New Challenges to International Marketing (Advances in International Marketing, Volume 20)

NEW CHALLENGES TO INTERNATIONAL MARKETING ADVANCES IN INTERNATIONAL MARKETING Series Editor: S. Tamer Cavusgil Recent...

Author: Tamer Cavusgil | Rudolf R. Sinkovics | Pervez N. Ghauri

109 downloads 1890 Views 3MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

NEW CHALLENGES TO INTERNATIONAL MARKETING

ADVANCES IN INTERNATIONAL MARKETING Series Editor: S. Tamer Cavusgil Recent Volumes: Volume 5:

Industrial Networks

Volume 6:

Export Marketing: International Perspectives

Volume 7:

Marketing in Asia Paciﬁc and Beyond

Volume 8:

Parts I & II

Volume 9:

International Marketing and Purchasing

Volume 10:

Globalization, the Multinational Firm, and Emerging Economies

Volume 11:

Reassessing the Internationalization of the Firm

Volume 12:

New Directions in International Advertising Research

Volume 13:

Study Abroad: Perspectives and Experiences from Business Schools

Volume 14:

Reviving Traditions in Research on International Market Entry

Volume 15:

Research on International Service Marketing: A State of the Art

Volume 16:

Relationship Between Exporters and their Foreign Sales and Marketing Intermediaries

Volume 17:

International Market Research: Opportunities and Challenges in the 21st Century

Volume 18:

Cross-Cultural Buyer Behaviour

Volume 19:

Michigan State University Contributions to International Business and Innovation

ADVANCES IN INTERNATIONAL MARKETING

VOLUME 20

NEW CHALLENGES TO INTERNATIONAL MARKETING EDITED BY

RUDOLF R. SINKOVICS Manchester Business School, Manchester, UK

PERVEZ N. GHAURI King’s College London, London, UK

United Kingdom – North America – Japan India – Malaysia – China

JAI Press is an imprint of Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK First edition 2009 Copyright r 2009 Emerald Group Publishing Limited Reprints and permission service Contact: [email protected] No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. No responsibility is accepted for the accuracy of information contained in the text, illustrations or advertisements. The opinions expressed in these chapters are not necessarily those of the Editor or the publisher. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-84855-468-9 ISSN: 1474-7979 (Series)

Awarded in recognition of Emerald’s production department’s adherence to quality systems and processes when preparing scholarly journals for print

CONTENTS LIST OF CONTRIBUTORS

ix

PART I: INTERNATIONALIZATION INTRODUCTION Rudolf R. Sinkovics and Pervez N. Ghauri THE PROPENSITY TO CONTINUE INTERNATIONALIZATION: A STUDY OF ENTREPRENEURIAL DECISION-MAKING IN AUSTRALIAN SERVICE FIRMS Pieter Pauwels, Paul G. Patterson, Ko de Ruyter and Martin Wetzels

3

11

FACTORS INFLUENCING THE INTERNATIONALISATION PROCESS OF UK FIRMS IN ASIA Kannika Leelapanyalert

37

TOURISM AS A LEVERAGE OF INTERNATIONALIZATION FOR CONSUMER GOODS FIRMS: A CASE STUDY APPROACH Konstantinos Poulis and Mo Yamin

69

PART II: ENTREPRENEURSHIP AND FUNCTIONAL ADVANCEMENTS RELATIONSHIPS BETWEEN THE DIMENSIONS OF INTERNATIONAL GROWTH ORIENTATION, ENVIRONMENTAL TURBULENCE, AND STRATEGIC ORIENTATIONS Sanna Sundqvist and Olli Kuivalainen v

89

vi

CONTENTS

INTERNATIONALISATION AND TECHNOLOGICAL INNOVATION: EMPIRICAL EVIDENCE ON THEIR MUTUAL RELATIONSHIP Diana A. Filipescu, Alex Rialp and Josep Rialp

125

THE ROLE OF INNOVATIVE AND ENTREPRENEURIAL BEHAVIOR IN INTERNATIONALIZATION PROCESSES Arild Aspelund and Tage Koed Madsen

155

BRAND IMAGE PERCEPTIONS ACROSS CULTURES: A STUDY OF SYMBOLIC AND FUNCTIONAL ASSOCIATIONS Laura Salciuviene, Pervez N. Ghauri, Audra I. Mockaitis and Claudio De Mattos

177

PART III: COLLABORATIVE RELATIONSHIPS DO FIRMS CREATE VALUE THROUGH INTERNATIONAL STRATEGIC ALLIANCES? Chia-Ling ‘Eunice’ Liu

195

WHAT CAUSES BREAK-UPS? FACTORS DRIVING THE DISSOLUTION OF MARKETING-ORIENTED INTERNATIONAL JOINT VENTURES Mehmet Berk Talay and M. Billur Akdeniz

227

GEOGRAPHIC MARKET DIVERSIFICATION: A PREMIUM OR DISCOUNT IN FIRM’S VALUE Z. Seyda Deligonul

257

PART IV: METHODOLOGICAL ADVANCES THE USE OF PARTIAL LEAST SQUARES PATH MODELING IN INTERNATIONAL MARKETING Jo¨rg Henseler, Christian M. Ringle and Rudolf R. Sinkovics

277

Contents

ADVANCING THE UNDERSTANDING OF CONSTRUCT VALIDITY AND CROSS-NATIONAL COMPARABILITY: ILLUSTRATED BY A FIVE-COUNTRY STUDY OF CORPORATE EXPORT INFORMATION USAGE Thomas Salzberger, Hartmut H. Holzmu¨ller and Anne Souchon THE DYNAMICS OF TECHNOLOGICAL READINESS IN MARKETING UNITS: WHY CROSS-CULTURAL EXAMINATION IS NECESSARY R. Glenn Richey, Daniel G. Bachrach, Michael G. Harvey and Hui Wang METHODOLOGICAL RIGOR OF ORDINARY LEAST SQUARES REGRESSION STUDIES IN MAINSTREAM MARKETING: AN ASSESSMENT OF RESEARCH REPORTED IN THE JOURNAL OF MARKETING Z. Seyda Deligonul, Brian R. Chabowski, Steven H. Seggie, Shichun Xu and S. Tamer Cavusgil

vii

321

361

389

LIST OF CONTRIBUTORS M. Billur Akdeniz

Michigan State University, East Lansing, MI, USA

Arild Aspelund

Norwegian University of Science and Technology, Trondheim, Norway

Daniel G. Bachrach

The University of Alabama, Tuscaloosa, AL, USA

S. Tamer Cavusgil

Georgia State University, Atlanta, GA, USA

Brian R. Chabowski

The University of Tulsa, Tulsa, OK, USA

Z. Seyda Deligonul

St. John Fisher College, Pittsford, NY, USA

Claudio De Mattos

Manchester Business School, Manchester, UK

Ko de Ruyter

Maastricht University, Maastricht, The Netherlands

Diana A. Filipescu

Autonomous University of Barcelona, Barcelona, Spain

Pervez N. Ghauri

King’s College London, London, UK

Michael G. Harvey

The University of Mississippi, Oxford, MS, USA

Jo¨rg Henseler

Radboud University Nijmegen, Nijmegen, The Netherlands

Hartmut H. Holzmu¨ller

University of Dortmund, Dortmund, Germany

ix

x

LIST OF CONTRIBUTORS

Tage Koed Madsen

University of Southern Denmark, Odense, Denmark

Olli Kuivalainen

Lappeenranta University of Technology, Lappeenranta, Finland

Kannika Leelapanyalert

Birkbeck College, University of London, London, UK

Chia-Ling ‘Eunice’ Liu

National Chung Cheng University, Chia-Yi, Taiwan

Audra I. Mockaitis

Victoria University of Wellington, Wellington, New Zealand

Paul G. Patterson

University of New South Wales, Sydney, Australia

Pieter Pauwels

Maastricht University, Maastricht, The Netherlands

Konstantinos Poulis

University College Cork, Cork, Ireland

Alex Rialp

Autonomous University of Barcelona, Barcelona, Spain

Josep Rialp

Autonomous University of Barcelona, Barcelona, Spain

R. Glenn Richey

The University of Alabama, Tuscaloosa, AL, USA

Christian M. Ringle

University of Hamburg, Hamburg, Germany

Laura Salciuviene

Lancaster University Management School, Lancaster, UK

Thomas Salzberger

Wirtschaftsuniversita¨t Wien, Wien, Austria

Steven H. Seggie

Ozyegin University, Istanbul, Turkey

Rudolf R. Sinkovics

Manchester Business School, Manchester, UK

xi

List of Contributors

Anne Souchon

Loughborough University, Loughborough, Leicestershire, UK

Sanna Sundqvist

Lappeenranta University of Technology, Finland

Mehmet Berk Talay

HEC Montreal, Montreal, Canada

Hui Wang

Guanghua School of Management, Beijing, China

Martin Wetzels

Maastricht University, Maastricht, The Netherlands

Shichun Xu

The University of Tennessee, Knoxville, TN, USA

Mo Yamin

Manchester Business School, Manchester, UK

PART I INTERNATIONALIZATION

INTRODUCTION Rudolf R. Sinkovics and Pervez N. Ghauri Labelling this volume ‘‘New challenges’’ proved to be a challenge in itself, as we have to visualize what we mean by ‘‘New’’. On the one hand, it is the very nature of our academic investigations that we are aiming to continuously push frontiers and explore phenomena which have been underexplored. On the other hand, there is nothing ‘‘new’’ in the fact that the economic, physical, cultural and political environment within which international marketers operate is continuously changing. Given this pretext, we still chose to label this volume ‘‘New challenges’’. This is because in the recent years we have seen a number of phenomena which have been emerging that we had not seen before. The anxieties regarding increasing prices of commodities, spatial reorganization of production processes (Buckley & Ghauri, 2004), outsourcing and offshoring of vital supply chain activities are some of the factors that are creating new challenges and are transforming the landscape of international marketing. Globalization, the all-encompassing force which transforms the fabric of our socio-cultural and economic interactions (Giddens, 2000, 2002) continues as a topic on the rise (Miller, Hartwick, & Le Breton-Miller, 2004) and is inﬂuencing marketing strategies of internationally active ﬁrms. Adding to this, developments such as the advancement in technology and modernity (Levitt, 1983) are heating up and revolutionizing the way companies do business in the international marketing arena (Jean, 2007). In fact, we see a faster pace of technological innovation and technological connectivity is transforming the way people live and consume. Mega-trends such as the rise of ﬁrms from China and India (Doh, 2005; Lewin & Peeters, New Challenges to International Marketing Advances in International Marketing, Volume 20, 3–10 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020003

3

4

RUDOLF R. SINKOVICS AND PERVEZ N. GHAURI

2006) as mature and valid competitors to Western ﬁrms (Aulakh, 2007; Dunning, 2006) require international marketers to reframe their traditional thinking. In the call for papers we encouraged submissions around these trends, e.g. the changing nature of technology-enabled international marketing research, the impact of environmental degradation, global warming and scarcity of resources on international marketing strategies, global account management, procurement and international supplier networks, internationalization of small and entrepreneurial ﬁrms, implications of outsourcing and offshoring, reliability and validity issues for construct measurement and advanced methodologies. The submissions selected for this volume cover some of these areas, while some of the topics remain open for future research. The book is organized in four parts; Part I presents a selection of papers that focus on the internationalization process of the ﬁrm. Part II combines studies with a small ﬁrm perspective and a focus on entrepreneurship. In Part III we offer three studies which examine value creation in strategic alliances, investigate reasons for international joint venture dissolution and ﬁnally a conceptual contribution on diversiﬁcation and the controversy within the literature which either suggests that diversiﬁcation results in a ﬁrm’s value premium or a ﬁrm’s value discount. The ﬁnal part in this edited book is devoted to methodological advancements. This part provides good examples of methodological challenges faced by today’s scholars and managers alike.

INTERNATIONALIZATION The ﬁrst chapter by Pieter Pauwels, Paul G. Patterson, Ko de Ruyter, and Martin Wetzels is entitled ‘‘The Propensity to Continue Internationalization: A Study of Australian Service Firms’’. The authors build on the process theory of internationalization and the theory of planned behavior and investigate a ﬁrm’s propensity to continue internationalization. They develop a theoretical model and test this using structural equation modeling using a sample of international service providers using partial least square (PLS). Their model conﬁrms the pivotal role of attitudes towards internationalization, relevant behavioral norms, and behavioral control factors as contributors to the propensity to continue internationalization. In the chapter ‘‘Factors Inﬂuencing the Internationalisation Process of UK Firms in Emerging Markets’’, Kannika Leelapanyalert explores the

Introduction

5

Marks & Spencer case and their entry into Hong Kong and Thailand. Building in rich interview and archival data, the author examines the contribution of ‘‘matching’’ in the ﬁrm internationalization. This concept, which builds and extends beyond the micro-level networking concept, apparently is very relevant in the context of ﬁrm internationalization and the achievement of a strong market position, speciﬁcally in emerging markets. Konstantinos Poulis and Mo Yamin present a case study within the tourism industry. In their examination of seven Greek ﬁrms with an existing export activity, they ﬁnd that international tourist encounters within their domestic home market helped them as a leverage to internationalize, expand into formally unknown markets, and learn about international markets in a kind of ‘‘follow the client’’ approach.

ENTREPRENEURSHIP AND FUNCTIONAL ADVANCEMENTS In the ﬁrst chapter, Sanna Sundqvist and Olli Kuivalainen empirically explore the relationships between international growth orientation, environmental turbulence, and strategic orientations. In their study of 783 Finnish exporters they ﬁnd that ﬁve different international growth orientations (IGO) exist. There are ﬁrms which emphasize export proﬁts, exporters highlighting export volume and proﬁtability, ﬁrms with no special growth focus, exporters growing by entering new foreign markets, and ﬁnally exporters emphasizing volume. Signiﬁcant differences can be detected across environmental factors and strategic orientations among different IGO patterns. In the second chapter, ‘‘Internationalization and Technological Innovation: Empirical Evidence on Their Reciprocal Relationship’’, Diana A. Filipescu, Josep Rialp, and Alex Rialp analyze internationalization and technological innovation through a multiple-case-study method. They propose a conceptual framework of reciprocal relationships between technological innovation and ﬁrm internationalization. Firms in this study, innovated more in products than in processes, and their technological innovations were more of incremental than radical nature. Additionally, the reciprocal relationship was reinforced in their empirical results, with technological innovation being an important ingredient for internationalization as well as being an international helping in the innovation process.

6

RUDOLF R. SINKOVICS AND PERVEZ N. GHAURI

Arild Aspelund and Tage Koed Madsen, in their chapter on ‘‘The Role of Innovative and Entrepreneurial Behavior in Internationalization Processes’’ present two strands of literature, the innovation-related internationalization model and the international entrepreneurship studies. They explore how these concepts have contributed to our understanding of internationalization processes of ﬁrms. Their conceptual deliberations also offer suggestions of how knowledge on innovative and entrepreneurial behavior from early research can be transferred to contemporary international marketing issues. Laura Salciuviene, Pervez N. Ghauri, Audra I. Mockaitis, Claudio De Mattos study brand image perceptions across cultures and examine consumer values on brand image perceptions in the mobile telephone market. Based on data from Lithuania and Denmark, they conclude that consumers perceive brand image differently depending upon their cognitive characteristics. While Danish consumers determine the value of brand based on its functional and symbolic image, the Lithuanian consumers perceive the brand value based only on its symbolic image. The chapter provides useful implications for managers.

COLLABORATIVE RELATIONSHIPS In this part the ﬁrst chapter, Chia-Ling ‘‘Eunice’’ Liu asks the question, ‘‘Do Firms Create Value Through Strategic Alliances?’’ investigates factors inﬂuencing the relationship between knowledge acquisition and capability enhancement. She develops a conceptual model and links alliance research, organizational learning, and network theory. The author then employs a structural equation model and examines a cross-section of 160 ﬁrms from the Taiwanese IT industry. Both antecedents of knowledge acquisition and its effect on ﬁrm’s capability and network position enhancement are tested. Her work is interesting in that while previous research on alliance learning has typically focused on knowledge acquisition per se, this study highlights the role of knowledge application in an intra-organizational context. Firms enhance their competences not only by acquiring knowledge but also integrating knowledge from alliance partners into the ﬁrm’s organizational routines. The second chapter by Mehmet Berk Talay and M. Billur Akdeniz examines causes for break-ups of ﬁrms and factors driving international joint venture dissolution. In their empirical analysis of a large sample of 3,038 IJVs in 63 different countries over a 21-year time period, they examine conceptualized antecedents of the dissolution of IJVs and reveal positive

7

Introduction

signiﬁcant effects of cultural distance and market potential as well as the negative signiﬁcant effects of equally shared ownership on the likelihood of IJV dissolution. Interestingly, their results reveal that competition between partners does not have a signiﬁcant effect on IJV dissolution. The third chapter in this part by Seyda Deligonul, ‘‘Geographic Diversiﬁcation: A Premium or Discount in Firm’s Value’’ extends the scope of marketing function to multiple geographical locations and comes up with two contradictory conclusions, that the ﬁrm strategy may result in (1) value premium, or (2) value discount. It discusses a number of controversies and ‘‘errors’’ in the existing literature and calls for an analysis with appreciation of organizational skills, resources, and knowledge assets that can be transferred across geographies to improve performance. It thus highlights the relationship between diversiﬁcation and performance and suggests that idiosyncratic aspects of different locations and business units must not be overlooked.

METHODOLOGICAL ADVANCES The ﬁrst chapter in this part, ‘‘The Use of Partial Least Squares Modeling in International Marketing’’, by Jo¨rg Henseler, Christian M. Ringle, and Rudolf R. Sinkovics reviews the PLS applications in the international marketing domain and comments on strengths and weaknesses of PLS. The authors provide guidance on this technique for international marketing work, discuss formative and reﬂective measurement differences and offer insights into PLS and multi-group analysis. PLS appears well suited for certain research situations and with the advancement of software such as SmartPLS, the methodology offers its strengths in a much more accessible way. The intriguing chapter by Thomas Salzberger, Hartmut H. Holzmu¨ller, and Anne Souchon, ‘‘A New Perspective of Construct Validity and CrossNational Comparability’’, explores export information use and its crosscultural equivalence in a ﬁve-country study (United Kingdom, United States, New Zealand, Austria, and Germany). Not only they provide an impressive data set, but they also offer a highly promising methodology, based on item response theory (IRT) and the Rasch model of measurement. The Rasch model, named after the Danish mathematician Georg Rasch, is an exciting framework which lends itself particularly well to the investigation of measurement equivalence, because it strongly emphasizes invariance. The analysis shows partial invariance for export information use, allowing

8

RUDOLF R. SINKOVICS AND PERVEZ N. GHAURI

for valid comparisons within and across cultures and offering a basis for the assessment of individual companies as well. The third chapter by R. Glenn Richey, Daniel G. Bachrach, Michael G. Harvey, and Hui Wang is called ‘‘The Dynamics of Technological Readiness in Marketing Units: Why Cross-Cultural Examination is Necessary’’. The authors follow recent literature and embrace the view that technological readiness in marketing is a key driver to operational efﬁciency, speciﬁcally for international supply chain and marketing managers. Yet, they contend that technological readiness will be a function of task interdependence and unit-level performance. To examine this notion, they conduct an experiment in the United States and the People’s Republic of China and measure ratings of technological readiness from a total of 383 evaluators. Their results via cross-cultural experimental examinations support a three-way interaction effect of task interdependence, unit performance, and culture on technological readiness estimations. They thus provide insights into dynamic differences in the management of units in both the United States and Peoples Republic of China. The ﬁnal chapter addresses the question: How methodologically rigorous is the empirical research reported in a major journal – the Journal of Marketing? This study makes a comprehensive effort at characterizing the strengths and weaknesses in the preparation, analysis, and reporting of empirical research by marketing scholars. As an example of a statistical technique, Deligonul and his colleagues investigate practices associated with the use of Ordinary Least Squares (OLS) regression – by far the most widely-used technique reported in the Journal of Marketing. The study explores trends in the use of OLS regression over a decade and a half, and also seeks to identify patterns associated with single versus multiple authors. An overall objective is to delineate the extent to which consensus exists in the use of a popular statistical technique by leading marketing scholars. This is important because consensus on the means of scholarship signals the degree of maturity of a ﬁeld.

ACKNOWLEDGMENTS Special thanks go to the Advances in International Marketing editor-in-chief, S. Tamer Cavusgil whose support for this initiative has helped to make this project happen. Furthermore, we are indebted to colleagues who were prepared to invest time and effort to review the chapters submitted to this special volume. Their contribution has been vital to the development of the

9

Introduction

chapters and to the overall coherence of this special volume. The reviewers are listed in alphabetical order. Billur Akdeniz, Michigan State University, USA Nigel Barrett, University of Technology, Sydney, Australia Talay Berk, Michigan State University, USA Francisco Jose´ Molina Castillo, University of Murcia, Spain Ulf Elg, Lund University, Sweden Georg Fassott, Technical University Kaiserslautern, Germany Richard Fletcher, University of Western Sydney, Australia Arne Floh, Wirtschaftsuniversita¨t Vienna, Austria Alexandra Ganglmair, University of Otago, New Zealand David Grifﬁth, Michigan State University, USA Anna Jonsson, Lund University, Sweden Tim Kiessling, Bond University, Australia Gary Knight, Florida State University, USA Glenn R. Richey, University of Alabama, USA Sami Saarenketo, Lappeenranta University of Technology, Finland Arnold Schuh, Wirtschaftsuniversita¨t Vienna, Austria Per Servais, University of Southern Denmark, Denmark Carl-Arthur Solberg, BI Norwegian School of Management, Norway Carlos Sousa, University College Dublin, Ireland Lawrence Welch, Melbourne Business School, Australia Mo Yamin, Manchester Business School, UK Attila Yaprak, Wayne State University, USA Anna Zueva, University of Bradford, UK

REFERENCES Aulakh, P. S. (Ed.) (2007). Editorial: Emerging multinationals from developing economies: Motivations, paths and performance. Journal of International Management, 13(3), 235–240. 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. Doh, J. P. (2005). Offshore outsourcing: Implications for international business and strategic management theory and practice. Journal of Management Studies, 42(3), 695–704. Dunning, J. H. (2006). Comment on dragon multinationals: New players in 21st century globalization. Asia Paciﬁc Journal of Management, 23(2), 139–141. Giddens, A. (2000). The Runaway World Debate [Online]. Available at http://www.lse.ac.uk/ Giddens/RWDdefault.htm (accessed on January 20, 2005). Giddens, A. (2002). Runaway world: How globalisation is reshaping our lives. London: Proﬁle Books.

10

RUDOLF R. SINKOVICS AND PERVEZ N. GHAURI

Jean, R.-J. ‘‘Bryan’’ (2007). The ambiguous relationship of ICT and organizational performance: A literature review. Critical Perspectives on International Business, 3(4), 306–321. Levitt, T. (1983). The globalization of markets. Harvard Business Review, 61(3), 92–102. Lewin, A. Y., & Peeters, C. (2006). Offshoring work: Business hype or the onset of fundamental transformation? Long Range Planning, 39(3), 221–239. Miller, D., Hartwick, J., & Le Breton-Miller, I. (2004). How to detect a management fad – and distinguish it from a classic. Business Horizons, 47(4), 7–16.

THE PROPENSITY TO CONTINUE INTERNATIONALIZATION: A STUDY OF ENTREPRENEURIAL DECISION-MAKING IN AUSTRALIAN SERVICE FIRMS Pieter Pauwels, Paul G. Patterson, Ko de Ruyter and Martin Wetzels Although behavioral in nature, internationalization process literature has long ignored the impact of strategic management and managerial decisionmaking. Internationalization has been explained as an incremental process of experiential learning that builds upon the reciprocal relationship of a ﬁrm’s growing international market commitment and its increasing foreign market knowledge (Johanson & Vahlne, 1977). Yet, ‘‘once it has started, [this process] will tend to proceed regardless of whether strategic decisions in that direction are made or not’’ (Johanson & Vahlne, 1990, p. 12). More than a decade later, however, an emergent perspective emphasizes entrepreneurial behavior as a new and complementary explanatory platform for internationalization (Knight, 1997; McDougall & Oviatt, 2000; Zahra, 2005). International entrepreneurship highlights the dominant role of innovative, proactive, and risk-seeking behavior in searching, selecting, entering, and

New Challenges to International Marketing Advances in International Marketing, Volume 20, 11–36 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020004

11

12

PIETER PAUWELS ET AL.

managing cross-border markets (Covin & Slevin, 1991; Jones & Coviello, 2005). Human volition, decisions, and mainly actions have become the units and focus of analysis (Jones & Coviello, 2005). To date, this approach has contributed considerably to the explanation of, for instance, apparently untypical internationalization processes of ‘‘Born Globals’’/‘‘International New Ventures’’ (Jones & Coviello, 2005; Knight & Cavusgil, 2004; McDougall & Oviatt, 2000). The empirical focus of international entrepreneurship literature is on the evidence of internationalization behavior: country selections and choices of entry modes, both with a speciﬁc timing in a continuous process of change (Le´vesque & Shepherd, 2004). However, managerial decision-making is not investigated as such it ‘‘may be determined post hoc from interpretation of internationalization behavior patterns and proﬁles’’ (Jones & Coviello, 2005, p. 16). The analysis of event-patterns and behavioral proﬁles is indeed a ﬁrst and important analytic step. We argue, though, that the explanation of internationalization on the basis of behavioral outcomes is incomplete without a close investigation of the antecedents of managerial choice and behavior in the internationalization process (Ucbasaran, Westhead, & Wright, 2001). Therefore, the further integration of international entrepreneurship in internationalization process theory requires a better understanding of managerial information-processing and (entrepreneurial) choice-making in the context of the internationalization process. Towards this goal, the present chapter focuses on the relationship between managerial attitudes and behaviorial intentions (i.e., the propensity) to (continue) internationalization. Extant literature on entrepreneurial decision-making legitimizes this focus on managerial attitudes, as it indicates that entrepreneurs are more susceptible to the use of decision-making heuristics (Brockhaus, 1980; Busenitz & Barney, 1997; Manimala, 1992; Mitchell et al., 2004). Lacking other information, an entrepreneur tends to rely on his/her own overall attitude to the general issue (i.e., internationalization) as an anchor while making decisions and is inclined to adjust perception and interpretation of new information to this attitudinal anchor (Slovic, Fischhoff, & Lichtenstein, 1977; Tversky & Kahneman, 1974). Hereafter, we argue that the overall attitude towards internationalization is a critical antecedent to behavioral outcomes in the internationalization process. Therefore, this study contributes to the literature as it formalizes this attitudinal perspective, and identiﬁes its sources and effect on the intentions towards internationalization. This results in a model of internationalization propensity that is rooted in the classic attitude-based psychological Theory of Planned Behavior (TPB) (Ajzen, 1991).

The Propensity to Continue Internationalization

13

The model is operationalized and tested on a sample of 193 Australian internationalizing service ﬁrms. This choice for service ﬁrms is waranted as cross-border marketing and delivery of services is lagging behind. Over the period 2001–2006, the global export of commercial services has increased by a stable average of 11% per year to an estimated $2.71 trillion in 2006 (WTO, 2007). While this is a signiﬁcant growth, commercial services’ exports expanded less rapidly than global cross-border merchandise trade – 15% per year to $11.76 trillion in 2006 (WTO, 2007). The Australian service sector is no exception to this trend.

1. DECISION-MAKING IN THE INTERNATIONALIZATION PROCESS To investigate a ﬁrm’s propensity to continue internationalization, the socalled Uppsala internationalization process model is a logical point of departure (Johanson & Vahlne, 1977, 1990). Building upon a behavioral theory of the ﬁrm (Cyert & March, 1963) and Penrose’s (1959) theory of the growth of the ﬁrm, the basic logic of the Uppsala or U-model is quite straightforward: The allocation of resources to foreign activities holds a certain risk yet induces experiential learning, which results in market-speciﬁc knowledge. The increasing stock of market-speciﬁc knowledge reduces this risk and stimulates additional allocation of resources (Eriksson, Johanson, Majkga˚rd, & Sharma, 1997). The U-model builds upon four (implicit) assumptions. First, ﬁrms maximize the expected future rent of their resources by allocating these to markets where doing business is judged to be least risky (Andersen, 1993). Second, risk reduces only through increasing market-speciﬁc knowledge. Third, ﬁrms acquire new market-speciﬁc knowledge through experiential learning (Johanson & Vahlne, 1977). Fourth, experiential learning and the reciprocal relationship of accumulating knowledge and market commitment are efﬁcient processes. Internationalization processes would not embark and/ or continue if experiential learning were an inefﬁcient process, i.e., when the cost of experiential learning would outweigh the beneﬁts reaped from increasing market commitment. In sum, the U-model explains internationalization as an incremental and risk adverse process that is solely driven by an efﬁcient experiential learning process (Johanson & Vahlne, 1977; Pedersen & Petersen, 1998; Forsgren, 2002). Since its conception in the mid-1970s, the impact of the U-model on international business research has only increased, both in product and service

14

PIETER PAUWELS ET AL.

contexts. To date, the U-model remains a popular theoretical – almost paradigmatic – point of reference (Johanson & Vahlne, 2006). Nevertheless, the model and its underlying logic have been challenged on theoretical (e.g., Andersen, 1993; Hadjikhani, 1997; Bjo¨rkman & Forsgren, 2000) as well as on empirical grounds (e.g., Sullivan & Bauerschmidt, 1990; Millington & Bayliss, 1990). For instance, apparently it cannot explain observed ﬁrm behavior such as leapfrogging, discontinuity, and (temporal) withdrawal (Benito & Welch, 1997; Pedersen & Petersen, 1998). One of the main grounds for critique is that the U-model actually fails to cope with managerial discretion that goes beyond the optimization of the experiential learning process (Buckley, 1996). As Johanson and Vahlne (1990) conclude: ‘‘The [U-Model] is rather skeptical in regard to strategy. Nevertheless, we think that internationalization processes are the result of a mixture of strategic thinking, strategic action, emergent developments, chance and necessity. We believe it is worthwhile to analyze the internationalization of ﬁrms with an open mind with regard to these factors’’ (Johanson & Vahlne, 1990, p. 22). Therefore, a focus on decision-making in the context of the internationalization process is warranted. Strategic decisions do not only affect the current course of a ﬁrm in international waters, they also outline if not deﬁne the learning process that underlies a long-term internationalization process (e.g., Bell, 1995; Chetty & Campbell-Hunt, 2004; Johanson & Vahlne, 1990; Larimo, 2001). 1.1. Internationalization through Heuristics The internationalization process of ﬁrm deﬁnes an enter- or intrapreneurial setting that is characterized by a sense of urgency, lack of information, and relatively high perceived risk (Knight & Cavusgil, 2004; Johanson & Vahlne, 1990). In this challenging environment, managers have to decide on investments, strategies, and priorities in overseas ventures. Preliminary ﬁndings on these decisions demonstrate their outspoken entrepreneurial nature. More in particular, decision related to the internationalization process seems to rely heavily on simplifying heuristics (Chetty & Campbell-Hunt, 2004; Etemad & Lee, 2003; Ibeh, 2003; Jones & Coviello, 2005; Le´vesque & Shepherd, 2004). Because of (perceived) lack of information, percieved risk, and a sense of urgency, fully elaborated decision-making may simply be impossible (Ellis & Pecotich, 2001). In this situation, an entrepreneur may still come to a decision based on heuristics whereas a non-entrepreneurial stance may obstruct any decision at all (Brockhaus, 1980; Busenitz & Barney, 1997; Manimala, 1992; Mitchell et al., 2004).

The Propensity to Continue Internationalization

15

Studying heuristics, Slovic, Fischhoff, and Lichtenstein (1977) and Einhorn and Hogarth (1981), among others, have shown that entrepreneurial decisionmakers typically manifest overconﬁdence, which mainly emanates from anchoring and adjustment (Hogarth, 1987; Schwenk, 1988). Lacking other information, the entrepreneur tends to rely on his/her own overall attitude towards the general issue (i.e., internationalization) as a pivotal anchor while making decisions and is inclined to adjust perception and interpretation of new information to this anchor (Slovic et al., 1977; Tversky & Kahneman, 1974). To understand a ﬁrm’s propensity to continue internationalization in the context of extant internationalization process theory, it is therefore imperative to focus on a decision-maker’s overall attitude towards internationalization. In the following sections, we elaborate on the attitudinal component of the U-model and accommodate entrepreneurial decisionmaking to explain the propensity to continue internationalization.

1.2. Attitudinal Commitment in the U-Model In the U-model, market commitment is composed of two factors – the amount of resources committed and the degree of commitment (Johanson & Vahlne, 1977). The amount of resources points at the economic factors allocated to a particular market. Market commitment increases when a ﬁrm allocates more inputs to manage and support foreign operations. The degree of commitment represents the difﬁculty of ﬁnding alternative uses for these resources (Andersen, 1993). Hadjikhani (1997) extends this conceptualization and argues that the amount of commitment points at the short-term tangible (i.e., ﬁnancial and institutional) forms of commitment, whereas the degree of commitment captures more the intangible long-term elements of commitment. Focusing on tangible and intangible commitment as separate constructs does not refute the U-model’s logic, yet promises a more dynamic perspective on internationalization. For instance, it could help explaining market withdrawal (decreasing tangible commitment) as a decision of product/ market portfolio optimization in the context of a progressing (increasing intangible commitment) internationalization process (Douglas & Craig, 1996; Pauwels & Matthyssens, 1999). Hadjikhani’s (1997) two-dimensional conceptualization of market commitment corresponds with an established understanding of commitment that distinguishes between attitudinal commitment and behavioral commitment (Meyer & Allen, 1991; Gundlach, Achrol, & Mentzer, 1995). Doing so, we upgrade the theoretical role of managers’ attitude and discretion in the

16

PIETER PAUWELS ET AL.

U-model. This paves the way for the explanation of the propensity to continue internationalization within the logic of the U-model.

2. THE PROPENSITY TO CONTINUE INTERNATIONALIZATION: A MODEL In line with the aforementioned logic, we argue that the propensity to continue internationalization is the outcome of a heuristics-based decisionmaking process that highly relies on the decision-makers’ attitude towards internationalization. To model and test this statement, we build upon classic TPB (Ajzen, 1988, 1991; Fishbein & Ajzen, 1975). This theory proposes that behavior is determined by behavioral intention as an outcome of attitudes, subjective norms, and perceived behavioral control (Ajzen, 1991; Ajzen & Fishbein, 1980). On its turn, the attitude towards a behavioral intention is determinent by salient beliefs about the consequences of the behavior multiplied by the evaluation of the consequences. This attitude equation represents an information-processing view of how attitudes are formed and change with the inﬂuence of external stimuli on the structure of the outcome beliefs (Ajzen & Fishbein, 1980). The more one beliefs that performing the behavior will lead to desired outcomes, the more favorable the attitude will be. As a second antecedent of behavioral intention, subjective norms are determined by normative beliefs. The latter rely on expectations of certain reference groups, and the motivation to comply with these expectations (Fishbein & Ajzen, 1975). Thirdly, perceived behavioral control captures the perception of ease or difﬁculty of performing this behavior (Ajzen, 1991). Behavioral intention is strongly inﬂuenced by the conﬁdence or belief in the ability to perform a behavior (Bandura, 1991). In line with TPB, we hypothesize that the overall attitude towards internationalization is a central antecedent to the propensity to continue internationalization. Hence, H1. The propensity to continue internationalization is positively related to the overall attitude towards internationalization. This attitude captures the subjective1 desire to continue internationalization and develops from salient beliefs: outcome beliefs and normative beliefs (Fishbein & Ajzen, 1975). Referring to Patterson, Cicic, and Shoham (1997) and Wang and Olsen (2002), it is expected that the disconﬁrmation of expectations of the outcomes of internationalization is a prime antecedent of

The Propensity to Continue Internationalization

17

a ﬁrm’s attitude and, eventually, of its intention to continue internationalization. Active internationalizing ﬁrms have at least implicit expectations about what beneﬁts internationalization might bring. These expectations shape the outcome beliefs as they serve as a benchmark against which subsequent performance is compared and (dis)conﬁrmed. Negative (positive) disconﬁrmation occurs when actual performance is perceived as falling short of (exceeding) expectations. Current performance that lags behind expectations mitigates future outcome beliefs. To the contrary, current performance that surpasses expectations strengthens future outcome beliefs. This disconﬁrmation process feeds the overall attitude towards a particular behavior (Anderson, 1973; Oliver, 1980; Oliver & DeSarbo, 1988). Hence, Hypothesis 2 summarizes the effect of outcome beliefs: H2. The overall attitude towards internationalization is positively related to the disconﬁrmation of expectations about past internationalization performance. Next to outcome beliefs, the overall attitude to internationalization builds on normative beliefs, which reﬂect a feeling of intrinsic obligation (Meyer & Allen, 1991). We conceive a structural and a strategic notion of these normative beliefs. The structural notion encompasses the degree of organizational institutionalization of a ﬁrm’s internationalization process (Selznick, 1957; Boeker, 1989; Sullivan, 1994). The more the internationalization process is ingrained in the organization, the more positive the overall attitude towards internationalization. The intensity of internationalization is suggested as a proxy variable to conﬁne the structural part of this normative belief. The strategic component relates to the perceived ﬁt between the strategic option to continue internationalization and the ﬁrm’s corporate strategy and goals. The better the internationalization route and its related consequences ﬁt with the corporate course, the more positive the attitude towards internationalization (Nelson & Winter, 1982; Boeker, 1989; Hamel & Prahalad, 1989). Accordingly, we propose: H3a. The overall attitude towards internationalization is positively related to the intensity of internationalization. H3b. The overall attitude towards internationalization is positively related to the perceived ﬁt between internationalization and the corporate strategy. In line with TPB, the overall attitude towards internationalization is complemented with subjective norms (Fishbein & Ajzen, 1975). Jointly, they affect the propensity to continue internationalization. In fact, these

18

PIETER PAUWELS ET AL.

subjective norms represent the acceptability of the intended behavior given current organizational routines, procedures, and structures and, as such, captures, organizational and strategic inertia (e.g., Kelly & Amburgey, 1991; Ross & Staw, 1993). Organizational inertia is conceived as the institutionalization of past behavior. It is expected that the mere intensity of the ﬁrm’s current international program contributes to a subjective norm that affects a ﬁrm’s intention to continue internationalization. Likewise, strategic ﬁt represents strategic inertia to affect behavioral intention to continue internationalization. Hypotheses 4a and 4b summarize the expected impact of the subjective norm: H4a. The propensity to continue service internationalization is positively related to the current intensity of internationalization. H4b. The propensity to continue service internationalization is positively related to the perceived ﬁt between internationalization and the corporate strategy. Consequently, the ﬁrm’s current internationalization intensity and the perceived strategic ﬁt play a double role in the model. First, they act as (proxies of ) normative belief structures that shape the attitude towards internationalization. Second, they capture the subjective organizational and behavioral norms to continue internationalization. This conceptualization is in line with Davis, Bagozzi, and Warshaw (1989), among others, who model normative belief structures to affect the intermediate attitudinal construct as well as behavioral intention directly. Apart from the overall attitude towards internationalization and extant subjective norms, the propensity to continue internationalization also depends upon perceived behavioral control (Ajzen, 1991). In the present study, we focus on both internal and market inhibitors to represent this behavioral control. This way, we capture a ﬁrm’s perception of the feasibility of internationalization in a certain business context. If one perceives high barriers within the ﬁrm (e.g., limited resources or a lack of relevant capabilities) or in the foreign market (e.g., level of competition), the propensity to continue internationalization is expected to be attenuated. Managers may claim that they are very much in favor of internationalization (i.e., a positive attitude), yet forgo the opportunity to commit resources to this strategic option because they perceive it to be unfeasible. In line with Sheppard, Hartwick, and Warshaw (1988), we model internal and market inhibitors as moderators that affect the relationship between attitude and the two normative constructs on the one hand, and the ﬁrm’s propensity on

The Propensity to Continue Internationalization

19

the other. Hence, Hypothesis 5 presents the expected impact of behavioral control as follows: H5a,b,c. The higher (lower) the degree of perceived internal and market inhibitors, the weaker (stronger) the direct impact of (a) intensity, (b) attitude, and (c) perceived strategic ﬁt on the propensity to continue internationalization. To complete and to further contextualize the model, we acknowledge the moderating effect of personal interaction with foreign customers during product/service delivery on the development of an attitude. In line with perception–knowledge–belief theory (Dretske, 2000), it is expected that a higher degree of personal interaction during product/service delivery enhances the relationship between salient beliefs and the emerging attitude. The more cross-border delivery is mediated by domestic or foreign channels, the less the internationalizing ﬁrm has personal contacts with its local market, the less salient beliefs impact upon the overall attitude towards internationalization. These interactions provide a higher perceived sense of expertness, which inﬂates the perceived truthfullness of extant beliefs. Accordingly, we propose: H6a,b,c. The higher (lower) the degree of personal interaction with the local market, the stronger (weaker) the impact of (a) intensity, (b) disconﬁrmation of expectations, and (c) perceived strategic ﬁt on the overall attitude towards internationalization. Fig. 1 summarizes the model graphically. It presents a contextualized understanding of a ﬁrm’s propensity to continue internationalization. The model builds upon TPB (Ajzen, 1991). It complements the U-model as it conceives the propensity to continue internationalization as the outcome of an entrepreneurial decision that highly relies upon simplifying heuristics with a pivotal role for the overall attitude (i.e., attitudinal commitment in the terminology of the U-model) towards internationalization (Slovic et al., 1977; Tversky & Kahneman, 1974).

3. RESEARCH DESIGN To test the hypotheses related to our model (Fig. 1), we studied data from 193 Australian service ﬁrms currently involved in international operations. Although in literature the internationalization of service ﬁrms has been studied largely apart from product ﬁrms (Buckley, Pass, & Prescott, 1991;

20

PIETER PAUWELS ET AL.

Personal Interaction Internationalization Intensity

H6a,b,c Internal Inhibitors

H3a H4a

Disconfirmation Of Expectations

H2

Propensity to continue Internationalization

Attitude to Internationalization H1

H4b H3b

H5a,b,c

H5a,b,c Market Inhibitors

Strategic Fit

Fig. 1.

The Propensity to Continue Internationalization.

Winsted & Patterson, 1998), we see no particular reason why the present model would not be valid in both a service and product setting. The underlying logic does not touch upon particularities of service or product markets. The Australian service sectors provide an ideal context for a study such as this. It has produced thriving service economies (especially professional services such as engineering, business consultancy, legal, and medical ﬁrms). Moreover, this sector has been especially aggressive in capitalizing on overseas market opportunities, especially in the Asia-Paciﬁc region. Today, the Australian exports of commercial services amounts to US$32 billion in 2006, with an average yearly increase of about 6%. However, as in the rest of the world, the dramatic increase of cross-border service delivery is lagging. Indeed, Australian exports of merchandized goods leveled off at US$123 billion in 2006 with an average yearly increase of 16% over the last six years (WTO, 2007). We employed a three-stage research design. The ﬁrst stage consisted of 15 semi-structured in-depth interviews with service ﬁrms involved in international operations. The purpose of this qualitative phase was twofold. First, we aimed at understanding how the antecedents of a ﬁrm’s propensity to continue internationalization were interrelated in practice. Second, the interviews provided a glimpse on the decision-making process that underlies

The Propensity to Continue Internationalization

21

this model. While this knowledge enhanced the interpretation of the analytical results, a discussion of the process elements of this model goes beyond the ambition of this chapter. Stage two involved a cross-sectional survey of service organizations. We mailed pre-tested, self-administered questionnaires to a non-probability sample of 1,264 organizations. The questionnaire was directed to the senior executive ofﬁcer responsible for the organization’s internationalization process or, in case no such ofﬁcer existed, to the senior marketing person. After one follow-up reminder letter and the incentive of receiving an executive summary of the ﬁndings, 347 usable questionnaires were returned, 183 were unusable due to incorrect address or incomplete information, and 67 reported they were only involved in marketing goods. The remaining 667 questionnaires were not returned and could not be traced. Of the 347 usable replies, 154 ﬁrms were considering export as strategy for growth, but were currently not operational across borders. After excluding these, the usable response was 193 service export cases, which means a response rate of 34.2%.2 An analysis of the questionnaires returned from a second wave mailing were compared with the ﬁrst wave, as recommended by Armstrong and Overton (1977), and revealed no statistically signiﬁcant differences (a ¼ 0.05) on any of the measures. Stage three of the analysis was organized after the statistical analysis of the survey. We asked experienced top managers of four internationally operating service ﬁrms to comment extensively on the outcomes of the survey and to assess alternative interpretations. This stage signiﬁcantly helped us to upgrade the managerial logic of the discussion section as well as the identiﬁcation of remaining study topics.

3.1. Measures Table 1 gives an overview of the indicators and scales used in the questionnaire. The intensity of internationalization is measured using a formative ratio of international to total sales (Sullivan, 1994). The measurement of the disconﬁrmation of expectations construct relies upon Anderson (1973), Oliver (1980), and Oliver and DeSarbo (1988). Strategic ﬁt is measured following Van de Ven and Drazin (1985). Measures for the attitude towards internationalization and the propensity to continue internationalization build upon instruments discussed in Sheppard et al. (1988). The moderating variables personal interaction and internal and external inhibitors were measured by use of three or four self-constructed items. Except for

22

PIETER PAUWELS ET AL.

Table 1. Label (type) Intensity of internationalization (Open)

Disconﬁrmation of expectations on internationalization (5-point Likert type)

Strategic ﬁt between internationalization and corporate strategy (5-point Likert type)

Attitude towards internationalization (5-point Likert type)

Personal interaction with customer base (5-point Likert type) Internal inhibitors to internationalization (5-point Likert type) Market inhibitors to internationalization (5-point Likert type) Propensity to continue internationalization (5-point Likert type)

Measures. Indicators

% of annual sales from international operations most recent ﬁscal year % of annual sales from international operations past three ﬁnancial years Overall performance of internationalization compared to expectations Accrued beneﬁts from internationalization compared to expectations Problems in internationalization compared to expectations Internationalization is consistent with our organization’s mission Financial returns to the company are greater if we internationalize International operations are potentially more proﬁtable than domestic operations Our mind-set towards internationalization is high Commitment to internationalization is high Internationalization is a desirable path Degree of overall interaction Degree of face-to-face interaction Frequency of interaction Lack of expertise Inappropriate services Inadequate resources Difﬁculty of gaining entry Lack of contacts Underdeveloped markets Competition abroad Likelihood of continuing internationalization Possibility of continuing internationalization Chance of continuing internationalization

Anchor Points n/a

Much worse/ much better

Strongly disagree/ strongly agree

Strongly disagree/ strongly agree

High/low

Hinder to great extent/not hindering at all

Very unlikely/ very likely

The Propensity to Continue Internationalization

23

internationalization intensity, which is a construct with two formative indicators, we measured the remaining non-moderating latent variables with multiple reﬂective indicators. When the dependent and independent variables are obtained from a single key informant, common method bias is a potential problem. We have examined the extent of any potential bias using Harman’s one-factor test. In the unrotated factor structure, we identify ﬁve factors with an Eigenvalue greater than 1. We ﬁnd no evidence of a general factor. Those ﬁve factors explain 72% of the total variance and the ﬁrst factor accounts for less than the majority of the variance explained. Following Podsakoff and Organ (1986), common method variance does not appear to be a concern.

3.2. Method We used Partial Least Squares (PLS) Path Modeling to estimate both the measurement and structural parameters in our structural equation model (Chin, 1998; Wold, 1985). PLS Path Modeling is a component-based structural equation modeling (SEM) approach that does not require multivariate normal data, places minimum requirements on measurement levels, and is more suitable for small samples (Chin, 1998; Fornell & Bookstein, 1982; Wold, 1985). In this study, covariance-based SEM approaches (e.g., LISREL, EQS) are less appropriate due to the existence of a construct with formative indicators (intensity).3 PLS Path Modeling allows assessing the psychometric properties of the measurement instruments. More in particular, we can examine reliability, convergent validity, and discriminant validity (Chin, 1998; Fornell & Larcker, 1981). For the structural model, we use the R2 value for the endogenous latent variables as a measure of model ﬁt (Chin, 1998). Tenenhaus, Vinzi, Chatelin, and Lauro (2005) have developed a global ﬁt measure for PLS, goodness of ﬁt (GoF). GoF is deﬁned as the geometric mean of average communality and average R2 (0pGoFp1), or more formally: qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ GoF ¼ communality R2 (1) Since communality equals average variance extracted (AVE) in the PLS approach, we propose a cut-off value of 0.5 for communality as suggested by Fornell and Larcker (1981). Moreover, in line with the effect sizes for R2 (small: 0.02; medium: 0.13; large: 0.26) proposed by Cohen (1988), we derive

24

PIETER PAUWELS ET AL.

the following GoF criteria for small, medium, and large effect sizes of R2: 0.1, 0.25, and 0.36. To test for the signiﬁcant effects of the moderator variables, we established two groups for each moderator on the basis of a median split as suggested by Duxbury and Higgins (1991). Subsequently, we use PLS to obtain parameter estimates and bootstrapping to provide the standard errors of the estimates and the accompanying t statistics. The equality of the path coefﬁcients for the two groups can be tested for signiﬁcance using a t test (Duxbury & Higgins, 1991; Howell & Aviolo, 1993). A conservative value of 0.01 is used for the signiﬁcance level.

4. ANALYSIS This section discusses the analytical results in four steps. First, we assess the psychometric properties of the measures. Second, we present the outcomes of the PLS analysis of the model as speciﬁed in Fig. 1. Third, we discuss the results of the moderator analysis. To assess the psychometric properties of the measures, we test a measurement model without structural paths in PLS-Graph Version 3.0 (Chin, 2001), which is analogous to conﬁrmatory factor analysis in covariance-based SEM. Convergent validity can be evaluated by inspecting the factor loadings of the measures on their respective constructs (Chin, 1998; Hulland, 1999; Tenenhaus et al., 2005). If we assume a minimum loading of 0.7 item reliability will be approximately 0.5 (Hulland, 1999). Reliability of the measures was assessed using composite reliability (CR) and AVE. For CR a threshold value of 0.70 is suggested (Chin, 1998; Hulland, 1999; Nunnally & Bernstein, 1994), whereas for AVE, Fornell and Larcker (1981) propose a threshold value of 0.50. Discriminant validity can be assessed by examining the crossfactor loadings of the measures (Chin, 1998; Howell & Aviolo, 1993; Hulland, 1999). The measures should not exhibit any loadings substantial in magnitude on constructs for which we have not hypothesized so. Moreover, the correlations of residual terms (Y) across blocks should not exceed 0.20 in absolute terms (Falk & Miller, 1992; Fornell & Bookstein, 1982). Additionally, a construct should share more variance with its measures than it shares with other constructs in the model (Chin, 1998; Hulland, 1999). Consequently, the square root of the AVE should exceed the intercorrelations of the construct with the other constructs in the model. In our study, all reﬂective measures show factor loadings exceeding 0.7 and no factor cross-loadings substantial in magnitude. The between-block

25

The Propensity to Continue Internationalization

correlations of the residual terms (Y) do not exceed 0.2 in absolute terms. Moreover, none of the intercorrelations of the constructs exceeds the square root of the AVE for the constructs. To assess the reliability of the measures, we calculated the composite scale reliability and the AVE (Chin, 1998; Fornell & Larcker, 1981; Hulland, 1999). As far as composite scale reliability is concerned, all reﬂective measures exceed a value of 0.90 and for AVE all reﬂective measures are higher than 0.60. Moreover, the square root of the AVE exceeds the intercorrelations of the construct with the other constructs in the model indicating discriminant validity (Chin, 1998; Hulland, 1999). Table 2 summarizes the psychometric properties of the measures. Next, we estimated the parameters in the structural model using PLS analysis as implemented in PLS-Graph Version 3.0 (Chin, 2001) and tested the substantive relationships as laid down in the conceptual model. Fig. 2 summarizes the parameter estimates. All relationships are signiﬁcant at a ¼ 0.01 except for the relationship between internationalization intensity and the propensity to continue internationalization, which is weaker yet signiﬁcant at a ¼ 0.05. The ﬁndings show that the propensity to continue internationalization relates signiﬁcantly to the overall attitude towards internationalization (H1 supported), to the internationalization intensity (H4a supported), and to the perceived strategic ﬁt between internationalization and the corporate strategy (H4b supported). The overall attitude to internationalization, on its turn, signiﬁcantly relates to salient beliefs on the disconﬁrmation of performance expectations (H2 supported), to the internationalization intensity (H3a supported), and to salient beliefs about the congruence between internationalization and the corporate strategy (H3b supported). Since PLS estimation does not generate standard errors for the parameter estimates, we resort to resampling procedures (Chin, 1998; Efron & Tibshirani, Table 2. Composite Reliability and Intercorrelations of Latent Variables. # of Items Composite Reliability

1. 2. 3. 4. 5.

Intensity Disconﬁrm Strategic ﬁt Attitude Propensity

2 3 3 3 3

0.92 0.91 0.90 0.91 0.96

Intercorrelations of Latent Variables 1

2

3

4

5

0.92 0.24 0.40 0.49 0.38

0.88 0.22 0.45 0.26

0.79 0.63 0.57

0.85 0.59

0.95

26

PIETER PAUWELS ET AL.

Internationalization Intensity 0.14* 0.21**

Disconfirmation Of Expectations

0.30**

Attitude to Internationalization R2=0.54

0.48**

0.31**

Propensity to continue Internationalization R2=0.42

0.31**

Strategic Fit

Fig. 2.

PLS Parameters Estimates of the Model. Note: * po0.05; ** po0.01.

1993). To test the effects and the statistical signiﬁcance of the parameters in the structural model, a bootstrapping procedure with 250 and 500 resamples was used. The resulting t statistics for both the 250 and 500 resample revealed a consistent pattern. In general, the model shows a more than acceptable variance explained for attitude (R2 ¼ 0.54) as well as for internationalization propensity (R2 ¼ 0.42). As a point of reference, Sheppard et al. (1988) indicated that models of planned behavior explain between 20 and 30% of the variance in behavioral intention. Finally, Tenenhaus et al.’s (2005) GoF measure for this model is 0.58, which clearly exceeds the proposed cut-off value of 0.36. Furthermore, we compared our model with the nested, fully mediated model. In the latter, we assume that the propensity to continue internationalization is only explained by attitude towards internationalization and no direct relations between the normative antecedents (i.e., intensity and strategic ﬁt) and propensity are postulated. The R2 of propensity in the fully mediated model is signiﬁcantly lower than the R2 of propensity in the partially mediated model (R2partial ¼ 0:42 vs. R2full ¼ 0:34; F2,189 ¼ 13.03, po0.001). This further supports the validity of our model. Table 3 indicates the standardized path coefﬁcients of the moderators as presented in Fig. 1. The moderator effects on the relationship between attitude and propensity (ATT-PROP) are signiﬁcant and attenuate or

27

The Propensity to Continue Internationalization

Table 3.

Results of Moderator Analysis a.

Relationships

Unmoderated Moderated Internal inhibitors Market inhibitors

Unmoderated Moderated Personal interaction

Standardized Path Coefﬁcients

Low (n ¼ 101) High (n ¼ 92) Low (n ¼ 114) High (n ¼ 79)

Low (n ¼ 86) High (n ¼ 50)

ATT-PROP

INT-PROP

FIT-PROP

0.31**

0.14*

0.31**

0.48** 0.20** 0.40** 0.19*

0.11 0.18** 0.32** 0.03

0.22* 0.38** 0.25** 0.44**

INT-ATT

DIS-ATT

FIT-ATT

0.21**

0.30**

0.48**

0.26* 0.40**

0.30** 0.41**

0.47** 0.27*

* po0.05. ** po0.01. a All t tests for the moderator analyses are signiﬁcant at ao0.01.

strengthen as expected. Compared to the non-moderated model, the relationship between attitude and propensity becomes signiﬁcantly stronger in case of low internal inhibitors (bIL ¼ 0.48 vs. b ¼ 0.31; po0.01) and signiﬁcantly weaker in case of high internal inhibitors (bIH ¼ 0.20 vs. b ¼ 0.31; po0.01). The same counts for the effect of market inhibitors on the relationship between attitude and propensity. These results support Hypothesis 5b, which postulated that the higher (lower) the degree of perceived internal and market inhibitors, the weaker (stronger) the direct impact of the attitude towards internationalization on the propensity to continue internationalization. As indicated in Table 3, the moderator effects on the relationship between intensity and propensity (INT-PROP) are insigniﬁcant for low internal and high market inhibitors and are opposite to the proposed direction in the case of internal inhibitors, albeit insigniﬁcantly. As such, these ﬁndings do not support Hypothesis 5a which stated that higher (lower) internal and external inhibitors attenuate (strengthen) the impact of intensity on propensity. The moderator effects on the relationship between strategic ﬁt and propensity (FIT-PROP) are all signiﬁcant, but opposite to the expected direction. This ﬁnding is interesting and will be addressed in the discussion section of this chapter. Nevertheless, the analytical results do not support Hypothesis 5c.

28

PIETER PAUWELS ET AL.

Finally, the lower part of Table 3 presents the results of the moderator effect of the degree of personal interaction with customers during service delivery (Hypothesis 6). Concerning the impact on the relationships between intensity and attitude (INT-ATT) and between disconﬁrmation and attitude (DIS-ATT), the results are signiﬁcant and support Hypotheses 6a and 6b, respectively. With respect to the impact on the relationship between strategic ﬁt and attitude (FIT-ATT), Table 3 shows signiﬁcant yet unexpected results. Contrary to what was postulated in Hypothesis 6c, a higher (lower) degree of personal interaction attenuates (strengthens) the impact of strategic ﬁt on attitude. Therefore, Hypothesis 6c cannot be supported. In sum, although some of the hypotheses related to the moderating variables cannot be supported, overall the moderators have a signiﬁcant impact on the model.

5. DISCUSSION The ﬁndings largely conﬁrm the validity of our model. All hypotheses related to the core model as well as the majority of the hypotheses related to the moderator effects are supported by the data. This way, the present study conﬁrms the viability of the TPB (Ajzen, 1991) in the context of the internationalization process. As expected, the propensity to continue internationalization is partially explained by the overall attitude towards internationalization, which, on its turn, relies on decision-makers’ normative and outcome beliefs. Complementary to the overall attitude towards internationalization, the propensity to continue is also enhanced by structural and strategic inertial forces. Both the core model and the moderating effects shed light on the entrepreneurial nature of managerial decisions in the course of the internationalization process. The relatively high variance explained supports the idea that managerial decision-making in this context is largely heuristicsbased, with the overall attitude to internationalization as a central anchor (Hogarth, 1987; Schwenk, 1988). Even some unexpected ﬁndings in the moderator analysis shore up this conclusion. Contrary to what was expected in Hypothesis 5c, higher internal and market inhibitors seem to strengthen the relationship between strategic ﬁt and propensity and, again contrary to what was put forward in Hypothesis 6c, the analysis shows that less personal interaction during service delivery strengthens the impact of strategic ﬁt on attitude. Although unexpected, both ﬁndings may possibly be explained in line with the comment of one respondent collected in the qualitative third stage our empirical study: ‘‘As long as we can play our own game, [ . . . ] the

The Propensity to Continue Internationalization

29

more barriers we see, the more we go for it. If we can cross the river in our way, we might be alone on the other side.’’ This interpretation concurs with Hamel and Prahalad (1989) who argue that strategic ﬁt materializes more strongly in case of perceived obstruction. The unexpected yet signiﬁcant ﬁndings on Hypotheses 5c and 6c point at apparent risk-proneness in decision-making on the internationalization process (Hogarth, 1987; Schwenk, 1988). More in particular, it can be expected that enterpreneurs are willing to bear risk with respect to their ability uncertain in new international ventures. It is, in cases of perceived strategic ﬁt, enterpreneurial decision-makers are expected to be overconﬁdent with respect to their abilities (March & Shapira, 1987; Wu & Knott, 2006). More analysis is required to investigate this interpretation. In all, the ﬁndings characterize the entrepreneurial and semi-volitional nature of the internationalization process. The decision-by-heuristics to continue internationalization is complemented with normative (i.e., inertial) antecedents outside the scope of entrepreneurial discretion. Both the institutionalization of the internationalization process and the perceived ﬁt between the internationalization course and the corporate strategy have a direct effect on the ﬁrm’s propensity to continue internationalization. The central role of managerial attitude warrant the integration of our model in the Uppsala model and, at the same time, upgrades the latter’s entrepreneurial perspective (Chetty & Campbell-Hunt, 2004; Johanson & Vahlne, 1990).

6. MANAGERIAL IMPLICATIONS This study is among the ﬁrst to unveil the entrepreneurial nature of decisionmaking in the internationalization process. More in particular, it highlights the pivotal role of the decision-maker’s attitude towards internationalization. Given its importance, it is crucial to understand how this attitude is formed and changes. We learned that the decision-makers’ attitude emerges from highly interpretative – and therefore subjective – insights on the ﬁt of internationalization with the strategic course of the ﬁrm. Even more challenging is the relevance of the disconﬁrmation of earlier expectations as an antecedent of this attitude. Apparently irrelevant and outdated expectations on the outcomes of earlier decisions do affect managers’ attitude that guides future decision-making. Finally, we show that accumulating commitment to internationalization (i.e., represented by the current intensity) increases perceived structural inertia on the future of internationalization.

30

PIETER PAUWELS ET AL.

In all, the nature of these antecedents highlight the subjectivity, instability, and fragility of decision-makers’ attitude towards internationalization – the main mental anchor to ground decisions on the future of the internationalization process of the ﬁrm. Accordingly, this attitude might not be shared equally within the organization. Disagreement on the future international course of a ﬁrm may be highly related to discrepancies that exist among decision-makers on the general attitude towards internationalization. At the same time, perceived strategic and structural inertia on the internationalization process may dramatically reduce the perceived degrees of freedom on future internationalization decisions. This calls for managerial efforts in building shared interpretations on past internationalization performance, on the strategic relevance of internationalization, as well as on the perceived level of current commitment to internationalization. In absence, managerial decision-making on the future of internationalization may be troublesome and full of conﬂict. At the same time, management should be aware of increasing blindness that emerges from a decision-by-heuristics process as well as from inertial structural and strategic forces. We suggest avoiding amnesia by regularly inviting outsiders into this decision-making process; preferably lower level personnel that operates at the outer limits of the ﬁrm. Needless to say, it is challenging to accommodate this need for stability in the perceptions on the antecedent of the overall attitude with a need to ﬁght inertial forces.

