 2000 to 2008 – which is consistent with the more recent research results, over shorter time periods, on the longitudinal development of home-regional success (Oh 2009: 341; Rugman and Oh 2007: 36–37; Rugman and Verbeke 2008a: 406; Rugman and Verbeke 2008c: 328).10 A comparison of this stable home-regional success development with our
Sect. 5.2.1 – is based on Rugman’s (2005b: 4) classification scheme according to the 50% home region and 20% host region thresholds. 7 For a review of further literature related to different home-regionally successful MNCs cf. Sect. 2.2.2. 8 Whereas the relative number of MNCs with tri-regional and even quad-regional success slightly also increased from 2000 to 2008, as outlined in Sect. 6.2.1, the share of bi-regionally successful firms rose much more strongly during this period, because of this, this form of regional success represents the main focus of our explanations here. 9 Cf. Sect. 6.2.1. 10 The methodological underpinnings of these findings on the home-regional success of MNCs are different from the classification of home-regionally successful firms, which we presented earlier (cf. Sect. 5.2.1 for a more detailed explanation of these methodological differences).
7.1 Discussion of Theoretical Implications
173
earlier finding – the decreasing number of firms being classified as home-regionally successful – indicates that MNCs increasingly generate their sales in another, mainly second, foreign host region over our 9-year sample period.11 This suggests that MNCs follow a sequential pattern in their regional expansion (Rugman 2005b: 76; Van Tulder et al. 2001: 67), to be regionally successful beyond their home region. As mentioned earlier, this however has been achieved by only relatively few bi-regionally successful firms. Most MNCs of our research sample are still mainly successful in their home region, while our results at the level of particular industries indicate that service MNCs tend to be even more home-regionally successful than manufacturing firms – which reflects the findings of other academic scholars (Rugman and Oh 2007: 39; Rugman and Sukpanich 2006a: 378; Rugman and Sukpanich 2006b: 147; Rugman and Verbeke 2008a: 406). In addition, we find that the least home-regionally successful, or most internationalized, industries are chemicals and pharmaceuticals – corresponding to the results of Rugman and Oh (2007: 39) – as well as computer, office and electronics, and food, drug and tobacco. This implies that the interpretation of the home-regional success of particular MNCs should always be sensitive to their industry.12 More broadly however, based upon the percentage of their home-regional sales and assets, only particular industries and relatively few MNCs achieve regional success beyond their home region. By means of additional home-regional figures of our sample firms, we found that the overall dominance of their home-regional success appears to be due to their relatively high home-regional resource commitments in the form of investments and employees. This supports Rugman’s (2005b: 229) contention that a MNC’s resource commitments are restricted almost exclusively to the home region, due to the high risks associated with attracting potential foreign customers and increasing sales in foreign regions.13 The preference of a home-regional resource concentration thus appears to be the outcome of careful cost-benefit calculations of our sample MNCs (Rugman 2005b: 76). This is substantiated by our findings about their resource employment – given by an increasing home-regional employee productivity over the most part of our sample period, where more recent declines in 2008 can be explained in light of the worldwide financial crisis at this time (Palazuelos and Ferna´ndez 2009: 9; Steil 2009: 9). While these observations provide reasons for the prevalence of home-regional success, our empirical findings on the home-regional profitability and growth rates do not unambiguously support this finding – even showing contradictory results. This indicates that MNCs may realize higher profits and growth rates beyond their home region – reflecting the benefits of a broader geographic scope of their activities (Rugman 2005b: 194; Tallmann and Yip 2009: 313). Consequently, while the MNCs’ commitment and 11
More specifically, this shows that the non-home-regional sales in one foreign host region increasingly exceed the 20% host region threshold for a bi-regional success classification. 12 This also shows that any classification scheme for different forms of particular MNCs’ regional success – or for their regional strategies, as outlined earlier in this chapter – should be utilized with caution, while more detailed analyses should be based on industry-by-industry comparisons. 13 Cf. Sect. 2.2.8 for a more detailed description of these risks.
174
7 Discussion of Results and Implications
employment of resources serve to explain the dominance of their home-regional success – as measured by the respective percentage of regional sales and assets – these firms also appear to derive a considerable part of their profits and growth from outside of their home region. More generally, this implies that depending upon the particular geographic scope of a MNC’s activities, different forms of regional success may have to be considered for evaluating its performance in a particular region. A closer examination of the regional strategy–performance relationship – thus the dependence of the MNCs’ performance on their geographic scope and configuration of activities – led to empirical evidence of an S-curve development of this relationship, as supposed by recent research on the multinationality-performance relationship (Contractor 2007b: 23; Contractor et al. 2003: 15–16; Lu and Beamish 2001: 582; Lu and Beamish 2004: 606; Thomas and Eden 2004: 106–107).14 This suggests that the varying regional geographic spread, or regional orientation, expressed in the regional strategies of MNCs reflects changes in their multinationality – leading to different performance implications. The sigmoid, or S-shaped, development of this relationship was found to be robust along various classifications of regional strategies,15 ranging from the MNCs’ sales, assets, investments, and employees – which correspond to the different elements of the depth and breadth of their multinationality (Thomas and Eden 2004: 92). Along this sigmoid development of the regional strategy–performance relationship, a home-regional strategy appears to represent the initial stage of possible cross-border expansion paths for MNCs – where they may realize a notable performance because of the low liability of foreignness they face (Contractor 2007b: 14; Lu and Beamish 2001: 582; Rugman 2005b: 229; Sukpanich 2007: 333–334). MNCs with host-regional strategies reach a comparably more advanced stage of regional expansion – where the orientation towards, or focus on, a foreign host region increases their multinationality. This seems to translate in a higher liability of foreignness, associated with a decline in firm performance – due to high additional costs of learning about the foreign cultures and markets, and the costs of achieving legitimacy and acceptance in the different institutional environments of this host region (Contractor 2007b: 14; Contractor et al. 2003: 16; Goerzen and Asmussen 2007: 79; Goerzen and Beamish 2003: 1302; Kostova and Zaheer 1999: 77; Rugman and Verbeke 2005: 14; Zaheer and Mosakowski 1997: 457). In the case of both a home-regional and a host-regional strategy, the regionally limited reach of
14
It should be noted that several academic scholars (Contractor 2007a: 471; Contractor 2007b: 19–21; Contractor et al. 2003: 16; Lu and Beamish 2004: 606) have shown that these recent research results on the S-curve development of the multinationality-performance relationship present a reconciliation of prior research findings about (upright) U-shaped patterns (e.g., Lu and Beamish 2001: 580; Ruigrok and Wagner 2003: 77) and about inverted-U-shaped developments (e.g., Geringer et al. 1989: 117; Hitt et al. 1997: 790) of this relationship. 15 Mainly for comparability reasons – as outlined earlier in this chapter and in Sect. 5.2.1 – the classification of different types of regional strategies is based on Rugman’s (2005b: 4) classification scheme according to the 50% home region and 20% host region thresholds.
7.1 Discussion of Theoretical Implications
175
the FSAs of MNCs constrains the development of their global scope, and thus the reaping of the related benefits (Contractor 2007b: 15; Rugman 2005b: 38, 226). MNCs with a bi-regional strategy, however, appear to exploit these advantages – by means of their ability to transfer non-location-bound FSAs and to complement these with location-bound CSAs in host regions – which imply considerable performance increases (Lu and Beamish 2001: 582; Rugman and Verbeke 1992: 766; Rugman and Verbeke 2003a: 128). The benefits of still further regional expansion by MNCs following a tri-regional strategy, though, may be less than the incremental costs of managing many operations and of offering competitive products/services in dissimilar regional markets – leading to a drop in firm performance (Bowen 2007: 120; Contractor 2007b: 16–17; Contractor et al. 2003: 16; Elango 2004: 439; Lu and Beamish 2001: 582; Lu and Beamish 2004: 601). Another plausible reason for this development is provided by Ruigrok et al. (2007: 363), who found that high degrees of internationalization limit the absorptive capacity of MNCs for both organizational and environmental complexities. These observations support our earlier findings of a dominant portion of MNCs with home-regional strategies – which decreases mainly in favor of an increasing share of companies pursuing bi-regional strategies – and of a negligible share of firms with any form of multi-regional (or global) strategies. More broadly, the existing insights about the multinationality-performance relationship appear to be applicable to the study of the regional strategy–performance relationship – where the lower level of aggregation of the latter analysis may contribute to a better understanding of the former research field (Hennart 2007: 447; Verbeke et al. 2009: 158–159). While the previous discussion of our observations on regions, regional strategies, regional success, and the regional strategy–performance relationship produced several new conceptual and theoretical insights about these research fields, such an analysis does not provide proof of a causal argument. However, we believe that these knowledge advances enhance our understanding about the variables, which we relate to each other to investigate causal relationships in the field of regionalization theory. Consequently, we drew upon these insights about the variables entering in our study of success factors of regional strategies for MNCs, to improve the rigor in the modeling of these causal relationships and the interpretation of its theoretical implications – which we will present in the following. In our investigation of the success factors of MNC’s regional strategies, we proposed that their regional management autonomy and regional product/service adaptation would have a positive influence on their regional success – taking into account the respective internal regional orientation of each firm and the external inter-regional distance it faces. Contrary to our research hypothesis H1,16 results suggest that high degrees of regional management autonomy granted by a MNC have a weakly significant negative impact on its regional success. Upon reflection, one can think of several reasons as to why this may be the case.
16
Cf. Sect. 4.1.1.5.
176
7 Discussion of Results and Implications
First, these adverse performance effects may be explained by the fact that a high regional management autonomy can lead to inefficiencies of coordination and control (Bartlett and Ghoshal 1989: 14) – where a regionally decentralized decision-making exceeds the degree that is necessary for the efficient management of regional products/services (Stopford and Wells 1972: 70). A possible reasoning behind this may be that, as we have outlined earlier in this chapter, these results are based to a large part on MNCs with home-regional strategies – where extensive regional management activities may lead to inefficiencies due to needless duplications of the coordination and control functions of corporate headquarters. More specifically, given that these MNCs derive more than 50% of sales from within their home region, usually most cash flows, assets, and employees should be concentrated in this region (Rugman and Verbeke 2008c: 329) – which are administered more efficiently and directly by corporate headquarters, instead of regional management. In addition, considering that their sales are mainly concentrated in the home region, nearly all important product/service decisions of these firms relate to customer preferences within that region – where the establishment of large regional management structures may add more bureaucratic nuisance than any value (Ohmae 1985: 188). In other words, the growth of benefits from additional, regional managerial efforts to address customer preferences at the downstream end of these MNCs’ value chains does not compensate for the growth of its associated costs (Proff 2000: 546). Second, the negative effects of regional management autonomy on performance may be explained also by the fact that – besides being largely home-regional – most firms in our study are European MNCs.17 As we have illustrated earlier in this chapter, due to their administrative heritage, these firms tend to coordinate their regional units by cooperative-regional approaches based on socialization mechanisms – which are associated with high costs for their implementation (Rugman 2005b: 201; Rugman and Verbeke 2008b: 306, 308; Verbeke et al. 2009: 157). The respective negative effects on performance may be compounded by the fact that such cooperativeregional coordination forms largely build upon subsidiary initiatives (Birkinshaw 2000: 8; Rugman 2005b: 205; Rugman and Verbeke 2001: 238; Rugman and Verbeke 2008b: 310),18 where a delegation of excessive decision-making autonomy to the regional level may hamper the exploitation of these subsidiary strengths (Morrison and Roth 1992: 45). Third, apart from the specific regional coordination mechanism applied, a decision-making by corporate headquarters gives subsidiaries the impression that the MNC is interested in their affairs (Vora et al. 2007: 613) – which in the case of considerable intermediate layers of coordination and control by a regional management may have adverse effects, with negative performance implications. These explanations show how a MNC’s regional success can be negatively affected by delegating decision-making autonomy to the regional level – if this does not
17
Cf. Sect. 6.1. Cf. Sect. 2.2.6.
18
7.1 Discussion of Theoretical Implications
177
account for the firm’s idiosyncratic market position in the home region, and the costs resulting from its applied coordination mechanisms and/or its administrative heritage (Rugman 2005b: 232). More broadly, these manifold influences in the regional organizational effectiveness of MNCs imply that the theoretical rationale behind regional management centers should not, or at least not exclusively, be related to their involvement in corporate decision-making processes. Furthermore, our empirical results suggest that, consistent with our hypothesis H2,19 a firm’s regional adaptation of its products/services has a highly significant positive impact on its regional success. This supports the contention of regionalization scholars that regions and their specific demands from customers, governments, and other stakeholders of MNCs have to be considered in their products/services – to realize success in regional markets (Rugman 2005b: 53; Rugman and Verbeke 2004: 5; Rugman and Verbeke 2008b: 307). In addition, the positive performance effects of increasing levels of regional product/service adaptation provide evidence for Ohmae’s (1985: 13, 165) view that MNCs have to achieve a deep penetration into each regional market in their regional expansion. Moreover, this finding shows that the perspective popularized by Bartlett and Ghoshal (1989: 59–61) – regarding the complementary, rather than opposing, logic of national responsiveness and global integration for the product/service strategies of MNCs – has to be extended by the region, in order to properly analyze a firm’s regional competitiveness (Rugman and Verbeke 2008b: 307; Rugman and Verbeke 2008c: 330). Concerning the elements that constitute a regional market strategy of MNCs, we found that the regional design, the regional functional scope, the regional offering portfolio, and the regional brands of their products/services highly influence the degree of their regional adaptation. This lends support to the proposition that MNCs – for a successful penetration of geographical regions – do develop region-bound company strengths, or FSAs, to adapt their market strategy approach to regional requirements and conditions (Rugman 2005b: 50; Rugman and Verbeke 2008b: 307). The ability of MNCs to develop region-specific capabilities can thus highly contribute to the explanation of their regional market success with their product/service offering, both within and beyond their home region. The MNCs’ building up of these regional competencies, therefore, offers an answer to the “why” question of their successful regional, instead of national or global, expansion patterns (Delios and Beamish 2005: 33; Rugman 2005b: 240). In particular the development of regional FSAs in the areas of design, functional scope, offering portfolio, and brands appears to determine both the limitations and opportunities of regional strategies along the dimensions of adaptation, aggregation, and arbitrage. More generally, this implies that MNCs achieve an optimal mix – or moderate levels of – global integration and national responsiveness by means of different combinations of these four regionbound FSAs in the regional adaptation of their products/services. The processes of a MNC’s value creation by its regional product/service adaptation on the basis of
19
Cf. Sect. 4.1.2.5.
178
7 Discussion of Results and Implications
regional FSAs should thus receive utmost attention in regionalization theory and its scholarly reflection on regional firm strategies. With respect to the organizational and environmental context faced by regionally operating MNCs, we found that the internal context in the form of their regional orientation has a positive direct influence on their regional success – while the external context given by their inter-regional distance exerts a direct negative effect on their regional performance. This may be explained by the fact that a MNC’s high orientation, or focus, on only one particular region and its low inter-regional distance lead to a low organizational and environmental complexity – which contributes to the survival of MNCs (Drazin and Van de Ven 1985: 516; Verbeke et al. 2009: 154–156). Conversely, this means that a firm’s exposure to higher internal and external complexities has to pay off – which serves to explain for instance that inward FDI flows into a particular region occur in high- rather than low-paying industries (Nachum and Zaheer 2005: 763). However, these direct influences of the organizational and environmental context on regional success are not able to address the more fundamental issue of regionalization theory – relating to those regional performance effects, which arise from the interaction of the internal and external contingencies with the regional strategy and structure of MNCs (Drazin and Van de Ven 1985: 514; Rugman 2005b: 235; Verbeke et al. 2009: 158). These interaction effects were hypothesized to emanate from a MNC’s regional orientation (H3a and H3b) and inter-regional distance (H4a and H4b) on the relationships of regional management autonomy and regional product/service adaptation with regional success. Concerning the interaction effect of the regional orientation of MNCs, as postulated in our research hypotheses H3a and H3b,20 our results provide no support for its influence hypothesized in H3a on the relationship between their regional management autonomy and regional success – where this internal contextual variable appears to have almost no impact at all. However, we found a highly significant positive interaction effect for the hypothesized impact in H3b of a MNC’s regional orientation on the relationship between its regional product/service adaptation and its regional performance. This shows that the regional orientation of a firm seems to be relatively unimportant for the impact that its regional management autonomy has on regional success – while the performance effects from its regional product/ service adaptation are highly positively affected by this internal contingency.21 The latter finding can be explained by the fact that a narrow geographic market focus of MNCs, given by their strong regional orientation, supports the development of those FSAs in the regional adaptation of their products/services – which are capable of highly satisfying customer needs within the region, leading to positive effects on regional success (Rugman 2005b: 212; Rugman and Verbeke 2008b: 312).
20
Cf. Sect. 4.1.3.2. Here, given the marginal, almost non-existent, interaction effect of regional orientation on the relationship between regional management autonomy and regional success – no further theoretical implications can be drawn from the rejection of our research hypothesis H3a.
21
7.1 Discussion of Theoretical Implications
179
In other words, MNCs with a geographically limited regional focus appear to be able to utilize their region-bound capabilities more effectively – as regards the successful regional adaptation of their products/services. This effective employment of their region-bound FSAs also reduces the MNCs’ overall amount of one-sided, locationspecific investments at the customer-end – which are required for the regional adaptation of their products/services (Rugman 2005b: 231) – constituting a further positive effect on regional performance. Thus, in contrast to Bartlett and Ghoshal’s (1989: 195, 212) claim for building a matrix in the minds of MNC managers – meaning the view of problems and opportunities from both local and global perspectives simultaneously – firms following a regional strategy approach appear to best handle such trade-offs at the regional level by a geographically limited mindset. This implies that a MNC’s successful regional adaptation of its products/services can be explained by a restricted regiocentric orientation of its managers – where the economics of concentration outweigh the economics of dispersion in the deployment of regional FSAs (Ghemawat 2005: 102). More broadly, a reduced regional orientation of MNCs appears to lower their ability, first, to mentally absorb such increases in their multinationality, and second – as a consequence – to effectively develop those FSAs that are required for a successful regional product/service adaptation. This interaction effect of a firm’s regional orientation thus advances the existing knowledge about the constraints of a successful regional product/ service adaptation, which enhances the theoretical understanding of the regional market strategies of MNCs. In our investigation of the interaction effect of a firm’s inter-regional distance, as suggested in our research hypotheses H4a and H4b,22 we found contradictory results to our hypothesis H4a – given by a weakly significant negative influence of this external contingency on the relationship between regional management autonomy and regional success. However, our results provide strong support for hypothesis H4b – in the form of a highly significant positive interaction effect of inter-regional distance on a MNC’s successful regional product/service adaptation. The rejection of hypothesis H4a means that a high inter-regional distance negatively affects the negative relationship between regional management autonomy and regional success. In other words, the negative impact that a highly autonomous regional management has on the regional success of MNCs is reduced in the case of a high inter-regional distance. Conversely, a low inter-regional distance only to a marginal extent, negatively affects the negative relationship between regional management autonomy and regional success. Thus, in this case, the negative performance effects that result from a highly autonomous management, at least to a large part, persist – as they are only reduced marginally by low values of this external context of the MNC. These relationships related to the rejection of hypothesis H4a can be explained by several reasons. First, at low levels of inter-regional distance, MNCs face only a minor liability of foreignness – where their respective familiarity with regional demands and requirements
22
Cf. Sect. 4.1.3.4.
180
7 Discussion of Results and Implications
does not require a high allocation of decisions to the regional level (Yeung et al. 2001: 169; Zaheer and Mosakowski 1997: 445). Therefore, as no discernible benefits are to be expected from a high level of regional management autonomy, its negative influence on a firm’s regional performance largely remains. In light of increasing cultural, administrative/ political/institutional, geographic, and economic differences between regions, however, MNCs experience higher difficulties in comprehending specific demands and requirements of a particular region (Zaheer 1995: 341–343, 357; Zaheer and Mosakowski 1997: 445). A firm’s costs associated with this higher liability of foreignness may be, to a certain extent, offset by benefits of a highly autonomous regional management. For example, such benefits may include a better access to regional information, or to institutional and regulatory domains in the respective regional environment (Zaheer 2002: 352) – which reduce the negative effects of regional management autonomy on regional success. Second, if a MNC is exposed to a high inter-regional distance, it requires a large cadre of experienced, cosmopolitan managers, to cope with the uncertainties resulting from high inter-regional differences in their decision-making (Ghemawat 2001: 147; Kostova and Zaheer 1999: 71; Rugman and Verbeke 2007: 202). In other words, such uncertainties lead to a strong reliance on cues and on the participating efforts from different actors within the organization in these decisions (Delios and Henisz 2003: 239; Henisz and Delios 2001: 466). Here, a delegation of decision-making to regional management may contribute to reducing these uncertainties related to the economic, socio-cultural and business environment – which facilitates the successful exploitation of regional FSAs (Collinson and Rugman 2008: 225–228). This effect alleviates the negative impact that regional management autonomy has on regional success. Conversely, in light of a low inter-regional distance, such uncertainties – related to regional cultural, administrative/political/ institutional, geographic, and economic differences – only exist to a minor part. In this case, the costs and investments associated with a delegation of large sets of decisions to the regional level (Rugman 2005b: 233) – due to the low levels of uncertainty, largely represent unnecessary overhead expenditures, which ultimately result in sunk costs. Therefore, given a low inter-regional distance, the negative effect of a highly autonomous regional management on a MNC’s regional success persists, at least for the most part. Third, at increased levels of inter-regional distance faced by a MNC, higher decision-making responsibilities at the regional level can contribute to a better coordination of intra-regional operations (Enright 2005b: 100; Ricart et al. 2004: 192) – such as, for example, national subsidiaries or regional production and distribution hubs (Buckley and Ghauri 2004: 87; Ghemawat 2005: 103–104; Ghemawat 2008: 148–149; Yeung et al. 2001: 169). The coordinative support of autonomous regional managers in reaping such regional agglomeration benefits contributes to reducing the negative effect of a high regional management autonomy on performance (Arregle et al. 2009: 103; Rugman and Verbeke 2001: 247). Conversely, given a low inter-regional distance, the relative “closeness” and similarity of regions reduces the bounded rationality of a firm’s corporate headquarters in the coordination of intra-regional organizations (Arregle et al. 2009:
7.1 Discussion of Theoretical Implications
181
90; Rugman and Verbeke 2005: 15). Therefore, in this case, no substantial benefits result from dispersed coordination competencies and capabilities of autonomous regional managers – which implicates that the negative influence of regional management autonomy on regional success remains largely unchanged. Overall, these explanations show that – in light of increases in a MNC’s interregional distance – a decision-making by regional management involves benefits that mitigate its negative effects on regional performance. At the same time however, a low inter-regional distance of MNCs may imply that a decision-making by corporate headquarters and/or national subsidiaries creates more regional value than a highly autonomous regional management (Ghemawat 2005: 104). This external contingency thus contributes to the theoretical explanation of limitations and opportunities in the delegation of decision-making autonomy to the regional level.23 The strong confirmation of the research hypothesis H4b indicates that the regional adaptation of products/services becomes increasingly important for regional success in the case of rising levels of inter-regional distance. This is consistent with existing accounts of regionalization theory that MNCs – which face high inter-regional differences in the various markets they serve – need to establish a regional insider position by their regionally adapted products/services to be successful (Ohmae 1985: 89, 92; Rugman 2005b: 66; Rugman and Verbeke 2004: 4). Thus in a semi-globalized world, different forms of inter-regional distance – along cultural, administrative/ political/institutional, geographic, and economic dimensions – constitute strong environmental pressures on MNCs to adjust to these differences by means of their regional product/service adaptations (Ghemawat 2001: 149; Ghemawat 2008: 110–111). MNCs realize these product/service adaptations, as we have shown before, mainly by their regional FSAs – meaning that region-specific capabilities, particularly those at the customer-end which enable a deep penetration of markets (Millar et al. 2005: 127; Rugman 2005b: 198), need to be employed for successfully addressing inter-regional differences. Here, the costs for developing such regional FSAs – as well as the associated location-bound investments for melding them with CSAs at the regional level – appear to pay off in light of high inter-regional differences, given the associated positive regional performance effects. While the MNCs in our study are mainly successful in their home region – these results are to a large part based on their most successful foreign host region.24 Recognizing further that most of our sample firms have a substantial regional
23
However, it should be noted that this influence of the external contingency inter-regional distance on the relationship between regional management autonomy and regional success – as described in Sect. 6.3 and depicted in Table A.12 – was not found to be significant after considering the control variables firm size and regional competition. Therefore, the degree to which inter-regional distance influences a delegation of decision-making to the regional level – depending on a MNC’s firm size and/or its regional competition – should be interpreted with caution (cf. Sect. 7.2.1). 24 Approximately two-thirds of the MNCs in our study have completed the answers to our survey for their most successful foreign region (cf. Sect. 6.1). Even if taking into consideration the
182
7 Discussion of Results and Implications
experience in regional markets,25 this implies that – even in highly distant hostregional markets – the costs associated with a MNC’s liability of foreignness, can be overcome in the long run by its insider position resulting from regional product/ service adaptations (Zaheer 2002: 353). Consequently, the MNCs’ commitment to distant foreign host regions – given by the development of region-specific FSAs for the regional adaptation of their products/services – constitutes an important element in their successful regional expansion across borders. Mainly due to the dominance of home-regional success of MNCs – which, as we have demonstrated before, also applies to the firms of our study – little attention has been devoted by regionalization scholars to their successful expansion into distant foreign regions. This work, however, indicates that this is possible, even for mainly home-regionally successful MNCs, by responding to inter-regional differences with regionally adapted products/services on the basis of regional FSAs. Therefore, and more generally, the interaction effect of inter-regional distance on the successful regional product/ service adaptation of MNCs lays the ground for a broad theoretical basis for the explanation of firms’ cross-border expansion into distant regions – irrespective of the main geographical source of their regional success. Furthermore, we found that – after a consideration of the interaction effects of a MNC’s regional orientation and its inter-regional distance – the hypothesis H1 is again rejected, given a very highly significant negative relationship between a MNC’s regional management autonomy and its regional success. At the same time, however, we could not find support for the hypothesized positive relationship in H2 between a MNC’s regional product/service adaptation and its regional success – while this has been confirmed earlier, in the absence of interaction effects from organizational and environmental contingencies. A MNC’s successful product/service adaptation thus appears to be highly dependent upon both its internal and external context – as shown by the strong support for the hypotheses H3b and H4b. This indicates that the regional market strategies of MNCs – given by the degree of their regional product/service adaptations – should correspond to the requirements of the organizational and environmental context they face. Therefore, the MNCs’ development of those FSAs that are necessary for their successful regional product/service adaptation is highly contingent upon these contextual influences. In other words, a MNC’s successful regional product/service adaptation seems to highly depend upon achieving a proper regional environment-strategystructure fit. This appears to be only partly true for a MNC’s regional management autonomy, whose impact on regional success was only found to be weakly negatively affected by the external contingency inter-regional distance (H4a) – while the
reductions in sample size for the modeling of our data – as explained in Sect. 5.2.3 – half of the modeling results are still based upon the MNCs’ respective most successful foreign region. 25 Nearly half of the MNCs in our study have been operating more than 30 years in either their most successful foreign region or in their home region (cf. Sect. 6.1). It should be noted that this also applies to the reduced sample in the modeling of our data – as outlined in Sect. 5.2.3 – where half of the firms in our modeling sample have such a high regional experience, almost equally, in either their most successful foreign region or in their home region.
7.2 Discussion of Practical Implications
183
firm’s regional orientation exerted no notable influence on this relationship (H3a). More specifically, the influence of a MNC’s regional management autonomy on its regional performance seems to be exclusively dependent upon its inter-regional distance – indicating a weaker form of the regional environment-strategy-structure fit. More generally, however, and consistent with the moderate voluntaristic approach of this work,26 this implies that there is no universal “best way” for implementing regional strategies, nor that their realization is solely situationspecific (Hambrick and Lei 1985: 765; Zheng Zhou et al. 2007: 305). This shows that the contingency approach is well-suited for the study of regionalization phenomena, as it offers both theoretical and practical insights into how the organizational and environmental context influence the success of regional strategies. By means of the explanations before, we have presented the theoretical implications of our research findings. In the following, we will discuss the practical relevance of these results – in the form of implications from the identified success factors of regional strategies, regional management autonomy and regional product/ service adaptation, for the corporate practice of MNCs.