7. CONCLUSIONS Based upon Ajzen’s (1991) TPB, we model the propensity to continue internationalization as an outcome of the overall attitude towards internationalization, relevant behavioral norms, and behavioral control factors. The ﬁndings support the pivotal role of the overall attitude towards internationalization, which largely depends upon the disconﬁrmation of expectations on the outcome of earlier internationalization decisions as well as on normative beliefs that internationalization ﬁts the strategy and current course of the organization. These ﬁndings emphasize the volitional nature of the model. Complementary, the data demonstrate the explanatory power of structural and strategic inertial factors – in disregard of managerial discretion. The importance of the institutionalization of internationalization as well as of the perceived ﬁt between the internationalization process and the corporate strategy emphasize the evolutionary nature of the model.

The Propensity to Continue Internationalization

31

The ﬁndings imply that the propensity to continue internationalization – and by extension the internationalization process – is neither the result of an extensive decision-making process, as implicitly assumed in some earlier work on internationalization (e.g., Young, Hamill, Wheeler, & Davis, 1989), nor is it the outcome of a purely endogenous process as assumed in the (more dominant) U-model (Johanson & Vahlne, 1977, 1990). To the extent that ongoing internationalization is the result of managerial discretion, the ﬁndings support the idea that internationalization results from compensatory interactions among belief structures, which shape the overall attitude towards internationalization. This general attitude serves as a pivotal anchor for entrepreneurial decision-making on the basis of simplifying heuristics (Kahneman, Slovic, & Tversky, 1982; Krabuanrat & Phelps, 1998).

7.1. Limitations and Future Research The present study has limitations that should be taken into consideration when judging its merits. First, we used cross-sectional data which prevent from testing causal relationships. More in-depth case study research would allow a more profound investigation of the decision-making heuristics. Indeed, the current model presumes rather than tests the existence of these heuristics. Second, the generalizability of the study remains restricted as our sample was limited to Australian service providers. While our model does not include service-speciﬁc variables that may hinder the interpretation and validity in a manufacturing goods business, replication of the model in different settings remains necessary. Finally, although a promising interpretation is provided, the unexpected ﬁndings for two of the hypotheses indicate that the current model may be improved when risk propensity and risk perception are included explicitly (Busenitz, 1999; Forlani & Mullins, 2000; Sitkin & Weingart, 1995). The ﬁndings of the present study invite future researchers to investigate the relative impact of and interaction between discretionary and non-discretionary antecedents of internationalization in high- and low-risk environments.

NOTES 1. The Theory of Planned Behavior (Ajzen, 1988, 1991) focuses on the individual person as the level of analysis. In the presents study, we abstract from internal consultation and political processes that might precede an executive manager’s

32

PIETER PAUWELS ET AL.

intention to continue internationalization. As such, we simplify the internationalization process and related decision-making as if it were the intended action of one individual. 2. To assess whether this sample is large enough, we take into account the effect size, signiﬁcance level (a) and desired level of power (1b). Based on Cohen (1988), a medium effect size ( f 2 ¼ 0.15) would require a sample of 76 cases. A large effect size ( f 2 ¼ 0.35) requires a sample size of 34 (assuming a ¼ 0.05 and (1b) ¼ 0.80). Using the approach suggested by Green (1991) under the same assumptions, a medium effect size ( f 2 ¼ 0.15) would require a sample size of 73 and 31 for a large effect size ( f 2 ¼ 0.35). As generally medium effect sizes are adopted for behavioral research (cf. Cohen, 1988), our sample size appears to be adequate for this study, even allowing for moderator analysis using a multigroup approach. 3. In fact, to ensure model identiﬁcation in the presence of a construct with formative indicators in covariance-based SEM, two requirements are to be satisﬁed (Jarvis, MacKenzie, & Podsakoff, 2003; MacCallum & Browne, 1993). The ﬁrst concerns establishing the scale of measurement for the construct by either ﬁxing one of the paths leading from the indicators to the construct to one or equivalently by constraining the disturbance term of the construct to one. The second requests a solution to the factor indeterminacy of the disturbance term of the construct with the formative indicators. To achieve this, the construct with the formative indicators needs to emit at least two paths to constructs with formative indicators or to contain at least two reﬂective indicators (or a mix of both). Since our conceptual model cannot accommodate for these requirements, the use of PLS Path Modeling seems imperative.

REFERENCES Ajzen, I. (1988). Attitudes, personality and behavior. Chicago: Dorsey Press. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–211. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall. Andersen, O. (1993). On the internationalization process of the ﬁrm: A critical analysis. Journal of International Business Studies, 24(2), 209–231. Anderson, R. E. (1973). The effect of disconﬁrmed expectancy on perceived product performance. Journal of Marketing Research, 10, 38–44. Armstrong, J. S., & Overton, T. S. (1977). Estimating non-response bias in mail surveys. Journal of Marketing Research, 14, 396–402. Bandura, A. (1991). Social-cognitive theory of self-regulation. Organizational Behavior and Human Decision Processes, 50, 258–287. Bell, J. (1995). The internationalization of small computer software ﬁrms: A further challenge to ‘stage’ theories. European Journal of Marketing, 29(8), 60–75. Benito, G. R. G., & Welch, L. S. (1997). De-internationalization. Management International Review, 2(Special Issue), 7–25. Bjo¨rkman, I., & Forsgren, M. (2000). Nordic international business research a review of its development. International Studies of Management and Organization, 30(1), 6–25.

The Propensity to Continue Internationalization

33

Boeker, W. (1989). The development and institutionalization of subunit power in organizations. Administrative Science Quarterly, 34, 388–410. Brockhaus, R. H. (1980). Risk taking propensity of entrepreneurs. Academy of Management Journal, 23, 509–520. Buckley, P. J. (1996). The role of management in international business theory: A meta-analysis and integration of the literature on international business and international management. Management International Review, 36(1 Special Issue), 7–54. Buckley, P. J., Pass, C. L., & Prescott, K. (1991). The internationalization of service ﬁrms: A comparison with the manufacturing sector. Scandinavian International Business Review, 1(1), 39–56. Busenitz, L. W. (1999). Entrepreneurial risk and strategic decision making. The Journal of Applied Behavioral Science, 35(3), 325–340. Busenitz, L. W., & Barney, J. B. (1997). Differences between entrepreneurs and managers in large organizations: Biases and heuristics in strategic decision-making. Journal of Business Venturing, 12, 9–30. Chetty, S., & Campbell-Hunt, C. (2004). A strategic approach to internationalization: A traditional versus a ‘‘Born-Global’’ approach. Journal of International Marketing, 12(1), 57–81. Chin, W. (1998). The partial least squares approach to structural equation modeling. In: G. A. Marcoulides (Ed.), Modern business research methods (pp. 295–336). Mahwah, NJ: Lawrence Erlbaum Associates. Chin, W. (2001). PLS-Graph User’s Guide Version 3.0. Houston, TX: C.T. Bauer College of Business, University of Houston. Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Covin, J. G., & Slevin, D. P. (1991). A conceptual model of entrepreneurship as ﬁrm behavior. Entrepreneurship Theory and Practice, 16(1), 7–25. Cyert, R. M., & March, J. G. (1963). A behavioral theory of the ﬁrm. Englewood Cliffs, NJ: Prentice Hall. Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982–1003. Douglas, S. P., & Craig, C. S. (1996). Executive insights: Global portfolio planning and market interconnectedness. Journal of International Marketing, 4(1), 93–110. Dretske, F. (2000). Perception, knowledge and belief. Cambridge, UK: Cambridge University Press. Duxbury, L. E., & Higgins, C. A. (1991). Gender differences in work-family conﬂict. Journal of Applied Psychology, 76(1), 60–74. Efron, B., & Tibshirani, R. J. (1993). An introduction to the bootstrap. New York: Chapman and Hall. Einhorn, H. J., & Hogarth, R. M. (1981). Behavioral decision theory: Processes of judgment and choice. Journal of Accounting Research, 19(1), 1–31. Ellis, P., & Pecotich, A. (2001). Social factors inﬂuencing export initiation in small and mediumsized enterprises. Journal of Marketing Research, 38, 119–130. Eriksson, K., Johanson, J., Majkga˚rd, A., & Sharma, D. D. (1997). Experiential knowledge and cost in the internationalization process. Journal of International Business Studies, 28(Second Quarter), 337–360. Etemad, H., & Lee, Y. (2003). The knowledge network of international entrepreneurship: Theory and evidence. Small Business Economics, 20, 5–23.

34

PIETER PAUWELS ET AL.

Falk, R. F., & Miller, N. B. (1992). A primer for soft modeling. Akron, OH: University of Akron Press. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Forlani, D., & Mullins, J. W. (2000). Percieved risks and choices in enterpreneurs. New Venture Decisions. Journal of Business Venturing, 15, 305–322. Fornell, C., & Bookstein, F. L. (1982). Two structural equation models: LISREL and PLS applied to consumer exit-voice theory. Journal of Marketing Research, 19(4), 440–453. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. Forsgren, M. (2002). The concept of learning in the uppsala internationalization process model: A critical review. International Business Review, 11, 257–277. Green, S. B. (1991). How many subjects does it take to do a regression analysis? Multivariate Behavioral Research, 26(3), 499–510. Gundlach, G. T., Achrol, R. S., & Mentzer, J. T. (1995). The structure of commitment in exchange. Journal of Marketing, 59(January), 78–92. Hadjikhani, A. (1997). A note on the criticisms against the internationalization process model. Management International Review, 37(2 Special Issue), 43–66. Hamel, G., & Prahalad, C. K. (1989). Strategic intent. Harvard Business Review, 67(3), 63–76. Hogarth, R. M. (1987). Judgement and choice: The psychology of decisions. New York: Wiley. Howell, J. M., & Aviolo, B. J. (1993). Transformational leadership, transactional leadership, locus of control, and support for innovation: Key predictors of consolidated-businessunit-performance. Journal of Applied Psychology, 78(6), 891–902. Hulland, J. (1999). Use of Partial Least Squares (PLS) in strategic management research: A review of four recent studies. Strategic Management Journal, 20, 195–204. Ibeh, K. I. N. (2003). Toward a contingency framework of export entrepreneurship: Conceptualisations and empirical evidence. Small Business Economics, 20, 49–68. 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 commitment. Journal of International Business Studies, 8(1), 23–32. Johanson, J., & Vahlne, J.-E. (1990). The mechanism of internationalization. International Marketing Review, 7(4), 11–24. Johanson, J., & Vahlne, J.-E. (2006). Commitment and opportunity development in the internationalization process: A note on the uppsala internationalization process model. Management International Review, 46(2), 165–178. Jones, M. V., & Coviello, N. E. (2005). Internationalisation: Conceptualising an entrepreneurial process of behavior in time. Journal of International Business Studies, 36(3), 284–303. Kahneman, D., Slovic, P., & Tversky, A. (1982). Judgement under uncertainty: Heuristics and biases. Cambridge, UK: Cambridge University Press. Kelly, D., & Amburgey, T. L. (1991). Organizational inertia and momentum: A dynamic model of strategic change. Academy of Management Journal, 34(3), 591–612. Knight, G. A. (1997). Emerging paradigm for international marketing: The born global ﬁrm. PhD Dissertation, Michigan State University, East Lansing, MI.

The Propensity to Continue Internationalization

35

Knight, G. A., & Cavusgil, S. T. (2004). Innovation, organizational capabilities, and the bornglobal ﬁrm. Journal of International Business Studies, 35, 124–141. Krabuanrat, K., & Phelps, R. (1998). Heuristics and rationality in strategic decision making: An exploratory study. Journal of Business Research, 41, 83–93. Larimo, J. (2001). Internationalization of SMEs – two case studies of Finnish born global ﬁrms. Paper presented at the CIMaR Annual Conference, Sydney. Le´vesque, M., & Shepherd, D. A. (2004). ‘Entrepreneurs’ choice of entry strategy in emerging and developed markets. Journal of Business Venturing, 19, 29–54. MacCallum, R. C., & Browne, M. W. (1993). The use of causal indicators in covariance structure models: Some practical issues. Psychological Bulletin, 114(3), 533–541. Manimala, M. J. (1992). Entrepreneurial heuristics: A comparison between high PI (PioneeringInnovative) and Low PI ventures. Journal of Business Venturing, 7, 477–504. March, J. G., & Shapira, Z. (1987). Managerial perspectives and risk taking. Management Science, 33(11), 1404–1418. McDougall, P. P., & Oviatt, B. M. (2000). International entrepreneurship: The intersection of two research paths. Academy of Management Journal, 43(5), 902–906. Meyer, J. P., & Allen, N. J. (1991). A three component conceptualization of organizational commitment. Human Resource Management Review, 1(1), 61–89. Millington, A. I., & Bayliss, B. T. (1990). The process of internationalization: UK companies in the EC. Management International Review, 30(l), 151–161. Mitchell, R. K., Busenitz, L., Lant, T., McDougall, P. P., Morse, E. A., & Smith, J. B. (2004). Toward a theory of entrepreneurial cognition: Rethinking the people side of entrepreneurial research. Entrepreneurship theory and practice, 27(1), 93–104. Nelson, R. R., & Winter, S. G. (1982). An evolutionary theory of economic change. Cambridge, MA: The Belknap Press of Harvard University Press. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill. Oliver, R. L. (1980). A cognitive model of the antecedents and consequences of satisfaction decisions. Journal of Marketing Research, 17, 460–469. Oliver, R. L., & DeSarbo, W. S. (1988). Response determinants in satisfaction judgments. Journal of Consumer Research, 14, 495–508. Patterson, P. G., Cicic, M., & Shoham, A. (1997). A temporal sequence model of satisfaction and export intentions of service ﬁrms. Journal of Global Marketing, 10(4), 23–43. Pauwels, P., & Matthyssens, P. (1999). A strategy process perspective on export withdrawal. Journal of International Marketing, 7(3), 10–37. Pedersen, T., & Petersen, B. (1998). Explaining gradually increasing resource commitment to a foreign market. International Business Review, 7, 483–501. Penrose, E. (1959). The theory of growth of the ﬁrm. New York: Wiley. Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12(4), 531–544. Ross, J., & Staw, B. M. (1993). Organizational escalation and exit: Lessons from the shoreham nuclear power plant. Academy of Management Journal, 36(4), 701–732. Schwenk, C. R. (1988). The cognitive perspective on strategic decision making. Journal of Management Studies, 25(1), 41–55. Selznick, P. (1957). Leadership in administration. Evanston, IL: Row, Peterson. Sheppard, B. H., Hartwick, J., & Warshaw, P. R. (1988). The theory of reasoned action: A meta-analysis of past research with recommendations for modiﬁcations and future research. Journal of Consumer Research, 15, 325–343.

36

PIETER PAUWELS ET AL.

Sitkin, S. B., & Weingart, L. R. (1995). Determinants of risky decision-making behavior: A test of the mediating role of risk perceptions and propensity. Academy of Management Journal, 38(6), 1573–1592. Slovic, P., Fischhoff, B., & Lichtenstein, S. (1977). Behavioral decision theory. Annual Review of Psychology, 28, 1–39. Sullivan, D. (1994). Measuring the degree of internationalization of a ﬁrm. Journal of International Business Studies, 25(Second Quarter), 325–342. Sullivan, D., & Bauerschmidt, A. (1990). Incremental international integration: A test of Johanson and Vahlne’s thesis. Management International Review, 30(1), 19–30. Tenenhaus, M., Vinzi, V. E., Chatelin, Y.-M., & Lauro, C. (2005). PLS path modeling. Computational Statistics & Data Analysis, 48, 159–205. Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185, 1124–1131. Ucbasaran, D., Westhead, P., & Wright, M. (2001). The focus of entrepreneurial research: Contextual and process issues. Entrepreneurship Theory and Practice, 25(3), 57–80. Van de Ven, A. H., & Drazin, R. (1985). The concept of ﬁt in contingency theory. Greenwich, CT: JAI Press. Wang, G., & Olsen, J. E. (2002). Knowledge, performance, and exporter satisfaction: An exploratory study. Journal of Global Marketing, 15(3/4), 39–64. Winsted, K. F., & Patterson, P. G. (1998). Internationalization of services: The service exporting decision. Journal of Services Marketing, 12(4), 294–311. Wold, H. (1985). Partial least squares. In: S. Kotz & N. L. Johnson (Eds), Encyclopedia of Statistical Sciences (Vol. 6, pp. 581–591). New York: Wiley. WTO. (2007). ‘Risks lie ahead following stronger trade in 2006. WTO reports (http://www. wto.org/english/news_e/pres07_e/pr472_e.htm) retrieved on November 12. Wu, B., & Knott, A. M. (2006). Entrepreneurial risk and market entry. Management Science, 52(9), 1315–1330. Young, S., Hamill, J., Wheeler, C., & Davis, J. R. (1989). International market entry and development strategies and management. Prentice Hall: Englewood Cliffs, NJ. Zahra, S. A. (2005). A theory of international new ventures: A decade of research. Journal of International Business Studies, 36, 20–28.

FACTORS INFLUENCING THE INTERNATIONALISATION PROCESS OF UK FIRMS IN ASIA Kannika Leelapanyalert A number of studies have been performed on the internationalisation of manufacturing and industrial product ﬁrms (Dunning, 1981; Johanson & Vahlne, 1977; Buckley & Ghauri, 1999; Ford, 1990). Many of these studies use network approach and relationship marketing as the starting point (Ha˚kansson & Johanson, 2001; Ha˚kansson & Snehota, 1989; Johanson & Mattsson, 1988; Sheth, 1995), while others use marketing strategies and management of unfamiliar environments through internal and external factors (Cavusgil, Ghauri, & Agarwal, 2002; McGoldrick, 2002; Root, 1992). Although some earlier studies have incorporated political actors in the network (Hadjikhani & Ghauri, 2001), these studies often deal with the management of political environments or with lobbying activities of multinational enterprises (Lenway & Murtha, 1994; Murtha, 1991). This study, however, will provide a better insight into factors that inﬂuence the internationalisation process of ﬁrms that deal directly with consumers in a foreign market and how these ﬁrms manage this process to achieve a solid foothold in a new market. This study explains the management of the internationalisation process through behavioural theory (Cyert & March, 1963). We endeavour to extend the network theory partly by extending it to retail ﬁrms and partly by bringing in a wider network of relationships that inﬂuence international activities of a ﬁrm. There is a need for New Challenges to International Marketing Advances in International Marketing, Volume 20, 37–67 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020005

37

38

KANNIKA LEELAPANYALERT

a fundamental understanding of the internationalisation process of the retail industry. Previous international retailing literature provides the allinclusive conceptualisation of the process models (Akehurst & Alexander, 1995; Alexander & Myers, 2000; Vida & Fairhurst, 1998). However, there are few retail studies which focus on networking and the matching activities. The focus is on the retail sector, especially during the retail internationalisation process. Previous matching concept studies focus on international expansion to Baltic States (Ghauri & Holstius, 1996), and to Eastern European countries (Rogers, Ghauri, & George, 2005), however, less on retail ﬁrms in developing countries in Asia. The existing gap in the literature on which this research aims to focus, is the study of networking and matching concepts during the international expansion of the retail sector, mainly in the Asian emerging market. This research seeks to explore this area in detail in order to develop a conceptual framework, which could provide useful insights and a valid approach to the various processes of retail internationalisation. The project aims to determine and analyse the factors that are crucial for successful establishment. This research studies the overall internationalisation process. The reasons for developing a speciﬁc framework for retail internationalisation are that retailers have been aggressively pursuing international expansion in the past decades, retail is the sector which has contact with customers directly and it provides product and service directly to its customers. The reason why this research focuses on matching and networking is because retailers do not produce anything by themselves, but subcontract the work out to their suppliers to produce the goods. Hence, the retailers rely heavily on their suppliers. It is interesting to study how retailers use matching/networking with their suppliers and stakeholders. Retail is also the industry which faces customers directly so it is interesting to study its relation to the consumer. It is not only necessary to understand customers’ needs but also to understand the risks, which are beyond ﬁrm’s control (Ghauri, Elg, & Sinkovics, 2004; Rogers et al., 2005). The main research questions in this research are presented below: What roles do networking and matching play during the internationalisation process of Western ﬁrms in emerging markets? How can dissimilarities in market environments be managed during the internationalisation process? How could different adaptation levels affect performance in different countries?

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

39

Based on the literature on the internationalisation process and entry strategies (Buckley & Casson, 1976; Buckley & Ghauri, 1999; Cavusgil, 1990; Cavusgil & Robert, 1990; Cavusgil et al., 2002; Root, 1987), networking (Johanson & Mattsson, 1987; Ha˚kansson & Snehota, 1989; Elg & Johansson, 1996; Johanson & Vahlne, 2006) and matching (Ghauri & Holstius, 1996; Holstius, 1991), a framework is developed to study the research questions. It is assumed that networking could be affected by matching factors and market characteristics, allowing retailers to better understand consumers and facilitate business activities, leading to a better performance.

RESEARCH BACKGROUND AND THEORETICAL FRAMEWORK Internationalisation leads to a radical process of change through which an organisation modiﬁes the focus of its operations, value system and cognitive framework so as to achieve a more internationally responsive structure (Whitehead, 1992). Earlier studies have investigated the internationalisation process considering the internal and external factors of the company and its market (Cavusgil et al., 2002; McGoldrick, 1998, 2002; McGoldrick & Davies, 1995; Treadgold & Davies, 1988). The internationalisation literature based on Uppsala studies about the internationalisation process of Swedish ﬁrms, (Johanson & Wiedersheim-Paul, 1975; Johanson & Vahlne, 1977) identiﬁed four different internationalisation stages called the ‘‘establishment chain’’, which also applied to the retail context (Davies & Fergusson, 1995, p. 99). The results show that different stages demand different resource commitment from the company (Johanson & Vahlne, 1977). It is assumed that a company initially lacks knowledge of the local market. The level of local market knowledge affects the company’s commitment decisions and its activities. The network approach (Johanson & Mattsson, 1988) could be applied to the study of vertical international relationships. It is, therefore, useful for the international sourcing activities of retailers (Dawson, 1994, p. 270) and provides a competitive advantage as well as ﬂexibility, in which each of the organisations in the network is working towards a common objective (McGoldrick, 2002, p. 571). In the past 20 years, many researchers have paid more attention to network relationships, which have become the main marketing strategy. They agree that the study of network relationships between companies, suppliers and customers is more important than the marketing mix (Ghauri, 1999).

40

KANNIKA LEELAPANYALERT

Akehurst and Alexander (1995) developed a framework of internationalisation of retailing by asking what, who, why, how and when retail internationalisation happened. Vida and Fairhurst (1998) suggested that decision-maker characteristics, ﬁrm characteristics and external environment are antecedent factors inﬂuencing retailers’ decisions on degree of involvement, mode of entry and market selection (Vida & Fairhurst, 1998). This research considers an important local market environment (Alexander & Myers, 2000; Vida & Fairhurst, 1998) and brings networking and matching activities to apply during the retail internationalisation process. Some of these studies, their theoretical underpinning and focal concerns are presented in Table 1. Matching allows an understanding of how the development of successful business relationships can be facilitated by initiating and developing relationships at different levels beyond ﬁrms’ immediate contacts (Ghauri & Holstius, 1996; Ghauri et al., 2004). Matching refers to activities of the ﬁrms and governments, where they try to match with each other to gain mutual beneﬁts. Companies try to match governments (home and host) to get preferential treatment to facilitate their entry into the particular market (Rogers et al., 2005). For instance, there might need to be an understanding between two countries to facilitate business between ﬁrms from each other’s country. According to Ghauri and Holstius (1996), there are three phases of a company’s international establishment process: the search phase, the project phase and the establishment phase. Actors, activities and resources play different roles in different stages. Actors relate to the companies involved in the relationship. The link existing between actors has an impact on the company’s proﬁt, cost control and strategic development. Activities concern the action taken to run company’s business and the effect it would have on each other. Resources involve company’s asset, skill, production output and capability (Ford, Gadde, & Ha˚kansson, 1998, p. 76). The search phase is about the company trying to get as much information as possible in that market and environment. It might be difﬁcult to get information from emerging market regarding undeveloped market research. The company builds on relationship with the local market in this stage (Ghauri & Holstius, 1996). The project phase is the stage in which the company creates network and relationship (Johanson & Mattsson, 1988). The company will analyse and make a decision about entering the new market and preparing to implement the project. The establishment phase is when the company understands its market and environment, and then the company begins its operation in that market.

Positive environment could be driving force but negative environment is obstacle for international growth

Relationship of two or more independent companies that take advantage of mutual resources and sharing of information

Business performance can be Performance involves three measured by two different key ﬁelds: Annual revenue approaches: judgmental/ growth, annual proﬁt subjective and objective growth and average return measurement on total assets. It is an evaluation of the perceived performance

Market environment

Networking approach

Performance

Areas of Concern

Judgmental perception is likely to enhance bias, due to the high degree of personal views involved

A network approach involves Company might focus three factors: Actors, only on short-term resources and activities beneﬁts and ignore long-term beneﬁts

Difﬁcult to implement in Processes that facilitate the terms of building trust development of business and commitment relationship between ﬁrms in dissimilar countries, at a global, macro and micro level Environmental factors (legal, Factors beyond cultural, political and company’s control economical factor) have impact on the companies’ activities and outcomes

Understanding of how to facilitate the development of successful business relationships

Matching

Focal Aspect

Underlying Logic

Alexander (1990), Dunne, Lusch, and Gable (1995), Ha˚kansson and Snehota (1989), Hollander (1970), Kacker (1985), Morganosky (1993), Salmon and Tordjman (1989), Sybrandy, Pirog, and Tuninga (1991), Treadgold (1990), Vida and Fairhurst (1998) and Williams (1992) Anderson, Ha˚kansson, and Johanson (1994), Coviello & McAuley (1999), Ford et al. (1998), Ghauri and Prasad (1995), Ha˚kansson and Snehota (1989), Johanson and Mattsson (1988) and Meyer and Skak (2002) Clark (2002), Day (1990), Dunne et al. (1995), Jaworski and Kohli (1993), Jennings and Young (1990), Katsikeas and Morgan (1998), Kohli and Jaworski (1990) and March and Sutton (1997)

Ghauri and Holstius (1996), Holstius (1991), Ghauri et al. (2004) and Rogers et al. (2005)

Examples of Studies

Theoretical Underpinnings to Develop Conceptual Framework.

Theory and Concept

Table 1.

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia 41

42

KANNIKA LEELAPANYALERT

Matching at the global level relates to international organisations involved in agreements, relationships and communication. It includes international activities that could assist the international operation among different countries (Holstius, 1991, p. 179). Institutions establish relationships through membership of global organisations, such as the World Trade Organisation (WTO) and International Bank of Reconstruction and Development (IBRD), and by working together to develop mutual beneﬁts for member countries. Matching at this level is beyond the control of the organisation; nevertheless, it still has an impact as organisations are subject to regulations resulting from international agreements. The IMF and the World Bank provided assistance to the Baltic States in 1992, so their ﬁnancial system illustrates matching at the global level. The European Union giving economic aid to each of the Baltic countries is another example of matching at the global level (Ghauri & Holstius, 1996). At the Macro level, matching relates to the level of support provided by governments towards business activities. Matching at these two levels, although outside the domain of the organisation, is orchestrated and planned by it. At the Micro level, matching refers to the activities the company needs to carry out in order to achieve a successful market entry. Matching at the micro level is closely related to networking. Johanson and Mattsson (1987) introduced a network approach to explain the international production process. The network approach looks at the inter-ﬁrm relationship, rather than the network itself. Also, it enhances the understanding of the strategies used in different markets. Different types of relationships exist between companies in the network; one is the member of a group or network. In contrast, a company also plays the role of the competitor to provide products to consumers. Companies in a network create competitive advantages by developing relationships with other companies. As long as each partner beneﬁts, the relationship will be pursued and grow stronger (Johanson & Vahlne, 2006). Ha˚kansson and Johanson (2001) mentioned the importance of personal contact and networking in market expansion. The buyer and seller networks might initially face a lack of knowledge about each other due to physical distance and culture differences. These gaps could be reduced by regular two-way communication and learning from each other, which lead to develop strong relationships and commitment. Matching is crucial, as internationalising companies’ problems are more pronounced in the earlier stages of the process. According to Ghauri and Holstius (1996), matching creates speciﬁc advantages for entering a new foreign market, as well as determining the process required to achieve a successful position after the establishment phase.

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

43

Market characteristics: The changing nature of the environment – legal factors, political forces, economic climate and cultural issues – ampliﬁes uncertainty in international retailing expansion (Salmon & Tordjman, 1989; Treadgold & Davies, 1988). The level of international expansion is increasing within emerging markets. As the European market is now saturated and has reached mature growth, many companies are expanding to developing countries throughout Asia. The nature of the developing countries, Asian economic growth and the level of high demand are attracting Western companies’ investment. The global market development and increasing sophistication of customer demand have made the world a more complex place (Brown & Burt, 1992). Consequently, ﬁrms have to develop strategies to respond quickly to different customers’ needs. Companies that are entering a foreign market have to take into consideration each of the environmental factors, as they have a potential impact on the companies’ activities and performance (Alexander, 1990; Kacker, 1985; Morganosky, 1993; Salmon & Tordjman, 1989; Treadgold, 1990; Vida & Fairhurst, 1998; Williams, 1992). Several countries have strict legislations concerning some industries such as retailing (Butler, 2007). For instance, there could be limitation on expansion, termination of employees, regulation about hours of operation and pricing. It is important to understand consumer perception and acceptance in order to determine an appropriate degree of standardisation or differentiation needed. The closer the culture between two countries, the greater the chance of success during retail expansion (Sternquist, 1988). This indicates that customers’ perception is very similar in countries with cultural proximity, requiring a relatively higher level of standardisation rather than a high level of adaptation. Business performance can be measured mainly by two different approaches: judgemental/subjective and objective measurement (Jaworski & Kohli, 1993). The ﬁrst one is viewed as a judgmental approach, which includes the view of all participants concerning the overall performance of the business and refers to an evaluation of the perceived performance (Jennings & Young, 1990). This subjective measurement is often adequate for research on a ﬁrm as a whole as well as research on Strategic Business Units (Narver & Slater, 1990). The second approach relates to the performance of the organisation in terms of objective measures. This includes the market share, market value, return on investment and proﬁtability (Ghauri et al., 2004; Kohli & Jaworski, 1990). This research aims at understanding how the performance of the business operations is affected by the company’s networking, matching and market characteristics. The above-mentioned studies of these variables have led us to a conceptual

44

KANNIKA LEELAPANYALERT

Matching • Global level • Macro level • Micro level

Networking Actors Performance Resources

Activities

1. Search Phase Market Characteristics • Legal • Culture • Political • Economic

Fig. 1.

2 Project Phase 3 Establishment Phase

Factors Inﬂuencing Internationalisation Process of UK Firms in Emerging Markets. Source: Based on Ghauri et al. (2004).

framework (Fig. 1) that we intend to use to analyse our studies. The research will focus mainly on subjective measures as the research has been conducted using qualitative methods. Nevertheless, objective measurement has also been looked at through secondary data such as sales and other performance indicators.

RESEARCH METHODOLOGY The methodological approach is based on qualitative research by using multiple case studies and interviews to allow an in-depth study of factors that inﬂuence ﬁrms while entering foreign markets. In-depth interview provides access to rich information. It allows the researcher to get a holistic picture of research problems (Ghauri, 2004). The case study approach has been selected because it provides a clear description of the management situation, which is suitable for theory development (Yin, 1994; Ghauri & Grønhaug, 2005). It is a suitable method to use when it concerns explaining how and why questions, when there is little control over events by the researcher, and when the emphasis is on a modern phenomenon within a real-life context (Yin, 1994). Case studies are also used in retailers and their internationalisation efforts in order to explore the issue in the speciﬁc organisation (Burt, Davies, Mcauley, & Sparks, 2002a; Burt, Davies, Mcauley, & Sparks, 2005). Case studies are suitable for this cross-cultural research; however, arguments are raised about the comparability of

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

45

information between different countries. In fact, interviews allow the researcher to understand the information in depth, and understand points that could not otherwise be explained through numbers and statistics (Sinkovics, Penz, & Ghauri, 2005). This method provides with an opportunity to ask further questions and check if all has been understood correctly. This method allows researcher to have a deeper understanding and to come across the issue of cultural differences (Ghauri, 2004). This research involves the study of a ﬁrm’s matching during the internationalisation process. We choose to study retailers because of the sector’s growth and aggressive international expansion. In addition, challenges faced in terms of local responsiveness, close customer contact requirements, localised decision-making and establishing relationships with stakeholders provides speciﬁc challenges. Marks & Spencer (M&S) has been selected as a suitable company to study, mainly due to its international activities in several markets as well as the importance of customer orientation and adaptation in this industry. M&S is a large organisation that has been established internationally for a long time. Also, the fact that the company has a record of both international successes and failures makes it an interesting company to study. Several multinational retailers are potential source of study but the access to the company is limited. The selection criteria were set towards large Western retailers that had invested in Asia. Thailand was chosen for study due to its rapid growth and market attractiveness and due to the researcher’s capability regarding the language and local market understanding. Hong Kong is interesting to study because of the very developed consumer markets in Asia. Comparative case studies between two countries have been conducted, as similar questions could be raised from each case and common outcomes could lead to draw conclusions (Miles & Huberman, 1994). This process allows a systematic comparison of the phenomenon studied within different cases and explores different dimensions of our research issues (Ghauri, 2004). Data is collected through various sources such as in-depth interviews, event attendance and observation, local ofﬁces and local stores visits. Fifteen interviews were conducted with M&S staff including senior managers, middle managers and frontline employees at the head ofﬁces in United Kingdom. Also, eight interviews were conducted with company representatives in Hong Kong and ﬁve in Thailand. All interviews have been tape recorded and fully transcribed. Some interviews have also been translated from Thai to English. As one of the researcher is a native speaker of Thai language as well as completely ﬂuent in English, the equivalence issues were properly dealt with before data collection and while translating

46

KANNIKA LEELAPANYALERT

the questions and answers. The data has been cross-checked with the secondary reports, media sources, company reports, brochures and newspaper articles. NVivo software is used to code the data and corroborate the analysis. The software is used to manage interview transcript, encode data and manage themes in order to analyse numerous qualitative data in more systematic way (Richards, 2000; Sinkovics et al., 2005). Triangulation occurred when data was collected in different ways, such as asking the same questions to different managers and cross checking the interviews with secondary data. By conducting case studies, it is intended to point out detailed, valuable and useful insights into the ﬁrm’s internationalisation process, which could not have been achieved by a survey method alone (Ghauri, 2004).

CASE STUDY 1: MARKS & SPENCER IN HONG KONG M&S is a leading UK clothing retailer. The company has diversiﬁed into food, housewares, beauty and ﬁnancial services in the United Kingdom. The company was established as partnership between Michael Marks and Tom Spencer in 1884. Company operates 219 stores managed under franchise in 34 countries including Europe, the Middle East, Asia, the Far East and Republic of Ireland, and 8 wholly owned stores in Hong Kong (Company Report, 2007). M&S entered to Hong Kong by franchise; Dodwell department store in 1980s. The company decided to establish their own subsidiary in 1988. In 1999, it became apparent that M&S’s global ambitions have failed and the company decided to return to its home market. By 2002, M&S has failed to sell its 10 Hong Kong stores to a local party and retained direct ownership of the stores. Despite the issues previously faced, M&S still operates wholly owned stores in Hong Kong. The business in Hong Kong is at a mature stage but considered as success in Hong Kong (Company Report, 2007). Customers perceive M&S as British, Western, traditional, reliable, good quality but high price. The positioning of M&S in Hong Kong is in the top-middle segment of the market.

Matching At the global level, the company does not seem to do matching with any international organisation. They do not have any section/people working on

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

47

these issues. ‘‘We don’t believe that we get any beneﬁt, support or inﬂuence from international organisations’’ (Asia Regional Manager). At the macro level, the relationship with the British Embassy is much more important than the government; however, M&S works closely with the Department of Trade and Industry (DTI). M&S has a strong network relationship with the DTI, British Consulate and British Embassy. M&S tend to work through the embassy’s trading ofﬁce, which understands and provides the sort of details on day-to-day information they need. The ﬁrst time M&S visited the country; they began to contact the DTI in the British embassy in each territory. The DTI could facilitate the collection of information on macro and micro economic, regulative or legislative environment including tariffs and duties, legal contracts, banking contracts, general knowledge of investment and background of consumers as well as the advice on local partners. The government of Hong Kong has a policy to stimulate the Hong Kong tourist department. Hong Kong Tourism Board tried to build Hong Kong to be the shopping destination in Asia as they offer tax-free shopping. Hong Kong government also introduced the new ‘‘individual visit’’ policy (IVS scheme) in July 2003, allowing the mainland Chinese tourists to enter Hong Kong easier. As a result, the total of 100 million mainland Chinese are eligible to travel to Hong Kong (Fig. 2). The total number of mainland

12000 10000

('000)

8000 6000 4000 2000 0 1997

1998

1999

2000

2001

2002

2003E

Year Mainland China

Fig. 2.

Others

Visitors Arrivals to Hong Kong. Source: Hong Kong Tourism Board (2003).

48

KANNIKA LEELAPANYALERT

tourists travelling under the IVS scheme has passed 6.6 million since the scheme was introduced. Mainland Chinese are the major tourists visitors, which represented 53.8% of total visitors in Hong Kong in 2006 (Hong Kong Tourism Board, 2006). This created a boost in the spending in the retail industry in Hong Kong (Hong Kong Trade Development Council, 2003). At the micro level, M&S is involved with different actors. The company gained information of the market from property companies and consultants. M&S also employed external consulting companies to conduct market research in the local market. It means that networking could facilitate the generation of market information. In terms of gathering information, external actors, suppliers and state institutions are used. The main market research was conducted to identify M&S customers and their needs. The recent market research conducted is about mainland China customers in Hong Kong, which has been seen as strong potential market for M&S products. M&S acknowledges the importance of supplier knowledge, as they are more experienced and specialised. For example, they discuss the trends in fashion, size ﬁtting and consumer needs with Triumph, the market leader in the lingerie sector in Asia. M&S also networks and is part of Hong Kong Retail Management Association (HKRMA). To generate information about the market environment, M&S is part of the retail management association so it could facilitate on retail environment information. The company also obtains market information from the local government statistics, secondary data, such as newspapers and databases. Landlords are the most important actor for M&S Hong Kong because it is very competitive to get a prime location. The company needs to reveal to the landlord that they perform well enough and maintain their brand image, therefore persuading the landlord to keep them in the mall. However, they need to balance so as not to reveal that they are doing too well, otherwise the landlord might increase the rent. The saying, Location Location Location has never been more accurate than here in Hong Kong . . . You need to be here in Tsim Sha Tsui, central, Admiralt, and Time Square. These 4 districts are the key selling points in Hong Kong. If you are in 4 locations, you got a chance, try to open in malls other than that is a risk. (Managing Director, M&S Hong Kong)

They maintain a relationship with the landlord by inviting them for business lunches, or to M&S events in order to show a positive brand image. M&S uses local managers to negotiate the rent with the landlord. As the UK

49

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

operation was not doing well in the home market, this affects the brand image in the international market. The HKRMA provides information, shares the information among the members and consults on business issues (HKRMA, 2004). The main discussion is to share factual information about the retail environment and give advice on current issues. Severe Acute Respiratory Syndrome (SARS), the disease that was widespread in Asia from March 2003 (Department of Health & Human Services, 2004), had a negative impact on the Hong Kong economy, leading to a drop in the level of tourists coming to Hong Kong in that year. As a result, all network members got together to negotiate a decrease in rent due to the outbreak. ‘‘Retail is an industry that was signiﬁcantly affected because few tourists were coming to Hong Kong. In Hong Kong people also did not go out for shopping so they negotiated with the landlord, asking to cut the lease rents’’ (Marketing Manager, M&S Hong Kong).

Performance

10000 8000 6000 4000 2000

HK$/sq.ft/p.a.

12000

1400 1200 1000 800 600 400 200 0 01/02

00/01

99/00

98/99

97/98

96/97

95/96

94/95

93/94

92/93

91/92

90/91

89/90

0 88/89

HK$'m

From 1988 to 1997, the company initially experienced fast growth and gradually increased its retail sales. The Asian crisis in 1998 affected the ﬁnancial performance dramatically (Fig. 3). The UK market was also facing the business recession period. The company considered whether to sell its Hong Kong subsidiary in 2001. However, by introducing a new strategy of cutting costs and improving its brand name, the company managed to become proﬁtable again and contribute signiﬁcantly to the operating proﬁt

Year Total Sales

Fig. 3.

RPF

The Total Sales and Return per Square Foot (RPT): M&S Hong Kong. Source: Company Report (2005).

50

KANNIKA LEELAPANYALERT

to the group in 2004. The 2004/2005 overall turnover of M&S international is d675.6 million which increased by 1.6% on the year (Company Report, 2005). M&S still has been a reactive company as their product trend often follows their competitors such as Hennes & Mauritz, Zara and French Connection. Due to problems in the domestic market, M&S changed their product design and introduced the ‘‘Per Una’’ womenswear range in order to attract a younger customer proﬁle. The ‘‘Per Una’’ merchandise has been developed to be more fashionable. The company promotes its brand image through advertisements on billboards, bus stop ads, fashion magazines and newspapers. Although, M&S continues to make a proﬁt in Hong Kong, the major challenge in the local market is to improve brand equity in order to keep the store in the prime locations. ‘‘M&S strategy is to focus on markets where there’s a strong demand for our products from a growing middle class’’ (Company Report, 2007). A few years after re-structuration, the international operating proﬁt increased from d60.7 million 2004/2005 to d87.5 million in 2006/2007. The wholly owned international business has increased by 14.0% in sales between 2006 and 2007 and operating proﬁt has increased to 26.8% reaching d45.4 million. Hong Kong contributed signiﬁcantly to this success (Company Report, 2007).

CASE STUDY 2: MARKS & SPENCER IN THAILAND M&S entered the Thai market in 1993 through franchising with the Central Department Store (CDS), a major retail player in Thailand. The CDS established 12 stores in Bangkok, Chiangmai, Haadyai and Phuket. In 2003, the store’s total revenue was over US $420 million (Company Brochure, 2005). M&S began to operate two of the department stores in 1994; six stores in 1997 and 10 stores in 2001 (Burt, Mellahi, Jackson, & Sparks, 2002b, p. 208). They currently operate 10 stores covering over 7,000 m2. The format of the M&S shop in Thailand vary from concession stands within CDS to stand-alone shops in the prime shopping centre in Bangkok.

Matching The franchisee does not do any matching at the global level. At the macro level, a network relationship with British Embassy and local government are useful as they provide information about the local market during the project

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

51

and the establishment phases. The company build their relationship with the local government ofﬁce through CDS, not only from the point of view of beneﬁting from the relationship but sometimes for obtaining a license to import, or when the trading ofﬁce demands a lot of information and take long process in order to register a product to sell, for example toiletries. At the micro level, M&S Thailand does not have an issue dealing with the landlord because the CDS owns majority of assets. The major actor involved the relationship between M&S and other brands in CDS, because they need to share resources with other brands in store. The meeting between different brands in the department store is useful to share information and best practices. M&S can get an idea and adapt into its brand. We use the intranet a lot. We put our information in there to share with other departments. For example we announce which product will be launched in this season and how will we launch it, what do we have in the catalogue and so on. (Buying Manager, M&S Thailand)

The M&S corporate level do share some information and discuss about local promotion in Asian markets, through emails, meeting at annual events, and some intranet. ‘‘We contact Singapore about what they are doing and we can learn from each other. When we have seen the best practice, we learn and bring it to improve ourselves. We contact each other in selling events and we also contact by email, for example, when we want to implement bra measurements’’ (Buying Manager, M&S Thailand ). M&S Thailand networks with customers by issuing loyalty cards so they have already established a system of collecting information about customers: ‘‘The Company improves the relationship with customers speciﬁcally by using a loyal customer base. We use marketing campaigns aiming at this group. We create activities for them and give special treatment to them, i.e. a gift set. We emphasise on that’’ (Marketing manager, M&S Thailand ).

Market Characteristics According to the Switzerland-based Institute for Management Development, the overall country competitiveness is high, ranked 27th in the world. The report showed the positive political environment by its ﬂexibility of policy towards business operation and efﬁcient policy achievement. Thailand also has low taxes: 7% consumption tax rates compared with the

52

KANNIKA LEELAPANYALERT

average of 17.17% worldwide (Nivatpumin, 2005). M&S experience difﬁculties with the food and drugs administration with regards to importing products due to delays in obtaining authorisation. The problem is that the shelf life of the food is often expired by the time the food arrives to the store. Hence, the company has appointed one employee to deal with the food regulations and to network with the government. Thailand has been particularly focusing on creating a favourable legal environment for foreign ﬁrms (Country Viewswire, 2002). The country has been launching incentives and tax concessions from 2002 to 2004, thus encouraging foreign investment. In the period 2002–2004, tariffs have been reduced within the structure of the Association of South-East Asian Nations (ASEAN) and the Asian Free Trade Area (AFTA). In order to protect Thai small businesses, the Thai government ﬁnanced 359 million baht ($9.45 million) to establish the Allied Retail Trade Company (ART). The purpose of this non-proﬁt organisation is to help family-owned retailers to compete with large foreign retailers like Tesco and Carrefour (Country Viewswire, 2002; Mass Market Retailers, 2002). The ART aims to build a national retail network, combining orders from members in order to reach a lower price from suppliers, provide low interest in ﬁnancing for store redecoration and give free management training. Later, in November 2002, Thailand conﬁrmed it had dropped the plans, which would discriminate foreignowned shops with a large share of the Thai retail market, because it feared retaliation under international trade rules (Wong-Anan, 2002). In the culture perspective, M&S uses local people who select the products for the local customer. It is a crucial role to select the right product to respond to a customer’s need; hence they need local people with experience. Economic factors are one of the major factors. Bangkok represents 50% of gross domestic product. It is the major reason why the retailing industry in Bangkok has developed and modernised, whereas the rest of the country still remains backward in style (Feeny, Vongpatasin, & Soonsatham, 1996).

Performance M&S’s entry into Thailand is considered successful. As we can see from the Fig. 4, the sales ﬁgure increased from the entry year from 398 million to 454 million Thai baht in a couple of years. ‘‘I think people open up to M&S because it is a good brand, well known. The economy in Thailand was booming at that time. Many factors inﬂuence M&S’s success. When M&S entered, there were not many competitors. People bought basic and comfortable

53

550 500 450 400 350

03 20

02 20

01 20

00 20

99 19

98

300 19

Local Retail Sales (Million Thai Baht)

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

Year Local Retail Sales in Thailand (Million Thai Baht)

Fig. 4.

Local Retail Sales in Thailand (Million Thai Baht). Source: Company Report (2005).

clothes. Nowadays, the clothing industry really emphasises fashion and we can see it clearly’’ (Buying Manager, Thailand). However, the company’s retail sales dropped from 2001 to 2003. This is because the increasing competition in clothing market from both small boutiques and international brands such as Mango and Zara. M&S offer a high price with less attractive products compared with its competitor ‘‘M&S always is a little bit behind others. Our competitors follow fashion trends about 3–6 months faster than us . . . Do not forget that we are not a trendsetter. We are not like MANGO and Zara that you have to launch only products in fashion trend immediately’’ (Buying Manager, M&S Thailand). Customers think that the price of the M&S products is too high in Thailand. M&S thus communicates with customers to explain about the quality and value of the brand: ‘‘Customers said we are expensive. When we put the advertisement, we change and try to show that we are premium and we have something to add, we have better service, we improve the customer relationship’’ (Marketing Manager, M&S Thailand).

DISCUSSION: CASE COMPARISON Comparing the similarities and differences in both markets, it is clear that both markets are fragmented. There are several small competitors providing cheap and cheerful products without a major brand. Both markets are, however, different in terms of the market characteristics (Table 2).

Small food and drug regulation issue Asian consumer behaviour Use experienced buyer and sales history

Do not have import duties, tax free

Buyer picks suitable range for local customers Focus on younger customer

Do not have issues in Asia

Deliver product and franchisees select them

Market characteristic: Legal factor Market characteristic: Cultural difference

Franchise – Use local expertise to build network

Landlord – Money talk, improve brand image Retail Association – Rarely share information due to business competitor

External consultant to conduct market research

Some contact with embassy -W marketing and promotion Local Government for speedy clearance of goods

Government supports tourism, mainland China entry to Hong Kong easier -W boost sales

Little or no impact

Matching – Micro level

Matching – Macro level

Little or no impact

‘‘We do not believe that they get any beneﬁcial support or inﬂuence from international organisations’’ (Asia Regional Manager) Department of Trade and Industry (DTI) Local government

Marks & Spencer Thailand

Matching – Global level

Marks & Spencer Hong Kong

The Inﬂuencing Factors to a Successful Business Operations Abroad.

Opinion from United Kingdom Towards International Expansion

Table 2.

54 KANNIKA LEELAPANYALERT

Performance

Networking in establishment phase

Market characteristic: Economic factor Networking in search phase and project phase

Company use British embassy, property company, and consultant to get information about local market such as economic Network with supplier to develop local adaptation product Network with M&S in other countries to share information and best practices through meeting, emails, intranet Network with local customer by royalty card to gain customer loyalty Use experience local staff to pick right product for the market Successful entry due to economy, Western brand but sales go down

Successful entry but encounter Asian crisis in 1997 Turn loss into proﬁt in 2003 Well-established in Hong Kong

Use British embassy to get information about potential franchisee

Economic growth and GDP increase

Use experience local staff to pick right product for the market

British embassy is the ﬁrst contact to get local knowledge about local market

Impact of SARS on economy in 2003

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia 55

56

KANNIKA LEELAPANYALERT

Legal aspect: Hong Kong does not have issues in legal matters or regulation. Hong Kong does not have import duties and is tax-free so it attracts customers to shopping in Hong Kong. The impact of the SARS outbreak made tourist and retail sales drop, but it has now improved. The HKRMA has been seen in a strong network to be able to succeed and negotiate the rent with the landlord at that time. The Thai regulation towards import duty is stricter, as compared to Hong Kong regulations. For instance, the standard tax rate in Hong Kong is very low in comparison with Thailand or other developed countries. The food and drug regulations might delay transport time so it could be negative for the shorter shelf life of food products. As a consequence, M&S in Thailand needs to network with the customs department in order to facilitate customs clearance procedure. Due to complex customs clearance procedures present in the Thai market, it is critical for business to take advantage of network with local customs actors. Culture: Thai people are more conservative and highly respect superiority. People do not give direct feedback to protect themselves against the concept of ‘‘loosing face’’. On the contrary, in Hong Kong, people are more openminded and independent. The approach towards networking was different because of the differences in culture. Hong Kong uses a direct approach and people clearly say what they want. However, Thai people use a more suggestive and indirect way to communicate. Political: After the SARS disaster, which affected the Hong Kong economy and tourism industry, more assistance was provided by the government. This policy to stimulate the tourism industry and economy included the promotion of Hong Kong as the Asian shopping destination. This was an element which helped to boost retail sales. On the contrary, there was little government assistance for the retail industry in Thailand. Matching: As illustrated in Table 3, M&S in both countries stated that they do not do much matching activities at the global level. At macro level, The British embassy is useful in providing local market information at the early stage of the process (search phase). The tourist policy from Hong Kong tourist board has stimulated purchasing power and increased retail sales in Hong Kong. In Thailand, the Thai government aborted the antiMNC retail law, which has been beneﬁcial for M&S. At micro level, the main important network actors were different in the two countries (Table 4). The landlord was the main inﬂuencing actor in the Hong Kong business as the retail premises in prime location were very important. The company is networking with the landlord through meeting, business lunch and inviting landlord to the brand events. However, the landlord was not an issue in Thailand because CDS owns its land and building. The main important

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

Table 3.

The Network Activities during M&S Expansion to Hong Kong and Thailand: Search Phase and Project Phase. Search Phase

Actors

Activities

Resources

Table 4.

57

Project Phase

Department of Trade and Industry (DTI) in the British embassy, consultant company M&S collect local market information about local economy, regulation, competitor, potential franchisee before entering to the market Business development manager contact DTI in the ﬁrst visited in each market

Local government

Local government intend to promote tourism industry and focus on shopping/the Thai government has recently aborted the anti-MNC retail law M&S is a member of retail association. The company create activities and follow the government policy

The Network Activities during M&S Expansion to Hong Kong and Thailand: Establishment Phase. Establishment Phase M&S Hong Kong

Actors

Landlord

Activities

Company build relationship through regular meeting, submit sales taken to the landlords and building the positive brand image to landlord in order to keep store footage in the prime location. M&S also be able to negotiate with the landlord to reduce the rent after SARS outbreak Managing director invite landlord to business lunch and company events such as fashion show

Resources

M&S Thailand Network with other brands under Central Department Store (CDS) M&S need to corporate and share resources and information with other brands in CDS

M&S share resources with other brands such as using main staff who works with different brands

actor in Thailand was the relationship with the M&S brand and other brands in the Central store, because M&S is one of the many brands that operate in Central store and they need to share resources. One of the reasons that M&S operates through franchising is to be able to use the franchisee’s local expertise, such as local market knowledge and network relationship at the macro and micro level.

58

KANNIKA LEELAPANYALERT

M&S Hong Kong has been more networks oriented as compared to Thailand. It is because the entry mode and organisation is different. As Hong Kong is a M&S subsidiary, they are more committed in the market, the size of the business is bigger and hence there are more people who are dedicated to the brand. Hong Kong uses external partners to conduct market research, and conduct some in-house but ad hoc market research, such as competitor shopping and mystery shopping. However, there is little systematic market research conducted in Thailand due to high costs and the small size of the business operation. M&S operate as one brand under CDS and share resources and people with other brands. There is smaller team dedicated to M&S in Thailand as it is a small department. As a result, networking with the external is less speciﬁc in Thai market. Both countries use network with local employees who have extensive experience in buying to select the product to respond to the local customer needs. Sales assistants play the role of collecting information not only about customers but also competitors. Customers would suggest what they like and do not like about M&S as well as suggesting which competitor’s product they like. The buying team always shop at the competitor’s places. The company adjusts the sizes, styles and colours of products to the local needs, such as offering Asian ﬁt bras and shorter sleeves on men’s shirts. Impact of the recession in the UK market: M&S, at the time of the study, was not doing very well in the UK market. M&S retail sales were 4% compared to the average British retail performance, which reached an annual percentage growth of 2% (Financial Times, 2004). This situation had a major impact on the Hong Kong business and less on Thailand. As Hong Kong operates as a subsidiary, the recession in the domestic market had a negative impact on M&S’s Hong Kong operation. However, M&S Thailand operates under a franchise to the CDS, and therefore customers perceive M&S as being under Central’s operation in Thailand ‘‘If you ask me whether the situations in UK affect us? It is very little, not much. People know that M&S here is only one part of Central. Customers do not really feel it to be a separate company’’ (Buying Manager, M&S Thailand). The issue, which the company is currently working on, is price. As the result of market research, Asian customers perceive M&S as expensive, not a high volume retailer like in the UK market. The M&S price is at the top of the middle end of the retailing market in both Thailand and Hong Kong, and the company is currently working together with their subsidiary and franchise partners to minimise costs and bring prices down. In Singapore, prices were reduced by 10 per cent and they also plan reducing prices in Hong Kong.

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

59

Research Propositions The intention of this research is to build a conceptual framework for directing future research. Hence, this research suggests the conceptual framework needs further development and presents some propositions developed from previous studies and this study. The future research suggested will test these propositions (Fig. 5).

A A higher level of matching would drive the company to become more network orientated The studies show that the matching company actively use market characteristics in order to understand local market (Rogers et al., 2005). Matching is one of the methods to assist company access customers and competitors’ information. Ghauri and Holstius (1996) introduced the role of matching which could facilitate international market entry. The company essentially networks with other companies and key stakeholders in the market (Johanson & Mattsson, 1987) as each company’s activities rely on each other within the network (Buckley & Ghauri, 1999). Ghauri and Holstius (1996) identify the actors, activities and resources in three stages of international entry process. They found that efﬁciently matching and networking with the right actors at different stages; global, macro and micro level; could beneﬁt companies during the entry process.

Matching • Global level • Macro level • Micro level

B Networking

A

Actors Performance

C

Resources

Activities

1. Search Phase Market Characteristics • Legal • Culture • Political • Economic

2 Project Phase

E

3 Establishment Phase

D

Fig. 5. Factors Inﬂuencing the Internationalisation Process of UK Firms in Emerging Markets: Propositions. Source: Based on Ghauri et al. (2004).

60

KANNIKA LEELAPANYALERT

B A higher degree of matching activities would facilitate a better performance in the host market M&S Thailand uses macro matching with local customs to facilitate the custom clearance process so the goods arrive to the store at the right time, which affect sales positively. In Hong Kong, M&S uses micro matching with local landlords through regular communication, report store performance and inviting them to company events in order to keep M&S stores in prime shopping locations.

C The higher degree of matching activities would better equip the company to overcome different market environments The result of Ghauri and Holstius’ (1996) studies shows that matching facilitates the company’s activities during the internationalisation process in emerging countries. Matching with local government helps the company to overcome the political issues. There is no study identifying the relationship between matching and company’s performance, so it is interesting to look at it in the further research.

D The more a company becomes embedded in the local network, the better equipped the company will be to overcome different market environments A dissimilar market environment between the host and home market, such as cultural differences (Hofstede, 1994), physic distance (O’Grady & Lane, 1996) and attitude of the foreign government, could become a barrier to international expansion (Ghauri & Holstius, 1996; Buckley & Ghauri, 1999).

E The higher level of networking in different stages would lead to better performance in the host market Government contacts need to be established at the early stage of internationalisation. Matching with government helps company to access to information, up-to-date information about policy and get permission (Rogers et al., 2005). The use of networking with external factors including suppliers assisting the company to access know how and perform better. This study agrees with Rogers et al.’s (2005) ﬁnding that network relationships with suppliers can allow a company to build a supply chain network. Networking with suppliers helps the company to provide low priced products for customers. As a result, the brand image of M&S has changed to become a more affordable brand image.

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

61

CONCLUSION There are several studies that have investigated the internationalisation of M&S particularly its de-internationalisation and withdrawal of M&S stores from several international markets (Burt et al., 2002a; Burt et al., 2002b; Jackson, Mellahi, & Sparks, 2004; Mellahi, Jackson, & Sparks, 2002). Mellahi et al. (2002) explained how the most successful retailers like M&S turned to the critical situation in 1997. These studies found that both internal and external factors inﬂuenced company performance. The company faced difﬁculty because of the external environment such as tough competition and changing consumer behaviour. To answer the question: What roles do networking and matching play during the internationalisation process of Western ﬁrms in emerging markets? This study conﬁrms that matching would assist the company to understand more about the local environment thus enabling it to adapt to the local market (Ghauri & Holstius, 1996). It also conﬁrms that the company ﬁnds it difﬁcult to adapt and respond to the changing environment. Management style and hierarchy are the main problem as the decision-making is very slow (Vida & Fairhurst, 1994). Company is not able to respond to the customer needs in providing the right product at the right price. It has been expected to ﬁnd a relationship between matching and networking. In these cases, it has been found that crucial networking with the right actors, such as suppliers’ associations, allow the company to access the market and customer information. This research further conﬁrms that the ability of understanding customers, which can be done through matching, are the main factors that inﬂuence performance in international operations. Companies in networks create competitive advantages by developing relationships with other companies. A company in the network could reduce its risk and uncertainty and create a more efﬁcient network (Elg & Johansson, 1996). How dissimilarities in the market environment can be managed during the internationalisation process? Effectively, the selection of a speciﬁc market to enter will be strongly inﬂuenced by environmental factors and company characteristics (Ghauri & Cateora, 2006; Vida & Fairhurst, 1998; Root, 1987). These environmental issues can encourage or deter entry into foreign markets (Cavusgil et al., 2002, p. 41). This study found that matching helps at many levels to reduce and avoid the risks involved with international operations (Ghauri & Holstius, 1996; Holstius, 1991). Although matching at different levels might vary in terms of importance, it is evident that it contributes positively to the performance of M&S in Hong Kong. This is illustrated by the networking taking place with the right actors including the

62

KANNIKA LEELAPANYALERT

local government, retail association and the landlord. Using matching and networking helped the company to manage the dissimilarities of the market environment between the host and home country and also uncontrollable factors (Rogers et al., 2005). The example is the retail association asking landlord and local government for assistance after the SARS outbreak in Hong Kong so the landlord reduced the rent and government has policy to promote tourist and retail industry in order to gain tourists conﬁdent. How could different adaptation levels affect performance in different countries? Market environment has signiﬁcant impact on retail internationalisation (Akehurst & Alexander, 1995; Alexander & Myers, 2000; Vida & Fairhurst, 1998). The study showed that reacting to the market environment allows the company to understand local market and be able to adapt to local markets. Western ﬁrms establish a market position in emerging markets through the use of matching with authorities, suppliers of goods and local actors. This is facilitated through inﬂuencing local actors through the British Embassy, Property Company and other ofﬁcials. Hong Kong is a subsidiary; therefore they have more resources and are more committed to the market. However, the franchise operation is dependent on its franchise partner to operate the business locally. The evidence of M&S Hong Kong conﬁrms Johanson and Vahlne (1977) that the more local knowledge a company has, the more committed to the market it is. Hong Kong expanded stepwise, ﬁrst through franchises and later by setting up its own subsidiaries. In addition, this study found that networking helps company to understand more about market and customers and therefore enhance that market knowledge. Buckley, Newbould, and Thurwell (1988) also studied the evolution approach of internationalisation and found that the company started direct export and used intermediate agency and own subsidiaries later. The purpose of that was to learn from the experience. The second purpose was to pulling out the operation in case of unsuccessful results in speciﬁc market. In these cases, the company used franchisee’s expertise in the market to help organisation learn about local market. In M&S Hong Kong case, company began from direct export to franchisee and ﬁnally when they saw the proﬁt and successful operation in Hong Kong, they set up their own subsidiary. After the 1997 Asian crisis combined with the recession of M&S brand in home market, the company tried to reverse the internationalisation process by announcing the sale of the Hong Kong business to its franchise partner. This process revealed less commitment by the company when its local operations showed negative results and also because the policy of the headquarters was to focus on improving the home market operations rather

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

63

than international operations. These were the major reasons for M&S’s withdrawal from most markets in Europe. As regards the managerial implications, this study will beneﬁt companies engaged in activities abroad, as well as those planning to engage in activities in international markets. It focuses on inﬂuencing factors, thus providing a better understanding of the requirements and conditions necessary for a successful foreign operation. It has been demonstrated that building relationship with key actors in the local market is crucial; also a better understanding of the customer’s needs helps signiﬁcantly in attracting more customers. From the theoretical point of view, this chapter re-evaluates Ghauri and Holstius’ (1996) matching concept in the context of international business operations and extends the scope of the analysis, using cases of UK retailer in the two emerging markets. This research also contributes towards theory development in this ﬁeld by providing a framework that could be a basis for future empirical investigations of international expansion. The ﬁndings also suggest how foreign market operations can be managed, thereby providing a way to achieve an improved understanding of the internationalisation process.