7.2
Discussion of Practical Implications
By applying the requirements of “good” success factor research in this work – as outlined in Sect. 2.4.1 – our research results not only aim to advance regionalization theory by academic rigor, but also try to achieve practical relevance by enhancing the dialogue about the existing application of regional strategies by MNCs (Kieser and Nicolai 2005: 278). Such an orientation towards corporate practice is important, as the field of management studies – to which this work belongs to – acquires a part of its legitimacy by ensuring that it is an applied social science (Kieser and Leiner 2009: 523; Whitley 1984: 371). Thus, in the following, we will elaborate the practical implications of our findings. In a world of regionalization, as suggested by our research results, MNCs are successful mainly in their home region, and to a certain extent in foreign host regions. This is the consequence of their home-regional, host-regional, bi-regional, and multi-regional strategies. By means of the study of this work, we could observe the application of these regional strategies by our sample MNCs in either their home region or in their most successful foreign region. The explorative results of these observations, about how MNCs apply different elements of their regional strategies in practice, have been presented in Sect. 6.2.2 – for which Table A.13 provides a comprehensive overview.27 On the basis of our investigation of causal 26
Cf. Sects. 2.3.1 and 2.4.2. This overview in Table A.13 summarizes Figs. A.26, A.31, A.32 and A.33 regarding the relative importance assigned by our survey sample MNCs, which – according to the five-point Likert scales of our survey – ranges from very low, low, moderate, high, to very high, for each observed element of their regional strategies. Those terms that are presented in brackets in Table A.13 show
27
184
7 Discussion of Results and Implications
relationships between different elements of regional strategies, we provided evidence on the factors that influence a MNC’s success at home or in a foreign host region. These success factors of regional MNC strategies – the degree to which their regional management is granted decision-making autonomy, and the extent of adaptation of their products/services to regional markets – were found to interact with their regional orientation and/or the inter-regional distance they face. These causal relationships will provide the basis for deriving respective practical implications in the following – regarding the appropriate degree of regional management autonomy and regional product/service adaptation to achieve regional success. In the course of this discussion of practical implications, we will reflect the existing practices of MNCs according to our explorative findings in light of our hypothesesbased research results.
7.2.1
Regional Management Autonomy as a Success Factor
The appropriate degree of regional management autonomy was found to influence regional success. In more detail, our research results suggest that MNCs should not delegate substantial decision-making autonomy to the regional level, as this negatively affects their regional performance.28 The practical implication of this finding is that a firm – which utilizes regional management centers for the implementation of its regional strategy – should, if at all, only allocate little decision-making autonomy to its regional managers. More specifically, regional decision-making autonomy should be granted with caution, with a high sensitivity to its associated costs and benefits. Furthermore, the degree to which a highly autonomous regional management improves regional success was found to be largely unaffected by a MNC’s orientation towards, or focus on, only one particular region or several regions.29 In other words, the extent to which MNCs have a limited, or enlarged, focus in their geographic expansion – does not influence the regional performance that results from an increasing autonomy of its regional management. This implies for corporate practice, that their regional orientation does not have to be considered in their decision about how much management autonomy they should grant to their regional managers to realize regional success.
the second, third, or fourth preference of our respondents for a certain element of their regional strategy – according to the illustrations in Figs. A.26, A.31, A.32 and A.33. 28 This finding corresponds to the weakly significant rejection of our research hypothesis H1 (cf. Sect. 6.3). 29 This is due to the fact that we could not find support for any notable influence of this internal contingency variable on the relationship between regional management autonomy and regional success – contrary to its hypothesized, respective positive effect in our research hypothesis H3a (cf. Sect. 6.3).
7.2 Discussion of Practical Implications
185
In addition, inter-regional distance was found to weakly negatively impact the negative effect that an autonomous regional management has on regional success.30 This implies for practitioners, first, that they should be aware of cultural, administrative/political/ institutional, geographic, and economic differences between their firm’s home region and foreign regions in the delegation of decision-making autonomy to the regional level. Second, if their firm increasingly faces such interregional differences – for example, as a result of its expansion into new regions, or due to changes in its regional institutional or economic environment – some benefits arise from delegating decision-making responsibilities to the regional level. While these benefits may alleviate the negative influence of regional management autonomy on their firm’s regional performance, practitioners should assess the precise extent of this compensating effect – to decide if an autonomous regional management should be installed. Third, at low levels of inter-regional distance, they should try to avoid decision-making at the regional level, and think about how regional concerns can be effectively handled at other organizational levels – such as the corporate headquarters and/or national subsidiaries. Concerning the extent to which these implications from our hypotheses-based research findings are reflected in corporate practice – as illustrated for our survey sample firms in Table A.13 – we find different results for MNCs following homeregional, host-regional, bi-regional, and tri-regional strategies. These four regional strategies, as indicated by our explorative analyses, are homogenous subsets regarding the main elements of a regional strategy – and thus serve to explain differences in the application of regional management autonomy. Taking into consideration our finding of the S-curve development along these regional strategies and their associated corporate performance,31 the following points can be made.32 First, our sample firms pursuing home-regional strategies – in comparison with MNCs following other regional strategies – have delegated a moderate degree of decision-making autonomy to the regional managers within their home region, or respectively, their most successful foreign region.33 This application of regional management autonomy does not fully correspond to the practical implications that we have outlined previously for this success factor of regional strategies. Furthermore, their perception of the inter-regional distance they face is moderate – indicating that an autonomous regional management’s benefits are not expected to offset its negative effects on regional success. This implies that these companies 30
This result relates to the weakly significant rejection of our research hypothesis H4a (cf. Sect. 6.3). 31 Cf. Sect. 6.2.2. 32 Due to the fact, as outlined earlier, that a MNC’s regional orientation does not have to be considered in its decision about the degree of regional management autonomy, we will not consider this internal contextual variable in this reflection of corporate practices. 33 In our study, several MNCs with home-regional strategies – despite being mainly active in their home region – have completed the answers of our survey for their most successful foreign region (cf. Sect. 6.1).
186
7 Discussion of Results and Implications
should reduce the degree of regional management autonomy, to reduce its associated negative influence on regional success. However, due to the fact that these home-regional firms currently have delegated moderate degrees of decision-making autonomy to their regional managers, such reductions do not have to be substantial. These reductions, rather, should be analyzed step-by-step and closely monitored, particularly regarding the abandoned benefits of regional decision-making in light of the moderate inter-regional distance they face. This could include, for example, cost-benefit calculations for the decisions that are to be reduced at the regional level – as well as a regular assessment of changes in their inter-regional distance and the respective consequences for the newly allocated decisions. By means of this reallocation of decisions within the firm – if applied with the described caution – the notable corporate performance, which is usually realized by companies pursuing home-regional strategies,34 can be protected against negative effects and possible risks from regional performance declines. Second, those MNCs of our sample with host-regional strategies – vis-a`-vis firms with other regional strategies – grant high levels of management autonomy to their regional managers. This contrasts with the practical implications that we have presented previously. However, if we consider their relatively high subjective rating for the inter-regional distance that they are exposed to, this means that notable benefits of an autonomous regional management may exist – which considerably contribute to alleviate its negative influences on regional success. However, any adverse effects of substantial regional decision-making responsibilities on these companies’ regional success should be avoided, given that MNCs with hostregional strategies usually realize low levels of corporate performance.35 This implies that these host-regional firms should try to further accentuate the benefits of an autonomous regional management, with respect to the inter-regional differences they face in the cultural, administrative/political/institutional, geographic, and economic sphere. They could achieve this by thoroughly analyzing how their autonomous regional management, for example, could further reduce the costs associated with their liability of foreignness.36 A reduction of costs related to these host-regional companies’ liability of foreignness, however, may involve other costs – for example related to improving the firm’s legitimacy in the institutional environment (Kostova and Zaheer 1999: 73; Zaheer and Mosakowski 1997: 461). Therefore, these MNCs should carefully evaluate respective opportunity costs of alternative scenarios. Third, our sample companies following bi-regional strategies allocate relatively high levels of regional management autonomy to the regional level. This is contradictory to the practical implications that we have outlined earlier. In addition – compared to firms with other regional strategies – they perceive to face a moderate inter-regional distance. Therefore, the negative impact of these MNCs’ highly
34
Cf. Sect. 6.2.2. Cf. Sect. 6.2.2. 36 Cf. Sect. 7.1. 35
7.2 Discussion of Practical Implications
187
autonomous regional management on their regional success is only moderately reduced – which means that these negative effects, at least to a certain extent, remain – in light of this external context. This implies that these bi-regional companies should reduce the decision-making autonomy that they have delegated to their regional management – to lower its adverse effects on their regional success. As MNCs following bi-regional strategies are usually highly successful at corporate level,37 reductions in regional management autonomy should not jeopardize total firm performance – and thus should be evaluated regarding its potential benefits on a region-by-region basis, and with respect to their overall impact on corporate success. Given the preference of these bi-regional firms for cooperative-regional coordination forms at the regional level,38 they could delegate decision-making power from their regional managers to the national subsidiaries in their regional markets – thereby further upgrading the role of their subsidiaries’ initiatives within the firm (Birkinshaw 1997: 207; Birkinshaw 1998: 356; Birkinshaw 2000: 20). Conversely, other bi-regional firms who – maybe due to their administrative heritage – prefer different mechanisms for the coordination of their regional operations, could reduce their regional management’s autonomy by more centralized decision-making at the corporate headquarters. Fourth, those firms of our sample pursuing tri-regional strategies – versus companies with other regional strategies – assign a high autonomy to their regional managers. This application of the regional success factor regional management autonomy is inconsistent with the practical implications presented previously. Furthermore, these tri-regional firms indicated they face relatively high levels of inter-regional distance. Considering this external context, several benefits may be associated with a highly autonomous regional management of these companies – which considerably reduce its negative influence on their regional performance. This implies that – corresponding to our earlier implication for firms pursuing hostregional strategies – they should try to further increase the benefits from delegating decision-making responsibilities to regional managers. However, compared with firms pursuing host-regional strategies, MNCs with tri-regional strategies have a much broader geographic scope. Therefore, potential further benefits of an autonomous regional management are to be expected, for example, from an improved intra-regional coordination of their highly dispersed operations within their three main regions.39 Taking into account the fact that tri-regional companies usually achieve low levels of total corporate success,40 the incremental benefits from such a better exploitation of their regional managers’ decision-making autonomy – could act as a counterbalance to the already high costs of their complex firm-internal knowledge diffusion systems.41
37
Cf. Sect. 6.2.2. Cf. Table A.13 and Sect. 6.2.2. 39 Cf. Sect. 7.1. 40 Cf. Sect. 6.2.2. 41 Cf. Table A.13, and Sects. 6.2.2 and 7.1. 38
188
7 Discussion of Results and Implications
More broadly, these explanations show that many of our sample firms can still improve their regional performance by more appropriate degrees of regional management autonomy,42 given by low levels of decision-making autonomy delegated to regional managers. A proper application of this success factor of regional strategies appears to depend upon the specific regional strategy and inter-regional distance of the company,43 its particular preferences – e.g., for coordinating regional operations, and/or for firm-internal knowledge diffusion – as well as its overall performance situation. However, taking into consideration these firm-specific circumstances, the practical implications of our research findings may offer relevant insights for corporate practice regarding appropriate degrees of regional management autonomy in the implementation of their regional strategies.
7.2.2
Regional Product/Service Adaptation as a Success Factor
Our research results further suggest that the extent to which MNCs regionally adapt their products/services highly affects their regional performance. More specifically, we found that a firm should substantially adapt its products/services to regional demands, as this has a highly positive effect on its regional success.44 Here, a MNC’s realized degree of regional adaptation of its products/services was found to be the result of their adjustments in the areas of regional design, regional functional scope, regional offering portfolio, and regional brands.45 This implies for corporate practice, that a firm has to develop and employ these four region-specific FSAs, or capabilities, to increase the regional adaptation of its products/services – to reap the respective regional performance improvements. Thus, a firm should try, along these four regional FSAs, to further increase the regional adaptation of its products/ services – which may require an extension of its respective abilities, either by better exploiting its existing strengths, or by building up additional competencies at the regional level. Furthermore, we found that increasing levels of a MNC’s regional orientation have a highly positive impact on the degree to which its regionally adapted products/
42
Here, we utilize the expression “many of our sample firms”, as our explanations above are largely based on a reflection of those current practices of our sample firms that – according to Table A.13 – are based on average values for different elements of their distinct regional strategies. Thus, it would be possible that some of our sample firms already apply this regional success factor in a proper manner. 43 It should be noted, however, that the extent to which a firm’s inter-regional distance has an influence on the delegation of decision-making autonomy to regional management – as outlined in Sects. 6.3 and 7.1 – depends on a MNC’s firm size and/or its regional competition. 44 This finding relates to the highly significant confirmation of our research hypothesis H2 (cf. Sect. 6.3). 45 Cf. Sect. 6.3.
7.2 Discussion of Practical Implications
189
services lead to regional success.46 Thus, a high orientation towards, or focus on, a particular region facilitates earning the benefits of highly adapted regional products/services. More specifically, a higher regional orientation of a MNC increases the additional value derived from its regional FSAs that it deploys to regionally adapt its products/services. The practical implication of this finding is that a MNC, which focuses on a geographically limited space for the exploitation of its FSAs for regional product/service adaptation, can augment the resulting regional performance potential – in comparison with more geographically diversified MNCs. In addition, our research findings show that increases in the inter-regional distance faced by a MNC highly positively influence the extent to which its regional product/service adaptations result in regional performance gains.47 Therefore, if a MNC is exposed to a rising inter-regional distance, its FSAs have to be aligned more strongly to regional demands, to achieve a higher level of regional product/ service adaptation – which improves its success at the regional level. For practitioners, this implies that they should closely match the degree of regional product/ service adaptation with the inter-regional differences faced by their firm in the cultural, administrative/political/institutional, geographic, and economic sphere. Here, they should respond to increasing differences between the firm’s home region and foreign regions by augmenting its region-specific capabilities that are required for a higher regional adaptation of their firm’s products/services. Similar to our observations in Sect. 7.2.1, the degree to which these implications from our hypotheses-based research findings are reflected in corporate practice – as illustrated for our survey sample firms in Table A.13 – leads to different results for MNCs pursuing home-regional, host-regional, bi-regional, and tri-regional strategies. If we analyze these differences in combination with the S-shaped development of the regional strategy–performance relationship, the following arguments result. First, our sample MNCs with home-regional strategies – compared to companies pursuing other regional strategies – have moderately adapted their products/ services to the demands within their home region, or respectively, their most successful foreign region.48 This does not, at least not completely, correspond to the practical implications outlined previously – showing that these firms do not yet tap the full potential of their regional FSAs. In addition, these home-regional MNCs perceive both their regional orientation and their inter-regional distance as being moderate. This implies that these companies should try to achieve higher levels of regional product/service adaptation, to improve their regional success – particularly as this, to some extent, is supported by their internal and external 46
This result corresponds to the highly significant confirmation of our research hypothesis H3b (cf. Sect. 6.3). 47 This finding is derived from the results for our research hypothesis H4b, for which we found a highly significant confirmation (cf. Sect. 6.3). 48 In our study – even though operating mainly within their home region – several MNCs with home-regional strategies have completed the answers of our survey for their most successful foreign region (cf. Sect. 6.1).
190
7 Discussion of Results and Implications
context. For example, these firms could analyze the inter-regional differences they face in the cultural, administrative/political/institutional, geographic, and economic sphere – to identify possible levers of FSAs that both increase the regional adaptation of their products/services and are able to better address the inter-regional distance they face. The latter form of FSA development could further accentuate the already existing inter-regional differences faced by these firms – which would increase the influence that the respective regional product/service adaptations have on their regional success. Here, the substantial corporate performance, which is usually achieved by home-regional MNCs,49 can provide the necessary funding that is needed for the development of new, or the better exploitation of existing, regional FSAs. Second, those firms of our sample following host-regional strategies – versus MNCs with other regional strategies – highly adapt their products/services to regional demands. This is consistent with the practical implications that we have presented previously – particularly if we consider that they perceive their interregional distance as being relatively high. At the same time, however, these hostregional companies indicated having a relatively low regional orientation. Thus, these MNCs face a favorable external context for the regional performance that results from their highly adapted regional products/services – whereas, respectively, no notable positive effects are to be expected from their regional orientation. As MNCs with host-regional strategies are normally characterized by a rather strong regional orientation towards, or focus on, a particular region,50 this implies for these firms that – in a first step – the reasons for their low regional orientation should be analyzed. A possible reasoning behind their currently low focus on a particular region might be that these MNCs try to improve their usually low levels of corporate performance,51 for example, by initiating an aggressive expansion strategy into further regional markets. In this case or in other situations, given their low regional orientation, these host-regional firms should ensure that they do not lose sight of the region – which constitutes the large bulk of their cash flows, assets, and employees. This requires a reorientation towards this region and/or more selectivity in their geographic growth ambitions. More specifically, to increase the regional success potential from their regional product/service adaptations, they should try to refocus their attention on the region – where the respective FSAs are mainly deployed. Third, our sample MNCs pursuing bi-regional strategies – vis-a`-vis firms following other regional strategies – consider the regional adaptation of their products/ services as being moderate. Such a degree of regional product/service adaptation does not fully reflect the practical implications that we have outlined earlier – particularly if we take into account their moderate subjective rating for their regional orientation and the inter-regional distance they face. This implies for these firms with bi-regional strategies that they should try to increase the extent of their
49
Cf. Sect. 6.2.2. Cf. Sect. 2.3.2.1. 51 Cf. Sect. 6.2.2. 50
7.2 Discussion of Practical Implications
191
regional product/service adaptations – corresponding to the earlier implication for the companies pursuing home-regional strategies. Due to their strong presence in two regions, bi-regional MNCs, however, can build on a much higher regional reach of their FSAs than home-regional firms – to increase the regional adaptation of their products/services. Therefore, these bi-regional companies should try to identify those needs of their regional customers in each of its two major regions – which are not yet fully satisfied by their existing regional product/service adaptations in the fields of regional design, regional functional scope, regional offering portfolio, and regional brands. Based upon the subsequent development and employment of respective regional downstream FSAs, these bi-regional MNCs improve their ability to link their region-specific capabilities with location-specific advantages, in the form of CSAs, within the distant regions.52 As a result, these biregional MNCs can access region-bound CSAs more easily, like for example different absolute cost levels for important firm resources. Consequently – apart from the positive impact on regional success of these regional product/service adaptations – these bi-regional MNCs can add further value by arbitrage, as they have improved their ability to exploit differences across these two regions. Over the long run, these firms’ additional value created by such inter-regional arbitrage advantages can highly contribute to, or even offset, the initial costs and investments required to regionally adapt their products/services. Thus, the financial risk incurred by these bi-regional firms from such regional product/service adaptations, considering that MNCs with bi-regional strategies usually are highly successful at the corporate level,53 appears to be manageable. Fourth, companies with tri-regional strategies – in comparison to firms with other regional strategies – moderately adapt their products/services to regional demands. This does not conform, at least to a certain extent, to the practical implications presented previously. In addition, these MNCs indicate they have a moderate regional orientation – while they perceive to be exposed to a relatively high inter-regional distance. This implies that these firms, particularly as they face high inter-regional differences, should try to increase the regional adaptation of their products/services. However, tri-regional MNCs can benefit from global integration benefits across three regions – which may arise from commonalities in the regional customers’ basic product/service requirements.54 These competitive aggregation advantages – in the form of economies of scale and competitive cost advantages – should not be jeopardized by regional product/service adaptations. In particular with respect to the already relatively low corporate performance levels, by which tri-regional firms are usually characterized,55 such countervailing effects in their efforts to create value should be avoided. In other words, these tri-regional
52
Such market-oriented investments are perceived by our bi-regional sample firms as a well-suited means to increase their regional commitment. Cf. Table A.13 and Sect. 6.2.2. 53 Cf. Sect. 6.2.2. 54 Cf. Sects. 2.1.1, 2.1.2.4, and 7.1. 55 Cf. Sect. 6.2.2.
192
7 Discussion of Results and Implications
MNCs should try to increase the regional adaptation of their products/services to a degree, which optimizes the mix between regional responsiveness and regional integration. For this appropriate degree of regional product/service adaptations, for example, a detailed analysis of these firms’ value chains could produce valuable insights – concerning both their downstream and upstream capabilities and the relative contributions of such FSA types to the corporate performance of these triregional firms. More specifically, this could be achieved by comparing the effects from regional performance increases – which result from higher product/service adaptations to regional markets on the basis of downstream FSAs – with efficiency gains from exploiting upstream FSAs across regions. For a particular product/ service, for instance, this may involve comparisons of, on the one hand, incremental revenues from adapting this product/service to regional customer demands – with, on the other hand, decreasing costs per unit from utilizing a more efficient production system for this product/service on the basis of its standardization across regional markets. Here, the high inter-regional distance faced by these tri-regional MNC provides an argument for exploiting the former, downstream FSAs at the customer-end – as the resulting regional adaptations of products/services have a stronger positive impact on regional success in light of this external context. More generally, these explanations demonstrate that many of our sample MNCs do not yet fully exploit the regional performance potential from appropriate degrees of regional product/service adaptation,56 given by highly regionally adapted products/services. An adequate application of this success factor of regional strategies appears to depend upon the particular regional strategy and capabilities of the firm, its regional orientation, inter-regional distance, and its corporate performance situation. However, the region-specific capabilities of a MNC, which are required to achieve the regional adaptation of its products/services – if properly combined with complementary assets at the regional level, and aligned to the internal and/or external context of the firm – constitute a key competitive advantage along adaptation, aggregation, and/or arbitrage dimensions. Thus, depending upon the particular firm situation, the practical implications of our research findings may offer relevant insights for practitioners of how to improve the regional competitiveness of their company by appropriate degrees of regional product/service adaptations.
7.3
Limitations
While our research points to relationships that may contribute to advancing the existing knowledge about regional strategies in theory and practice, the results and implications of our study need to be interpreted with caution, as they are subject to 56
Here, we use the term “many of our sample MNCs”, as our explanations are largely based upon a reflection of those current practices of our sample MNCs that – according to Table A.13 – are based on average values for different elements of their distinct regional strategies. Therefore, maybe some our sample companies already apply this regional success factor appropriately.
7.3 Limitations
193
several limitations. These limitations are mainly related to methodological and theoretical matters – which we will outline in the following. With respect to the methodological limitations of this work, six aspects in particular must be highlighted. First, the data for our research was collected solely from the Fortune Global 500 firms, which results in a bias of our findings towards large companies according to the criterion sales revenue. While the respective database is comprised of, on average, approximately 20,000 data points each year,57 they solely represent 663 MNCs in 16 industries, over a total time period of only 9 years from 2000 to 2008.58 Furthermore, these companies originate almost exclusively from within the broad triad of North America, Europe, and the AsiaPacific, and only to a minor part from South America – while our database does not include any African firm at all.59 These restrictions of our data set do not allow us to generalize our findings to small and medium-sized enterprises (SMEs), to other industries, to different points in time, and to firms from all regions of the world. Second, this focus of our research on exclusively 663 firms resulted in a small sample size of 96 companies – for which we were able to produce mainly descriptive and explorative insights.60 In the PLS-based modeling of our data however, solely 36 firms could be considered.61 Whereas 36 cases are sufficient to carry out PLS calculations,62 this very small sample size severely constrains the generalizability of our findings – particularly if we take into consideration the restrictions of our data set that we have outlined before. Furthermore, some of our implications are based upon a joint interpretation of our findings for both sample sizes – which methodologically limits the validity of these implications. Third, we collected our secondary data by means of archival information from annual financial reports and/or from SEC filings, while our primary data was derived from a web-based survey. For both types of data sources, the applied methods for data collection can be challenged. With respect to our secondary data sources, even though we took several measures to improve the comparability of the information provided in our sample companies’ financial statements,63 the application of accounting and tax rules may highly diverge from one firm to the other. This can, at least to some extent, distort the measures applied in our research. In the collection of our primary data with a web-based questionnaire, by directly contacting our respondents prior to its transmission, we were able to shortly describe the
57
More specifically, the figure of 20,000 data points is derived from dividing the total number of 179,010 data points in our database by the 9 years of our sample period – which results in 19,890 data points in each year (cf. Sect. 5.1.2). 58 Cf. Sect. 5.1.1. 59 Cf. Sect. 6.1. 60 Cf. Sects. 6.1 and 6.2. 61 Cf. Sect. 5.2.3. 62 Cf. Sects. 3.1 and 5.2.3. 63 Cf. Sect. 5.1.2.
194
7 Discussion of Results and Implications
background of our research.64 However, this procedure did not allow for a detailed explanation of the questions to our respondents. Therefore, even though we conducted two in-depth pretests,65 our respondents may still have experienced difficulties in understanding the questions of our survey. This might have reduced, to a certain degree, the meaningfulness of their responses and/or the extent to which their responses are comparable with each other – particularly as we were not able to detect differences in the respondents’ interpretation of the questions. Fourth, our secondary data relates to the years 2000–2008, while our primary data was collected in 2009. This limits the timely consistency of our research, as we relate our secondary data, mainly for the period from 2004 to 2008, with our primary data from 2009 – to derive several of our explorative findings, and our PLS-based research results.66 However, it should be noted that the questions on the current practical application of different elements of regional strategies in our survey are the result of our sample firms’ strategic decisions in the past. This leads to a convergence, at least approximately, of the different time intervals applied in our research – whereas their complete synchronization appears to be rather unlikely. Fifth, our primary data is based on the subjective perceptions of single managers at the corporate headquarters of our sample firms. While we have controlled for this key informant bias by various measures,67 influences from the respondents’ subjectivity could only be reduced to a certain degree, by a dyadic research design – which grounds on both primary and secondary data sources.68 The focus specifically on respondents from the corporate headquarters of our sample MNCs, however, neglects the perceptions of national or regional managers at other levels within these firms’ organizations. Sixth, the PLS method that we applied as the SEM technique is only one of various statistical methods to analyze our data. Even though we showed that PLS is the appropriate modeling technique for this work, it has methodological disadvantages,69 which other statistical methods could have avoided – and that thus, perhaps, would have led to different results. Furthermore, while PLS enables us to analyze causal relationships, definite conclusions about causality are not possible. We exclusively built on theory for deriving our assumptions about the causal relationships between exogenous and endogenous variables, whereas additional methods to more systematically examine their causality – e.g., experimental research – are not employed. In addition to methodological concerns, there exist theoretical limitations of this study that are related to five issues. First, the knowledge offered by our work is
64
Cf. Sect. 5.2.2.2. Cf. Sect. 5.2.2.2. 66 Cf. Sects. 5.2.1 and 5.2.3. 67 Cf. Sect. 5.2.2.1. 68 Cf. Chap. 5. 69 Cf. Sect. 3.1. 65
7.3 Limitations
195
limited within the cognitive theory frameworks of Stopford and Wells (1972), Ohmae (1985), Bartlett and Ghoshal (1989), and particularly Rugman (2005b). Furthermore, our findings are the result of our own, subjective interpretation of the research object – success factors of regional strategies as appropriate degrees of regional management autonomy and regional product/service adaptation. Consequently, if accepting multiple paradigms and multiple truths, the knowledge offered by this work does not represent a final truth, builds on the basis of own assumptions – and therefore might be incommensurable with the knowledge provided by other studies (Scherer 2006: 40–43). Second, within the field of regionalization theory, we took a continental perspective focusing almost exclusively on the level of the region,70 to theoretically explain success factors of regional strategies. However, regional strategies can also be interpreted at other geographical layers like the global, subcontinental, national, intra-national, or local level (Ghemawat 2005: 107) – which are largely neglected in this work. Third, the success factors of regional strategies were conceptually derived, in the case of regional management autonomy, mainly from Enright’s (2005a, b) findings on the functions of regional management centers.71 Our approach, however, does not consider other influences in explaining the effect of regional management autonomy on regional success – such as the processes of delegating decisions, or influences from different forms of decision-making (e.g., consensus-driven versus hierarchical). Furthermore, our theoretical explanation for the regional adaptation of products/services is largely based on insights from international marketing theory and from Rugman’s (2005b) interpretation of the resource-based view.72 This results in a relatively narrow theoretical perspective on the complex phenomenon of how MNCs regionally adapt their products/services to realize regional performance. Fourth, a foundation in organization-theory for deriving success factors of regional strategies was achieved exclusively by contingency theory.73 Furthermore, we limited the theoretical contribution of this approach – for the explanation of changes in the effectiveness of a regional organization – solely to one internal contextual variable, a MNC’s regional orientation, and to only one external contingency variable, the inter-regional distance faced by the firm. Fifth, based on success factor theory, we have formulated requirements for “good” success factor research.74 However, concerning both practical relevance and academic rigor, these requirements are presumably incomplete. By means of the previous explanations, we have presented our main concerns regarding methodological and theoretical aspects of our work. Recognizing these
70
Cf. Sect. 1.2. Cf. Sects. 2.2.7 and 4.1.1. 72 Cf. Sects. 2.2.1, 2.2.8, and 4.1.2. 73 Cf. Sects. 2.3 and 4.1.3. 74 Cf. Sect. 2.4.1. 71
196
7 Discussion of Results and Implications
limitations, our study of success factors of regional strategies – being both rigorous and relevant – however offers a new approach to think about the geographic competitiveness of MNCs. In the following chapter, our main insights related to this novel perspective will be briefly summarized.