REFERENCES Akehurst, G., & Alexander, N. (1995). Developing a framework for the study of the internationalisation of retailing. In: G. Akehurst & N. Alexander (Eds), The internationalisation of retailing (pp. 204–209). London: Frank Cass. Alexander, N. (1990). Retailers and international markets: Motives for expansion. International Marketing Review, 7(4), 75–85. Alexander, N., & Myers, H. (2000). The retail internationalisation process. International Marketing Review, 17(4/5), 334–353. Anderson, J. C., Ha˚kansson, H., & Johanson, J. (1994). Dyadic business relationships within a business network context. Journal of Marketing, 58(4), 1–15. Brown, S., & Burt, S. L. (1992). Conclusion-retail internationalization: Past imperfect, future imperative. European Journal of Marketing, 26(8–9), 80–84. Buckley, P. J., & Casson, M. (1976). The future of the multinational enterprise. New York: The Macmillan Company. Buckley, P. J., & Ghauri, P. N. (1999). The internationalisation of the ﬁrm: A reader. London: Thomson Learning. Buckley, P. J., Newbould, G. D., & Thurwell, J. (1988). Foreign direct investment by smaller UK ﬁrms. London: Macmillan. Burt, S., Davies, K., Mcauley, A., & Sparks, L. (2002a). The failure of retail internationalisation in Marks and Spencer. European Retail Digest, 35, 1–4. Burt, S., Davies, K., Mcauley, A., & Sparks, L. (2005). Retail internationalisation: From formats to implants. European Management Journal, 23(2), 195–202.

64

KANNIKA LEELAPANYALERT

Burt, S., Mellahi, K., Jackson, T., & Sparks, L. (2002b). Retail internationalisation and retail failure: Issues from the case of Marks and Spencer. The International Review of Retail, Distribution and Consumer Research, 12(2), 191–219. Butler, S. (2007). New clampdown will make it tougher for Tesco in Thailand. The Times, 9 May. Cavusgil, S. T. (1990). On the internationalization process of ﬁrms. In: H. B. Thorelli & S. T. Cavusgil (Eds), International marketing strategy (pp. 147–159). Oxford: Pergamon Press. Cavusgil, S. T., & Robert, W. N. (1990). Assessment of company readiness to export. In: H. B. Thorelli & T. Cavusgil (Eds), International marketing strategy (pp. 129–139). Oxford: Pergamon Press. Cavusgil, T., Ghauri, P., & Agarwal, M. (2002). Doing business in emerging markets: Entry and negotiation strategies. Thousand Oaks: Sage Publications. Clark, B. (2002). Measuring performance: The marketing perspective. In: A. Neely (Ed.), Business performance measurement: Theory and practice. Cambridge: Cambridge University Press. Company Brochure. (2005). Central retail corporation: Company brochure. Available at http:// www.central.co.th/web/html/eng/links.html. Accessed on 29 June. Company Report. (2005). International franchise. Available at http://www2.marksandspencer.com/thecompany/index.shtml. Accessed on 5 June. Company Report. (2007). Annual review 2007: International. Available at http://www. marksandspencer.com/gp/node/n/57009031?ie=UTF8&mnSBrand=core. Accessed on 5 December. Country ViewsWire. (2002). Thailand industry: Hypermarket worry small retailers. Country ViewsWire, 16 November. Coviello, N. E., & McAuley, A. (1999). Internationalization of the smaller ﬁrm: A review of contemporary empirical research. Management International Review, 39, 223–256. Cyert, R., & March, J. (1963). A behavioral theory of the ﬁrm. New Jersey: Blackwell. Davies, K., & Fergusson, F. (1995). The international activities of Japanese retailers. In: G. Akehurst & N. Alexander (Eds), The internationalisation of retailing (pp. 97–117). London: Frank Cass. Dawson, J. A. (1994). Internationalization of retailing operations. Journal of Marketing Management, 10, 267–282. Day, G. S. (1990). Market driven strategy. New York: The Free Press. Department of Health and Human Services. (2004). Global SARS outbreak. Available at http://www.cdc.gov/ncidod/sars/faq.htm. Accessed on 1 November. Dunne, P., Lusch, R., & Gable, M. (1995). Retailing (2nd ed.). Ohio: South-Western College. Dunning, J. H. (1981). International production and the multinational enterprise. London: Allen and Unwin. Elg, U., & Johansson, U. (1996). Networking when national boundaries dissolve. European Journal of Marketing, 30(2), 61–74. Feeny, A., Vongpatasin, T., & Soonsatham, A. (1996). Retailing in Thailand. International Journal of Retailing and Distribution Management, 24(8), 38–44. Financial Times Business. (2004). Rose faces an extensive ‘to-do’ list to out M&S closet in order. Financial Times, 10 November, p. 21. Ford, D. (1990). Understanding business markets: Interaction, relationships and networks. London: Academic Press.

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

65

Ford, D., Gadde, L., & Ha˚kansson, H. (1998). Managing business relationships. Chichester: Wiley. Ghauri, P. (2004). Designing and conducing case studies in international business research handbook of qualitative research methods for international business. In: R. MarchanPiekkari & C. Welch (Eds), Handbook of qualitative research methods for international business (pp. 109–124). Cheltenham: Edward Elgar. Ghauri, P., & Cateora, P. (2006). International marketing (2nd ed.). London: McGraw Hill. Ghauri, P., Elg, U., & Sinkovics, R. (2004). Foreign direct investment – Location attractiveness for retailing ﬁrms in the European union. In: L. Oxelheim & P. Ghauri (Eds), European Union and the race for foreign direct investment in Europe (pp. 407–428). Oxford: Pergamon. Ghauri, P., & Grønhaug, K. (2005). Research methods in business studies: A practical guide (3rd ed.). London: Prentice Hall. Ghauri, P., & Holstius, K. (1996). The role of matching in the foreign market entry process in the Baltic States. European Journal of Marketing, 30(2), 75–88. Ghauri, P., & Prasad, B. (1995). Probing Asia’s inter-ﬁrm linkages: A network approach. Advance in International Comparative Management, 10, 63–78. Ghauri, P. N. (1999). Advances in international marketing: International marketing purchasing. Connecticut: JAI Press. Hadjikhani, A., & Ghauri, P. (2001). The behaviour of international ﬁrms in socio-political environments in the European union. Journal of Business Research, 52(3), 263–275. Ha˚kansson, H., & Johanson, J. (2001). Business network learning. New York: Pergamon. Ha˚kansson, H., & Snehota, I. (1989). No business is an island: The network concept of business strategy. Scandinavian Journal of Management, 5(3), 189–200. HKRMA. (2004). Hong Kong retail management association. Available at http:// www.hkrma.org/en. Accessed on 25 October. Hofstede, G. (1994). The business of international business is culture. International Business Review, 3(1), 1–14. Hollander, S. C. (1970). Multinational retailing. East Lansing: Michigan State University. Holstius, K. (1991). A matching concept for international business: The case of Finland and China. Advance in Chinese Industrial Studies, 2, 177–185. Hong Kong Tourism Board. (2003). A statistical review of Hong Kong tourism. Hong Kong Tourism Board, pp. 12–13. Hong Kong Tourism Board. (2006). A statistical review of Hong Kong tourism. Hong Kong Tourism Board, pp. 5–12. Hong Kong Trade Development Council. (2003). Hong Kong: China’s consumption city. Available at http://www.tdctrade.com/econforum/bea/bea031001.htm. Accessed on 1 October. Jackson, P., Mellahi, K., & Sparks, L. (2004). Shutting up shop: Understanding the international exit process in retailing. The Service Industries Journal, 25(3), 355–371. Jaworski, B. J., & Kohli, A. K. (1993). Market orientation: Antecedents and consequences. Journal of Marketing, 57(3), 53–70. Jennings, D., & Young, D. (1990). An empirical comparison between objective and subjective measures of the product innovation domain of corporate entrepreneurship. Entrepreneurship: Theory and Practice, 15(1), 301–344. Johanson, J., & Mattsson, L. (1987). Internationalization in industrial systems – A network approach compared with the transaction-cost approach. International Studies of Management and Organization, 17(1), 34–48.

66

KANNIKA LEELAPANYALERT

Johanson, J., & Mattsson, L. (1988). Internationalization in industrial systems-A network. In: N. Hood & J.-E. Vahlne (Eds), Strategies in global competition (pp. 287–318). New York: Croom Helm. Johanson, J., & Vahlne, J. (1977). The internationalization process of the ﬁrm – A model of knowledge development and increasing foreign market commitment. Journal of International Business Studies, 8(1), 23–32. Johanson, J., & Vahlne, J. (2006). Commitment and opportunity development in the internationalization process: A note on the Uppsala internationalization process model. Management International Review, 46(2), 165–178. Johanson, J., & Wiedersheim-Paul, F. (1975). The internationalization of the ﬁrm: Four Swedish cases. Journal of Management Studies (October), 305–322. Kacker, M. (1985). Transatlantic trends in retailing. Westport, CT: Greenwood Press. Katsikeas, C. S., & Morgan, N. (1998). Assessing ﬁrms’ export performance: A literature review and guidelines for future research. Working Paper. Cardiff Business School, University of Wales. Kohli, A. K., & Jaworski, B. J. (1990). Market orientation: The construct, research propositions and management implications. Journal of Marketing, 54(2), 1–18. Lenway, S., & Murtha, T. (1994). Country capabilities and the strategic state: How national political institutions affect multinational corporations’ strategies. Strategic Management Journal, 15(Summer), 113–129. March, J. G., & Sutton, R. I. (1997). Organizational performance as a dependent variable. Organization Science, 8(6), 698–706. Mass Market Retailers. (2002). Thailand state ﬁrm. Mass Market Retailers, 19(12), 11–12. McGoldrick, P., & Davies, G. (1995). International retailing: Trends and strategies. London: Pitman Publishing. McGoldrick, P. J. (1998). Spatial and temporal shifts in the development of international retail images. Journal of Business Research, 42(2), 189–196. McGoldrick, P. J. (2002). Retail marketing (2nd ed.). Berkshire: McGraw Hill Education. Mellahi, K., Jackson, P., & Sparks, L. (2002). An exploratory study into failure in successful organizations: The case of Marks & Spencer. British Journal of Management, 13(1), 15–30. Meyer, K., & Skak, A. (2002). Networks, serendipity and SME entry into Eastern Europe. European Management Journal, 20(2), 179–188. Miles, M. B., & Huberman, M. A. (1994). Qualitative data analysis: An expanded sourcebook. London: Sage. Morganosky, M. A. (1993). Opportunities and problems associated with direct marketing: A perspective by US retailers. Journal of Direct Marketing, 7(2), 41–52. Murtha, T. (1991). Surviving industrial targeting: State credibility and public policy contingencies in multinational. Journal of Law, Economics, and Organization, 7(1), 117–143. Narver, J. C., & Slater, S. F. (1990). The effect of a market orientation on business proﬁtability. Journal of Marketing, 54(4), 20–35. Nivatpumin, C. (2005). Thailand up two notches in world rankings. Bangkok Post. Available at http://www.siamfuture.com/ThaiNews/ThNewsTxt.asp?tid ¼1562. O’Grady, S., & Lane, H. (1996). The psychic distance paradox. Journal of International Business Studies, 27(2), 309–333.

Factors Inﬂuencing the Internationalisation Process of UK Firms in Asia

67

Richards, L. (2000). Using NVivo in qualitative research (2nd ed.). Melbourne: QSR International. Rogers, H. P., Ghauri, P., & George, K. (2005). The impact of market orientation on the internationalization of retailing ﬁrms: Tesco in Eastern Europe. The International Review of Retail, Distribution and Consumer Research, 15(1), 53–74. Root, F. R. (1987). Entry strategies for international markets. Massachusetts: Health and Company. Root, F. R. (1992). International strategic management: Challenges and opportunities. Washington, DC: Taylor and Francis. Salmon, W. J., & Tordjman, A. (1989). The internationalization of retailing. International Journal of Retailing, 4(2), 2–16. Sheth, J. (1995). Relationship marketing in consumer markets: Antecedents and consequences. Journal of the Academy of Marketing Science, 23(4), 255–271. Sinkovics, R., Penz, E., & Ghauri, P. (2005). Analysing textual data in international marketing research. Qualitative Market Research: An International Journal, 8(1), 9–38. Sternquist, B. (1998). International retailing. New York: Fairchild Publications. Sybrandy, A., Pirog, S. F. D., & Tuninga, R. S. J. (1991). The output of distributive systems: A conceptual framework. Journal of Macromarketing (Fall), 19–26. Treadgold, A., & Davies, R. L. (1988). The internationalisation of retailing. Harlow: Longman. Treadgold, A. D. (1990). The developing internationalization of retailing. International Journal of Retailing and Distribution Management, 18(2), 4–11. Vida, L., & Fairhurst, A. (1998). International expansion of retail ﬁrms: A theoretical approach for future investigations. Journal of Retailing and Customer Services, 5(3), 143–151. Whitehead, M. B. (1992). Internationalization of retailing: Developing new perspectives. European Journal of Marketing, 26(8/9), 269–285. Williams, D. E. (1992). Retailer internationalization: An empirical inquiry. European Journal of Marketing, 26(8/9), 8–24. Wong-Anan, N. (2002). Thailand withdraws anti-foreign retail law. Reuters, 19 November. Yin, R. K. (1994). Case study research: Design and methods (2nd ed.). London: Sage.

TOURISM AS A LEVERAGE OF INTERNATIONALIZATION FOR CONSUMER GOODS FIRMS: A CASE STUDY APPROACH Konstantinos Poulis and Mo Yamin 1. INTRODUCTION Large-scale incoming tourism potentially creates a multinational market within the domestic economy of the recipient countries. More speciﬁcally, in a number of countries, there is a large inﬂux of ‘foreign’ consumers, or tourists, from many countries and for a signiﬁcant part of the year. As can be seen from Table 1, for countries such as France, Spain, Austria, or Greece the annual inﬂux of tourists exceeds the population of these countries by very large margins. For such countries, tourism transforms a nationally homogeneous consumer market into a multinational or multicultural consumer market during the tourist season, which can be most of the year. This is likely to be particularly signiﬁcant for FMCG ﬁrms in the tourist-receiving countries. For example, tourists consume beers, soft drinks, shampoos, and ice creams while on holiday and thus, the size of the market for FMCGs will expand signiﬁcantly over the tourist season. Additionally, it is likely that in their consumption decisions, tourists from particular countries reﬂect traits of their

New Challenges to International Marketing Advances in International Marketing, Volume 20, 69–85 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020006

69

70

KONSTANTINOS POULIS AND MO YAMIN

Table 1.

France Spain Italy Austria Greece

Incoming Tourism and Domestic Population.

Tourists Arrivals (Millions of Tourists)

Population of Country (Millions of People)

% of Tourists to the Domestic Population

77 50 40 18 13

60 40 57 8 11

128 125 70 225 118

Source: World Tourism Organization.

nationality and hence be somewhat distinct from not only the ‘native’ populations but also from each other. Consequently, companies not only face a much larger market but one that is also likely to be highly heterogeneous in terms of consumer tastes and preferences; FMCG ﬁrms essentially operate in a multinational consumer market during the tourist season. This study seeks to explore the implications of this fact for the internationalization of ﬁrms domiciled in the tourist-receiving countries.1 The aim of this chapter is to contribute to the internationalization literature by considering the possible inﬂuence of inward tourism in stimulating the internationalization of SMEs headquartered in the tourist-receiving countries. The insights of this chapter rely on the existing literature and ﬁndings from case studies of internationalizing SMEs in the FMCG sectors in Greece. The structure of the chapter is as follows. Section 2 considers the implications of tourism for internationalization in the context of the existing literature on internationalization, particularly the literature on the relevance of knowledge. In this section, we attempt to show how in responding to the presence of tourist as a signiﬁcant component of their domestic markets, SMEs effectively gain information/knowledge that is also ‘international’ in character. For example, they may be able to gain brands with signiﬁcant international appeal or equity even before they operate in foreign countries (or even advertise) there. Section 3 explains the case selection methodology and reports ﬁndings on the link between internationalization and inbound tourism revealed by the experience of ﬁrms included in the study. Section 4 develops ﬁve propositions, based on the extant literature and our case study ﬁndings, relating to the impact of incoming tourism on the internationalization of SME in the tourist recipient countries. The propositions emphasize knowledge development; the leveraging of local, ‘ethnobrands’ and the key role played by the Internet to maintain communication with tourist customers and the role that this plays in ﬁrms’ internationalization.

Tourism as a Leverage of Internationalization for Consumer Goods Firms

71

Section 5 concludes the chapter indicating the limitations of the present study and drawing out implications for future research.

2. KNOWLEDGE AND INTERNATIONALIZATION Lack of knowledge about foreign markets is recognized as a key impediment to ﬁrms’ internationalization. This is because the accumulation of necessary knowledge is very time-consuming, country-speciﬁc with signiﬁcant differences of needed information across markets and depends on the degree of the ﬁrm’s internationalization (Johanson & Vahlne, 1977, 1990; Luostarinen, 1979; Denis & Depelteau, 1985; Johanson & Wiedersheim-Paul, 1975; Eriksson & Chetty, 2003). Research, focusing speciﬁcally on export barriers also highlights the importance of knowledge. Leonidou (2004), by pulling together results from 32 empirical studies, ranked a number of barriers in terms of the overall impact. Lack of knowledge is invariably ranked as the factor with the highest negative impact on SME internationalization. However, once achieved, market knowledge can provide a platform for further internationalization of the ﬁrm. Researchers highlight both internal and external knowledge accumulation and how these lead to international growth of the ﬁrm by internalizing information. An increasing number of studies dealing with the issue of knowledge creation in the internationalization literature suggest, therefore, that the knowledge variable is of critical importance for MNEs (Kogut & Zander, 1993; Yang, Leone, & Alden, 1992; Liesch & Knight, 1999). As a result, knowledge gained through market screening and involvement is an essential driver of internationalization, which can deﬁne its degree and pace (Hadley & Wilson, 2003). Slow rates of generating this experiential knowledge can lead to reduced internationalization pace and vice versa (Johanson & Vahlne, 1977), while empirical results show that knowledge as a result of being exposed to a variety of market knowledge sources, creates more opportunities than relying solely on internal capabilities (Inkpen, 1996; Osland & Yaprak, 1995).

2.1. Incoming Tourism: Reduction in the Knowledge Barrier to Internationalization Given the importance of market knowledge for export development and internationalization, it is somewhat surprising that the relevance of incoming tourism for ﬁrm internationalization has not been seriously

72

KONSTANTINOS POULIS AND MO YAMIN

considered in prior research. Extant research on the barriers to exports and internationalization speciﬁcally indicates that lack of contact with foreign consumers is a major barrier to export initiation and development. The study by Leonidou (2004), classiﬁes this as a ‘very high impact’ export barrier; more speciﬁcally he pointed out that Identifying customers in overseas markets constitutes a serious impediment for many would be and current exporters. This can be attributed to three major factors (1) large geographic distances separating sellers and buyers in foreign markets, hampering communication between them; (2) the characteristically half hearted approach taken by many ﬁrms towards carrying out market research in overseas markets and (3) the limited exposure to sources listing potential customers.

Clearly ﬁrms in tourist-receiving countries have an unusual opportunity to become acquainted with (some) foreign consumer behaviors, attitudes, and needs without actually having to move beyond their borders or having to invest as much in international market research and screening as would otherwise be necessary. Firms in tourist-receiving countries can assess the attractiveness of their offerings/brands among foreign consumers; they can gain knowledge of what kind of product categories are usually purchased by which ethnic or demographic groups helping deﬁne the product portfolio by which they may enter a speciﬁc foreign market; similarly, they may gain knowledge that enables them to specify standardized or adapted promotional campaigns and product attributes. Additionally, research ﬁndings that a ﬁrm’s exposure to culturally diverse markets can facilitate its ability to conduct international operations (Hadley & Wilson, 2003; Eriksson & Chetty, 2003; Ghoshal, 1987) suggest that ﬁrms operating in a tourismreceiving country can capitalize on this inﬂux of an additional consumer base which is multicultural in character. In general, a key feature of macro-environment in countries with largescale inbound tourism is the opportunity to create ﬁrm-speciﬁc knowledge, relating speciﬁcally to consumer behavior that is simultaneously both domestic and international. This is a valuable potential advantage compared with the environments in which tourism is not an important feature; in the latter context the relevance of domestic knowledge for international expansion is usually highly uncertain and risky (Johanson & Vahlne, 1977). 2.2. Tourism and the Brand Enhancement SMEs, especially those in small economies such as Greece, usually carry a strong country-of-origin image on their shoulders unlike their global

Tourism as a Leverage of Internationalization for Consumer Goods Firms

73

counterparts, which often have a seamless entity that is not anchored on speciﬁc countries. This national rather than global image is often perceived as an inherent drawback of local ﬁrms in their effort to internationalize their operations. While there is an assertion in the literature that all ﬁrms strive to develop a global image for their brands (Yu, 2003; Craig & Douglas, 2000), inadequate ﬁnancial resources usually impedes SMEs to further enhance their brand equity among foreign consumers who could otherwise provide a platform of demand for international expansion. However, the branding disadvantage of SMEs may only be valid for a ﬁrm selling in a foreign country to consumers within their country of residence. In this case, there are inherent problems for foreign SMEs associated primarily with the lack of brand awareness, ﬁnancial scarcity, and the lack of access to distribution channels. However, in cases when the international consumers are within the country of operation of the ﬁrm – where it has an established access to distribution channels, i.e. in the case of foreign (tourists) consumers in the home country of the SME – the ‘locality’ of the brand may not be as great a disadvantage and may even become an advantage. There is an opportunity for ﬁrms to address to an increasing number of foreign consumers domestically in a way that would otherwise cost enormous amounts of money, i.e. launch an international promotional campaign to develop brand awareness and enhance international brand equity. After all, a main premise behind the alleged brand superiority of global brands is the familiarity concept (Cheng, Blankson, Wu, & Chen, 2005). Consumers are familiar with the properties of a brand and this leads to favorable purchasing. We do not know yet though whether the familiarity concept is equally applicable for a foreigner consuming products in a country beyond his/her usual habitat. Familiarity may be replaced by other constructs and attitudes, such as experimentation, curiosity, culturally oriented exploration, or simply a deliberate get-away from the normal daily routines back home. After all, the essence of being a tourist according to one of the most cited sociologists of tourism, MacCannell (1976) is to ﬁnd reality ‘out there’ through authentic experiences. If this is the case indeed and this unitary type of tourist exists, local ﬁrms not only have an edge for local consumption over their multinational rivals who enjoy an established brand reputation but also this consumption can lead to satisfaction, repurchasing, and even loyalty when the tourist gets back home. Therefore, the opportunity for the ﬁrm lies in the fact that there is low cost way (compared to the ‘usual’ internationalization process) to develop and enhance the international brand equity and consequently, create and retain a signiﬁcant number of customers abroad.

74

KONSTANTINOS POULIS AND MO YAMIN

3. METHODOLOGY AND INTERVIEW ANALYSIS Seven Greek ﬁrms with an already existing exporting activity were approached to serve as case studies. Following Patton (1990), sampling was based on a maximum variation logic trying to include ﬁrms from a diversiﬁed range of industries such as food, beverage, and cosmetics since the nature of the industry is an ubiquitous factor in the international business literature (Theodosiou & Leonidou, 2003). Four of these ﬁrms have never really thought of tourism as a leverage to internationalize their activities. They exploit tourism as a sales opportunity domestically but they have not speciﬁcally targeted incoming tourists as qualitatively distinct segments. In particular, respondents in these four companies were quite explicit that they were not particularly interested or focused on tourist segments of their respective markets. In effect, these ﬁrms treated sales to tourists as an additional bonus to their main business. Although these ﬁrms had signiﬁcant export activity, such activity was overwhelmingly in the Balkans and they have minimal or no export activities in countries where the majority of tourists to Greece emanate (e.g. Western Europe and, increasingly, Asian countries). These ﬁrms’ export activity exactly mirrors the dominant pattern of exporting in Greece. The remaining three are ﬁrms who have realized the importance of tourism and they do implement programs towards a better exploitation of tourism as a platform of internationalization. One ﬁrm is active in the beverages sector and two ﬁrms in the cosmetics sector. All are SMEs based in Greece, an archetypal tourism destination for international tourists. Greece was chosen as the ﬁeld of study due to the fact that it is a mature tourism destination with a relatively stable tourism clientele every year, it has a huge tourism population compared with the domestic market, has a large number of SMEs striving for internationalization, and is the home country of one of the authors, something which assisted towards the implementation of a deliberate, maximum variation sampling logic. Semi-structured interviews took place with key people in the ﬁrms resulting to seven interviews. An ofﬁcial request was made indicating that we would like the ﬁrm to decide who is the ideal person to answer our question with respect to the research problem under scrutiny. Due to the fact that all three cases are SMEs, only a few number of people deal with this issue. Table 2 describes the key characteristics of the three sampled ﬁrms.

Tourism as a Leverage of Internationalization for Consumer Goods Firms

Table 2.

75

Cases Descriptive Information. The Cases

Case X Industry: cosmetics International activity: 13 countries Interviews: 2 Respondents: Human Resources and Export Marketing Managers

Case Y Industry: beverages International activity: 30 countries Interviews: 3 Respondents: Trade and Marketing, Area Sales and Senior Brand Manager

Case Z Industry: cosmetics International activity: 21 countries Interviews: 2 Respondents: Marketing Manager, Group Product Manager

3.1. Inbound Tourism and Firm Internationalizations: Interview Analyses In this section, based on responses of the managers interviewed, we consider the role tourism plays for the international activity of the ﬁrms; additionally, we highlight the active reliance on the Internet by these ﬁrms in order to maintain communication and build relationship with their tourist customers. 3.1.1. Tourism and Export Development The key beneﬁt of tourism presence in the mother country of the cases is the knowledge factor. This knowledge covers several issues ranging from preferable brand strategies to consumers’ needs and behaviors. As an executive from ﬁrm X stressed: . . . for all these (strategic decisions of international scope) tourism was the source. The source of information . . .

Tourists are a convenient way to get feedback on foreign consumers’ needs and behaviors. This feedback generates knowledge that leads to necessary improvements for competition in the international arena. We certainly use the feedback . . . especially us who are not in the position to quantify any info we would like due to lack of available research ﬁndings . . . feedback is something we use a lot as a means of improvement . . .

A main theme from the interviews is how this knowledge may help to develop international product strategies. The question was whether to standardize the brand and its attributes across all markets or adapt it in several of them. And if adapted, what is the decisive feature of this

76

KONSTANTINOS POULIS AND MO YAMIN

adaptation. How is the brand going to be positioned in these markets? Interaction with and responses from tourists was a key factor leading to the exploitation of the national heritage of the brand. In other words, the ﬁrms customized their brand positioning strategy for the tourism population and then extended this ‘tourist’ logic abroad. Therefore, now, there is a uniform, purely standardized image of the brands all over the world, which aims to communicate the message of ‘Greekness’ as a source of differentiation in foreign markets. The following excerpt is characteristic of this A basic element of Z brand is the word [ . . . ] which reﬂects the whole concept of Greece. It is something we use everywhere, in communication, advertising, the brands themselves, packaging. It is a conscious choice of the ﬁrm that is not predicted to change. It is an element of the brand and has made all tourists visiting Greece to associate the brand with something Greek. And this is something we aim for, that Z is a Greek ﬁrm . . . we have built all our brand and communication on that.

We can call such brands as Ethnobrands since they capitalize on the ethnic origin of the company and its accompanying properties and thus, promote culture-speciﬁc values communicated by the product. Such Ethnobrands are perceived by ﬁrms as powerful sources of differentiation whose market power was empirically tested through both the domestic, ‘tourism-loaded’ market and the international operations. International tourism generated the knowledge on what foreign consumers value most, something which otherwise would demand extensive international market research, probably several failures through experimentation and huge costs associated with research and analysis of foreign cultures. Our case interviews clearly indicate that these ﬁrms were actively engaged with the tourist segments of their markets and saw promotional communications among tourists as a vehicle to help their internationalization: Tourism is used as a means of promotion abroad since our orientation is entirely and clearly towards exporting . . . It is an opportunity for communicational approach of potential customers abroad.

An interesting insight yielded by the case studies is the utilization of domestic distribution as a means of international promotional strategy. Distribution within Greece was adapted to a more ‘tourist’ logic both in order to increase domestic sales and in order to serve the purpose of the exporting orientation of the ﬁrm. So, channels of distribution were integrated within the already existing network to work in favor of the promotion of the corporate and brands’ image. Two of the ﬁrms created

Tourism as a Leverage of Internationalization for Consumer Goods Firms

77

their own distribution outlet and became retailers themselves in order to exploit tourism as a powerful means of communication abroad. They opened two ﬂagship stores in areas of high tourism mobility. It is a spot near the city center’s hotels and where the tourists go for their morning stroll, coffee or shopping. We had to do something like it, too . . . as promotion, as image. An investment of tens of thousands of euros . . . but you are obliged to do it. In order to show your uniﬁed image you need to take them to a good spot.

The companies made large investments to open these stores even though they expected them to be unproﬁtable; they did so to enhance their exporting orientation by having a high-quality promotional platform that would serve the global product policy of the ﬁrm (i.e. a retail channel that would reinforce the high-quality, Greek brand concept). The reason was clearly communicational. We knew it beforehand that it would be a loss for us in terms of money. The target was to lose as less as possible and we believed that all this investment could be worth it through the communicational beneﬁts we would have . . . Our target was to promote the brand and everything that this represents abroad; when you have the possibility to develop the brand the way you want in such a premium area and with such a large mobility of tourists, then it would be crazy to let it go.

The importance of such a strategic deviation in the trade policy of the ﬁrms in order to serve the internationalization goal is highlighted by the fact that such actions were implemented despite the fact that they damaged enormously long-established relations with their already existing network of retailers domestically. You may not believe it, but we lost around 100 customers (retailers) when we opened the store at . . .

Another distribution channel was the hotels sector and the amenity products they offer to their guests in the rooms. Firms create a whole range of product categories speciﬁcally aiming the hotel market in order to promote their image among tourists. The interesting point is that again this channel is not worthwhile in terms of proﬁt margins and the only reason was to enhance their position within the tourists segment. . . . the most important (issue) is communication . . . we wanted to exploit the channel of hotels and respectively tourists in order to communicate the brand.

The third ﬁrm also adapted its distribution strategy within Greece in order to serve the purpose of additional sales and communication abroad. The intensiﬁcation of distribution in tourist areas so as to increase the

78

KONSTANTINOS POULIS AND MO YAMIN

degree of contact tourists have with the product was deemed essential in order to increase brand awareness abroad and thus create an initial platform demand for further consumption in their home countries. We try to have high distribution in tourists’ areas so as to have an eponymous demand created but also to be able to respond to that by being at the place where tourists will ask for us.

3.1.2. Internet-Based Communications with Tourists Last but not least, tourism was also the reason the ﬁrms standardized their Internet promotional strategy. After creating the initial connection with the consumer, this relationship had to be further enhanced when tourists went back home. And the best possible way to achieve that is through the Internet. This means of communication presented an excellent opportunity for the ﬁrm to retain the customers ‘gained’ in the tourism destination abroad. There is a special category of emails; an amazing number of 70% of these emails was from tourists. Not from our regular consumer in e.g. Hong Kong but from a tourist who visited Greece and wanted to ask something.

With such an increasing number of tourists communicating with the ﬁrm, the ﬁrms had to take respective actions. We decided to re-activate and upgrade the level of the site which was inactive until then, ﬁll it with information in English only . . . there is no Greek site . . . and give information in which country we are, in which shop we are, by which ways the consumers can buy the product etc. . . . serve the tourists who needed information but they couldn’t ﬁnd. So, we had a demand that wasn’t in our plans and that’s why we reactivated the site. To give the information needed and create a selling point we didn’t have. We can now serve a Swiss tourist. But we don’t have a selling point in Switzerland.

Tourism, therefore, altered the whole logic through which foreign markets are served and gave the impetus to internationalize ﬁrms’ activities after having a platform demand created within Greece. Internet became the medium for this new form of internationalization of the ﬁrm. They (tourists) made us sell through Internet . . . to have in www.x.com a special place to sell because we lose sales at this point. Moreover, in Germany we don’t have a selling point in Munich, only in Berlin and Frankfurt. And especially from Munich we had mails asking for foam baths, other from Paris, other from Italy etc. So, this is in our plans now because we lose sales. We said . . . People who happened to come in contact with X while on holidays in Greece, what do we do with them afterwards?

Firm Y exploited the opportunities presented by the Internet even more by utilizing it as a bridge of communication and relationship development

Tourism as a Leverage of Internationalization for Consumer Goods Firms

79

with prospective tourists before coming to Greece. They adapted their promotional strategy by utilizing online channels in order to integrate tourism with their exporting orientation in the best possible way. The phases through which a tourist passes pre (e.g. information search), during, and post trip were used as steps for the implementation of a holistic customer relationship management program which aims both to increase sales domestically and serve the purpose of internationalization. The question is how we catch them here (pointing abroad – at tourists’ home). Not when they are already here but before they are . . . This must be combined with the exporting budget so even if he does not come you will make him consume there.

Therefore, there is a circular line of focused, adapted promotional actions taken by the ﬁrm which start at the country of residence of tourists and the market where the ﬁrm is active through exports, within Greece while tourists take their holidays and ﬁnally when tourists return home through Internet. We get a lot of emails from foreign tourists who tell us congratulations, I drank your beverage, etc. We have communication with them, we send them gifts and we combine it with our exports . . . we tell them where they can ﬁnd the product so as to continue consuming it.

So, the trial level within Greece does not stop but is retained abroad in addition to the already existing customers who reinforce their relationship with the company even further when they return back home. the volume (of emails) is huge . . . either by people abroad who want to come to Greece for holidays and want to know where they will ﬁnd the products . . . or by people who have been to Greece, met the products here and ask where they can ﬁnd them abroad.

Therefore, the Internet may serve as a tool for companies to overcome environmental uncertainty and may help improve strategic and operational performance for the Greek companies in pursuit of their international customers (Jean, Sinkovics, & Kim, 2008; Yamin & Sinkovics, 2007). Summarizing, we saw that tourism worked along several lines. It created an initial platform of demand, it enhanced the corporate image, it made ﬁrms become knowledgeable of foreign needs and behaviors and altered the channels through which products are distributed by utilizing the Internet as the next step along the continuum between tourism destination and home country. It assisted to evaluate current business practices and improve them, assess levels of critical demand abroad, value the importance of knowledge for foreign operations, and tunnel the creativity of individuals towards innovative ways of thinking.

80

KONSTANTINOS POULIS AND MO YAMIN

Table 3.

Tourism–Exports Linkage. Tourism–Exports Linkage

Contribution of tourism

Respective actions towards international operations

Knowledge Creation of an additional Enhancement of about foreign international consumer base International consumers’ domestically and customer Brand Equity needs and retention abroad behaviors Standardized Standardized Customized Customized brand website utilized promotional tourismpositioning as a tool for a strategies related and product holistic within channels of strategy – The customer Greece distribution Ethnobrand relationship and promotion management within Greece program

Table 3 summarizes the contribution of tourism towards the internationalization of ﬁrms and the respective actions that were implemented to exploit the tourism-exporting linkage.

4. DISCUSSION AND PROPOSITIONS Existing research indicates that the dominant pattern of internationalization in Greece (the country context for the present study) is towards countries considered to be at a relatively low psychic distance from Greece, notably in the Balkans (Louri, Papanastassiou, & Lantouris, 2000), rather than countries that most incoming tourists emanate from. Thus, while tourism can facilitate internationalization through enhanced knowledge of foreign consumer behavior and through brand enhancement, the majority of ﬁrms do not necessarily take advantage of these opportunities. This can be explained by the fact that market knowledge is clearly much broader than knowledge of consumers, and the presence of foreign tourists does not per se reduce (and certainly does not eliminate) other types of knowledge barriers with respect to foreign markets such as knowledge relating to institutional aspects, distribution systems and the business environment of foreign countries more generally (Evans & Mavondo, 2002). Secondly, as our cases appear to indicate, knowledge of ‘foreign’ customer needs to be acquired, through the SMEs engaging with the tourist

Tourism as a Leverage of Internationalization for Consumer Goods Firms

81

segments and is not necessarily generated automatically or through the mere presence of foreign tourists in the domestic market. Our case interviews indicate that ﬁrms treating incoming tourists as simply an additional ‘bonus’ and a source of ‘effortless’ additional revenue without incorporating tourism in their marketing strategy and practice are unlikely to gain sufﬁcient insight into tourist behavior or gain knowledge beneﬁcial to develop exports to the tourists’ home countries. For these ﬁrms, an internationalization pattern based on low psychic distance may be perceived as being much less risky than that based on ‘following’ tourists to the home countries. We can thus put forward the following propositions with regard to the relationship between inbound tourism and ﬁrm internationalization: Proposition 1. Inbound tourism can create ﬁrm knowledge relating to foreign consumer behavior that can help reduce barriers to developing exports to the home country of tourists. Proposition 2. The more ﬁrms focus on tourist segments in their domestic marketing, the more likely it is that they will gain knowledge beneﬁcial to developing exports to the tourists’ home countries. Our case studies also suggest that SMEs can leverage the ‘locality’ of their brands in their export development; in fact, in the context of our study, the development and use of ‘Ethnobrands’ is the concrete manifestation of ﬁrmspeciﬁc knowledge acquired by SMEs through their interactions with their tourist consumers. Thus: Proposition 3. Inbound tourism provides an opportunity for local brand enhancement and internationalization. An interesting and rather unexpected aspect of our ﬁndings is the ﬁrms reliance on Internet communication to build and maintain relationship with their tourist customers; this ﬁnding is particularly interesting, in the light of recent discussions of Internet-enabled ‘disintermediation’ and the effective bypassing of middlemen who are usually cited as a major obstacle for the success of internationalizing ﬁrms (Tiessen, Wright, & Turner, 2001; Mahadevan, 2000). By a focused use of the Internet, SMEs can circumvent previously unavoidable channels of promotion and distribution in a foreign country, which made the internationalization process either extremely costly or not feasible due to the inadequacy to reach these channels. This is particularly relevant in the present context. As noted above, although, incoming tourism can reduce knowledge gap with respect to

82

KONSTANTINOS POULIS AND MO YAMIN

foreign consumer behavior, it does not reduce knowledge gaps with respect to the business and distribution system of foreign countries. However, our interviews indicate that by reliance on Internet-based communications with tourists, ﬁrms can positively leverage their internationalization (Yamin & Sinkovics, 2006). Reﬂecting on our ﬁndings, it is perhaps not a coincidence that ﬁrms that developed their exporting activity in the home countries of tourists use the Internet extensively as a communication and transaction medium. However, the four ﬁrms that do not utilize tourism as a vehicle for internationalization do not emphasize the use of the Internet in any thing like the same degree. Thus, we put forward the following propositions: Proposition 4. Tourism-based ﬁrm internationalization is likely to be hampered by lack of knowledge about the distribution system in the tourists’ home countries. Proposition 5. Reliance on e-commerce can beneﬁt tourism-based internationalization by reducing dependence on the distribution system in the tourists’ home countries.

5. CONCLUDING REMARKS Empirical ﬁndings support the notion that internationalization processes are situation-speciﬁc and inﬂuenced by the context in which they originate, although it is also generally accepted that the pace of internationalization is increasing (Zafarullah, Ali, & Young, 1998; Coviello & McAuley, 1999). The present study is an exemplar of the ‘peculiar’ environment of a touristoriented economy that may generate signiﬁcant opportunities for SMEs that otherwise face signiﬁcantly more arduous barriers to internationalization; a tourism-oriented economy is clearly such a ‘peculiar’ context that deserves closer scrutiny. A particularly interesting insight generated from this study is how the combination of ‘mobile’ consumers and the development of the Internet as a medium for online internationalization (Yamin & Sinkovics, 2006) may be reinforcing each other to positively impact opportunities for rapid internationalization by otherwise ‘disadvantaged’ SMEs. Another interesting ﬁnding is perhaps the identiﬁcation of ‘Ethnobrands’ and their impact on SME internationalization. Despite the fact that local SMEs address to a multicultural audience, the Ethnobrand seems to have an edge over its global counterparts in speciﬁc product categories. The rise of such

Tourism as a Leverage of Internationalization for Consumer Goods Firms

83

ethnic products in music, health treatments, garments, and food among others supports a prediction that Levitt’s (1983) thesis regarding the supremacy of a global brand may be at stake. Needless to say, this research is an exploratory one and does not aim to generalize ﬁndings to other contexts or product categories. The selection of cases was purposeful in order to highlight a not-cited phenomenon so, further research is needed which would include more ﬁrms in more product categories such as food which is a heavily culture-bound product category and would yield very interesting results. Additionally, Greece may be an archetypal tourism destination with a lot of internationalizing SMEs but it is a country which carries a strong and speciﬁc country-of-origin image among foreigners and this may affect their stance towards a Greek product. It is necessary to include ﬁrms originating in countries with diverse country and product images among visitors in order to test the effect of such a factor on the effectiveness of the effort to exploit the qualitative impact of tourism on consumer goods ﬁrms.

NOTE 1. It is also possible that ﬁrm may internationalize by ‘following their customers abroad’. The manufacturer of ‘Iron Brew’ – a popular soft drink in Scotland apparently ﬁrst developed its exports following Scottish holidaymakers to Mediterranean resorts – we are grateful to Stephen Young for this anecdote.

REFERENCES Cheng, J. M. S., Blankson, C., Wu, P. C. S., & Chen, S. S. M. (2005). A stage model of international brand development: The perspectives of manufacturers from two newly industrialized economies – South Korea and Taiwan. Industrial Marketing Management, 34, 504–514. Coviello, N., & McAuley, A. (1999). Internationalisation processes and the smaller ﬁrm: A review of contemporary empirical research. Management International Review, 39(3), 223–256. Craig, C. S., & Douglas, S. P. (2000). International marketing research (2nd ed.). Chichester, UK: Wiley. Denis, J. E., & Depelteau, D. (1985). Market knowledge, diversiﬁcation and export expansion. Journal of International Business Studies, 16, 77–89. Eriksson, K., & Chetty, S. (2003). The effect of experience and absorptive capacity on foreign market knowledge. International Business Review, 12(6), 673–695.

84

KONSTANTINOS POULIS AND MO YAMIN

Evans, J., & Mavondo, F. T. (2002). Psychic distance and organizational performance: An empirical examination of international retailing operations. Journal of International Business Studies, 33(3), 515–532. Ghoshal, S. (1987). Global strategy: An organizing framework. Strategic Management Journal, 8(5), 425–440. Hadley, R. D., & Wilson, H. I. M. (2003). The network model of internationalisation and experiential knowledge. International Business Review, 12(6), 697–717. Inkpen, A. C. (1996). Creating knowledge through collaboration. California Management Review, 39(1), 123–140. Jean, R. J. B., Sinkovics, R. R., & Kim, D. (2008). Information technology and organizational performance within international business to business relationships – a review and an integrated conceptual framework. International Marketing Review, 25(5), 563–583. Johanson, J., & Vahlne, J.-E. (1977). The internationalisation process of the ﬁrm – a model of knowledge development and increasing foreign commitments. Journal of International Business Studies, 8(1), 23–32. Johanson, J., & Vahlne, J.-E. (1990). The mechanism of internationalisation. International Marketing Review, 7(4), 11–24. Johanson, J., & Wiedersheim-Paul, F. (1975). The internationalisation of the ﬁrm-four Swedish cases. Journal of Management Studies, 12(3), 305–322. Kogut, B., & Zander, U. (1993). Knowledge of the ﬁrm and the evolutionary theory of the multinational corporation. Journal of International Business Studies, 24, 625–645. Leonidou, L. C. (2004). An analysis of the barriers hindering small business export development. Journal of Small Business Management, 24(3), 279–302. Levitt, T. (1983). The globalization of markets. Harvard Business Review (May–June), 92–102. Liesch, P., & Knight, G. (1999). Information internationalization and hurdle rates in small and medium enterprise internationalization. Journal of International Business Studies, 30(2), 383–394. Louri, H., Papanastassiou, M., & Lantouris, J. (2000). FDI in the EU periphery: A multinomial logit analysis of Greek ﬁrm strategies. Regional Studies, 34, 419–427. Luostarinen, R. K. (1979). The internationalization of the ﬁrm. Helsinki: Acta Academic Oeconomicae Helsingiensis. MacCannell, D. (1976). The tourist: A new theory of the leisure class. New York: Schocken. Mahadevan, B. (2000). Business models for Internet based e-commerce: An anatomy. California Management Review, 42(4), 55–69. Osland, G. E., & Yaprak, A. (1995). Learning through strategic alliances: Processes and factors that enhance marketing effectiveness. European Journal of Marketing, 29(3), 52–66. Patton, M. Q. (1990). Qualitative evaluation and research methods. Newbury Park, CA: Sage. Theodosiou, M., & Leonidou, L. C. (2003). Standardization versus adaptation of international marketing strategy: An integrative assessment of the empirical research. International Business Review, 12(2), 141–171. Tiessen, J. H., Wright, R. W., & Turner, I. (2001). A model of e-commerce use by internationalizing SMEs. Journal of International Management, 7, 211–233. Yamin, M., & Sinkovics, R. (2006). Online internationalization, psychic distance reaction and the virtuality trap. International Business Review, 15(4), 339–360.

Tourism as a Leverage of Internationalization for Consumer Goods Firms

85

Yamin, M., & Sinkovics, R. (2007). ICT and MNE reorganization – the paradox of control. Critical Perspectives on International Business, 3(4), 322–336. Yang, Y. S., Leone, R. P., & Alden, D. L. (1992). A market expansion ability approach to identify potential exporters. Journal of Marketing, 56, 84–96. Yu, L. (2003). The global brand advantage. Sloan Management Review (Spring), 13. Zafarullah, M., Ali, M., & Young, S. (1998). The internationalization of the small ﬁrm in developing countries – exploratory research from Pakistan. Journal of Global Marketing, 11(3), 21–38.

PART II ENTREPRENEURSHIP AND FUNCTIONAL ADVANCEMENTS

RELATIONSHIPS BETWEEN THE DIMENSIONS OF INTERNATIONAL GROWTH ORIENTATION, ENVIRONMENTAL TURBULENCE, AND STRATEGIC ORIENTATIONS Sanna Sundqvist and Olli Kuivalainen The topic of organizational growth as a focus of entrepreneurship study has attracted considerable attention over the years (see, e.g. Delmar, Davidsson, & Gartner, 2003). However, after a plethora of studies stemming, e.g. from Edith Penrose’s (1959) work on growth of the ﬁrm, theoretical development of the ﬁeld remains fragmented. No consensus has been reached regarding the factors that lead to ﬁrm growth (e.g. Weinzimmer, 2000). It is generally accepted that growth is a multidimensional and, alas, a challenging phenomenon as changes in marketing environment, technological development, global competition, and changing global customer preferences create new challenges for international marketers, for example, and ﬁrms need to adjust their growth strategies and objectives accordingly. Today we know that organizational growth can be achieved in a number of different ways (e.g. in the form of organic growth, acquisitions, etc.) and that the pattern of ﬁrm growth over time can look very different across ﬁrms (see, e.g. Delmar et al., 2003). In this chapter we focus on the organic growth in foreign markets, i.e. on international growth (see also Yli-Renko, New Challenges to International Marketing Advances in International Marketing, Volume 20, 89–123 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020007

89

90

SANNA SUNDQVIST AND OLLI KUIVALAINEN

Autio, & Tontti, 2002), as in many cases the internationalization of the ﬁrm is part of and inseparable from the overall growth and development process of the ﬁrm (Jones, 1999). However, the heterogeneity of ﬁrm characteristics suggests that also internationalization process may be unique to individual ﬁrms (ibid.), and therefore it should be acknowledged that internationalization and international growth can happen in many different ways. For example, there is a difference between the internationalization of so-called born-global ﬁrms that internationalize rapidly after the foundation and sell their outputs in multiple countries and traditionally internationalizing ﬁrms in terms of pace and degree of internationalization with regard to antecedents and ﬁrm characteristics, respectively (see, e.g. Knight & Cavusgil, 1996, 2004). Also knowledge-intensive ﬁrms are reported to have different growth objectives than so-called traditional ﬁrms (Bell, Crick, & Young, 2004). Furthermore, Bell et al. (2004) studied the patterns, processes, and pace of internationalization of smaller knowledge intensive and traditional ﬁrms within the manufacturing sector, and found that even growth objectives and international orientations of ﬁrms differed across these ﬁrms. Correspondingly, previous research suggests that the industry and the orientation and experience of entrepreneurs inﬂuence the nature of internationalization (see, e.g. Jones, 1999). In general, situational variables are known to affect the strategy choice and growth strategy. Several studies have emphasized the fact that growth orientation is a precondition to growth (see, e.g. Autio, Sapienza, & Almeida, 2000). So ﬁrms with modest growth aspirations may also show modest growth ﬁgures, making growth orientation a critical element of ﬁrm growth. Already Penrose (1959) linked ﬁrm’s growth with managerial aspirations (i.e. growth orientation) by saying that: ‘‘the growth of ﬁrms is connected with the attempts of a particular group of human beings to do something’’. Taken that there exist several routes for a ﬁrm to grow in export markets, and the fact that growth orientation precedes growth, we highlight here the importance of identifying different patterns of growth orientation, as they may well determine the actual growth patterns of the ﬁrm and thus have signiﬁcant role in understanding ﬁrm growth behavior in general. Surprisingly, there does not exist to our knowledge any studies focusing on this important topic. Also Thomas A. Stewart, the editor of Harvard Business Review (HBR), has highlighted the challenges that relate to growth decisions. As he put it: ‘‘Tell an experienced manager to cut costs 10%, . . . and he’ll know exactly what to do. He might not like the job, but he’ll know what tools to use, and their handles will feel familiar in his grip. Give a manager the order to grow 10%, however, and he’ll be on alien ground, with few recognizable

Dimensions of IGO, Environmental Turbulence, and EO and MO

91

landmarks or well-marked paths. In part, this is because cause-and-effect relationships are inherently less clear-cut on the upside than on the downside. A search for new customers may come up empty; but a cut in ad spending affects the bottom line immediately. It’s easier to reduce inventory than to increase sales. As Jeffrey Immelt, the CEO of General Electric, says in this month’s lead article, the decisions that produce growth are harder to identify – and harder to make – than the decisions that improve productivity’’ (Stewart, 2006, p. 12). What makes identifying ﬁrm’s growth orientation even more critical is the notion that ﬁrms are likely to select a strategy that best exploits its resources and capabilities relative to external opportunities (see, e.g. Grant, 1991). As different growth patterns require different sets of resources and capabilities, it might be that for a ﬁrm some growth aspirations are totally inappropriate taken its resources and capabilities. For example, a ﬁrm may wish to increase its market share on its main foreign target market (which may also be a very competitive one). It may have size (and thus resource) disadvantage to other larger players in the market, and thus its growth orientation pattern may not be ideal taken the ﬁrm’s internal factors and market external factors. It has been suggested that in situations like this, small ﬁrms should rely more on market niche strategy (see, e.g. Pelham & Wilson, 1996) and they should not challenge market leaders in their game (see Moore, 1998), because in these latter cases ﬁrms just end up wasting their scarce resources. Thus, present study not only identiﬁes the different patterns of international growth orientation (IGO), but explores the internal and external factors that relate to these patterns. Additionally this study will increase our understanding of differences across international ﬁrms’ growth orientations, for example, along their international experience. Interesting question here is whether there exist some common growth patterns that vary over time (or along the ﬁrm experience). We might expect that start up ﬁrms are ﬁrst looking at mainly ways how to enter new foreign markets, and only later will focus more on increasing their export volumes or improving their proﬁtability in these markets (see, e.g. Madsen, 1989). Additionally, this type of thinking is in line with the behavioral or so-called Nordic school of internationalization (cf. e.g. Johanson & Vahlne, 1977) that emphasizes learning-by-doing, and subsequently new knowledge gained from international market operations increases market commitment and internationalization. This, in turn, should make a ﬁrm change its international growth objectives. Consequently, the purpose of the present study is to explore what kind of IGO dimensions exist and are utilized by exporting ﬁrms, and how these

92

SANNA SUNDQVIST AND OLLI KUIVALAINEN

patterns relate to a set of external (environmental) and internal (managerial) factors. In this chapter we propose that the relationship between IGO and actual growth may be dependent on both situational variables and managerial perceptions of drivers of international growth. Although it was only recently that researchers started to identify managerial perceptions of drivers of international growth beside the factors (e.g. small home markets, market niche, knowledge intensity) that ‘‘force/drive’’ companies to internationalize their operations, their role may be inﬂuential. First, the IGO is thought to be a useful construct for differentiating companies according to their motivation to seek growth in international markets, and also for identifying the factors behind the chosen growth strategies. Second, strategic orientations such as market orientation (MO) and entrepreneurial orientation (EO) may also provide a ﬁrm foundation how a ﬁrm should grow. Such orientations can be seen as capabilities on that a ﬁrm builds its international competitive strategy or even key determinants of success (see, e.g. Knight & Cavusgil, 2004). By recognizing the differences in ﬁrms’ IGO patterns and their links with several internal and external factors of the ﬁrms, the chapter aims to enrich the further development in the ﬁeld of internationalization research. The rest of the chapter is organized as follows: First the multidimensional nature of growth orientation and IGO are discussed. Second, the links between growth orientation and international environment and two important strategic orientations, EO and MO are analyzed. In this section of the chapter three research propositions are presented. Third, the research methodology and results of the empirical study focusing on 783 exporting ﬁrms are explained. Finally, we will conclude with the discussion and future research directions.

MULTIDIMENSIONAL INTERNATIONAL GROWTH AND GROWTH ORIENTATION Linking International Growth Orientation and Firm Growth A common weakness of most growth models is the assumption that growth is a desired objective for entrepreneurs (see, e.g. Bird, 1989). However, not all entrepreneurial ﬁrms seek growth as their primary objective (Covin, Slevin, & Covin, 1990; Porter, 1996), and further there is no reason to expect that all entrepreneurs want their businesses to grow in similar ways

Dimensions of IGO, Environmental Turbulence, and EO and MO

93

(Liao & Welsch, 2003). Thus, in many cases researchers have separated growth orientation from actual growth. In previous studies growth orientation has been deﬁned as precondition to growth (see, e.g. Autio et al., 2000). Growth orientation is an attitudinal concept based on subjective evaluation (Nummela, Puumalainen, & Saarenketo, 2005) and is considered to be particularly important for international growth (Yli-Renko et al., 2002). IGO is thought to be a useful construct for differentiating companies according to their motivation to seek growth in international markets, and also for identifying the factors behind the chosen growth strategies (Nummela et al., 2005). Another common weakness in growth research is the fact that growth is either deﬁned vaguely or operationalized differently across studies making it difﬁcult to make theoretical or empirical generalizations (see, e.g. Birley & Westhead, 1990; Delmar et al., 2003). The most often used indicators of growth are based on sales data (see, e.g. Weinzimmer, 2000; Weinzimmer, Nystrom, & Freeman, 1998); other measures that have been used include employment, proﬁts, and assets (see, e.g. Birley & Westhead, 1990; Weinzimmer et al., 1998). In many cases growth is seen as an absolute ﬁgure. However, it is important to note that growth rate measures focus on change over time, for example, relative changes in size of the ﬁrm or revenues. If the growth rate measure is used instead of absolute growth, smaller ﬁrms tend to show higher growth ﬁgures than larger ﬁrms and vice versa (Weinzimmer et al., 1998). Research ﬁeld on the growth orientation is heterogeneous, and across studies various different growth orientation-related concepts are applied: growth orientation (Yli-Renko et al., 2002), growth aspiration (Liao & Welsch, 2003), growth intention (Hay & Kamshad, 1994), growth motivation (Davidsson, 1991), goal orientation (Paswan, 2003), and attitudes toward growth (Wiklund, Davidsson, & Delmar, 2003). These concepts have been operationalized very differently in the extant research, making the comparison of different studies very challenging. For example, Wiklund et al. (2003) have studied the relationship between expected consequences of growth and overall attitude toward growth. They used a single item measure to illustrate the business managers’ attitude toward growth, and operationalized attitudes toward growth by asking the respondents to evaluate whether a 100% increase in the number of employees in 5 years time would be mainly negative or positive. Liao and Welsch (2003) studied the growth aspiration of the technology-based new ventures. They operationalized growth aspiration by entrepreneurs’ projected revenues of the new venture for the ﬁrst year and ﬁfth year of

94

SANNA SUNDQVIST AND OLLI KUIVALAINEN

operation. Both Wiklund et al.’s (2003) and Liao and Welsch’s (2003) growth orientation measures focus on motivational factors behind the ﬁrm growth. Wiklund and Shepherd (2003) measured growth aspiration with four questions. Their measure covers two dimensions of growth: growth in number of employees, and growth in sales, and assess whether managers consider certain growth rates as positive, and their view of the ideal size of the company, i.e. ideal growth. In their study of international growth of Finnish electronics ﬁrms, Autio et al. (2000) included growth orientation as a control variable. They measured growth orientation by combining two statements on the absolute importance of growth to the ﬁrm with the relative measure of the importance of growth in comparison to four other strategic goals. In the relative measure, respondents were asked to allocate 100 points between 5 typical strategic goals to indicate how important they had been to the ﬁrm in the recent past. The goals include maximizing sales growth, proﬁtability, technical superiority, maximizing the values of the ﬁrm for a potential acquisition, and the stability and longevity of the ﬁrm. Yli-Renko et al. (2002) have adapted the growth orientation measure presented by Autio et al. (2000) in their study of international growth of technology-based new ﬁrms. To sum up, in the extant literature the importance of growth orientation is noticed but heterogeneity prevails when the measures are concerned. It has to be pointed out that prior empirical research on IGO is even more limited than research focusing on general growth orientation. Nummela et al. (2005), in their study of IGO of Finnish knowledge-intensive small- and medium-sized enterprises (SME), introduced a multi-item measure for IGO, and found that companies with high orientation were more internationalized than others. However, the measure developed was unidimensional, and measurement items represented managerial mindset, and focused mainly on the fact whether the managers aspired growth from international markets in general. Thus, their measurement items did not cover the dimensions of IGO, i.e. we do not know whether the growth-oriented entrepreneurs are looking for organic growth, or what is the desired way they want to grow internationally. Thus, we propose that IGO research should go a step further, and not only identify whether entrepreneurs and their ﬁrms are internationally growth oriented, but also study the dimensions of international growth intentions, i.e. how do they want to grow. Taken together, although there is a reasonably large group of studies focusing on growth and growth orientation; there is no clear consensus on what are the best single or composite measures of growth and growth orientation. In this study, we use a relative measure to capture the growth

Dimensions of IGO, Environmental Turbulence, and EO and MO

95

orientation: consequently respondents had to prioritize their objectives of the international growth.

Identifying the Patterns of International Growth and Growth Orientation As mentioned earlier, one additional weakness of the growth models is the assumption that there is only one route to growth. Recent research on the topic has conﬁrmed that growth is a multidimensional concept, and that even among high growth ﬁrms there is heterogeneity in growth patterns (Delmar et al., 2003). If we agree that growth is a multidimensional concept, we should critically evaluate whether our current measures for growth orientation are valid, as they do not capture this multidimensional nature of growth and growth orientation but treat growth (orientation) as unidimensional. This view is supported by Hay and Kamshad (1994) who found that the factors that affect the ability of the ﬁrm to grow or earn proﬁts differ from the courses that affect the desirability of growth and proﬁts. Delmar et al. (2003) have noted that the causes and consequences of growth may be different for different forms of growth. As all ﬁrms do not grow in the same way, different forms of growth should be measured with different growth measures. In his extensive review Madsen (1987) classiﬁed the export growth/performance into four groups: exports proﬁtability (the amount of proﬁts), export volume (the absolute value of exports or exports as a percentage of total sales), export growth (the absolute or percentage value of export growth), and various other indicators (e.g. perception of success). Thurik (1993) has suggested that there exist many different kinds of winning strategies for penetrating or developing foreign markets. Additionally, Pelham (1997) has proposed that ﬁrm performance should be operationalized as a multidimensional concept consisting of three dimensions: (a) ﬁrm effectiveness, (b) growth/share (sales level, growth rate, and market share), and (c) proﬁtability. Also in marketing literature, some of the more commonly used growth measures include market effectiveness, sales growth, market share, and proﬁtability (Narver, Jacobson, & Slater, 1999; Pelham, 2000). However, in entrepreneurship studies, the traditional aspect to examine the growth of ﬁrm relates, e.g. to product innovation, market differentiation, or market breadth. In their study of small ﬁrms’ growth strategies, Hay and Kamshad (1994) found diversifying growth patterns among UK ﬁrms. Their study indicated that proﬁt maximization was the single most important objective, while sales growth came in second. Thus, based on the discussion earlier, we adopt

96

SANNA SUNDQVIST AND OLLI KUIVALAINEN

a marketing perspective, and focus on the four most important dimensions of international growth: (a) proﬁtability (see, e.g. Autio et al., 2000; Hay & Kamshad, 1994; Madsen, 1987; Narver & Slater, 1990; Pelham, 2000), (b) volume (Madsen, 1987), (c) market share (Pelham, 1997), and (d) market entry (Thurik, 1993).

Linking Growth Orientation to Firm International Environment and Strategic Orientations Firms do not end up with particular growth patterns at random. Even though the underlying causes of growth may be the same, the situational and managerial/internal factors may determine the speciﬁc form of growth a ﬁrm engages in. Firms are likely to select a strategy that best exploits its resources and capabilities relative to external opportunities (see, e.g. Grant, 1991). In other words, ﬁrms’ growth patterns are related to characteristics of the ﬁrms and their environments (Delmar et al., 2003). Delmar et al. (2003) explored the differential representation of growth pattern for ﬁrms of different demographic afﬁliation in terms of ﬁrm size, ﬁrm age, industry, and ownership/governance. Demographic variables were clearly related to certain growth patterns as large and older ﬁrms from traditional industry sectors dominated the cluster that had grown through acquisitions, for example. As they (Delmar et al., 2003, p. 192) note ‘‘if ﬁrms grow in different ways we can also assume that the reasons leading to growth and the outcome of growth may also be different’’. Instead of just studying demographics, we focus on situational and managerial factors, i.e. on the role of external environmental turbulence and the two most commonly studied forms of strategic orientations (i.e. EO and MO). Role of External Environment To some extent the growth of the ﬁrm will be determined by the growth opportunities in the ﬁrm’s markets, and changes in the business environment can offer new growth opportunities (Drucker, 1985). Dynamic environment is likely to generate opportunities, which the growth-oriented businesses can take advantage of (Covin & Slevin, 1991; Zahra, 1993). In their study of independent Swedish small businesses, Wiklund and Shepherd (2003) found that environmental dynamism magniﬁes the effect the managers’ growth aspirations have on the realization of growth. The less growth-oriented businesses, however, are less alert to opportunities in dynamic environments and can thus even face negative growth in such

Dimensions of IGO, Environmental Turbulence, and EO and MO

97

environments (Wiklund & Shepherd, 2003). Also Hay and Kamshad (1994) have found that growth aspirations differ across external environments, like across different industries. Thus, it is proposed here that: P1. Firms under different environmental conditions will seek export growth differently. Role of Strategic Orientation of a Firm Firms’ internal drivers of growth orientation have been studied earlier. Wiklund et al. (2003) studied the expected consequences of growth on attitude toward growth among a sample of small ﬁrms, and found that small business managers’ overall attitude toward growth was inﬂuenced by the consequences they expected from growth. However, Atuahene-Gima and Ko (2001) argue that an alignment of market and EO processes and practices enables the ﬁrm to adapt to and manage its market environment to meet current and emerging customer needs. They build their argument on Child’s (1972) notion of strategic choice, and Chakravarthy’s (1982) framework of organizational adaptive states. It is believed that ﬁrms have different adaptive orientations based on their pattern of choices with respect to strategy and structure and, as a consequence, adopt different operational procedures and activities (Conant, Mokwa, & Varadarajan, 1990; Miles & Snow, 1978). These adaptive orientations can be called strategic orientations of the ﬁrm (see Gatignon & Xuereb, 1997). The needed orientation may differ depending on the challenges a ﬁrm faces in its marketing environment. A chosen strategic orientation provides a foundation of guidelines for ﬁrms upon which to compete and improve performance (Sinkovics & Roath, 2004). As pointed out earlier the two focal orientations in this chapter are MO and EO. The latter (EO) is an environmental management capability by which ﬁrms embark on proactive and aggressive initiatives to alter the competitive landscape to their advantage. MO, in contrast, is an adaptive capability by which ﬁrms react or respond to conditions in the market environment (Hunt & Morgan, 1995; Kohli & Jaworski, 1990). These strategic orientations are approached here from the international business perspective, i.e. we focus on international EO and export market-oriented behaviors. EO can be seen as a multidimensional construct (consisting of innovativeness, autonomy, risk taking, proactiveness, and competitive aggressiveness (cf. e.g. Covin, Green, & Slevin, 2006; Lumpkin & Dess, 1996). Interestingly, although the EO has often been measured as a single construct, Lumpkin and Dess (1996, 2001) argue that many of the

98

SANNA SUNDQVIST AND OLLI KUIVALAINEN

dimensions are independent. Thus, it is important to study the independent relationships between different EO dimensions and IGO as it is possible that a ﬁrm would only demonstrate one or two dimensions of EO. And, as some of the dimensions are independent, their effect may vary across different types of external environments and contexts (Kreiser, Marino, & Weaver, 2002). Thus, this study focuses on the following three dimensions of EO: (1) risk taking, which represents managers’ willingness to pursue opportunities that carry a reasonable risk of costly failure (Miller & Friesen, 1982); (2) proactiveness, which refers to a posture of anticipating and acting on future wants and needs in the market and would enable a ﬁrm to gain a ﬁrst mover advantage vis-a`-vis its competitors (Lumpkin & Dess, 1996); and (3) competitive aggressiveness, which in its turn reﬂects the intensity of a ﬁrm’s operations in creating the ability to outperform rivals within the industry. The characteristics of this type of behavior can be seen in the way the ﬁrm responds to competitors’ actions (Lumpkin & Dess, 2001). MO represents a set of activities that reﬂect an organization’s degree of adoption of the marketing concept philosophy (Atuahene-Gima, 1996). MO is conceptualized by both Kohli and Jaworski (1990) and Narver and Slater (1990) as a multidimensional organizational phenomenon, where each dimension represents a different feature of MO. It has been proposed that MO’s most critical manifestation is in the form of market-oriented behaviors (Cadogan, Cui, & Li, 2003). Speciﬁcally, market-oriented behaviors comprise the generation of information about the ﬁrm’s markets, the dissemination of this information to relevant decision makers within the ﬁrm, and development and implementation of appropriate responses to the information (Kohli & Jaworski, 1990). MO involves generation and dissemination of market intelligence that is composed of information about the external environment confronting an organization, sharing of this information among all functions in an organization, and rapid managerial action in response to this information (Kohli & Jaworski, 1990). Hence, the use of intelligence is the key to MO (Kumar, Subramanian, & Yauger, 1998). Already in the seminal papers of MO, it is linked with ﬁrm growth, and Slater and Narver (1994a) found the MO signiﬁcantly inﬂuenced ﬁrm growth. The theoretical linkage between MO and ﬁrm behavior (like growth and performance) is based on the concept of sustainable competitive advantage (see, e.g. Day & Wensley, 1983), i.e. on market-oriented ﬁrms’ ability to create long-term superior value for their customers. The concept of sustainable competitive advantage provides a strong base for expectation

Dimensions of IGO, Environmental Turbulence, and EO and MO

99

that MO can offer a ﬁrm the ability to outperform its competitors (Pelham & Wilson, 1996), and thus achieve higher growth rates. Furthermore, Pelham and Wilson (1996) claim that MO leads to more effective marketing decision making, and greater customer retention, which result in greater sales growth and market share. Although main evidence of the ﬁrm growth – MO relationship is positive, also contradictory views are presented. For example, Golann (2006) and Pelham (2000) have proposed that especially for small ﬁrms it might be impossible to achieve high levels of MO during the rapid growth phase of the ﬁrm. It is believed that different growth patterns require different sets of resources and capabilities, and that ﬁrms’ external environment may have an effect on the chosen growth strategy. For example, in growing markets it is relatively easier to increase sales volume, especially if the competition is limited. Increase in volume can thus be achieved by simply selling more to current customers or to new customers. Whereas increasing market share may require ﬁrms to make more effort. Additionally, we propose that different dimensions of international EO and export market-oriented behaviors may relate differently to growth orientation patterns. MO is generally believed to have a positive effect on ﬁrms’ proﬁtability (see, e.g. Greenley, 1995; Jaworski & Kohli, 1993; Slater & Narver, 1994b), and conceptually market-oriented behaviors are based on understanding customer needs. Thus, market-oriented behaviors may be more critical for ﬁrms whose growth aspirations relate proﬁtability and volume. However, international EO is believed to be especially beneﬁcial for ﬁrms seeking growth. It is assumed here that proactiveness and competitive aggressiveness could be essential if ﬁrms are looking for new market entries or increases in market shares. Based on the discussion earlier, following exploratory propositions are presented: P2. The dimensions of international entrepreneurial orientation applied by ﬁrms differ along the international growth orientation patterns. P3. International growth orientation patterns may vary along the different dimensions of export market orientation. In Fig. 1 the three research propositions are presented. It is suggested that international EO and export MO and the environment ﬁrms operate and vary across the IGO patterns. To test these propositions, a survey using mailed questionnaires was conducted. The method of the study is explained in the following section of the chapter.

100

SANNA SUNDQVIST AND OLLI KUIVALAINEN

ENVIRONMENTAL CONDITIONS

ENTREPRENEURIAL ORIENTATION

P3

Fig. 1.

P1

DIMENSIONS OF INTERNATIONAL GROWTH ORIENTATION

MARKET ORIENTATION

P2

The Research Model and Propositions.

METHODOLOGY Data Collection The sample of this survey was drawn from the Kompass Finland database. Companies involved in exporting and employing more than 50 persons were chosen. The database contained 1,205 companies that met the earlier-mentioned requirements. Each ﬁrm was contacted by telephone to (a) determine eligibility and (b) elicit cooperation in the study. The informant in the ﬁrm was either export director/manager, CEO, marketing director/manager or some other person who had the best knowledge of the export function. Of the original 1,205 contacts, 237 proved to be ineligible, leaving a total of 968 eligible. Of these, 21 refused to participate in the study. The questionnaire, together with a cover letter, was mailed within 24 h of the telephone call to those who agreed to participate in the study. Ten days after the initial mailing, a reminder card was sent to non-respondents (the questionnaires were stamped making the identiﬁcation of non-respondents possible). Seven days after the reminder card was mailed, a second questionnaire with a modiﬁed cover letter was mailed to non-respondents. In total, 783 usable responses were obtained, corresponding to the effective response rate of 81% (i.e. 783/968). To test for non-respondent bias the results obtained from the late responses were compared to those of early responses (Armstrong & Overton, 1977). No signiﬁcant differences were found between these two sets of respondents on the variables of interest.

Dimensions of IGO, Environmental Turbulence, and EO and MO

101

Measurement Items IGO was measured by asking respondents to illustrate the weight of importance, which they attached to achieving the export growth objectives by sharing 100 points across these objectives (export sales volume, export market share, export proﬁtability, and market entry). The reasoning behind multiple-indicator measures is that different indicators of growth are attributes of the same underlying theoretical concepts of growth and therefore tend to be correlated. The use of this kind of relative measure is recommended as it avoids the tendency of respondents to give the socially desirable responses that all of the goals are extremely important. Three environment measures were adapted from Cadogan, Diamantopoulos, and Siguaw (1998), who used measures initially developed by Jaworski and Kohli (1993), and subsequently modiﬁed them for use in an export setting. Speciﬁcally, the scales captured technological turbulence (to capture changes and opportunities occurring in the ﬁrm’s export markets as a result of technology), competitor intensity (which assessed the ease with which a ﬁrm can differentiate in its export markets), and customer dynamism (capturing changes in export customer preferences and needs, customer demand and market growth). Technological turbulence scale consisted of four items, customer dynamism was captured with ﬁve items, and competitor intensity scale included six items. All the environmental turbulence items were measured with 7-point Likert scales. The scale anchors ranged from 1 ‘‘very strongly disagree’’ to 7 ‘‘very strongly agree’’. As we were interested to study the effect of separate international EO dimensions, Jambulingam, Kathuria, and Doucette (2005) EO scale was applied. First, we used Jambulingam et al.’s (2005) proactiveness scale, adapted for the international business (exporting) context and extended to identify the degree that managers seized the opportunities in the anticipation of future market conditions. Proactiveness was measured with three items. We then assessed the degree to which managers took risks using items drawn from Jambulingam et al.’s (2005) risk-taking scale, again adapted to gauge the role of risk taking as part of the ﬁrm’s internationalization strategy. Our risk-taking scale consisted of three items. The competitive aggressiveness measure was based on items from Narver and Slater’s (1990) competitor orientation scale, and Jaworski and Kohli’s (1993) market responsiveness scale, and included four items. International EO variables were all measured with 9-point Likert scales. Scale anchors varied from 1 ‘‘strongly disagree’’ to 9 ‘‘strongly agree’’.

102

SANNA SUNDQVIST AND OLLI KUIVALAINEN

The measures of export MO were adopted from Cadogan, Diamantopoulos, and de Mortanges’ (1999) export MO scales. Three measures were used here to capture the quality of the ﬁrms’ export market intelligence generation, dissemination and responsiveness behaviors, respectively. Intelligence generation scale was based on ﬁve items, intelligence dissemination scale consisted of nine items, and responsiveness scale had nine items. All items were measured using 7-point Likert scales with scale anchors 1 ‘‘strongly disagree’’ and 7 ‘‘strongly agree’’. In order to test for unidimensionality, all the items were subjected to an exploratory factor analysis (principal factor analysis with varimax rotation). Those that cross-loaded above the 0.40 level, or that did not load on any factor above that level, were eliminated from the scale. The ﬁnal score for each scale was an average of all the items included in it. The ﬁnal factor loadings for export MO, international EO and environmental turbulence are reported in Tables 1, 2, and 3, respectively. The reliability assessments showed (see Table 4) that all the ﬁnal scales exceeded the level of 0.70 recommended by Nunnally (1978). Furthermore, we control for ﬁrm size and their experience. Firm size has long been associated with a variety of important individual and organizational characteristics (see, e.g. Schminke, Ambrose, & Cropanzano, 2000), and is thought to be an approximation of ﬁrm resources that also relate with the probability of international activity (Aaby & Slater, 1989; Katsikeas, 1994). Firm size was measured by the turnover (mh). The role of ﬁrm’s experience has long been acknowledged in theories of experiential learning and stages theories of internationalization (Johanson & Wiedersheim-Paul, 1975; Welch & Luostarinen, 1988). Thus, we expect that more experienced ﬁrms are more likely to seek export proﬁts as they have already grown internationally, and are now at the phase of improving their export market positions. Firm’s experience was measured by the years in business, years exporting, and the number of export countries.

Analyses and Results First the two-step cluster analysis was applied to identify the different dimensions of IGO. Two-step cluster analysis was performed using the four measures describing the importance of each growth dimension as the clustering variates. Five distinct set of export growth patterns were identiﬁed. As Table 5 reveals cluster 1 consists of ﬁrms that are very proﬁt oriented, and they consider export proﬁts as the most important element of

.455 .571 .584 .490 .544

.593 .701 .746 .660 .704

.414

.483

.580

.629

.699

.738

.725

.803

.527

.483

.564

.691

.714

.835

PC3

.400

PC2

Communalities

.491

PC1

Rotated Pattern Loading

Factor Analysis of Export Market Orientation Items.

In this company, we generate a lot of information concerning trends (e.g. regulation, technological developments, political, economy) in our export markets We are slow to detect fundamental shifts in our export environment (e.g. regulation, technology, economy) (reversed) We periodically review the likely effect of changes in our export environment (e.g. regulation, technology) We do not generate enough reliable/relevant information concerning our competitors’ activities in our export markets (reversed) We generate a lot of information in order to understand the forces which inﬂuence our overseas customers’ needs and preferences There is minimal communication between the export and other departments concerning foreign market developments (e.g. regulation, technology) (reversed) When one department ﬁnds out something important about foreign market competitors, it is slow to alert other departments (reversed) Too much information concerning our export market competitors is discarded before it reaches decision-makers (reversed) All information concerning our export competition is shared within this company Information which can inﬂuence the way we serve our export customers takes forever to reach export personnel (reversed) Important information about our export customers is often ‘‘lost in the system’’ (reversed) Information about our export competitors’ activities often reaches relevant personnel too late to be of any use (reversed) Important information concerning export market trends (regulation, technology) is often discarded as it makes its way along the communication chain (reversed)

Item

Table 1.

Dimensions of IGO, Environmental Turbulence, and EO and MO 103

8.315 34.645 34.645

.538

PC1

.486 .353 .434 .431

.524 .625 .577 .512

.592 .505 .577 .479

.679 .767 .754 .638 2.168 9.032 43.677

.332 .346

1.589 6.622 50.300

PC3

Communalities

.460

PC2

Rotated Pattern Loading

Note: Extraction method: principal component (PC) analysis. Rotation method: Oblimin with Kaiser normalization.

Export sales personnel rarely share their experiences of dealing with customers with others (reversed) For one reason or another, we tend to ignore changes in our foreign customers’ product or service needs (reversed) We periodically review our product development efforts to ensure that they are in line with what foreign customers want The product lines we sell to foreign markets depend more on internal politics rather than real market needs (reversed) When we ﬁnd out that export customers are unhappy with the quality of our service, we take corrective actions immediately Several ‘‘departments’’ get together periodically to plan a response to changes taking place in our foreign business environment (e.g. regulation, technology, etc.) Our export business strategies are driven by our beliefs about how we can create greater value for export customers Our export strategy for competitive advantage is based on our understanding of export customer needs Our export business objectives are driven primarily by customer satisfaction We give close attention to after-sales service in our export markets Eigenvalue Percentage of variance Cumulative percentage

Item

Table 1. (Continued )

104 SANNA SUNDQVIST AND OLLI KUIVALAINEN

2.114 21.144 57.812

Note: Extraction method: principal component analysis. Rotation method: Oblimin with Kaiser normalization.

3.667 36.668 36.668

.521 .680 .449 .699

.712 .811 .635 .790 1.301 13.014 70.826

.806 .798

.757

.825

.896 .887

.792

.867

.794

.787

PC3

.856

PC2

Communalities

.860

PC1

Rotated Pattern Loading

Factor Analysis of International Entrepreneurial Orientation Items.

Export managers in this company usually take action in anticipation of future export market conditions Export managers in this company try to shape our business environment to enhance our presence in the export market Export managers in this company continually seek out new opportunities, because export market conditions are changing Export managers in this company see taking gambles as part of our strategy for export success Export managers in this company take above average risks Export managers in this company see taking chances as an element of our export strategy If a major competitor were to launch an intensive campaign targeted at our foreign customers, we would implement a response immediately We are quick to respond to signiﬁcant changes in our competitors’ price structures in foreign markets We take forever to decide how to respond to our export competitors’ price changes (reversed) We rapidly respond to competitive actions that threaten us in our export markets Eigenvalue Percentage of variance Cumulative percentage

Item

Table 2.

Dimensions of IGO, Environmental Turbulence, and EO and MO 105

1.844 12.296 56.569

.701 .844 .609

.796 2.678 17.856 44.273

.522 .726 .402

.777 .604

.882 3.962 26.416 26.416

.594 .423 .597 .458 .264 .443 .642 .367

PC3

.776 .741 .733

.769 .630 .766 .652 .503 .664

PC2

Communalities

.870 .858 .833

PC1

Rotated Pattern Loading

Factor Analysis of International Environmental Turbulence Items.

Industry competition in our export markets is cut-throat In our export markets, there are many ‘‘promotion wars’’ In our foreign markets, price competition is a hallmark of our industry One hears of a new competitive move in our export markets almost every day Our export competitors are relatively weak (reversed) In our foreign markets, aggressive selling is the norm Our export customers’ product preferences change quite a bit over time New export customers tend to have product-related needs that are different form those of our existing export customers Our export customers tend to look for new products all the time Our export customers tend to have stable product preferences (reversed) We are witnessing changes in the type of products/services demanded by our export customers The technology in our industry in changing rapidly Technological changes provide big opportunities in our industry A large number of new product ideas have been made possible through technological breakthroughs in our industry Technological developments in our industry are rather minor (reversed) Eigenvalue Percentage of variance Cumulative percentage

Item

Table 3.

106 SANNA SUNDQVIST AND OLLI KUIVALAINEN

Dimensions of IGO, Environmental Turbulence, and EO and MO

Table 4. Scale

107

Scale-Reliability Analyses. Cronbach’s Alpha

Number of Items

Mean Score

Standard Deviation

Environment Competitor environment Customer environment Technology environment

.748 .772 .896

6 5 4

4.946 4.306 4.334

1.024 1.174 1.504

Entrepreneurial orientation Proactiveness Risk taking Competitive aggressiveness

.856 .872 .750

3 3 4

6.056 3.979 4.930

1.509 1.775 1.095

Market orientation Intelligence generation Intelligence dissemination Intelligence responsiveness

.807 .869 .839

5 9 9

4.883 5.115 5.114

1.129 1.049 .943

their international growth. Firms in cluster 2 put also a strong emphasis on export proﬁtability, but compared to ﬁrms in cluster 1, these ﬁrms will also seek for high export volume. Firms who did not emphasize any particular growth dimension formed the third cluster. Thus, cluster 3 consists of ﬁrms who did not have a speciﬁc growth orientation focus, but considered all dimensions equally important. They seek simultaneously high export volume and proﬁts, and want to enter new markets and have large market shares. However, compared to their exporting counterparts, these ﬁrms seem to have a bit stronger emphasis on increasing their export market shares. Firms in cluster 4 consider new market entries as the most important dimension of international growth. These ﬁrms do not want to be market leaders, i.e. ﬁrms in this cluster are probably ﬁrms entering into the new market areas with steady growth strategy. Whereas ﬁrms in cluster 5 are the ones seeking mainly export volume. In the next stage, we explored demographical variables separating these ﬁrms. General examination of demographics (Table 5) revealed that growth rate among ﬁrms in cluster 1 (i.e. ﬁrms emphasizing the export proﬁtability) was the most modest (9.4%), and that their international EO was the lowest among the studied ﬁrms. Additionally, the external environment of these ﬁrms appeared to be quite stable. Firms in cluster 2 represent ‘‘safe mediocrity’’. Whereas ﬁrms in cluster 3 stand out based on their international experience-related demographics: These ﬁrms have the longest

72.333 8.350

export proﬁtability (%) international market

9.988 7.576

8.249 7.067

Standard deviation

1.046 1.536 1.018 1.315 1.126

4.843 4.317 4.690 4.207 3.113

28.842

20.96

5.158 4.364 5.058 4.373 3.138

5.078

21.20

12.8 26.67

29.497

9.4 27.13

.999

52.2

4.588

6.318 7.667

9.218 7.400

Standard deviation

.844 1.443 .984 1.176 1.166

1.008

19.817

24.356

Food, drinks, tobacco (4.8%) Textiles (4.8%) Forest (15.7%) Chemical (11.3%) Metal (14.3%) Machinery (20.9%) Electronics (10.4%) Other manufacturing (7.4%) Services (10.4%) Industrial goods (71.1%)

46.369 12.661

28.291 12.584

Mean

45.0

Food, drinks, tobacco (1.7%) Textiles (3.4%) Forest (12.1%) Chemical (7.8%) Metal (18.1%) Machinery (22.4%) Electronics (8.6%) Other manufacturing (6.9%) Services (19.0%) Industrial goods (67.2%)

12.368 6.863

export volume (%) export market shares

Type of product (industrial versus consumer goods) Percentage of sales from international markets (%) Growth (% during the past 3 years) International experience (years exporting) Scope of international activities (number of export countries) International entrepreneurial orientation International market orientation Technology environment turbulence Competitor environment turbulence Customer environment turbulence Regulatory environment turbulence

Importance of Importance of (%) Importance of Importance of entry (%) Industry

Mean

Cluster 2: Export Volume and Proﬁtability Emphasized N ¼ 232

7.971 9.746

8.810 9.273

Standard deviation

5.241 4.503 5.031 4.296 3.209

5.183

30.86

14.7 28.84

54.0

.724 1.467 1.012 1.128 1.284

1.008

30.062

32.907

Food, drinks, tobacco (4.4%) Textiles (6.0%) Forest (10.9%) Chemical (10.4%) Metal (15.8%) Machinery (21.9%) Electronics (12.6%) Other manufacturing (6.6%) Services (11.5%) Industrial goods (74.6%)

25.540 16.269

28.761 29.043

Mean

Cluster 3: Not a Speciﬁc Growth Focus N ¼ 186

10.929 13.961

10.962 6.239

Standard deviation

4.770 4.433 5.802 4.196 3.057

4.883

13.15

15.3 16.61

32.2

.933 1.640 1.121 1.059 1.227

1.149

17.998

16.967

Food, drinks, tobacco (9.1%) Textiles (2.3%) Forest (8.0%) Chemical (15.9%) Metal (10.2%) Machinery (21.6%) Electronics (6.8%) Other manufacturing (3.4%) Services (22.7%) Industrial goods (67.4%)

25.811 44.833

21.333 8.022

Mean

Cluster 4: Market Entry Focused N ¼ 90

Cluster Solutions of International Growth Orientation Dimensions.

Cluster 1: Export Proﬁtability Focused N ¼ 117

Table 5.

12.382 7.989

9.554 6.762

Standard deviation

4.915 3.952 4.981 4.381 2.953

5.040

21.11

11.0 28.64

47.3

.823 1.495 .995 1.173 1.085

1.055

21.214

32.879

Food, drinks, tobacco (7.6%) Textiles (5.6%) Forest (18.1%) Chemical (8.3%) Metal (13.9%) Machinery (19.4%) Electronics (9.0%) Other manufacturing (7.6%) Services (10.4%) Industrial goods (62.3%)

29.726 8.702

54.664 6.661

Mean

Cluster 5: Export Volume Focused N ¼ 146

Dimensions of IGO, Environmental Turbulence, and EO and MO

109

international experience (they have been exporting almost for 29 years), the scope of their international activities was widest (they operated in almost 30 countries), and they derived 54% of their turnover from foreign markets. Furthermore, they seem to be the most entrepreneurially oriented and market oriented. Additionally, their external environment appears to be quite turbulent (especially their markets can be described with high technology turbulence). Contrary to cluster 3 ﬁrms, ﬁrms in cluster 4 have the least international experience (only 16 years exporting with 13 exporting countries). Firms in cluster 4 scored low on export MO. Their external environment can be characterized with high competitor environment turbulence and low levels of customer environment turbulence. Yet, these market entry focused ﬁrms experience the highest growth (15.3%). Export volume focused ﬁrms in cluster 5 did not have any particular demographic features, but their customer environment was the most turbulent. Examination of industry distributions did not reveal differences across ﬁrms in different clusters. Thus, it seems that so-called push effect from home markets may be similar for each cluster. One of the reasons for this result may be that all of the ﬁrms are from the small home market, i.e. Finland. The most interesting outcome here is that service ﬁrms seem to be more market entry focused. As an additional check, an analysis of variance (ANOVA) was performed to compare ﬁrm demographics (i.e. size, experience, international experience, and scope of international activities). As Table 6 reveals variables reﬂecting ﬁrm size and experience do not vary signiﬁcantly across the ﬁve clusters. However, signiﬁcant differences were detected for variables reﬂecting international experience and the scope of international activities. Especially, ﬁrms in cluster 3 that did not have any speciﬁc single growth objective but which put a slightly more emphasis on increasing their export market shares, appear to be more international than ﬁrms in other clusters. However, ﬁrms in cluster 4 emphasizing importance of market entries as their vehicle for export growth appeared to have the most limited international experience. Then, at the third stage we studied if there were signiﬁcant differences across external (environmental) variables. In line with our expectations, it seems that environmental differences relate to different export growth patterns (see Table 7). Both competitive environment turbulence and technology environment turbulence were signiﬁcantly different across clusters. Cluster 3 ﬁrms, which did not have any speciﬁc growth emphasis (although these ﬁrms highlighted relatively more the importance of gaining large export market shares), seem to operate in signiﬁcantly more turbulent exports markets when it comes to technological developments. Whereas,

Size (number of employees) Size (turnover) Experience (years in business) International experience (years exporting) Number of export countries

678.90

204.33 53.09

27.13

20.96

115 116

116

116

Mean

116

N

28.84

29.50

867.60 41.45

2144.59

Standard deviation

Cluster 1: Export Proﬁtability Focused

228

230

227 228

230

N

21.20

26.67

244.93 51.49

795.23

Mean

19.82

24.36

1253.26 55.66

3252.49

Standard deviation

184

184

184 182

185

N

30.86

28.84

205.12 56.79

1244.76

Mean

30.06

32.91

1051.86 48.32

5546.41

Standard deviation

Cluster 3: Not a Speciﬁc Growth Focus

88

88

88 88

90

N

13.15

16.61

75.51 42.30

440.71

Mean

18.00

16.97

141.85 34.12

867.29

Standard deviation

Cluster 4: Market Entry Focused

Comparison of Demographics across Clusters.

Cluster 2: Export Volume and Proﬁtability Emphasized

Table 6.

144

146

145 145

146

N

21.11

28.64

210.57 57.96

938.82

Mean

21.21

32.88

1092.46 43.98

4221.24

Standard deviation

Cluster 5: Export Volume Focused

9.168

3.183

.424 1.854

.825

F

.000

.013

.791 .117

.509

Signiﬁcance

ANOVA

Competitive environment turbulence Customer environment turbulence Technology environment turbulence

Mean

4.690

4.207

4.312

N

115

116

116

1.536

1.315

1.018

Standard deviation

232

232

231

N

4.364

4.373

5.058

Mean

1.443

1.176

.984

Standard deviation

Cluster 2: Export Volume and Proﬁtability Emphasized

185

185

185

N

4.503

4.296

5.031

Mean

1.467

1.128

1.012

Standard deviation

Cluster 3: Not a Speciﬁc Growth Focus

90

89

90

N

4.433

4.196

4.802

Mean

1.640

1.059

1.121

Standard deviation

Cluster 4: Market Entry Focused

145

145

145

N

3.952

4.381

4.981

Mean

1.495

1.173

.995

Standard deviation

Cluster 5: Export Volume Focused

F

3.064

.742

.016

.564

.010

Signiﬁcance

ANOVA

3.343

Linking External Environmental Turbulence and International Growth Orientation Patterns.

Cluster 1: Export Proﬁtability Focused

Table 7.

112

SANNA SUNDQVIST AND OLLI KUIVALAINEN

technology environment was most stable among ﬁrms in cluster 5, i.e. among ﬁrms that pursued to grow via increases in export volume. The competitive environment turbulence was highest among ﬁrms that emphasized both export volume and proﬁtability in cluster 2, and lowest among ﬁrms that focused on export proﬁtability (cluster 1). However, no signiﬁcant differences were detected across customer environment turbulence. Thus, P1 is only partially supported. In the ﬁnal stage of the analyses, we studied the relationships between internal factors to ﬁrm, i.e. strategic orientations and ﬁrms’ international growth patterns. As Table 8 reveals, strategic orientations differed across clusters, i.e. ﬁrms striving for different growth patterns utilized different strategies (which is a result that supports both P2 and P3). Signiﬁcant differences can be detected in all other variables except proactiveness. Firms emphasizing export proﬁtability appear to be least entrepreneurially oriented (no risks taken). Cluster 2 ﬁrms were the most proactive, and additionally they seem to be good at generating and disseminating export intelligence and in responsiveness. Firms in cluster 3 were aggressive in competition and what it comes to export market-oriented behaviors they generated most export market information and appeared to be very active in responding to the collected information. Market entry focused ﬁrms in cluster 4 scored relatively low in EO and their levels of export market intelligence generation and dissemination were the lowest among the studied companies. Firms emphasizing the export volume as their vehicle for international growth were the bold risk takers, but did not perform well in responding the market intelligence. This type of behavior may relate to the fact that ﬁrms may have to compete with the price and the focal customer segment may be large. In the following chapter, the demographics and external and internal factors of each cluster are summarized (see also Table 9). It appears that especially ﬁrms in cluster 3 and cluster 4 distinguished from others based on their past experiences in international markets, as ﬁrms in cluster 3 had signiﬁcantly longer experience and as their scope of international operations was very wide, whereas cluster 4 ﬁrms were at the very beginning of their internationalization. Additionally, our results indicate that high levels of EO and MO relate to higher levels of international activities, and that ﬁrms with limited international experience do not behave entrepreneurially or in market-oriented way. Interestingly, the rest three clusters (cluster 1, cluster 2, and cluster 4) did not differ signiﬁcantly based on their past international behavior. Thus, a detailed examination of their external and internal factors is needed.

Proactiveness Risk taking Competitive aggressiveness Intelligence generation Intelligence dissemination Intelligence responsiveness

Table 8.

5.802 3.431 4.547

4.602

4.945

4.979

116

115

117

Mean

116 116 117

N

1.077

1.254

1.241

1.617 1.622 1.213

Standard deviation

Cluster 1: Export Proﬁtability Focused

229

228

232

229 228 231

N

5.215

5.246

5.006

6.221 3.946 5.036

Mean

.890

.995

1.127

1.436 1.673 1.070

Standard deviation

Cluster 2: Export Volume and Proﬁtability Emphasized

184

185

185

186 185 185

N

5.307

5.223

5.197

6.143 4.189 5.199

Mean

.772

.896

.963

1.371 1.766 .986

Standard deviation

Cluster 3: Not a Speciﬁc Growth Focus

90

88

89

89 89 89

N

5.022

4.812

4.445

5.835 4.082 4.730

Mean

1.027

1.184

1.217

1.678 1.968 1.079

Standard deviation

Cluster 4: Market Entry Focused

146

146

145

145 144 146

N

4.887

5.097

4.779

6.044 4.208 4.875

Mean

.964

1.003

1.032

1.544 1.884 1.054

Standard deviation

Cluster 5: Export Volume Focused

5.732

4.055

10.205

2.170 4.141 8.093

F

.000

.003

.000

.071 .003 .000

Signiﬁcance

ANOVA

Linking International Entrepreneurial Orientation and Export Market Orientation with International Growth Orientation Patterns.

High in responsiveness

Low

Entrepreneurial orientation

Most proactive

High

High Moderate High competitive environment turbulence

Moderate Low Relatively stable

Market orientation

Moderate

Cluster 2: Export Volume and Proﬁtability Emphasized

Low

Cluster 1: Export Proﬁtability Focused

High in intelligence generation and responsiveness Most entrepreneurially oriented. Very aggressive in competition

Very high High High technology environment and regulatory environment turbulence

High

Cluster 3: Not a Speciﬁc Growth Focus

Relatively low

Low High High levels of competitor environment turbulence and stable customer environment Low

Very low

Cluster 4: Market Entry Focused

Bold risk taking

Very low in responsiveness

Moderate Moderate High customer environment turbulence

High

Cluster 5: Export Volume Focused

Summary of the External and Internal Factors across International Growth Orientation Clusters.

International experience Export dependency Growth (%) Environment

Table 9.

114 SANNA SUNDQVIST AND OLLI KUIVALAINEN

Dimensions of IGO, Environmental Turbulence, and EO and MO

115

Cluster 1: Export Proﬁtability Focused Firms These ﬁrms considered export proﬁts as the single most important element of the export growth strategy. This is well reﬂected in their growth rates, which were the most modest (9.4%) of the studied ﬁrms. Additionally, the external environment of these ﬁrms appeared to be quite stable. There seems not to be that much competitive pressure and ﬁrms may be able to utilize skimming as their pricing strategy. And furthermore, as their international EO was lowest, it appears that for these ﬁrms internationalization may play a lesser role in their corporate strategy. Cluster 2: Firms Emphasizing Proﬁtability and Volume Firms in cluster 2 put a strong emphasis on export proﬁtability, but compared to ﬁrms in cluster 1, these ﬁrms seek also increase in export volume. Their international experience was on the average level (26 years exporting, and operation in 21 foreign markets). Also their growth rate was moderate (12.8%). These ﬁrms operate in export markets that can be described as highly competitive. Thus, it appears that ﬁrms in cluster 2 challenge their competitors by following proactive export behavior and strategies. Additionally, they seem to be good at generating and disseminating export intelligence and in responding to foreign market intelligence. This positive linkage between MO and high environmental turbulence has been reported previously by Day and Wensley (1988), and Kohli and Jaworski (1990) who proposed that the relationship between MO and performance will be moderated by the competitive environment. Cluster 3: Firms with No Clear Growth Focus These ﬁrms seek simultaneously high export volume and proﬁts, and want to enter new markets and have large market shares. What separates them from other ﬁrms in the sample is their stronger emphasis on increasing their export market shares. Examination of their international experience revealed that these ﬁrms have long experience in exporting and thus their scope of international activities is also wide, as they operated in almost 30 countries, and as 54% of their turnover was generated from foreign markets. Maybe due to the wide scope of export markets, they felt that their external environment is very turbulent – especially the technology environment. This is a logical consequence of operating in many distant markets with different technology standards. Furthermore, they were the most entrepreneurially oriented and market oriented. Their entrepreneurial

116

SANNA SUNDQVIST AND OLLI KUIVALAINEN

behavior is mainly characterized by aggressive competitive strategies. Whereas their market-oriented behavior can be described with active export intelligence generation and responsiveness. To sum up, with the objective of gaining larger markets shares and high EO and MO, many ﬁrms in this cluster seem to be in such a phase in their internationalization that they are challenging a market leader in some of their chosen key markets.

Cluster 4: Market Entry Focused Firms Firms in cluster 4 consider new market entries as the most important dimension in their international growth. This growth orientation can easily be explained by the fact that these ﬁrms were most inexperienced in international business. Yet, these market entry focused ﬁrms experienced the highest growth (15.3%), which is inline with the notion of Weinzimmer et al. (1998) who pointed out that smaller ﬁrms tend to show higher growth rate ﬁgures than larger ﬁrms. Their external environment can be characterized with high competitor environment turbulence and low levels of customer environment turbulence, i.e. it is difﬁcult for these ﬁrms to differentiate in their export markets characterized with relative stable customer needs. This indicates that these ﬁrms are operating in relatively stable markets, with established products and competitive structures. Market entry focused ﬁrms scored relatively low in EO and export MO. Probably, partly due to the environmental stability of their technology and customer environment, ﬁrms in this cluster would not beneﬁt of developing high levels of entrepreneurial or MO. An interesting outcome here is that service ﬁrms seem to be mainly market entry focused, although service ﬁrms are represented in other clusters as well.

Cluster 5: Export Volume Focused Firms Firms in this cluster are seeking mainly export volume. They do not have any particular demographic features, but can be described as having quite high levels of international experience, and moderate international growth. Their technology environment is the most stable across the clusters, but their customers’ needs are changing rapidly. Under these conditions, these ﬁrms will take above average risks. But, surprisingly they seem to neglect responsiveness, which should be the key for ﬁrms under such conditions. It may be that these ﬁrms operate with a standardized products and markets – volume is more important than adapting to local customer needs.

Dimensions of IGO, Environmental Turbulence, and EO and MO

117

DISCUSSION AND CONCLUSIONS Delmar et al. (2003) have noted that the study of ﬁrm growth is heterogeneous in nature, and that more research is needed here, as the extant empirical research on organizational growth has, with a few exceptions, failed to recognize this heterogeneous nature of growth and growth orientation. Consequently, the objective of this study was to explore what kind of IGO patterns exist and how these patterns relate to a number of internal and external factors reﬂecting ﬁrms’ behavior and strategies. It is of importance to note that our focus is on exporting, which is one of the most common ways to internationalize; as most of the small ﬁrms start their international operations and international growth through exporting. We do not suggest though that all exporting ﬁrms would grow in a same way; in line with Bell et al. (2004) among others, we propose that exporters are a heterogeneous set of ﬁrms, and that their international growth strategies, orientations, and patterns differ. In this research context (i.e. Finnish exporting ﬁrms and their IGO) probably one of the most surprising results is that ﬁrms within cluster 3, which did not seek any speciﬁc type of growth and thus competed on ‘‘all fronts’’ instead, were most entrepreneurially and market oriented. In other words, ﬁrms in cluster 3 seek to grow in multiple ways in international markets. Their export intensity and scope of internationalization were also higher than among ﬁrms in other clusters. Although this chapter does not focus on actual export performance, it can be speculated that these ﬁrms already know better what it takes to compete successfully in international markets. International EO and management and personnel motives have been emphasized in many studies (see, e.g. Autio et al., 2000; Knight & Cavusgil, 2004) as key determinants for success. Adoption of export marketoriented behavior is also seen as one of the possible route for superior export performance (Cadogan, Paul, Salminen, Puumalainen, & Sundqvist, 2001). Market-oriented behavior may also be linked to the wider geographic scope: there is a need to generate and utilize market information of distant markets as well when the number of markets increases. The distance concept implies that the greater the difference in consumer behavior, language, cultural standards, or purchasing power, for example, the greater the distance between the markets (see, e.g. Sto¨ttinger & Schlegelmilch, 1998); a ﬁrm operating in distant markets needs market information to overcome the inﬂuence of psychic or geographical distance. Bell et al. (2004) have, for example, noted that environmental variables (such as domestic and international market conditions) seem to be particularly important in initial business development. The signiﬁcant effects

118

SANNA SUNDQVIST AND OLLI KUIVALAINEN

that became apparent in our study were higher technological turbulence among the ﬁrms in cluster 4 (market entry focused ﬁrms), and high competitive environment turbulence among ﬁrms in cluster 2 (i.e. among ﬁrms that emphasized export volume and proﬁtability). As the industry distribution (see Table 1) was quite equal across the clusters, this result cannot be explained by the industry structure factors, although some of the previous studies have proposed that there is meaningful inference in international growth across, for example, high technology versus traditional manufacturing ﬁrms. As mentioned earlier these ﬁrms were operating in multiple markets, and the large number of markets may create more turbulence for these ﬁrms or it may be that some of the focal markets are more turbulent than the markets the ﬁrms in other clusters were operating in. Regarding other differences among the clusters, it is worthwhile to note that ﬁrms in cluster 3 were younger and smaller than ﬁrms in other clusters. These ﬁrms were focusing on market entry. Naturally this relates to the fact that their export intensity was still rather low (approximately 32.2%). However, it can also be pointed out that, for example, Madsen (1989) found that it would actually be better for very small ﬁrms to spread their efforts over several markets, as they may not have the resources to follow a market concentration strategy successfully. Thus, our results support Madsen’s study. To conclude, the present study conﬁrms the ﬁndings of general growth studies suggesting that growth is a multidimensional phenomenon, and expands our knowledge in export growth studies by introducing four IGO patterns. One explanation to the results of this study seems to be that ﬁrms adapt their growth objectives and orientations in a way that these match the challenges the ﬁrms are facing. For example, high scores in the strategic orientations such as EO and MO are responses to internal and external challenges and problems. MO, and especially good exporting/foreign market intelligence, was seen of importance for ﬁrms that emphasized proﬁtability, for example. To be able to follow high-price strategy in a competitive environment there is often a need to create extra value for one’s customers. In such a situation close relationships should help: to be able to build close customer relationships understanding customers’ needs and tailoring one’s offerings to those needs is often a necessity. Here market-oriented behaviors should be helpful. In line with Jones and Coviello (2005) we also suggest that the type of entrepreneurial internationalization is linked to various aspects of ﬁrm behavior. One explanatory factor to the question, what type of strategic orientation is of importance, is clearly the internationalization phase or stage a ﬁrm is in and what are the key challenges of that phase. This type of ‘‘matching behavior’’ found in our sample seems rather logical to us and

Dimensions of IGO, Environmental Turbulence, and EO and MO

119

consequently, practicing managers were wise to analyze the environment a ﬁrm is operating and to adjust their strategies accordingly. Here, the managers could do well if they would be capable to develop and foster an organizational culture that would be able to support the chosen growth objectives and growth strategy in the most efﬁcient manner. The challenge lies in many cases in the ability to recruit the right people and in the ability to make changes in the organization in the right time. In summary, we believe that the primary value of this study is the insight that future ﬁrm growth research would beneﬁt by recognizing differences in ﬁrm IGO patterns. The good understanding of ﬁrms’ capabilities and resources are also of importance as they provide a stable basis for ﬁrms to develop orientations and strategies that engender superior export performance (Knight & Cavusgil, 2005; Teece, Pisano, & Shuen, 1997). Limitations and Further Research There are some limitations concerning our study. First, it is important to note that our analysis is based on cross-sectional data and we do not know the future paths of the ﬁrms or possible changes in the IGO. Second, the cause–effect relationships were not studied in detail. Thus, present study probably provides more questions than it offers answers. Further studies should focus on investigating the reasons why ﬁrms have different growth aspirations, i.e. study the causal relationships between IGO patterns and their antecedents, as well as consequences. One explanation for the detected differences is offered by the ﬁrm life cycle theory: ﬁrms at different growth phases may need to adapt their growth orientation patterns, and thus future studies should control for ﬁrms’ international growth phase. Here the focus could also be on critical incidents or phases of growth (see, e.g. Greiner, 1998) and subsequent changes in IGO. Additionally, in this chapter we have left out the performance consequences. However, in future studies the focus should be on examining the relationships between external and internal factors, growth and growth orientation, and performance.

REFERENCES Aaby, N. E., & Slater, S. F. (1989). Management inﬂuences on export performance: Review of the empirical literature, 1978–1988. International Marketing Review, 6(4), 7–26. Armstrong, S. J., & Overton, T. S. (1977). Estimating non-response bias in mail surveys. Journal of Marketing Research, 14(3), 396–402.

120

SANNA SUNDQVIST AND OLLI KUIVALAINEN

Atuahene-Gima, K. (1996). Market orientation and innovation. Journal of Business Research, 35(2), 93–103. Atuahene-Gima, K., & Ko, A. (2001). An empirical investigation of the effect of market orientation and entrepreneurship orientation alignment on product innovation. Organization Science, 12(1), 54–74. Autio, E., Sapienza, H. J., & Almeida, J. G. (2000). Effects of age at entry, knowledge intensity, and imitability on international growth. Academy of Management Journal, 43(5), 909–924. Bell, J., Crick, D., & Young, S. (2004). Small ﬁrm internationalization and business strategy. An exploratory study of knowledge intensive and traditional manufacturing ﬁrms in the UK. International Small Business Journal, 22(1), 23–56. Bird, B. J. (1989). Entrepreneurial behavior. Glenview, IL: Scott, Foresman & Company. Birley, S., & Westhead, P. (1990). Growth and performance contrasts between ‘‘types’’ of small ﬁrms. Strategic Management Journal, 11(7), 535–557. Cadogan, J. W., Cui, C. C., & Li, E. K. Y. (2003). Export market-oriented behavior and export performance. The moderating roles of competitive intensity and technological turbulence. International Marketing Review, 20(5), 493–513. Cadogan, J. W., Diamantopoulos, A., & de Mortanges, C. P. (1999). A measure of export market orientation: Scale development and cross-cultural validation. Journal of International Business Studies, 30(4), 689–707. Cadogan, J. W., Diamantopoulos, A., & Siguaw, J. A. (1998). Export market-oriented behaviour, their antecedents, consequences, and moderating factors: Evidence from the U.S. and the U.K. In: P. Andersson (Ed.), Marketing: Research and practice Vol. 2, Proceedings of Annual Conference of the European Marketing Academy (May 20–23, pp. 449–451), Sweden: Stockholm. Cadogan, J. W., Paul, N., Salminen, R. T., Puumalainen, K., & Sundqvist, S. (2001). Key antecedents to ‘export’ market-oriented behaviors: A cross-national empirical examination. International Journal of Research in Marketing, 18(3), 261–282. Chakravarthy, B. S. (1982). Adaptation: A promising metaphor for strategic management. Academy of Management Review, 7(1), 35–44. Child, J. (1972). Organization structure, environments and performance: The role of strategic choice. Sociology, 6(1), 1–22. Conant, J. S., Mokwa, M. P., & Varadarajan, P. R. (1990). Strategic types, distinctive marketing competencies and organizational performance: A multiple measures-based study. Strategic Management Journal, 11(5), 365–383. Covin, J. G., Green, K. M., & Slevin, D. P. (2006). Strategic process effects on the entrepreneurial orientation – sales growth rate relationship. Entrepreneurship Theory and Practice, 31(1), 57–81. Covin, J. G., & Slevin, D. P. (1991). A conceptual model of entrepreneurship as ﬁrm behavior. Entrepreneurship Theory and Practice, 16(1), 7–25. Covin, J. G., Slevin, D. P., & Covin, T. J. (1990). Content and performance of growth-seeking strategies: A comparison of small ﬁrms in high- and low-technology industries. Journal of Business Venturing, 5(6), 391–412. Davidsson, P. (1991). Continued entrepreneurship: Ability, need, and opportunity determinants of small ﬁrm growth. Journal of Business Venturing, 6(6), 405–429. Day, G. S., & Wensley, R. (1983). Marketing theory with a strategic orientation. Journal of Marketing, 47, 79–89.

Dimensions of IGO, Environmental Turbulence, and EO and MO

121

Day, G. S., & Wensley, R. (1988). Assessing advantage: A framework for diagnosing competitive superiority. Journal of Marketing, 52(2), 1–20. Delmar, F., Davidsson, P., & Gartner, W. B. (2003). Arriving at the high-growth ﬁrm. Journal of Business Venturing, 18(2), 189–216. Drucker, P. (1985). Innovation and entrepreneurship. New York: Harper and Row. Gatignon, H., & Xuereb, J.-M. (1997). Strategic orientation of the ﬁrm and new product performance. Journal of Marketing Research, 34, 77–90. Golann, B. (2006). Achieving growth and responsiveness: Process management and market orientation in small ﬁrms. Journal of Small Business Management, 44(3), 369–385. Grant, R. M. (1991). The resource-based theory of competitive advantage: Implications for strategy formulation. California Management Review, 33(3), 114–135. Greenley, G. E. (1995). Market orientation and company performance: Empirical evidence from UK companies. British Journal of Management, 6(1), 1–13. Greiner, L. E. (1998). Evolution and revolution as organizations grow. Harvard Business Review, 76(3), 55–67. Hay, M., & Kamshad, K. (1994). Small ﬁrm growth: Intentions, implementation and impediments. Business Strategy Review, 5(3), 49–68. Hunt, S. D., & Morgan, R. M. (1995). The comparative advantage theory of competition. Journal of Marketing, 59(2), 1–15. Jambulingam, T., Kathuria, R., & Doucette, W. R. (2005). Entrepreneurial orientation as a basis for classiﬁcation within a service industry: The case of retail pharmacy industry. Journal of Operations Management, 23(1), 23–42. Jaworski, B. J., & Kohli, A. K. (1993). Market orientation: Antecedent and consequences. Journal of Marketing, 57(3), 53–70. 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., & Wiedersheim-Paul, F. (1975). The internationalization of the ﬁrm: Four Swedish cases. Journal of Management Studies, 12(3), 305–322. Jones, M. V. (1999). The internationalization of small high-technology ﬁrms. Journal of International Marketing, 7(4), 15–41. Jones, M. V., & Coviello, N. E. (2005). Internationalisation: Conceptualising an entrepreneurial process of behaviour in time. Journal of International Business Studies, 36, 284–303. Katsikeas, C. S. (1994). Exporter competitive advantages: The relevance of ﬁrm characteristics. International Marketing Review, 11(3), 33–53. Knight, G., & Cavusgil, S. T. (1996). The born global ﬁrm: A challenge to traditional internationalization theory. Advances in International Marketing, 8, 11–26. Knight, G., & Cavusgil, S. T. (2004). Innovation, organizational capabilities, and the bornglobal ﬁrm. Journal of International Business Studies, 35(2), 124–141. Knight, G., & Cavusgil, S. T. (2005). A taxonomy of born-global ﬁrms. Management International Review, 45(3), 15–35. Kohli, A. K., & Jaworski, B. J. (1990). Market orientation: The construct, research propositions, and managerial implications. Journal of Marketing, 54(2), 1–18. Kreiser, P. M., Marino, L. D., & Weaver, K. M. (2002). Assessing the psychometric properties of the entrepreneurial orientation scale: A multi-country analysis. Entrepreneurship Theory and Practice, 26(Summer), 71–94.

122

SANNA SUNDQVIST AND OLLI KUIVALAINEN

Kumar, K., Subramanian, R., & Yauger, C. (1998). Examining the market orientationperformance relationship: A context-speciﬁc study. Journal of Management, 24(2), 201–233. Liao, J., & Welsch, H. (2003). Social capital and entrepreneurial growth aspiration: A comparison of technology- and non-technology-based nascent entrepreneurs. Journal of High Technology Management Research, 14(1), 149–170. Lumpkin, G. T., & Dess, G. G. (1996). Clarifying the entrepreneurial orientation construct and linking it to performance. Academy of Management Review, 21(1), 135–172. Lumpkin, G. T., & Dess, G. G. (2001). Linking two dimensions of entrepreneurial orientation to ﬁrm performance: The moderating role of environment and industry life cycle. Journal of Business Venturing, 16(5), 429–451. Madsen, T. K. (1987). Empirical export performance studies. Advances in International Marketing, 2(2), 178–198. Madsen, T. K. (1989). Successful export marketing management: Some empirical evidence. International Marketing Review, 6(4), 41–57. Miles, R. E., & Snow, C. C. (1978). Organizational strategy, structure and process. New York: McGraw-Hill. Miller, D., & Friesen, P. H. (1982). Innovation in conservative and entrepreneurial ﬁrms: Two models of strategic momentum. Strategic Management Journal, 3, 1–25. Moore, G. (1998). Gorilla game: Picking the winners in high technology. New York, NY: Harper Business. Narver, J. C., Jacobson, R. L., & Slater, S. F. (1999). Market orientation and business performance: An analysis of panel data. In: R. Deshpande (Ed.), Developing a market orientation (pp. 195–216). Thousand Oaks, CA: Sage Publications. Narver, J. C., & Slater, S. F. (1990). The effect of a market orientation on business proﬁtability. Journal of Marketing, 54(5), 20–35. Nummela, N., Puumalainen, K., & Saarenketo, S. (2005). International growth orientation of knowledge-intensive SMEs. Journal of International Entrepreneurship, 3(1), 5–18. Nunnally, J. C. (1978). Psychometric theory. New York: McGraw-Hill Book Company. Paswan, A. K. (2003). Channel support activities and perceived goal orientation: An exploration in the Indian market. Asia Paciﬁc Journal of Marketing and Logistics, 15(4), 19–41. Pelham, A. M. (1997). Mediating inﬂuences on the relationship between market orientation and proﬁtability in small industrial ﬁrms. Journal of Marketing Theory and Practice, 5(3), 55–76. Pelham, A. M. (2000). Market orientation and other potential inﬂuences on performance in small and medium-sized manufacturing ﬁrms. Journal of Small Business Management, 38(1), 48–67. Pelham, A. M., & Wilson, D. T. (1996). A longitudinal study of the impact of market structure, ﬁrm structure, strategy, and market orientation culture on dimensions of small-ﬁrm performance. Journal of the Academy of Marketing Science, 24(1), 27–43. Penrose, E. T. (1959). The theory of the growth of the ﬁrm. Oxford: Oxford University Press. Porter, M. E. (1996). What is strategy? Harvard Business Review, 74(6), 61–78. Schminke, M., Ambrose, M. L., & Cropanzano, R. S. (2000). The effect of organizational structure on perceptions of procedural fairness. Journal of Applied Psychology, 85(2), 294–304. Sinkovics, R., & Roath, A. S. (2004). Strategic orientation, capabilities, and performance in manufacturer – 3PL relationships. Journal of Business Logistics, 25(2), 43–64.

Dimensions of IGO, Environmental Turbulence, and EO and MO

123

Slater, S. F., & Narver, J. C. (1994a). Does competitive environment moderate the market orientation–performance relationship? Journal of Marketing, 58, 46–55. Slater, S. F., & Narver, J. C. (1994b). Market orientation, customer value, and superior performance. Business Horizon, 37(2), 22–28. Stewart, T. A. (2006). From the editor: Growth’s uncharted road. Harvard Business Review, 84(6), 12. Sto¨ttinger, B., & Schlegelmilch, B. B. (1998). Explaining export development through psychic distance: Enlightening or elusive? International Marketing Review, 15(5), 357–372. Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533. Thurik, R. (1993). Exports and small business in the Netherlands: Presence, potential, and performance. International Small Business Journal, 11(3), 49–59. Weinzimmer, L. G. (2000). A replication and extension of organizational growth determinants. Journal of Business Research, 48(1), 35–41. Weinzimmer, L. G., Nystrom, P. C., & Freeman, S. J. (1998). Measuring organizational growth: Issues, consequences and guidelines. Journal of Management, 24(2), 235–262. Welch, L. S., & Luostarinen, R. (1988). Internationalization: Evolution of a concept. Journal of General Management, 14(2), 36–64. Wiklund, J., Davidsson, P., & Delmar, F. (2003). What do they think and feel about growth? An expectancy-value approach to small business managers’ attitudes towards growth. Entrepreneurship Theory and Practice, 27(3), 247–270. Wiklund, J., & Shepherd, D. (2003). Aspiring for, and achieving growth: The moderating role of resources and opportunities. Journal of Management Studies, 40(8), 1919–1941. Yli-Renko, H., Autio, E., & Tontti, V. (2002). Social capital, knowledge, and the international growth of technology-based new ﬁrms. International Business Review, 11(3), 279–304. Zahra, S. A. (1993). A conceptual model of entrepreneurship as ﬁrm behaviour: A critique and extension. Entrepreneurship Theory and Practice, 17(4), 5–21.

INTERNATIONALISATION AND TECHNOLOGICAL INNOVATION: EMPIRICAL EVIDENCE ON THEIR MUTUAL RELATIONSHIP Diana A. Filipescu, Alex Rialp and Josep Rialp Internationalisation and innovation of the ﬁrms have been considered two of the most important factors determining business success over the last decade (Buckler & Zien, 1996; Wind & Mahajan, 1997; Zahra & George, 2002; Vila & Kuster, 2007). In fact, the scientiﬁc literature refers to ‘‘technoglobalism’’ for analysing the relationship between technological innovations and internationalisation of ﬁrms (Archibugi & Michie, 1995). Internationalisation is the process in which ﬁrms gradually increase their international involvement (Johanson & Vahlne, 1977) and also the process of adapting ﬁrms’ operations to international environments (Prashantham, 2005), this being an important issue for ﬁrms that often results in vital growth, useful learning outcomes and enhanced ﬁnancial performance. Furthermore, it is also described as an innovation of the ﬁrm that often entails decision-making under conditions of uncertainty, where knowledge is vital. Innovation, as it is deﬁned by Acs, Morck, and Yeung (2001), is the effort to create purposeful, focused change in a ﬁrm’s economic or social potential. According to Knight and Cavusgil (2004), the relevance of the internationalisation process in the last several years is likely powered by New Challenges to International Marketing Advances in International Marketing, Volume 20, 125–154 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020008

125

126

DIANA A. FILIPESCU ET AL.

two key trends that have reduced the transactions costs of foreign market expansion: globalisation of markets and technological advances. However, in the opinion of Jones and Coviello (2005), there are three questions that stand out concerning the internationalisation of the ﬁrm: the internationalmarket selection (where), the entry mode choice (how) and the pace of internationalisation (when). Recently, researchers have focused on the time aspect and some results indicate that time may not be the only explanation as to why ﬁrms start to internationalise (Rialp, Rialp, & Knight, 2005a). Today, many ﬁrms internationalise soon after their establishment, which has led to the emergence of the concept of born-global ﬁrms. International markets are also characterised by a greater competitive pressure than national markets (Lo´pez & Garcı´ a, 2005). Under these conditions, having competitive advantages allows ﬁrms to compete in a more active way in the markets, even more when they interact in different foreign markets (Eusebio & Rialp, 2002). In this context, both knowledge and experience obtained during ﬁrms’ international activities represent important competitive advantages, facilitating the internationalisation process (Johanson & Vahlne, 1977; Johanson & Wiedersheim-Paul, 1975). However, technology represents another important factor in increasing the national and international competitiveness of ﬁrms (Lachenmaier & Wo¨ssmann, 2006). Patterson, Grimm, and Corsi (2003) suggest as well that consideration of new technologies and a willingness to adopt them are the starting point and deﬁnitely contribute to ﬁrm success. Nevertheless, as stressed by Parasuraman (2000) and later by Richey, Daugherty, and Roath (2007), as being the ﬁrm’s technological readiness, this implies not only to possess technology but also to be ready to use it. Being an innovative ﬁrm allows it, on one hand, to obtain products with superior characteristics than the ones offered by the competitors (product innovations) and, on the other hand, to reduce costs of production and, consequently, prices (process innovation). Furthermore, innovation helps explain ﬁrm-level heterogeneity in export behaviour of ﬁrms (Basile, 2001). In this way, the innovative ﬁrms obtain some competitive advantages that give them the possibility to compete in an active way in different markets. As stated by Hoffman, Parejo, Bessant, and Perren (1998), the ﬁrms’ innovative capability is a key driver for sustainable competitive advantage in today’s rapidly changing markets. Some authors (Grossman & Helpman, 1991; Young, 1991) state that an important cause of innovative activities may be represented by exports themselves, as predicted by global economy models of endogenous innovation and growth. Other possible sources of information for technological

Internationalisation and Technological Innovation

127

innovation are internal sources, market sources, research sources and generally available sources of information (Amara & Landry, 2005). In summary, technology is one of the most important factors in increasing the national and international competitiveness of ﬁrms, thus technological innovations represent competitive advantages that give ﬁrms the possibility to compete in an active way in different markets. However, the international activity that ﬁrms develop can embody more innovativeness, considering that ﬁrms’ presence in new markets offers new perspectives and, therefore, new knowledge acquisition (Prashantham, 2005; Jean, Sinkovics, & Kim, 2008). Hence, we argue that there seem to be a mutual relationship between technological innovation and internationalisation, where one process is feeding the other one and vice versa. Although this connection seems obvious, we believe it has not been fully addressed by academics and this constitutes the main purpose of our research: to determine the existence of a reciprocal, interdependent relationship between technological innovation and internationalisation of the ﬁrm and, if it exists, to establish what it is like and which factors impact on it. More precisely, by means of the casestudy approach, we will analyse some export ﬁrms in terms of patterns of technological innovation and internationalisation. We believe this investigation is useful both to academics as well as to managers. The former can observe how the intuitive relationship between both processes is produced and the latter can consider the study as a guide in order to improve their international activities by innovating and developing their technological, innovative behaviour through international activities. To achieve the above-mentioned objective, this chapter is structured as follows. The next two sections provide a short overview of the research on ﬁrm internationalisation and innovation and provide theoretical frameworks that could be considered for explaining the relationship among both processes. Then, the research methodology is described, having as a base a systematic application of the case-study approach. Each case is individually described and a cross-analysis is presented. Finally, several conclusions and future lines of investigation are outlined.