Chapter 8
Conclusions
8.1
Conclusion
In the course of the growing regionalization of worldwide markets, the regional strategies of MNCs have increasingly received attention as a road map to crossborder competition. However, there exists only limited knowledge in both theory and practice about how MNCs should design their regional strategies to be successful at the regional level. The objective of this work was to enhance this understanding about the factors that drive the success of MNCs in their pursuit of regional strategies. Apart from exploring the main variables that influence these success factors of regional strategies, we particularly intended to identify their underlying causal relationships – taking into consideration interaction effects from the internal and external context of MNCs. On the basis of different IB theories, and particularly the recent academic work in the field of regionalization theory, we could derive regional management autonomy and regional product/service adaptation as success factors of regional strategies. The contingency approach, as an organization theory for MNCs, provided the conceptual basis for the investigation of their internal and external context – in the form of their regional orientation and the inter-regional distance they face. Building on theoretical and empirical insights from different regionalization scholars, studies about regional management centers, and international marketing theory, we generated hypotheses about success factors of regional strategies. High degrees of regional management autonomy and of regional product/service adaptation were both proposed to have a positive effect on regional success. These relationships were assumed to be positively influenced by interaction effects from the MNCs’ regional orientation and the inter-regional distance they face. To test these hypotheses, we collected data for 663 companies of the Fortune Global 500 firms from 2000 to 2008. This data and 96 responses of these firms to a survey-based inquiry enabled us, on the one hand, to carry out an explorative analysis of the main variables that influence the success factors of regional strategies – and, on the other hand, to analyze their hypothesized causal relationships in a structural equation model. The PLS approach was applied to estimate this model.
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1_8, # Springer-Verlag Berlin Heidelberg 2011
197
198
8 Conclusions
Our explorative findings show that, while MNCs regionally expand across all continents of the world, their regional strategies and regional success are mainly restricted to their home region. A longitudinal study indicated a growth outside of this region by the MNCs’ gradual expansion into another foreign host region. A possible reason for this trend was provided by our evidence for an S-curve development of the regional strategy–performance relationship, where bi-regional strategies were associated with the relatively highest corporate performance. Sensitivity analyses supported the existence of multi-regional strategies, while truly global strategies and success patterns of MNCs were observed to be far from common. The PLS-based results provide evidence for the existence of success factors of regional strategies for MNCs, given by appropriate degrees of regional management autonomy and the regional adaptation of products/services. The extent to which these success factors of regional strategies can be effectively applied by MNCs was found to be contingent on the regional orientation of MNCs and/or the inter-regional distance they face. Contrary to our expectations, high levels of regional management autonomy were found to negatively impact the regional success of MNCs – indicating that, if at all, rather low degrees of decision-making autonomy should be granted to regional management. As suggested, the evidence supports the fact that a high regional adaptation of products/services by MNCs positively affects their regional performance. Region-specific capabilities, or FSAs, were found to be an important influence in the regional adaptation of the MNCs’ products/services – given by their regional design, regional functional scope, regional offering portfolio, and regional brands. Our results only provided partial support for the interaction effects of the MNCs’ internal and external context. The negative influence of high levels of regional management autonomy on regional success was found to be mitigated by the MNCs’ inter-regional distance – while no notable effect from the regional orientation of MNCs could be identified. However, in conformity with our hypotheses, the positive regional performance effects resulting from high degrees of regional product/service adaptations by MNCs were found to be highly positively affected by both their regional orientation and the inter-regional distance they face. Based on theoretical and practical implications of these research findings, we showed how low degrees of regional management autonomy and high levels of regional product/service adaptation, based on region-specific FSAs, can improve the MNCs’ success at the regional level. Limits and opportunities of these success factors of regional strategies were discussed in the light of the MNCs’ regional orientation and the inter-regional distance they are exposed to. Ultimately, the realization of a proper regional environment-strategy-structure fit was found to be highly dependent upon each MNC’s specific situation. We hope that the findings of our study make a valuable contribution to the existing knowledge in theory and practice about the framing and implementation of regional strategies in MNCs. Apart from providing a practical guidance for managers to better understand the success factors of regional strategies, and to deal with certain interaction effects, our results can act as a starting point for further empirical
8.2 Recommendations for Further Research
199
research. In the following, we will give some direction on several possible avenues for such research advances.
8.2
Recommendations for Further Research
With respect to the limitations of this work,1 we can recommend five possible lines of further research in particular. First, future research could investigate the factors that influence the regional success of individual firms, or of specific industries. To achieve a more profound understanding of the firm- and sector-specific nature of success factors of regional strategies, we suggest the application of qualitative research designs for such analyses – such as case studies or interviews. Second, researchers could change the perspective that we applied in this work for the examination of success factors of regional strategies. This could involve a focus on SMEs,2 other geographical levels than the region, and/or on the perceptions of national and/or regional managers instead of the corporate headquarters of MNCs. Third, there is a rich avenue of work to be pursued on the different forms of successfully managing regional operations, particularly if decision-making responsibilities are allocated from the region to other organizational levels. This may include, for instance, an analysis of how corporate headquarters and/or national subsidiaries configure intra-regional coordination mechanisms and intra-/interregional knowledge transfers, to improve MNC performance at the regional level. Valuable insights for such research activities may be derived, for example, by applying existing accounts on strategic firm networks (e.g., Gilbert 2003, 2005) and/or on flagship networks (e.g., Rugman and D’Cruz 2000) to the regional level. Fourth, further research could be directed towards the processes that are associated with the regional adaptation of products/services. This could be achieved by investigating how region-specific capabilities are developed to meet regional demands, for example, related to the regional design, regional functional scope, regional offering portfolio, and/or regional brands of products/services. Apart from the development of these FSAs, such an analysis could examine how each of these four FSA is linked with region-bound/location-bound advantages in the form of CSAs (Rugman 2005b: 235). In the respective research design, it could be worthwhile to consider the perceptions of regional customers, governments, suppliers, employees, and other stakeholders of MNCs. Fifth, a fruitful field for future research is provided by extending the analysis of the regional environment-strategy-structure fit by additional internal and external contextual variables. Possible internal contingency variables, besides a MNC’s regional 1
Cf. Sect. 7.3. Valuable research findings on regional activities of SMEs are provided, for example, by Bassen et al. (2001: 420) and Behnam and Gilbert (2001: 107–108). 2
200
8 Conclusions
orientation, could include a firm’s regional commitment (e.g., its staying power while a region faces periods of difficulty), and/or its focus on different mechanisms for value creation (e.g., adaptation, aggregation, and arbitrage) (e.g., Ghemawat 2007a, 2008). Influences from the external context of MNCs could include the degree of competition they are exposed to, for example, at the intra-national, national, and/or global level – or within their particular industries. Any of these different research avenues could lead to a more profound understanding of the success factors of regional strategies. Such insights would further advance the existing knowledge in both theory and practice about how a regional extent of the market can be translated into an improved geographic competitiveness of MNCs. Ultimately, this would contribute to answering the question of how MNCs can be regionally successful by: “[. . .] appropriate strategies at the level of each broad region [. . .]” (Rugman 2009: 293).
Appendix
Fig. A.1 (continued)
P. Heinecke, Success Factors of Regional Strategies for Multinational Corporations, Contributions to Management Science, DOI 10.1007/978-3-7908-2640-1, # Springer-Verlag Berlin Heidelberg 2011
201
202
Fig. A.1 (continued)
Appendix
Appendix
Fig. A.1 (continued)
203
204
Fig. A.1 (continued)
Appendix
Appendix
Fig. A.1 (continued)
205
206
Fig. A.1 (continued)
Appendix
Appendix
Fig. A.1 (continued)
207
208
Fig. A.1 (continued)
Appendix
Appendix
Fig. A.1 (continued)
209
210
Fig. A.1 (continued)
Appendix
Appendix
Fig. A.1 (continued)
211
212
Fig. A.1 (continued)
Appendix
Appendix
Fig. A.1 (continued)
213
214
Fig. A.1 (continued)
Appendix
Appendix
Fig. A.1 (continued)
215
216
Appendix
Fig. A.1 Survey of this work Source: Own illustration
No.of MNCs 210
North America 234 (35.3%)
Asia-Pacific 192 (29.0%) Africa -
Europe 230 (34.7%)
108
45 45 43
South America 7 (1.1%)
36 19 18 16 16 13 11 10 9
8
7
6
6
6
5
5
2
2
2
2
2
2
Fig. A.2 Home countries and home regions of MNCs Source: Own illustration based on the total research sample of this work
1
1
1
1
1
1
1
1
1
Appendix
217
No. of MNCs
359 (54.1%) 304 (45.9%)
122 72 14
28
46
36
41
30
64
62
36
35
29
33
9
6
Fig. A.3 Industry membership of MNCs Source: Own illustration based on the total research sample of this work
In millions of US dollars, unless otherwise stated 250.000
200.000
150.000
100.000
50.000
0 2000
2001
2002
2003
2004
Sales revenue
Profits
Assets
Company data (in millions of US dollars, unless otherwise stated) Sales revenue
2005
2006
2007
2008
Employees (number of)
2000
2001
2002
2003
2004
2005
2006
2007
2008
29.426
28.940
28.786
30.982
34.638
38.595
42.144
47.381
50.562
Profits
1.486
721
283
1.490
1.929
2.468
3.086
3.192
1.674
Assets
94.163
97.865
104.465
120.355
134.429
145.129
167.123
207.547
199.235
101,926
101,356
98.054
94,621
98,299
102,554
105,888
108,568
113,524
Employee (number of)
Fig. A.4 Company data of the total research sample Source: Own illustration based on the total research sample of this work and the data for these Fortune Global 500 firms according to the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009)
218
Appendix
In % 70 60 50 40 30 20 10 0 2000 US GAAP
2001 IFRS
2002
2003
2004
2005
2006
2007
2008
Local accounting standard (e.g., Japanese GAAP, Korean GAAP, Australian GAAP)
Fig. A.5 Applied accounting standards Source: Own illustration based on the total research sample of this work and the information provided in the annual reports of these firms from 2000 to 2008
(No.of MNCs) 49.0% (47) 35.3% (234)
34.7% (230)
29.0% 29.2% (192) (28)
20.8% (20)
1.1% 1.0% (7) (1)
North America
Europe
Asia-Pacific
Total research sample
South America
-
-
Africa
Survey sample
Fig. A.6 Home regions of MNCs Source: Own illustration based on the total research sample and the survey sample of this work
Appendix
219
(No.of MNCs)
Total research sample
Survey sample
Fig. A.7 Industry membership of MNCs Source: Own illustration based on the total research sample and the survey sample of this work
In % 100 90 80 70 60 50 40 30 20 10 0
2000
2001
2002
2003
2004
Home-regional success
Host-regional success
Quad-regional success
Global success
2005
2006
Bi-regional success
Rest-regional success
Fig. A.8 Regional success based on the regional sales spread of MNCs Source: Own illustration based on the total research sample of this work
2007
2008
Tri-regional success
220
Appendix
In % 100 90 80 70 60 50 40 30 20 10 0 2000
2001
2002
2003
2004
Home-regional success
Host-regional success
Quad-regional success
Global success
2005
2006
Bi-regional success
2007
Tri-regional success
Rest-regional success
Fig. A.9 Regional success based on the regional asset dispersion of MNCs Source: Own illustration based on the total research sample of this work
In % 80 75 70 65 60 2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100%-HR MNCs
Excl. 100%-HR MNCs
Fig. A.10 Percentage of intra-regional sales within the home region Source: Own illustration based on the total research sample of this work
2008
Appendix
221 In % 85 80 75 70 65 60 2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100% -HR MNCs
Excl. 100% -HR MNCs
Fig. A.11 Percentage of intra-regional assets within the home region Source: Own illustration based on the total research sample of this work
In %
90 85 80 75 70 65 60 55 50 2000
2001
2002
2003
2004
2005
2006
2007
2008
Manufacturing sector incl. 100%-HR MNCs
Service sector incl. 100%-HR MNCs
Manufacturing sector excl. 100%-HR MNCs
Service sector excl. 100%-HR MNCs
Fig. A.12 Percentage of intra-regional sales within the home region by industry category Source: Own illustration based on the total research sample of this work
222
Appendix
In % 95 90 85 80 75 70 65 60 55 50
2000
2001
2002
2003
2004
2005
2006
2007
2008
Manufacturing sector incl. 100%-HR MNCs
Service sector incl. 100% -HR MNCs
Manufacturing sector excl. 100% -HR MNCs
Service sector excl. 100% -HR MNCs
Fig. A.13 Percentage of intra-regional assets within the home region by industry category Source: Own illustration based on the total research sample of this work
In % 100 90 80 70 60 50 40 30 20 2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100%-HR MNCs
Excl. 100%-HR MNCs
Fig. A.14 Percentage of intra-regional profits within the home region Source: Own illustration based on the total research sample of this work
Appendix
223 In % 85 80 75 70 65 60 55 50 2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100%-HR MNCs
Excl. 100%-HR MNCs
Fig. A.15 Percentage of intra-regional investments within the home region Source: Own illustration based on the total research sample of this work
In % 9 8 7 6 5 4 3 2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100%-HR MNCs
Excl. 100%-HR MNCs
Fig. A.16 Investment quota for the home region Source: Own illustration based on the total research sample of this work
224
Appendix In percentage points 16 14 12 10 8 6 4 2 0 –2
2000 2001 2002 2003 2004 2005 Incl. 100%-HR MNCs
2006 2007 2008
Excl. 100%-HR MNCs
Fig. A.17 Difference of home-regional to corporate investment quota Source: Own illustration based on the total research sample of this work
In % 85 80 75 70 65 60 55 50 2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100%-HR MNCs
Excl. 100%-HR MNCs
Fig. A.18 Percentage of intra-regional employees within the home region Source: Own illustration based on the total research sample of this work
Appendix
225 In millions of US dollars 3
2
1
0 2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100%-HR MNCs
Excl. 100%-HR MNCs
Fig. A.19 Home-regional sales revenue per employee Source: Own illustration based on the total research sample of this work
In thousands of US dollars 80 60 40 20 0 –20 – 40 – 60 – 80
2000 2001 2002 2003 2004 2005 2006 2007 2008 Incl. 100%-HR MNCs
Excl. 100%-HR MNCs
Fig. A.20 Home-regional profit per employee Source: Own illustration based on the total research sample of this work
226
Appendix In percentage points 120 100 80 60 40 20 0 –20 –40 –60 –80
2001
2002
2003
2004
Incl. 100% -HR MNCs
2005
2006
2007
2008
Excl. 100% -HR MNCs
Fig. A.21 Difference of home-regional to corporate revenue growth Source: Own illustration based on the total research sample of this work
In percentage points
80 60 40 20 0 –20 –40 –60 –80
2001
2002
2003
2004
Incl. 100%-HR MNCs
2005
2006
2007
2008
Excl. 100%-HR MNCs
Fig. A.22 Difference of home-regional to corporate profit growth Source: Own illustration based on the total research sample of this work
Appendix
227
In % 90 80 70 60 50 40 30 20 10 0 50%
55%
60%
65%
70%
Home-regional
Host-regional
Quad-regional
Global
75%
80%
Bi-regional
85%
90%
95%
Tri-regional
Rest-regional
Fig. A.23 Sensitivity analyses of regional strategies based on 20% host region threshold Source: Own illustration based on the total research sample of this work
In % 90 80 70 60 50 40 30 20 10 0 50%
55%
60%
65%
70%
Home-regional
Host-regional
Quad-regional
Global
75%
80%
Bi-regional
85%
90%
Tri-regional
Rest-regional
Fig. A.24 Sensitivity analyses of regional strategies based on 15% host region threshold Source: Own illustration based on the total research sample of this work
95%
228
Appendix
In % 90 80 70 60 50 40 30 20 10 0 50%
55%
60%
65%
70%
Home-regional
Host-regional
Quad-regional
Global
75%
80%
Bi-regional
85%
90%
95%
Tri-regional
Rest-regional
Fig. A.25 Sensitivity analyses of regional strategies based on 10% host region threshold Source: Own illustration based on the total research sample of this work
Likert scale value 4.5 4.0 3.5 3.0 2.5 2.0 1.5 Home-regional
Host-regional
Regional product/service adaptation Inter-regional distance
Bi-regional
Tri-regional
Regional management autonomy
Regional orientation
Fig. A.26 Regional strategies of MNCs and survey responses on regional management autonomy, regional product/service adaptation, regional orientation, and inter-regional distance Source: Own illustration based on the survey sample of this work
Appendix
229
Likert scale value 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
Fig. A.27 Regional strategies of MNCs and survey responses on regional management autonomy by industry Source: Own illustration based on the survey sample of this work
Likert scale value 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
Fig. A.28 Regional strategies of MNCs and survey responses on regional product/service adaptation by industry Source: Own illustration based on the survey sample of this work
230
Appendix
Likert scale value 4 3.5 3 2.5 2 1.5 1 0.5 0
Fig. A.29 Regional strategies of MNCs and survey responses on regional orientation by industry Source: Own illustration based on the survey sample of this work
Likert scale value 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
Fig. A.30 Regional strategies of MNCs and survey responses on inter-regional distance by industry Source: Own illustration based on the survey sample of this work
Appendix
231
Likert scale value 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 Home-regional Directiveregional
Host-regional Adaptiveregional
Bi-regional Trans-regional (world mandate)
Tri-regional Cooperative-regional
Fig. A.31 Regional strategies of MNCs and survey responses on the different coordination forms of further regional strategies Source: Own illustration based on the survey sample of this work
Likert scale value 4.5 4 3.5 3 2.5 2 Home-regional
Host-regional
Know-how transfer from home region to foreign region
Bi-regional
Tri-regional
Know-how transfer from foreign region to home region
Intra-regional knowledge sharing
Fig. A.32 Regional strategies of MNCs and survey responses on the inter-regional know-how transfer and on the intra-regional knowledge sharing Source: Own illustration based on the survey sample of this work
232
Appendix
Likert scale value 5 4.5 4 3.5 3 2.5 2 Home-regional
Host-regional
Management continuity
Bi-regional
Market-oriented investments
Tri-regional Staying power
Fig. A.33 Regional strategies of MNCs and survey responses on different aspects of regional commitment Source: Own illustration based on the survey sample of this work
In billions of US dollars 100 80 60 40 20 0 Home-regional
Host-regional
Bi-regional
Sales-based classification
Profit-based classification
Assets-based classification
Employees-based classification
Tri-regional
Fig. A.34 Regional strategies and total corporate performance Source: Own illustration based on the total research sample of this work and the data for these Fortune Global 500 firms according to the magazine Fortune (2005, 2006, 2007, 2008, 2009)
Appendix Table A.1 Industry classification This work Manufacturing sectors 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
Service sectors Banks Entertainment, printing and publishing
Merchandisers
Other financial services
Other services
233
Fortune Global 500 Aerospace and defense 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 Pipelines Beverages Food consumer products Food production Food services Tobacco Automotive retailing, services Motor vehicles and parts Forest and paper products Mining, crude-oil production Apparel Household and personal products Industrial and farm equipment Scientific, photo, control equipment Banks: Commercial and savings Entertainment Hotels, casinos, resorts Publishing, printing Food and drug stores General merchandisers Specialty retailers Wholesalers: Electronics and office equipment Wholesalers: Health care Wholesalers: Other Diversified financials Insurance: Life, health (mutual) Insurance: Life, health (stock) Insurance: Property and casualty (mutual) Insurance: Property and casualty (stock) Securities Computer services and software Diversified outsourcing services (continued)
234
Appendix
Table A.1 (continued) This work
Fortune Global 500 Health care: Insurance and managed care Health care: Other Homebuilders Information technology services Internet services and retailing Oil and gas equipment, services Temporary help Trading Telecommunications and utilities Telecommunications Utilities Transportation services Airlines Mail, package and freight delivery Railroads Shipping Source: Own illustration on the basis of Rugman and Oh (Rugman 2005b: 79–182; Rugman and Oh 2007: 42–43), Hoover’s (2009) industry information, and the Fortune Global 500 firms from 2000 to 2008 according to the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009)
Table A.2 Categories of geographical segment reporting No. of MNCs Sales revenue Revenue Sales or turnover Net sales or net turnover (Net) operating revenue Net revenue Gross sales revenue Segment revenue Income (Total) ordinary income Gross or net premium income Net interest income Gross or net banking income
603 260 143 115 53 26 3 3 60 37 11 8 4
91.0% 39.2% 21.6% 17.3% 8.0% 3.9% 0.5% 0.5% 9.0% 5.6% 1.7% 1.2% 0.6%
Profit Profit or earnings Net income Net profit or net surplus Net earnings Profit after taxes Segment result Operating profit Operating income Operating profit or earnings EBIT
396 235 124 19 16 1 1 208 108 64 22
59.7% 35.4% 18.7% 2.9% 2.4% 0.2% 0.2% 31.4% 16.3% 9.7% 3.3% (continued)
Appendix Table A.2 (continued) No. of MNCs Earnings from operations EBITDA Operating profit before depreciation and amortization Operating profit before taxes EBITA Net operating income Profit before taxes Profit or earnings before income taxes (Gross) income before taxes
235
4 4 2 2 1 1 59 31 28
0.6% 0.6% 0.3% 0.3% 0.2% 0.2% 8.9% 4.7% 4.2%
Investments Capital expenditures or capital investments (Net) investments Capital additions Additions to long-lived assets Investments in property, plant, and equipment (Net) purchases of property, plant, and equipment Additions to property, plant, and equipment Investments in, or expenditures for, property, plant, and equipment Capitalized fixed asset additions Additions to, or purchases of, fixed assets Investments in property, plant and equipment, and intangible assets Investments in property, plant and equipment, intangible assets (and investment property) Acquisition of fixed assets and intangible assets Additions to property, plant and equipment, and additions to acquired intangible assets Additions to fixed assets and intangible assets Purchases of fixed and intangible assets
584 453 128 2 1 44 21 9 9 3 2 35
88.1% 68.3% 19.3% 0.3% 0.2% 6.6% 3.2% 1.4% 1.4% 0.5% 0.3% 5.3%
13
2.0%
8 7
1.2% 1.1%
4 3
0.6% 0.5%
Assets (Total) assets Segment assets Long-lived assets (Total) non-current assets Property, plant, and equipment, and intangible assets Net assets Average assets Identifiable assets Fixed assets (Net) property, plant, and equipment (Net) fixed assets Operating assets
610 439 68 62 20 9 8 2 2 45 38 7 8
92.0% 66.2% 10.3% 9.4% 3.0% 1.4% 1.2% 0.3% 0.3% 6.8% 5.7% 1.1% 1.2%
Employees All firms used the word “employees” (or a synonym, e.g. headcount)
663
100%
Source: Own illustration based on the total research sample of this work and the information provided in the annual reports of these firms from 2000 to 2008
236 Table A.3 Company information of the survey sample firms Company information Frequencies Absolute number Regional operating experience (in yrs.) <1 3 1–3 3 4–5 5 6–10 8 11–20 18 21–30 12 31–50 22 >50 22 Don’t know 3 Regional management center experience (in yrs.) <1 8 1–3 4 4–5 7 6–10 11 11–20 15 21–30 15 31–50 17 >50 11 Don’t know 8 Regional subsidiary density (no. of subsidiaries) 1–2 15 3–4 17 5–6 15 7–8 5 9–10 7 11–20 9 21–30 2 >30 18 Don’t know 8
Appendix
Percentage 3% 3% 5% 8% 19% 13% 23% 23% 3% 8% 4% 7% 11% 16% 16% 18% 11% 8% 16% 18% 16% 5% 7% 9% 2% 19% 8%
Source: Own illustration based on the survey sample of this work
Table A.4 Respondent information of the survey sample firms Respondent information Frequencies Absolute number Age (in yrs.) 20–30 11 31–40 28 41–50 37 51–60 17 >60 3 Management level Middle management 59 Senior management 37
Percentage 11% 29% 39% 18% 3% 61% 39% (continued)
Appendix Table A.4 (continued) Respondent information Management experience (in yrs.) <1 1–3 4–5 6–10 11–20 21–30 Functional area Administration Business development or strategy Communications Finance General management Marketing, sales and after sales Research & development (R&D)or service development Functional experience (in yrs.) <1 1–3 4–5 6–10 11–20 21–30
237
Frequencies Absolute number
Percentage
5 22 19 30 15 5
5% 23% 20% 31% 16% 5%
2 65 2 7 10 8 2
2% 68% 2% 7% 10% 8% 2%
8 23 22 26 15 2
8% 24% 23% 27% 16% 2%
Source: Own illustration based on the survey sample of this work Table A.5 Ratio of home-regional to total sales and ratio of home-regional to total assets by industry Average values from 2000 to Ratio of home-regional to total Ratio of home-regional to total 2008 (in %) sales (HRS/TS) assets (HRA/TA) Incl. 100%Excl. 100%Incl. 100%Excl. 100%HR MNCs HR MNCs HR MNCs HR MNCs Manufacturing sector 65.6 61.3 70.8 66.4 Aerospace and defense 63.1 63.1 72.1 72.1 Chemicals and pharmaceuticals 56.7 53.7 65.2 62.6 Computer, office and 54.5 53.6 65.1 64.3 electronics Construction, building 74.8 69.5 79.1 73.7 materials and glass Energy, petroleum and refining 88.3 79.1 86.6 75.0 Food, drug and tobacco 60.9 53.5 62.3 54.5 Motor vehicle and parts 62.7 59.7 69.2 67.0 Natural resource manufacturing 67.8 60.8 69.0 62.9 Other manufacturing 61.7 58.6 68.7 65.7 Service sector 84.0 75.0 87.2 77.5 Banks 84.7 80.4 82.3 76.4 Entertainment, printing and 84.8 80.1 93.3 90.8 publishing Merchandisers 87.1 78.6 86.4 76.8 Other financial services 77.5 67.3 84.8 73.5 Other services 78.4 59.9 78.4 57.2 (continued)
238
Appendix
Table A.5 (continued) Average values from 2000 to 2008 (in %)
Telecommunications and utilities Transportation services
Ratio of home-regional to total sales (HRS/TS) Incl. 100%Excl. 100%HR MNCs HR MNCs 92.1 84.0
Ratio of home-regional to total assets (HRA/TA) Incl. 100%Excl. 100%HR MNCs HR MNCs 91.2 79.6
83.7
93.7
74.5
87.8
Source: Own illustration based on the total research sample of this work Table A.6 Number of MNCs within each region Region Number of MNCs Absolute North America 470 Europe 400 Asia-Pacific 399 Rest 310 South America 141 Africa 65
Percentage 70.9% 60.3% 60.2% 46.8% 21.3% 9.8%
Source: Own illustration based on the total research sample of this work Table A.7 Geographic expansion of MNCs by region Region Europe North AsiaAmerica Pacific European MNCs 30% 22% 17% North American MNCs 19% 39% 15% Asian-Pacific MNCs 18% 20% 42% South American MNCs 0% 17% 0% African MNCs – – –
South America 7% 8% 0% 50% –
Africa Rest Total 5% 1% 1% 0% –
18% 18% 20% 33% –
100% 100% 100% 100% –
Source: Own illustration based on the total research sample of this work Table A.8 Administrative heritage of MNCs and preferences for different coordination forms of further regional strategies Home regions of non-home- Directive- Adaptive- Trans-regional Cooperative- No. of regional MNCs from regional regional (world mandate) regional MNCs Europe 3.00 2.52 1.64 3.09 33 North America 2.42 3.33 2.42 3.25 12 Asia-Pacific 3.67 3.11 1.67 3.44 9 South America 1.00 3.00 1.00 5.00 1 Source: Own illustration based on the survey sample of this work Table A.9 Evaluation of the reflective measurement models Latent constructs Regional product/ service adaptation
Indicators
Factor loadings Regional Appearance 0.910 design Packaging 0.876 Innovativeness of 0.815 the design Regional Specifications 0.829 functional Quality standards 0.835 scope 0.835
t-values
Cronbach’s alpha 25.346*** 0.836 17.273*** 15.110***
Internal AVE consistency 0.902 0.753
12.200*** 0.846 7.697*** 9.911***
0.896
0.683
(continued)
Appendix
239
Table A.9 (continued) Latent constructs
Indicators
Regional offering portfolio
Regional brands
Regional orientation
Inter-regional distance
Regional success
Innovativeness of features Number of features Number of products/ services Newness of products/ services Life cycle of products/ services Factors that influence the brand image Brand positioning Brand development Brand name Objective marketrelated regional orientation Objective activitiesrelated regional orientation Subjective marketrelated regional orientation Cultural distance Administrative/ political/ institutional distance Geographic distance Economic distance Regional sales to total sales Regional profits to total profits Regional assets to total assets
Factor t-values loadings
Cronbach’s alpha
Internal AVE consistency
0.645
0.806
0.594
0.935
0.783
0.722
0.466
0.895
0.924
0.753
0.922
0.799
0.806
6.788***
0.505
1.907**
0.905
32.157***
0.841
9.611***
0.932
34.643*** 0.907
0.894
16.784***
0.884
15.312***
0.827 0.759
12.173*** 4.070*** 0.476
0.616
2.857***
0.665
2.104**
0.823 0.845
6.295*** 8.894***
0.909
32.439***
0.890
12.339***
0.941
38.616*** 0.872
0.784
9.605***
0.947
52.202***
Asterisks denote significance levels of one-sided t-test: *indicates significance at the 10% level, **at the 5% level, and ***at the 1% level Source: Own illustration based on the survey sample of this work
240
Appendix
Table A.10 Evaluation of the formative measurement models Latent construct Indicators Weights tvalues Regional Regional strategy 0.521 1.040 management development autonomy Regional market and product/ 0.847 1.447 service development Regional market coordination 0.865 1.486 Regional operational 0.350 0.746 administration
Tolerance VIF
CI
0.605
1.654 12.273
0.453
2.206 12.075
0.430 0.742
2.328 14.679 1.347 14.493
Asterisks denote significance levels of two-sided t-test: * indicates significance at the 10% level, ** at the 5% level, and *** at the 1% level. Source: Own illustration based on the survey sample of this work Table A.11 Evaluation of the basic structural model R2
Model
Latent constructs
Basic structural model
Regional management 0.162 autonomy Regional product/ 0.171 service adaptation Regional design Regional functional scope Regional offering portfolio Regional brands Regional orientation 0.065 Inter-regional distance 0.220 Regional success 0.460 Regional management 0.129 autonomy Regional product/ 0.134 service adaptation Regional design Regional functional scope Regional offering portfolio Regional brands Regional orientation 0.077 Inter-regional distance 0.134 Regional success 0.489 Firm size 0.053 Regional competition 0.009
Basic structural model incl. control variables
Effect size
Path t-values coefficients 0.308
1.893*
0.315
2.190**
0.282 0.295
7.689*** 4.086***
0.205
6.099***
0.438 0.200 0.364
5.736*** 1.675* 3.238***
0.290
2.017**
0.277
1.905*
0.281 0.294
6.727*** 4.235***
0.205
5.703***
0.439 0.213 0.311
6.371*** 1.562 2.366**
0.183 0.073
1.391 0.565
StoneGeissertest
0.292
0.284
Asterisks denote significance levels of two-sided t-test: *indicates significance at the 10% level, **at the 5% level, and ***at the 1% level Source: Own illustration based on the survey sample of this work
Appendix
241
Table A.12 Evaluation of the extended structural model Model
Latent constructs
Extended structural model
Regional management autonomy Regional product/service adaptation Regional design Regional functional scope Regional offering portfolio Regional brands Regional orientation Inter-regional distance Regional orientation regional management autonomy Regional orientation regional product/service adaptation Inter-regional distance regional management autonomy Inter-regional distance regional product/service adaptation Regional success Regional management autonomy Regional product/service adaptation Regional design Regional functional scope Regional offering portfolio Regional brands Regional orientation Inter-regional distance Regional orientation regional management autonomy Regional orientation regional product/service adaptation Inter-regional distance regional management autonomy Inter-regional distance regional product/service adaptation Regional success Firm size Regional competition
Extended structural model incl. control variables
R2
0.403
Path t-values Stonecoefficients Geissertest 0.472 2.972***
0.020
0.100
0.456
0.001 0.399 0.013
0.282 0.295 0.205 0.438 0.017 0.419 0.076
5.309*** 4.060*** 4.813*** 5.252*** 0.122 2.518** 0.586
0.312
0.438
2.533**
0.104
0.268
1.886*
0.268
0.400
2.041**
0.383
0.455
2.188**
0.014
0.084
0.363
0.003 0.300 0.024
0.282 0.295 0.205 0.438 0.041 0.404 0.106
8.528*** 4.112*** 5.329*** 7.628*** 0.175 1.887* 0.492
0.284
0.415
2.136**
0.116
0.284
1.392
0.246
0.390
1.715*
0.016 0.024
0.093 0.103
0.445 0.697
Effect size
0.625
0.414
0.640
0.395
Asterisks denote significance levels of two-sided t-test: *indicates significance at the 10% level, **at the 5% level, and ***at the 1% level Source: Own illustration based on the survey sample of this work
242
Appendix
Table A.13 Observed application of regional strategies in corporate practice Element of regional strategy Regional management autonomy Regional product/ service adaptation Regional orientation Inter-regional distance Preferred coordination mechanism for regional operations
Home-regional strategy Moderate
Host-regional strategy High
Bi-regional strategy High
Tri-regional strategy High
Moderate
High
Moderate
Moderate
Moderate Moderate Cooperativeregional [Directiveregional] [Adaptiveregional]
Low Moderate High Moderate Adaptive-regional Cooperative[Trans-regional regional (world [Adaptivemandate)] regional] [Directiveregional]
Regional know-how transfer/exchange
Intra-regional know-how sharing [Know-how transfer from home to foreign] Staying power [Management continuity]
Know-how transfer Intra-regional from home to know-how foreign [Intrasharing regional know[Know-how how sharing] transfer from home to foreign] Staying power Staying power [Market[Management oriented continuity] investments] [Marketoriented investments]
Regional commitment
Moderate High Directive-regional [Cooperativeregional] [Trans-regional (world mandate)] [Adaptiveregional] Know-how transfer from home to foreign and from foreign to home [Intraregional knowhow sharing] Staying power [Management continuity]
Source: Own illustration based on the survey sample of this work Table A.14 Total research sample Company
Country
Home region
Industry
1
3M
USA
2
Denmark
3
A.P. Moller-Maersk Group ABB
Switzerland
North America Chemicals and pharmaceuticals Europe Transportation services Europe Other manufacturing
4
Abbott Laboratories
USA
5 6
Accenture Acciona
USA Spain
7
ACS
Spain
8 9
Adecco Aegon
Switzerland Netherlands
10 11
AEON AES Corporation
Japan USA
12
Aetna
USA
Industry category Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector
North America Chemicals and pharmaceuticals North America Other services Europe Construction, building materials and glass Europe Construction, building materials and glass Europe Other services Europe Other financial services Asia-Pacific Merchandisers Service sector North America Telecommunications Service sector and utilities North America Other services Service sector
(continued)
Appendix
243
Table A.14 (continued) Company
Country
Home region
13
AFLAC
USA
14
China
Service sector
France
Europe
Service sector
16
Agricultural Bank of China Air France-KLM Group Aisin Seiki
North America Other financial services Asia-Pacific Banks
Japan
Asia-Pacific
17
Akzo Nobel
Netherlands
Europe
18
Alcatel-Lucent
France
Europe
19
Alcoa
USA
North America
20 21
Alfresa Holdings All Nippon Airways
Japan Japan
Asia-Pacific Asia-Pacific
22
Allegheny Energy
USA
North America
23 24
Alliance Boots Allianz
Switzerland Germany
Europe Europe
25
Allstate
USA
North America
26
Alstom
France
Europe
27
Altria Group
USA
North America
28
Aluminum Corp. of China Amazon.com America Movil
China
Asia-Pacific
USA Mexico
North America North America
USA
North America
USA
North America
15
29 30 31 32 33
American Electric Power American Express
Industry
Transportation services Motor vehicle and parts Chemicals and pharmaceuticals Computer, office and electronics Construction, building materials and glass Merchandisers Transportation services Energy, petroleum and refining Merchandisers Other financial services Other financial services Computer, office and electronics Food, drug and tobacco Construction, building materials and glass Other services Telecommunications and utilities Energy, petroleum and refining Other financial services Other financial services
USA
North America
34 35
American International Group AmerisourceBergen AMP
USA Australia
36
AMR
USA
37
Anglo American
38
Anheuser-Busch
United Kingdom USA
39
Apple
USA
North America Merchandisers Asia-Pacific Other financial services North America Transportation services Europe Natural resource manufacturing North America Food, drug and tobacco North America Computer, office and electronics
Industry category Service sector
Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Service sector
Service sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector
(continued)
244
Appendix
Table A.14 (continued) Company
Country
40 41
Arcandor ArcelorMittal
Germany Luxembourg
42
Archer Daniels Midland AREVA
USA
USA
45
Arrow Electronics, Inc. Asahi Glass
Industry category Europe Merchandisers Service sector Europe Construction, building Manufacturing materials and glass sector North America Food, drug and Manufacturing tobacco sector Europe Energy, petroleum and Manufacturing refining sector North America Merchandisers Service sector
Japan
Asia-Pacific
46
Asahi Kasei
Japan
47
Japan
49
Asahi Mutual Life Insurance Assicurazioni Generali Astrazeneca
50
Asustek Computer
United Kingdom Taiwan
51
AT&T
USA
52
Australia and New Zealand Banking AutoNation
Australia
43 44
48
53 54 55
Aviation Industry Corp. of China Aviva
56 57
Avnet Inc. AXA
58
BAE Systems
59
Banco Bilbao Vizcaya Argentaria Banco Bradesco Banco do Brasil Bank of America Corp. Bank of China Bank of Communications Bank of Ireland Group Bank of Montreal Bank of Nova Scotia Baosteel Group
60 61 62 63 64 65 66 67 68
France
Italy
USA
Home region
Industry
Construction, building materials and glass Asia-Pacific Chemicals and pharmaceuticals Asia-Pacific Other financial services Europe Other financial services Europe Chemicals and pharmaceuticals Asia-Pacific Computer, office and electronics North America Telecommunications and utilities Asia-Pacific Banks
Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector
United Kingdom Spain
North America Motor vehicle and Manufacturing parts sector Asia-Pacific Aerospace and defense Manufacturing sector Europe Other financial Service sector services North America Merchandisers Service sector Europe Other financial Service sector services Europe Aerospace and defense Manufacturing sector Europe Banks Service sector
Brazil Brazil USA
South America Banks South America Banks North America Banks
Service sector Service sector Service sector
China China
Asia-Pacific Asia-Pacific
Banks Banks
Service sector Service sector
Ireland
Europe
Banks
Service sector
Canada Canada China
North America Banks Service sector North America Banks Service sector Asia-Pacific Construction, building Manufacturing materials and glass sector
China United Kingdom USA France
(continued)
Appendix
245
Table A.14 (continued) Company
Country
Home region
Industry
69
Barclays
Europe
Banks
70
BASF
United Kingdom Germany
Industry category Service sector
Europe
71
Bayer
Germany
Europe
72
Germany
Europe
Manufacturing sector Manufacturing sector Service sector
73
Bayerische Landesbank BCE
Chemicals and pharmaceuticals Chemicals and pharmaceuticals Banks
Canada
74
Berkshire Hathaway
USA
75
Bertelsmann
Germany
76 77
Best Buy BG Group
78
Bharat Petroleum
USA United Kingdom India
79
BHP Billiton
Australia
80
BMW
Germany
81 82
BNP Paribas Boeing
France USA
North America Telecommunications and utilities North America Other financial services Europe Entertainment, printing and publishing North America Merchandisers Europe Natural resource manufacturing Asia-Pacific Energy, petroleum and refining Asia-Pacific Natural resource manufacturing Europe Motor vehicle and parts Europe Banks North America Aerospace and defense
83
Bombardier
Canada
North America Aerospace and defense
84
Bouygues
France
Europe
85
BP
86
Bridgestone
United Kingdom Japan
87
Bristol-Myers Squibb USA
88
British Airways
89 90
British American Tobacco BT
91
Bunge
92 93
Burlington Northern Santa Fe Cable and Wireless
94
Caltex Australia
United Kingdom United Kingdom United Kingdom USA USA United Kingdom Australia
Construction, building materials and glass Europe Energy, petroleum and refining Asia-Pacific Motor vehicle and parts North America Chemicals and pharmaceuticals Europe Transportation services Europe Food, drug and tobacco Europe Telecommunications and utilities North America Food, drug and tobacco North America Transportation services Europe Telecommunications and utilities Asia-Pacific Energy, petroleum and refining
Service sector Service sector Service sector
Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector
(continued)
246
Appendix
Table A.14 (continued) Company
Country
Home region
95
Canada
North America Banks
Japan
Asia-Pacific
USA USA France Mexico
North America North America Europe North America
USA
North America
Taiwan
Asia-Pacific
USA
North America
104 Cemex
Mexico
North America
105 Centex 106 Central Japan Railway 107 Centrica
USA Japan
North America Asia-Pacific Europe
108 Cepsa
United Kingdom Spain
Europe
109 CFE
Mexico
North America
110 Chevron
USA
North America
111 China Communications Construction 112 China Construction Bank 113 China First Automotive Works 114 China Huaneng Group 115 China Life Insurance
China
Asia-Pacific
China
96
Canadian Imperial Bank of Commerce Canon
97 98 99 100
Capital One Financial Cardinal Health Carrefour Carso Global Telecom 101 Caterpillar 102 Cathay Financial Holdings 103 CBS (formerly Viacom)
116 China Metallurgical Group 117 China Minmetals 118 China Mobile Communications 119 China National Offshore Oil 120 China National Petroleum
Industry
Computer, office and electronics Banks Merchandisers Merchandisers Telecommunications and utilities Other manufacturing
Industry category Service sector
Manufacturing sector Service sector Service sector Service sector Service sector Manufacturing sector Service sector
Other financial services Entertainment, printing and publishing Construction, building materials and glass Other services Transportation services Telecommunications and utilities Energy, petroleum and refining Telecommunications and utilities Energy, petroleum and refining Construction, building materials and glass
Manufacturing sector Manufacturing sector
Asia-Pacific
Banks
Service sector
China
Asia-Pacific
Motor vehicle and parts
Manufacturing sector
China
Asia-Pacific
Service sector
China
Asia-Pacific
China
Asia-Pacific
China
Asia-Pacific
China
Asia-Pacific
China
Asia-Pacific
China
Asia-Pacific
Telecommunications and utilities Other financial services Construction, building materials and glass Construction, building materials and glass Telecommunications and utilities Natural resource manufacturing Energy, petroleum and refining
Service sector
Manufacturing sector Service sector Service sector Service sector Manufacturing sector Service sector
Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector
(continued)
Appendix
247
Table A.14 (continued) Company
Country
Home region
Industry
Asia-Pacific
Construction, building materials and glass Construction, building materials and glass Motor vehicle and parts Telecommunications and utilities Construction, building materials and glass Telecommunications and utilities Telecommunications and utilities Other manufacturing
121 China Railway China Construction 122 China Railway Group China
Asia-Pacific
China
Asia-Pacific
China
Asia-Pacific
China
Asia-Pacific
China
Asia-Pacific
123 China South Industries Group 124 China Southern Power Grid 125 China State Construction 126 China Telecommunications 127 China United Telecommunications 128 Christian Dior
China
Asia-Pacific
France
Europe
129 CHS
USA
130 Chubb
USA
North America Food, drug and tobacco North America Other financial services Asia-Pacific Telecommunications and utilities Europe Other financial services North America Other services North America Computer, office and electronics Asia-Pacific Other financial services North America Banks North America Energy, petroleum and refining Europe Other financial services North America Food, drug and tobacco North America Food, drug and tobacco Asia-Pacific Food, drug and tobacco North America Telecommunications and utilities Europe Banks Asia-Pacific Banks
131 Chubu Electric Power Japan 132 Cie Nationale a Portefeuille 133 Cigna 134 Cisco Systems
Belgium USA USA
135 Citic Group
China
136 Citigroup 137 CMS Energy Corporation 138 CNP Assurances
USA USA France
139 Coca-Cola
USA
140 Coca-Cola Enterprises 141 COFCO
USA China
142 Comcast
USA
143 Commerzbank Germany 144 Commonwealth Bank Australia of Australia 145 Compass Group United Kingdom 146 Computer Sciences USA 147 ConAgra Foods USA 148 ConocoPhillips
USA
Europe
Food, drug and tobacco North America Other services North America Food, drug and tobacco North America Energy, petroleum and refining
Industry category Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector Service sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Service sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector
(continued)
248
Appendix
Table A.14 (continued) Company
Country
Home region
Industry
149 Constellation Energy
USA
150 Continental
Germany
151 Cosmo Oil
Japan
152 Costco Wholesale 153 CPC
USA Taiwan
154 Credit Agricole 155 Credit Industriel & Commercial 156 Credit Suisse 157 CRH
France France
North America Energy, petroleum and refining Europe Motor vehicle and parts Asia-Pacific Energy, petroleum and refining North America Merchandisers Asia-Pacific Energy, petroleum and refining Europe Banks Europe Banks
Switzerland Ireland
Europe Europe
158 CVRD
Brazil
South America
159 CVS Caremark 160 D.R. Horton 161 Dai Nippon Printing
USA USA Japan
North America North America Asia-Pacific
162 Dai-ichi Mutual Life Insurance 163 Daimler
Japan
Asia-Pacific
Germany
Europe
Banks Construction, building materials and glass Natural resource manufacturing Merchandisers Other services Entertainment, printing and publishing Other financial services Motor vehicle and parts Other services Motor vehicle and parts Banks Other manufacturing
164 Daiwa House Industry Japan 165 Dana Corporation USA
Asia-Pacific North America
166 Danske Bank Group 167 Deere
Denmark USA
Europe North America
168 Delhaize Group 169 Dell
Belgium USA
170 Delphi
USA
171 Delta Air Lines
USA
172 DENSO
Japan
173 Dentsu 174 Deutsche Bahn
Japan Germany
175 Deutsche Bank 176 Deutsche Post
Germany Germany
177 Deutsche Telekom
Germany
178 Dexia Group 179 Diageo
Belgium United Kingdom
Europe Merchandisers North America Computer, office and electronics North America Motor vehicle and parts North America Transportation services Asia-Pacific Motor vehicle and parts Asia-Pacific Other services Europe Transportation services Europe Banks Europe Transportation services Europe Telecommunications and utilities Europe Banks Europe Food, drug and tobacco
Industry category Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector Service sector
Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Service sector Service sector Service sector Service sector Manufacturing sector
(continued)
Appendix
249
Table A.14 (continued) Company
Country
Home region
Industry
180 DirecTV Group
USA
181 Dominion Resources
USA
182 Doosan
South Korea
183 Dow Chemical
USA
184 Dresdner Bank 185 Duke Energy
Germany USA
186 DuPont
USA
187 Dynegy Inc.
USA
188 DZ Bank 189 E.ON
Germany Germany
190 EADS
Netherlands
North America Telecommunications and utilities North America Telecommunications and utilities Asia-Pacific Construction, building materials and glass North America Chemicals and pharmaceuticals Europe Banks North America Telecommunications and utilities North America Chemicals and pharmaceuticals North America Energy, petroleum and refining Europe Banks Europe Energy, petroleum and refining Europe Aerospace and defense
191 East Japan Railway
Japan
Asia-Pacific
192 Eastman Kodak
USA
193 Edeka Zentrale 194 Edison International
Germany USA
195 Edison SpA
Italy
196 EDP-Energias de Portugal 197 Eiffage
Portugal France
198 El Paso
USA
199 Electricite de France
France
200 Electrolux
Sweden
201 Eli Lilly
USA
202 Emerson Electric
USA
203 EnCana
Canada
204 Endesa
Spain
205 Enel
Italy
206 Energie BadenW€urttemberg
Germany
Transportation services North America Other manufacturing Europe Merchandisers North America Telecommunications and utilities Europe Energy, petroleum and refining Europe Telecommunications and utilities Europe Construction, building materials and glass North America Energy, petroleum and refining Europe Telecommunications and utilities Europe Computer, office and electronics North America Chemicals and pharmaceuticals North America Computer, office and electronics North America Natural resource manufacturing Europe Telecommunications and utilities Europe Telecommunications and utilities Europe Telecommunications and utilities
Industry category Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Service sector
(continued)
250
Appendix
Table A.14 (continued) Company
Country
Home region
207 ENI
Italy
Europe
208 Enterprise GP Holdings 209 Erste Bank 210 Evonik Industries
USA
North America
Austria Germany
Europe Europe
211 Evraz Group
Russia
Europe
212 Exelon
USA
North America
213 Express Scripts 214 Exxon Mobil
USA USA
North America North America
215 Fannie Mae
USA
North America
216 Faros 217 FedEx
France USA
Europe North America
218 Fiat
Italy
Europe
219 Finmeccanica
Italy
Europe
220 FirstEnergy
USA
North America
221 Flextronics International 222 Fluor
Singapore
Asia-Pacific
USA
North America
Spain
Europe
France USA
Europe North America
223 Fomento de Construcciones 224 Fonciere Euris 225 Ford Motor 226 Formosa Petrochemical 227 Fortis 228 Fortum
Taiwan
Asia-Pacific
Belgium Finland
Europe Europe
229 FPL Group
USA
North America
230 France Telecom
France
Europe
231 Franz Haniel 232 Freddie Mac
Germany USA
Europe North America
233 Freeport-McMoRan Copper & Gold 234 Friends Provident
USA
North America
United Kingdom 235 Fuji Heavy Industries Japan
Europe Asia-Pacific
Industry
Industry category Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Banks Service sector Chemicals and Manufacturing pharmaceuticals sector Construction, building Manufacturing materials and glass sector Telecommunications Service sector and utilities Other services Service sector Energy, petroleum Manufacturing and refining sector Other financial Service sector services Other services Service sector Transportation Service sector services Motor vehicle and Manufacturing parts sector Aerospace and defense Manufacturing sector Telecommunications Service sector and utilities Computer, office and Manufacturing electronics sector Construction, building Manufacturing materials and glass sector Construction, building Manufacturing materials and glass sector Merchandisers Service sector Motor vehicle and Manufacturing parts sector Energy, petroleum Manufacturing and refining sector Banks Service sector Energy, petroleum Manufacturing and refining sector Telecommunications Service sector and utilities Telecommunications Service sector and utilities Merchandisers Service sector Other financial Service sector services Natural resource Manufacturing manufacturing sector Other financial Service sector services Motor vehicle Manufacturing and parts sector
(continued)
Appendix
251
Table A.14 (continued) Company
Country
Home region
236 Fujifilm Holdings
Japan
Asia-Pacific
237 Fujitsu
Japan
Asia-Pacific
238 Galp Energia
Portugal
Europe
239 Gap 240 Gas Natural SDG
USA Spain
North America Europe
241 GasTerra
Netherlands
Europe
242 Gazprom
Russia
Europe
243 General Dynamics
USA
North America
244 General Electric
USA
North America
245 General Motors
USA
North America
246 George Weston 247 Georgia-Pacific
Canada USA
North America North America
248 GlaxoSmithKline
United Kingdom USA
Europe North America
USA
North America
USA
North America
USA USA
North America North America
France
Europe
249 GMAC 250 Goldman Sachs Group 251 Goodyear Tire & Rubber 252 Google 253 Great Atlantic & Pacific Tea 254 Groupama 255 Groupe Auchan 256 Groupe Caisse d’E´pargne 257 Groupe Danone
France France
Europe Europe
France
Europe
258 Grupo Ferrovial
Spain
Europe
259 GS Holdings
South Korea
Asia-Pacific
260 Halliburton 261 Hanwha
USA South Korea
North America Asia-Pacific
262 Hartford Financial Services 263 HCA 264 Health Net
USA
North America
USA USA
North America North America
Industry
Industry category Other manufacturing Manufacturing sector Computer, office Manufacturing and electronics sector Manufacturing Energy, petroleum and refining sector Merchandisers Service sector Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Aerospace and defense Manufacturing sector Other financial Service sector services Motor vehicle and Manufacturing parts sector Merchandisers Service sector Natural resource Manufacturing manufacturing sector Chemicals and Manufacturing pharmaceuticals sector Other financial Service sector services Other financial Service sector services Motor vehicle and Manufacturing parts sector Other services Service sector Merchandisers Service sector Other financial services Merchandisers Banks Food, drug and tobacco Construction, building materials and glass Energy, petroleum and refining Other services Chemicals and pharmaceuticals Other financial services Other services Other services
Service sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Service sector
(continued)
252
Appendix
Table A.14 (continued) Company
Country
Home region
Industry
265 HeidelbergCement
Germany
Europe
266 Heineken Holding
Netherlands
Europe
267 Henkel
Germany
Europe
Construction, building materials and glass Food, drug and tobacco Other manufacturing
268 Heraeus Holding
Germany
Europe
269 Hess
USA
North America
270 Hewlett-Packard
USA
North America
271 Hindustan Petroleum
India
Asia-Pacific
272 Hitachi
Japan
Asia-Pacific
273 Hochtief
Germany
Europe
274 Holcim
Switzerland
Europe
275 Home Depot 276 Hon Hai Precision Industry 277 Honda Motor
USA Taiwan
North America Asia-Pacific
Japan
Asia-Pacific
USA
North America
Construction, building materials and glass Energy, petroleum and refining Computer, office and electronics Energy, petroleum and refining Computer, office and electronics Construction, building materials and glass Construction, building materials and glass Merchandisers Computer, office and electronics Motor vehicle and parts Aerospace and defense
United Kingdom USA Canada
Europe
Banks
278 Honeywell International 279 HSBC Holdings 280 Humana 281 Husky Energy 282 Hutchison Whampoa 283 Hypo Real Estate Holding 284 Hyundai Heavy Industries 285 Hyundai Motor
South Korea
North America Other services North America Natural resource manufacturing Asia-Pacific Merchandisers Europe Other financial services Asia-Pacific Other manufacturing
South Korea
Asia-Pacific
286 Iberdrola
Spain
Europe
287 Idemitsu Kosan
Japan
Asia-Pacific
288 Imperial Tobacco Group 289 Inbev
United Kingdom Belgium
Europe Europe
290 Indian Oil
India
Asia-Pacific
China Germany
291 Industrial & China Commercial Bank of China
Asia-Pacific
Motor vehicle and parts Telecommunications and utilities Energy, petroleum and refining Food, drug and tobacco Food, drug and tobacco Energy, petroleum and refining Banks
Industry category Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector
(continued)
Appendix
253
Table A.14 (continued) Company
Country
Home region
Energy, petroleum and refining Asia-Pacific Motor vehicle and parts South America Banks
Industry category Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector
292 ING Group 293 Ingram Micro 294 Intel
Netherlands USA USA
Europe Banks North America Merchandisers North America Computer, office and electronics North America Computer, office and electronics North America Natural resource manufacturing Europe Banks Europe Other manufacturing