PREVIOUS RESEARCH ON FIRM INTERNATIONALISATION AND INNOVATION Broadly speaking, internationalisation means the entry to new-country markets. It may, therefore, be described as a process of innovation (Bilkey & Tesar, 1977;

128

DIANA A. FILIPESCU ET AL.

Andersen, 1993; Casson, 2000). Faced with increasing international competition, innovation has become a central focus in ﬁrms’ long-term strategies. Firms competing in global markets face the challenges and opportunities of change in markets and technologies. One important aspect within innovation management is the optimal integration of external knowledge, since innovation increasingly is derived from a network of companies interacting in a variety of ways (Veugelers & Cassiman, 1999). According to Lo´pez and Garcı´ a (2005), technological resources can generate a two-fold competitive advantage for a ﬁrm. On one hand, they confer competitive advantages in costs, via the development of new and more efﬁcient productive processes. On the other hand, they generate competitive advantages based on differentiation, by means of product innovations, allowing ﬁrms to tailor products according to customer requirements, or develop products of a higher quality. Regarding this last point, Styles and Ambler (1994) point out that product strength – in terms of quality and uniqueness – is one of the key elements in export success. The role of innovation in export behaviour has already been analysed in the literature. Wakelin (1998), considering the UK case, ﬁnds that the number of innovations used at the sector level is positively and signiﬁcantly related to the probability of exporting. Moreover, it is observed that ﬁrms with a large number of innovations are more likely to export, indicating heterogeneity even within the group of innovating ﬁrms. Following the same research line, Basile (2001) analyses and compares the relation between innovation capabilities and export behaviour of Italian ﬁrms in different exchange-rate regimes. The results of his study show that innovation is a very important competitive factor and helps explain ﬁrm heterogeneity in export behaviour among Italian ﬁrms. The product innovation strategies have a positive effect on the export intensity only after currency, before Euro, devaluation. The author also demonstrated that the relation between innovation strategies and export behaviour of southern Italian ﬁrms is weaker than that found for the national average. However, Castellani and Zanfei (2007) highlight how intra-industry heterogeneity in Italy is associated with different internationalisation modes. They show that ﬁrms with high engagement in foreign activities also exhibit better economic and innovative performances. Firms with the highest international involvement, namely those with manufacturing activities abroad, are characterised by both higher productivity and higher R&D efforts and innovative performances. Lachenmaier and Wo¨ssmann (2006) have the possibility to directly test whether innovation causes exports, by considering a sample of German

Internationalisation and Technological Innovation

129

ﬁrms. The authors mention that a causal relationship between innovation and export is expected, focusing on the product-cycle features and the endogenous-growth models. However, their results can only show one part of the relationship, the one in which the fact stands out of being innovative causes ﬁrms to have substantially larger export shares than non-innovative ﬁrms in the same sector. So, considering this evidence, it is obvious that more extensive research is needed in order to accomplish the objective of analysing the causal relationship between these two factors. According to Vila and Kuster (2007), ﬁrms start to think about innovation because they want to offer different things in different markets. The authors focus their attention on the assessment of whether more innovative and less innovative ﬁrms correspond to high internationalised and less internationalised ﬁrms. With this purpose, different degrees of internationalisation and kinds of innovation are described, defending the idea that greater degrees of internationalisation lead to more innovation. Although only partial support for the argument establishing that internationalisation is associated with some kinds of innovation is provided, the empirical results appear to conﬁrm that the extent of a ﬁrm’s internationalisation inﬂuences innovation in terms of strategy and processes. In addition, Vila and Kuster (2007) state that many ﬁrms can be international or innovative, but only a small group is highly internationalised and has superior innovations, speciﬁcally those that invest abroad and use more resources to innovate in four dimensions (products, strategies, processes and markets). This statement is consistent with previous studies (Zou & Ozsomer, 1999) that also proposed that companies with high levels of innovation reﬂected a high degree of dependence on export markets and vice versa.1 Although the literature has presented different types of classiﬁcation for the innovation process (Rothwell, 1994; Henard & Szymanski, 2001; De Propris, 2002) we shall focus on the one stated by De Propris (2002), according to whom innovation is classiﬁed into four types: product, process, incremental and radical innovation. Product innovation corresponds to the introduction into the market of a new or improved product, whereas process innovation relates to the sequences and nature of the production process. Process innovation is often more difﬁcult to detect, but it is very important especially for buyer–supplier transactions. Freeman and Perez (1988) deﬁne radical innovations as discontinuous events, which are the result of a deliberate research and development activity. Incremental innovations refer to improvements due to use or experience; it can often take the form of smaller enhancements around major, radical innovations. Both radical and

130

DIANA A. FILIPESCU ET AL.

incremental innovations can be either in the product or process. This classiﬁcation of innovation coincides with the one that the Oslo Manual (OECD, 1997) states as being technological innovation and is deﬁned as an iterative process initiated by the generation of new products and processes or of signiﬁcant technological improvements in current products and processes. In summary, the concept of innovation is another cornerstone in modern international marketing strategy because in today’s competitive market environment, many senior managers have become increasingly concerned with the need to be ﬁrst, fast and on-time (Wong, 2002).

THEORETICAL FRAMEWORKS AND PRELIMINARY CONCEPTUAL MODEL Academics explaining ﬁrms’ internationalisation and innovation of ﬁrms have considered and used various theories. Among these, we can mention the gradual internationalisation theory of the ﬁrm (Andersen, 1993; Molero, 1998; Rialp, Rialp, Urbano, & Vaillant, 2005b), the resource-based view (Teece, Pisano, & Shuen, 1997; Knight & Cavusgil, 2004; Lo´pez & Garcı´ a, 2005; Galende, 2006; Hewitt-Dundas, 2006), transaction cost theory (Anderson & Gatignon, 1986; Galende, 2006) and the organisational capabilities dynamic perspective (Kogut & Zander, 1993; Day, 1994; Madhok, 1997; Knight & Cavusgil, 2004). Considering the purpose of this study, the gradual internationalisation theory of the ﬁrm, jointly with the resource-based view (referred to as RBV from now on), seem to be appropriate theoretical frameworks. The gradual internationalisation theory helps to explain when and how a ﬁrm begins to develop international activities, and RBV helps to explain how, in the context of an innovative culture, knowledge and the resultant organisational capabilities are developed and leveraged by enterprising ﬁrms. The gradual internationalisation theory of the ﬁrm, also known as the Uppsala Model (Johanson & Vahlne, 1977, 1990), develops the ﬁrms’ advantages, emphasising the knowledge of international markets, which can be considered as an advantage against competitors and the level of commitment with them from a perspective of greater personalisation of the managers of the ﬁrm. Hence, it posits that, as ﬁrms learn more about a certain market, they become more committed to it by investing more resources into that market. Johanson and Wiedersheim-Paul’s (1975) study

Internationalisation and Technological Innovation

131

identiﬁes the lack of knowledge and/or resources, and the resulting uncertainty to the ﬁrm, as the principle obstacle to internationalisation. With time, the ﬁrm gradually progresses through a series of learning and commitment stages, also deemed as the ‘‘establishment chain’’, in the following order: beginning to export to a country via an agent, establishing a subsidiary, beginning production in the host country. Therefore, ﬁrms improve their foreign market knowledge through initial expansion characterised by low risk and indirect exporting approaches targeting ‘‘psychologically close’’ markets. Johanson and Vahlne (1977, p. 24) deﬁne the psychological distance as ‘‘the sum of factors preventing the ﬂow of information to and from the market’’. Although this concept of psychic distance is being recently questioned due mostly to the new possibilities opened by the Internet and other similar on-line internationalisation mechanisms (Yamin & Sinkovics, 2006), many ﬁrms have gradually extended their foreign activities to markets of an increasingly greater psychological distance based upon the accumulation of experimental knowledge abroad over time. RBV has its origins in Penrose’s (1959) seminal work and suggests that the best way of understanding a ﬁrm is to consider it as a collection of productive resources, imperfectly imitable and speciﬁc to each ﬁrm, which allows it to compete successfully against other ﬁrms. The capacity of ﬁrms to generate sustainable competitive advantages depends on their particular set of resources. Barney (1991) mentions that resources that generate competitive advantages must fulﬁl four conditions: they must be valuable, scarce, inimitable and non-substitutable. Consequently, RBV helps to explain how, in the context of an innovative culture, knowledge and the resultant organisational capabilities are developed and leveraged by enterprising ﬁrms (Knight & Cavusgil, 2004). RBV also lends great importance to the ﬁrm’s technological capability. It points out that the innovative capability does not come from skill in exploiting external technologies, which are easily accessible for competitors and therefore insufﬁcient for sustaining a competitive advantage (Barney, 1991). It rather comes from the generation of internal innovation, which implies the possession of heterogeneous and speciﬁc technological resources, and the capability to generate other new resources and to build basic technological capabilities. Considering the above-mentioned frameworks, we propose a model (Fig. 1) which is based on the assumption that internationalisation and technological innovation exist in a mutual, interdependent relation. This model is in line with the one of Lo´pez and Garcı´ a (2005), who suggest

132

DIANA A. FILIPESCU ET AL.

Firm’s technology

Knowledge acquisition

Fig. 1.

Innovation

Competitive advantage

Internationalisation

Preliminary Model of the Relation between Internationalisation and Innovation. Source: Self-Elaborated.

a conceptual model which also considers technological resources and export behaviour with a focus on their relationship, and reinforcing the importance of intangible resources as sources of sustainable competitive advantages. In addition, Lo´pez and Garcı´ a (2005) state that ﬁrms possessing these kinds of resources will have a superior competitive potential and hence a superior capacity to gain access to international markets. This idea is reinforced by Itami (1987) and Eusebio and Rialp (2002), who consider technological resources and export behaviour as the key resources for business success. Regarding the knowledge acquired by ﬁrms during their international activity and its role in developing and/or improving their technological resources, the proposal of Prashantham (2005), according to whom there would seem to be a mutually reinforcing effect between knowledge and organisational capabilities, is considered. Explicitly, on one hand, organisational capabilities enable ﬁrms to create and leverage knowledge (Eisenhardt & Martin, 2000), and, on the other hand, the creation of new knowledge by ﬁrms leads to the development of organisational capabilities (Nelson & Winter, 1982). Moreover, knowledge is the most important resource, and the integration of individuals’ specialised knowledge is the essence of organisational capabilities (Nelson & Winter, 1982). We will divide knowledge into two: product and market knowledge (Grant, 1996; Li & Calantone, 1998). The latter may refer to industry knowledge, competitor knowledge and macro-environmental knowledge. Bearing in mind Amara and Landry’s (2005) classiﬁcation of information sources for technological innovation, we may also argue that knowledge acquirement represents an information source which has a certain degree of inﬂuence upon the technological innovation of ﬁrms. In conclusion, the proposed model is based on the assumption that internationalisation and technological innovation exist in an interdependent

Internationalisation and Technological Innovation

133

relation. This relation starts with the idea that the technology possessed by the ﬁrm helps to innovate in order to create the needed competitive advantages for competing and succeeding in an international market. Once the ﬁrm develops activities in international markets, it gains knowledge about the existing environment and competition, and this knowledge will be very helpful in maintaining competitive advantages and creating resource and market opportunities. Improving or creating competitive advantages implies more innovation. Consequently, the relation between innovation and internationalisation may be considered as a mutual one, and this is the core of our investigation. In reinforcing the idea of a mutual phenomenon, Edquist and McKelvey (2000) and Lundvall (1992) argue that the innovation process should rather be considered as a circular and complex system embracing interactive elements. In addition, Lewin and Massini (2003) contend that ﬁrms with superior innovation and knowledge-creation processes are more sophisticated, and have more highly developed and elaborated knowledge-creation routines and learning regimes.

METHODOLOGY The empirical part of this research is based upon a systematic application of the multiple-case study approach to an export context in which ﬁve Spanish exporters were ﬁrst judgementally chosen and then individually examined. Consistent with several qualitative methodologists (Eisenhardt, 1989; Yin, 1994; Maxwell, 2005), multiple-case-based investigations serve as a basis for either empirically testing previous theories or building new theoretical explanations of the researched phenomenon. The sample selection was made taking into account two important facts: ﬁrms should be widely recognised for their intense international activity and they should also be highly innovative. Table 1 shows which the selected ﬁrms were and some of their characteristics. Concerning the percentage of the export sales of the total sales, it can be observed that in only one case is this number extremely high (Comexi), representing almost the entire revenue of the ﬁrm. The other four companies indeed show a good level of export activity, this being between 15% and 30%. The main source of information acquisition in case studies was the semistructured interviews with general managers, export/commercial department managers and/or R&D managers of the selected ﬁrms. The contact was

134

DIANA A. FILIPESCU ET AL.

Table 1.

Description of the Case Studies.

Europastry

Year of foundation 1963 Ofﬁcially awarded for n.a. its international activity Industry products Baking industry: frozen dough

Indo

Pinturas Lobo/ Euroquı´ mica

Comexi

Tecnitoys/ Scalextric

1937 n.a.

1972 2003

1954 1998

1990 n.a.

Chemical Optical industry: industry: technical lenses, paintings eyewear, equipment, decorations 254,951,240 128,891,000 3,641,176

Operating revenue (h) Number of employees 2,150 Sales (h) 246,786,342 Foreign sales (%) 15

1,686 127,133,000 30

17 3,634,591 15

Converting Toy industry: sector: circuit, cars ﬂexo-press

57,551,280

22,830,759

241 56,233,127 90

44 21,439,520 25

Note: n.a., not applicable. Source: Self-elaborated.

carried out by means of a telephone call, the potential interviewees being informed about characteristics of the investigation and being asked for collaboration. Later on, detailed information about the investigation and the protocol of the interview was sent. The interviews took 45 min, on average. They were recorded with the consent of the interviewees, and afterwards full write-ups were constructed on each company in the form of a detailed case study, focusing on the speciﬁc characteristics of each case situation. As a requirement to achieve construct validity (Rialp et al., 2005b), a combined use of multiple secondary sources of information (such as information from the company website, internal documentation provided by the company, product and ﬁrm brochures, etc.) was made. Also, reliability requirements were assured by the use of the same protocol for each speciﬁc company and by the development of a complete database in the data collection phase. Once the transcription of the interviews was made, they were sent to the interviewees for the approval of the received information. All data sources, applied for each company, were used in order to edit one report with all of the information of the ﬁrm. That purpose was to obtain a clearer analysis and to allow the comparison of the different cases (Eisenhardt, 1989). Furthermore, the triangulation process guarantees

Internationalisation and Technological Innovation

135

internal validity of the investigation (Yin, 1994). According to Rialp (1998), the quality of the design of the study should be guaranteed by introducing a series of methods and tests of validity and reliability along different methodological phases.

ANALYSIS OF THE CASES The analysis of the case information is divided into two different parts: the individual analysis of the companies and the crossed analysis among the companies. The analysis of each company is carried out by studying the information that each company provided us following the same order of the items that were taken into consideration in the interview, therefore according to the protocol of the interviews. The cross-case analysis allows us to see similarities or differences among the results of all of the interviewed companies.

Individual Analysis Europastry Europastry is a family business which has developed its activity in the baking industry since the 1950s. Nowadays, the company aims to transform the baking industry, providing agile solutions to satisfy the requirements of professionals and customers through the use of new refrigeration technologies, being the Spanish market leader in frozen dough. Frozen dough offers an optimum, non-perishable product, achieving high quality at the best price with maximum simplicity of use. Europastry can be considered an innovative company due to the fact that it has introduced a variety of new and very competitive products, having its own R&D department. Developing internal R&D allows the ﬁrm to be more ﬂexible and to have greater barriers to imitation. Innovation within the ﬁrm is a product one, both incremental and radical. Regarding its internationalisation, the company started this process in 1998, creating an export department. The motivation to become international resides in the necessity of surviving in a hyper-competitive global world. The ﬁrst export countries were Germany, France and Portugal. Geographic as well as cultural factors were important for market selection. It has used different types of entry modes depending on the maturity grade of markets: agent (Germany), branch (France) and joint venture (Portugal).

136

DIANA A. FILIPESCU ET AL.

Considering the effect that an entry mode in a foreign market could have on the innovation process, the interviewee mentions that when buying an existent ﬁrm the investments in R&D are discouraged. In this case, the buyer needs to absorb the innovation of the acquired ﬁrm and integrate it within its own innovation system. As for organic growth toward international markets, this compels greater innovation in order to enter in speciﬁc markets where competition exists and, in this way, the innovation is longer and more effective. The ﬁrm possesses great knowledge of foreign markets and must be able to adapt its products immediately to every single market, since product adaptability is vital for success in these markets. However, being international has helped the company to innovate more and more, becoming very competitive. Thus, the specialisation of the ﬁrm is a very important factor in order to maintain its competitiveness. Indo Indo is an optical ﬁrm founded in 1937. In 1939, it began manufacturing its own products and implementing its policy of technological and scientiﬁc development. The ﬁrst industrial facilities for production of frames and sunglasses were built, and the manufacture of the ﬁrst fused bifocal lenses began. It also took the ﬁrst steps in the fusion of optical glass. Thus, we are dealing with an innovative ﬁrm from its inception. During the 1960s, the ﬁrm experienced a growth period in all areas of ophthalmic optics and technological development. Afterwards, the company began a process of decentralising its productive structure as the ﬁrst step toward becoming a multinational company. For each of the three business units (lenses, eyeglasses and equipment-goods), a speciﬁc exporting department operates which has the role to reach the markets where neither the subsidiaries nor the distributors do. The innovation within the ﬁrm is both incremental and radical in product and process. On one hand, the company owns two special materials of lenses and puts an emphasis on the design geometry of eyeglasses (product innovation). On the other hand, it differentiates itself from the others by having an anti-reﬂex treatment whose properties are different from others (process innovation). The main reason for the company to start exporting was the idea of maintaining its stability and growth. Foreign markets are selected depending on their economic potential, more than on geographical or cultural factors (Germany, France and Morocco were the ﬁrst countries where the ﬁrm exported). As the company has put a great emphasis on

Internationalisation and Technological Innovation

137

product differentiation it chooses entry modes which make it possible. For instance, it has several branches (Germany, France, USA, Chile and Morocco), foreign distributors in many countries and ﬁve factories (two in Spain, and one each in Thailand, China and Morocco). The ﬁrm has learnt from its international activity that the capacity to adapt its product to each market seems to be a valuable resource. The ﬁrm obtained a wider vision, a greater competitiveness, and a greater power of innovation, having a propensity for changes and being more agile, by becoming increasingly international. Taking into account the ﬁrm’s opinion, the key success factors are the continuous innovation that is realised, the proximity to the client, and also a great product experience since the ﬁrm has been working in the optical industry for a long time and has been a leader in the local market.

Pinturas Lobo/Euroquı´mica Pinturas Lobo is a chemical company that began its activity in 1972. It develops the chemical products according to the requirements of each customer, in order to respond to the need for aesthetics and for the most demanding performance (e.g. mechanical and chemical resistance, electrical resistance and conductivity, resistance to UV, etc.). It was purchased in 1998 by Euroquı´ mica, which has kept its ideology since the employees have continued to be the shareholders of the ﬁrm. As the products of the two ﬁrms were complementary (different types of industrial paints: paints for plastics, for glass, etc.), Euroquı´ mica decided to commercialise the whole range of products in an extensive way since new markets were opened. In terms of market share, although the company is a small one, it has always been a leader for two main reasons: its own R&D activities and self-ﬁnancing capacity (it has never depended either technologically or ﬁnancially on anyone else). An R&D department has existed within the company since its foundation, allowing for both incremental and radical product innovations. For instance, the company investigates in order to introduce a completely new product (radical innovation), but it also takes into account the clients’ opinions for improving the product (incremental innovation). The company began to export immediately after being purchased, that is, in 1998. The motivation for becoming international came with the idea of being competitive in the domestic market both economically and qualitatively. There were also some outside stimuli which showed their interest in the products. Moreover, being an innovative ﬁrm has helped in

138

DIANA A. FILIPESCU ET AL.

achieving new markets and become an international one. It mainly works with agents and foreign distributors. The ﬁrst country where the company exported to was Portugal, but only because the Portuguese contact spoke Spanish. So the company had no intention to become international until its purchase. After that moment, an international department has begun to operate being endowed with different specialised persons. As the product is a very speciﬁc one, it is very difﬁcult to ﬁnd a market which needs it or which does not already have a supplier. Therefore, the selection of markets is made through the creation of some concentric circles, but also considering geographical distance. Comexi Comexi is a family business with leader spirit in the converting sector and specialised in printing and converting ﬂexible packaging materials. Since its foundation in 1954, it has been characterised by the development of innovative and highly reliable technologies, adapted to client needs and market demands. Its expansion has never stopped since its foundation, becoming a leader at the international level and having only one serious competitor, from Germany. The company has always followed a policy of innovation, growth and leadership in the world market. Innovation has been achieved with creativity and ﬂexibility. Being a leader implies continuous innovation, with the R&D team of the company having a very important role in this effort. It is stressed: ‘‘We listen to everybody in the company; the future depends on the ability to get ahead of it; our team has been the developer of many of the latest milestones of recent years in the sector’’. Actually, the company has a person who is responsible for the innovation that coordinates all innovation processes within the ﬁrm. Recently, the ﬁrm has carried out innovation not only in the internal processes, but also in products. This type of innovation is an incremental one due to the fact that a process is improved, not a new technology. But it also realised radical innovation like in 1996, when it produced a total revolution in the whole area of engines by introducing another way of doing work into the process. Basically, innovation is developed depending on the level of knowledge that the ﬁrm has acquired during its international activities. Put differently, once being present in many markets with an aggressive strategy, it was essential to improve the product by adding more innovative components. Thinking of having a stable activity and regular sales, it began to export in 1960s and rapidly achieved a level of 50% of its production for international

Internationalisation and Technological Innovation

139

markets. The ﬁrst export countries were France and some countries of South America. As the founder had no skills for foreign languages, the selection of the markets was made taking into consideration the similarity of the languages and geographical proximity. Considering the entry modes, the company usually appeals to agents and branches (USA, Italy and Mexico). There is one case – Brazil – where it opened a subsidiary because of the high customs fees which made export there impossible).

Tecnitoys/Scalextric This company began its activity in the toy industry in 1990 and became international in 1992 once it purchased a well-known international brand from the same activity, Scalextric. Its main products are circuits, cars, tracks and accessories. The company has had an R&D department since the beginning of its activity. Two important characteristics of this department are that people who work there were hired in 1990s (they have great experience inside the ﬁrm) and are also very fond of the product. Every year, the ﬁrm introduces around 50 new and competitive products into the markets, so it is considered an innovative ﬁrm. Actually, it fulﬁls a quality rigor, meaning that not only are its products high-quality, but they are also environment friendly. Innovation within the ﬁrm is both in product and process, and both incremental and radical. The basis of these innovations stands out in the knowledge acquired, which can be product and market knowledge. For example, the digital version of the main product represents an incremental innovation since it is an improvement of the product. The remote control units that the ﬁrm has just introduced into the market represent a radical innovation. With the phenomenon of globalisation, exporting became important for the company in order to maintain stability and to grow. As it is a brand that can exist or not in other markets, the entry mode strategy is different: in the case that the brand already exists, the ﬁrm changes the name and competes directly through high quality; in the case that the brand does not exist but is it known, the ﬁrm competes through the brand; in the case that neither the brand exists nor it is known, the ﬁrm makes publicity for itself by means of public contests. Therefore, it has been able to select the entry modes in every foreign market, although the ﬁrm enters especially through agents and distributors.

140

DIANA A. FILIPESCU ET AL.

The selection of the international markets depends on both geographic and cultural distance, and the knowledge acquired has helped the ﬁrm to gain advantages in new markets. Since the company only offers a new and high-quality product in the markets where it enters, it is important to have a high innovation level. Therefore, innovation helps in achieving new markets. However, the ﬁrm has been able to internalise information which has helped it to keep innovating, thus the fact of being an international ﬁrm has helped it to continue innovating, bringing different and competitive products into the markets. In Table 2, information related to innovation and internationalisation of the ﬁve companies taken under consideration in this investigation is summarised.

Cross-Case Analysis and Discussion The ﬁve companies under analysis do not show a marked contrast regarding innovation and international phenomena. On the contrary, in spite of the fact that their activity is developed in completely different industries (baking, optical, toy industry, converting sector and chemical industry), they share almost the same opinion about the importance of innovation for becoming international and also the importance of being international for continuing the innovation process. Three of the ﬁve companies under consideration in this investigation were created around the middle of the past century and were leaders in the local market, two of them also being extremely innovative (Europastry and Indo); the other two ﬁrms are recent, one of them having the innovation process highly developed (Comexi), and the second beginning to innovate immediately after being present in other foreign markets. This last case is the one of Tecnitoys that, due to the fact that it became international almost immediately after its foundation by buying an existing international brand (Scalextric), it is difﬁcult to afﬁrm that the innovation was inﬂuenced by the internationalisation process of the ﬁrm. Nevertheless, it was emphasised by the interviewee that if the initial ﬁrm had been at least equally innovative, it surely would have helped more in beginning its international activities. As the international activities of the innovative ﬁrms from the chemical, optical and baking industries have begun after many years since their foundation, it was emphasised that being an innovative ﬁrm has helped in achieving other markets.

Motivation to begin exporting Export

Innovation

Export/innovation

Pinturas Lobo/ Euroquı´ mica

Comexi

Tecnitoys/Scalextric

It began to export It began to export in 1992, when it also in 1960 and afterwards began began to innovate to innovate It has an R&D It has had an R&D department for department since the every ﬁrm inside beginning of its the group and a activity. Innovation person responsible allowed the ﬁrm to for innovation. It select the entry has the possibility modes in different to select different markets. It realises entry modes. It product and process realises process, innovation, product, incremental and incremental and radical ones radical innovation Necessity to survive Necessity to survive Desire to be Necessity to have a Important for the in a global world in a global world competitive in the stable activity and stability of the ﬁrm domestic market regular sales Export department Three export Export department Export department Export department since its beginning. departments. The since its beginning. since its beginning. since its beginning. Both the geographic markets are selected Geographic distance Both cultural and Both geographic and and the cultural depending on their is important in order geographic cultural distances are factors have been economic potential. to select new proximity are taken into account. important. It uses Emphasis on product markets. Agents and important. The The modes of entry different modes of differentiation. The foreign distributors entry modes used used are through

Indo

Case-Studies Results.

It began to export in It began to export in It began to export 1998. It has always 1998. It has always around 1950. It has been innovative been an innovative always been innovative ﬁrm It has an R&D It has an R&D&I It has an R&D department for every department. department which product family. Innovation allowed allows it to achieve Innovation allowed the ﬁrm to select the new markets and the ﬁrm to select the entry modes in different entry entry modes in different markets. modes. The different markets. The innovation innovation realised The innovation within the ﬁrm is is a product one, realised is a product, both in product and both radical and incremental and in process, both incremental radical one incremental and radical

Europastry

Table 2.

Internationalisation and Technological Innovation 141

Commitment/type of knowledge

Knowledge/ adaptation

Indo

Pinturas Lobo/ Euroquı´ mica

Comexi

Tecnitoys/Scalextric

entry modes used agents and are the entry modes are agents, are: foreign used branches and own distributors distributors, subsidiary branches and factories The knowledge The ﬁrm has great The ﬁrm possesses Being for so many The knowledge acquired in foreign knowledge of the vast knowledge years on the acquired in other markets enables it to markets and it adapts about the markets. international markets has helped adapt its products its product The product cannot market, it has the ﬁrm to gain immediately to every be adapted too much gained experience advantages in new market because of its which helps in the ones speciﬁcity creation of new products Less and more Less and more Less and more Less and more Less and more commitment. commitment. commitment. commitment. commitment. Product knowledge Product knowledge Product knowledge Product Product knowledge and market/product and market/product and market/product knowledge and and market/product knowledge knowledge knowledge market/product knowledge knowledge

entry (agent, branch, joint venture)

Europastry

Table 2. (Continued )

142 DIANA A. FILIPESCU ET AL.

143

Internationalisation and Technological Innovation

All ﬁve companies have well-developed R&D departments which have appeared since the beginning of their innovation process. This is an important issue since different ﬁrm-level studies have used R&D expenditures as a proxy to innovation (Kumar & Siddharthan, 1994). Employees in these departments are well-prepared and have superior qualiﬁcations and experience in the ﬁrm; thus, the rate of personnel rotation is quite low. Being innovative has had a strong inﬂuence in the ﬁve ﬁrms’ international activities, especially in the selection of entry modes. Step-by-step, considering their international activities, the ﬁrms has become more innovative, and this fact determines the commitment that they are willing to have towards foreign markets. This fact is consistent with Eusebio and Rialp (2002), according to whom the realisation of innovations, both product and process ones, is positively related to export behaviour. As can be seen in Table 3, all of the ﬁrms innovate more in products than in processes, and the type is incremental more than radical. So it can be said that ﬁrms prefer to develop their product more step-by-step, this being in conformity with the aspects revealed by the interviewees who emphasised the customers’ reactions and opinions about every new product in order to improve it. Consequently, ﬁrms from optical and toy industries and also the one from the converting sector develop a complete innovation process, meaning both radical and incremental, focused on both product and process. According to Hewitt-Dundas (2006), the development of radical new products is dominated by large, and typically, multinational enterprises. Our results coincide with this afﬁrmation, since all three companies from this study

Table 3.

Types of Innovation within the Companies.

Company

Innovation Product and Radical Innovation

Product and Incremental Innovation

Process and Radical Innovation

Process and Incremental Innovation

X X X X X

X X X X X

n.a. X n.a. X X

n.a. X n.a. X X

Europastry Indo Pinturas Lobo Comexi Tecnitoys Note: n.a., not applicable. Source: Self-elaborated.

144

DIANA A. FILIPESCU ET AL.

which realise radical innovation are multinationals. In the other cases, the innovation process is a more speciﬁc one, with a preferred focus on product innovation. In four cases, the motivation to begin international activities is quite the same: in a global world it is compulsory to be global and have a stable activity in order to survive. Only Pinturas Lobo has begun to export, thinking about its competitiveness in the domestic market. Regarding the selection of markets, this decision is predominant due to cultural and geographic distance (four cases of ﬁve); as for the other case, Indo, the economic potential of the market is more important, although it sometimes considers geographic distance as well. The entry modes most used by all ﬁve companies are the ones by agents. Two of the ﬁve companies (Indo and Comexi) also use branches and own subsidiary in order to enter a market. Besides this, in the case of Comexi, the purchase of some other companies or the opening of a plant in countries where exporting is difﬁcult is also used. Tecnitoys also uses foreign distributors for its products, depending on the markets. As for Pinturas Lobo, it is common to select markets through the creation of some concentric circles. With regard to the relationship between the commitment that the ﬁrms have to the new markets and the knowledge they acquire, Europastry and Pinturas Lobo choose modes of entry which imply basically less commitment. In the other cases, it is observed that ﬁrms enter the new markets by modes with less commitment and, when time goes by and/or the markets become different, they choose to act with more commitment to the markets. In this way, they obtain both product knowledge and market knowledge, being able to develop radical product and process innovations. Taking into consideration the Uppsala model (Johanson & Vahlne, 1977, 1990), it can be argued that two of the ﬁve companies have not followed a traditional process of internationalisation, beginning to export in the years immediately following foundation. Comexi started its international operations in 1960, only 6 years after being founded, and Tecnitoys started to export only 2 years after its foundation. Both of them have sought to gain a signiﬁcant competitive advantage from its exporting activities, putting great emphasis on the quality of the product and, implicitly, on the innovation process. Based on previous research regarding the phenomenon of internationalisation, it could be afﬁrmed that these are born-global ﬁrms. Nevertheless, due to ﬁrms’ regional restriction at the beginning of their operations, we have decided to call them ﬁrms which started with an international orientation.

Internationalisation and Technological Innovation

145

Once they have internationalised, the ﬁve ﬁrms have obtained great knowledge – product and market knowledge – as a consequence of the mode of entry chosen which can imply less and/or more commitment to markets, and they also have gained experience with the competition; the knowledge acquired has been very important in order to adapt their products so they may become more competitive and obtain a greater market share. As mentioned by one of the interviewees, product adaptability is vital for success in these markets, and only by knowing the environment is adaptability possible, being in conformity with Oviatt and McDougall (2005), who consider knowledge to be the core of the internationalisation process of the ﬁrm. Keeping in mind the objective of this investigation – if there is a mutual, interdependent relationship between internationalisation and technological innovation processes of the ﬁrms – it can be answered that in all of the cases this relationship is observed and conﬁrmed, although one of the interviewees adds that innovation is not an indispensable condition for a ﬁrm to become international, but it is a very important factor. This assessment is coherent with Hall’s (1992) work, which emphasises the importance of innovation as a source of competitive advantage. However, after analysing the in-depth interviews, we are able to develop our initial model. We now argue that the mutual relationship supposedly existing between internationalisation and technological innovation processes are largely driven by the level of international commitment associated with different entry modes into the selected foreign markets. This conceptual development of our preliminary model is illustrated in Fig. 2. Therefore, according to the proposed model shown in Fig. 2, the potential alignment between the internationalisation and technological innovation processes can be more fully achieved when these two key sources of competitive advantage, instead of acting separately, actually expand and mutually reinforce each other: not only technological innovation, either in product or process, but also internationalisation itself, are on the basis of competitive advantage development. In addition, as demonstrated in some of the investigated ﬁrm cases, the higher the degree of internationalisation, in terms of both foreign market and entry mode selection processes, the more likely radical, and not only incremental, innovation can be. However, the investigated ﬁrms mostly differ in terms of the moment in time in which internationalisation and technological innovation jointly become two powerful and highly combinable sources of competitive advantage. This alignment or relationship can exist sooner or later.

146

DIANA A. FILIPESCU ET AL.

Competitive advantage

Technological Innovation

Technological Innovation

Internationalisation Radical Innovation

Incremental Innovation Entry mode selection

Less commitment/ risk (agents, foreign distributors)

More commitment/ risk (local branches, jointventure, sales/ production subsidiary).

Foreign knowledge acquisition (product/ market) Low

Fig. 2.

High

Revised Conceptual Relation between Innovation and Internationalisation. Source: Self-Elaborated.

In three out of the ﬁve ﬁrm cases analysed, those following a more slow and gradual pattern of internationalisation (the ﬁrms competing in the baking, optical and chemical industries, respectively) technological innovation preceded internationalisation and only after both were combined. However, for the other two ﬁrms which began with an international orientation from inception (the company competing in the toy industry and the one acting in the converting sector), this internationalisation and technological innovation alignment started nearly at the same time with the founding of these business. Accordingly, these two companies, Comexi and Tecnitoys, are international ﬁrms which have sought to exploit their technological competitive advantages and have been present at many different markets abroad almost from inception. This type of company, usually referred to as a born-global ﬁrm in the literature, (see Rialp et al., 2005a, for an extensive literature review) is mostly characterised by extending itself very soon and

Internationalisation and Technological Innovation

147

rapidly to the most signiﬁcant markets worldwide, also generating signiﬁcant value in this process and coordinating complex ﬂows of knowledge, capital and products among its deeply interrelated units (Pla & Leo´n, 2006). According to Acs et al. (2001), ﬁrms become multinationals because they see and capture proﬁtable international opportunities. For a multinational to compete abroad, it needs an advantage of its own to offset local ﬁrms’ home-court advantage. Actually, a very important issue also highlighted by Acs et al. (2001) is that multinational ﬁrms can quickly and simultaneously introduce an innovation in many countries, greatly magnifying the innovation’s return, being able to do this without exposing or losing control over their intellectual property.

CONCLUSIONS The purpose of this investigation was to analyse the existence of a relationship between technological innovation and internationalisation processes of the ﬁrm, having as a theoretical background the internationalisation theory of the ﬁrm (Johanson & Vahlne, 1977, 1990), the RBV of the ﬁrm (Barney, 1991) and the literature review related to these phenomena. From RBV’s perspective, generating and sustaining competitive advantages resides in the set of strategic resources and capabilities available to the ﬁrm, among these strategic resources being the intangible ones. Among intangible resources, technological resources are particularly signiﬁcant (Lo´pez & Garcı´ a, 2005). These provide the ﬁrm with an innovative capacity, for both products and processes, and are important for the creation of competitive advantages based on differentiation. These advantages give a ﬁrm a superior competitiveness to act in the international markets. This is consistent with the idea of Mahoney (1995), who stresses that differentiation is a fundamental strategy in order to achieve competitive advantages. Concerning the technological resource of the ﬁrms in this study, all ﬁrms showed a propensity for internationalisation due to the technological innovation developed within them. However, the fact of developing international activities has inﬂuenced, in a positive way, technological innovation. Actually, Basile (2001) emphasises the role of technology and innovation as one of the main factors contributing to facilitate entry into international markets, at the same time as boosting the ﬁrm’s export performance. As mentioned by Molero (1998), in the cases of Italy and Spain, the presence abroad of non-innovative ﬁrms suggests the existence

148

DIANA A. FILIPESCU ET AL.

of a relative divorce between the processes of innovation and internationalisation. Therefore, the results of our study are in line with the ones of Basile (2001) and Molero (1998), although this idea is not supported by Wakelin (1998), who ﬁnds that being an innovative ﬁrm in the UK has a negative impact on the probability of exporting, concluding that innovative ﬁrms are more inclined to use their innovation to exploit the domestic (UK) market rather than to enter foreign markets. It is also observed that the analysed ﬁrms use different entry modes, beginning with an agent, subsidiary and ending with owning a subsidiary. In this last case, it is emphasised that it discourages investments in R&D since the ﬁrm buys an existing company together with its innovation and/or its knowledge. We veriﬁed that when the ﬁrm chooses an entry mode which implies less commitment to the market, it is more common that an incremental and product innovation will be realised. This is the case of all of the ﬁve companies here analysed. However, when more commitment is chosen, then the radical innovations in products and also processes are more probable, and this happens to three of the ﬁve companies. In any case, it can be noticed that the predominant entry mode chosen by the ﬁrm is through agents and foreign distributors, representing the ﬁrst step of the internationalisation process of the ﬁrm, as explained by Johanson and Vahlne (1977, 1990). Actually, three of the ﬁve ﬁrms have followed a traditional process of internationalisation, the remaining two beginning to export very soon after their foundation. However, it would be convenient to study this phenomenon more in-depth due to the fact that it is not easy to establish a sequential order of the entry modes. As mentioned in the cross-case analysis, the innovation that all ﬁve ﬁrms subsequently realise is the incremental product one by adapting their products for the international markets and expanding these markets due to the product knowledge they gain. As Ferna´ndez (2005) describes, this kind of innovation is based on the improvement of the current components of the product, maintaining the same structure of relationships among the components. The incremental innovation is not usually visible, since the external aspect of the product and the functions that it carries out are the same ones. However, it can have a fundamental incidence in production costs, providing the company an important competitive advantage in costs. Actually, the need to acquire foreign product and market knowledge and the importance of organisational learning for entering or expanding in the international marketplace were recognised by several scholars, such as Andersen (1993) and Zahra, Ireland, and Hitt (2000), among others. The motivation in most cases to extend their markets has appeared owing to

Internationalisation and Technological Innovation

149

the necessity of surviving in a global market and of having a more stable competitive position.

FUTURE RESEARCH LINES AND LIMITATIONS OF THE STUDY The main idea highlighted in the results of this investigation is that once the ﬁrms have entered the foreign markets through different entry modes, depending on the commitment level to markets, they have gained experience but, most of all, they have acquired knowledge – product and/or market knowledge – and with that knowledge the ﬁrms are able to realise more technological innovation. Furthermore, depending on the entry mode, the companies acquire a certain type of knowledge (product/market) which leads to a certain type of technological innovation (incremental/radical). For instance, if one company chooses less commitment to a market, it will gain mostly product knowledge, which will imply more incremental product/process innovation. On the contrary, if a company chooses more commitment to a market, it will gain market knowledge, which will imply more radical product/process innovation. Thus, besides the traditional sources of technological innovation mentioned at the beginning (Amara & Landry, 2005), there are also new sources among which the commitment of entry mode in a new foreign market is considered. In other words, the internationalisation of the ﬁrm, more precisely the commitment of the entry modes in new markets, is considered to be an important source for promoting technological innovation. Surely, it is at least equally important to focus not only on the commitment of resources associated with the different entry modes but also on the market driven factors such as geographic and cultural proximity, country/market size, competitors and other socio-economic factors. Hopefully, further studies which also approach this issue through quantitative and not only qualitative research methods will be highly relevant to verify the empirical soundness of the suggested model. Moreover, following Parasuraman (2000) and Richey et al. (2007), it is also essential to implement the technological capabilities developed by a ﬁrm, since only forging them does not guarantee success. This argument is also known as ‘‘ﬁrm technological readiness’’, which is deﬁned as the ability of the ﬁrm to embrace and make use and effective of new technological assets (Richey et al., 2007). This, in our opinion, underdeveloped area focused on

150

DIANA A. FILIPESCU ET AL.

technology implementation issues, and their relationship with the internationalisation process of the ﬁrm, constitutes another stream of research in which much more effort should be deﬁnitely devoted in the near future. Referring to other limitations of this investigation, those fundamental characteristics of the methodology of the case-study stand out. This methodology produces a bias introduced by the same investigator in the collection process and analysis of the information (Rialp, 1998). The criticism regarding the lack of both statistical validity and representativeness is also assumed, but it is considered that the objective of the investigation is not the one to generalise but rather to study the knowledge of the theme of the study in depth, and, therefore, the methodology used is believed to be the right one.

NOTE 1. Other studies focused on Spain are: Buesa and Molero (1993), where different patterns of innovation and technological strategy are presented; Molero (1998), who presents patterns of internationalisation of innovative ﬁrms; Eusebio and Rialp (2002), who relate technological innovation and exporter result and ﬁnally Lo´pez and Garcı´ a (2005), who analyse technological export behaviour.

ACKNOWLEDGMENTS The authors would like to thank the reviewers of earlier versions of this chapter for their very helpful comments and suggestions as well as the General Manager of Europastry, R&D&I Manager of Indo, General Manager and International Division Manager of Pinturas Lobo, Commercial Manager and Corporate Marketing Responsible of Comexi and General Manager of Tecnitoys for their support in developing this research. Also the support received from the Universities and Research Comissioner from the Innovation, Universities and Firm Department of the Catalan Government and European Social Funds is especially acknowledged.

REFERENCES Acs, Z. J., Morck, R. K., & Yeung, B. (2001). Entrepreneurship, globalisation, and public policy. Journal of International Management, 7(3), 235–251. Amara, N., & Landry, R. (2005). Sources of Information as determinants of novelty of innovation in manufacturing ﬁrms: Evidence from the 1999 statistics Canada innovation survey. Technovation, 25(3), 245–259.

Internationalisation and Technological Innovation

151

Andersen, O. (1993). On the internationalisation process of ﬁrms: A critical analysis. Journal of International Business Studies, 24(2), 209–231. Anderson, E., & Gatignon, H. (1986). Modes of foreign entry: A transaction cost analysis and prepositions. Journal of International Business Studies, 17(3), 1–26. Archibugi, D., & Michie, J. (1995). Technology and innovation: An introduction. Cambridge Journal of Economics, 19(1), 1–4. Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120. Basile, R. (2001). Export behaviour of italian manufacturing ﬁrms over the nineties: The role of innovation. Research Policy, 30(8), 1185–1201. Bilkey, W. J., & Tesar, G. (1977). The export behaviour of smaller sized wisconsin manufacturing ﬁrms. Journal of International Business Studies, 8(1), 93–98. Buckler, S. A., & Zien, K. A. (1996). The spirituality of innovation: Learning from stories. Journal of Product Innovation Management, 13(5), 391–405. Buesa, M., & Molero, J. (1993). Patrones de Innovacio´n y Estrategias Tecnolo´gicas en las Empresas Espan˜olas. In: J. L. Garcı´ a Delgado (Ed.), Espan˜a, economı´a. Madrid: Edicio´n aumentada y actualizada, Espasa Calpe. Casson, M. (2000). Economics of international business. Cheltenham: Edward Elgar. Castellani, D., & Zanfei, A. (2007). Internationalisation, innovation and productivity: How do ﬁrms differ in Italy? World Economy, 30(1), 156–176. Day, G. S. (1994). The capabilities of market-driven organizations. Journal of Marketing, 58(4), 37–52. De Propris, L. (2002). Types of innovation and inter-ﬁrm co-operation. Entrepreneurship & Regional Development, 14(4), 337–353. Edquist, C., & McKelvey, M. (2000). Systems of innovation: Growth, competitiveness and employment. Chletenham: Edward Elgar. Eisenhardt, K. (1989). Building theories from case study research. Academy of Management Review, 14(4), 532–550. Eisenhardt, K., & Martin, J. A. (2000). Dynamic capabilities: What are they? Strategic Management Journal, 21(10/11), 1105–1121. Eusebio, R., & Rialp, A. (2002). Innovacio´n Tecnolo´gica y Resultado Exportador: Un Ana´lisis Empı´rico Aplicado al Sector Textil-Confeccio´n Espan˜ol. Document de treball, n 4. Spain: Autonomous University of Barcelona. Ferna´ndez-Sa´nchez, E. (2005). Estrategia de innovacio´n. Spain: Thomson. Freeman, C., & Perez, C. (1988). Structural crises of adjustment: Business cycles and investment behaviour. In: G. Dosi, C. Freeman, R. Nelson & L. Soete (Eds), Technical change and economic theory. England: London Pinter. Galende, J. (2006). Analysis of technological innovation from business economics and management. Technovation, 26(3), 300–311. Grant, R. M. (1996). Toward a knowledge-based theory of the ﬁrm. Strategic Management Journal, 17(Winter, Special Issue), 109–122. Grossman, G. M., & Helpman, E. (1991). Endogenous product cycles. Economic Journal, 101(408), 1214–1229. Hall, R. (1992). The strategic analysis of intangible resources. Strategic Management Journal, 13(2), 135–144. Henard, D., & Szymanski, D. (2001). Why some new products are more successful than others. Journal of Marketing Research, 38(3), 362–375.

152

DIANA A. FILIPESCU ET AL.

Hewitt-Dundas, N. (2006). Resource and capability constraints to innovation in small and large plants. Small Business Economics, 26(3), 257–277. Hoffman, K., Parejo, M., Bessant, J., & Perren, L. (1998). Small ﬁrms, R&D technology and innovation in the UK: A literature review. Technovation, 18(1), 39–55. Itami, H. (1987). Mobilizing invisible assets. Cambridge, MA: Harvard University Press. Jean, R.-J., Sinkovics, R. R., & Kim, D. (2008). Information technology and organizational performance within international business to business relationships – a review and an integrated conceptual framework. International Marketing Review, 25(5), 563–583. Johanson, J., & Vahlne, J. (1977). The internationalisation process of the ﬁrm: A model of knowledge development and increasing foreign commitments. Journal of International Business Studies, 8(1), 23–32. Johanson, J., & Vahlne, J. (1990). The mechanism of internationalisation. International Marketing Review, 7(4), 11–24. Johanson, J., & Wiedersheim-Paul, F. (1975). The internationalisation process of the ﬁrm: Four Swedish case studies. Journal of Management Studies, 12(3), 305–322. Jones, M. V., & Coviello, N. E. (2005). Internationalisation: Conceptualising an entrepreneurial process of behaviour in time. Journal of International Business Studies, 36(3), 284–303. Knight, G. A., & Cavusgil, S. T. (2004). Innovation, organisational capabilities, and the bornglobal ﬁrm. Journal of International Business Studies, 35(2), 124–141. Kogut, B., & Zander, U. (1993). Knowledge of the ﬁrm and the evolutionary theory of the multinational corporation. Journal of International Business Studies, 24(4), 625–645. Kumar, N., & Siddharthan, N. S. (1994). Technology, ﬁrm size and export behaviour in developing countries: The case of Indian enterprise. Journal of Development Studies, 31(2), 288–309. Lachenmaier, S., & Wo¨ssmann, L. (2006). Does innovation cause export? Evidence from exogenous innovation impulses and obstacles using German micro data. Oxford Economic Press, 58(2), 317–350. Lewin, A. Y., & Massini, S. (2003). Knowledge creation and organizational capabilities of innovating and imitating ﬁrms. In: H. Tsoukas & N. Mylonopoulos (Eds), Organizations as knowledge systems. Basingstoke: Palgrave. Li, T., & Calantone, R. J. (1998). The impact of market knowledge competence on new product advantage: Conceptualization and empirical examination. Journal of Marketing, 62(October), 13–29. Lo´pez, R. J., & Garcı´ a, R. M. (2005). Technology and export behaviour: A resource-based view approach. International Business Review, 14(5), 539–557. Lundvall, B. A. (1992). National systems of innovation: Toward a theory of innovation and interactive learning. England: London Pinter. Madhok, A. (1997). Cost, value and foreign market entry mode: The transactions and the ﬁrm. Strategic Management Journal, 18(1), 39–61. Mahoney, J. T. (1995). The management of resources and the resource of management. Journal of Business Research, 60(4), 19–38. Maxwell, J. A. (2005). Qualitative research design: An interactive approach (2nd edn). Thousand Oaks, CA: Sage Publications. Molero, J. (1998). Patterns of internationalisation of Spanish innovatory ﬁrms. Research Policy, 27(5), 541–558. Nelson, R., & Winter, S. (1982). An evolutionary theory of economic change. Cambridge, MA: Belknap Press.

Internationalisation and Technological Innovation

153

OECD. (1997). Proposed guidelines for collecting and interpreting technological innovation data. Oslo Manual, Paris. Oviatt, B. M., & McDougall, P. P. (2005). The internationalisation of entrepreneurship. Journal of International Business Studies, 36(1), 2–8. Parasuraman, A. (2000). Technology readiness index (TRI), a multiple-item scale to measure readiness to embrace new technologies. Journal of Services Research, 2(4), 307–320. Patterson, K. A., Grimm, C. M., & Corsi, T. M. (2003). Adopting new technologies for supply chain management. Transportation Research: Part E, 39(2), 95–121. Penrose, E. T. (1959). The theory of the growth of the ﬁrm. Oxford: Blackwell. Pla Barber, J., & Leo´n Darder, F. (2006). Direccio´n de Empresas Internacionales. Spain: Prentice Hall, Ed. Pearson. Prashantham, S. (2005). Toward a knowledge-based conceptualisation of internationalisation. Journal of International Entrepreneurship, 3(1), 37–52. Rialp, A. (1998). El Me´todo del Caso como Te´cnica de Investigacio´n y Su Aplicacio´n al Estudio de la Funcio´n Directiva. IV Taller de Metodologı´a ACEDE, La Rioja. Rialp, A., Rialp, J., & Knight, G. (2005a). The phenomenon of early internationalizing ﬁrms: What do we know after a decade (1993–2003) of scientiﬁc inquiry? International Business Review, 14(2), 147–166. Rialp, A., Rialp, J., Urbano, D., & Vaillant, Y. (2005b). The born-global phenomenon: A comparative case study research. Journal of International Entrepreneurship, 3(2), 133–171. Richey, R. G., Daugherty, P. J., & Roath, A. S. (2007). Firm technological readiness and complementarity: Capabilities impacting logistics service competency and performance. Journal of Business Logistics, 28(1), 195–228. Rothwell, R. (1994). Towards ﬁfth generation process innovation. International Marketing Review, 11(1), 7–31. Styles, C., & Ambler, T. (1994). Successful export practice: The UK experience. International Marketing Review, 11(6), 23–47. Teece, D. J., Pisano, G., & Shuen, R. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533. Veugelers, R., & Cassiman, B. (1999). Make and buy in innovation strategies: Evidence from Belgian manufacturing ﬁrms. Research Policy, 28(1), 63–80. Vila, N., & Kuster, I. (2007). The importance of innovation in international textile ﬁrms. European Journal of Marketing, 41(1/2), 17–36. Wakelin, K. (1998). Innovation and export behaviour at the ﬁrm level. Research Policy, 26(7-8), 829–841. Wind, J., & Mahajan, V. (1997). Issues and opportunities in new product development: An introduction to the special issue. Journal of Marketing Research, 34(1), 1–12. Wong, V. (2002). Antecedents of international new product rollout timeliness. International Marketing Review, 19(2), 120–132. Yamin, M., & Sinkovics, R. R. (2006). Online Internationalisation, psychic distance reduction and the virtuality trap. International Business Review, 15(4), 339–360. Yin, R. (1994). Case study research: Design and methods (3rd edn). Thousand Oaks, CA: Sage Publications. Young, A. (1991). Learning by doing and the dynamic effects of international trade. Quarterly Journal of Economics, 106(2), 369–405.

154

DIANA A. FILIPESCU ET AL.

Zahra, S., Ireland, R., & Hitt, M. (2000). International expansion by new venture ﬁrms: International diversity, mode of market entry, technological learning, and performance. Academy of Management Journal, 43(5), 925–950. Zahra, S. A., & George, G. (2002). International entrepreneurship: The current status of the ﬁeld and future research agenda. In: M. Hitt, R. Ireland, M. Camp & D. Sexton (Eds), Strategic leadership: Creating a new mindset (pp. 255–288). London: Blackwell. Zou, S., & Ozsomer, A. (1999). Global product R&D and the ﬁrm’s strategic position. Journal of International Marketing, 7(1), 57–76.

THE ROLE OF INNOVATIVE AND ENTREPRENEURIAL BEHAVIOR IN INTERNATIONALIZATION PROCESSES Arild Aspelund and Tage Koed Madsen During the past decade, the research streams focusing on entrepreneurship and internationalization have to some extent converged. This is a reﬂection of what has happened in business life. During the 1970s and 1980s, numerous empirical studies supported the notion that ﬁrms initially gain a strong foothold in their domestic market before they leap into international business. Accordingly, entrepreneurship occurred in domestic market settings. This is no longer the case. Newly established ﬁrms increasingly engage in international marketing right from their inception, a fact that has been reﬂected in the literature on entrepreneurship as well as on internationalization processes (for early contributions see Oviatt & McDougall, 1994; Knight & Cavusgil, 1996, for more extensive evidence see Rialp, Rialp, & Knight, 2005 and Zahra, 2005). This development has given rise to a lot of studies that have looked on various aspects of innovation and entrepreneurship in relation to international marketing. Especially, two lines of research have contributed with many of these studies, namely the early Innovation-Related Internationalization model (IM) and the more recent International Entrepreneurship model (IE). The objective of this study is to examine how ‘innovative behavior’ has been New Challenges to International Marketing Advances in International Marketing, Volume 20, 155–175 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020009

155

156

ARILD ASPELUND AND TAGE KOED MADSEN

treated in the IM model and how ‘entrepreneurial behavior’ has been treated in the IE model in order to gain new knowledge on the impact on both strands of research and to point out their relevance to new challenges in international marketing. Such examination is timely as the literature streams have existed virtually separated and the extensive body of research from the IM tradition has been largely ignored in the more contemporary research in IE. We believe that innovative behavior and entrepreneurship are valuable in order to solve the new challenges of international marketing. Environmental dynamics, new information and communication technologies as well as more open, global market conditions are drivers of challenges for international marketing mangers who have to deal with dynamic changes with regard to demand as well as processes and structure of competition and collaboration. Continuous change requires new solutions and other procedures and routines when designing marketing channels, communication with customers, servicing products, etc., i.e. innovative marketing behavior. The role of innovative behavior has been emphasized since the birth of behavioral internationalization research in the 1960s and the 1970s. This early literature on export development processes embodies theoretical models that attempt to explain how ﬁrms develop from domestically oriented to full-ﬂedged internationally oriented organizations. One of the most prominent of the early models, The North American IM, conceptualizes internationalization processes as proceeding in gradual and sequential stages, with an explicit focus on management characteristics and their impact on the process and its outcome (e.g. Bilkey, 1978; Cavusgil, 1980). The IM researchers supplement the explicit focus on managers with the notion that internationalization should be regarded as an innovation and, hence, the internationalization process as a process of adopting an organizational innovation. An important part of this approach is therefore theories of how innovations are adopted by organizations. The IE model reﬂects the emergence of a new type of ﬁrms that target international markets right from their inception. Such ‘instant internationalization’ is recognized to represent a serious challenge to the existing internationalization models because the latter assume that the ﬁrm matures on the domestic market before starting international expansion (Oviatt & McDougall, 1994; Knight & Cavusgil, 1996). This evident inconsistency has triggered the interest of a broad range of researchers and consequently caused an increase in the academic literature on the phenomenon (Rialp et al., 2005; Zahra, 2005; Aspelund, Madsen, & Moen, 2007); consequently, international entrepreneurship has been established as a recognized research strand of its own with entrepreneurial behavior and internationalization of new ﬁrms ﬁrmly placed in the core (Fletcher, 2004).

Role of Innovative and Entrepreneurial Behavior

157

Here, we will present the IM and the IE models and show how they have used the concepts of innovative and entrepreneurial behavior. Second, we will discuss whether future research on internationalization processes of ﬁrms may beneﬁt by drawing upon both strands of research to solve the new challenges of international marketing. Finally, we will discuss whether practitioners can improve their international marketing decisions by doing the same.

METHODOLOGY The study is based on reviews of the relevant parts of the IM and IE literatures. We only selected articles for review that explicitly deal with innovative or entrepreneurial behavior. The IM literature dates back from the mid-1970s. In order to identify relevant articles about this perspective and its theoretical roots, we searched for articles written by some of the founders and prominent authors of the perspective (Waren J. Bilkey, George Tesar, and S. Tamer Cavusgil). We searched for articles in which innovative behavior was mentioned explicitly and traced the theoretical roots through the references mentioned. The articles identiﬁed dated back to the early 1960s with regard to the theoretical roots, and we did not identify relevant articles after 1990 which is probably because the European perspective on internationalization processes of ﬁrms (see Andersen, 1993) became dominant. However, the rapidly declining interest in the IM perspective in the 1990s might also be related to the apparent incompatibility with the emerging IE perspective. With regard to the IE literature, we have used an identical list of studies as a recent literature review on International New Ventures (INVs) by Aspelund et al. (2007). This study reviewed a set of 21 core journals within the ﬁelds of management, marketing, and entrepreneurship yielding a total sample of 41 articles in the phenomenon of rapid internationalization of new ﬁrms published between 1995 and 2005. For this study, we have selected the studies that are related to innovative and entrepreneurial behavior and included them in our presentation of the IE model below.

THE ‘INNOVATION-RELATED INTERNATIONALIZATION MODEL’ The literature on internationalization processes of ﬁrms grew out of an attempt to describe empirically how individual ﬁrms become international.

158

ARILD ASPELUND AND TAGE KOED MADSEN

In Europe (starting in Sweden, see for example Johanson & Vahlne, 1977), basic ideas about experiential learning in neighboring countries with low so-called psychic distance were seen as the background for slow and incremental processes involving increasing international knowledge and commitments, a process during which the ﬁrm would attempt to avoid too much risk-taking. Researchers in North America, especially at the University of Wisconsin, conceived similar ideas. As Andersen (1993) notes, however, the latter group of researchers considered international marketing to be an innovation for the ﬁrm.

Theoretical Foundations The early models established in North America built upon the idea that an internationalization process would unfold itself in stages just like the adoption of an innovation. Different models involving alternative numbers and deﬁnitions of such stages were developed (see for example Bilkey & Tesar, 1977; Cavusgil, 1980). In one of his most important articles, Warren J. Bilkey (1978), one of the initiating researchers of American internationalization studies, refers to Rogers (1962) and Schumpeter (1961) as important theoretical roots. It is also clearly demonstrated in his article that the American approach was very much inspired by several empirical studies carried out in the UK during the 1960s, mainly by Simmonds and Smith (1968). In an early contribution, Simmonds and Smith (1968) suggested to analyze the ﬁrst export order as a marketing innovation. They refer to other British studies (e.g. Tookey, 1964), which demonstrate that active, powerful, and aggressive top managers were often drivers of the ﬁrst export activities. Apparently, Simmonds and Smith regarded these individuals as a kind of early adopters of innovative export marketing activities. They argue that opening up a new export market is just as innovative as the adoption of a new production process, for which reason they refer to Rogers (1962) when trying to identify a theoretical framework to understand the adoption of such new practices. Inspired by these ideas, the American scholars modeled internationalization processes as proceeding in stages just like the innovation-decision process developed by Rogers (1962). The IM researchers never explicitly explain how they developed the stages model. On the basis of their theoretical foundation, however, the export activity must be considered to be the innovation which the individuals and the organization initially have to

Role of Innovative and Entrepreneurial Behavior

159

regard as possible. Subsequently, the ﬁrm/manager must gain insights into the possibilities, at later stages form attitudes to them, and ﬁnally decide to adopt (or reject) them. Then, the manager may develop and implement the ﬁrm’s export and internationalization strategy. The authors behind the IM viewed the internationalization process as taking place at such stages in the individual ﬁrms. They also applied the stages as framework in crosssectional studies in which they categorized each ﬁrm in order to compare management characteristics, strategies, performance, etc. at the different stages of internationalization (e.g. Cavusgil, 1980). Simmonds and Smith (1968) note that exporting can only be characterized as an intra-ﬁrm innovation. This argument shows that the early internationalization literature drew upon the literature on adoption of innovations, but they did not adopt the view of an innovation proposed by Schumpeter (1961), according to whom innovative behavior is seen as being a new combination of known resources so that a new type of problem-solving is created, i.e. the ability of the organization to revolutionize products and processes. In order to regard exporting as innovative according to Schumpeter’s deﬁnition, it is a condition that decision makers in the ﬁrm have no prior experience with exporting and that the product exported should be new to the foreign market. As empirical research demonstrated, this was only rarely the case at that time. So, even though the IM formulated in the 1970s claims to build upon theories of diffusion and adoption of innovations, it does not deal with genuine innovative behavior in a Schumpeterian sense. Relying mainly on Rogers (1962), Bilkey and Tesar (1977) formulated a six-stage model of export development building on the ideas of adoption of innovations. The deﬁnition of stages was based on management characteristics (e.g. interest in exporting, exploration of the feasibility of exporting) and ﬁrm behavior (e.g. proactivity to exporting, geographical orientation). Also Lee and Brasch (1978) consider the adoption of export as an innovative strategy. They point out that technological innovations have been researched far more than innovations related to new services, strategies, and market opportunities. They investigate the initiating force for exporting which they hypothesize to be inside the ﬁrm (perceived problems or needs) rather than external. They also hypothesize that the adoption process is rational because the decision on exporting and its implications are potentially costly. In their study, they investigate the process leading to adoption, i.e. the stage at which the ﬁrm explores the feasibility of exporting. They do so with reference to the stages model developed by Bilkey and Tesar (1977).