Asia-Pacific Europe
Other services Merchandisers
Service sector Service sector
304 J.C. Penney 305 J.P. Morgan Chase & Co. 306 Japan Airlines
Japan United Kingdom USA USA
North America Merchandisers North America Banks
Service sector Service sector
Japan
Asia-Pacific
Service sector
307 Japan Post
Japan
308 Japan Tobacco
Japan
309 Jardine Matheson 310 JFE Holdings
China Japan
311 Jiangsu Shagang Group 312 Johnson & Johnson
China USA
313 Johnson Controls
USA
314 Kajima
Japan
295 International Business USA Machines 296 International Paper USA 297 Intesa Sanpaolo 298 Invensys 299 Israel Corp.
Italy United Kingdom Israel
300 Isuzu Motors
Japan
301 Itausa-Investimentos Itau 302 Itochu 303 J. Sainsbury
Brazil
Industry
Asia-Pacific
316 Kawasaki Heavy Industries, Ltd. 317 KBC Group 318 KDDI
Japan
Transportation services Asia-Pacific Transportation services Asia-Pacific Food, drug and tobacco Asia-Pacific Merchandisers Asia-Pacific Construction, building materials and glass Asia-Pacific Construction, building materials and glass North America Chemicals and pharmaceuticals North America Motor vehicle and parts Asia-Pacific Construction, building materials and glass Asia-Pacific Telecommunications and utilities Asia-Pacific Other manufacturing
Belgium Japan
Europe Asia-Pacific
Manufacturing sector Service sector Service sector
319 KFW Bankengruppe 320 Kimberly-Clark
Germany USA
321 Kingfisher
United Kingdom
Europe
Service sector Manufacturing sector Service sector
315 Kansai Electric Power Japan
Banks Telecommunications and utilities Europe Banks North America Other manufacturing Merchandisers
Service sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector
(continued)
254
Appendix
Table A.14 (continued) Company
Country
Home region
322 Kintetsu
Japan
Asia-Pacific
Industry category Service sector
323 Kirin Holdings
Japan
324 Kobe Steel
Japan
325 Koc Holding
Turkey
326 Kohl’s 327 Komatsu
USA Japan
Asia-Pacific
Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector
328 Korea Electric Power South Korea 329 Korea Gas
South Korea
330 Kraft Foods
USA
331 Kroger 332 KT
USA South Korea
333 Kyocera
Japan
334 Kyushu Electric Power 335 L.M. Ericsson
Japan Sweden
336 La Poste
France
337 Ladbrokes
United Kingdom
338 Lafarge
France
339 Lagardere Groupe
France
340 L’Air Liquide
France
341 Landesbank BadenW€urttemberg 342 Lear
Germany USA
343 Legal & General Group 344 Lennar 345 Lenovo Group
United Kingdom USA China
346 LG
South Korea
347 LG International 348 Liberty Mutual Insurance Group
South Korea USA
Industry
Transportation services Asia-Pacific Food, drug and tobacco Asia-Pacific Construction, building materials and glass Asia-Pacific Motor vehicle and parts North America Merchandisers Asia-Pacific Other manufacturing Telecommunications and utilities Asia-Pacific Energy, petroleum and refining North America Food, drug and tobacco North America Merchandisers Asia-Pacific Telecommunications and utilities Asia-Pacific Computer, office and electronics Asia-Pacific Telecommunications and utilities Europe Computer, office and electronics Europe Transportation services Europe Entertainment, printing and publishing Europe Construction, building materials and glass Europe Entertainment, printing and publishing Europe Chemicals and pharmaceuticals Europe Banks North America Motor vehicle and parts Europe Other financial services North America Other services Asia-Pacific Computer, office and electronics Asia-Pacific Computer, office and electronics Asia-Pacific Other services North America Other financial services
Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector
Manufacturing sector Service sector
Manufacturing sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector
(continued)
Appendix
255
Table A.14 (continued) Company
Country
Home region
Industry
349 Linde Group
Germany
Europe
350 Lloyds TSB Group
Europe
351 Lockheed Martin
United Kingdom USA
Chemicals and pharmaceuticals Banks
352 Loews
USA
353 L’Oreal
France
354 Lowe’s 355 Lufthansa Group
USA Germany
356 Lukoil
Russia
357 LVMH
France
358 LyondellBasell Industries 359 Macy’s 360 Magna International
Netherlands USA Canada
361 MAN Group
Germany
362 Manpower 363 ManuLife Financial
USA Canada
364 Mapfre Group
Spain
365 Marathon Oil
USA
366 Marks & Spencer
United Kingdom Germany
367 Marquard & Bahls
368 Marriott USA International, Inc. 369 Marsh & McLennan
USA
370 Marubeni 371 Maruhan
Japan Japan
372 Masco
USA
373 Massachusetts Mutual USA Life Insurance 374 Mazda Motor Japan 375 McDonald’s
USA
376 McKesson
USA
Industry category Manufacturing sector Service sector
North America Aerospace and defense Manufacturing sector North America Other financial Service sector services Europe Other manufacturing Manufacturing sector North America Merchandisers Service sector Europe Transportation Service sector services Europe Energy, petroleum Manufacturing and refining sector Europe Other manufacturing Manufacturing sector Europe Chemicals and Manufacturing pharmaceuticals sector North America Merchandisers Service sector North America Motor vehicle and Manufacturing parts sector Europe Motor vehicle and Manufacturing parts sector North America Other services Service sector North America Other financial Service sector services Europe Other financial Service sector services North America Energy, petroleum Manufacturing and refining sector Europe Merchandisers Service sector Europe
Energy, petroleum and refining North America Entertainment, printing and publishing North America Other financial services Asia-Pacific Other services Asia-Pacific Entertainment, printing and publishing North America Other manufacturing North America Other financial services Asia-Pacific Motor vehicle and parts North America Food, drug and tobacco North America Merchandisers
Manufacturing sector Service sector
Service sector Service sector Service sector
Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector
(continued)
256
Appendix
Table A.14 (continued) Company
Country
Home region
Industry
377 Medco Health Solutions 378 Mediceo Paltac Holdings 379 Meiji Yasuda Life Insurance 380 Merck
USA
North America Other services
Industry category Service sector
Japan
Asia-Pacific
Merchandisers
Service sector
Japan
Asia-Pacific
Service sector
USA
North America
381 Merrill Lynch
USA
North America
382 Metalu´rgica Gerdau
Brazil
South America
383 MetLife
USA
North America
384 Metro 385 Michelin
Germany France
Europe Europe
386 Microsoft 387 Migros 388 Mirant Corporation
USA Switzerland USA
389 Mitsubishi 390 Mitsubishi Chemical Holdings 391 Mitsubishi Electric
Japan Japan
Other financial services Chemicals and pharmaceuticals Other financial services Construction, building materials and glass Other financial services Merchandisers Motor vehicle and parts Other services Merchandisers Energy, petroleum and refining Other services Chemicals and pharmaceuticals Computer, office and electronics Other manufacturing
Manufacturing sector Service sector Manufacturing sector Service sector
392 Mitsubishi Heavy Industries 393 Mitsubishi Materials
Japan
394 Mitsubishi Motors
Japan
395 Mitsubishi UFJ Financial Group 396 Mitsui 397 Mitsui Fudosan 398 Mitsui OSK Lines
Japan
Service sector Manufacturing sector North America Service sector Europe Service sector North America Manufacturing sector Asia-Pacific Service sector Asia-Pacific Manufacturing sector Asia-Pacific Manufacturing sector Asia-Pacific Manufacturing sector Asia-Pacific Construction, building Manufacturing materials and glass sector Asia-Pacific Motor vehicle and Manufacturing parts sector Asia-Pacific Banks Service sector
Japan Japan Japan
Asia-Pacific Asia-Pacific Asia-Pacific
Japan
Asia-Pacific
Japan
Asia-Pacific
Hungary
Europe
Japan Japan
399 Mitsui Sumitomo Insurance 400 Mizuho Financial Group 401 Mol Hungarian Oil & Gas 402 Morgan Stanley
USA
403 Motorola
USA
404 Munich Re Group
Germany
405 Murphy Oil
USA
Other services Other services Transportation services Other financial services Banks
Energy, petroleum and refining North America Other financial services North America Computer, office and electronics Europe Other financial services North America Energy, petroleum and refining
Service sector Service sector Service sector Service sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector
(continued)
Appendix
257
Table A.14 (continued) Company
Country
Home region
Industry
406 N.V. Nederlandse Gasunie 407 National Australia Bank 408 National Grid
Netherlands
Europe
Australia
Asia-Pacific
Energy, petroleum and refining Banks
Europe
409 Nationwide
United Kingdom USA
410 Nationwide Building Society 411 Nec
United Kingdom Japan
412 Neste Oil
Finland
413 Nestle
Switzerland
414 New York Life Insurance 415 News Corp.
USA USA
416 Nike
USA
Computer, office and electronics Europe Energy, petroleum and refining Europe Food, drug and tobacco North America Other financial services North America Entertainment, printing and publishing North America Other manufacturing
417 Nippon Express
Japan
Asia-Pacific
418 Nippon Life Insurance 419 Nippon Mining Holdings 420 Nippon Oil
Japan
Asia-Pacific
Japan
Asia-Pacific
Japan
Asia-Pacific
421 Nippon Steel
Japan
Asia-Pacific
422 Nippon Telegraph & Telephone 423 Nippon Yusen
Japan
Asia-Pacific
Japan
Asia-Pacific
424 Nissan Motor
Japan
Asia-Pacific
425 Noble Group 426 Nokia
China Finland
Asia-Pacific Europe
427 Nomura Holdings
Japan
Asia-Pacific
428 Norddeutsche Landesbank Girozentrale 429 Nordea Bank 430 Norinchukin Bank 431 Norsk Hydro
Germany
Europe
Sweden Japan Norway
Europe Asia-Pacific Europe
Telecommunications and utilities North America Other financial services Europe Banks Asia-Pacific
Transportation services Other financial services Energy, petroleum and refining Energy, petroleum and refining Construction, building materials and glass Telecommunications and utilities Transportation services Motor vehicle and parts Other services Computer, office and electronics Other financial services Banks
Industry category Manufacturing sector Service sector Service sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector
Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector
Banks Service sector Banks Service sector Construction, building Manufacturing materials and glass sector
(continued)
258
Appendix
Table A.14 (continued) Company
Country
Home region
432 Nortel Networks Corporation 433 Northrop Grumman
Canada
North America
USA
North America
434 Northwestern Mutual
USA
North America
435 Novartis
Switzerland
Europe
436 Nucor
USA
North America
437 Obayashi
Japan
Asia-Pacific
438 Occidental Petroleum USA
North America
439 Office Depot 440 OfficeMax 441 Oil & Natural Gas
USA USA India
North America North America Asia-Pacific
442 Oji Paper
Japan
Asia-Pacific
443 Old Mutual
Europe
444 OMV Group
United Kingdom Austria
Europe
445 Onex
Canada
North America
446 Oracle 447 Otto Group 448 PACCAR
USA Germany USA
North America Europe North America
449 Panasonic
Japan
Asia-Pacific
450 PDVSA
Venezuela
South America
451 Pemex
Mexico
North America
452 PepsiCo
USA
North America
453 Petrobras
Brazil
South America
454 Petro-Canada
Canada
North America
455 Petronas
Malaysia
Asia-Pacific
456 Petroplus Holdings
Switzerland
Europe
457 Peugeot
France
Europe
458 Pfizer
USA
North America
459 PG & E Corp.
USA
North America
Industry
Industry category Computer, office Manufacturing and electronics sector Aerospace and defense Manufacturing sector Service sector Other financial services Chemicals and Manufacturing pharmaceuticals sector Construction, building Manufacturing materials and glass sector Construction, building Manufacturing materials and glass sector Natural resource Manufacturing manufacturing sector Merchandisers Service sector Merchandisers Service sector Natural resource Manufacturing manufacturing sector Natural resource Manufacturing manufacturing sector Other financial Service sector services Energy, petroleum and Manufacturing refining sector Other financial Service sector services Other services Service sector Merchandisers Service sector Motor vehicle and Manufacturing parts sector Computer, office Manufacturing and electronics sector Energy, petroleum Manufacturing and refining sector Natural resource Manufacturing manufacturing sector Food, drug and Manufacturing tobacco sector Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Motor vehicle and Manufacturing parts sector Chemicals and Manufacturing pharmaceuticals sector Telecommunications Service sector and utilities
(continued)
Appendix
259
Table A.14 (continued) Company
Country
Home region
460 PKN Orlen Group
Poland
Europe
461 Plains All American Pipeline 462 POSCO
USA
North America
South Korea
Asia-Pacific
463 Poste Italiane
Italy
Europe
464 Power Corp. of Canada 465 PPR 466 Premafin Finanziaria
Canada
North America
France Italy
Europe Europe
467 Procter & Gamble
USA
North America
468 Progressive
USA
North America
469 Prudential
Europe
470 Prudential Financial
United Kingdom USA
North America
471 PTT
Thailand
Asia-Pacific
472 Public Service Enterprise Group 473 Publix Super Markets 474 Pulte Homes 475 Quanta Computer
USA
North America
USA USA Taiwan
North America North America Asia-Pacific
476 Qwest Communications 477 Rabobank 478 RAG
USA
North America
Netherlands Germany
Europe Europe
479 Rallye 480 Randstad Holding 481 Raytheon
France Netherlands USA
Europe Europe North America
482 Reliance Industries
India
Asia-Pacific
483 Reliant Energy
USA
North America
484 Renault
France
Europe
485 Repsol YPF
Spain
Europe
486 Resona Holdings 487 Rexel 488 Ricoh
Japan France Japan
Asia-Pacific Europe Asia-Pacific
489 Rio Tinto Group
United Kingdom
Europe
Industry
Industry category Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Construction, building Manufacturing materials and glass sector Transportation Service sector services Other financial Service sector services Merchandisers Service sector Other financial Service sector services Other manufacturing Manufacturing sector Other financial Service sector services Other financial Service sector services Other financial Service sector services Energy, petroleum Manufacturing and refining sector Telecommunications Service sector and utilities Merchandisers Service sector Other services Service sector Computer, office and Manufacturing electronics sector Telecommunications Service sector and utilities Banks Service sector Natural resource Manufacturing manufacturing sector Merchandisers Service sector Other services Service sector Aerospace and defense Manufacturing sector Energy, petroleum Manufacturing and refining sector Energy, petroleum Manufacturing and refining sector Motor vehicle and Manufacturing parts sector Energy, petroleum Manufacturing and refining sector Banks Service sector Merchandisers Service sector Computer, office Manufacturing and electronics sector Natural resource Manufacturing manufacturing sector
(continued)
260
Appendix
Table A.14 (continued) Company
Country
Home region
490 Rite Aid 491 Robert Bosch
USA Germany
492 Roche Group
Switzerland
493 Rosneft Oil
Russia
North America Merchandisers Europe Motor vehicle and parts Europe Chemicals and pharmaceuticals Europe Energy, petroleum and refining Europe Other financial services Europe Merchandisers North America Banks
Service sector Service sector
Europe
Banks
Service sector Manufacturing sector Service sector
494 Royal & Sun Alliance United Kingdom 495 Royal Ahold Netherlands 496 Royal Bank of Canada Canada 497 Royal Bank of United Scotland Kingdom 498 Royal Dutch Shell Netherlands
Industry
511 Santander Central Hispano Group 512 Sanyo Electric
Spain
Energy, petroleum and refining Europe Telecommunications and utilities Europe Transportation services Europe Computer, office and electronics Europe Energy, petroleum and refining Asia-Pacific Chemicals and pharmaceuticals Europe Food, drug and tobacco North America Merchandisers Europe Construction, building materials and glass Asia-Pacific Other services Asia-Pacific Computer, office and electronics Asia-Pacific Other financial services Europe Chemicals and pharmaceuticals Europe Banks
Japan
Asia-Pacific
513 Sara Lee
USA
514 Sberbank 515 Schering-Plough Corporation 516 Schlumberger 517 Schneider Electric
Russia USA USA France
518 Scottish & Southern Energy
United Kingdom
499 Royal KPN
Netherlands
500 Royal Mail Holdings
United Kingdom Netherlands
501 Royal Philips Electronics 502 RWE Aktiengesellschaft 503 Sabic 504 SABMiller
Germany Saudi Arabia
505 Safeway, Inc. 506 Saint-Gobain
United Kingdom USA France
507 Samsung 508 Samsung Electronics
South Korea South Korea
509 Samsung Life Insurance 510 Sanofi-Aventis
South Korea France
Europe
Computer, office and electronics North America Food, drug and tobacco Europe Banks North America Chemicals and pharmaceuticals North America Other services Europe Computer, office and electronics Europe Telecommunications and utilities
Industry category Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector
Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector
(continued)
Appendix
261
Table A.14 (continued) Company
Country
Home region
519 Sears Holdings 520 Seiko Epson
USA Japan
521 Sekisui House 522 Seven & I Holdings 523 SeverStal
Japan Japan Russia
Industry
525 Sharp
Japan
526 Shimizu
Japan
527 Shinhan Financial Group 528 Showa Shell Sekiyu K.K. 529 SHV Holdings
South Korea
North America Merchandisers Asia-Pacific Computer, office and electronics Asia-Pacific Other services Asia-Pacific Merchandisers Europe Construction, building materials and glass Asia-Pacific Motor vehicle and parts Asia-Pacific Computer, office and electronics Asia-Pacific Construction, building materials and glass Asia-Pacific Banks
Japan
Asia-Pacific
Netherlands
Europe
530 Siemens
Germany
Europe
531 Sinochem 532 Sinopec
China China
Asia-Pacific Asia-Pacific
533 Sinosteel
China
Asia-Pacific
534 SK Holdings
South Korea
Asia-Pacific
535 Skandinaviska Enskilda Banken 536 Skanska
Sweden
Europe
Sweden
Europe
537 SNCF
France
538 Snow Brand Milk Products 539 Societe Generale 540 Sodexo
Japan France France
541 Softbank
Japan
542 S-Oil
South Korea
543 Sompo Japan Insurance 544 Sony
Japan Japan
545 Southern
USA
546 Sprint Nextel
USA
524 Shanghai Automotive China
Industry category Service sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector
Energy, petroleum and refining Energy, petroleum and refining Computer, office and electronics Other services Energy, petroleum and refining Construction, building materials and glass Energy, petroleum and refining Banks
Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector
Construction, building materials and glass Europe Transportation services Asia-Pacific Food, drug and tobacco Europe Banks Europe Food, drug and tobacco Asia-Pacific Telecommunications and utilities Asia-Pacific Energy, petroleum and refining Asia-Pacific Other financial services Asia-Pacific Computer, office and electronics North America Telecommunications and utilities North America Telecommunications and utilities
Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector
(continued)
262
Appendix
Table A.14 (continued) Company
Country
Home region
Industry
547 Standard Chartered Bank 548 Standard Life
United Kingdom United Kingdom USA India USA
Europe
Banks
Europe
Other financial services North America Merchandisers Asia-Pacific Banks North America Other financial services Asia-Pacific Telecommunications and utilities Europe Energy, petroleum and refining Europe Natural resource manufacturing Europe Energy, petroleum and refining Asia-Pacific Other services Asia-Pacific Chemicals and pharmaceuticals Asia-Pacific Computer, office and electronics Asia-Pacific Other financial services Asia-Pacific Construction, building materials and glass Asia-Pacific Banks
549 Staples 550 State Bank of India 551 State Farm Insurance Cos. 552 State Grid
China
553 Statoil Hydro
Norway
554 Stora Enso
Finland
555 Suez
France
556 Sumitomo 557 Sumitomo Chemical
Japan Japan
558 Sumitomo Electric Industries 559 Sumitomo Life Insurance 560 Sumitomo Metal Industries Ltd. 561 Sumitomo Mitsui Financial Group 562 Sun Life Financial
Japan
Canada
563 Sun Microsystems
USA
564 Suncor Energy
Canada
565 Sunoco
USA
566 Suntory Limited
Japan
567 Supervalu 568 Surgutneftegas
USA Russia
569 Suzuki Motor
Japan
570 Svenska Cellulosa
Sweden
North America Other financial services North America Computer, office and electronics North America Energy, petroleum and refining North America Energy, petroleum and refining Asia-Pacific Food, drug and tobacco North America Merchandisers Europe Natural resource manufacturing Asia-Pacific Motor vehicle and parts Europe Other manufacturing
571 Swiss Life
Switzerland
Europe
572 Swiss Reinsurance
Switzerland
573 Swisscom
Switzerland
574 Sysco
USA
Japan Japan Japan
Other financial services Europe Other financial services Europe Telecommunications and utilities North America Merchandisers
Industry category Service sector Service sector Service sector Service sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Service sector Service sector
(continued)
Appendix
263
Table A.14 (continued) Company
Country
Home region
575 T&D Holdings
Japan
Asia-Pacific
576 Taisei
Japan
577 Takashimaya 578 Takenaka Corporation 579 Target 580 Tata Steel
Japan Japan USA India
581 Tech Data 582 Telecom Italia
USA Italy
583 Telefonica
Spain
584 Telstra
Australia
585 Tenet Healthcare 586 Tesco 587 Tesoro
USA United Kingdom USA
588 Texas Instruments
USA
589 Textron, Inc.
USA
590 Thales Group
France
591 ThyssenKrupp
Germany
592 TIAA-CREF
USA
593 Time Warner
USA
594 TJX 595 TNK-BP Holding
USA Russia
596 TNT
Netherlands
597 Tohoku Electric Japan Power 598 Tokyo Electric Power Japan 599 Tokyu 600 Toppan Printing
Japan Japan
601 Toronto-Dominion Bank
Canada
Industry
Other financial services Asia-Pacific Construction, building materials and glass Asia-Pacific Merchandisers Asia-Pacific Construction, building materials and glass North America Merchandisers Asia-Pacific Construction, building materials and glass North America Merchandisers Europe Telecommunications and utilities Europe Telecommunications and utilities Asia-Pacific Telecommunications and utilities North America Other services Europe Merchandisers
Industry category Service sector Manufacturing sector Service sector Manufacturing sector Service sector Manufacturing sector Service sector Service sector Service sector Service sector Service sector Service sector
North America Energy, petroleum and Manufacturing refining sector North America Computer, office and Manufacturing electronics sector North America Aerospace and defense Manufacturing sector Europe Aerospace and defense Manufacturing sector Europe Construction, building Manufacturing materials and glass sector North America Other financial Service sector services North America Entertainment, Service sector printing and publishing North America Merchandisers Service sector Europe Energy, petroleum Manufacturing and refining sector Europe Transportation Service sector services Asia-Pacific Telecommunications Service sector and utilities Asia-Pacific Telecommunications Service sector and utilities Asia-Pacific Merchandisers Service sector Asia-Pacific Entertainment, Service sector printing and publishing North America Banks Service sector
(continued)
264
Appendix
Table A.14 (continued) Company
Country
Home region
Industry
602 Toshiba
Japan
Asia-Pacific
603 Total
France
Europe
604 Toyota Industries
Japan
Asia-Pacific
605 Toyota Motor
Japan
Asia-Pacific
606 Toyota Tsusho Corporation 607 Toys ‘R’ Us 608 TransCanada Pipelines Limited 609 Travelers Cos.
Japan
Asia-Pacific
Computer, office and electronics Energy, petroleum and refining Motor vehicle and parts Motor vehicle and parts Other services
USA Canada USA
610 TUI
Germany
611 Tyco International
USA
612 Tyson Foods
USA
613 U.S. Bancorp 614 U.S. Postal Service
USA USA
615 UAL
USA
616 UBS 617 UES of Russia
Switzerland Russia
618 UniCredit Group 619 Unilever
Italy United Kingdom USA
North America Merchandisers North America Energy, petroleum and refining North America Other financial services Europe Entertainment, printing and publishing North America Computer, office and electronics North America Food, drug and tobacco North America Banks North America Transportation services North America Transportation services Europe Banks Europe Telecommunications and utilities Europe Banks Europe Food, drug and tobacco North America Transportation services North America Transportation services North America Construction, building materials and glass North America Aerospace and defense
620 Union Pacific
621 United Parcel Service USA 622 United States Steel
USA
623 United Technologies
USA
624 UnitedHealth Group 625 UNY 626 UPM-Kymmene
USA Japan Finland
627 Valero Energy
USA
628 Vattenfall
Sweden
629 Veolia Environnement
France
North America Other services Asia-Pacific Merchandisers Europe Natural resource manufacturing North America Energy, petroleum and refining Europe Telecommunications and utilities Europe Telecommunications and utilities
Industry category Manufacturing sector Manufacturing sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Service sector Service sector
Manufacturing sector Manufacturing sector Service sector Service sector Service sector Service sector Service sector Service sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector
(continued)
Appendix
265
Table A.14 (continued) Company
Country
Home region
630 Verizon Communications 631 Vinci
USA France
632 Visteon
USA
633 Vivendi
France
634 Vodafone 635 Volkswagen
United Kingdom Germany
636 Volvo
Sweden
637 Walgreen 638 Wal-Mart Stores 639 Walt Disney
USA USA USA
640 Waste Management
USA
641 642 643 644
USA USA Australia Japan
North America Telecommunications and utilities Europe Construction, building materials and glass North America Motor vehicle and parts Europe Telecommunications and utilities Europe Telecommunications and utilities Europe Motor vehicle and parts Europe Motor vehicle and parts North America Merchandisers North America Merchandisers North America Entertainment, printing and publishing North America Telecommunications and utilities North America Other services North America Banks Asia-Pacific Merchandisers Asia-Pacific Transportation services Europe Banks Asia-Pacific Banks North America Natural resource manufacturing North America Computer, office and electronics Europe Entertainment, printing and publishing Europe Merchandisers
Wellpoint Wells Fargo Wesfarmers West Japan Railway
645 WestLB 646 Westpac Banking 647 Weyerhaeuser
Germany Australia USA
648 Whirlpool
USA
649 William Hill
United Kingdom
650 William Morrison Supermarkets 651 Wilmar International
United Kingdom Singapore
652 Winn-Dixie Stores 653 Wolseley
USA United Kingdom Australia South Korea
654 Woolworths 655 Woori Finance Holdings 656 Wyeth
USA
657 XCEL Energy Inc.
USA
658 Xerox
USA
Industry
Industry category Service sector Manufacturing sector Manufacturing sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector Service sector Service sector
Service sector Service sector Service sector Service sector Service sector Service sector Service sector Manufacturing sector Manufacturing sector Service sector
Service sector
Asia-Pacific
Food, drug and tobacco North America Merchandisers Europe Construction, building materials and glass Asia-Pacific Merchandisers Asia-Pacific Banks
Manufacturing sector Service sector Manufacturing sector Service sector Service sector
North America Chemicals and pharmaceuticals North America Telecommunications and utilities North America Computer, office and electronics
Manufacturing sector Service sector Manufacturing sector
(continued)
266
Appendix
Table A.14 (continued) Company
Country
Home region
Industry
659 Xstrata
Switzerland
Europe
660 Yamada Denki 661 Yamaha Motor
Japan Japan
Asia-Pacific Asia-Pacific
662 ZF Friedrichshafen
Germany
Europe
663 Zurich Financial Services
Switzerland
Europe
Natural resource manufacturing Merchandisers Motor vehicle and parts Motor vehicle and parts Other financial services
Industry category Manufacturing sector Service sector Manufacturing sector Manufacturing sector Service sector
Source: Own illustration on the basis of the Fortune Global 500 firms from 2000 to 2008 according to the magazine Fortune (2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009)
References
Agarwal, J., Malhotra, N. K., & Wu, T. (2002). Does NAFTA influence Mexico’s product image? A theoretical framework and an empirical investigation in two countries. Management International Review, 42(4), 441–471. Albaum, G., & Tse, D. K. (2001). Adaptation of international marketing strategy components, competitive advantage, and firm performance: A study of Hong Kong exporters. Journal of International Marketing, 9(4), 59–81. Albers, S., & G€otz, O. (2006). Messmodelle mit Konstrukten zweiter Ordnung in der betriebswirtschaftlichen Forschung. Die Betriebswirtschaft, 66(6), 669–677. Albers, S., & Hildebrandt, L. (2006). Methodische Probleme bei der Erfolgsfaktoren-forschung – Messfehler, formative versus reflektive Indikatoren und die Wahl des StrukturgleichungsModells. Zeitschrift f€ ur betriebswirtschaftliche Forschung, 58(2), 2–33. Alden, D. L., Steenkamp, J.-B. E. M., & Batra, R. (1999). Brand positioning through advertising in Asia, North America, and Europe: The role of global consumer culture. Journal of Marketing, 63(1), 75–87. Anderson, J. C., & Gerbing, D. W. (1991). Predicting the performance of measures in a confirmatory factor analysis with a pretest assessment of their substantive validities. Journal of Applied Psychology, 76(5), 732–740. Arregle, J.-L., Beamish, P. W., & He´bert, L. (2009). The regional dimension of MNEs’ foreign subsidiary localization. Journal of International Business Studies, 40(1), 86–107. Asmussen, C. G. (2009). Local, regional, or global? Quantifying MNE geographic scope. Journal of International Business Studies, 40(7), 1192–1205. Asmussen, C. G., Pedersen, T., & Dhanaraj, C. (2009). Host-country environment and subsidiary competence: Extending the diamond network model. Journal of International Business Studies, 40(1), 42–57. Athey, S., & Schmutzler, A. (1995). Product and process flexibility in an innovative environment. RAND Journal of Economics, 26(4), 557–574. Backhaus, K., Blechschmidt, B., & Eisenbeiß, M. (2006). Der Stichprobeneinfluss bei Kausalanalysen. Die Betriebswirtschaft, 66(6), 711–726. Backhaus, K., & Ebers, M. (2006). Die Anwendung der Kausalanalyse in der betriebswirtschaftlichen Forschung. Die Betriebswirtschaft, 66(6), 607–608. Backhaus, K., Erichson, B., Plinke, W., & Weiber, R. (2008). Multivariate Analyse-methoden: Eine anwendungsorientierte Einf€ uhrung. Berlin/Heidelberg: Springer. Baden-Fuller, C. W. F., & Stopford, J. M. (1991). Globalization frustrated: The case of white goods. Strategic Management Journal, 12(7), 493–507. Bagozzi, R. P. (1977). Structural equation models in experimental research. Journal of Marketing Research, 14(2), 209–226.
267
268
References
Bagozzi, R. P. (1998). A prospectus for theory construction in marketing: Revisited and revised. In L. Hildebrandt & C. Homburg (Eds.), Die Kausalanalyse: Ein Instrument der empirischen betriebswirtschaftlichen Forschung (pp. 45–81). Stuttgart: Sch€affer-Poeschel Verlag. Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74–94. Balderjahn, I. (1986). Das umweltbewusste Konsumentenverhalten: Eine empirische Studie. Berlin: Duncker & Humblot. Banalieva, E. R., & Santoro, M. D. (2009). Local, regional, or global? Geographic orientation and relative financial performance of emerging market multinational enterprises. European Management Journal, 27(5), 344–355. Barkema, H. G., Bell, J., & Pennings, J. (1996). Foreign entry, cultural barriers and learning. Strategic Management Journal, 17(2), 151–166. Barkema, H. G., & Drogendijk, R. (2007). Internationalising in small, incremental or larger steps? Journal of International Business Studies, 38(7), 1132–1148. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182. Bartlett, C. A. (1983). MNCs: Get off the reorganization merry-go-round. Harvard Business Review, 61(2), 138–146. Bartlett, C. A. (1986). Building and managing the transnational: The new organizational challenge. In M. E. Porter (Ed.), Competition in global industries (pp. 367–401). Boston, MA: Harvard Business School Press. Bartlett, C. A., & Ghoshal, S. (1987a). Managing across borders: New organizational responses. Sloan Management Review, 29(1), 43–53. Bartlett, C. A., & Ghoshal, S. (1987b). Managing across borders: New strategic requirements. Sloan Management Review, 28(4), 7–17. Bartlett, C. A., & Ghoshal, S. (1988). Organizing for worldwide effectiveness: The transnational solution. California Management Review, 31(1), 54–74. Bartlett, C. A., & Ghoshal, S. (1989). Managing across borders: The transnational solution. Boston, MA: Harvard Business School Press. Bartlett, C. A., & Ghoshal, S. (2002). The transnational and beyond: Reflections and perspectives at the millennium. In M. A. Hitt & J. L. C. Cheng (Eds.), Managing transnational firms: Resources, market entry and strategic alliances (Advances in international management, Vol. 14, pp. 3–36). Oxford: Elsevier, JAI Press. Bartlett, C. A., Ghoshal, S., & Birkinshaw, J. (2003). Transnational management: Text, cases, and readings in cross-border management. Boston, MA: McGraw-Hill/Irwin. Bassen, A., Behnam, M., & Gilbert, D. U. (2001). Internationalisierung des Mittelstands. Ergebnisse einer Studie zum Internationalisierungsverhalten deutscher mittelst€andischer Unternehmen. Zeitschrift f€ ur Betriebswirtschaft, 71(4), 413–432. Batra, R., Ramaswamy, V., Alden, D. L., Steenkamp, J.-B. E. M., & Ramachander, S. (2000). Effects of brand local and nonlocal origin on consumer attitudes in developing countries. Journal of Consumer Psychology, 9(2), 83–95. Bauer, H. H., & Sauer, N. E. (2004). Die Erfolgsfaktorenforschung als schwarzes Loch? Die Betriebswirtschaft, 64(5), 621–623. Baumgartner, H., & Homburg, C. (1996). Applications of structural equation modeling in marketing and consumer research: A review. International Journal of Research in Marketing, 13(2), 139–161. Bausch, A., Fritz, T., & Boesecke, K. (2007). Performance effects of internationalization strategies: A meta-analysis. In A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 143–176). Amsterdam: Elsevier. Bausch, A., & Krist, M. (2007). The effect of context-related moderators on the internationalization-performance relationship: Evidence from meta-analysis. Management International Review, 47(3), 319–347.
References
269
Beath, J., & Katsoulacos, Y. (1991). The economic theory of product differentiation. New York, NY: Cambridge University Press. Behn, B. K., Nichols, N. B., & Street, D. L. (2002). The predictive ability of geographic segment disclosures by U.S. companies: SFAS no. 131 vs. SFAS no. 14. Journal of International Accounting Research, 1(31), 31–44. Behnam, M., & Gilbert, D. U. (2001). Internationalization strategies of medium-sized German companies: An empirical survey. Journal of International Business and Economy, 2(Fall), 95–109. Beleska-Spasova, E., & Glaister, K. W. (2009). The geography of British exports: Country-level versus firm-level evidence. European Management Journal, 27(5), 295–304. Belsey, D. A., Kuh, E., & Welsch, R. E. (1980). Regression diagnostics: Identifying influential data and sources of collinearity. New York, NY: John Wiley and Sons. Belsey, D. A., Kuh, E., & Welsch, R. E. (2004). Regression diagnostics: Identifying influential data and sources of collinearity. New York, NY: John Wiley and Sons. Benkenstein, M., & Stenglin, A. V. (2006). Innovationsmanagement im Service-Marketing: Neue Gesch€afte f€ur den Service erschließen. In: H.-J. Bullinger & A.-W. Scheer (Eds.), Service Engineering: Entwicklung und Gestaltung innovativer Dienstleistungen (pp. 271–295). Berlin/ Heidelberg: Springer. Bennis, W. G., & O’Toole, J. (2005). How business schools lost their way. Harvard Business Review, 83(5), 96–104. Berger, R., & Steger, U. (1998). Auf dem langen Weg zur Europ€aischen Unternehmensf€uhrung. In: R. Berger & U. Steger (Eds.), Auf dem Weg zur Europ€ aischen Unternehmensf€ uhrung: Ein Lesebuch f€ ur Manager und Europ€ aer (pp. 11–38). M€ unchen: Beck. Bergh, D. D., Perry, J., & Hanke, R. (2006). Some predictors of SMJ article impact. Strategic Management Journal, 27(1), 81–100. Bettis, R. A. (1991). Strategic management and the straightjacket: An editorial essay. Organization Science, 2(3), 315–319. Betzin, J., & Henseler, J. (2005). Einf€ uhrung in die Funktionsweise des PLS-Algorithmus. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 49–69). Stuttgart: Sch€affer-Poeschel Verlag. Birkinshaw, J. (1997). Entrepreneurship in multinational corporations: The characteristics of subsidiary initiatives. Strategic Management Journal, 18(3), 207–229. Birkinshaw, J. (1998). Corporate entrepreneurship in network organizations: How subsidiary initiative drives internal market efficiency. European Management Journal, 16(3), 355–364. Birkinshaw, J. (2000). Entrepreneurship in the global firm. London: Sage. Birkinshaw, J., Braunerhjelm, P., Holm, U., & Terjesen, S. (2006). Why do some multinational corporations relocate their headquarters overseas? Strategic Management Journal, 27(7), 681–700. Birkinshaw, J. M., & Morrison, A. J. (1995). Configurations of strategy and structure in subsidiaries of multinational corporations. Journal of International Business Studies, 26(4), 729–754. Birkinshaw, J. M., & Pedersen, T. (2009). Strategy and management in MNE subsidiaries. In: A. M. Rugman (Ed.), The Oxford handbook of international business (pp. 367–388). Oxford: Oxford University Press. Bliemel, F., Eggert, A., Fassott, G., & Henseler, J. (2005). Die PLS-Pfadmodellierung: Mehr als eine Alternative zur Kovarianzstrukturanalyse. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 9–16). Stuttgart: Sch€affer-Poeschel Verlag. B€ ohmann, T., & Krcmar, H. (2004). Servicedatenmanagement f€ur modulare Dienstleistungen. In: H. Luczak (Ed.), Betriebliche Tertiarisierung: Der ganzheitliche Wandel vom Produktionsbetrieb zum dienstleistenden Probleml€ oser (pp. 149–175). Wiesbaden: DUV. Bohrnstedt, G. W. (1970). Reliability and validity assessment in attitude measurement. In: G. F. Summers (Ed.), Attitude measurement (pp. 80–99). London: Wiley.
270
References
Bollen, K. A. (1989). Structural equations with latent variables. New York, NY: John Wiley and Sons. Bollen, K. A., & Lennox, R. (1991). Conventional wisdom on measurement: A structural equation perspective. Psychological Bulletin, 110(2), 305–314. Bollen, K. A., & Ting, K. (2000). A tetrad test for causal indicators. Psychological Methods, 5(1), 3–22. Bortz, J. (1993). Statistik f€ ur Sozialwissenschaftler. Berlin/Heidelberg: Springer. Bouquet, C., Morrison, A. J., & Birkinshaw, J. (2009). International attention and multinational enterprise performance. Journal of International Business Studies, 40(1), 108–131. Bowen, H. P. (2007). The empirics of multinationality and performance. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 113–142). Amsterdam: Elsevier. Bruhn, M., Georgi, D., & Hadwich, K. (2008). Customer equity management as formative secondorder construct. Journal of Business Research, 61(12), 1292–1301. Buckley, P. J. (2002). Is the international business research agenda running out of steam? Journal of International Business Studies, 33(2), 365–373. Buckley, P. J., Clegg, J., Forsans, N., & Reilly, K. T. (2001). Increasing the size of the “country”: Regional economic integration and foreign direct investment in a globalised world economy. Management International Review, 41(3), 251–274. Buckley, P. J., Devinney, T. M., & Louviere, J. J. (2007). Do managers behave the way theory suggests? A choice-theoretic examination of foreign direct investment location decisionmaking. Journal of International Business Studies, 38(7), 1069–1094. Buckley, P. J., & Ghauri, P. N. (2004). Globalisation, economic geography and the strategy of multinational enterprises. Journal of International Business Studies, 35(2), 81–98. Burns, T., & Stalker, G. M. (1961). The management of innovation. London: Tavistock Publications. Calantone, R. J., Cavusgil, S. T., Schmidt, J. B., & Shin, G.-C. (2004). Internationalization and the dynamics of product adaptation – An empirical investigation. Journal of Product Innovation Management, 21(3), 185–198. Calantone, R. J., Kim, D., Schmidt, J. B., & Cavusgil, S. T. (2006). The influence of internal and external firm factors on international product adaptation strategy and export performance: A three-country comparison. Journal of Business Research, 59(2), 176–185. Carmines, E. G., & Zeller, R. A. (1979). Reliability and validity assessment. Beverly Hills, CA: Sage. Carnap, R. (1966). Philosophical foundations of physics: An introduction to the philosophy of science. New York, NY: Basic Books. Cassel, C., Hackl, P., & Westlund, A. H. (1999). Robustness of partial least-squares method for estimating latent variable quality structures. Journal of Applied Statistics, 26(4), 435–446. Caves, R. E. (1971). International corporations: The industrial economics of foreign investment. Economica, 38(149), 1–27. Caves, R. E. (1996). Multinational enterprise and economic analysis. Cambridge: Cambridge University Press. Cavusgil, S. T., & Kirpalani, V. H. (1993). Introducing products into export markets: Success factors. Journal of Business Research, 27(1), 1–15. Cavusgil, S. T., Zou, S., & Naidu, G. M. (1993). Product and promotion adaptation in export ventures: An empirical investigation. Journal of International Business Studies, 24(3), 479–506. Cerrato, D. (2009). Does innovation lead to global orientation? Empirical evidence from a sample of Italian firms. European Management Journal, 27(5), 305–315. Chakrabarti, A., Singh, K., & Mahmood, I. (2007). Diversification and performance: Evidence from East Asian firms. Strategic Management Journal, 28(2), 101–120. Chandler, A. D. (1962). Strategy and structure. Cambridge, MA: MIT Press.
References
271
Chandler, A. D. (1990). The enduring logic of industrial success. Harvard Business Review, 68(2), 130–140. Chang, S.-J., Van Witteloostuijn, A., & Eden, L. (2010). From the Editors: Common method variance in international business research. Journal of International Business Studies, 41(2), 178–184. Chao, P., Samiee, S., & Yip, L. S.-C. (2003). International marketing and the Asia-Pacific region: Developments, opportunities, and research issues. International Marketing Review, 20(5), 480–492. Chase, R. B., & Erikson, W. J. (1988). The service factory. The Academy of Management Executive, 2(3), 191–196. Chen, H. (1999). International performance of multinationals: A hybrid model. Journal of World Business, 34(2), 157–170. Chen, S. (2007). Testing regional effects in the internationalization-performance relationship in Asian service firms. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 337–358). Amsterdam: Elsevier. Chin, W. W. (1998a). Issues and opinion on structural equation modeling. Management Information Systems Quarterly, 22(1), 7–16. Chin, W. W. (1998b). The partial least squares approach to structural equation modeling. In: G. A. Marcoulides (Ed.), Modern methods for business research (pp. 295–336). Mahwah, NJ: Lawrence Erlbaum Associates. Chin, W. W., Marcolin, B. L., & Newsted, P. R. (1996). A partial least squares latent variable modeling approach for measuring interaction effects: Results from a Monte Carlo simulation study and voice mail emotion/adoption study. Available from http://disc-nt.cba.uh.edu/chin/ icis96.pdf. Accessed 26.09.09. Chin, W. W., Marcolin, B. L., & Newsted, P. R. (2003). A partial least squares latent variable modeling approach for measuring interaction effects: Results from a Monte Carlo simulation study and an electronic-mail emotion/adoption study. Information Systems Research, 14(2), 189–217. Chin, W. W., & Newsted, P. R. (1999). Structural equation modeling analysis with small samples using partial least squares. In: R. H. Hoyle (Ed.), Statistical strategies for small sample research (pp. 307–342). Thousand Oaks, CA: Sage. Churchill, G. A. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, 16(1), 64–73. Churchill, G. A. (1987). Marketing research: Methodological foundations. Fort Worth, TX/ Chicago, IL: Dryden Press. Claessens, S., Dobos, G., Klingebiel, D., & Laeven, L. (2003). The growing importance of networks in finance and its effects on competition. In: A. Nagurney (Ed.), Innovations in financial and economic networks (pp. 110–135). Northampton, MA: Edward Elgar. Clark, T., Knowles, L. L., & Hodis, M. (2004). Global dialogue: A response to the responders in the special globalization issue of JIM. Journal of International Management, 10(4), 511–514. Coase, R. H. (1937). The nature of the firm. Economica, 4(16), 386–405. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale: Lawrence Erlbaum Associates. Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35(1), 128–152. Collinson, S., & Rugman, A. M. (2008). The regional nature of Japanese multinational business. Journal of International Business Studies, 39(2), 215–230. Collis, D., Young, D., & Goold, M. (2007). The size, structure, and performance of corporate headquarters. Strategic Management Journal, 28(4), 383–405. Contractor, F. J. (2007a). Is international business good for companies? The evolutionary or multistage theory of internationalization vs. the transaction cost perspective. Management International Review, 47(3), 453–475. Contractor, F. J. (2007b). The evolutionary or multi-stage theory of internationalization and its relationship to the regionalization of firms. In: A. M. Rugman (Ed.), Regional aspects of
272
References
multinationality and performance (Research in global strategic management, Vol. 13, pp. 11–29). Amsterdam: Elsevier. 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. Costner, H. L. (1969). Theory, deduction, and rules of correspondance. American Journal of Sociology, 75(2), 245–263. Creswell, J. W. (2009). Research design: Qualitative, quantitative, and mixed method approaches. Thousand Oaks, CA: Sage. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. Czinkota, M. R., & Ronkainen, I. A. (2009). Trends and indications in international business. Management International Review, 49(2), 249–265. Daniel, D. R. (1961). Management information crisis. Harvard Business Review, 39(5), 111–121. Daniels, J. D. (1987). Bridging national and global marketing strategies through regional operations. International Marketing Review, 4(3), 29–44. Daniels, J. D., Pitts, R. A., & Tretter, M. J. (1984). Strategy and structure of U.S. multinationals: An exploratory study. Academy of Management Journal, 27(2), 292–307. Daniels, J. D., Pitts, R. A., & Tretter, M. J. (1985). Organizing for dual strategies of product diversity and international expansion. Strategic Management Journal, 6(3), 223–237. Davidson, W. H., & Haspeslagh, P. (1982). Shaping a global product organization. Harvard Business Review, 60(4), 125–132. Dawes, J. (1999). The relationship between subjective and objective company performance measures in market orientation research: Further empirical evidence. Marketing Bulletin, 10 (Research Note 3), 65–75. De Bu´rca, S., Fletcher, R., & Brown, L. (2004). International marketing: An SME perspective. Harlow: Financial Times Prentice Hall. De Vaus, D. A. (2002). Surveys in social research. London: Routledge. Delios, A., & Beamish, P. W. (2005). Regional and global strategies of Japanese firms. Management International Review, 45(1), 19–36. Delios, A., & Henisz, W. J. (2003). Policy uncertainty and the sequence of entry by Japanese firms, 1980–1998. Journal of International Business Studies, 34(3), 227–241. Dess, G. G., & Robinson, R. B. (1984). Measuring organizational performance in the absence of objective measures: The case of the privately-held firm and conglomerate business unit. Strategic Management Journal, 5(3), 265–273. Diamantopoulos, A., & Papadopoulos, N. (2010). Assessing the cross-national invariance of formative measures: Guidelines for international business researchers. Journal of International Business Studies, 41(2), 360–370. Diamantopoulos, A., Riefler, P., & Roth, K. P. (2008). Advancing formative measurement models. Journal of Business Research, 61(12), 1203–1218. Diamantopoulos, A., & Winklhofer, H. M. (2001). Index construction with formative indicators: An alternative to scale development. Journal of Marketing Research, 38(2), 269–277. Diller, H. (2006). Probleme der Handhabung von Strukturgleichungsmodellen in der betriebswirtschaftlichen Forschung. Die Betriebswirtschaft, 66(6), 611–617. Donaldson, L. (2001). The contingency theory of organizations. Thousand Oaks, CA: Sage. Doty, D. H., Glick, W. H., & Huber, G. P. (1993). Fit, equifinality, and organizational effectiveness: A test of two configurational theories. Academy of Management Journal, 36(6), 1196–1250. Doz, Y. L. (1980). Strategic management in multinational companies. Sloan Management Review, 21(2), 27–46. Doz, Y. L. (1986). Strategic management in multinational companies. Oxford: Pergamon Press. Doz, Y. L., & Prahalad, C. K. (1991). Managing diversified MNCs: A search for a new paradigm. Strategic Management Journal, 12(Special issue 1), 145–164.
References
273
Drazin, R., & Van de Ven, A. H. (1985). Alternative forms of fit in contingency theory. Administrative Science Quarterly, 30(4), 514–539. Dunning, J. H. (1993). Internationalizing Porter’s Diamond. Management International Review, 33(2), 7–15. Dunning, J. H. (2001). Global capitalism at bay? London: Routledge. Dunning, J. H., Fujita, M., & Yakova, N. (2007). Some macro-data on the regionalisation/ globalisation debate: A comment on the Rugman/Verbeke analysis. Journal of International Business Studies, 38(1), 177–199. Eberl, M. (2006). Formative und reflektive Konstrukte und die Wahl des Strukturgleichungsverfahrens. Die Betriebswirtschaft, 66(6), 651–668. Ebers, M. (1981). Aufgaben und Ziele der Organisationsforschung. In: A. Kieser (Ed.), Organisationstheoretische Ans€ atze. M€ unchen: Vahlen. ECB (2009). Euro foreign exchange reference rates. Available from http://www.ecb.int/stats/ exchange/eurofxref/html/index.en.html. Accessed 28.08.09. Eden, L. (2009). Letter from the editor-in-chief: Time in international business. Journal of International Business Studies, 40(4), 535–538. Efron, B., & Gong, G. (1983). A leisurely look at bootstrap, the jacknife and cross-validation. The American Statistican, 37(1), 36–48. Efron, B., & Tibshirani, R. J. (1993). An introduction to the bootstrap. New York, NY: Chapman & Hall. Egelhoff, W. G. (1982). Strategy and structure in multinational corporations: An informationprocessing approach. Administrative Science Quarterly, 27(3), 435–458. Egelhoff, W. G. (1988a). Organizing the multinational enterprise: An information-processing perspective. Cambridge, MA: Ballinger. Egelhoff, W. G. (1988b). Strategy and structure in multinational corporations: A revision of the Stopford and Wells model. Strategic Management Journal, 9(1), 1–14. Eggert, A., & Fassott, G. (2005). Zur Verwendung formativer und reflektiver Indikatoren in Strukturgleichungsmodellen: Bestandsaufnahme und Anwendungsempfehlungen. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 31–47). Stuttgart: Sch€affer-Poeschel Verlag. Eggert, A., Fassott, G., & Helm, S. (2005). Identifizierung und Quantifizierung mediierender und moderierender Effekte in komplexen Kausalstrukturen. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 101–116). Stuttgart: Sch€affer-Poeschel Verlag. Elango, B. (2004). Geographic scope of operations by multinational companies: An exploratory study of regional and global strategies. European Management Journal, 22(4), 431–441. Ellis, P. D. (2007). Distance, dependence and diversity of markets: Effects on market orientation. Journal of International Business Studies, 38(3), 374–386. Emrich, C. (2007). Interkulturelles marketing-management: Erfolgsstrategien, Konzepte, Analysen. Wiesbaden: DUV. Enright, M. J. (2005a). Regional management centers in the Asia-Pacific. Management International Review, 45(1), 59–82. Enright, M. J. (2005b). The roles of regional management centers. Management International Review, 45(1), 83–102. Enright, M. J., & Subramanian, V. (2007). An organizing framework for MNC subsidiary typologies. Management International Review, 47(6), 895–924. Epstein, B. J., & Jermakowicz, E. K. (2009). IFRS converges to U.S. GAAP on segment reporting. Journal of Accountancy, 207(4), 50. Fassott, G. (2005). Die PLS-Pfadmodellierung: Entwicklungsrichtungen, M€oglichkeiten, Grenzen. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 19–29). Stuttgart: Sch€affer-Poeschel Verlag. Fassott, G. (2006). Operationalisierung latenter Varialben in Strukturgleichungsmodellen: Eine Standortbestimmung. Zeitschrift f€ ur betriebswirtschaftliche Forschung, 58(2), 67–88.
274
References
Feils, D. J., & Rahman, M. (2008). Regional economic integration and foreign direct investment: The case of NAFTA. Management International Review, 48(2), 147–163. Fincham, R., & Clark, T. (2009). Introduction: Can we bridge the rigour-relevance gap? Journal of Management Studies, 46(3), 510–515. Fornell, C. (1987). A second generation of multivariate analysis: Classification of methods and implications for marketing research. In: M. J. Houston (Ed.), Review of marketing (pp. 407–450). Chicago, IL: American Marketing Association. Fornell, C. (1989). The blending of theoretical and empirical knowledge in structural equations with unobservables. In: H. Wold (Ed.), Theoretical empiricism: A general rationale for scientific model-building (pp. 153–173). New York, NY: Paragon House. Fornell, C., & Cha, J. (1994). Partial least squares. In: R. P. Bagozzi (Ed.), Advanced methods of marketing research (pp. 52–78). Cambridge, MA: Blackwell. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. Fortanier, F., Muller, A., & Van Tulder, R. (2007). Internationalization and performance: The moderating role of strategic fit. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 177–200). Amsterdam: Elsevier. Fortune. (2001). The Fortune Global 500. Fortune, 144(2), F1–F10. Fortune. (2002). The Fortune Global 500. Fortune, 146(2), F1–F10. Fortune. (2003). The Fortune Global 500. Fortune, 148(2), F1–F10. Fortune. (2004). The Fortune Global 500. Fortune, 150(2), F1–F10. Fortune. (2005). The Fortune Global 500. Fortune, 152(2), F1–F10. Fortune. (2006). The Fortune Global 500. Fortune, 154(2), F1–F10. Fortune. (2007). The Fortune Global 500. Fortune, 156(2), F1–F10. Fortune. (2008). The Fortune Global 500. Fortune, 158(2), F1–F10. Fortune. (2009). The Fortune Global 500. Fortune, 160(2), F1–F10. Fouraker, L. E., & Stopford, J. M. (1968). Organizational structure and the multinational strategy. Administrative Science Quarterly, 13(1), 47–64. Franko, L. G. (1976). The European multinationals: A renewed challenge to American and British big business. Stamford, CT: Greylock Publishers. Freeland, R. F. (1996). The myth of the M-form? Governance, consent, and organizational change. American Journal of Sociology, 102(2), 483–526. Fritz, W. (2004a). Die Erfolgsfaktorenforschung – Ein Misserfolg? Die Betriebswirtschaft, 64(5), 623–625. € Fritz, W. (2004b). Uber Tautologien, gesicherte Erkenntnisse und ein problematisches Wissenschaftsverst€ andnis. Available from http://dialog-erfolgsfaktorenforschung.de/ReplikFritz. pdf. Accessed 30.09.09. Geisser, S. (1975). The predictive sample reuse method with applications. Journal of the American Statistical Association, 70(350), 320–328. Gerdin, J., & Greve, J. (2004). Forms of contingency fit in management accounting research – A critical review. Accounting, Organizations & Society, 29(3/4), 303–326. Geringer, M. J., Beamish, P. W., & daCosta, R. C. (1989). Diversification strategy and internationalization. Strategic Management Journal, 10(2), 109–119. Geringer, M. J., & Hebert, L. (1991). Measuring performance of international joint ventures. Journal of International Business Studies, 22(2), 249–263. Ghemawat, P. (1991). Commitment: The dynamic of strategy. New York, NY: The Free Press. Ghemawat, P. (2001). Distance still matters: The hard reality of global expansion. Harvard Business Review, 79(8), 137–147. Ghemawat, P. (2003). Semiglobalization and international business strategy. Journal of International Business Studies, 34(2), 138–152. Ghemawat, P. (2005). Regional strategies for global leadership. Harvard Business Review, 83(12), 98–108.