160

ARILD ASPELUND AND TAGE KOED MADSEN

As it appears, the theoretical roots referred to (Schumpeter, Rogers) are reﬂected only partially in the IM literature. Exporting processes are innovative to the ﬁrm/manager and not to the market as such. We will proceed by examining how the empirical ﬁndings of the IM researchers ﬁt with the theoretical approach.

Empirical Findings Simmonds and Smith (1968) expected that the generation of the innovation (the export activity) would be created by an innovator inside the ﬁrm. They carried out a number of empirical case studies to investigate this hypothesis, but in six of nine cases, however, they found that the initiative came from outside (parent companies or customers with international backgrounds). In the remaining three cases, the internal innovator originated from a foreign country and/or had substantial international experience. So, in fact the ﬁrst export activities in these ﬁrms cannot be characterized as an innovation, and they did not seem to stem from any marketing orientation as assumed by the authors. Despite their initial ideas of modeling export development as an adoption of an innovation, Simmonds and Smith never really return to this issue, perhaps because their empirical evidence pointed in other directions. Lee and Brasch (1978) ﬁnd results similar to those of Simmonds and Smith, i.e. about 65% of export activities are initiated by external change agents, and about the same percentage of ﬁrms seem to follow a non-rational decision-making process. Bilkey (1978) considered exporting to be a developmental process in which management characteristics are very important for success. This is very much in accordance with the stages model developed by Rogers (1962) in which he argues that prior conditions (such as previous practice, personality variables, and norms of the social system) are important factors in the initial stages of the adoption of an innovation. Bilkey (1978) refers to empirical results from studies involving surveys or interviews among a total of about 1000 ﬁrms in Wisconsin, Indiana, Georgia, Oregon, and British Columbia, carried out as part of six Ph.D. dissertations written during the years 1968–1976. He refers to a similar amount of studies reported in the business press and still other empirical studies conducted by departments of commerce, or similar public organizations. The empirical evidence suggests that the initiation of exporting is driven by external as well as internal change agents. Managerial experiences, attitudes, and expectations seem to be of decisive importance to the further development of the

Role of Innovative and Entrepreneurial Behavior

161

internationalization process of a ﬁrm. Cavusgil (1980, 1984) also stresses the importance of managerial and organizational characteristics. As a consequence of their focus on managers, the North American researchers carried out empirical studies in which they attempted to identify the top management characteristics typically associated with successful internationalization processes. Axinn (1988) notes, however, that the empirical work in relation to the IM has mainly focused on ﬁrm-related variables, attempting to explain export behavior. She examines six major studies in the period of 1964–1983 and ﬁnds that only one-third of more that 60 explanatory variables are manager-related. Most frequently, manager-related characteristics of exporters are compared with the same characteristics of non-exporters. In her own US-based study, Axinn examines managers’ perceptions of exporting as well as their level of education and overseas work experience. She also includes ﬁrm-related factors such as geographical scope, technology, size, growth, and proﬁtability. Managers’ perception of export advantages and non-complexity as well as their overseas work experience are signiﬁcantly related to the export performance of the ﬁrm.

Concluding Remarks and Formulation of Propositions The focus of the IM is on already existing ﬁrms with long experience on the domestic market, the innovation being the exporting process. As we have seen, the empirical studies show that only a minority of decision makers are innovators in a true sense. Most of them are simply reacting to external stimuli, e.g. a foreign customer that approaches them (e.g. Lee & Brasch, 1978). Only rarely they ﬁnd new types of solutions to problem-solving. Most often they may even attempt to apply their existing problem-solving methods (Axinn, 1988). So, the IM focuses very much on processes that cannot be regarded as innovative in the marketplace as such. Rather, the export activity may be new to the ﬁrm and its mangers that have to adopt different ways of organizing and marketing their existing products. The exporting and internationalization process in the IM is mainly taking place as incremental developments based on existing knowledge. Therefore, managerial perceptions of exporting as well as their prior international experience become very important (Bilkey, 1978). The adoption of exporting typically involves the use of low commitment governance structures such as export agents or foreign distributors. We know from the literature (Bell, McNaughton, & Young, 2001; Rialp et al., 2005; Aspelund et al., 2007) that also many ﬁrms in present market

162

ARILD ASPELUND AND TAGE KOED MADSEN

conditions base their international marketing activities on previous experience and network relationships. Therefore, even ﬁrms with rapid and profound international development may rely very much on external stimuli and previous routines. Also low commitment entry modes are often reported to be used. This lead to the following propositions: P1. External stimuli and change agents are important for international market development even in today’s market conditions. P2. International marketing managers still attempt to apply existing experience and routines when they enter new markets. P3. Low commitment entry modes are still commonplace to use. P4. Even rapidly internationalizing ﬁrms may develop incrementally. If these propositions are reﬂected in real life, the IM model still has something to offer in the effort to meet and respond until the new challenges in international marketing.

THE ‘INTERNATIONAL ENTREPRENEURSHIP MODEL’ Firms that internationalize right from their inception represent the core phenomenon of the new theory building in international entrepreneurship. The new international ﬁrms have been given different names in the literature such as Innate Exporters (Granitsky, 1989), Global Start-Ups (Jolly, Alahutha, & Jeannet, 1992), Born Globals (Rennie, 1993; Knight & Cavusgil, 1996; Madsen & Servais, 1997; Moen, 2002), INVs (McDougall, Shane, & Oviatt, 1994; Oviatt & McDougall, 1994), and Instant Internationals (McAuley, 1999; Preece, Miles, & Baetz, 1999). With the introduction of ‘new ventures’ in the internationalization literature, the concepts of ‘entrepreneurship’ and ‘entrepreneurial behavior’ also emerges in the literature and to some extent replaces the traditional use of the terms ‘innovation’ and ‘innovative behavior’. This change of vocabulary is a natural consequence of the new focus on newly established ﬁrms as well as the crossing of the research paths of international business and entrepreneurship, which leads numerous entrepreneurship scholars to focus on issues related to international marketing. There exists a multitude of entrepreneurship deﬁnitions (Davidsson, 2004). However, there seems to be at least some agreement on the applicability

Role of Innovative and Entrepreneurial Behavior

163

of Lumpkin and Dess’ (1996) deﬁnition of entrepreneurship as new entry. This is also a deﬁnition of entrepreneurship that suits the internationalization scholars as internationalization deals with international new market entry and, hence, can be viewed as an act of entrepreneurship (Ibeh & Young, 2001). Hence, on the basis of Lumpkin and Dess (1996), we deﬁne entrepreneurial behavior as the conduct of processes, practice, and decisionmaking activities that lead to new entry. Under the inﬂuence of the entrepreneurship scholars and Lumpkin and Dess’ deﬁnition of entrepreneurship, the IE literature has shifted focus from the narrow deﬁnition of ‘innovative behavior’ as an act of adopting innovations to a far broader deﬁnition of ‘entrepreneurial behavior’ where internationally oriented entrepreneurs supply international markets with new innovations and use various entrepreneurial strategies to set up an innovative organization to exploit international opportunities. This broadening of perspectives has also lead to a broader focus in terms of research. On the one hand, IE model lifts the focus out of the organization and view the internationalization process as an engine to provide new innovations in international markets. On the other hand, there has been far more focus on the individual entrepreneurs (or entrepreneurial team) and their behavior and impact in the internationalization process.

Theoretical Roots The ﬁrst serious effort to create a new model adapted to INVs was published by Oviatt and McDougall in 1994 in Journal of International Business Studies. This study has turned out to be a seminal work and is cited in all but one of the studies of INVs recently reviewed by Aspelund et al. (2007). It is also the study that forms the basis of what we refer to as the IE model in this study. The model is an integrative framework based on theories from strategic management, international business, and entrepreneurship. Basically, the model depicts four necessary and sufﬁcient conditions for the existence of INVs. One of the conditions is taken from the entrepreneurship literature and deals speciﬁcally with entrepreneurial behavior in terms of resource management associated with the general resource poverty of new organizations. Owing to the lack of resources that is generally associated with new ﬁrms (Vesper, 1990), new ventures often employ alternative governance structures (or hybrid structures) to control productive resources. This is also the case regarding INVs, which frequently use hybrid structures for international market access (Gabrielsson & Kirpalani, 2004).

164

ARILD ASPELUND AND TAGE KOED MADSEN

Another condition for the emergence of INVs is unique resources. This argument stems from the resource-based view of the ﬁrm which has its roots in Penrose’s (1959) theory of the growth of the ﬁrm. Penrose’s perspective of looking at organizations as bundles of resources and capabilities is now being applied to a still higher degree in the entrepreneurship literature. Oviatt and McDougall (1994) apply resource-based arguments and argue that if the new ﬁrm want to survive, it must introduce an innovative market offer that is both valuable to the customer and hard for competitors to imitate or substitute. In conclusion, Oviatt and McDougall (1994) point to two features of entrepreneurial behavior that are preconditions for the existence of INVs. First, management must be entrepreneurial in their resource management in order to reach new business space rapidly. Second, the introduction of the new ﬁrm’s product and services must have innovative features on the market to secure a certain level of competitive ability.

Empirical Evidence In the following section, we will examine how this view of entrepreneurial behavior has been reﬂected in the empirical work of the IE researchers on three levels. First, we will present evidence on how the IE model views INVs as innovative change agents by spreading new innovations into new markets. Second, we will see how international entrepreneurs use entrepreneurial resource management in order to overcome resource constraints. Finally, we will see some ﬁndings on the role of the international entrepreneur as an individual inﬂuence the internationalization process.

INVs as Bearers of Innovations The IE model research brings forward a market-oriented view of innovations in comparison with the IM model that treats innovation only on the organizational level. Several studies in international entrepreneurship bring forward the role of new technologies in the internationalization process. First, the fact that many INVs are technology oriented, at least in terms of their competitive advantages (Aspelund & Moen, 2005), has brought many researcher to look at the role of new technologies in internationalization processes. The fact that a new ﬁrm is trying to introduce a new technology into a market inﬂuences the internationalization process in several ways.

Role of Innovative and Entrepreneurial Behavior

165

First, both Bloodgood and Sapienza (1996) and Aspelund and Moen (2005) found that a dominant part of internationalizing new ventures grew their competitive advantage out of technological innovation. Hence, the primary driver behind the international expansion of the ﬁrms is to proﬁt from disseminating of new technologies in multiple markets. Zahra, Matherne, and Carleton (2003) found that new technologies in its own right facilitated internationalization as INVs frequently used their technological networks and technological reputation to push a more rapid and extensive international expansion. Oesterle (1997) also identiﬁed a push factor related to technology. In his case study of Daimler, he saw that Daimler’s innovative technologies represented such a destructive potential for domestic incumbents that the domestic market turned hostile toward the new entrant. Hence, domestic market hostility pushed the new technology entrant into less hostile international markets. Finally, Jolly et al.’s (1992) study of Global Start-Ups is a ﬁne example of the role the INVs play in order to disseminate new technologies. They ﬁnd that many INVs time their global market expansion in industry shifts where new technologies are replacing old and the incumbents’ competitive position is weakened. Even though few studies mention this speciﬁcally; the underlying motivation for studying INVs is INVs’ role as global change agents that increase efﬁciency in international markets by disseminating new technologies.

Entrepreneurial Resource Management One of the factors that distinguish the IE model from the IM model is that the former lifts the focus out of the expanding ﬁrm and into the network context in which it operates. Entrepreneurs in INVs need control of resources that are too expensive to internalize and, hence, seek to gain control of resources through entrepreneurial arrangements with business partners. Two examples are found in Coviello and Munro’s (1997) study of rapidly internationalizing Canadian software ﬁrms and in Eriksson and Chetty’s (2003) study. They both ﬁnd that the experiential knowledge needed to conduct international business can be accumulated in business networks rather than solely internalized in the ﬁrm itself. This is an innovative approach to knowledge management and a feature that is closely tied to entrepreneurial behavior (Oviatt & McDougall, 1994). The latter study also points to another aspect frequently used by entrepreneurs, namely the

166

ARILD ASPELUND AND TAGE KOED MADSEN

tendency to use social relations rather than ownership to control the behavior of their business partners. Hence, international entrepreneurs tend to select trustworthy business partners from their social network. Another study exemplifying INVs’ use of entrepreneurial arrangements for market entry is Gabrielsson and Kirpalani’s (2004) study of new resource-constrained, knowledge-intensive ﬁrms. They identify and exemplify four alternative routes to international markets that are frequently used and well ﬁtted for resource-constrained new ﬁrms with international ambitions. All these four arrangements are hybrid structures, which depend on either the involvement of an independent business partner or the extensive use of the Internet as a global marketing channel.

The Role of the Entrepreneurs Some of the recent studies of international entrepreneurship attempt to disengage from the traditional views of internationalization. For example, experiential market knowledge is treated only as one of many ﬁrm resources that form the internationalization process of ﬁrms. This perspective is often referred to as the holistic perspective because of the view that all ﬁrms form idiosyncratic internationalization patterns shaped by the resources and capabilities the ﬁrm possesses and the environmental opportunities and constraints it faces. Empirical studies such as McDougall et al. (1994), Bell (1995), Jones (1999), McAuley (1999), Crick and Jones (2000), and Moen and Servais (2002) have documented that new ﬁrms often exhibit far more heterogeneity in their manners of internationalization than the traditional, incremental internationalization models predict, and hence call for a holistic model of ﬁrm internationalization that incorporates the diversity of internationalization patterns. Basically, such diversity develops because the individual entrepreneur has such an important role to play. Good examples of how entrepreneurial behavior affects the internationalization process of the ﬁrm can be identiﬁed in studies of opportunity recognition. A pioneering study of that kind is McDougall et al.’s (1994) study of 24 INVs. They investigated three fundamental questions, namely who are INV founders, why are their new ventures international, and how do they organize their international activities. They concluded that INV founders are entrepreneurs that are able to see international opportunities from successful combinations of resources from different markets and that they tend to have this ability because of their extensive international experience. This ﬁnding is supported by a quantitative

Role of Innovative and Entrepreneurial Behavior

167

study by Harveston, Kedia, and Davis (2000), which concludes that managers in INVs have a more global mindset, more extensive international experience, and a higher risk tolerance than managers in gradually internationalizing ﬁrms. Both studies conclude that managers with international experience are more likely to perceive and act on international business opportunities. Three studies that take an holistic perspective and emphasize the role of the entrepreneur are McAuley’s (1999) study of ﬁrms in the Scottish arts and craft sector, Crick and Jones’ (2000) study of UK high-tech ﬁrms, and Crick, Chaudhry, and Batstone’s (2001) study of minority-owned ﬁrms in the UK clothing industry. All these studies conclude that the behavior of the entrepreneur is central in the internationalization process of new ﬁrms, both for the speed of internationalization and for the subsequent international behavior of the ﬁrm. They also conclude that unplanned internationalization is widespread and that entrepreneurs are often pulled into internationalization without initial strategic intent. It is interesting, however, to see the great variety of motivation for quick and extensive internationalization in these studies. Crick et al.’s (2001) minority-owned ﬁrms expanded early to exploit their international business network in the area of their ethnic origin. Crick and Jones’ (2000) high-tech ﬁrms exhibited often a more planned internationalization, but it was primarily motivated by the insufﬁcient size of the domestic market. McAuley’s (1999) ﬁrms internationalized in an ad hoc manner, often through contacts made at trade fairs or through personal networks and seemed to be motivated by the opportunity for increased sales and proﬁt.

Concluding Remarks and Formulation of Propositions The new theory building in the internationalization literature seems to have a strong foothold within the ﬁeld of entrepreneurship. Especially three aspects of entrepreneurial behavior in the IE model are supplementary to the concept of innovative behavior. The ﬁrst is the focus on market innovations as sources of competitive advantage and motivation for international expansion rather than innovations on the organizational level as new ways of doing business. Second, we see focus on international entrepreneurs’ tendencies to form hybrid structures of international resource and knowledge management to overcome resource constraints. Hybrid structures illustrate how international entrepreneurs are able to establish and use efﬁcient international sales

168

ARILD ASPELUND AND TAGE KOED MADSEN

and marketing channels that they cannot afford to control by ownership. Furthermore, we see the tendency to rely on alternative governance strategies than ownership to control international market channel behavior. A major consequence of this development is that the research attention has been lifted out of the organization and into the business network in which it operates. The third is the effect of the behavior of the individual entrepreneurs and their tendencies for opportunity recognition, which is a concept that has migrated from the entrepreneurship literature. It exempliﬁes how both entrepreneurs’ former educational and business experience and ethnical and cultural background affect the cognitive processes of opportunity recognition and guides both the speed and direction of its exploitation. Hence, the characteristics of the entrepreneur or the founding team have consequences, both for the likelihood of establishing INVs and for the subsequent internationalization process of the ﬁrm. A result of this development is that the entrepreneur or the founding team has more frequently been used as the unit of analysis in studies of internationalization. These observations lead to the following propositions: P5. Genuine market innovations are presently important drivers of international competitive advantage. P6. In today’s market conditions, low commitment entry modes are supplemented by hybrid structures involving some degree of control with partners’ resources and competences. P7. The characteristics of the entrepreneurs and founding teams are of decisive importance for the ﬁrm’s response to international marketing challenges.

DISCUSSION AND IMPLICATIONS FOR INTERNATIONAL MARKETING As demonstrated earlier, the IM and the IE models have similarities, but they also have very different ways of studying behavior, especially in terms of motivation. IM researchers were focusing on the ﬁrst international steps of existing ﬁrms with a strong foothold in their domestic markets. As a consequence, they primarily studied ﬁrm-related characteristics and

169

Role of Innovative and Entrepreneurial Behavior

Table 1.

Innovative and Entrepreneurial Behavior Studied. Innovation-Related Internationalization Model

Theoretical foundation

Unit of analysis Driver of development Decision-making process Entry modes Speed of development Degree of newness The effect of the innovation

Adoption of innovations in organizations (mainly Rogers, 1962) Existing ﬁrms Growth-driven, external agents Reactive Traditional governance structures Slow, gradual, incremental New to the ﬁrm Existing/slightly modiﬁed products in new markets

International Entrepreneurship Model Entrepreneurship, resource-based theory, and international business Entrepreneurs, new ﬁrms, and business networks Opportunity-driven, internal change agents Proactive Hybrid governance structures Often very fast Often new to the market as well as to the ﬁrm New products and technologies in new markets

included managerial characteristics only as a supplement. They focused on the adoption of innovations and had the ﬁrm, not the individual, as their primary unit of analysis. The IE researchers focus on the founder(s) of a new ﬁrm as well as the effects of introducing new technologies. For that reason, they build on entrepreneurship theories of the innovative, proactive, and risk-seeking individual that are able to recognize and exploit international market opportunities in collaboration with partners in their network. Table 1 summarizes some of the most important differences between the two approaches. In the ensuing we will discuss these special characteristics and relate them to new challenges in international marketing in order to give more ‘ﬂesh and blood’ to the propositions developed earlier. We have structures the discussion according to behavior at the individual level, behavior at the organizational level, and ﬁnally, the effects of innovative and entrepreneurial behavior on international markets.

Innovative and Entrepreneurial Behavior at the Individual Level In a recent review, Aspelund et al. (2007) reported that one of the most consistent ﬁndings from their review of the INV literature is the vital and

170

ARILD ASPELUND AND TAGE KOED MADSEN

long-lasting effects of the international entrepreneurs and their strategic decisions in early phases of new international ventures. ‘‘Even though many studies have identiﬁed this phenomenon, few studies have been designed to go into depth on the strategic processes that occur near founding, . . . , in particular on how key decisions in the establishment process inﬂuence on the strategic and marketing related decisions of the ﬁrm and how these decisions inﬂuence on the development of the ﬁrm’’ (p. 1442). The undisclosed relationship between entrepreneurial acts in founding processes and international marketing strategies remains a challenge for both scholars and practitioners as consequences are apparent. The internal change agent as the driver of the internationalization process within the ﬁrm was a key concept within the IM as mentioned earlier. This ‘ﬁrm champion’ worked with the organization in order to make it adopt international business. Many IE researchers, however, redeﬁned this concept into an entrepreneur that rather designed his/her organization for international business from the outset. They often consider the decisive role of the entrepreneur’s previous professional experience of international opportunity recognition and the importance of personal network for market selection and choice of entry modes. This resembles very much the empirical ﬁndings of the IM researchers with regard to the importance of managers’ perceptions of and experience with international markets. The role of the entrepreneur is to be the person that works with the organization to bring continuous change. Findings suggest that it is the organization’s ability to learn and change according to international market demands that ultimately bring international performance, not the decision to internationalize in itself (McDougall & Oviatt, 1996; Zahra, Ireland, & Hitt, 2000; Aspelund & Moen, 2005).

Innovative and Entrepreneurial Behavior at the Organizational Level Other recent reviews have pointed to challenges on the organizational level that should be sorted out. One of these is Rialp et al. (2005) that argues that focus should be put toward identifying organizational capabilities that ‘‘ . . . explain and interpret not only the emergence of early internationalizing ﬁrms but also its further development in the form of a rapid and sustained international growth’’ (p. 162). Along these lines, IM studies may supply new ideas to the study of established ﬁrms that suddenly go through rapid internationalization, often labeled Born-Again Globals (Bell et al., 2001). It is likely that managers in these types of ﬁrms may be reacting on external stimuli just as the IM

Role of Innovative and Entrepreneurial Behavior

171

researchers demonstrated was the case 30 or 40 years ago. Hence, the IM ideas of internal change agents and reactive behavior may prove to be relevant when studying all internationalization processes. Moreover, it might be very advisable for entrepreneurs to be more open toward possible external agents that may change the development path of the ﬁrm. Especially in today’s global market conditions where such external change agents are widely available. Furthermore, the focus of the IM model on the organizational level is highly relevant for INVs as well. During the ﬁrst years of existence, the entrepreneurs themselves are very important for the development of the ﬁrm. Rapid growth is often the result, but after that many INVs do not grow very fast (Rennie, 1993). One reason may be that the entrepreneurs are not aware that the adoption of new innovations in the ﬁrm is depending on the norms and attitudes of the whole organization. Being alert with regard to forming attitudes among all employees is important in the process of adopting new products and processes. As Zahra (2005) concludes: ‘‘There is a dearth of empirical studies on the evolution of INVs, as most prior research has focused on the founding and operations of these ﬁrms’’ (p. 27). It may be argued that future research on innovative behavior and internationalization processes must combine the two approaches and have a dual unit of analysis, i.e. the individual as well as the ﬁrm. Where the entrepreneur traditionally has been ‘the lonesome wolf ’, it has been demonstrated that the foundation process of Born Global Firms in most cases involves several entrepreneurs as well as several other persons that play an active role. Therefore, it is relevant to study some kind of organization. It is also well documented (see e.g. Bell et al., 2001) that many ﬁrms nowadays experience quick and highly individual international marketing processes that resemble those of Born Global Firms. So, in order to completely understand the internationalization processes taking place under the present market conditions, it seems to be necessary to carry out analyses at both individual and ﬁrm level as well as the business network from which the internationalizing ﬁrm gain access to important resources.

The Effect of Innovative and Entrepreneurial Behavior on International Markets Finally, we would like to draw attention to a remaining gap in the literature on international marketing, namely that of the effects of the diffusion of managerial, organizational, and technological innovations that is embedded

172

ARILD ASPELUND AND TAGE KOED MADSEN

in the internationalization of ﬁrms. The IM primary view of innovation is that international business is a new way of doing things for the ﬁrm and for that reason an innovation that resides within the ﬁrm. It is recognized that exporting is not an innovative act in the marketplace as such; many ﬁrms have already international experience and are used to marketing similar goods on many international markets. However, one could also argue that all organizations bear with them distinct managerial and organizational routines that may be viewed as innovations in the markets which they enter and with time are adopted by local ﬁrms. The IE has another point of departure, namely a newly established ﬁrm, and oftentimes a ﬁrm that has developed a technologically advanced product. Hence, the innovation in this case is not only something internal to the ﬁrm. The product or technology is also new to the market, which implies that an innovation is taking place as new solutions or technologies are introduced. This perspective of international entrepreneurship has been all but neglected in the literature on international marketing and we know very little of the effects of dissemination of managerial, organizational, and technological innovations through internationalization of ﬁrms.

Directions for Future Research Observing the year of publishing of the reviewed articles in this study, the reader might be led to believe that the IM perspective has passed its expiry date and that the knowledge developed from this perspective is drowned in the extensive, recent IE literature on INVs. Actually, published contributions of the former strand are quite limited over the past 10–15 years. This is a loss to the debate, and we would like to conclude with some of the features of innovative behavior that may add to our understanding of challenges of international marketing and for that reason still deserve attention from academics. First of all, the IM literature’s focus on the organization as such should not be forgotten. The impact of the entrepreneurship literature has shifted the focus to the individual entrepreneurs and their networks in the marketplace. However, also INVs have to develop into ongoing concerns, so organizational issues are deﬁnitely important to incorporate into the research on the phenomenon. One excellent example of this is given in McDougall and Oviatt (1996) study of internationalization and performance of ﬁrms. They found that internationalization per se is not associated with performance. Only the companies that are capable of strategic change

Role of Innovative and Entrepreneurial Behavior

173

to adapt to international market conditions can turn their internationalization efforts into higher proﬁts. This is even more the case if the models developed are more general, including also the phenomenon of Born-Again Globals. This type of ﬁrm is an existing organization that completely changes its scope from local to global. In that process, where managers seek to turn a purely domestic organization into a global player, organizational issues related to adoption of new ways of doing business are highly relevant. Second, we ﬁnd that also the IM focus on reactive behavior may have merit to the international marketing challenges of INVs. The empirical work on the IM demonstrates that external change agents are often very important and that in their internationalization process decision makers in the ﬁrm often react to the behavior of such external agents. A similar situation often arises when small INVs react to competitive actions from large incumbents. This may be new products, other pricing strategies, or even an offer to take over the INV. Therefore, the idea of reactive behavior should not be forgotten, even if we are dealing with entrepreneurs that are mainly depicted as being autonomous, innovative, risk-taking, proactive, and strategically aggressive. We have formulated seven propositions that may be investigated in future research. In doing so, it is important to notice that empirical studies have to be carried out at multiple levels. As the propositions show, it may be the efforts of entrepreneurs, founding teams, the entrepreneurial organization, and its combined network relationships that are relevant when studying the best practices in the attempt to cope with the new challenges to international marketing.

REFERENCES Andersen, O. (1993). On the internationalization process of ﬁrms: A critical analysis. Journal of International Business Studies (Second Quarter), 209–231. Aspelund, A., Madsen, T. K., & Moen, Ø. (2007). International new ventures: Review of conceptualizations and ﬁndings. European Journal of Marketing, 41(11/12), 1423–1474. Aspelund, A., & Moen, Ø. (2005). Small international ﬁrms: Typology, performance and implications. Management International Review, 45(3), 37–57. Axinn, C. N. (1988). Export performance: Do managerial perceptions make a difference? International Marketing Review (Summer), 61–71. Bell, J. (1995). The internationalization of small computer software ﬁrms: A further challenge to ‘‘stage’’ theories. European Journal of Marketing, 29(8), 60–75. Bell, J., McNaughton, R., & Young, S. (2001). ‘‘Born-Again Global’’ ﬁrms: An extension to the ‘‘Born Global’’ phenomenon. Journal of International Management, 7, 173–189.

174

ARILD ASPELUND AND TAGE KOED MADSEN

Bilkey, W. J. (1978). An attempted integration of the literature on the export behavior of ﬁrms. Journal of International Business Studies (Spring–Summer), 33–46. Bilkey, W. J., & Tesar, G. (1977). The export behavior of smaller-sized Wisconsin manufacturing ﬁrms. Journal of International Business Studies (Spring–Summer), 93–98. Bloodgood, J. M., & Sapienza, H. J. (1996). The internationalization of new high-potential U.S. ventures: Antecedents and outcomes. Entrepreneurship Theory & Practice (Summer), 61–76. Cavusgil, S. T. (1980). On the internationalisation process of ﬁrms. European Research (November), 273–281. Cavusgil, S. T. (1984). Differences among exporting ﬁrms based on their degree of internationalization. Journal of Business Research, 12, 195–208. Coviello, N., & Munro, H. (1997). Network relationships and the internationalization process of small software ﬁrms. International Business Review, 6(4), 361–386. Crick, D., Chaudhry, S., & Batstone, S. (2001). An investigation into the overseas expansion of small Asian-owned U.K. ﬁrms. Small Business Economics, 16, 75–94. Crick, D., & Jones, M. (2000). Small high-technology ﬁrms and international high-technology markets. Journal of International Marketing, 8(2), 63–85. Davidsson, P. (2004). Researching entrepreneurship. New York: Springer. Eriksson, K., & Chetty, S. (2003). The effect of experience and absorptive capacity on foreign market knowledge. International Business Review, 12, 673–695. Fletcher, D. (2004). International entrepreneurship and the small business. Entrepreneurship & Regional Development, 16(July), 289–305. Gabrielsson, M., & Kirpalani, V. H. M. (2004). Born globals: How to reach new business space rapidly. International Business Review, 13, 555–571. Granitsky, J. (1989). Strategies for innate and adoptive exporters: Lessons from Israel’s case. International Marketing Review, 6(5), 50–65. Harveston, P. D., Kedia, B. L., & Davis, P. S. (2000). Internationalization of born global and gradual globalizing ﬁrms: The impact of the manager. Advances in Competitive Research, 8(1), 92–98. Ibeh, K. I. N., & Young, S. (2001). Exporting as an entrepreneurial act – An empirical study of Nigerian ﬁrms. European Journal of Marketing, 35(5/6), 566–586. 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, 23–32. Jolly, V. K., Alahutha, M., & Jeannet, J.-P. (1992). Challenging the incumbents: How high technology start-ups compete globally. Journal of Strategic Change, 1, 71–82. Jones, M. V. (1999). The internationalization process of small high-technology ﬁrms. Journal of Marketing Management, 7(4), 15–41. Knight, G. A., & Cavusgil, S. T. (1996). The born global ﬁrm: A challenge to traditional internationalization theory. Advances in International Marketing, 8, 11–26. Lee, W. Y., & Brasch, J. J. (1978). The adoption of export as an innovative strategy. Journal of International Business Studies (Spring–Summer), 85–104. Lumpkin, G. T., & Dess, G. G. (1996). Clarifying the entrepreneurial orientation construct and linking it to performance. Academy of Management Review, 21(1), 135–172. Madsen, T. K., & Servais, P. (1997). The internationalization of born globals: An evolutionary process? International Business Review, 6(6), 561–583. McAuley, A. (1999). Entrepreneurial instant exporters in the Scottish arts and crafts sector. Journal of International Marketing, 7(4), 67–82.

Role of Innovative and Entrepreneurial Behavior

175

McDougall, P. P., & Oviatt, B. M. (1996). New venture internationalization, strategic change, and performance: A follow-up study. Journal of Business Venturing, 11, 23–40. McDougall, P. P., Shane, S., & Oviatt, B. M. (1994). Explaining the formation of international new ventures: The limits of theories from international business research. Journal of Business Venturing, 9, 469–487. Moen, Ø. (2002). The born globals: A new generation small European exporters. International Marketing Review, 19(2), 156–175. Moen, Ø., & Servais, P. (2002). Born global or gradual global? Examining the export behavior of small and medium-sized enterprises. Journal of International Marketing, 10(3), 49–72. Oesterle, M.-J. (1997). Time-span until internationalization: Foreign market entry as a built-inmechanism of innovation. Management International Review, 37(2), 125–149. Oviatt, B. M., & McDougall, P. P. (1994). Toward a theory of international new ventures. Journal of International Business Studies, 25(1), 45–64. Penrose, E. (1959). The theory of the growth of the ﬁrm. Oxford: Blackwell. Preece, S. B., Miles, G., & Baetz, M. C. (1999). Explaining the international intensity and global diversity of early-stage technology-based ﬁrms. Journal of Business Venturing, 14(3), 259–281. Rennie, M. W. (1993). Global competitiveness: Born global. The McKinsey Quarterly (4), 45–52. Rialp, A., Rialp, J., & Knight, G. A. (2005). The phenomenon of early internationalizing ﬁrms: What do we know after a decade (1993–2003) of scientiﬁc enquiry. International Business Review, 14, 147–166. Rogers, E. M. (1962). Diffusion of innovations. New York: The Free Press. Schumpeter, J. A. (1961). History of economic analysis. New York: Oxford University Press. Simmonds, K., & Smith, H. (1968). The ﬁrst export order: A marketing innovation. British Journal of Marketing (Summer), 93–100. Tookey, D. A. (1964). Factors associated with success in exporting. The Journal of Management Studies (March), 48–64. Vesper, K. H. (1990). New venture Strategies: Revised edition. Englewood Cliffs, NJ: Prentice Hall. Zahra, S. A. (2005). A theory of international new ventures: A decade of research. Journal of International Business Studies, 36(1), 20–28. Zahra, S. A., Ireland, R. D., & Hitt, M. A. (2000). International expansion by new venture ﬁrms: International diversity, mode of market entry, technological learning, and performance. Academy of Management Journal, 43(5), 925–950. Zahra, S. A., Matherne, B. P., & Carleton, J. M. (2003). Technology resource leveraging and the internationalization of new ventures. Journal of International Entrepreneurship, 1, 163–186.

BRAND IMAGE PERCEPTIONS ACROSS CULTURES: A STUDY OF SYMBOLIC AND FUNCTIONAL ASSOCIATIONS Laura Salciuviene, Pervez N. Ghauri, Audra I. Mockaitis and Claudio De Mattos Long-term strategies aiming at a successful brand image positioning depend, in part, on how the unique attributes that correspond to the core values and tastes of consumers in different countries manifest themselves (Hsieh, Pan, & Setiono, 2006). Companies that match their brand image to consumer values will more easily attain a competitive advantage over those that do not (Hofstede, Steenkamp, & Wedel, 1999). A brand’s image creates associations with a consumer’s culture and his/her daily surroundings; managers need to know and understand those particular brand associations that best appeal to their consumers with respect to their values (Hsieh, 2002; Brangule-Vlagsma, Pieters, & Wedel, 2002). With the enlargement of the European Union, the debate about the convergence of consumer perceptions and tastes has become even more important (Shintaro, Charles, & Jonathan, 2007). Although previous studies have investigated factors of brand image management, few of them have explored the role of personal values in brand image perceptions of consumers. Moreover, as with many other marketing concepts, much of New Challenges to International Marketing Advances in International Marketing, Volume 20, 177–191 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020010

177

178

LAURA SALCIUVIENE ET AL.

what is known about brand image is based primarily on studies conducted in Western countries. As suggested by Patterson and Smith (2003), ‘‘while it is well understood that there is a high level of importance accorded the role Western cultures have played in building consumer behaviour models, it would seem there is little manifest interest in validating their explanatory powers in markedly different cultures’’ (p. 108). Indeed, little research exists on ‘‘consumer reactions in the new European Union member countries’’ (Vida & Reardon, 2008, p. 36). Hence, our study addresses this underresearched area. More speciﬁcally, it examines how personal values affect perceptions of brand image in relation to the functional and symbolic attributes of the brand in different cultural contexts. Speciﬁcally, our study aims to provide answers to the following questions: 1. How do consumers perceive a brand image and how these perceptions are related to consumer values? 2. Which speciﬁc consumer values correlate with speciﬁc attributes of a particular product and brand image in each country? 3. How perceptions of brand image attributes differ in each country? By assessing the extent to which personal values affect brand image perceptions in two EU countries, one in Eastern Europe (Lithuania) and the other in Western Europe (Denmark), our study responds to recent calls in the literature to foster cross-cultural research by testing existing marketing concepts in new countries (Burgess & Steenkamp, 2006; Yaprak, 2007). Thus, our study aims to further our understanding about how brand image attributes in different cultural contexts are perceived and how those perceptions, as a function of consumer personal values, differ. The chapter is structured as follows. We ﬁrst review existing knowledge on key issues involving brand image and link this to consumer values. The methods, ﬁndings and discussion follow in the next section. The chapter ends with the conclusions, implications for international marketing practice and guidelines for further research.

THEORETICAL BACKGROUND Brand Image The concept of brand image has received considerable attention in marketing (Batra & Homer, 2004; Dhar & Wertenbroch, 2000; Roth, 1992; Thompson, Rindﬂeisch, & Arsel, 2006; van Reijmersdal, Neijens, & Smith, 2007;

Brand Image Perceptions across Cultures

179

van Rekom, Jacobs, & Verlegh, 2006), yet there is still little agreement on its deﬁnition and operationalisation in the literature. As Dobni and Zinkhan (1990) observed, despite the frequent use by scholars of the term ‘‘brand image,’’ its deﬁnitions in the literature tend to focus on different elements. It is possible to group deﬁnitions of brand image into different categories. For example, brand image has been deﬁned as (a) an attitude extending its meaning beyond the physical product (e.g., Reynolds & Gutman, 1984) and (b) perception, relating brand image to psychological aspects of a product’s tangible attributes (e.g., Keller, 1998). One generally accepted view is that brand image can be deﬁned as perceptions regarding a brand as reﬂected by the cluster of associations that consumers connect to the brand name in memory (Herzog, 1963). This is consistent with an associative network memory model. Thus, ‘‘brand associations are the other informational nodes linked to the brand node in memory and contain the meaning of the brand for consumers’’ (Keller, 2003, p. 66). Drawing on Allen (2001) (holistic evaluation), Keller (1998) (strength, favourability) and Park, Jaworski, and MacInnis (1986) (functional, symbolic and experiential), we hold that a brand image consists of the collection of perceived functional and symbolic associations (attributes) that vary in strength and favourability in accordance with consumer values. Thus, as part of the overall branding strategy, a brand’s image must manifest itself by incorporating those attributes that consumers prefer in a speciﬁc cultural context. Attributes deﬁning a brand have consequences for the individual, and these reinforce individual values. Symbolic brand image attributes are associated with symbolic and intangible beneﬁts offered by a brand, such as status, group membership or self-image. Functional brand image attributes relate to the tangible beneﬁts offered by the brand, such as the ability to perform certain functions or a high level of technical standards. Moreover, attributes allow consumers to achieve speciﬁc beneﬁts, and consumers attach their personal values to brand beneﬁts in considering what the brand can do for them (del Rio, Vazquez, & Iglesias, 2001). In other words, consumers prefer brands, the values of which correspond to their own values (de Chernatony, 2002).

The Role of Personal Values in Consumer Behaviour Consumer behaviour literature recognises values being a powerful force shaping consumers’ motivations, lifestyles and product choices. This topic has received extensive attention over the past three decades

180

LAURA SALCIUVIENE ET AL.

(Honkanen, Verplanken, & Olsen, 2006; Marandi, Little, & Sekhon, 2006; Gullestrup, 2006). However, value studies have also been constantly criticised for their fragmented nature and inadequate value conceptualisation. As Fulton and Lipscomb (1996) point out ‘‘few studies offer a clariﬁcation of the value concept and several researchers have noted ambiguity in the use of terms of value or values’’ (p. 25). One of the main difﬁculties in deﬁning values is the disagreement on what constitutes a value and what value priorities are (Reisinger & Turner, 2003). For instance, Rokeach (1973) differentiated between ‘‘instrumental’’ and ‘‘terminal’’ value dimensions. According to Vinson, Scott, and Lamont (1977), values stimulate motivation for behavioural response in respect to product attributes. In addition, Schwartz and Bilsky (1987) suggest ﬁve central features of values: ‘‘values: (a) are concepts or beliefs, (b) about desirable end states or behaviours, (c) that transcend a speciﬁc situation, (d) guide the selection or evaluation of behaviour and events, and (e) are ordered by relative importance’’ (p. 551). On the basis of previous work, we hold that ‘‘values are generalised, relatively enduring and consistent priorities for how we want to live’’ (Zetterberg, 1998, p. 995) and how we want ourselves to be perceived.

Relating Brand Image Perceptions to Personal Values One stream of research suggests that consumer perceptions of brands are based on consumer needs (see Maslow, 1959). Needs are determined by personal values and certain values are considered as more important to consumers in one country than in others (Kim, Forsythe, Gu, & Moon, 2002). Consumers try to satisfy functional needs by searching for functional brands (Park et al., 1986). Functional brands, as asserted by Bhat and Reddy (1998), fulﬁl immediate and practical needs. Symbolic needs such as self-enhancement or self-image are fulﬁlled by symbolic brands (Thompson et al., 2006). Another stream of research proposes that brands can have different attributes to be related to perceived brand image beneﬁts (Bhat & Reddy, 1998). Allen (2001) states that functional brand image attributes are of utmost importance for consumers. Thompson et al. (2006) go further and argue that a focus only on functional attributes of brands will not guarantee long-term competitive advantage in international markets as the brand should meet emotional consumer expectations as well. Therefore, perception of the brand image has become separated from its original use, and it is

Brand Image Perceptions across Cultures

181

attached with new, emotional and symbolic perception such as heightened self-esteem, beauty or happiness (Firat, 1991; Zaltman, 2003). Two further points are important in the context of this study. First, D’Andrade (1986) outlines the contemporary shift in culture theory from the view of culture as patterns of behaviour to one of culture as knowledge, meanings and symbols. This shift has important implications. For example, Ricoeur (1969) reports that symbols do not simply arise from societal conventions or culture. Second, consumers might have different preferences for brands in different cultures. For example, Arnould, Price, and Zinkhan (2004) state that brands should reﬂect cultural meanings in concrete forms by providing associations with tangible beneﬁts. On the basis of an investigation of studies addressing brand image and personal value issues, we assert that brand image perceptions are inextricably linked with the inner characteristics of consumers, namely personal values. Despite previous research in the ﬁeld of brand management, scholars have rarely explored how brand image is perceived from the consumers’ perspective or offered explanations as to what is preferred by consumers in view of their personal values. Therefore, our study addresses these issues.

METHOD In respect to sampling procedures, the guidelines for two sampling levels namely the unit of analysis and the selection of countries (including the number of countries, market size and geographic distance) in cross-cultural studies were closely followed (Craig & Douglas, 2006; Reynolds, Simintiras, & Diamantopoulos, 2003; Sinkovics, Penz, & Ghauri, 2005; Sekaran, 1983; van de Vijver & Leung, 1997). We chose to conduct our study on mobile telephone brands. Mobile telephones are noticeable and visible consumer products and carry an element of fashion. These combined with technical product attributes may have an inﬂuence on brand choice. The products chosen contain associations of both a ‘‘symbolic’’ and ‘‘functional’’ attributes for consumers (Bhat & Reddy, 1998). Since price plays an important role in consumer perceptions (Kim & Kramer, 2006), this study intends to minimise the inﬂuence of price on consumer perceptions, as it includes only those respondents who purchased a mobile telephone for the negligible prices of 1Lt (0.33 Euros) or 1DKK (0.16 Euros) when signing a long-term contract with a telecommunication company.

182

LAURA SALCIUVIENE ET AL.

Table 1. Country

Lithuania Denmark

Sample Characteristics.

Sample Size

Percent of Females

Education

Age Range

150 140

52 54.3

Undergraduate students in business management and administration

20–23 20–24

We conducted surveys in both Lithuania and Denmark because both the countries have a relatively similar market size (populations of approximately three and a half million in Lithuania and ﬁve million in Denmark). Lithuania and Denmark are relatively close geographically (Dunning, 1993) and are relatively culturally homogeneous (Hofstede, 2001). Data were collected only from respondents who have had previous experience in purchasing these products and sufﬁcient knowledge about leading brands, as they have tried more than one brand in the mentioned product group. To control for homogeneity of demographics, the sample is composed of university students. We wanted to survey consumers that are most attractive to the particular product and are between the age of 20 and 30. The use of students should ensure sampling equivalence that is usually required in cross-cultural research (Kumar, 2000). Students are representatives of typical mobile phones consumers (Lages & Fernandes, 2005). As suggested by Cunningham, Young, Lee, and Ulaga (2006) ‘‘while students might not be representative of the populations of their respective countries, they are representative of a consuming population with inﬂuences of their respective culture’’ (p. 197). Table 1 summarises sample characteristics. We distributed 800 questionnaires in each country under investigation. Before distributing the questionnaires, we explained the main purpose of the study to the students and gave them the opportunity to participate in a lottery. This can explain the relatively high response rates of 19.6% in Lithuania and 18.7% in Denmark. After a careful inspection, a total of 290 valid questionnaires, that is, 150 questionnaires from Lithuanian students and 140 from Danish students were deemed appropriate for use.

Measures The survey questionnaire was developed carefully following the requirements for conducting a cross-cultural study (Ghauri & Grønhaug, 2005). First, due to the risk of cultural accommodation when using English-language

Brand Image Perceptions across Cultures

183

questionnaires in cross-national surveys (Harzing, Salciuviene, & 31 countries collaborators, 2005) and in order to avoid misinterpretation of some items, an original English version of the questionnaire was translated into both Lithuanian and Danish. The questionnaire was back translated by bilingual translators (Brislin, Lonner, & Thorndike, 1973). Second, in order to detect possible discrepancies in meaning (Nijssen & Douglas, 2004; Salzberger & Sinkovics, 2006) and ensure comprehension and conceptual equivalence (Craig & Douglas, 2000), all measurement items were carefully inspected by bilingual translators. Finally, a pre-test of the questionnaires was conducted with small groups of Lithuanian- and Danish-speaking respondents in both countries. Constructs were measured on 5-Likert scales ranging from 1 (completely disagree/completely unimportant) to 5 (completely agree/very important). A consumer personal values scale was derived from Lages and Fernandes (2005), and respondents were asked to indicate what use of their brand allows them to achieve in their lives. The items for the perceived brand image beneﬁt scale were derived from Allen (2001), Allen and Ng (1999), Bhat and Reddy (1998), Martinez and de Chernatony (2004) and Sweeney and Soutar (2001). Respondents were asked to indicate the extent of importance they place on items pertaining to the functional and symbolic beneﬁt of their own brand of mobile phone (Table 2). Our measures were puriﬁed by means of a series of conﬁrmatory factor models (Gerbing & Anderson, 1998). Our initial analysis indicated that all items loaded signiﬁcantly for the Danish sample. Few items did not load signiﬁcantly for the Lithuanian sample. To continue the analysis, these items were eliminated. The results of conﬁrmatory factor analysis for the ﬁnal scales indicated a good model ﬁt in both countries. All items had high loadings, and they all were statistically signiﬁcant (po 0.05). The chi-square, goodness of ﬁt index and root mean square error of approximation (RMSEA) indicated satisfactory ﬁt for both countries. Cronbach’s alpha coefﬁcients were acceptable for both countries and ranged from 0.74 to 0.76. Measurement invariance was tested with conﬁrmatory factor analysis using multi-group analysis. The results indicate that the same factor structure is appropriate for either sample. Thus, conﬁgural invariance is established for the Danish–Lithuanian comparison (RMSEA ¼ 0.031; Bentler’s comparative ﬁt index [CFI] ¼ 0.96). As expected, full metric invariance was not established in this study (Reardon, Miller, Foubert, Vida, & Rybina, 2006; Steenkamp & Baumgartner, 1998). As suggested by Stain, Lee, and Jones (2006), ‘‘in practice, full invariance of all parameters

184

LAURA SALCIUVIENE ET AL.

Table 2.

Operationalisation of the Variables.

Variables/Items

Source

Consumer personal value scale The use of my brand allows me to achieve More harmony and stability in life Greater family security More tranquility More respect from others The feeling that the world is more agreeable More social recognition Higher status An inspirational, exciting and adventurous life A higher integration into my group Better relationships (e.g. social and family) Stengthened friendship

Lages and Fernandes (2005)

Perceptions of functional and symbolic brand image attributes How important are each of the following aspects with regard to a brand of the mobile phone? I think that this brand gives me an image of someone important This brand changes my image in other people’s eyes This brand reﬂects my success in life My friends choose it Ability to send SMS Technical functions It is a good phone for a good price

Adapted from Allen (2001), Allen and Ng (1999), Bhat and Reddy (1998), Martinez and de Chernatony (2004), Sweeney and Soutar (2001)

Note: The table summarises items that were retained in the ﬁnal scale. Deleted items are not included.

and even invariance of only key parameters is difﬁcult to obtain’’ (p. 251). According to Horn (1991, p. 125), metric invariance is ‘‘a condition to be striven for, not one expected to be fully realized’’ (in Reardon et al., 2006, p. 125). Reardon et al. (2006) consider ‘‘full metric invariance as scientiﬁcally unrealistic’’ (p. 125). Since this study does not ‘‘compare means of measures across countries, scalar invariance was not tested’’ (Reardon et al., 2006, p. 125; Steenkamp & Baumgartner, 1998, p. 80). The study’s aim is not to compare absolute perceptions of brand image attributes in different countries; therefore, differences in stylistic responding (Reardon et al., 2006; Baumgartner & Steenkamp, 2001) were also not assessed. Finally, multicolinearity was assessed, and the results satisfactorily indicated

Brand Image Perceptions across Cultures

185

that multicolinearity is not an issue here (variance-inﬂation factor [VIF] o5, as suggested by Hair, Black, Babin, Anderson, & Tatham, 2006).

ANALYSIS AND FINDINGS The effects of personal values on brand image perceptions were tested using regression analysis. For the Danish sample, the results indicate that the peaceful life value dimension has a signiﬁcant effect on perceptions of both symbolic attributes (b ¼ 0.35, p o 0.05) and functional attributes (b ¼ 0.20, p o 0.05). This dimension explains 43% of variation in perceptions of both symbolic and functional attributes. This was not so for the Lithuanian sample. For the Lithuanian sample, the results indicate that the peaceful life value dimension has a signiﬁcant effect on perceptions of symbolic attributes (b ¼ 0.21, p o 0.05) only. The total variance explained is 36%. The other two dimensions of personal values (social recognition and social integration personal values) have no signiﬁcant effects on brand image attributes in each country. These values do not inﬂuence brand image perceptions possibly because brands of mobile phones do not aid towards social recognition or integration as other more expensive high involvement shopping goods might do (e.g., BMW automobiles). We also aimed to assess how brand image was perceived and whether the symbolic and functional attributes of brand image were truly perceived by respondents as distinct. Testing these perceptions values were not taken into consideration. As a result, t tests (p o 0.05) on the factor scores indicate signiﬁcant differences between the perceptions of symbolic and functional brand image attributes of mobile telephones in each country. Additionally, for the Lithuanian sample, the symbolic brand image attributes of mobile phones are highly important, whereas functional attributes are moderately important. For the Danish sample, respondents report that they ﬁnd symbolic attributes more important than functional ones.

DISCUSSION OF FINDINGS Despite the stream of research of brand management studies, consumer brand image perceptions that are a function of consumer personal values have received insufﬁcient attention in the marketing literature. In addressing this shortcoming, we investigated the effect of consumer values on

186

LAURA SALCIUVIENE ET AL.

perceptions of brand image attributes in a context of two EU countries: one Eastern European-Lithuania and another Western European-Denmark. Our ﬁndings indicate that speciﬁc consumer personal values inﬂuence brand image perceptions in the market of mobile phones. Our ﬁndings suggest that the functional and symbolic attributes represent brand images of mobile telephones. The symbolic attribute explains consumers’ perceptions as regards intangible and symbolic beneﬁts offered by brand of mobile phones (e.g., status, role in a group), whereas the functional attribute explains consumers’ functional evaluations of brand image of mobile phones (e.g., practicality, usefulness). In respect to the symbolic brand image attributes, the results indicate that respondents in each country score higher on symbolism without taking into consideration personal values. A high score on the symbolic attribute indicates a preference for recommendations by friends regarding what to buy and fashionable, expensive brands; these should also correspond to their selfimage. Although the functional brand image attributes are important in each sample, the respondents score lower on this dimension compared with the symbolic one. A lower score on the functional attribute indicates that a choice of practical and useful products is of secondary role. Although our ﬁnding of two distinct brand image attributes was expected, it is consistent with the notion that brand image comprises both symbolic and functional attributes (Allen & Ng, 1999; Kivetz & Simonson, 2002). Our results also reveal a signiﬁcant effect of consumer values on perceptions of brand image attributes. On the one hand, Lithuanian students favouring peaceful life values over social recognition and social integration values expect brand image to convey the symbolic beneﬁts. On the other hand, Danish students favouring peaceful life values over social recognition and social integration values expect brand image to convey both the symbolic and the functional beneﬁts. The variations in effects indicate that brand image perceptions vary among respondents. These ﬁndings are in line with the notion that personal values play an important role in consumer perceptions and provide a deeper understanding of consumers (Kim et al., 2002; Marandi et al., 2006). In respect to the effect of personal values in brand image perceptions, we draw two conclusions on the basis of our ﬁndings. First, although the literature suggests diverse conceptualisations of brand image, our ﬁndings are similar to those of Bhat and Reddy (1998), Kivetz and Simonson (2002) and Martinez and de Chernatony (2004) that brand image consists of two brand image attributes and the symbolic and functional attributes of the brand may represent images of mobile telephone brands. Second, exploring

Brand Image Perceptions across Cultures

187

the conceptualisations of personal values, our results support the notion that personal values inﬂuence perceptions (Brangule-Vlagsma et al., 2002; Honkanen et al., 2006; Marandi et al., 2006). An assessment of how speciﬁc values affect brand image perceptions of mobile telephone brands provides an indication of the extent to which those perceptions differ among respondents. Through the understanding of how brand image is perceived from the consumers’ perspective, companies can build a relevant brand image that will convey speciﬁc beneﬁts offered by brands. Such knowledge will aid the process of strategic decision-making regarding the adaptation or standardisation of brand image and its maintenance in a cross-cultural context. Hence, the results of our study provide a more informed and systematic basis on which to develop brand image positioning strategy by providing a framework for selecting relevant brand image attributes.

MANAGERIAL IMPLICATIONS AND FURTHER RESEARCH SUGGESTIONS The relative impact of personal values on brand image perception can be of interest to brand managers, as the consideration of these factors can lead to more successful brand image formation strategies for the Lithuanian and Danish markets. Our results suggest that brands of mobile phones can be positioned in terms of the functional-symbolic reciprocity in view of consumer values for Danish market, whereas the symbolic positioning in view of consumer values is suggested for Lithuanian market. Understanding the issue that symbolic and functional attributes are perceived as multidimensional facets of brand image and that different attributes are inﬂuenced by speciﬁc consumer values may point to those variables that brand managers need to inﬂuence when managing brand image in the long run and developing more effective brand positioning strategies that are culturally adapted. Although it is difﬁcult to predict all relationships between consumer values and perceptions of brand image attributes, to the extent that this has been done in our study, we have demonstrated that it is worth taking into consideration the application of other strategies than those of standardisation only, even in such small markets as Lithuania or Denmark. Value-based consumer segmentation can be an option. Another point is that three-attribute brand image management may not always be necessary (Park et al., 1986), as

188

LAURA SALCIUVIENE ET AL.

our study has reported that only two attributes are associated with the brand image of mobile telephones. If companies continually extend the geographical area of their business activities, the degree of acceptance by the local consumers should be one of the major determinants of the success or failure of their operations. Therefore, the decision by companies as to whether they standardise, and thus neglect cognitive characteristics of local consumers, or adapt brand image locally is a central strategic one (Hsieh et al., 2006). Our study has several limitations that suggest future research directions. First, our study focuses on well-known brands in one product category of high-involvement shopping goods. Thus, caution is warranted when attempting to generalise the ﬁndings to less familiar brands for different product categories. Next, our study was conducted in two countries and the targeted respondents were students. Thus, young people in the sample might bias the ﬁndings, as the older generation might hold different values that inﬂuence their perceptions of brand image attributes in different ways, especially in the region that was the Soviet Union. Finally, our study does not include a sender– receiver communication model, or a consideration of advertising design. Thus, further research would provide us with additional insights into the ﬁeld. First, research needs to be extended to other Eastern European countries in order to generalise the results to these markets. Such a study would offer a deeper insight into the consumers’ perceptions and personal values in new countries that are joining the European Union. Second, our study includes relatively small and homogeneous countries; thus, future studies should include countries with signiﬁcant ethnic minorities. Third, other studies could focus on other types of products and brands with which consumers are less familiar. Finally, cross-cultural studies that take into account consumer preferred communication modes or consumer response to advertising campaigns would also add to furthering our understanding of consumer brand image perceptions.

REFERENCES Allen, M. W. (2001). A practical method for uncovering the direct and indirect relationships between human values and consumer purchase. Journal of Consumer Marketing, 18(2), 102–120. Allen, M. W., & Ng, S. H. (1999). The direct and indirect inﬂuences of human values on product ownership. Journal of Economic Psychology, 20(1), 5–39. Arnould, E., Price, L., & Zinkhan, G. (2004). Consumers (2nd ed.). Boston, MA: McGraw-Hill/ Irwin.

Brand Image Perceptions across Cultures

189

Batra, R., & Homer, P. M. (2004). The situational impact of brand image beliefs. Journal of Consumer Psychology, 14(3), 318–330. Baumgartner, H., & Steenkamp, J.-B. E. M. (2001). Response styles in marketing research: A cross-national investigation. Journal of Marketing Research, 38, 143–156. Bhat, S., & Reddy, S. K. (1998). Symbolic and functional positioning of brands. Journal of Consumer Marketing, 15(1), 32–43. Brangule-Vlagsma, K., Pieters, R. G. M., & Wedel, M. (2002). The dynamics of value segments: Modelling framework and empirical illustration. International Journal of Research in Marketing, 19(3), 267–285. Brislin, R., Lonner, W., & Thorndike, R. (1973). Cross-cultural research methods. New York, NY: Wiley. Burgess, S. M., & Steenkamp, J.-B. E. M. (2006). Marketing renaissance: How research in emerged markets advances marketing science and practice. International Journal of Research in Marketing, 23, 337–356. Craig, S., & Douglas, S. P. (2006). Beyond national culture: Implications of cultural dynamics for consumer research. International Marketing Review, 23(3), 322–342. Craig, S. C., & Douglas, S. P. (2000). International marketing research (2nd ed.). West Sussex: Wiley. Cunningham, L. F., Young, C. E., Lee, M., & Ulaga, W. (2006). Customer perceptions of service dimensions: Cross-cultural analysis and perspective. International Marketing Review, 23(2), 192–210. d’Andrade, R. G. (1986). Cultural meaning system. In: R. A. Shweder & R. A. Levine (Eds), Culture theory: Essays on mind, self, and emotion (pp. 88–122). New York, NY: Cambridge University Press. de Chernatony, L. (2002). From brand vision to brand evaluation. Oxford: ButterworthHeinemann. del Rio, A. B., Vazquez, R., & Iglesias, V. (2001). The effects of brand associations on consumer response. The Journal of Consumer Marketing, 18(5), 410–425. Dhar, R., & Wertenbroch, K. (2000). Consumer choice between hedonic and utilitarian goods. Journal of Marketing Research, 37(1), 60–71. Dobni, D., & Zinkhan, G. M. (1990). In search of brand image: A foundation analysis. Advances in Consumer Research, 17, 110–119. Dunning, J. B. (1993). Multinational enterprises and the global economy. Reading, MA: Addison-Wesley. Firat, F. A. (1991). The consumer in postmodernity. Advances in Consumer Research, 22, 239–267. Fulton, D. C., & Lipscomb, J. (1996). Wildlife value orientations: A conceptual and measurement approach. Human Dimensions of Wildlife, 1(2), 24–47. Gerbing, D., & Anderson, J. (1998). An updated paradigm for scale development incorporating unidimensionality and its assessment. Journal of Marketing Research, 25, 186–192. Ghauri, N. P., & Grønhaug, K. (2005). Research methods in business studies (3rd ed.). Harlow: Prentice Hall. Gullestrup, H. (2006). Cultural analysis – towards cross-cultural understanding. Aalborg: Aalborg University Press. Hair, J. F., Jr., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis with readings (6th ed.). New Jersey: Prentice-Hall International, Inc.

190

LAURA SALCIUVIENE ET AL.