References
275
Ghemawat, P. (2007a). Managing differences: The central challenge of global strategy. Harvard Business Review, 85(3), 59–68. Ghemawat, P. (2007b). Why the world isn’t flat. Foreign Policy, 159(Mar/Apr), 54–61. Ghemawat, P. (2008). Redefining global strategy: Crossing borders in a world where differences still matter. Boston, MA: Harvard Business School Press. Ghemawat, P., & Ghadar, F. (2006). Global integration 6¼ global concentration. Industrial and Corporate Change, 15(4), 595–623. Ghislanzoni, G., Penttinen, R., & Turnbull, D. (2008). The multilocal challenge: Managing crossborder functions. McKinsey Quarterly, 17(2), 70–81. Ghoshal, S. (1987). Global strategy: An organizing framework Strategic Management Journal, 8(5), 425–440. Ghoshal, S. (2005). Bad management theories are destroying good management practices. Academy of Management Learning & Education, 4(1), 75–91. Ghoshal, S., & Nohria, N. (1989). Internal differentiation within multinational corporations. Strategic Management Journal, 10(4), 323–337. Ghoshal, S., & Nohria, N. (1993). Horses for courses: Organizational forms for multinational corporations. Sloan Management Review, 34(2), 23–35. Giddens, A. (2001). Entfesselte Welt: Wie die Globalisierung unser Leben ver€ andert. Frankfurt am Main: Suhrkamp. Giere, J., Wirtz, B. W., & Schilke, O. (2006). Mehrdimensionale Konstrukte. Die Betriebswirtschaft, 66(6), 678–695. Giering, A. (2000). Der Zusammenhang zwischen Kundenzufriedenheit und Kundenloyalit€ at: Eine Untersuchung moderierender Effekte. Wiesbaden: Gabler. Gilbert, D. U. (1998). Konfliktmanagement in international t€ atigen Unternehmen. Ein diskursethischer Ansatz zur Regelung von Konflikten im interkulturellen Management. Berlin: Sternenfels. Gilbert, D. U. (2003). Vertrauen in strategischen Unternehmensnetzwerken – Ein strukturationstheoretischer Ansatz. Wiesbaden: DUV. Gilbert, D. U. (2005). Kontextsteuerung und Systemvertrauen in strategischen Unternehmensnetzwerken. Die Unternehmung, 59(5), 407–422. Ginsberg, A., & Venkatraman, N. (1985). Contingency perspectives of organizational strategy: A critical review of the empirical research. Academy of Management Review, 10(3), 421–434. Girod, S. J. G., & Rugman, A. M. (2005). Regional business networks and the multinational retail sector. Long Range Planning, 38(4), 335–357. Glaum, M., & Oesterle, M.-J. (2007). 40 years of research on internationalization and firm performance: More questions than answers? Management International Review, 47(3), 307–317. Goerzen, A., & Asmussen, C. G. (2007). The geographic orientation of multinational enterprises and its implications for performance. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 65–83). Amsterdam: Elsevier. Goerzen, A., & Beamish, P. W. (2003). Geographic scope and multinational enterprise performance. Strategic Management Journal, 24(13), 1289–1306. G€ otz, O., & Liehr-Gobbers, K. (2004). Analyse von Strukturgleichungsmodellen mit Hilfe der Partial-Least-Squares(PLS)-Methode. Die Betriebswirtschaft, 64(6), 714–738. Grace, J. B. (2006). Structural equation modeling and natural systems. New York, NY: Cambridge University Press. Griffith, D. A., Cavusgil, S. T., & Xu, S. (2008). Emerging themes in international business research. Journal of International Business Studies, 39(7), 1220–1235. Grosse, R. (1981). Regional offices in mutinational enterprise: The Latin American case. Management International Review, 21(2), 48–56. Grosse, R. (2005). Are the largest financial institutions really “global”. Management International Review, 45(1), 129–144.
276
References
Gulati, R. (2007). Tent poles, tribalism, and boundary spanning: The rigor-relevance debate in management research. Academy of Management Journal, 50(4), 775–782. Haenecke, H., & Forsmann, D. (2006). Erfolgsfaktorenforschnung als Instrument des MarketingControllings. In: C. Zerres & M. P. Zerres (Eds.), Handbuch marketing-controlling (pp. 45–56). Berlin/Heidelberg: Springer. Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis. Upper Saddle River, NJ: Prentice-Hall. Hambrick, D. C., & Lei, D. (1985). Toward an empirical prioritization of contingency variables for business strategy. Academy of Management Journal, 28(4), 763–788. Hamel, G., & Prahalad, C. K. (1983). Managing strategic responsibility in the MNC. Strategic Management Journal, 4(4), 341–351. Hamel, G., & Prahalad, C. K. (1985). Do you really have a global strategy? Harvard Business Review, 63(4), 139–148. Hartmann, F. G. H., & Moers, F. (1999). Testing contingency hypotheses in budgetary research: An evaluation of the use of moderated regression analysis. Accounting, Organizations and Society, 24(4), 291–315. Harzing, A.-W. (1997). Response rates in international mail surveys: Results of a 22-country study. International Business Review, 6(6), 641–665. Harzing, A.-W. (2000). Cross-national industrial mail surveys: Why do response rates differ between countries? Industrial Marketing Management, 29(3), 243–254. Hauser, R. M., & Goldberger, A. S. (1971). The treatment of unobservable variables in path analysis. In: H. L. Costner (Ed.), Sociological methodology (pp. 81–117). San Francisco, CA: Jossey-Bass. Heenan, D. A. (1979). The regional headquarters decision: A comparative analysis. Academy of Management Journal, 22(2), 410–415. Heitmann, M. (2006). Security risks and business opportunities in in-car entertainment. In: K. Lemke, C. Paar, & M. Wolf (Eds.), Embedded security in cars: Securing current and future automotive IT applications (pp. 233–247). Berlin/Heidelberg: Springer. Hejazi, W. (2007). Reconsidering the concentration of US MNE activity: Is it global, regional or national? Management International Review, 47(1), 5–27. Helm, S. (2005). Entwicklung eines formativen Messmodells f€ur das Konstrukt Unternehmensreputation. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 241–254). Stuttgart: Sch€affer-Poeschel Verlag. Hempel, C. G. (1951). The concept of cognitive significance: A reconsideration. Proceedings of the American Academy of Arts and Sciences, 80(1), 61–77. Henisz, W. J., & Delios, A. (2001). Uncertainty, imitation, and plant location: Japanese multinational corporations, 1990–1996. Administrative Science Quarterly, 46(3), 443–475. Hennart, J. F. (2000). Transaction costs theory and the multinational enterprise. In: C. N. Pitelis & R. Sugden (Eds.), The nature of the transnational (pp. 72–118). London: Routledge. Hennart, J.-F. (2007). The theoretical rationale for a multinationality-performance relationship. Management International Review, 47(3), 423–452. Henseler, J. (2005). Einf€ uhrung in die PLS-Pfadmodellierung. Wirtschaftswissen-schaftliches Studium, 34(2), 70–75. Henseler, J., Ringle, C. M., & Sinkovics, R. R. (2009). The use of partial least squares path modeling in international marketing. In: R. R. Sinkovics & P. N. Ghauri (Eds.), New challenges to international marketing (Advances in international marketing, Vol. 20, pp. 277–319). Bingley: Emerald Group Publishing Limited. Herrmann, A., Huber, F., & Kressmann, F. (2006). Varianz- und kovarianzbasierte Strukturgleichungsmodelle – Ein Leitfaden zu deren Spezifikation, Sch€atzung und Beurteilung. Zeitschrift f€ ur betriebswirtschaftliche Forschung, 58(2), 34–66. Hildebrandt, L. (2004). Strukturgleichungsmodelle f€ ur die Konsumentenverhaltensforschung – Methodische Trends und Softwareentwicklungen. In: A. Gr€oppel-Klein (Ed.), Konsumentenverhaltensforschung im 21. Jahrhundert (pp. 541–564). Wiesbaden: Gabler.
References
277
Hildebrandt, L., & Temme, D. (2006). Probleme der Validierung mit Strukturgleichungs-modellen. Die Betriebswirtschaft, 66(6), 618–639. Hinrichs, M. (2008). Anforderungen an ein regionales Controlling – Dargestellt am Beispiel des Volkswagen Konzerns f€ ur die Region Nordamerika. Saarbr€ucken: Verlag Dr. M€uller. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversification: Effects on innovation and firm performance. Academy of Management Journal, 40(4), 767–798. Hodgkinson, G. P., Herriot, P., & Anderson, N. (2001). Re-aligning the stakeholders in management research: Lessons from industrial, work and organizational psychology. British Journal of Management, 12(Suppl. 1), S41–S48. Hodgkinson, G. P., & Rousseau, D. M. (2009). Bridging the rigour-relevance gap in management research: It’s already happening! Journal of Management Studies, 46(3), 534–546. Hofer, C. W. (1975). Toward a contingency theory of business strategy. Academy of Management Journal, 18(4), 784–810. Hofstede, G. (1983). The cultural relativity of organizational practices and theories. Journal of International Business Studies, 14(2), 75–89. Hofstede, G. (1993). Interkulturelle Zusammenarbeit. Kulturen-Organisationen-Management. Gabler: Wiesbaden. Hofstede, G. (1997). Lokales Denken, globales Handeln: Kulturen, Zusammenarbeit und Management. M€unchen: Deutscher Taschenbuch Verlag. Homburg, C. (1992). Die Kausalanalyse: Eine Einf€ uhrung. Wirtschaftswissenschaftliches Studium, 21(10), 499–508. 541–544. Homburg, C., & Baumgartner, H. (1995). Die Kausalanalyse als Instrument der Marketingforschung: Eine Bestandsaufnahme. Zeitschrift f€ ur Betriebswirtschaft, 65(10), 1091–1108. Homburg, C., & Baumgartner, H. (1998). Beurteilung von Kausalmodellen: Bestandsaufnahme und Anwendungsempfehlungen. In: L. Hildebrandt & C. Homburg (Eds.), Die Kausalanalyse: Ein Instrument der empirischen betriebswirtschaftlichen Forschung (pp. 343–369). Stuttgart: Sch€affer-Poeschel Verlag. Homburg, C., & Dobratz, A. (1998). Iterative Modellselektion in der Kausalanalyse. In: L. Hildebrandt & C. Homburg (Eds.), Die Kausalanalyse: Ein Instrument der empirischen betriebswirtschaftlichen Forschung (pp. 447–474). Stuttgart: Sch€affer-Poeschel Verlag. Homburg, C., & Giering, A. (1996). Konzeptualisierung und Operationalisierung komplexer Konstrukte: Ein Leitfaden f€ ur die Marketingforschung. Marketing – Zeitschrift f€ ur Forschung und Praxis, 18(1), 5–24. Homburg, C., & Giering, A. (2001). Personal characteristics as moderators of the relationship between customer satisfaction and loyalty – An empirical analysis. Psychology & Marketing, 18(1), 43–66. Homburg, C., & Hildebrandt, L. (1998). Die Kausalanalyse: Bestandsaufnahme, Entwicklungsrichtungen, Problemfelder. In: L. Hildebrandt & C. Homburg (Eds.), Die Kausalanalyse: Ein Instrument der empirischen betriebswirtschaftlichen Forschung (pp. 15–43). Stuttgart: Sch€affer-Poeschel Verlag. Homburg, C., & Klarmann, M. (2006). Die Kausalanalyse in der empirischen betriebswirtschaftlichen Forschung – Problemfelder und Anwendungsempfehlungen. Die Betriebswirtschaft, 66(6), 727–748. Homburg, C., & Klarmann, M. (2009). Multi Informant-Designs in der empirischen betriebswirtschaftlichen Forschung. Die Betriebswirtschaft, 69(2), 147–171. Homburg, C., & Krohmer, H. (2004). Die Fliegenpatsche als Instrument des wissenschaftlichen Dialogs. Die Betriebswirtschaft, 64(5), 626–631. Hoover’s (2009). Company information. Available from http://www.hoovers.com/. Accessed 01.09.09. Hoque, Z. (2004). A contingency model of the association between strategy, environmental uncertainty and performance measurement: Impact on organizational performance. International Business Review, 13(4), 485–502.
278
References
Huber, F., Heitmann, M., & Herrmann, A. (2006). Ans€atze zur Kausalmodellierung mit Interaktionseffekten. Die Betriebswirtschaft, 66(6), 696–710. Huber, F., Herrmann, A., Meyer, F., Vogel, J., & Vollhardt, K. (2007). Kausalmodellierung mit Partial Least Squares: Eine anwendungsorientierte Einf€ uhrung. Wiesbaden: Gabler. Hulland, J. (1999). Use of partial least squares (PLS) in strategic management research: A review of four recent studies. Strategic Management Journal, 20(2), 195–204. Hult, G. T. M., Ketchen, D. J., Griffith, D. A., Chabowski, B. R., Hamman, M. K., Dykes, B. J., et al. (2008). An assessment of the measurement of performance in international business research. Journal of International Business Studies, 39(6), 1064–1080. Hult, G. T. M., Ketchen, D. J., Griffith, D. A., Finnegan, C. A., Gonzalez-Padron, T., Harmancioglu, N., et al. (2008). Data equivalence in cross-cultural international business research: Assessment and guidelines. Journal of International Business Studies, 39(6), 1027–1044. Hurrle, B., & Kieser, A. (2005). Sind Key Informants verl€assliche Datenlieferanten? Die Betriebswirtschaft, 65(6), 584–602. H€ uttner, M., & Schwarting, U. (1999). Explorative Faktorenanalyse. In: C. Homburg & A. Herrmann (Eds.), Marktforschung: Methoden – Anwendungen – Praxisbeispiele (pp. 381–412). Wiesbaden: Gabler. Hutzschenreuter, T., & Gr€ one, F. (2009). Product and geographic scope changes of multinational enterprises in response to international competition. Journal of International Business Studies, 40(7), 1149–1170. Hymer, S. H. (1976). The international operations of national firms: A study of direct foreign investment. Cambridge, MA: MIT Press. Jarvis, C. B., MacKenzie, S. B., & Podsakoff, P. M. (2003). A critical review of construct indicators and measurement model misspecification in marketing and consumer research. Journal of Consumer Research, 30(2), 199–218. 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), 1–24. Johnson, J. L., & Arunthanes, W. (1995). Ideal and actual product adaptation in US exporting firms. International Marketing Review, 12(3), 31–46. Jones, G., & Khanna, T. (2006). Bringing history (back) into international business. Journal of International Business Studies, 37(4), 453–468. J€ oreskog, K. G. (1971). Statistical analysis of sets of congeneric tests. Psychometrika, 36(2), 109–133. J€ oreskog, K. G. (1973). A general method for estimating a linear structural equation system. In: A. S. Goldberger & O. D. Duncan (Eds.), Structural equation models in the social sciences (pp. 85–112). New York, NY: Academic Press. J€ oreskog, K. G., & S€ orbom, D. (1988). LISREL 7: A guide to the program and applications. Chicago, IL: SPSS, Inc. J€ oreskog, K. G., & Wold, H. (1982). The ML and PLS technique for modeling with latent variables: Historical and comparative aspects. In: K. G. J€oreskog & H. Wold (Eds.), Systems under indirect observation: Causality, structure, prediction (pp. 263–270). Amsterdam: NorthHolland. Keillor, B. D., Hult, G. T. M., & Kandemir, D. (2004). A study of the service encounter in eight countries. Journal of International Marketing, 12(1), 9–35. Kieser, A. (2004). Haltet die beiden Diebe, meine Messer stecken in ihren R€ ucken! Available from http://dialog-erfolgsfaktorenforschung.de/Fritz2.pdf. Accessed 30.09.09. Kieser, A. (2006). Der situative Ansatz. In: A. Kieser & M. Ebers (Eds.), Organisationstheorien (pp. 215–245). Stuttgart: Kohlhammer. Kieser, A., & Kubicek, H. (1992). Organisation. Berlin/New York, NY: Walter de Gruyter.
References
279
Kieser, A., & Leiner, L. (2009). Why the rigour-relevance gap in management research is unbridgeable. Journal of Management Studies, 46(3), 516–533. Kieser, A., & Nicolai, A. T. (2003). Mit der Theorie die wilde Praxis reiten, valleri, vallera, valleri? Die Betriebswirtschaft, 63(5), 589–594. Kieser, A., & Nicolai, A. T. (2005). Success factor research: Overcoming the trade-off between rigor and relevance? Journal of Management Inquiry, 14(3), 275–279. Kindleberger, C. P. (1969). American business abroad: Six lectures on direct investment. New Haven, CT: Yale University Press. Kirstein, S. (2009). Unternehmensreputation: Corporate social responsibility als strategische Option f€ ur deutsche Automobilhersteller. Wiesbaden: Gabler. Kleinbaum, D. G., Kupper, L., & Muller, K. E. (1988). Applied regression analysis and other multivariate methods. Boston: PWS-Kent Publishing Company. Kobrin, S. J. (1991). An empirical analysis of the determinants of global integration. Strategic Management Journal, 12(Special issue 1), 17–31. Kogut, B. (1991). Country capabilities and the permeability of borders. Strategic Management Journal, 12(Summer), 33–47. Kogut, B. (2002). Country capabilities and the permeability of borders. In: A. M. Rugman (Ed.), International business: Critical perspectives on business and management (pp. 66–89). London: Routledge. Kolk, A. (2005). Environmental reporting by multinationals from the triad: Convergence or divergence. Management International Review, 45(1), 145–166. 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., & Jiang, C. X. (2009). Contemporary research trends in international marketing: The 2000s. In: A. M. Rugman (Ed.), The Oxford handbook of international business (pp. 447–501). Oxford: Oxford University Press. Kotler, P., & Armstrong, G. (2004). Principles of marketing. Upper Saddle River, NJ: Pearson Prentice Hall. Kotler, P., Keller, K. L., & Bliemel, F. (2007). Marketing-management. Strategien f€ ur wertschaffendes Handeln. M€ unchen: Pearson Studium. Krafft, M. (1999). An empirical investigation of the antecedents of sales force control systems. Journal of Marketing, 63(3), 120–134. Krafft, M., G€otz, O., & Liehr-Gobbers, K. (2005). Die Validierung von Strukturgleichungsmodellen mit Hilfe des Partial-Least-Squares (PLS)-Ansatzes. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 71–86). Stuttgart: Sch€affer-Poeschel Verlag. Kreikebaum, H. (1997). Strategische Unternehmensplanung. Stuttgart: Kohlhammer. Kreikebaum, H., Gilbert, D. U., & Reinhardt, G. O. (2002). Organisationsmanagement internationaler Unternehmen – Grundlagen und moderne Netzwerkstrukturen. Wiesbaden: Gabler Verlag. Krugman, P. (1993). Regionalism versus multilateralism: Analytical notes. In: J. De Melo & A. Panagariya (Eds.), New dimensions in regional integration (pp. 58–79). Cambridge: Cambridge University Press. Kulic, D. (2009). Automobilindustrie zwischen Globalisierung und Regionalisierung: Ist der Freihandel nur eine Illusion? Hermannstal: Diplomica Verlag. Kumar, N., Stern, L. W., & Anderson, J. C. (1993). Conducting interorganizational research using key informants. Academy of Management Journal, 36(6), 1633–1651. Langguth, H. (2006). Segmentberichterstattung am deutschen Kapitalmarkt. Der Betrieb, 59(12), 625–632. Lasserre, P. (1996). Regional headquarters: The spearhead for Asia Pacific markets. Long Range Planning, 29(1), 30–37. Lasserre, P., & Probert, J. (1998). Competing in Asia Pacific: Understanding the rules of the game. Long Range Planning, 31(1), 30–50.
280
References
Lawrence, P., & Lorsch, J. (1967). Organization and environment: Managing differentiation and integration. Boston, MA: Harvard University Press. Lee, D. Y. (2000). Retail bargaining behaviour of American and Chinese customers. European Journal of Marketing, 34(1/2), 190–206. Lee, H.-Y., Mande, V., & Son, M. (2008). A comparison of reporting lags of multinational and domestic firms. Journal of International Financial Management & Accounting, 19(1), 28–56. Lehrer, M., & Asakawa, K. (1999). Unbundling European operations: Regional management and corporate flexibility in American and Japanese MNCs. Journal of World Business, 34(3), 267–286. Lehrer, M., & Behnam, M. (2009). Modularity vs programmability in design of international products: Beyond the standardization-adaptation tradeoff? European Management Journal, 27(4), 281–292. 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), 37–57. Li, L., & Li, D. (2007). Testing the global and regional strategies of multinational enterprises. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 263–296). Amsterdam: Elsevier. Lim, L. K. S., Acito, F., & Rusetski, A. (2006). Development of archetypes of international marketing strategy. Journal of International Business Studies, 37(4), 499–524. Lindsay, W. M., & Rue, L. W. (1980). Impact of the organization environment on the long-range planning process: A contingency view. Academy of Management Journal, 23(3), 385–404. Lohm€oller, J.-B. (1984). LVPLS, program manual. Latent variables path analysis with partial least-squares estimation. K€ oln: Zentralarchiv f€ ur empirische Sozialforschung. Lohm€oller, J.-B. (1989). Latent variable path modeling with partial least squares. Heidelberg: Physica-Verlag. 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. Luhmann, N. (1986). The autopoiesis of social systems. In: F. Geyer & J. Van der Zouwen (Eds.), Sociocybernetic paradoxes: Observation, control and evolution of self-steering systems (pp. 172–192). London: Sage. Luhmann, N. (1998). Die Wissenschaft der Gesellschaft. Frankfurt am Main: Suhrkamp. Luhmann, N. (2005). The concept of autopoiesis. In: D. Seidl & K. H. Becker (Eds.), Niklas Luhmann and organization studies (pp. 54–63). Kristianstad: Liber & Copenhagen Business School Press. Luther, S. (1998). Herausforderungen an die Betriebswirtschaftslehre – Die Perspektive der Praxis. Die Betriebswirtschaft, 58(6), 701–707. Macharzina, K., Oesterle, M.-J., & Wolf, J. (1998). Europ€aische Managementstile – Eine kulturorientierte Analyse. In: R. Berger & U. Steger (Eds.), Auf dem Weg zur Europ€ aischen Unternehmensf€ uhrung: Ein Lesebuch f€ ur Manager und Europ€ aer (pp. 137–164). M€unchen: Beck. Malhotra, N. K. (1993). Marketing research: An applied orientation. Englewood Cliffs, NJ: Prentice-Hall. Malhotra, N. K., Ulgado, F. M., Agarwal, J., Shainesh, G., & Wu, L. (2005). Dimensions of service quality in developed and developing economies: Multi-country cross-cultural comparisons. International Marketing Review, 22(3), 256–278. Maloney, P. (2007). Absatzmittlergerichtetes, identit€ atsbasiertes Markenmanagement: Eine Erweiterung des innengerichteten, identit€ atsbasierten Marken-managements unter besonderer Ber€ ucksichtigung von Premiummarken. Wiesbaden: Gabler. Manea, J., & Pearce, R. (2006). MNEs’ strategies in Central and Eastern Europe: Key elements of subsidiary behaviour. Management International Review, 46(2), 235–255.
References
281
Markides, C. (2007). In search of ambidextrous professors. Academy of Management Journal, 50(4), 762–768. Mauri, A. J., & Phatak, A. V. (2001). Global integration as inter-area product flows: The internalization of ownership and location factors influencing product flows across MNC units. Management International Review, 41(3), 233–249. Mauri, A. J., & Sambharya, R. B. (2003). The performance implications of a global integration strategy in global industries: An empirical investigation using inter-area product flows. Management International Review, 43(Special issue 2), 27–45. McGahan, A. M. (2007). Academic research that matters to managers: On zebras, dogs, lemmings, hammers, and turnips. Academy of Management Journal, 50(4), 748–753. McLarney, C., & Dastrala, R. (2001). Socio-political structures as determinants of global success: The case of Enron Corporation. International Journal of Social Economics, 28(4), 349–367. McNaughton, R. B. (2002). The use of multiple export channels by small knowledge-intensive firms. International Marketing Review, 19(2/3), 190–203. Meffert, H., Burmann, C., & Kirchgeorg, M. (2008). Marketing: Grundlagen marktorientierter Unternehmensf€ uhrung. Konzepte – Instrumente – Praxisbeispiele. Wiesbaden: Gabler. Melewar, T. C., & Saunders, J. (1999). International corporate visual identity: Standardization or localization? Journal of International Business Studies, 30(3), 583–598. Meyer, A. D., Tsui, A. S., & Hinings, C. R. (1993). Configurational approaches to organizational analysis. Academy of Management Journal, 36(6), 1175–1195. Millar, C. C. J. M., Choi, C. J., & Chen, S. (2005). Globalization rediscovered: The case of uniqueness and “creative industries”. Management International Review, 45(1), 121–128. Miller, S. R., & Richards, M. (2002). Liability of foreignness and membership in a regional economic group: Analysis of the European Union. Journal of International Management, 8(3), 323–337. Mintzberg, H., Ahlstrand, B., & Lampel, J. (2007). Strategy Safari: Eine Reise durch die Wildnis des strategischen managements. Heidelberg: Redline Wirtschaft. Molina-Morales, F. X. (2001). European industrial districts: Influence of geographic concentration on performance of the firm. Journal of International Management, 7(4), 277–294. Moore, K., & Rugman, A. M. (2003). Upfrontcomment: US multinationals are regional not global. Business Strategy Review, 14(4), 2. Morrison, A. J., Ricks, D. A., & Roth, K. (1991). Globalization versus regionalization: Which way for the multinational? Organizational Dynamics, 19(3), 17–29. Morrison, A. J., & Roth, K. (1992). The regional solution: An alternative to globalization. Transnational Corporations, 1(2), 37–55. Morschett, D. (2007). Institutionalisierung und Koordination von Auslandseinheiten: Analyse von Industrie- und Dienstleistungsunternehmen. Wiesbaden: DUV. Nachum, L., & Zaheer, S. (2005). The persistence of distance? The impact of technology on MNE motivations for foreign investment. Strategic Management Journal, 26(8), 747–767. Nacif, R. C. (2003). Online customer loyalty. Wiesbaden: DUV. Nebenzahl, I. D., & Jaffe, E. D. (1996). Measuring the joint effect of brand and country image in consumer evaluation of global products. International Marketing Review, 13(4), 5–22. Nicolai, A. T., & Kieser, A. (2002). Trotz eklatanter Erfolglosigkeit: Die Erfolgsfaktoren-forschung weiter auf Erfolgskurs. Die Betriebswirtschaft, 62(6), 579–596. Nicolai, A. T., & Kieser, A. (2004). Von Konsensgenerierungsmaschinen, Nebelkerzen und “the Operation called ‘Verstehen’”. Die Betriebswirtschaft, 64(5), 631–635. Nunally, J. C. (1978). Psychometric theory. New York, NY: McGraw Hill. Oh, C. H. (2009). The international scale and scope of European multinationals. European Management Journal, 27(5), 336–343. Oh, C. H., & Rugman, A. M. (2006). Regional sales of multinationals in the world cosmetics industry. European Management Journal, 24(2–3), 163–173. Oh, C. H., & Rugman, A. M. (2007). Regional multinationals and the Korean cosmetics industry. Asia Pacific Journal of Management, 24(1), 27–42.
282
References
Ohmae, K. (1985). Triad power: The coming shape of global competition. New York, NY: Free Press. Ohmae, K. (1989a). Managing in a borderless world. Harvard Business Review, 67(3), 152–161. Ohmae, K. (1989b). The global logic of strategic alliances. Harvard Business Review, 67(2), 143–154. Onkvisit, S., & Shaw, J. (2009). International marketing: Strategy and theory. London: Routledge. Osegowitsch, T., & Sammartino, A. (2007). Exploring trends in regionalisation. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 45–64). Amsterdam: Elsevier. Osegowitsch, T., & Sammartino, A. (2008). Reassessing (home-)regionalisation. Journal of International Business Studies, 39(2), 184–196. Palazuelos, E., & Ferna´ndez, R. (2009). Demand, employment, and labour productivity in the European economies. Structural Change and Economic Dynamics, 20(1), 1–15. Papadakis, V. M., Lioukas, S., & Chambers, D. (1998). Strategic decision-making processes: The role of management and context. Strategic Management Journal, 19(2), 115–147. Pappu, R., Quester, P. G., & Cooksey, R. W. (2007). Country image and consumer-based brand equity: Relationships and implications for international marketing. Journal of International Business Studies, 38(5), 726–745. Pearce, R., & Papanastassiou, M. (2006). To “almost see the world”: Hierarchy and strategy in Hymer’s view of the multinational. International Business Review, 15(2), 151–165. Perlmutter, H. V. (1969). The tortuous evolution of the multinational enterprise. Columbia Journal of World Business, 4(1), 9–18. Peter, S. I. (1997). Kundenbindung als Marketingziel: Identifikation und Analyse zentraler Determinanten. Wiesbaden: Gabler. Pla-Barber, J. (2002). From Stopford and Wells’s model to Bartlett and Ghoshal’s typology: New empirical evidence. Management International Review, 42(2), 141–156. Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. Poon, J. P. H., Thompson, E. R., & Kelly, P. F. (2000). Myth of the triad? The geography of trade and investment “blocs”. Transactions of the Institute of British Geographers, 25(4), 427–444. Porter, M. E. (1985). Competitive advantage. New York, NY: Free Press. Porter, M. E. (1986). Competition in global industries: A conceptual framework. In: M. E. Porter (Ed.), Competition in global industries (pp. 15–60). Boston, MA: Harvard Business School Press. Porter, M. E. (1990). The competitive advantage of nations. New York, NY: Free Press. Poynter, T. A., & Rugman, A. M. (1982). World product mandates: How will multinationals respond? Business Quarterly, 47(3), 54–61. Poynter, T. A., & White, R. E. (1984). The strategies of foreign subsidiaries: Responses to organizational slack. International Studies of Management & Organization, 14(4), 91–106. Prahalad, C. K. (1976). The strategic process in a multinational corporation. Boston, MA: Unpublished doctoral dissertation, School of Business Administration, Harvard University Prahalad, C. K., & Doz, Y. L. (1987). The multinational mission: Balancing global integration with local responsiveness. New York, NY: Free Press. Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Harvard Business Review, 68(3), 79–91. Prather-Kinsey, J., & Meek, G. K. (2004). The effect of revised IAS 14 on segment reporting by IAS companies. European Accounting Review, 13(2), 213–234. Proff, H. (2000). Hybrid strategies as a strategic challenge – The case of the German automotive industry. Omega, 28(5), 541–553. Proff, H. (2002). Business unit strategies between regionalisation and globalisation. International Business Review, 11(2), 231–250.