Harzing, A. W., & Salciuviene, L., & 31 countries collaborators. (2005). Does the use of English language questionnaires in cross-national research obscure national differences? International Journal of Cross-cultural Management, 5(2), 213–224. Herzog, H. (1963). Behavioural science concept for analyzing the consumers. In: P. Bliss (Ed.), Marketing and the behavioural sciences (pp. 76–86). Boston, MA: Allyn and Bacon Inc. Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviours, institutions, and organizations across nations (2nd ed.). London: Sage. Hofstede, F. T., Steenkamp, J.-B. E. M., & Wedel, M. (1999). International market segmentation based on consumer-product relations. Journal of Marketing Research, 36, 1–17. Honkanen, P., Verplanken, B., & Olsen, S. O. (2006). Ethical values and motives driving organic food choice. Journal of Customer Behaviour, 5(5), 420–430. Horn, J. (1991). Comments on issues on factorial invariance. In: L. Collins & J. Horn (Eds), Best methods for the analysis of change (pp. 114–125). Washington, DC: American Psychological Association. Hsieh, M. H. (2002). Identifying brand image dimensionality and measuring the degree of brand globalization: A cross-national study. Journal of International Marketing, 10(2), 46–67. Hsieh, M.-H., Pan, Sh.-L., & Setiono, R. (2006). Product-, corporate-, and country-image dimensions and purchase behavior: A multicounty analysis. Academy of Marketing Science, 32(3), 251–270. Keller, K. L. (1998). Strategic brand management. Upper Saddle River, NJ: Prentice Hall. Keller, K. L. (2003). Building, measuring, and managing brand equity (2nd ed.). Upper Saddle River, NJ: Prentice Hall. Kim, H. M., & Kramer, T. (2006). ‘‘Pay 80%’’ versus ‘‘get 20% off ’’: The effect of novel discount presentation on consumers deal perceptions. Marketing Letters, 17(4), 311–321. Kim, J.-O., Forsythe, S., Gu, Q., & Moon, S. J. (2002). Cross-cultural consumer values, needs and purchase behaviour. Journal of Consumer Marketing, 19(6), 481–502. Kivetz, R., & Simonson, I. (2002). Earning the right to indulge: Effort as a determinant of customer preferences towards frequency program reward. Journal of Marketing Research, 39(2), 155–170. Kumar, V. (2000). International marketing research. Upper Saddle River, NJ: Prentice Hall. Lages, L. F., & Fernandes, J. C. (2005). The SERPVAL scale: A multi-item instrument for measuring service personal values. Journal of Business Research, 58(11), 1562–1572. Marandi, E., Little, E., & Sekhon, Y. (2006). The Impact of personal values on perception of service provider empathy and customer loyalty. The Business Review, 5(2), 339–343. Martinez, E., & de Chernatony, L. (2004). The effect of brand extension strategies upon brand image. Journal of Consumer Marketing, 21(1), 39–50. Maslow, A. H. (1959). New knowledge in human values. New York: Harper. Nijssen, E. J., & Douglas, S. P. (2004). Examining the animosity model in a country with a high level of foreign trade. International Journal of Research in Marketing, 21(1), 23–38. Park, W. C., Jaworski, B. J., & MacInnis, D. J. (1986). Strategic brand concept image management. Journal of Marketing, 50(4), 135–145. Patterson, P. G., & Smith, T. (2003). A cross-cultural study of switching barriers and propensity to stay with service providers. Journal of Retailing, 79, 107–120. Reardon, J., Miller, Ch., Foubert, B., Vida, I., & Rybina, L. (2006). Antismoking messages for the international teenage segment: The effectiveness of message valence and intensity across different culture. Journal of International Marketing, 14(3), 115–138.

Brand Image Perceptions across Cultures

191

Reisinger, Y., & Turner, L. W. (2003). Cross-cultural behaviour in tourism: Concepts and analysis. Oxford: Elsevier Science Limited. Reynolds, T. J., & Gutman, J. (1984). Advertising is image management. Journal of Advertising Research, 24, 27–37. Reynolds, N. L., Simintiras, A. C., & Diamantopoulos, A. (2003). Theoretical justiﬁcation of sampling choices in international marketing research: Key issues and guidelines for researchers. Journal of International Business Studies, 34(1), 80–89. Ricoeur, P. (1969). The symbolism of evil. Boston, MA: Beacon Press. Rokeach, M. J. (1973). The nature of human values. New York: Free Press. Roth, M. S. (1992). Depth versus breath strategies for global brand image management. Journal of Advertising, 21(2), 25–36. Salzberger, T., & Sinkovics, R. R. (2006). Reconsidering the problem of data equivalence in international marketing research. International Marketing Review, 23(4), 390–417. Schwartz, S. H., & Bilsky, W. (1987). Toward a universal psychological structure of human values. Journal of Personality and Social Psychology, 53, 550–562. Sekaran, U. (1983). Methodological and theoretical issues and advancements in cross-cultural research. Journal of International Business Studies, 14(Fall), 61–73. Stain, J. A., Lee, J. W., & Jones, P. S. (2006). Assessing cross-cultural differences through use of multi-group invariance analysis. Journal of Personality Assessment, 87(3), 249–258. Steenkamp, J. E. M., & Baumgartner, H. (1998). Assessing measurement invariance in crossnational consumer research. Journal of Consumer Research, 25, 78–90. Shintaro, O., Charles, R. T., & Jonathan, P. D. (2007). Market convergence and advertising standardization in the European Union. Journal of World Business, 42, 384. Sinkovics, R., Penz, E., & Ghauri, P. (2005). Analyzing textual data in international marketing research. Qualitative Market Research: An International Journal, 8(1), 9–38. Sweeney, J. C., & Soutar, G. N. (2001). Consumer perceived value: The development of a multiple item scale. Journal of Retailing, 77(2), 203–220. Thompson, C. J., Rindﬂeisch, A., & Arsel, Z. (2006). Emotional branding and the strategic value of the Doppelga¨nger brand image. Journal of Marketing, 70(January), 50–64. Van de Vijver, F., & Leung, K. (1997). Methods and data analysis for cross-cultural research. London: Sage. van Reijmersdal, E. A., Neijens, P. C., & Smith, E. G. (2007). Effects of television brand placement on brand image. Psychology & Marketing, 24(5), 403–420. van Rekom, J., Jacobs, G., & Verlegh, P. W. J. (2006). Measuring and managing the essence of a brand personality. Marketing Letters, 17(3), 181–192. Vida, I., & Reardon, J. (2008). Domestic consumption: Rational, affective or normative choice? Journal of Consumer Marketing, 25(1), 34–44. Vinson, D. E., Scott, J. E., & Lamont, L. M. (1977). The role of personal values in marketing and consumer behaviour. Journal of Marketing, 42(April), 44–50. Yaprak, A. (2007). Culture study and international marketing: A literature review and suggestions for future research. In: R. R. Sinkovics & P. Ghauri (Eds), Proceedings of the CIMaR Conference, Manchester, United Kingdom. Zaltman, G. (2003). How customers think: Essential insights into the mind of the market. Cambridge, MA: Harvard University Press. Zetterberg, H. L. (1998). Cultural values in market and opinion research. In: C. McDonald & P. Vangelder (Eds), ESOMAR handbook of marketing and opinion research (pp. 995–1013). The Netherlands: Amsterdam.

PART III COLLABORATIVE RELATIONSHIPS

DO FIRMS CREATE VALUE THROUGH INTERNATIONAL STRATEGIC ALLIANCES? Chia-Ling ‘Eunice’ Liu In recent years, interﬁrm collaborations have inspired an abundant literature in marketing and strategy (Lambe, Spekman, & Hunt, 2002; Sarkar, Raj, Cavusgil, & Aulakh, 2001; Susan, Ron, & Bill, 2006). Alliances, collaborative arrangements between ﬁrms that comprise the exchange, sharing and codevelopment of products, services and knowledge, are now eagerly sought as a means of adding to ﬁrm’s value and integral part of the strategy (Contractor & Lorange, 2002a; Kandemir, Yaprak, & Cavusgil, 2006; Lorange & Roos, 1992). Moreover, these alliances are more and more knowledge-intensive and often involve competing ﬁrms (Duysters, Kok, & Vaandrager, 1999). Many researchers have suggested that an important explanatory factor for the growing trend toward alliances is that these provide a platform for organizational learning, giving ﬁrms access to the knowledge and resources of their partners (Chen & Chen, 2003; Inkpen, 2002; Inkpen & Beamish, 1997; Mol, van Tulder, & Beije, 2005; Norman, 2004; Weerawardena, O’Cass, & Julian, 2006). Small ﬁrms in a resource-limited country have a limited capacity to compete in the knowledge-intensive and highly globalized industries. However, the experience of Taiwan tells a different story. Taiwan has become the world’s biggest producer of notebook PCs and has established itself as a world-class supply source for various computer-related products, key components and New Challenges to International Marketing Advances in International Marketing, Volume 20, 195–225 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020011

195

196

CHIA-LING ‘EUNICE’ LIU

knowledge-intensive services (Einhorn & Burrows, 2003; Ernst, 2000). The size-related disadvantages have not prevented Taiwan from becoming a successful competitor in the IT industry, which requires a wide range of technological and organizational capabilities. Ernst (2000) argues that the answer to this puzzle is knowledge acquisition through various interorganizational linkages. As markets today have grown more and more fragmented, marketers related to each other increasingly in the roles of suppliers, customers, collaborators and competitors (Kandemir et al., 2006). The issue of how ﬁrms form meaningful collaborative relationship, how they maintain partnerships along the value chain and how they manage knowledge ﬂows in such environments has become critical marketing skills in achieving better performance (Day & Montogomery, 1999). This research focuses on examining how these Taiwanese IT suppliers enhance their overall capabilities through cooperating with multinational corporation (MNC) buyers, which may shed light on the complicated issues of cross-border alliances. This research examines the factors inﬂuencing learning between crossborder alliance partners. The management and processing of knowledge among organizations are increasingly viewed as important determinants of innovations and the success of alliances (Frost & Zhou, 2005). I build on antecedents to knowledge acquisition, learning dimensions such as learning intent, absorptive capacity and protection on the one hand, and relational dimensions such as trust and interaction on the other. First, most of the literature on learning has restricted itself to learning within the organization. There is less understanding about how the alliances trigger learning when organizations collaborate and interact (Nonaka, Von Krogh, & Voelpel, 2006; Phan & Peridis, 2000). It has become more and more critical, especially in high-tech industries, to collaborate with other organizations to advance in new technologies and innovations (Frost & Zhou, 2005; Kandemir et al., 2006). Accordingly, this research develops a more comprehensive and integrative model to link inter-organizational and intra-organizational learning to investigate the determinants of the capability enhancement through cross-border alliances. Second, although most researchers have focused on alliance motivations, forms of collaborations and outcomes (Kandemir et al., 2006; Zineldin & Dodourova, 2005), process-related questions have received less attention. This research addresses determinants of the knowledge acquisition process by recognizing the role of inter-partner relationships. The framework for accomplishing this involves integrating the relational view, which emphasizes the inter-organizational interactions and mutual trust between partners (Cullen & Johnson, 2000; Dyer & Singh, 1998).

Do Firms Create Value through International Strategic Alliances?

197

Finally, conﬂicting results have been obtained for alliances in developing and developed countries (Ramamurti, 2004). Most research on developing countries adopts the perspective of the MNCs and excludes the view of the developing country’s partner (De Mattos, Sanderson, & Ghauri, 2002; Ernst, 2000; Yan & Gray, 1994). In recent years, the vertical link between the MNC buyers and their offshore suppliers in less-developed countries/ newly industrialized countries (LDCs/NICs) is particularly strong, but there has been little systematic research into this emerging relationship (Chen & Chen, 2003; Jao, 1996; Mol et al., 2005). Hence, I focus on the perspective of the resource-limited suppliers in NICs and examine whether they can enhance their capabilities and obtain competitive advantages through alliance learning. On the basis of the above rationale, this research considers the following questions: What factors inﬂuence knowledge acquisition between crossborder alliance partners? Do inter-ﬁrm relationships inﬂuence knowledge acquisition in alliances? From the suppliers’ perspective, can alliance learning enhance the capabilities of the ﬁrm in a quasi-market, and asymmetric alliances such as contract manufacturing alliances? Can the ﬁrm enhance its network position after alliance learning? I develop a theoretical model, build hypotheses and continue to test the model using structural equations modeling. I end by discussing the results and suggesting the directions for future research.

THEORETICAL MODEL Antecedents of Knowledge Acquisition Learning Intent Intent refers to a ﬁrm’s initial propensity to view collaboration as an opportunity to learn (Hamel, 1991, p. 90). Comparing the intent to form alliances between Western and Japanese ﬁrms, Hamel indicates that most Western ﬁrms possess substitution intent to substitute their competitiveness in a speciﬁc area for their own lack of skills, whereas the Japanese partners seem to have explicit learning intent to actually internalize their partners’ skills. When the internalization intent is strong in a company, the skills and knowledge acquired from the partner are important to the growth of the whole company (Hamel, 1991). However, if both partners possess equal intent to internalize the other’s skill, distrust and conﬂict may occur to threaten the stability of alliances (Hamel, Doz, & Prahalad, 1989; Madhok, 2006).

198

CHIA-LING ‘EUNICE’ LIU

Hamel (1991, p. 92) proposes the factors that might account for the observed differences in intent: (1) whether the ﬁrm views collaboration as a more or less permanent alternative to competition or as a temporary vehicle for improving its competitiveness vis-a`-vis its partners; (2) its relative resource position as compared to its partners and other industry participants’ position; (3) its calculation of the pay-off from learning; and (4) its preference for symmetric vs. asymmetric dependence within the alliance (Kemp & Ghauri, 2001). Many scholars agree that the organization must consider learning as an explicit objective, and collective learning in the organization will enhance the ﬁrm’s capability to survive (Andersson, Bjorkman, & Forsgren, 2005; Ghoshal, 1987; Hamel, 1991; Norman, 2004). Even though there may be both intentional and unintentional knowledge acquisition in an alliance process, studies have consistently found a positive correlation between ‘intent’ and ‘organizational learning’ (Kandemir, Ghauri, & Cavusgil, 2002; Schacht, 1999). Although strategic alliances are fundamentally considered as dyadic exchange relationships, alliance participation over time results in increasingly complex inter-ﬁrm networks (Ghauri & Prasad, 1995; Gulati, 1999; Zaheer & Bell, 2005). Embedded relationships with other partners in the network enable ﬁrms to acquire valuable information of new business opportunities, which can be particularly important in a global setting (Ghauri, Hadjikhani, & Johanson, 2005; Hakansson & Johanson, 2001). The explicit goal to learn from the partnership helps the ﬁrm to move quickly in a rich information exchange network. Nowadays, the ﬁrm’s position in the network is considered more critical because it channels knowledge transfer and technology sharing (Contractor, Kim, & Beldona, 2002; Contractor & Lorange, 2002b). Hence, I propose the following hypothesis: H1. Suppliers with higher learning intent will experience more knowledge acquisition from their foreign partners. Absorptive Capacity If intent establishes the desire to learn, receptivity determines the capacity to learn (Hamel, 1991, p. 96). In that context, absorptive capacity can be deﬁned as a ﬁrm’s ability to ‘recognize the value of new, external knowledge, assimilate and apply it to commercial ends’ (Cohen & Levinthal, 1990, p. 128). It is argued that a ﬁrm’s absorptive capacity is ‘largely a function of ﬁrm’s prior related knowledge’ (Cohen & Levinthal, 1990, p. 128). The organization’s absorptive capacity will start at the individual level; hence, it depends not only on the interface with the external environment but also on the knowledge

Do Firms Create Value through International Strategic Alliances?

199

transfer across and within subunits (Scott-Kennel & Enderwick, 2004). The ﬁrm’s absorptive capacity results from a long process of investments in knowledge accumulation and is inﬂuenced by its participation in product development, research and development (R&D) and other technological activities (Frost & Zhou, 2005; Mowery, Oxlley, & Silverman, 2002). Without absorptive capacity, it will be difﬁcult to learn from their partners and add value to the ﬁrm. Lane and Lubatkin (1998) propose that the learning partner’s absorptive capacity will depend on (1) what type of new knowledge is offered by the teaching partner; (2) the similarity between the student and teacher ﬁrm’s capability and organization structures and (3) the learning partner’s familiarity with the teaching partner’s organizational problems. A supplier’s absorptive capacity is expected to be positively correlated with knowledge acquisition from its partners. Therefore I propose: H2. The higher the level of the supplier’s absorptive capacity, the greater the amount of knowledge acquisition from the foreign partners. Protection Transparency refers to the ‘knowability’ or openness of each partner and thus the potential for learning (Hamel, 1991, p. 90). Transparency can be determined from two aspects: organization and skills (Hamel, 1991). Norman (2004) claims that transparency could be inﬂuenced through the design of organizational interfaces, the attitudes toward outsiders, the partner’s relative pace of skill-building and the protective behavior of individuals. However, the challenge is to share enough knowledge to create competitive advantages while preventing the whole transfer of core skills (Baughn, Denekamp, Stevens, & Osborn, 1997; Hamel et al., 1989). The role of gatekeepers, i.e. people who control the knowledge ﬂow, is to protect the unintended and informal transfer of proprietary intellect assets. Szulanski (1996, p. 31) also identiﬁes the lack of motivation of the source of knowledge (e.g., protective behavior) as a likely source of internal stickiness (Szulanski, 1996; Wu & Cavusgil, 2006). If there is a high competitive overlap, the ﬁrms may be unwilling to share their knowledge due to the risk of knowledge loss (Inkpen, 1998). Knowledge loss arises when valuable knowledge spills out to the partner who then may use the knowledge to gain competitive advantage (Cohen & Levinthal, 1990). Von Hippel (1994) indicates that certain attributes inﬂuence knowledge stickiness, e.g., specialized personnel such as technology gatekeepers, specialized organization structures such as transfer groups, the pricing and the access to proprietary information. These attributes are indicators of the degree of the

200

CHIA-LING ‘EUNICE’ LIU

protection requested by information transmitters vis-a`-vis their knowledge base (Simonin, 1999a, 2004). The foreign partners may put barriers to limit local partners to access their speciﬁc skill area (Inkpen & Beamish, 1997; Simonin, 2004). The loss of knowledge can expose them to the threat of a new and stronger competitor. The fear of losing competitive advantage and proprietary assets in an alliance will impede knowledge acquisition. The partner’s protective behavior will hinder knowledge acquisition. Hence H3. The more the protection by the foreign partner, the less the knowledge acquisition by the local partner. Trust Trust reﬂects the belief that a partner’s word or promise is reliable and that a partner will fulﬁll its obligations in the relationship (Boersma, Buckley, & Ghauri, 2003; Inkpen, 2000). Many researchers have indicated the important role of trust in the performance of a collaborative relationship (Dhanaraj, Lyles, Steensma, & Tihanyi, 2004; Inkpen, 2000; Schoorman, Mayer, & Davis, 2007). However, most of these studies focus on interpersonal trust; the way in which close relationships among managers, scientist and engineers develop and little attention is paid to interorganizational trust (Dodgson, 1993). One of the most ambitious global research projects on the next generation of computer chips united by over 200 scientists from Siemens AG of Germany, Toshiba of Japan and IBM of the US provides a good example of the signiﬁcance of organizational trust in alliances (Browning, 1994). In principle, diversity in the cultural and scientiﬁc background is supposed to stimulate the creation of new technology. Nevertheless, language barriers, the lack of cross-culture training and the differences in management philosophy may result in mistrust of partners and terminate the alliance (Svejenova, 2006). Research has shown that trust and personal ties are more important than explicit contracts in the network relationship, and these features reduce monitoring costs and enhance ‘thick’ information exchange of tacit and proprietary know-how (Perry, Sengupta, & Krapfel, 2004; Smitka, 1991; Uzzi, 1997). Trust promotes the access to difﬁcult-to-price resources that enrich the ﬁrm’s ability to compete and overcome problems (Boersma et al., 2003; Ghauri et al., 2005; Uzzi, 1997). High-level trust will eliminate unnecessary safeguarding mechanism and is expected to have a positive impact on knowledge acquisition. Hence H4. The greater the trust between the alliance partners, the greater the effect of knowledge acquisition.

Do Firms Create Value through International Strategic Alliances?

201

Interaction The level of knowledge acquisition will be enhanced if the partners get to know where the important expertise resides within each ﬁrm. In order to achieve this, the alliance partners can design interﬁrm routines that facilitate information-sharing and increase socio-technical interactions over time (Dyer & Singh, 1998; Eriksson & Chetty, 2003). The interaction between different value-added functions will encourage productivity and innovation (Chetty & Wilson, 2003; Jao, 1996). Various interaction modes such as on-site visits, product concept reviews, technical meetings and joint training programs are encouraged to improve the quality of relationships and facilitate knowledge acquisition. The transfer of tacit knowledge is very communication-intensive, involving several months of frequent interactions between alliance partners. Subramaniam and Venkatraman (2001) also indicate that rich informationprocessing mechanisms such as face-to-face contact with overseas managers will increase the effectiveness of transferring and deploying tacit overseas knowledge in MNCs. Therefore, the level of partner interaction is expected to be positively related with knowledge acquisition by the alliance partners. If suppliers interact continuously and closely with their MNC buyers, they obtain better intelligence about the buyer’s requirements and the competitors’ moves. The interaction of external activities and technologies becomes important in sustaining strategic advantages in global competitive battles in high-technology industries (Teece, Pisano, & Shuen, 1997). MNC buyers nowadays increasingly involve suppliers in the design process and technology development activities such as joint product development and joint technology sharing (Kandemir et al., 2002; Mol et al., 2005). These joint actions bind suppliers to MNC buyers. Suppliers learn about the buyers’ requests for products, culture, management practices and decisionmaking patterns, which enable them to adjust and apply their resources in the ways that have the greatest beneﬁt (Vonderembse & Tracey, 1999). Relational capital not only facilitates learning through one-to-one interaction but also helps companies to balance the acquisition of new capabilities with the protection of proprietary assets between alliance partners (Ghauri et al., 2005; Kale, Singh, & Perlmutter, 2000). Hence, I propose H5. The level of partner interaction is positively related to knowledge acquisition from foreign partners. Knowledge Acquisition and Knowledge Application Knowledge allows the ﬁrms to develop their core competences from collective learning and enhance their ability to survive and grow (Hamel, 1991).

202

CHIA-LING ‘EUNICE’ LIU

Many studies have been concerned with knowledge acquisition but less is known about the process of knowledge application – the stage in which knowledge becomes institutionally available (Nevis, DiBella, & Gould, 1995; Rindﬂeisch & Moorman, 2001). The distinction between these two processes is crystallized in Demsetz’s (1993) observation. Knowledge acquisition requires that individuals are specialized in speciﬁc areas of knowledge, but knowledge application requires the combination of many areas of specialized knowledge. Learning may take place in planned or unintended ways. In some cases, knowledge acquired from foreign partners may exist only in certain functional teams such as R&D and product management. If knowledge is only acquired, but not processed and internalized in the organization, then its effect in enhancing a core competence is limited (Helleloid & Simonin, 1994; Kandemir et al., 2002). The ﬁrm needs to ensure that the routines continue to be carried out to leverage the learning of the individual members (Crossan & Berdrow, 2003; Miller, Fern, & Cardinal, 2007). A broader scope of organizational learning includes knowledge acquisition, information distribution, interpretation, and institutionalization (Crossan, Lane, & White, 1999). These processes should be integrated to support ‘actionable learning system’ to link the individual, group and organizational levels. Learning is not something that occurs in an organization only through ‘acquisition effort’ (Grant & Baden-Fuller, 2002; Nevis et al., 1995) but has to be looked at through all the phases simultaneously. Therefore, I suggest the following hypothesis: H6a. The greater the knowledge acquisition from the foreign partners, the greater the knowledge application by local partners. Knowledge Acquisition and Capability Enhancement It has been widely recognized that learning through alliances is the basis of a ﬁrm’s value creation (Helleloid & Simonin, 1994; Inkpen, 2002). Even though learning has been identiﬁed as an important motivation to form alliances in much research, few studies have seriously investigated the relationship between learning and alliance outcomes (Inkpen, 2002; Norman, 2004). If ﬁrms do beneﬁt from alliance learning, how will the learning enhance their capability? How will a ﬁrm achieve ‘competitive advantages’ through inter-ﬁrm learning? There are thus many challenges to assess the learning and its impact more precisely. One reason for these shortcomings in the literature is that there is a lack of availability of measures of ﬁrm-speciﬁc capabilities (Mowery, Oxley, & Silverman, 1996).

Do Firms Create Value through International Strategic Alliances?

203

In this research, I use the ‘enhancement of capability’ as the indicator of learning outcome. Those capabilities that are fundamental to a ﬁrm’s competitive advantage and difﬁcult to imitate by competitors – are the collective learning in the organization, especially about how to coordinate diverse production skills and integrate multiple streams of technologies (Prahalad & Hamel, 1990, p. 82). The ﬁrm’s ‘dynamic’ and ‘ﬂexible-response’ capabilities have been deemed critical in responding to various changing market and environmental conditions (Grant, 1996; Miller et al., 2007). Learning opportunities from various inter-organizational linkages with MNC partners enable suppliers in resource-limited countries to upgrade rapidly from relatively simple to increasingly complex forms of international arrangements such as product design and product innovation. Hence, I hypothesize that H6b. Suppliers with a higher level of knowledge acquisition will experience more capability enhancement. Knowledge Application and Capability Enhancement Knowledge acquired from foreign partners may not directly ﬁt the local ﬁrm (Lane, Salk, & Lyles, 2001). The ﬁrm needs to disseminate and adapt the knowledge learned from its foreign partners and apply it to organizational routines (Crossan et al., 1999; Kandemir et al., 2002). Huber (1991) asserts that ‘organizational memory’ is the means by which knowledge is stored for future use and is an important aspect of knowledge application. The organization members can quickly access knowledge through the use of an efﬁcient storing and retrieving information system (Gold, Malhotra, & Segars, 2001). The assimilation and utilization process refers to the stage at which knowledge becomes institutionally available, as opposed to being the assets of select individuals or groups (Nevis et al., 1995). It takes time to transfer knowledge from individuals to groups and from groups to the organization. The sharing and dissemination of personal experiences will facilitate the creation of organizational knowledge (Gold et al., 2001; Inkpen & Dinur, 1998). Institutionalizing is the process of embedding learning that has occurred through individuals and groups into the institutions of the organization including systems, structures, procedures and strategy (Crossan & Berdrow, 2003). The institutionalizing process aims to apply the learning systematically and enables the organizational routines to be continually carried out (Crossan et al., 1999). The use of in-house conferences, symposia and

204

CHIA-LING ‘EUNICE’ LIU

product seminars encourage knowledge sharing between different departments and individuals. The training and development of the organization’s staff has a positive impact on the application of knowledge. Although individuals may come and go, what they have learned does not necessarily leave with them. Hence, I hypothesize that H7. Suppliers with a higher level of knowledge application will experience greater capability enhancement. Capability Enhancement and Network Position Enhancement The network approach has contributed to the development of theories in business research and resulted in several publications concerning inter-ﬁrm relationships (Anderson & Dahlqvist, 2002; Dhanaraj & Parkhe, 2006; Ghauri, 1999; Hakansson & Johanson, 2001). Strategic issues typically resolved in a networking context include the following: turnkey contracts, make-lease-buy decisions, joint ventures, strategic alliances, mergers and acquisitions, vertical integration and some forms of outsourcing (Brass, Galaskiewicz, Greve, & Tasai, 2004). The position is a location of power to create and/or inﬂuence networks and is decided by three major factors: the domain of the company, the position of the company in other networks and the power of the company relative to the participants in the focal network (Thorelli, 1986). Prior research indicates that alliance networks can be considered as conduits of information, learning and knowledge (Goerzen & Beamish, 2005; Zaheer & Bell, 2005). From the analysis of Toyota production systems, Kogut (2000) shows how capabilities were transferred in the network and how the increased knowledge in the supplier network forced a radical disintegration of American auto assemblers. Globalization in the electronics and IT industry has culminated in the concept of ‘global production network’ that captures the spread of the broader systems of international production which cut across different stages of the value chain (Ernst & Guerrieri, 1998). The main purpose is to exploit complementarities, and the participation in these networks may enhance the migration of knowledge across ﬁrm boundaries and national borders (Ernst, 2000). Most IT companies participate in global production networks and greatly enhance their capabilities through repeated interactions and the acquisition of skills between ﬁrms in the networks. Over time, the suppliers accumulate knowledge and evolve toward greater independence because of an increase in the scope of their product design activities in the network. This dynamic and continuous learning leads to improvement in capabilities

Do Firms Create Value through International Strategic Alliances?

205

and enhancement of the ﬁrms’ positions in the network. Thus, I suggest the following hypothesis: H8. The level of capability enhancement is expected to be positively related with the enhancement of the supplier’s network position.

METHODS Sample and Data Collection The sampling list was obtained from various sources such as Taiwan Electronics and Appliance Manufacturers Associates (TEEMA), Taiwan Computer Association (TCA) and the Top 1000 Manufacturing Firms List. There are three main criteria for selecting the ﬁrms from these databases: (1) manufacturing ﬁrms with a capital of over US$3 million in 2003; (2) companies with experience of cross-border alliances and (3) secondary data of the companies are publicly accessible to obtain more detailed information. The content and format of the questionnaire were developed on the basis of the preliminary interviews and a thorough literature review, which was pretested using several on-site meetings with experienced alliance managers (Ghauri & Gronhaug, 2005). A mixed-mode strategy (Dillman, 2000) of combining e-mail and paper surveys was chosen, which increased the ﬂexibility and raised the response rate. The paper questionnaire was printed in a booklet that included a covering letter as its front page. A selfaddressed, pre-paid envelope was also included in the package to encourage the replies. Most of the electronic format questionnaires were sent to the people who agreed to answer by e-mail in the previous contact. Some companies with multiple alliance partners agreed to answer more than one questionnaire. The combined ﬁnal sample included 609 ﬁrms from six primary categories in the IT industry: communication products, systems, peripherals, card/board, semiconductors and components. I adopted two selection criteria for the key informants in this research. The ﬁrst criterion was the informant’s position to generalize ‘about patterns of behavior, after summarizing either observed or expected organizational relations’ (Seidler, 1974, p. 817). The second criterion was knowledge of collaborative phenomenon that respondents have been personally involved with alliances and were sufﬁciently knowledgeable to answer the questions. The majority of the respondents are top executives (i.e. presidents, CEOs,

206

CHIA-LING ‘EUNICE’ LIU

general managers, vice-presidents, directors and senior managers) and the others are functional or project managers.

Respondents and Alliance Proﬁles Throughout the three phases of follow-up, continuous effort and repeated contacts, 160 usable questionnaires were received; this sets the valid response rate at 26.3%. Of these 160 questionnaires, 97 (60.7%) were returned by post and the remaining 63 (39.3%) were sent back through e-mail. Much of the published research on inter-organizational relationships reports a lower response rate of around 15–20%. In terms of country of origin of the alliance partners, 58.8% were US ﬁrms, 17.5% were European ﬁrms, 16.9% were Japanese ﬁrms and the remaining 6.8% were ﬁrms based in other areas such as Korea, Israel, etc. In the review of empirical International Strategic Alliance (ISA) research, many researchers noted that the short lifespan of an ISA operation has become a limitation on sample generation. However, most of the ISA collaborations in this research have reached a satisfactory degree of stability. Over 70% of responding ﬁrms reported that their ISA operation duration was more than three years and 8.8% of the ﬁrms even maintained a long-term cooperative relationship for more than 10 years. Of the 160 responding ﬁrms, 125 (78.1%) ﬁrms not only manufacture but also design products for MNC buyers. They are usually labeled as original design manufacturing (ODM) suppliers. Most of Taiwanese IT suppliers already enhanced their capability in product design for the MNC buyers. Instead of only manufacturing or designing products for global buyers, 84 (52.5%) of the sample ﬁrms also market their own products under their own brand. Over 33% of the companies included in this research had a sales volume greater than US$300 million. Half the sample ﬁrms had a workforce larger than 1,000 employees. Following the procedure suggested by Armstrong and Overton (1977), I conducted two tests to examine the possibility of nonresponse bias. First, I compared the respondents’ characteristics (e.g., number of employees, annual sales revenue, capital and age of the company) to those of the original population sample. Second, I compared early with late responses. The ﬁrst 75% of the returned questionnaires were deﬁned as early responses and the last 25% were regarded as late responses. No signiﬁcant differences were found from comparing the means of the nine constructs in two groups. Therefore, I assume that nonresponse bias was not a signiﬁcant problem in the data analysis for this research.

207

Do Firms Create Value through International Strategic Alliances?

All of the measures used in this research were collected through the same questionnaire, which introduced the possibility of common method variance (Simonin, 1999a). Harman’s one-factor test (Konrad & Linnehan, 1995; Podsakoff & Organ, 1986; Scott & Bruce, 1994) was used to address the issue of common method variance. I performed principal components factor analysis on the questionnaire measurement items, extracting eight factors with eigenvalues greater than 1.0, which accounted for 71% of the total variance. Since factor one did not account for the majority of the variance (only 13%). The results of this test indicate that common method variance does not appear to be a problem in this study.

Measures This research adopts multiple indicators since they can offer a much ﬁner distinction. Table 1 reports the descriptive statistics and correlations among all the constructs. Most of the items are measured using 7-point Likert Scales, which is a widely used method in organizational research (Bryman, 2000; Churchill & Iacobucci, 2004). The items used in each construct, along with their reliability and conﬁrmatory factor analysis results, are outlined in Table 2. Antecedents of Knowledge Acquisition The measurement of learning intent is well developed and carefully operationalized in some research (see, e.g., Contractor et al., 2002). Hence, Table 1.

Means, Standard Deviations, and Correlations of the Constructs. Mean S.D. (1)

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Learning intent Absorptive capacity Protection Trust Interaction Knowledge acquisition Knowledge application Capability enhancement Network position enhancement

5.58 5.73 4.96 5.90 5.50 4.86 5.31 5.14 5.67

1.21 0.93 1.25 0.90 1.17 1.21 1.09 1.08 1.00

1.00 0.59 0.61 0.51 0.51 0.66 0.41 0.57 0.60

(2)

(3)

(4)

(5)

1.00 0.30 0.78 0.61 0.50 0.56 0.48 0.62

1.00 0.36 0.42 0.37 0.26 0.40 0.39

1.00 0.51 0.39 0.51 0.43 0.65

1.00 0.62 0.41 0.49 0.59

(6)

(7)

(8)

(9)

1.00 0.40 1.00 0.65 0.56 1.00 0.49 0.58 0.69 1.00

Notes: Correlations above 0.138 are signiﬁcant at po0.05; correlations above 0.181 are signiﬁcant at po0.01, based on n ¼ 160.

208

CHIA-LING ‘EUNICE’ LIU

Table 2.

Measurement Model and Conﬁrmatory Factor Analysis.

Items Dimension: Learning intent The explicit goal to learn from partners was stated during cooperation This collaboration was very important for us to learn from the foreign partners Senior management commitment underscored the importance of learning during cooperation Dimension: Absorptive capacity We have enough in-house expertise to fully use our partner’s knowledge We know what can be learned from our foreign partner We know who knows what and where the critical expertise resides in our partner Dimension: Protection Our partner has intentional procedures, routines, and policies to restrict the sharing of relevant information concerning its technology/process know-how The partner is very protective of its technology/ process know-how The partner clearly identiﬁes knowledge that is restricted in use Dimension: Trust A good faith relationship has developed over time in my ﬁrm’s dealings with the alliance partner My ﬁrm and this alliance partner understand each other well The partner and its employees are trustworthy; we would not break our promises in pursuit of our own interest Dimension: Interaction On-site visits and face-to-face communication Technology sharing Joint new product design Dimension: Knowledge acquisition New R&D expertise New product development Managerial practice Dimension: Knowledge application The ﬁrm makes efforts to establish formal policies, such as documentation, SOP, knowledge banks and expert systems, to facilitate the utilization of knowledge

l

t-Value

0.92

14.84

0.91

14.65

0.80

12.00

0.67

9.07

0.88

12.90

0.66

8.76

0.61

7.64

0.84

11.03

0.67

8.48

0.87

13.50

0.87

13.50

0.86

13.19

0.88 0.93 0.70

13.62 14.72 9.74

0.94 0.93 0.53

15.49 15.16 7.02

0.69

9.21

a

r

AVE

0.91

0.91

0.77

0.77

0.78

0.55

0.74

0.75

0.51

0.90

0.90

0.75

0.86

0.88

0.71

0.82

0.86

0.68

0.82

0.83

0.61

209

Do Firms Create Value through International Strategic Alliances?

Table 2. (Continued ) Items The ﬁrm offers on-job training The ﬁrm is able to locate and apply knowledge to change competitive conditions Dimension: Capability enhancement R&D capability in terms of product differentiation and functionality Speed to introduce new products Capability to perceive new technology development/market trends Dimension: Network position enhancement The ﬁrm gains more opportunities to form alliances with prominent global partners The ﬁrm increases the penetration of important growth markets The ﬁrm enhances its reputation

l

t-Value

0.86 0.79

12.46 11.05

0.83

12.55

0.90 0.89

14.28 14.08

0.84

12.79

0.88

13.72

0.88

13.78

a

r

AVE

0.90

0.91

0.76

0.90

0.90

0.75

Notes: l, factor-loading l; a, Cronbach’s a; r, Joreskog’s r; AVE, average variance extracted. Measurement ﬁt: Satorra–Bentler-scaled w2 ¼ 373.76; df ¼ 288; po0.001; CFI ¼ 0.948; BBNNFI ¼ 0.936; Bollen’s IFI ¼ 0.949; RMSEA ¼ 0.059, N ¼ 160.

learning intent in this research is measured by the ﬁrm’s inherent inclination for learning, the strategic importance of learning and top-management commitment to the learning goals (Schacht, 1999; Simonin, 1999a). On the basis of previous studies that contribute to the measures of absorptive capacity, the following three indicators were developed (Dyer & Singh, 1998; Schacht, 1999): the available in-house expertise, knowing what can be learned and where the critical expertise resides in the partner. Protection was measured using a three-item scale that were adapted from alliance learning literature (Gold et al., 2001; Simonin, 1999a): the intentional policies to restrict information sharing, the protectiveness of technology/process knowhow and the restriction of knowledge in use. Trust is conceptualized as ‘the belief that a partner’s promise is trustworthy and that a partner will accomplish its obligations in the relationship’ (Cullen & Johnson, 2000; Inkpen, 2000). The exchange of information and interpersonal contacts are viewed as effective ways of facilitating knowledge acquisition. The measures of interaction involve selectively applying the items developed by Bresman, Birkinshaw, and Nobel (1999) and Subramaniam and Venkatraman (2001). Respondents were asked to indicate the frequency of three forms of interaction: on-site visits and face-to-face communication, technology sharing and joint new product design.

210

CHIA-LING ‘EUNICE’ LIU

Knowledge Acquisition and Knowledge Application Knowledge acquisition is decided by the extent to which you have learned from your partner. The scale was designed to assess knowledge acquisition from the foreign partner across various areas, including (1) R&D expertise, (2) product development and (3) managerial practice. Knowledge application is deﬁned as the integration of learning, so it is broadly available and can be systematically exploited in the organization’s routines (Crossan et al., 1999; Nevis et al., 1995). Our research adapted their work on developing the measures of knowledge application as follows: the establishment of formal policies to facilitate knowledge utilization, on-job training and the ability to locate and apply knowledge to change competitive conditions. Capability and Network Position Enhancement Capability is deﬁned as ‘the ability to consistently and successfully integrate internal and external knowledge to lead to the growth and survival of the ﬁrm’ in this research. Capability enhancement was measured with three items comprising (1) R&D capability in terms of product differentiation and functionality (Jao, 1996), (2) speed to introduce new products (Subramaniam & Venkatraman, 2001) and (3) capability to perceive new technological development/market trends (Thomas, 1993). Network position is deﬁned as ‘the location of power to inﬂuence the decisions or actions of other ﬁrms and the base to provide development possibilities of the ﬁrm in the network’ (Ghauri et al., 2005; Hakansson & Johanson, 2001). The measurements were developed building on previous network studies by Johanson and Mattsson (1987) and Ernst (2000). Items include the better opportunities to form alliances with prominent global partners, the penetration of important growth markets and the enhancement of the ﬁrm’s reputation. A model of hypothesized relationships is presented in Fig. 1.

ANALYSIS Structural equation modeling (SEM) has become a popular methodology due to the ﬂexible interplay between theory and data and because it bridges theoretical and empirical knowledge for a better understanding of the real world (Simonin, 2004). SEM was considered to be a suitable statistical technique because a series of separate but interrelated dependence relationships were estimated simultaneously in this research. Bentler and Wu’s (2003) EQS 6.1 program was used to estimate structural equation models in this research.

Do Firms Create Value through International Strategic Alliances?

Fig. 1.

211

Diagram of Hypothesized Relationships.

Assessment of Measurement Model The ﬁrst stage of the SEM analysis was concerned with the assessment of the measurement items and how well they specify the relations between the observed measures and their posited underlying constructs (Anderson & Gerbing, 1988). The results of the measurement model are reported in Table 2. An inspection of Cronbach’s a for nine constructs reveals that all the coefﬁcients are greater than 0.70, which indicates acceptable reliability (Nunnally, 1978). Conﬁrmatory factor analysis was used to assess convergent and discriminant validity. Convergent validity was computed following Fornell and Larcker’s (1981) formula and evaluated based on the signiﬁcance of factor loading and shared variance. As shown in Table 2, factor loadings range from 0.53 to 0.94 (po0.01) and also share variance coefﬁcients ranging from 0.51 to 0.77, thus the constructs demonstrate sufﬁcient convergent validity. I adopted the procedure recommended by Jo¨reskog (1971) to assess discriminant validity. For example, the set of measures for learning intent was paired with absorptive capacity. I ran the model twice, once constraining the estimated correlation parameter between them to 1.0 and once freeing this parameter. A series of w2 difference tests were performed for one pair of

212

CHIA-LING ‘EUNICE’ LIU

factors at a time. For example, the w2 difference for constraining absorptive capacity and trust increases by 36.70 (po0.01) and for learning intent and absorptive capacity it raises to 80.13 (po0.01). A signiﬁcantly lower w2 value for the unconstrained models suggests that the constructs exhibit discriminant validity (Bagozzi & Phan, 1982). The Satorra–Bentler-scaled w2 test, which corrects for distortions of the normal theory method when data are non-normal (Satorra & Bentler, 1994), is statistically signiﬁcant (Satorra–Bentler-scaled w2 ¼ 373.76, df ¼ 288; po0.001). The other goodness-of-ﬁt indices suggest a good ﬁt of the conﬁrmatory measurement model: Comparative ﬁt index (CFI) ¼ 0.948; Bentler–Bonett non-normed ﬁt index (BBNNFI) ¼ 0.936; Bollen’s incremental ﬁt index (IFI) ¼ 0.949; root mean-square error of approximation (RMSEA) ¼ 0.059. Although the overall w2 statistic for the measurement model is signiﬁcant, this might be due to this test’s sensitivity to sample size (Bagozzi & Yi, 1988; Stump & Heide, 1996). However, the ratio of w2 to df (1.30, less than 3) corresponds to a satisfactory ﬁt (Hair, Anderson, Tatham, & Black, 2002). Overall, the measurement model represents a good ﬁt with the data.

Assessment of Structural Model The assessment of the structural model (see Fig. 1) provided a borderline model ﬁt (Satorra–Bentler-scaled w2 ¼ 453.448; df ¼ 307; po0.001). An alternative model strategy was therefore applied to improve model ﬁt, allowing the antecedent factors learning intent, trust and interaction to directly impact on the performance dimensions capability enhancement and network position enhancement. In our original model, these relationships were hypothesized to relate to the performance dimensions only in a fully mediated form through knowledge acquisition, i.e. I hypothesized no direct relationship. Model respeciﬁcation to improve model ﬁt is common in the social sciences because a priori models often do not adequately ﬁt the data, but respeciﬁcation is also controversial (Shook, Ketchen, Hult, & Kacmar, 2004). Anderson and Gerbing (1988) advocate the use of theory and content considerations for respeciﬁcation. Chin (1998) asserts that respeciﬁcation may be appropriate when validated with new samples. The additional introduction of direct relationships from learning intent to capability enhancement and network position enhancement, trust on network position enhancement and interaction on capability enhancement and network position enhancement resulted in a signiﬁcantly improved ﬁt

213

Do Firms Create Value through International Strategic Alliances?

Table 3. Structural Parameter Estimates and Model Fit Indices. Paths Learning intent - knowledge acquisition Learning intent - capability enhancement Learning intent - network position enhancement Absorptive capacity - knowledge acquisition Protection - knowledge acquisition Trust - knowledge acquisition Trust - network position enhancement Interaction - knowledge acquisition Interaction - capability enhancement Interaction - network position enhancement Knowledge acquisition - knowledge application Knowledge acquisition - capability enhancement Knowledge application - capability enhancement Capability enhancement - network position enhancement

Hypothesis (Expected Direction)

Standardized Coefﬁcient

t-Value

H1 (þ) H1b (þ) H1c (þ)

0.106 0.074 0.058

5.169** 2.096* 1.276

H2 (þ)

0.237

0.057

H3 (–) H4 (þ) H4b (þ) H5 (þ) H5b (þ) H5c (þ)

–0.119 –0.177 0.079 0.099 0.075 0.061

–1.293 –0.600 4.609** 4.602** 0.498 2.102*

H6a (þ)

0.056

4.436**

H6b (þ)

0.086

3.367**

H7 (þ)

0.108

3.973**

H8 (þ)

0.072

5.278**

Notes: Model ﬁt indices: Satorra–Bentler-scaled w2 (302) ¼ 413.63; po0.001; N ¼ 160; CFI ¼ 0.935; BBNNFI ¼ 0.925; IFI ¼ 0.936; RMSEA ¼ 0.064. * po0.05; ** po0.01.

(Satorra–Bentler-scaled w2 difference ¼ 39.818; df ¼ 5; po0.000). This is why I adopted the revised model. Table 3 reports the parameter estimates and goodness-of-ﬁt indicators of the revised structural equation model. Although the w2 statistic is signiﬁcant (Satorra–Bentler-scaled w2 ¼ 413.63; df ¼ 302; po0.001), the sufﬁciently low ratio of w2 to df (1.37, less than 3) reports a satisfactory ﬁt. Furthermore, CFI, BBNFI and Bollen’s IFI all indicate an excellent ﬁt for the structural model (CFI ¼ 0.935; BBNNFI ¼ 0.925; IFI ¼ 0.936). RMSEA ¼ 0.064 also exhibits adequate model ﬁt. Of the ﬁve postulated antecedents, the data analysis result displays a signiﬁcant effect on knowledge acquisition: learning intent (g ¼ 0.106, t ¼ 5.169) and interaction (g ¼ 0.099, t ¼ 4.602) in support of H1 and H5 respectively. That is, the greater (the smaller) the degree of learning intent

214

CHIA-LING ‘EUNICE’ LIU

and interaction between the alliance partners, the greater (the smaller) the level of the knowledge acquisition in the ISA. The result of H1b (g ¼ 0.074, t ¼ 2.096), which was introduced in the revised model, suggests that leaning intent has a positive inﬂuence on capability enhancement. Meanwhile, while H4 does not demonstrate a direct relationship between trust and knowledge acquisition, the direct support of H4b (g ¼ 0.079, t ¼ 4.609) plays out the importance of trust in enhancing alliance partner’s network position. Furthermore, H5c (g ¼ 0.061, t ¼ 2.102) afﬁrms interaction between the alliance partners strengthen the supplier’s network position directly and not only through knowledge acquisition. The link between knowledge acquisition and knowledge application is signiﬁcantly supported in H6a (b ¼ 0.056, t ¼ 4.436). The acceptance of H6b (b ¼ 0.086, t ¼ 3.367) and H7 (b ¼ 0.108, t ¼ 3.973) indicates that both knowledge acquisition and knowledge application contribute to capability enhancement. Last of all, the notable result of H8 (b ¼ 0.072, t ¼ 5.278) veriﬁes that capability enhancement exerts a strong positive inﬂuence on the improvement of network position (see Fig. 2).

Fig. 2. Final Revised Model and Examined Relationships. Notes: * Signiﬁcant at po0.05 Level; ** Signiﬁcant at po0.01. Dotted Arrows Represent Relationship which were Originally Hypothesized to Zero, but Introduced in the Revised Model.

Do Firms Create Value through International Strategic Alliances?

215

DISCUSSION In recent years, marketing scholars have expressed a considerable degree of interest in the antecedents and consequences of collaborative relationships (Kandemir et al., 2006; Krishnan, Martin, & Noorderhaven, 2006; Tsang, Nguyen, & Erramilli, 2004), whereas others have linked the alliance form (horizontal vs. vertical) to knowledge sharing between partners (Rindﬂeisch & Moorman, 2001) or alliance outcomes (Si & Bruton, 1999). Other studies have focused on the nature of governance in interﬁrm relationship, especially in buyer–seller contexts (Heide & John, 1990; Nobeoka, Dyer, & Madhok, 2002). On the basis of past research, my study proposes a framework integrating different streams of theories to show how different factors affect interﬁrm collaborations through both knowledge acquisition from alliance partners and the application of such knowledge to enhance the ﬁrm’s capability and network position. In this section, I highlight the key implications of my ﬁndings.

Key Determinants of Knowledge Acquisition The results indicate that the most signiﬁcant determinant of knowledge acquisition is learning intent, closely followed by partner interaction. A ﬁrm with ‘learning intent’ views collaboration as an opportunity for learning and competence acquisition from their alliance partners (Hamel, 1991). The SEM result is consistent with Tsang (2002), who states that a ﬁrm with strong learning motivation is more likely to allocate resource and commit management involvement to knowledge acquisition. The empirical results strongly support the conceptual postulations that inter-partner interaction has a positive effect on knowledge acquisition in the ISA context. This result emphasizes the importance of relationship marketing and management in dynamic environment. Relationship between local suppliers and foreign partners plays an important role on alliance success. This also conﬁrms the relational view that the ﬁrms should establish an on-going relationship that can foster learning to create value with the alliance partners (Kotabe, Martin, & Domoto, 2003). The SEM results show that absorptive capacity is only weakly associated with knowledge acquisition. This suggests that prior knowledge acquired from their foreign partners did not signiﬁcantly help the suppliers to acquire new knowledge. Although absorptive capacity theory is not supported, this ﬁnding is consistent with Lane et al. (2001), who found that the prior

216

CHIA-LING ‘EUNICE’ LIU

knowledge acquired by International Joint Ventures (IJVs) was of less help in facilitating current learning. One possible explanation is that prior knowledge expires quickly under the conditions of rapid technological change and short product life cycle in the IT industry. Another interesting ﬁnding is that knowledge protection is only marginally associated with knowledge acquisition. Most of the literature suggests that the ﬁrm must have an incentive to deploy shielding to prevent leakage of some of the core competences to the other partner (Simonin, 2004). However, this safeguard orientation may obstruct the development of mutually beneﬁcial relationships and limit learning opportunities (Madhok & Tallman, 1998). One possible explanation is that knowledge protection impedes the achievement of alliance goals (Norman, 2004). Most alliances between foreign buyers and Taiwanese IT suppliers involve product development. If such goals are not fulﬁlled due to the limitation on sharing knowledge between partners, the alliance provides less value. The linkage between trust and knowledge acquisition is surprisingly negative. In most of the ISA literature, trust leads to a shared understanding between the alliance partners (Dhanaraj et al., 2004) and allows greater access to resources (Uzzi & Lancaster, 2003). However, our result suggests that the level of trust between buyers and suppliers does not seem to inﬂuence knowledge acquisition. It is possible that since this study focuses on the contractual relationships between MNC buyers and NIC suppliers, there is less risk of opportunistic behavior since the NIC suppliers are typically less powerful than the MNC buyers.

Integration of External and Internal Learning Organizational learning is a complex issue. Most studies of alliance learning focus on the acquisition of knowledge externally from partners and, to a lesser extent, on the utilization of knowledge internally in the organization (Nevis et al., 1995). Although some knowledge acquired externally can be acted on immediately, it is more likely that it will have to be adapted and disseminated internally before it can be applied to commercial ends (Lane et al., 2001). As McEvily and Zaheer (1999) note, knowledge acquisition is not a separate event but rather a part of a multistage process that includes the ﬁnal utilization of this knowledge to achieve organizational goals. The marketing literature has also indicated the importance of information use activities (Baker, Simpson, & Siguaw, 1999; Bulent & Seigyoung, 2006; Kohli & Jaworski, 1990). Following these propositions, this research

Do Firms Create Value through International Strategic Alliances?

217

proposes a framework to connect knowledge acquisition, and knowledge application and examines the combined effects on capability enhancement. Network Position Enhancement Since the extant alliance learning literature is concerned with the dyadic relationship and suffers from what Gulati (1998, p. 307) called ‘dyadic reductionism’, this research makes a signiﬁcant contribution by systematically examining the network perspective of collaborations. Many ﬁrms today are involved in a web of interorganizational arrangements. The ﬁrm’s position in the network is increasingly critical because this facilitates the acquisition of skills and knowledge from other members. However, unlike previous studies that were mostly focused on the impact of network position on alliance learning (Contractor et al., 2002; Goerzen & Beamish, 2005; Grant & Baden-Fuller, 2002), this research reverses the direction to examine the impact of the latter on the former. A positive relationship between capability enhancement and network position enhancement is signiﬁcantly supported to afﬁrm this notion.

LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH This chapter provides novel insights into a more comprehensive structure of the knowledge acquisition process and integrates the learning and relationship strands of the literature with knowledge acquisition and application. This inter-partner learning, in turn, demonstrates to be effective within the context of contractual collaborative partnerships between buyers and suppliers. Although the alliance structures and similarities between the alliance partners may facilitate knowledge acquisition, our empirical results suggest that learning opportunities are not limited by the mode or symmetry of the partnerships. This research complements prior studies by systematically examining the impact of alliance learning on the ﬁrm’s capability enhancement (Doz, 1996; Dussauge, Bernard, & Will, 2000). It takes a further step of linking with network-related research to consider the inﬂuence of alliance learning on the ﬁrm’s position in the network because many organizations are increasingly involved in a multi-partner web of alliances (Ghauri et al., 2005). The limitations that inﬂuence the interpretation of this study are also well recognized. First, several factors related to alliance learning in previous

218

CHIA-LING ‘EUNICE’ LIU

studies, such as alliance duration (Anderson & Dahlqvist, 2002; Simonin, 1999b), size (Kotabe et al., 2003), alliance experience (Lambe et al., 2002) and the country of origin of the alliance partner (Mowery et al., 1996) only received scant attention. Further work on these variables would be a natural extension of the current research. Second, a potential problem is caused by the one-sided survey that depends on the suppliers’ perceptions and judgments about the alliances. Dyadic data (including both sides of an alliance in the research) would have been more desirable. Future work will beneﬁt by adding balancing views from the MNC buyers. Third, it is important to understand the evolutionary role of collaboration in order to advance the theoretical understanding of business strategy and alliance outcome (Dussauge et al., 2000). This study is cross-sectional, since all the variables included in this research were measured at the same time. Given the challenges involved in designing longitudinal research – which would certainly help to further explore the dynamic nature of the process and causes and effects of alliance learning – replications at other points in time and involving related industries are encouraged to raise conﬁdence in the ﬁndings. The study involved SEM, which is a methodologically rigorous, yet challenging approach, particularly when a relative large number of variables and smaller sample sizes are involved. The integration of various streams of literatures in this study also involved issues of model respeciﬁcation to improve model ﬁt. Although this procedure is quite common in the social sciences (Shook et al., 2004), replication of results on other samples is encouraged to validate hypothesized relationships. Finally, the issue of ‘interface-design’ in ISAs seems promising for future work (Foss & Pedersen, 2004). As ﬁrms join forces to achieve mutually beneﬁcial goals through their alliances, the relationship between the alliance partners plays an important role in knowledge management (Inkpen, 1996). Inter-partner interactions allow mutual information to be shared and joint actions to be speciﬁed. It may be useful to consider the multiple learning interfaces between and within organizations. It would be very useful to acquire a better understanding of the impact interface-design may have on knowledge ﬂows in the learning process; and how interface-design co-evolves with the partnerships.

REFERENCES Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychologyical Bulletin, 103(3), 411–423.

Do Firms Create Value through International Strategic Alliances?

219

Anderson, U., & Dahlqvist, J. (2002). Business-governed product development: Knowledge utilization in business relationships. In: H. Hakansson & J. Johanson (Eds), Business network learning (pp. 53–68). Amsterdam: Elsevier Science Ltd. Andersson, U., Bjorkman, I., & Forsgren, M. (2005). Managing subsidiary knowledge creation: The effect of control mechanisms on subsidiary local embeddedness. International Business Review, 14(5), 521–538. Armstrong, S. J., & Overton, T. S. (1977). Estimating nonresponse bias in mail surveys. Journal of Marketing Research, 14(3), 396–402. Bagozzi, R. P., & Phan, P. H. (1982). Representing and testing organizational theories: A holistic construal. Administrative Science Quarterly, 27, 459–489. Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74–94. Baker, T. L., Simpson, P. M., & Siguaw, J. A. (1999). The impact of suppliers’ perceptions of reseller market orientation on key relationship constructs. Journal of the Academy of Marketing Science, 27(1), 50. Baughn, C. C., Denekamp, J. G., Stevens, J. H., & Osborn, R. N. (1997). Protecting intellectual capital in international alliances. Journal of World Business, 32(2), 103–117. Bentler, P. M., & Wu, E. J. C. (2003). EQS 6.0 user’s guide. Encino, CA: Multivariate Software Inc. Boersma, M. F., Buckley, P. J., & Ghauri, P. N. (2003). Trust in international joint venture relationships. Journal of Business Research, 56(12), 1031–1042. Brass, D. J., Galaskiewicz, J., Greve, H. R., & Tasai, W. (2004). Taking stock of networks and organizations: A multilevel perspective. Academy of Management Journal, 47(6), 795–817. Bresman, H., Birkinshaw, J., & Nobel, R. (1999). Knowledge transfer in international acquisitions. Journal of International Business Study, 30(3), 439–462. Browning, E. S. (1994). Side by side: Computer chip project brings rivals together, but the cultures clash. The Wall Street Journal (May 3), A1–A6. Bryman, A. (2000). Research methods and organization studies. London: Routledge. Bulent, M., & Seigyoung, A. (2006). Creating a ﬁrm-level dynamic capability through capitalizing on market orientation and innovativeness. Journal of the Academy of Marketing Science, 34(1), 63–73. Chen, H., & Chen, T.-J. (2003). Governance structures in strategic alliances: Transaction cost versus resource-based perspective. Journal of World Business, 38, 1–14. Chetty, S. K., & Wilson, H. I. M. (2003). Collaborating with competitors to acquire resources. International Business Review, 12(1), 61–81. Chin, W. W. (1998). Issues and opinion on structural equation modeling. Management Information Systems Quarterly, 22(1), 7–16. Churchill, G. A., & Iacobucci, D. (2004). Marketing research: Methodological foundations (9th ed.). Mason, OH: Thomson Corporation, South Western. Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35, 128–152. Contractor, F. J., Kim, C.-S., & Beldona, S. (2002). Interﬁrm learning in alliance and technology networks: An empirical study in the global pharmaceutical and chemical industries. In: F. J. Contractor & P. Lorange (Eds), Cooperative strategies and alliances (pp. 493–516). Amsterdam: Elsevier. Contractor, F. J., & Lorange, P. (2002a). Cooperative strategies and alliances. Amsterdam: Elsevier.

220

CHIA-LING ‘EUNICE’ LIU

Contractor, F. J., & Lorange, P. (2002b). The growth of alliances in the knowledge-based economy. International Business Review, 11, 485–502. Crossan, M. M., & Berdrow, I. (2003). Organizational learning and strategic renewal. Strategic Management Journal, 24, 1087–1105. Crossan, M. M., Lane, H. W., & White, R. E. (1999). An organizational learning framework: From intuition to institution. Academy of Management Review, 24(3), 522–537. Cullen, J. B., & Johnson, J. L. (2000). Success through commitment and trust: The soft side of strategic alliance management. Journal of World Business, 35(3), 223–240. Day, G. S., & Montogomery, D. B. (1999). Charting new directions for marketing. Journal of Marketing, 63, 3–13. De Mattos, C., Sanderson, S., & Ghauri, P. (2002). Negotiating alliances in emerging markets – Do partners’ contributions matter? Thunderbird International Business Review, 44(6), 701–728. Demsetz, H. (1993). The theory of the ﬁrm revisited. In: O. E. Williamson & S. G. Winter (Eds), The nature of the ﬁrm (pp. 159–178). New York, NY: Oxford University Press. Dhanaraj, C., Lyles, M. A., Steensma, H. K., & Tihanyi, L. (2004). Managing tacit and explicit knowledge transfer in IJVs: The role of relational embeddedness and the impact on performance. Journal of International Business Studies, 35, 428–442. Dhanaraj, C., & Parkhe, A. (2006). Orchestrating innovation networks. Academy of Management Review, 31(3), 659–669. Dillman, D. A. (2000). Mail and internet surveys: The tailored design method. New York, NY: Wiley. Dodgson, M. (1993). Learning, trust, and technological collaboration. Human Relations, 46(1), 77–95. Doz, Y. (1996). The evolution of cooperation in strategic alliances: Initial conditions or learning process. Strategic Management Journal, 17(Summer Special Issue), 55–84. Dussauge, P., Bernard, G., & Will, M. (2000). Learning from competing partners: Outcomes and durations of scale and link alliances in Europe, North America and Asia. Strategic Management Journal, 21(2), 99–126. Duysters, G., Kok, G., & Vaandrager, M. (1999). Crafting successful strategic technology partnerships. R&D Management, 29(4), 343–351. Dyer, J. H., & Singh, H. (1998). The relational view: Cooperative strategy and sources of interorganizational competitive advantage. Academy of Management Review, 23(4), 660–679. Einhorn, B., & Burrows, P. (2003). Climbing the tech ladder: Taiwan is becoming a center for R&D as well as manufacturing. Business Week, October 6, pp. 24–25. Eriksson, K., & Chetty, S. (2003). The effect of experience and absorptive capacity on foreign market knowledge. International Business Review, 12(6), 673–695. Ernst, D. (2000). Inter-organizational knowledge outsourcing: What permits small Taiwanese ﬁrms to compete in the computer industry. Asia Paciﬁc Journal of Management, 17(2), 223–255. Ernst, D., & Guerrieri, P. (1998). International production networks and change trade patterns in East Asia: The case of the electronics industry. Oxford Development Studies, 26(2), 191–212. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50.