References
283
Pugh, D. S., & Hickson, D. J. (1976). Organizational structure in its context. The Aston programme I. Westmead: Saxon House. Pycraft, M., Singh, H., Phihlela, K., Slack, N., Chambers, S., Harland, C., et al. (2000). Operations management. London: Pearson Education. Rasche, A. (2007). The paradoxical foundation of strategic management. Berlin: Physica/ Springer. Rasche, A., & Behnam, M. (2009). As if it were relevant: A systems theoretical perspective on the relation between science and practice. Journal of Management Inquiry, 18(3), 243–255. Reinartz, W., Krafft, M., & Hoyer, W. D. (2004). The customer relationship management process: Its measurement and impact on performance. Journal of Marketing Research, 41(3), 293–305. Reinecke, J. (2005). Strukturgleichungsmodelle in den Sozialwissenschaften. M€unchen/Wien: Oldenbourg. Riahi-Belkaoui, A. (1999). Performance results of multinationality. Westport, CT: Greenwood Publishing Group. Ricart, J. E., Enright, M. J., Ghemawat, P., Hart, S. L., & Khanna, T. (2004). New frontiers in international strategy. Journal of International Business Studies, 35(3), 175–200. Richter, N. (2007). Intra-regional sales and the internationalization and performance relationship. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 359–381). Amsterdam: Elsevier. Rigdon, E. E. (1998). Structural equation modeling. In: G. A. Marcoulides (Ed.), Modern methods for business research (pp. 251–294). Mahwah, NJ: Lawrence Erlbaum Associates. Rindskopf, D. (1984). Using phantom and imaginary latent variables to parameterize constraints in linear structural models. Psychometrika, 49(1), 37–47. Ringle, C. M. 2004. G€ utemaße f€ ur den Partial Least Squares-Ansatz zur Bestimmung von Kausalmodellen, Working paper no. 16. Hamburg: Institut f€ur Industriebetriebslehre und Organisation, Universit€at Hamburg. Ringle, C. M., Boysen, N., Wende, S., & Will, A. (2006). Messung von Kausalmodellen mit dem Partial-Least-Squares-Verfahren. Das Wirtschaftsstudium, 35(1), 81–87. Ringle, C. M., Wende, S., & Will, A. 2005. SmartPLS: Release 2.0 M3 (beta). http://www. smartpls.de. University of Hamburg, Germany. Rockart, J. F. (1979). Chief executives define their own data needs. Harvard Business Review, 57(2), 81–93. Ronen, S., & Shenkar, O. (1985). Clustering countries on attitudinal dimensions: A review and synthesis. Academy of Management Review, 10(3), 435–454. Rossiter, J. R. (2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19(4), 305–335. Roth, K., & Morrison, A. J. (1990). An empirical analysis of the integration-responsiveness framework in global industries. Journal of International Business Studies, 21(4), 541–564. Roth, K., & Morrison, A. J. (1992). Business-level competitive strategy: A contingency link to internationalization. Journal of Management, 18(3), 473–487. Rugman, A. M. (1981). Inside the multinationals: The economics of internal markets. New York, NY: Columbia University Press. Rugman, A. M. (2000). The end of globalization. London: Random House. Rugman, A. M. (2003a). Regional strategies for service sector multinationals. European Business Journal, 15(1), 1–9. Rugman, A. M. (2003b). Regional strategy and the demise of globalization. Journal of International Management, 9(4), 409–417. Rugman, A. M. (2003c). The regional solution: Triad strategies for multinationals. Business Horizons, 46(6), 3–6. Rugman, A. M. (2005a). A further comment on the myth of globalization. Journal of International Management, 11(3), 441–445. Rugman, A. M. (2005b). The regional multinationals: MNEs and “global” strategic management. Cambridge: Cambridge University Press.
284
References
Rugman, A. M. (2007). Introduction. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 1–8). Amsterdam: Elsevier. Rugman, A. M. (2008). Redefining global strategy: Crossing borders in a world where differences still matter. Journal of International Business Studies, 39(6), 1091–1093. Rugman, A. M. (2009). Is international business strategy global or regional? European Management Journal, 27(5), 293–294. Rugman, A. M., & Bennett, J. (1982). Technology transfer and world product mandating in Canada. Columbia Journal of World Business, 17(4), 58–62. Rugman, A. M., & Brain, C. (2004). Regional strategies of multinational pharmaceutical firms. Management International Review, 44(Special issue 3), 7–25. Rugman, A. M., & Collinson, S. (2004). The regional nature of the world’s automotive sector. European Management Journal, 22(5), 471–482. Rugman, A. M., & Collinson, S. (2005). Multinational enterprises in the new Europe. Are they really global? Organizational Dynamics, 34(3), 258–272. Rugman, A. M., & Collinson, S. (2006). Asian business is regional, not global. In: M. Fratianni (Ed.), Regional economic integration (Research in global strategic management, Vol. 12, pp. 167–199). Amsterdam: Elsevier. Rugman, A. M., & D’Cruz, J. R. (2000). Multinationals as flagship firms: Regional business networks. Oxford: Oxford University Press. Rugman, A. M., & D’Cruz, J. R. (1993). The “double diamond” model of international competitiveness: The Canadian experience. Management International Review, 33(2), 17–39. Rugman, A. M., & Girod, S. J. G. (2003). Retail multinationals and globalization: The evidence is regional. European Management Journal, 21(1), 24–37. Rugman, A. M., & Hodgetts, R. (2001). The end of global strategy. European Management Journal, 19(4), 333–343. Rugman, A. M., & Kirton, J. J. (1998). Multinational enterprise strategy and the NAFTA trade and environment regime. Journal of World Business, 33(4), 438–454. Rugman, A. M., Kirton, J. J., & Soloway, J. (1999). Environmental regulations and corporate strategy. A NAFTA perspective. Oxford: Oxford University Press. Rugman, A. M., Kudina, A., & Yip, G. S. (2007). The regional dimension of UK multinationals. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 297–315). Amsterdam: Elsevier. Rugman, A. M., & Moore, K. (2001). The myths of globalization. Ivey Business Journal, 66(1), 64–68. Rugman, A. M., & Oh, C. H. (2007). Multinationality and regional performance, 2001–2005. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 31–43). Amsterdam: Elsevier. Rugman, A. M., & Sukpanich, N. (2006a). Firm-specific advantages intra-regional sales and performance of multinational enterprises. International Trade Journal, 20(3), 355–382. Rugman, A. M., & Sukpanich, N. (2006b). Intra-regional sales and performance of multinational enterprises. In: M. Fratianni (Ed.), Regional economic integration (Research in global strategic management, Vol. 12, pp. 131–150). Amsterdam: Elsevier. Rugman, A. M., & Verbeke, A. (1991). Environmental change and global competitive strategy in Europe. In: A. M. Rugman & A. Verbeke (Eds.), Global competition and the European community (Research in global strategic management, Vol. 2, pp. 3–27). Greenwich, CT: JAI Press. 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. (1993). Foreign subsidiaries and multinational strategic management: An extension and correction of Porter’s single diamond framework. Management International Review, 33(2), 71–84.
References
285
Rugman, A. M., & Verbeke, A. (1998a). Corporate strategy and international environmental policy. Journal of International Business Studies, 29(4), 819–833. Rugman, A. M., & Verbeke, A. (1998b). Multinational enterprises and public policy. Journal of International Business Studies, 29(1), 115–136. Rugman, A. M., & Verbeke, A. (2001). Subsidiary-specific advantages in multinational enterprises. Strategic Management Journal, 22(3), 237–250. Rugman, A. M., & Verbeke, A. (2003a). Extending the theory of the multinational enterprise: Internalization and strategic management perspectives. Journal of International Business Studies, 34(2), 125–137. Rugman, A. M., & Verbeke, A. (2003b). The World Trade Organization, multinational enterprise, and civil society. In: M. Fratianni, P. Savona, & J. J. Kirton (Eds.), Sustaining global growth and development: G7 and IMF governance (pp. 81–97). Aldershot: Ashgate. 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(1), 3–15. Rugman, A. M., & Verbeke, A. (2006). Regional multinational corporations and triad strategy. In: J. J. Kirton & P. I. Hajnal (Eds.), Sustainability, civil society and international governance: Local, North American, and global contributions (pp. 123–136). Aldershot: Ashgate. Rugman, A. M., & Verbeke, A. (2007). Liabilites of regional foreignness and the use of firm-level versus country-level data: A response to Dunning et al. (2007). Journal of International Business Studies, 38(1), 200–205. Rugman, A. M., & Verbeke, A. (2008a). A new perspective on the regional and global strategies of multinational services firms. Management International Review, 48(4), 397–411. Rugman, A. M., & Verbeke, A. (2008b). A regional solution to the strategy and structure of multinationals. European Management Journal, 26(5), 305–313. Rugman, A. M., & Verbeke, A. (2008c). The theory and practice of regional strategy: A response to Osegowitsch and Sammartino. Journal of International Business Studies, 39(2), 326–332. Ruigrok, W., Amann, W., & Wagner, H. (2007). The internationalization-performance relationship at Swiss firms: A test of the S-shape and extreme degrees of internationalization. Management International Review, 47(3), 349–368. Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43(1), 63–83. Rynes, S. L. (2007). Carrying Sumantra Ghoshal’s torch: Creating more positive, relevant, and ecologically valid research. Academy of Management Journal, 50(4), 745–747. Samiee, S. (1994). Customer evaluation of products in a global market. Journal of International Business Studies, 25(3), 579–604. Sarkar, S., & Pfeifer, J. (2005). The philosophy of science: An encyclopedia. London: Routledge. Schein, E. H. (1995). Unternehmenskultur: Ein Handbuch f€ ur F€ uhrungskr€ afte. Frankfurt am Main/ New York, NY: Campus-Verlag. Schendel, D. (1995). Notes from the editor-in-chief. Strategic Management Journal, 16(1), 1–2. Scherer, A. G. (2006). Kritik der Organisation oder Organisation der Kritik? – Wissenschaftstheoretische Bemerkungen zum kritischen Umgang mit Organisationstheorie. In: A. Kieser & M. Ebers (Eds.), Organisationstheorien (pp. 19–61). Stuttgart: Kohlhammer. Schildbach, T. (2008). Der Konzernabschluss nach HGB, IFRS und US-GAAP. M€unchen: Oldenbourg Wissenschaftsverlag. Schlie, E., & Yip, G. S. (2000). Regional follows global: Strategy mixes in the world automotive industry. European Management Journal, 18(4), 343–353. Schmidt, R. (2004). Marketing-Management und Unternehmenserfolg von Wachstumsunternehmen der TIMES-Branchen – Eine empirische Analyse am Neuen Markt. Frankfurt am Main: Peter Lang – Europ€aischer Verlag der Wissenschaften. Schnell, R., Hill, P. B., & Esser, E. (2005). Methoden der empirischen Sozialforschung. M€unchen: Oldenbourg Wissenschaftsverlag.
286
References
Scholderer, J., & Balderjahn, I. (2005). PLS versus LISREL: Ein Methodenvergleich. In: F. Bliemel, A. Eggert, G. Fassott, & J. Henseler (Eds.), Handbuch Pfadmodellierung: Methode – Anwendung – Praxisbeispiele (pp. 87–98). Stuttgart: Sch€affer-Poeschel Verlag. Scholderer, J., Balderjahn, I., & Paulssen, M. (2006). Kausalit€at, Linearit€at, Reliabilit€at: Drei Dinge, die Sie nie € uber Strukturgleichungsmodelle wissen wollten. Die Betriebswirtschaft, 66(6), 640–650. Scholz, C. (1998). Organisations- und Personalarbeit in Europa: Von operativer Oberfl€achlichkeit zur eurostrategischen Grundausrichtung. In R. Berger & U. Steger (Eds.), Auf dem Weg zur Europ€ aischen Unternehmensf€ uhrung: Ein Lesebuch f€ ur Manager und Europ€ aer (pp. 181–199). M€unchen: Beck. Schoonhoven, C. B. (1981). Problems with contingency theory: Testing assumptions hidden within the language of contingency “theory”. Administrative Science Quarterly, 26(3), 349–377. Schrey€ogg, G. (2008). Organisation: Grundlagen moderner Organisationsgestaltung; mit Fallstudien. Wiesbaden: Gabler. Schuiling, I., & Kapferer, J.-N. (2004). Executive insights: Real differences between local and international brands: Strategic implications for international marketers. Journal of International Marketing, 12(4), 97–112. Schulze, P. M. (1998). Beschreibende Statistik. M€ unchen/Wien: Oldenbourg. Sch€utte, H. (1997). Strategy and organisation: Challenges for European MNCs in Asia. European Management Journal, 15(4), 436–445. Schween, C. (2006). Standardentwurf des IASB zur Segmentberichterstattung: ED IFRS 8 – Operating Segments. Die Wirtschaftspr€ ufung, 59(8), 516–517. Seltin, N., & Keeves, J. P. (1994). Path analysis with latent variables. In: T. Husen & T. Postlethwaite (Eds.), The international encyclopedia of education (pp. 4352–4359). Oxford: Oxford University Press. Seno-Alday, S. (2009). Market characteristics and regionalisation patterns. European Management Journal, 27(5), 366–376. Sethi, D. (2009). Are multinational enterprises from the emerging economies global or regional? European Management Journal, 27(5), 356–365. Sheldon, D. H. (2006). World class sales & operations planning: A guide to successful implementation and robust execution. Fort Lauderdale, FL: J. Ross Publishing. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: New procedures and recommendations. Psychological Methods, 7(4), 422–445. Sillince, J. A. A. (2005). A contingency theory of rhetorical congruence. Academy of Management Review, 30(3), 608–621. Simon, H. A. (1957). Administrative theory: A study of decision-making processes in administrative organization. New York, NY: Macmillan. Simon, H. A. (1967). The business school: A problem in organizational design. Journal of Management Studies, 4(1), 1–16. Smith, A. (1776). An inquiry into the nature and causes of the wealth of nations. London: W. Strahan and T. Cadell. Smith, L. F., Gratz, Z., & Bousquet, S. G. (2008). The art and practice of statistics. Belmont, CA: Cengage Learning. Sosik, J. J., Kahai, S. S., & Piovoso, M. J. (2009). Silver bullet or voodoo statistics?: A primer for using the partial least squares data analytic technique in group and organization research. Group and Organization Management, 34(5), 5–36. Sousa, R., & Voss, C. A. (2008). Contingency research in operations management practices. Journal of Operations Management, 26(6), 697–713. SPSS. (2009). SPSS for Windows: Release 17 (11 March 2009). http://www.spss.com/. Chicago, IL: SPSS, Inc. Stanko, B. B., Utterback, J., & Fitzgerald, J. (2002). Segment reporting: The aftermath effects of statement of financial accounting standards no. 131. Journal of Applied Business Research, 18(4), 97–105.
References
287
Starkey, K., Hatchuel, A., & Tempest, S. (2009). Management research and the new logics of discovery and engagement. Journal of Management Studies, 46(3), 547–558. Steil, B. (2009). Lessons of the financial crisis. Available from http://www.cfr.org/content/publications/attachments/Financial_Regulation_CSR45.pdf. Accessed 17.12.09. Stevens, M. J., & Bird, A. (2004). On the myth of believing that globalization is a myth: Or the effects of misdirected responses on obsolescing an emergent substantive discourse. Journal of International Management, 10(4), 501–510. Stone, M. (1974). Cross-validatory choice and assessment of statistical predictions. Journal of the Royal Statistical Society (Series B), 36(2), 111–147. Stopford, J. M., & Wells, L. T. (1972). Managing the multinational enterprise: Organization of the firm and ownership of the subsidiaries. New York, NY: Basic Books. Sukpanich, N. (2007). Intra-regional sales and performance. In: A. M. Rugman (Ed.), Regional aspects of multinationality and performance (Research in global strategic management, Vol. 13, pp. 317–336). Amsterdam: Elsevier. Sullivan, D. (1994). Measuring the degree of internationalization of a firm. Journal of International Business Studies, 25(2), 325–342. Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics. New York, NY: Harper and Row. Takeuchi, H., & Porter, M. E. (1986). Three roles of international marketing in global strategy. In: M. E. Porter (Ed.), Competition in global industries (pp. 111–146). Boston, MA: Harvard Business School Press. Takeuchi, H., & Porter, M. E. (1989). Die drei Aufgaben des internationalen Marketing im Rahmen einer globalen Unternehmensstrategie. In: M. E. Porter (Ed.), Globaler Wettbewerb: Strategien der neuen Internationalisierung (pp. 127–164). Wiesbaden: Gabler. Tallmann, S. B., & Yip, G. S. (2009). Strategy and the multinational enterprise. In: A. M. Rugman (Ed.), The Oxford handbook of international business (pp. 307–340). Oxford: Oxford University Press. Temme, D., & Paulssen, M. (2009). Common method variance. Die Betriebswirtschaft, 69(2), 123–146. Tenenhaus, M., Vinzi, V. E., Chatelin, Y.-M., & Lauro, C. (2005). PLS path modeling. Computational Statistics & Data Analysis, 48(1), 159–205. Thomas, D. E., & Eden, L. (2004). What is the shape of the multinationality-performance relationship? Multinational Business Review, 12(1), 89–110. Thompson, J. D. (1967). Organizations in action. New York, NY: McGraw-Hill. Tong, T. W., & Reuer, J. J. (2007). Real options in multinational corporations: Organizational challenges and risk implications. Journal of International Business Studies, 38(2), 215–230. Tribe, K. (1995). National liberty and laissez faire: How Adam Smith became a free trade ideologue. In: S. Copley & K. Sutherland (Eds.), Adam Smith’s wealth of nations: New interdisciplinary essays (pp. 23–44). Manchester: Manchester University Press. Trommsdorff, V. 1990. Erfolgsfaktorenforschung, Produktinnovation und Schnittstelle Marketing – F&E, Discussion paper no. 143. Berlin: Wirtschaftswissenschaftliche Publikation, Fachbereich 18, Technische Universit€at Berlin. Tushman, M., & O’Reilly, C. (2007). Research and relevance: Implications of Pasteur’s quadrant for doctoral programs and faculty development. Academy of Management Journal, 50(4), 769–774. UNCTAD. (2008). Handbook of statistics. Available from http://www.unctad.org/en/docs/tdstat33_en.pdf. Accessed 14.11.09. United Nations. (1973). Multinational corporations in world development. In: UN Department of Economic and Social Affairs (Ed.), Vol. ST/ECA/190 (E.73.II.A.11). New York, NY. Van Pham, K.-Q. (2006). Strategic offshoring from a decomposed COO’s perspective: A crossregional study of four product categories. Journal of American Academy of Business, Cambridge, 8(2), 59–66. Van Tulder, R., Van den Berghe, D., & Muller, A. (2001). The world’s largest firms and internationalization. Rotterdam: Rotterdam School of Management/Erasmus University.
288
References
Venaik, S., Midgley, D. F., & Devinney, T. M. (2005). Dual paths to performance: The impact of global pressures on MNC subsidiary conduct and performance. Journal of International Business Studies, 36(6), 655–675. Venkatraman, N. (1989a). Strategic orientation of business enterprises: The construct, dimensionality, and measurement. Management Science, 35(8), 942–962. Venkatraman, N. (1989b). The concept of fit in strategy research: Toward verbal and statistical correspondence. Academy of Management Review, 14(3), 423–444. Venkatraman, N., & Camillus, J. C. (1984). Exploring the concept of “fit” in strategic management. Academy of Management Review, 9(3), 513–525. Venkatraman, N., & Ramanujam, V. (1986). Measurement of business performance in strategy research: A comparison of approaches. Academy of Management Review, 11(4), 801–814. Venzin, M., Kumar, V., & Kleine, J. (2008). Internationalization of retail banks: A micro-level study of the multinationality-performance relationship. Management International Review, 48(4), 463–485. Verbeke, A. (2009). International business strategy: Rethinking the foundations of global corporate success. Cambridge: Cambridge University Press. Verbeke, A., & Greidanus, N. S. (2009). The end of the opportunism vs trust debate: Bounded reliability as a new envelope concept in research on MNE governance. Journal of International Business Studies, 40(9), 1471–1495. Verbeke, A., & Kenworthy, T. P. (2008). Multidivisional vs metanational governance of the multinational enterprise. Journal of International Business Studies, 39(6), 940–956. Verbeke, A., Li, L., & Goerzen, A. (2009). Toward more effective research on the multinationality-performance relationship. Management International Review, 49(2), 149–161. Vermeulen, F. (2007). “I shall not remain insignificant”: Adding a second loop to matter more. Academy of Management Journal, 50(4), 754–761. Vermeulen, F., & Barkema, H. G. (2002). Pace, rhythm, and scope: Process dependence in building a profitable multinational corporation. Strategic Management Journal, 23(7), 637–653. Vernon, R. (1966). International investment and international trade in the product cycle. Quarterly Journal of Economics, 80(2), 190–207. Vernon, R., & Wells, L. T. (1986). Manager in the international economy. Englewood Cliffs, NJ: Prentice-Hall. von der Lippe, P., & Kladroba, A. (2002). Repr€asentativit€at von Stichproben. Marketing – Zeitschrift f€ ur Forschung und Praxis, 24(2), 139–145. von Hayek, F. A. (1969). Freiburger Studien. T€ ubingen: Mohr Siebeck. Vora, D., Kostova, T., & Roth, K. (2007). Roles of subsidiary managers in multinational corporations: The effect of dual organizational identification. Management International Review, 47(4), 595–620. € Walter, G., & Z€urn, M. (1998). Uber die These von der Erosion des Nationalstaates. In: R. Berger & U. Steger (Eds.), Auf dem Weg zur Europ€ aischen Unternehmensf€ uhrung: Ein Lesebuch f€ ur Manager und Europ€ aer (pp. 41–63). M€ unchen: Beck. Ward, J. J. (1973). Product and promotion adaptation by European firms in the U.S. Journal of International Business Studies, 4(1), 79–85. Weiber, R., & Adler, J. (2002). Hemnisfaktoren im Electronic Business. Ansatzpunkte einer theoretischen Systematisierung und empirische Evidenzen. Marketing Zeitschrift f€ ur Forschung und Praxis, 24(Special Issue “E-Marketing”), 5–12. Werts, C. E., Linn, R. L., & J€ oreskog, K. G. (1974). Intraclass reliability estimates: Testing structural assumptions. Educational and Psychological Measurement, 34(1), 25–33. Westney, D. E. (2006). Book review of “the regional multinationals: MNEs and ‘global’ strategic management”. Journal of International Business Studies, 37(3), 445–449. Wetzels, M., Odekerken-Schr€ oder, G., & Van Oppen, C. (2009). Using PLS path modeling for assessing hierarchical construct models: Guidelines and empirical illustration. Management Information Systems Quarterly, 33(1), 177–195. White, R. E., & Poynter, T. A. (1984). Strategies for foreign-owned subsidiaries in Canada. Business Quarterly, 48(4), 59–69. Whitley, R. (1984). The scientific status of management research as a practically-oriented social science. Journal of Management Studies, 21(4), 369–390.
References
289
Whittington, R. (2006). Organizational structure. In: D. O. Faulkner & A. Campbell (Eds.), The Oxford handbook of strategy: A strategy overview and competitive strategy (pp. 811–840). Oxford: Oxford University Press. Williams, C. R. (1967). Regional management overseas. Harvard Business Review, 45(1), 87–91. Williamson, O. E. (1975). Markets and hierarchies: Analysis and anti-trust implications. New York, NY: Free Press. Williamson, O. E. (1981). The Economics of organization: The transaction cost approach. The American Journal of Sociology, 87(3), 548–577. Williamson, O. E. (1996). The mechanisms of governance. Oxford, New York: Oxford University Press. Wind, Y., Douglas, S. P., & Perlmutter, H. V. (1973). Guidelines for developing international marketing strategies. Journal of Marketing, 37(2), 14–23. Wold, H. (1966). Estimation of principal components and related models by iterative least squares. In: P. R. Krishnaiah (Ed.), Multivariate analysis (pp. 391–420). New York, NY: Academic Press. Wold, H. (1980). Model construction and evaluation when theoretical knowledge is scarce: Theory and application of PLS. In: J. Kmenta & J. B. Ramsey (Eds.), Evaluation of econometric models (pp. 47–74). New York, NY: Academic Press. Wold, H. (1982). Soft modeling: The basic design and some extensions. In: K. G. J€oreskog & H. Wold (Eds.), Systems under indirect observation: Causality, structure, prediction (pp. 1–54). Amsterdam: North-Holland. Wolf, J., & Egelhoff, W. G. (2002). A reexamination and extension of international strategystructure theory. Strategic Management Journal, 23(2), 181–189. Woodward, J. (1965). Industrial organization: Theory and practice. London: Oxford University Press. Worrell, D. L. (2009). Assessing business scholarship: The difficulties in moving beyond the rigorrelevance paradigm trap. Academy of Management Learning & Education, 8(1), 127–130. Xu, D., & Shenkar, O. (2002). Institutional distance and the multinational enterprise. Academy of Management Review, 27(4), 608–618. Yang, Z., Wang, X., & Su, C. (2006). A review of research methodologies in international business. International Business Review, 15(6), 601–617. Yaveroglu, I. S., & Donthu, N. (2002). Cultural influences on the diffusion of new products. Journal of International Consumer Marketing, 14(4), 49–63. Yeniyurt, S., & Townsend, J. D. (2003). Does culture explain acceptance of new products in a country? International Marketing Review, 20(4), 377–396. Yeung, H. W.-C., Poon, J., & Perry, M. (2001). Towards a regional strategy: The role of regional headquarters of foreign firms in Singapore. Urban Studies, 38(1), 157–183. Yin, E., & Choi, C. J. (2005). The globalization myth: The case of China. Management International Review, 45(1), 103–120. Yip, G. S., Rugman, A. M., & Kudina, A. (2006). International success of British companies. Long Range Planning, 39(3), 241–264. Zaheer, S. (1995). Overcoming the liability of foreignness. Academy of Management Journal, 38 (2), 341–363. Zaheer, S. (2002). The liability of foreignness, redux: A commentary. Journal of International Management, 8(3), 351–358. 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. Zeithaml, V. A., Varadarajan, P. R., & Zeithaml, C. P. (1988). The contingency approach: Its foundations and relevance to theory building and research in marketing. European Journal of Marketing, 22(7), 37–64. Zentes, J. (1998). Strategieformulierung und -umsetzung: Ein Perspektivenwechsel. In: R. Berger & U. Steger (Eds.), Auf dem Weg zur Europ€ aischen Unternehmensf€ uhrung: Ein Lesebuch f€ ur Manager und Europ€ aer (pp. 167–180). M€ unchen: Beck. Zheng Zhou, K., Brown, J. R., Dev, C. S., & Agarwal, S. (2007). The effects of customer and competitor orientations on performance in global markets: A contingency analysis. Journal of International Business Studies, 38(2), 303–319. € Zinnbauer, M., & Eberl, M. (2005). Uberpr€ ufung der Spezifikation und G€ute von Strukturgleichungsmodellen. Wirtschaftswissenschaftliches Studium, 34(10), 566–572.