Do Firms Create Value through International Strategic Alliances?

221

Foss, N. J., & Pedersen, T. (2004). Organizing knowledge process in the multinational corporation: An introduction. Journal of International Business Studies, 35, 340–349. Frost, T. S., & Zhou, C. (2005). R&D co-practice and ‘reverse’ knowledge integration in multinational ﬁrms. Journal of International Business Studies, 36(6), 676–687. Ghauri, P. N. (1999). International marketing and purchasing (Vol. 9). Stamford, CT: JAI Press Inc. Ghauri, P. N., & Gronhaug, K. (2005). Research methods in business studies: A practical guide (3rd ed.). London: Prentice Hall International. Ghauri, P. N., Hadjikhani, A., & Johanson, J. (2005). Managing opportunity development in business networks. Basingstroke: Plagrave MacMillan. Ghauri, P. N., & Prasad, S. B. (1995). A network approach to probing Asia’s interﬁrm linkages. Advances in International Comparative Management, 10, 63–77. Ghoshal, S. (1987). Global strategy: An organizing framework. Strategic Management Journal, 5, 425–440. Goerzen, A., & Beamish, P. W. (2005). The effect of alliance network diversity on multinational enterprise performance. Strategic Management Journal, 26, 335–354. Gold, A. H., Malhotra, A., & Segars, A. H. (2001). Knowledge management: An organizational capabilities perspective. Journal of Management Information Systems, 18(1), 185–214. Grant, R. M. (1996). Toward a knowledge-based theory of the ﬁrm. Strategic Management Journal, 17(Winter Special Issue), 109–122. Grant, R. M., & Baden-Fuller, C. (2002). The knowledge-based view of strategic alliance formation: Knowledge accessing versus organizational learning. In: F. J. Contractor & P. Lorange (Eds), Cooperative strategies and alliances (pp. 418–436). Amsterdam: Elsevier. Gulati, R. (1998). Alliances and network. Strategic Management Journal, 19(4), 293–317. Gulati, R. (1999). Network location and learning: The inﬂuence of network resources and ﬁrm capabilities on alliance formation. Strategic Management Journal, 20(5), 397–420. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (2002). Multivariate data analysis (6th ed.). Upper Saddle River, NJ: Prentice-Hall. Hakansson, H., & Johanson, J. (2001). Business network learning. Amsterdam: Elsevier Science Ltd. Hamel, G. (1991). Competition for competence and inter-partner learning within international strategic alliances. Strategic Management Journal, 12(Special Issue), 83–103. Hamel, G., Doz, Y. L., & Prahalad, C. K. (1989). Collaborate with your competitors and win. Harvard Business Review, 67(1), 133–139. Heide, J. B., & John, G. (1990). Alliances in industrial purchasing: The determinants of joint action in buyer–supplier relationships. Journal of Marketing Research, 27(1), 24–36. Helleloid, D., & Simonin, B. L. (1994). Organizational learning and a ﬁrm’s core competence. In: G. Hamel & A. Heene (Eds), Competence-based competition (pp. 213–239). New York, NY: Wiley. Huber, G. P. (1991). Organizational learning: The contributing processes and the literatures. Organization Science, 2(1), 88–115. Inkpen, A. C. (1996). Creating knowledge through collaboration. California Management Review, 39(1), 123–140. Inkpen, A. C. (1998). Learning and knowledge acquisition through international strategic alliances. Academy of Management Executive, 12(4), 69–80.

222

CHIA-LING ‘EUNICE’ LIU

Inkpen, A. C. (2000). Learning through joint ventures: A framework of knowledge acquisition. Journal of Management Studies, 37(7), 1019–1043. Inkpen, A. C. (2002). Learning, knowledge management, and strategic alliances: So many studies, so many unanswered questions. In: F. J. Contractor & P. Lorange (Eds), Cooperative strategies and alliances (pp. 267–289). Amsterdam: Pergamon. Inkpen, A. C., & Beamish, W. P. (1997). Knowledge, bargaining power, and the instability of international joint ventures. Academy of Management Executive, 22(1), 177–202. Inkpen, A. C., & Dinur, A. (1998). Knowledge management processes and international joint ventures. Organization Science, 9(4), 454–468. Jao, I.-W. (1996). Capability enhancement through cross-border contract manufacturing alliances: A study of Taiwan information technology ﬁrms. Unpublished Ph.D. dissertation, London Business School, London. Johanson, J., & Mattsson, L.-G. (1987). Interorganizational relations in industrial systems: A network approach compared with the transaction-cost approach. International Studies of Management & Organization, 7(1), 34–48. Jo¨reskog, K. G. (1971). Statistical analysis of sets of congeneric tests. Psychometrika, 36, 109–133. Kale, P., Singh, H., & Perlmutter, H. (2000). Learning and protection of proprietary assets in strategic alliances: Building relational capital. Strategic Management Journal, 21(3), 217–237. Kandemir, D., Ghauri, P. N., & Cavusgil, T. S. (2002). The strategic role of organizational learning in relationship quality in strategic alliances. In: F. J. Contractor & P. Lorange (Eds), Cooperative strategic alliance (pp. 799–828). Oxford, UK: Pergamon. Kandemir, D., Yaprak, A., & Cavusgil, S. T. (2006). Alliance orientation: Conceptualization, measurement, and impact on market performance. Journal of the Academy of Marketing Science, 34(3), 324–340. Kemp, R. G. M., & Ghauri, P. N. (2001). Interdependency in joint-ventures: The relationship between dependence asymmetry and performance. Journal of Chain and Network Science, 1(2), 101–110. Kogut, B. (2000). The network as knowledge: Generative rules and the emergence of structure. Strategic Management Journal, 21(3), 405–425. Kohli, A. K., & Jaworski, B. J. (1990). Market orientation: The construct, research propositions, and managerial implications. Journal of Marketing, 54(2), 1–18. Konrad, A. M., & Linnehan, F. (1995). Formalized HRM structures: Coordinating equal employment opportunity or concealing organizational practices? Academy of Management Journal, 38(3), 787–820. Kotabe, M., Martin, X., & Domoto, H. (2003). Gaining from vertical partnerships: Knowledge transfer, relationship duration, and supplier performance improvement in the U.S. and Japanese automotive industries. Strategic Management Journal, 24(4), 293–316. Krishnan, R., Martin, X., & Noorderhaven, N. G. (2006). When does trust matter to alliance performance? Academy of Management Journal, 49(5), 894–917. Lambe, J. C., Spekman, R. E., & Hunt, S. D. (2002). Alliance competence, resources, and alliance success: Conceptualization, measurement, and initial test. Journal of the Academy of Marketing Science, 30(2), 141–158. Lane, P. J., & Lubatkin, M. (1998). Relative absorptive capacity and interorganizational learning. Strategic Management Journal, 19(5), 461–477.

Do Firms Create Value through International Strategic Alliances?

223

Lane, P. J., Salk, J. E., & Lyles, M. A. (2001). Absorptive capacity, learning, and performance in international joint ventures. Strategic Management Journal, 22(12), 1139–1161. Lorange, P., & Roos, J. (1992). Strategic alliances: Formation, implementation and evolution. Oxford: Blackwell. Madhok, A. (2006). How much does ownership really matter? Equity and trust relations in joint venture relationships. Journal of International Business Studies, 37(1), 4–11. Madhok, A., & Tallman, S. B. (1998). Resources, transactions and rents: Managing value through interﬁrm collaborative relationships. Organization Science, 9(3), 326–339. McEvily, B., & Zaheer, A. (1999). Bridging ties: A source of ﬁrm heterogeneity in competitive capabilities. Strategic Management Journal, 12(May/June), 1133–1156. Miller, D. J., Fern, M. J., & Cardinal, L. B. (2007). The use of knowledge for technological innovation within diversiﬁed ﬁrms. Academy of Management Journal, 50(2), 308–326. Mol, M. J., van Tulder, R. J. M., & Beije, P. R. (2005). Antecedents and performance consequences of international outsourcing. International Business Review, 14(5), 599–617. Mowery, D. C., Oxley, J. E., & Silverman, B. S. (1996). Strategic alliances and interﬁrm knowledge transfer. Strategic Management Journal, 17(Winter), 77–91. Mowery, D. C., Oxlley, J. E., & Silverman, B. S. (2002). The two faces of partner-speciﬁc absorptive capacity: Learning and cospecialization in strategic alliances. In: F. J. Contractor & P. Lorange (Eds), Cooperative strategies and alliances (pp. 381–418). Amsterdam: Pergamon. Nevis, E. C., DiBella, A. J., & Gould, J. M. (1995). Understanding organizations as learning systems. Sloan Management Review, 36(2), 73–85. Nobeoka, K., Dyer, J. H., & Madhok, A. (2002). The inﬂuence of customer scope on supplier learning and performance in the Japanese automobile industry. Journal of International Business Studies, 33(4), 717–736. Nonaka, I., Von Krogh, G., & Voelpel, S. (2006). Organizational knowledge creation theory: Evolutionary paths and future advances. Organization Studies, 27(8), 1179–1208. Norman, P. M. (2004). Knowledge acquisition, knowledge loss, and satisfaction in high technology alliances. Journal of Business Research, 57(6), 610–619. Nunnally, J. C. (1978). Psychometric theory. New York, NY: McGraw-Hill. Perry, M. L., Sengupta, S., & Krapfel, R. (2004). Effectiveness of horizontal strategic alliances in technologically uncertain environments: Are trust and commitment enough? Journal of Business Research, 57(9), 951–956. Phan, P. H., & Peridis, T. (2000). Knowledge creation in strategic alliances: Another look at organizational learning. Asia Paciﬁc Journal of Management, 17(2), 201–222. Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12(4), 531–544. Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Harvard Business Review, 68(3), 79–91. Ramamurti, R. (2004). Developing countries and MNEs: Extending and enriching the research agenda. Journal of International Business Studies, 35(4), 277–283. Rindﬂeisch, A., & Moorman, C. (2001). The acquisition and utilization of information in new product alliances: A strength-of-ties perspective. Journal of Marketing, 65(April), 1–18. Sarkar, M. B., Raj, E. S., Cavusgil, T., & Aulakh, P. S. (2001). The inﬂuence of complementarity, compatibility, and relationship capital on alliance performance. Journal of the Academy of Marketing Science, 29(4), 358–373.

224

CHIA-LING ‘EUNICE’ LIU

Satorra, A., & Bentler, P. M. (1994). Corrections to test statistics and standard errors in covariance structure analysis. In: A. von Eye & C. C. Clogg (Eds), Latent variables analysis: Applications for developmental research (pp. 399–419). Thousand Oaks, CA: Sage. Schacht, O. (1999). The evolution of organisational learning in international strategic alliances in biopharmaceuticals. Unpublished Ph.D. dissertation, Cambridge University, Cambridge. Schoorman, D. F., Mayer, R. C., & Davis, J. H. (2007). An integrative model of organizational trust: Past, present, and future. Academy of Management Review, 32(2), 344–354. Scott, S. G., & Bruce, R. A. (1994). Determinants of innovative behavior: A path model of individual innovation in the workplace. Academy of Management Journal, 37(3), 580–607. Scott-Kennel, J., & Enderwick, P. (2004). Inter-ﬁrm alliance and network relationships and the eclectic paradigm of international production: An exploratory analysis of quasiinternalisation at the subsidiary level. International Business Review, 13(4), 425–445. Seidler, J. (1974). On using informants: A technique for collecting quantitative data and controlling for measurement error in organizational analysis. American Sociological Review, 39(6), 816–831. Shook, C. L., Ketchen, D. J. Jr., Hult, G. T. M., & Kacmar, K. M. (2004). An assessment of the use of structural equation modeling in strategic management research. Strategic Management Journal, 25(4), 397–404. Si, S. X., & Bruton, G. D. (1999). Knowledge transfer in international joint ventures in transitional economies: The China experience. Academy of Management Executive, 13(1), 83–90. Simonin, B. L. (1999a). Ambiguity and the process of knowledge transfer in strategic alliances. Strategic Management Journal, 20(7), 595–623. Simonin, B. L. (1999b). Transfer of marketing know-how in international strategic alliances: An empirical investigation of the role and antecedents of knowledge ambiguity. Journal of International Business Study, 30(3), 463–490. Simonin, B. L. (2004). An empirical investigation of the process of knowledge transfer in international strategic alliances. Journal of International Business Studies, 35(5), 407–427. Smitka, M. (1991). Competitive ties: Subcontracting in the Japanese automotive industry. New York, NY: Columbia University Press. Stump, R. L., & Heide, J. B. (1996). Controlling supplier opportunism in industrial relationships. Journal of Marketing Research, 33(4), 431–441. Subramaniam, M., & Venkatraman, N. (2001). Determinants of transnational new product development capability: Testing the inﬂuence of transferring and deploying tacit overseas knowledge. Strategic Management Journal, 22(4), 359–378. Susan, F., Ron, E., & Bill, S. (2006). How smaller born-global ﬁrms use networks and alliances to overcome constraints to rapid internationalization. Journal of International Marketing, 14(3), 33. Svejenova, S. (2006). How much does trust really matter? Some reﬂections on the signiﬁcance and implications of Madhok’s trust-based approach. Journal of International Business Studies, 37(1), 12–20. Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practice within the ﬁrm. Strategic Management Journal, 17(Winter Special Issue), 27–43. Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533.

Do Firms Create Value through International Strategic Alliances?

225

Thomas, R. J. (1993). New product development-managing and forecasting for strategic success. New York, NY: Wiley. Thorelli, H. B. (1986). Networks: Between markets and hierarchies. Strategic Management Journal, 7(1), 37–51. Tsang, E. W. K. (2002). Acquiring knowledge by foreign partners from international joint ventures in a transition economy: Learning-by-doing and learning myopia. Strategic Management Journal, 23(9), 835–854. Tsang, E. W. K., Nguyen, D. T., & Erramilli, M. K. (2004). Knowledge acquisition and performance of international joint ventures in the transition economy of Vietnam. Journal of Information Marketing, 12(2), 82–103. Uzzi, B. (1997). Social structure and competition in interﬁrm networks: The paradox of embeddedness. Administrative Science Quarterly, 42(1), 35–67. Uzzi, B., & Lancaster, R. (2003). Relational embeddedness and learning: The case of bank loan managers and their clients. Management Science, 49(4), 383–399. Vonderembse, M. A., & Tracey, M. (1999). The impact of supplier selection criteria and supplier involvement on manufacturing performance. Journal of Supply Chain Management, 35(3), 33–39. von Hippel, E. (1994). ‘‘Sticky information’’ and the locus of problem solving: Implications for innovation. Management Science, 40(4), 429–439. Weerawardena, J., O’Cass, A., & Julian, C. (2006). Does industry matter? Examining the role of industry structure and organizational learning in innovation and brand performance. Journal of Business Research, 59(1), 37–45. Wu, F., & Cavusgil, S. T. (2006). Organizational learning, commitment, and joint value creation in interﬁrm relationships. Journal of Business Research, 59(1), 81–89. Yan, A., & Gray, B. (1994). Bargaining power, management control, and performance in United States–China joint ventures: A comparative case study. Academy of Management Journal, 37(6), 1478–1517. Zaheer, A., & Bell, G. G. (2005). Beneﬁting from network position: Firm capabilities, structural holes, and performance. Strategic Management Journal, 26(9), 809–825. Zineldin, M., & Dodourova, M. (2005). Motivation, achievements and failure of strategic alliances: The case of Swedish auto-manufacturers in Russia. European Business Review, 17(5), 460–470.

WHAT CAUSES BREAK-UPS? FACTORS DRIVING THE DISSOLUTION OF MARKETINGORIENTED INTERNATIONAL JOINT VENTURES Mehmet Berk Talay and M. Billur Akdeniz In recent years, accelerating globalization of markets along with the breakdown of national barriers have rendered international joint ventures (IJVs) ubiquitous and strategically signiﬁcant tools for managing international business operations (Anderson, 1990; Dhanaraj & Beamish, 2004; Geringer & Hebert, 1991; Harrigan, 1987; Lu & Beamish, 2006). As such they have become viable means of achieving economies of scale, critical mass, reducing risks, learning new skills and technologies, and facilitating effective resource sharing (Beamish & Berdrow, 2003; Bleeke & Ernst, 1993; Harrigan, 1988; Yip, 1992). These trends stimulated intensive research on IJVs in scholarly international business literature. An IJV involves two or more legally distinct organizations each of which shares the decision-making activities of the jointly owned entity (Geringer, 1988). Based on this deﬁnition, an IJV has to meet at least one of the three following criteria: (1) at least one parent ﬁrm is headquartered outside the IJV’s country of operation, (2) a joint venture has a signiﬁcant level of operations in more than one country, and (3) parent ﬁrms are from different New Challenges to International Marketing Advances in International Marketing, Volume 20, 227–256 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020012

227

228

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

countries (Geringer & Hebert, 1989). Since the early 1980s there has been an explosive growth of IJVs. Although, these inter-ﬁrm collaborations have become a critical part of corporate strategy and ﬁrm growth, they also entail serious competitive risks and managerial difﬁculties in implementing the cooperative relationships, which hinder the achievement of collective goals and lead to the dissolution of these entities (Park & Ungson, 2001). In addition, as a consequence of the globalizing world, businesses also choose to break-up as a more muniﬁcent strategic choice. A substantial literature has been materialized with the purpose of examining the motivations for IJV formation, the factors behind their strategic success, and measuring their performance. However, it is surprising that relatively less number of studies have been conducted on the dissolution of these entities despite the fact that the dissolution of IJVs is as likely as their survival. Parkhe (1993) indicates that at least 50 percent of equity joint ventures either fail or dissolve, which may cause various adverse outcomes for the partner ﬁrms. This relative lack of research drives our attention to the dissolution of these IJVs, which constitutes a new avenue for international business scholars and practitioners regarding the existing and newly arising challenges in the management of this speciﬁc entry mode. In this study, we aim to examine the ﬁrm- and country-level antecedents of IJV dissolution. This research sets out to shed light on two important research questions: (1) What is the role of competition–commitment relationship between parent ﬁrms on the likelihood of IJV dissolution? and (2) How do the country-level variables affect IJV dissolution? This study is expected to contribute to the IJV literature in several ways. First, IJV dissolution has been regarded as an outcome of bad performance, ﬁnancial incapability, undesirable alliance partners, or organizational failure. However, this study conceptualizes dissolution from a more recent perspective in alignment with the developments in the international business. We do not necessarily deﬁne dissolution in terms of failure or bad performance but as a strategic choice of the parent ﬁrms to bring the IJV to an end. This can represent a successful achievement of preset goals for the IJV and/or changes in the corporate strategies of the parents that render the IJV redundant. Recently, the practice of associating dissolution with organizational failure or poor performance has become more questionable in line with the changes in the contemporary business environment. For instance, when the functional foods joint venture, Altus Food was formed by Novartis and Quaker Oats in 2000, the global market for functional foods was estimated to be growing by around 15 percent per year, against 3 percent for most of the food industry. At the time of its

What Causes Break-Ups?

229

inception, the goal for Altus was to bring together Quaker’s mass marketing of food and beverages with Novartis’ research in nutrition science and to launch new products in this attractive business segment. However, the expectations about the functional foods business did not turn out to be as high as anticipated – just ﬁve percent in a good year; and Altus operations were ceased in 2002. In this case, the main reason for the termination of Altus can be attributed more to a strategy change in the partner ﬁrms due to disappointing environmental conditions rather than to the failure of the IJV. Second, we focus on marketing-oriented IJVs (i.e., the IJVs that are formed with the objective of joint selling, distributing, and promoting goods or services that are either jointly or individually manufactured by the partner ﬁrms). Among other knowledge management and transfer issues in cross-border alliances, marketing knowledge deserves a special attention for several reasons. One of them is that it has been relatively neglected and requires more attention compared to technology, manufacturing, know-how types of knowledge (Wong, Maher, & Luk, 2002). Another reason rises from the fact that marketing skills and expertise are more tacit than production, and technology (Shenkar & Li, 1999; Lane, Salk, & Lyles, 2001), are embedded and not easily codiﬁed in formulas (Zander & Kogut, 1995). Finally, in the literature, marketing knowledge is deﬁned as idiosyncratic and cultural/social-context dependent (Hau & Evangelista, 2007; Zander & Kogut, 1995). Third, we test our hypotheses with a comprehensive set of IJV dissolutions formed by both U.S. and non-U.S. ﬁrms; on contrary to most of the studies, which are limited to only the IJVs formed by U.S. companies. The organization of the chapter is as follows. In the next section, we explicate the theoretical background and our hypotheses. Then, in the methodology section, we explain the data and the covariates used in the analyses. After we present the main ﬁndings, we elaborate on the major implications for IJV literature and practitioners along with the limitations and suggestions for future research.

LITERATURE REVIEW In tandem with the drastic increase of IJVs, the academic research regarding various issues about them has also augmented. Myriad of studies examining the antecedents and outcomes of partner/location selection, different control and safeguarding mechanisms, performance, and stability of IJVs have appeared in revered scholarly outlets. A review of the previous research on IJVs reveals that a vast amount of these studies is focused on their formation

230

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

(e.g., Kogut, 1988) and the advantages of them as governance structures (e.g., Hennart, 1988). There has been a signiﬁcant debate on the drivers of IJV formations in the literature. Scholars, particularly from the ﬁelds of economics, strategic management, and international business, have provided various conceptualizations of the IJV formations based on various theories. Among those conceptualizations, the ones based on transaction cost theory and resourcebased view (RBV) have received more interest in the literature (Tsang, 2000). Transaction Cost Analysis (TCA) has been extended by Williamson (1991) to explain the hybrid modes of governance among ﬁrms (e.g., joint ventures, strategic alliances, etc.), which are regarded as a compromise between the two ends of the market-hierarchy continuum. Williamson (1991) suggests that hybrid governance structures entail better incentives for partners to avoid opportunistic behavior and prove superior against uncertainty due to their more adaptive and ﬂexible structure than hierarchies, while providing better control and monitoring than markets. In his seminal work on joint ventures, Kogut (1988) describes the conditions ‘‘best suited to a joint venture’’ as ‘‘high uncertainty over specifying and monitoring performance, in addition to a high degree of asset speciﬁcity.’’ As Tsang (2000) suggests due to the fact that the high degree of asset speciﬁcity rules out market transactions and that the high uncertainty precludes precise speciﬁcations – and hence ex ante stipulations – of the complications and contingencies regarding a) performance monitoring and b) safeguarding against opportunism, a joint venture becomes a viable option since it provides the partners with better alignment of incentives by mutual allocation of resources and sharing the residual value of the venture. However, according to the RBV of the ﬁrm, drivers of joint ventures are two-fold: joint development and mutual exploitation of resources (Tsang, 2000). RBV places particular emphasis on the signiﬁcance of efﬁcient development of resources as well as the efﬁcient utilization of extant resources (Prahalad & Hamel, 1990). As Ohmae (1989) argues in today’s competitive environment, critical know-how content of products and services reached such levels that it is infeasible for many ﬁrms to retain cutting-edge sophistication in all of them, which obligates seamless access to external sources of know-how (Tsang, 2000). Due to the resources required for, and various drawbacks of, acquisitions, they may not be the best way to obtain a certain capability. Moreover, by their very nature, capabilities may not be readily available in the market (Teece, 1986; Kogut & Zander, 1992). These two factors may render a joint venture an optimum choice. According to the RBV, exploitation of complementary resources is another motive behind the joint ventures. Joint ventures entail sharing of

What Causes Break-Ups?

231

complementary resources in order to create synergy (Tsang, 2000). The joint venture between General Motors and Toyota, New United Motor Manufacturing Incorporated (NUMMI) is a textbook example for sharing of complementary resources as each partner contributes to the joint venture with a distinct capability which is not only very costly to develop in-house, but also impossible to obtain from the market. Speciﬁcally, General Motors contributed to NUMMI with its vast distribution network, while Toyota contributed its superior capability to design small cars as well as cutting-edge manufacturing methods. In sum, ﬁrms utilize joint ventures (1) to gain quicker access to requisite skills to perform certain tasks that requires hefty and consequential investments, (2) to share investment risks, and (3) to decrease the total amount they invest (Kogut, 1991). Based on these drivers, Kogut (1991) argues that ‘‘due to its beneﬁts of sharing risk and of reducing overall investment costs, joint ventures serve as an attractive mechanism to invest in an option to expand in risky markets’’ and offers an explanation of joint venture phenomenon based on real options theory. Originating in ﬁnance, real options theory has also received attention from strategic management researchers as well (Reuer & Tong, 2005). According to Kogut and Kulatilaka (2001) real options, which illustrate a ﬁrm’s investments in non-ﬁnancial assets, enable the ﬁrm to safeguard against, and react to, future events, be favorable or unfavorable, in a contingent fashion. Therefore it is suggested that a portfolio of real options provide the ﬁrm with the right, rather than the obligation, to take a ‘‘future speciﬁed’’ action, and hence enabling the ﬁrm not only to decrease loss in case of a unfavorable situation, but also to increase its gains in an event of an unexpected opportunity in the future (Reuer & Tong, 2005; Bowman & Hurry, 1993; McGrath, 1997). Based on this rationale, real options theory represents a viable means to capture the recursive nature of managerial decision-making where managers amend their future strategies consistently with regard to changing market conditions (Trigeorgis, 1996). Joint ventures, from a real options theory perspective, can be regarded as mechanisms that provide the ﬁrm with the ﬂexibility to proactively manage its long- and short-term strategies with regard to market uncertainty and risk (Kogut, 1991). Joint ventures are ‘‘real options’’ because partner ﬁrms can (1) limit their downside losses with the initial investment and/or with a limited amount as they can divest from the joint venture more easily than a wholly owned subsidiary in case of an unfavorable situation and (2) maintain an ‘‘option’’ to expand and increase their investment if the market conditions become unexpectedly favorable (Reuer & Tong, 2005). Kogut (1991) suggests and empirically shows that ﬁrms expand by using

232

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

their option by fully acquiring the joint venture and making it a wholly owned subsidiary. This interest in the formation of joint ventures is indeed expected, when it is considered that the study of the motivation for corporate strategic choices is a milestone in business policy search (Park & Russo, 1996). However, what is unexpected is the relative lack of research on the failure and/or dissolution of these entities, although several studies report high failure rates such as 7 in 10 (Coopers & Lybrand, 1986), 2 in 3 (Auster, 1986; Kogut, 1989), and 1 in 2 (Harrigan, 1988). Based on these records, it is obvious that more thorough analyses should be conducted on the break-ups of IJVs. A fundamental idea about these entities is that they are not expected to last indeﬁnitely. Kogut (1989) argues that dissolution of an IJV reﬂects a business failure or an irresolvable conﬂict among parent ﬁrms. Porter (1987) puts forward that dissolution is signiﬁcant because companies generally do not shut down a successful IJV and dissolution happens only when the company is not ﬁnancially viable. Lane and Beamish (1990) associate dissolution with less tangible adverse outcomes, such as loss of reputation and add that in the long-term dissolution can create political tensions. Franko (1971) also asserts that the reason for dissolution might be the parties being undesirable alliance partners for others or the host country that sponsors the local ﬁrm might not be considered a suitable place for investment. While recent studies have more comprehensive approaches, they maintain the same stance about IJV dissolution with their predecessors. Ring and Van de Ven (1994) suggest that ‘‘it is not only in the economic but also in the psychological best interests of the organizational parties to ﬁnd ways to preserve their socially embedded relationship’’ (p. 107). Park and Ungson (1997) examine the effects of national culture, organizational complementarity, and economic motivation on the instability of IJVs and conclude that opportunistic threat and rivalry appear to be a stronger indication of the dissolution than organizational variables. In addition, Geringer and Hebert (1991) ﬁnd that IJVs that are perceived to be successful by their parent ﬁrms live longer compared to the others evaluated to be unsuccessful. Recently, Dhanaraj and Beamish (2004) search for the relationship between equity ownership and survival of IJVs and conﬁrms the declining, nonlinear, and asymmetrical relationship. Along with the studies that view IJV terminations as failures and undesirable outcomes, another line of scholarly inquiry takes a different stance and posits that the termination of an IJV may be due to a realized or prospective favorable context as well. For instance, Gomes-Casseres (1987) identify various ‘‘adaptive’’ drivers of joint venture termination. Speciﬁcally,

What Causes Break-Ups?

233

he argues that joint ventures may be terminated when (1) one of the partners acquire or develop new capabilities, (2) one of the partners enhance their network which stimulates a change in the ownership structure of the joint venture to be able to exploit new economies of scope, and (3) a change in government occurs. Kogut (1991) reports that unexpected growth in the product market, which is a favorable prospect rather than a failure or undesirable outcome may lead to joint venture termination via acquisition; whereas, unexpected problems may have no effect on the likelihood of termination. While Kogut’s (1991) results reveal that the underlying dynamics of termination are more complex to be simpliﬁed to ‘‘undesirable outcome’’ or ‘‘failure,’’ Makino, Chan, Isobe, and Beamish (2007) enhances this view of termination by integrating the concept of ‘‘intended termination.’’ They also distinguish between intended termination, in which the initial purposes of IJV are either achieved or disappeared; and unintended termination, in which unexpected contingencies emerge that lead to the termination. The critical contribution of Makino et al. (2007) is the idea that the termination of an IJV might be a positive (i.e., intended termination) as well as a negative (i.e., unintended termination) outcome in contrast to most of the previous studies viewing the termination as a result of poor performance, undesirable alliance partners, or organizational failure (e.g., Franko, 1971; Kogut, 1988; Ring & Van de Ven, 1994). Prior to Makino et al. (2007), Park and Russo (1996) differentiate ‘‘failure’’ from instability by stating that, instability, which can be signaled through dissolution and acquisition, can be perceived as a positive outcome if the end status of the IJV is acquisition. In sum, undesirable outcomes and failures are not the only reasons triggering the joint venture terminations since partners may terminate the joint venture due to several reasons including unexpected success, change in ﬁrm strategy, development of new capabilities, and adapting to changes in the environment. In the arguments that follow, our main direction of research is on the dissolution of marketing-oriented IJVs (i.e., the IJVs that are formed with the objective of joint selling, distributing, or promoting goods or services that are either jointly or individually manufactured by the partner ﬁrms). In this study, we do not deﬁne dissolution in terms of failure or bad performance as most previous studies do but conceptualize it as a strategic outcome, which is contingent on the competition–commitment relationships between parent ﬁrms, cultural characteristics of both parties and IJV, and also the host country attractiveness. In doing so, we avoid the questionable practice of associating dissolutions with bad performing IJVs, which is not always the case in current global business world.

234

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

THEORETICAL FRAMEWORK AND HYPOTHESES In this study, we adopt TCA and agency theory perspectives to examine the ﬁrm- and country-level antecedents of IJV dissolution. TCA has received substantial consideration in international economics, management (Anderson & Gatignon, 1986; Dunning, 1988; Caves, 1996), and marketing (Rindﬂeisch & Heide, 1997) ﬁelds. This view posits that IJVs are chosen when the transaction costs involved in an exchange are too high for using market mechanism, but not high enough to form a hierarchy (Hennart, 1991; Kogut, 1988; Teece, 1986). They can be explained as a form of hybrid governance structure that shares the attributes of markets and hierarchies in order to avoid and weaken the hazards of each. They are vulnerable to transaction hazards, which threaten the success of relationships between the parent ﬁrms (Park & Russo, 1996). Firm-Level Antecedents IJVs are designed to meet the goals of each parent and the collective undertaking in an international context, therefore, will be successful when the value of collective outcomes exceeds opportunity costs incurred by the parent ﬁrms, and when the sharing of both is fair (Jarillo, 1988). Given similar strategic objectives and resource bases, a partner can identify, appreciate, and then assimilate another partner’s know-how (Cohen & Levinthal, 1990). In the context of marketing-oriented IJVs, managers – when expanding to foreign countries – can choose to partner with those competitors having local marketing expertise, which can provide access to established distribution channels to market a new product (Beamish, 1993). Furthermore, companies can choose to form marketing collaborations with competitors to pursue short-term integration of similar product and market positioning strategies. Key to IJV survival is striking the right balance between individual competitive motives and cooperation to meet collective interests (Luo & Park, 2004). If this balanced and equitable-contributions system with its beneﬁts and safeguards is jeopardized, so is the IJV. Incentives to deceive can be easily activated and partners may behave opportunistically to attain their own competitive goals instead of the collective goals of the IJV. In this context, we expect that IJVs are more likely to dissolve when the parental goals conﬂict directly (Kogut, 1988) due to the competition between the partner ﬁrms, who both seek to maximize their shareholder value. For example, in a setting where the parents of an IJV are competing

What Causes Break-Ups?

235

in the same market, the opportunistic hazard is inevitable, because present friends can easily become future foes (Harrigan, 1988; Morris & Hergert, 1987). It is difﬁcult to develop cooperation in the face of self-interest and competitive uncertainty (Park & Ungson, 2001). Luo and Park (2004) suggest that without cooperation, mutual forbearance becomes less appealing to the partners, as they lack a long-term view. Therefore, IJV becomes a costly governance structure to arrange inter-ﬁrm transactions due to increased safeguarding mechanisms against opportunistic hazards. Thus Hypothesis 1. International joint ventures are more likely to dissolve when the partner ﬁrms are direct competitors. Agency theory, together with TCA, forms another theoretical basis for this study. The combination of both theories is a good choice in examining the IJV dissolution since they form the theoretical background for the existence of ﬁrms. The ownership conﬁguration of the IJV is also a variable of interest since it reﬂects the control and coordination issues. There has always been a debate on how inequality should affect the dissolution of an IJV. Since equity has been interpreted as a measure of control and dominance, the more absolute control implies the less potential conﬂict as decisions can be made easily by the majority partner (Killing, 1983; Blodgett, 1992). In the context of marketing-oriented IJVs, the ownership and control are particularly important regarding the development of new products for different countries. Especially, in high-technology industries, strategic ﬂexibility to remove potential bureaucratic inferences in different countries has favored dominant equity ownership over equal equity partnerships (Calantone & Zhao, 2000; Ding, 1997). When there exists a dominant ownership of a single parent, the remaining parents have an a priori perception that the achievement of their individual interests may be jeopardized. Therefore, dominant ownership can, paradoxically, simplify the control process. However, the dominant ownership of a parent ﬁrm can cause agency problems. In an IJV, where the dominant parent ﬁrm plays the role of a principal and the other party as an agent, several agency costs might be incurred in the form of either adverse selection or moral hazard. Since these processes create an outcome uncertainty, neither control nor commitment can be maintained easily and therefore the likelihood of IJV dissolution increases. In the literature, there is also evidence that it is possible to control an IJV with an equally shared ownership structure (Geringer & Hebert, 1989; Mjoen & Tallman, 1997; Schaan, 1988). This perspective suggests equity as a measure of commitment and involvement, with the implication that as the ownership is equally shared, the likelihood of instability, allowing for

236

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

a higher level of involvement by the parent ﬁrms (Beamish, 1985; Mjoen & Tallman, 1997). Moreover, the control and commitment interpretations need not be at the expense of another: a control perspective focuses on a singular point of majority ownership, whereas a commitment perspective treats equity as a continuum across the domain of possible ownership levels (Dhanaraj & Beamish, 2004). Thus Hypothesis 2. International joint ventures are more likely to dissolve when the partner ﬁrms do not have an equal equity ownership. The internal uncertainty in an IJV is, in part, a function of the parent ﬁrms’ cultural differences (Reuer, 2001), where culture refers to ‘‘patterns of beliefs and values that are manifested in practices, behaviors, and various artifacts shared by members of an organization or a nation’’ (Hofstede, 1980; Trice & Beyer, 1993). Since organizations are entities embedded in larger societies, it is assumed that the parent ﬁrms of an IJV represent the values and institutions of organized societies in boundaries deﬁned by nationality (Ronen & Shenkar, 1985; Shan & Hamilton, 1991; Park & Ungson, 1997). A study by Child, Markoczy, and Cheung (1994) has shown that national culture affects managerial behavior, philosophy, and moderates the relationship between structural variables and the performance of IJVs. For an IJV to survive, parent ﬁrms should be able to work together, agree on goals, and re-agree on them with respect to ex-post changes in the internal and external environment (Doz, 1996); and these will be more difﬁcult, when IJV partners are from culturally different nations (Hennart & Zeng, 2002). It is suggested that IJVs suffer from commitment and cooperation problems emerging from the differences in the values and behaviors of partners (Harrigan, 1988; Mohr & Spekman, 1994; Parkhe, 1991) that lead to differences in management philosophies, organizational and administrative practices, employee expectations, and responses to strategic issues (Kogut & Singh, 1988; Schneider & De Meyer, 1991). These challenges stem partly from the lack of shared norms or values (Park & Ungson, 1997), which can reduce effective communication (Rao & Schmidt, 1998), trust (Aulakh, Kotabe, & Sahay, 1996; Doney, Cannon, & Mullen, 1998), and knowledge sharing (Parkhe, 1991; Mohr & Spekman, 1994) between parent ﬁrms. This issue becomes even more signiﬁcant in the case of marketing-oriented IJVs where the impact of culture on marketing practices is vital regarding the partner ﬁrms’ understanding of, and approach to, marketing orientation. When entered into an IJV, partner ﬁrms attempt to apply their own marketing policies regarding product differentiation, pricing and promotion, and selling strategies, which hinder the achievement

237

What Causes Break-Ups?

of collective goals. While operating in a foreign market, partner ﬁrms’ standardization versus adaptation policies can also cause some problems regarding the marketing and selling of goods and services in that market. Child et al. (1994) ﬁnd a signiﬁcant relationship between partner ﬁrm nationality and IJV failure whereas Lane and Beamish (1990) conclude that cultural compatibility between partners is the most important factor for the endurance of an alliance relationship. Overall, the uncertainty due to cultural differences makes it costly to negotiate and transfer management philosophies and practices (Pothukuchi, Damanpour, Choi, Chen, & Park, 2002). IJV partners from different national cultures experience greater difﬁculty in their relationships (Lane & Beamish, 1990), which increases the likelihood of IJV dissolution. Thus Hypothesis 3. International joint ventures are more likely to dissolve when the partner ﬁrms have dissimilar management philosophies. Country-Level Antecedents As explained earlier, the impact of cultural similarity between IJV parent ﬁrms on the likelihood of IJV dissolution has been widely studied, and the extant literature unequivocally reports that IJVs between partners from different cultures, ceteris paribus, are more likely to dissolve than the IJVs between partners from similar cultures (Park & Russo, 1996; Park & Ungson, 1997; Hennart & Zeng, 2002). Nonetheless, the impact of the cultural differences between the host country and the IJV partners on the likelihood of IJV termination are yet to be analyzed despite the well-documented fact that this phenomenon is rather diagnostic in foreign market entry decisions (Barkema, Bell, & Pennings, 1996; Reuer & Koza, 2000; Zhao, Luo, & Suh, 2004; Tihanyi, Grifﬁth, & Russell, 2005), subsidiary management (Boyacigiller, 1990; Roth & O’Donnell, 1996), performance (Mjoen & Tallman, 1997; Brouthers, 2002), and the impact of IJV on the shareholder value of the partner companies (Merchant & Schendel, 2000; Reuer, 2001; Hanvanich, Richards, Miller, & Cavusgil, 2003). In the IJV literature, it is suggested that the differences in the ‘‘cultures’’ between the host country and the IJV partners has a complicating impact on the ability of IJV to operate effectively in the host country (e.g., Barkema et al., 1996; Yan & Zeng, 1999). Some of the studies in the extant literature associate international operations of ﬁrms in culturally dissimilar countries with increased intra-organizational conﬂicts and limited realization of organizational plans due to the incongruities in norms, values, and

238

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

institutions in the host country (Tihanyi et al., 2005). For instance, Root (1987) posits that higher information needs coupled with the operating costs from encoding and decoding gaps in languages due to the cultural distance between the partner ﬁrm’s home country and the subsidiary’s host country might lead to higher transaction costs. As such, managers shy away from operating in culturally distant markets due to the considerable difﬁculty in utilization of the extant operating routines (Kogut & Singh, 1988). We expect all these perplexities to have a detrimental impact on IJV. Speciﬁcally, an IJV operated in a culturally distant market entails the additional costs and risks due to the unfamiliarity with the local environment (e.g., differences in morals, manners, laws and regulations, and legal and political practices). Put differently, operating in a culturally distant market increases the ‘‘liability of foreignness,’’ which can be deﬁned as all additional costs a ﬁrm operating in a market overseas incurs that a local ﬁrm would not incur (Hymer, 1976). Thus Hypothesis 4. International joint ventures are more likely to dissolve when the cultural distance between the host country and the IJV partners increases. While expanding to foreign countries, ﬁrms incur with a variety of complications which aggravate the risk of their investments including general political risk (e.g., risks due to the political instabilities), ownership and control risks (e.g., expropriation, intervention), operational risks (e.g., price and inventory control and local content requirements), and transfer risks (e.g., arbitrage risk and remittance control) (Hill, Hwang, & Kim, 1990). These risks reﬂect the ambiguity as to the stability of the current political and economic conditions and government policies, which might prove substantial for the proﬁtability and survival of the investment in that country (Agarwal & Ramaswami, 1992). The entry mode literature submits that ﬁrms choose, ceteris paribus, licensing and joint venture to enter the foreign market, for they enable the ﬁrms to decrease its exposure to the risks associated with investing in a foreign country via limiting their transaction speciﬁc assets, and via enabling them to divest their investment relatively more easily without incurring consequential losses (Agarwal & Ramaswami, 1992; Brouthers, 2002; Hennart, 1991). Put differently, ﬁrms tend to limit their investments in a country with a risky environment despite the fact that this also limits their potential gains from that market (Reuer, 2001). However, countries with low investment risk and high potential, offer greater opportunities for the IJV partners. Such an environment facilitates the extraction of an IJV’s competitive potential (Merchant & Schendel, 2000) and

239

What Causes Break-Ups?

secures the repatriation of earnings as well as expropriation of assets (Agarwal & Ramaswami, 1992). Firms not only have the opportunity to achieve economies of scale (Sabi, 1988) along with a willingness to establish a long-term presence in markets with high potential, but also seek ways to extract as much beneﬁt as possible, which will be hindered in the presence of a partner. Firms in host country markets enter into marketing IJVs as a way of utilizing an established network (Contractor & Lorange, 1988). They seek to beneﬁt from a partner’s sales forces, market information, and local knowledge necessary to operate in the host country. Since future success is contingent on the continued availability of such infrastructure, these ﬁrms may also prefer to internalize this knowledge and acquire their joint ventures (Lele-Pingle, 1998). As such, marketing IJVs located at markets with high attractiveness are expected to be provided with better opportunities in terms of competitive selling, advertising, promotion, etc. in a muniﬁcent environment. Thus Hypothesis 5. International joint ventures are more likely to dissolve when they are located in a country with low market potential. Along with the aforementioned ﬁrm- and country-level antecedents, we also control for possible industry effects that might affect the likelihood of IJV dissolution. Speciﬁcally, we differentiate between IJVs with parents (1) in high- vs. low-tech industries and (2) in manufacturing vs. non-manufacturing industries. When an IJV depends on know-how or technology that cannot be easily codiﬁed and requires intimate human contact for exchange, venture partners can appropriate ﬁrm-speciﬁc competitive advantages (Hamel, Doz, & Prahalad, 1989). We want to examine whether an IJV characterized as a high-tech ﬁrm is more vulnerable compared to the one that involves more discrete contributions like ﬁnancing or physical resources (Teece, 1986). Moreover, we also look at the effect of parental ﬁrms’ coming from a manufacturing industry on the likelihood of IJV dissolution since these industries may differ signiﬁcantly in terms of the types of resources they use, supply chain dynamics, and marketing practices. The conceptual framework of the aforementioned hypothesized relationships is represented in Fig. 1.

DATA AND METHOD Sample This study uses the dissolution of an IJV as the unit of analysis. IJV dissolution data are extracted from Thomson Financial Security

240

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

Inter-partner Competition

Firmrelated variables

Equity Ownership Structure

Management Philosophy

Countryrelated variables

Likelihood of IJV Dissolution

Parent FirmsHost Country Cultural Distance

Host Country Potential

Manufacturing vs. Non-manufacturing Control variables

Fig. 1.

High-tech. vs. Low-tech

Conceptual Framework of the Dissolution of International Joint Ventures.

Database (TFSD). To be included in the study, an IJV has to meet three criteria. First, it has to involve a cross-border investment. A cross-border investment is deﬁned as the one, which involves ﬁrms from the same country forming an IJV in a foreign country, or ﬁrms from different countries forming an IJV either in a third country or in one of the partners’ home country. Second, the joint venture has to involve a marketing aspect (e.g., an IJV with the objective of jointly selling, distributing, and promoting goods

241

What Causes Break-Ups?

or services that are either jointly or individually manufactured). Third, following Park and Ungson (1997), Merchant (2000), Reuer and Koza (2000), Pothukuchi et al. (2002), and many others, IJVs with more than two partners are excluded to be able to more precisely capture and better measure key variables like inter-partner competition and inter-partner cultural differences. A search with these three criteria in TFSD for the period of 1980–2004 yields 5,196 IJVs in 123 different countries. This dataset is reﬁned by excluding observations in which IJV is between a company and an international ﬁnancial institution (e.g., IBRD and UN) and by eliminating IJVs formed in countries for which data are unavailable. For instance, an IJV between US and Liberian companies is excluded since cultural distance is one of the independent variables of this study and cannot be calculated for United States and Liberia because values of Hofstede’s cultural dimensions for Liberia do not exist (Hofstede, 1980). Final sample includes 3,038 observations complete in every aspect. Table 1 represents the yearly distribution of ‘‘non-dissolved’’ and dissolved IJVs. Once the sample has been identiﬁed, country-level data on the economic and demographic characteristics of 63 potential host-country markets have been added to the database. Although there exists more than 200 countries and territories during the sample period (i.e., 1980–2004), 63 countries included in the study represent more than 90 percent of the total domestic product and population of the world during the sample period.

Variables Dependent Variable The dissolution of an IJV is captured by a dichotomous variable which takes the value of 1 if the IJV dissolves during the observation period, and 0 if the IJV is still active at the end of the observation period. Independent Variables We deﬁne two categories for the independent variables: (1) parent ﬁrmrelated variables and (2) country-related variables. (1) Parent ﬁrm-related variables: In the previous section, we posit that an IJV is more likely to dissolve when (1) the competition among the parent ﬁrms diminishes their cooperation, (2) the parent ﬁrms do not have equal share ownership (i.e., 50 percent each) in the IJV, and (3) the partner ﬁrms have dissimilar management philosophies. Competition (COMP) is measured

1,979

1,494

Total

1,544

6 8 7 14 8 79 191 94 235 313 364 183 140 74 85 70 61 45 1

11 12 18 16 19 90 136 71 194 259 313 152 88 34 46 28 35 22

4 6 3 9 6 55 148 76 166 217 258 142 111 67 68 67 49 40 1

Low technology 1

Manufacturing

1

Nonmanufacturing

Formations

1,059

9 10 14 11 17 66 93 53 125 163 207 111 59 27 29 25 23 17

High technology

3,038

15 18 21 25 25 145 284 147 360 476 571 294 199 101 114 95 84 62 1

1

All

47

1 4 1 4 8 2 2 1 10 2 4 2 5 1

Nonmanufacturing

68

3 8 7 1 5 9 1 6 8 4 5 8 3

Manufacturing

79

4 6 2 5 11 5 3 4 14 3 6 7 7 1

1

Low technology

Terminations

Yearly Distribution of IJV Formations and Terminations.

1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Year

Table 1.

37

3 4 3 3 3 1

2 6

6 6

High technology

116

4 12 8 5 13 11 3 7 18 6 9 10 8 1

1

All

242 MEHMET BERK TALAY AND M. BILLUR AKDENIZ

243

What Causes Break-Ups?

by a dummy variable equal to 1 when the 4-digit SIC codes of the parents are equal and 0 otherwise. The ownership structure (OWN) is measured with a dummy variable coded as 1 for IJVs where each parent company has equal equity ownership in IJV, and 0 otherwise. The differences between management philosophies (MANPH) are captured via the cultural distance between the parents of the IJV. Cultural distance (CD) is measured using Kogut and Singh’s (1988) approach, where the distance is deﬁned as the accumulated deviation between the variance-adjusted cultural scores of the IJV parents. Cultural scores were obtained from the scores reported by Hofstede (1980), where cultural sub-dimensions for uncertainty avoidance, individuality, power distance, and masculinity are provided. These scales are used in the following formula: CDjk ¼

4 1X ðI ij I ik Þ2 Vi 4 i¼1

where the left-hand side represents the cultural distance between the country j and k. Each cultural sub-dimension is represented by i, whereas Vi is the sample variance of scale i. Therefore, Iij and Iik denote the score of the parents of the IJV on cultural sub-dimension i, respectively. (2) Country-related variables: We also posit that an IJV is more likely to dissolve when (1) the cultural distance between the host country and the IJV partners increases and (2) the IJV is located in a country with low market potential. Cumulative cultural distance (CUMCD) is measured in the same way as the MANPH is calculated. Using Kogut and Singh’s (1988) method and Hofstede’s cultural dimensions, we ﬁrst calculate the cultural distance between the home country of each parent ﬁrm and the host country (i.e., the country where the IJV was formed), then we add these two cultural distance values to ﬁnd the cumulative cultural distance (CUMCD). According to the OECD (1983), the indicators of market potential (MARKPOT) include the size of the market, the level of GNP, the growth of GNP, manufacturing sales, value added in manufacturing industries, etc. (OECD, 1983). Moreover, a widely cited resource that employs an index approach to measuring market potential is Business International (Cavusgil, 1997), which uses three indicators for world geographic regions and individual counties: (i) market size, (ii) market intensity, and (iii) market growth. A recent example of signiﬁcant factors of a market’s potential are size as measured by the population, economic well-being as measured by GDP per capita, and market dynamism as indicated by the annual GDP growth rate (Cavusgil, Kiyak, & Yeniyurt, 2004). Further

244

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

a number of studies (Kwack, 1972; Ahmed, 1979; Agarwal, 1980) highlight the relationship between market potential and FDI inﬂow. Therefore, we also use the total FDI inﬂow as a percentage of GDP of that country in the year IJV as a proxy for the market potential in the country where the IJV is formed. The data are extracted from the World Development Indicators Database of the World Bank for the period of 1980–2004. Control Variables We control for a series of other factors that may have an effect on the likelihood of dissolution of IJV. In particular, two mutually exclusive dichotomous variables are deﬁned: (1) high- vs. low-tech industries and (2) in manufacturing vs. non-manufacturing industries. Each company is categorized into 11 industry clusters based on their 2-digit SIC codes. Then, two different super-clusters are deﬁned: high tech vs. low tech and manufacturing vs. non-manufacturing. Following the nomenclature of SIC code system, high-technology vs. low-technology classiﬁcation is based on Calantone and Schatzel (2000) in which companies with two digit SIC codes of 28, 35, 36, 37, and 38 are named as ‘‘high tech’’ companies, hence the remaining are ‘‘low tech’’ companies. The companies with 2-digit SIC codes between 20 and 39 are classiﬁed as ‘‘manufacturing’’ companies whereas the remaining are classiﬁed as ‘‘non-manufacturing’’ companies. The descriptive statistics for, and the correlations among, the variables are presented in Table 2.

Method Since our dependent variable is dichotomous, binomial logit model is used to test for the probability of IJV dissolution. The regression coefﬁcients estimate the impact of the independent variables on the probability of IJV dissolution. A positive sign of the coefﬁcient indicates that the variable associated with that coefﬁcient increases the probability of dissolution, and vice versa. The model can also be expressed as: Pðyi ¼ 1Þ ¼

1 1 þ expða X i BÞ

where yi is the dependent variable, Xi the vector of independent variables for the ith observation, a the intercept, and B the vector of regression coefﬁcients (Hastings, 1986).

0.2472 0.6527 2.1730 2.3797 345.7204 12.6781 4.2641 2.3817 0.5086 0.3486 8.1158 6.6686

* indicates po.05.

COMP OWN MANPH CUMCD Population GDP per capita GDP growth FDI inﬂow MANU TECH CUMCD^2 MANPH^2

Mean

0.4315 0.4762 1.3954 1.5665 456.5144 9.7130 5.7124 2.6978 0.5000 0.4766 11.3116 8.2262

Standard Deviation

1.0000 0.0019 0.0375* 0.0449* 0.0011 0.0144 0.0361 0.0024 0.0444* 0.0388* 0.0236 0.0250

1

2

3

4

5

1.0000 0.3509* 0.3251* 0.1714* 0.0829* 0.0695* 0.1124*

6

1.0000 0.4188* 0.1268* 0.1306* 0.0334 0.0448*

7

Descriptive Statistics and Correlations.

1.0000 0.0536* 1.0000 0.0434* 0.8161* 1.0000 0.1166* 0.0623* 0.0317 1.0000 0.1541* 0.2047* 0.1702* 0.5396* 0.0816* 0.0926* 0.0839* 0.5781* 0.0941* 0.0330 0.0564* 0.3598* 0.0477* 0.0861* 0.0760* 0.1785* 0.0613* 0.0971* 0.0889* 0.1714* 0.0141 0.5938* 0.8801* 0.0296 0.0407* 0.8881* 0.7367* 0.0018

Table 2.

1.0000 0.0577* 0.0243 0.0415* 0.0280

8

10

11

12

1.0000 0.7191* 1.0000 0.0361 0.0441* 1.0000 0.0430* 0.0551* 0.6796* 1.0000

9

What Causes Break-Ups? 245

246

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

RESULTS The results of a series of binomial logistic regression for the sample are presented in Tables 3 and 4. The estimated coefﬁcients represent the likelihood of dissolution over non-dissolution: a positive coefﬁcient indicates that the independent variable increases the probability of dissolution, whereas a negative coefﬁcient indicates the opposite. In our ﬁrst model (Model 1), where we test for the linear relationships, we ﬁnd that direct competition between the parent ﬁrms (COMP) has no signiﬁcant effect on the probability of dissolution (b ¼ .273, pW.10), therefore our Hypothesis 1 is rejected. Nevertheless, the relationship we ﬁnd between the equal ownership (OWN) supports Hypothesis 2. In particular, we posit that if the ﬁrms share equal levels of equity ownership, the IJV dissolution will be more likely, and our result suggests that the equal ownership structure in an IJV signiﬁcantly decreases the probability of dissolution (b ¼ 1.288, po.01). We ﬁnd that the difference between management philosophies (MANPH) of the parent companies has no signiﬁcant effect on the likelihood of IJV dissolution (b ¼ .170, pW.10), therefore Hypothesis 3 is rejected. Regarding the country-level variables the results provide partial support to the postulated relationships. We ﬁnd no statistically signiﬁcant evidence to Table 3. Partial Likelihood Estimates of the Covariates on the Likelihood of IJV Termination (Linear Effects). Variables COMP OWN MANPH CUMCD Population GDP per capita GDP growth FDI inﬂow MANU TECH Observations Log-pseudo likelihood LR-w2 df Signiﬁcance

Coefﬁcient

Robust Standard Error

z

PW|z|

0.273 1.289 0.170 0.056 0.001 0.039 0.050 0.073 1.068 0.655

0.226 0.213 0.140 0.124 0.000 0.013 0.032 0.055 0.272 0.277

1.210 6.060 1.210 0.450 2.980 2.940 1.580 1.320 3.930 2.370

0.227 0.000 0.226 0.654 0.003 0.003 0.113 0.187 0.000 0.018

2,324 382.123 87.940 10 0.000

247

What Causes Break-Ups?

Table 4. Partial Likelihood Estimates of the Covariates on the Likelihood of IJV Termination (Linear and Quadratic Effects). Variables COMP OWN MANPH CUMCD Population GDP per capita GDP growth FDI inﬂow MANU TECH MANPH^2 CUMCD^2 Observations Log-pseudo likelihood LR-w2 df Signiﬁcance

Coefﬁcient

Robust Standard Error

z

PW|z|

0.272 1.295 1.083 1.136 0.001 0.041 0.057 0.076 1.066 0.678 0.129 0.195

0.227 0.213 0.539 0.588 0.000 0.013 0.033 0.055 0.272 0.278 0.092 0.101

1.200 6.080 2.010 1.930 3.060 3.040 1.750 1.370 3.920 2.440 1.390 1.920

0.231 0.000 0.044 0.053 0.002 0.002 0.080 0.169 0.000 0.015 0.163 0.054

2,354 379.738 92.710 12 0.000

support Hypothesis 4. Speciﬁcally, the relationship between the cumulative cultural distance of the partners to the host country of the IJV (CUMCD) and the likelihood of IJV dissolution is insigniﬁcant (b ¼ .056, pW.10). However, regarding the market potential (MARKPOT), we ﬁnd that the population of a country has a signiﬁcant negative impact (b ¼ .001, po.01), whereas the GDP per capita has a signiﬁcant positive impact on the probability of dissolution. GDP growth rate and FDI inﬂow have insigniﬁcant effects on IJV dissolution (b ¼ .039, po.01; b ¼ .050, pW.10, respectively). Therefore, Hypothesis 5 is supported for population. As per control variables, we ﬁnd that marketing IJVs in manufacturing industries are more likely to dissolve when compared to their nonmanufacturing counterparts (b ¼ 1.068, po.01). However, the results indicate that the marketing IJVs in high tech industries are less likely to dissolve (b ¼ .655, po.05) than the IJVs in low-tech industries. In our second model (Model 2), we test for the quadratic effects of the differences between the management philosophies (MANPH^2) and the cumulative cultural distance (CUMCD^2). Speciﬁcally, we ﬁnd that the linear effects of the differences between management philosophies (MANPH) are negative and statistically signiﬁcant (b ¼ 1.083, po.05), whereas the

248

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

quadratic effects of the same variable (MANPH^2) has a positive and signiﬁcant effect (b ¼ .129, pW.10). In addition, the results indicate that both the linear (b ¼ 1.136, pW.10) and quadratic (b ¼ .195, pW.10) effects of cumulative cultural distance are signiﬁcant. Positive sign of the linear effect (CUMCD), and the negative sign of the quadratic effect (CUMCD^2) indicate that the relationship between the cumulative cultural distance and the likelihood IJV dissolution can be characterized by an inverted U-shaped curve.

CONCLUSION AND DISCUSSIONS This research examines the impact of a set of ﬁrm- and country-related antecedents, which trigger IJV dissolution. Based on TCA and agency theory frameworks, we analyze the effect of the competition, the share of equity ownership, and the similarity of management philosophies between parent ﬁrms as ﬁrm-related variables; and the cultural distance between parent ﬁrms and the host country and host country potential as countryrelated variables. First, we ﬁnd that there is no signiﬁcant relationship between interpartner competition and the likelihood of IJV dissolution. This result is contrary to what we propose in relation with the TCA literature, which argues that direct competition between partner ﬁrms increases the likelihood of IJV dissolution. However, our ﬁnding implies that competition between parent ﬁrms, although it can create some detrimental impact on (1) exchanging ﬁrm-speciﬁc assets, (2) achieving collective goals of the independent entity, and (3) enabling the formation and maintenance of a trust-based relationship, does not signiﬁcantly contribute to the explanation of IJV dissolution. This unexpected result is even more interesting when interpreted together with our second ﬁnding: the negative impact of equal ownership structure on IJV dissolution. This ﬁnding suggests that the equal equity ownership structure in an IJV signiﬁcantly reduces the likelihood of dissolution. Although there are competing views in the literature regarding the effects of dominant versus equal ownership on the dissolution, our results support that when the ownership is equally shared, the involvement and commitment of parent ﬁrms for the achievement of the collective goals of the independent entity increase, and the likelihood of IJV dissolution decreases. Furthermore, since this structure leads to an increasing trust between partners, moral hazard decreases leading to reduced safeguarding costs as well. This highly signiﬁcant result along with the insigniﬁcant result

What Causes Break-Ups?

249

of the impact of competition between parent ﬁrms can lead to a notable argument in the IJV literature. It can be asserted that commitment between parent ﬁrms tend to be more important than the competition in explaining the destiny of the IJV. Ceteris paribus, the extent of the parent ﬁrms’ cooperation efforts, regardless of the competitive nature, makes the difference in the likelihood of IJV dissolution. In a similar fashion, it might also be argued that parent ﬁrms tend to emphasize commitment over competition, which suggests that even the imperfect cooperation between competitors (Brandenburger & Nalebuff, 1996) can affect the end outcome of inter-ﬁrm collaborations signiﬁcantly. The third result is the insigniﬁcant relationship between management philosophy and the dissolution of IJV. In the IJV literature, cultural differences between parent ﬁrms constitute a major block in the development of successful relationship whereas our result challenges this view. A plausible explanation might be the difference between our study and previous research in terms of the approach to the dissolution of IJVs. One commonly held perspective in most of the previous empirical studies is their conceptualization of the dissolution from a failure or poor performance point of view. However, as noted in the beginning, in line with the current global business world, we delineate dissolution as the partner ﬁrms’ strategic choice to bring the IJV to an end that can represent either a poor performance or a successful outcome. According to this conceptualization, an IJV relationship can dissolve when it accomplishes the targets set at the time of its inception or when the partners stop considering the host country as a potential market to invest due to some emerging contingencies; and under these circumstances, the similarity of partner ﬁrm cultures might not play a signiﬁcant role any more. Another interesting result of this research is the inverted U-shaped relationship of the cultural distance between the nationalities of the IJV parent ﬁrms and the host country, which reinforces the role of cultural distance as an important covariate. Cultural distance represents the extent of dissimilarities between the IJV parents and the cultural context in the host country. These differences increase the transaction costs since IJV parent ﬁrms should bear extra costs and risks of unacquaintedness to differences in morals, manners, laws and regulations, and legal and political practices, along with an increased ‘‘liability of foreignness’’ in the host country. However, the negative quadratic effect found in the analyses reveals a diminishing impact of cultural distance, suggesting that the perplexities due to high cultural distance are likely to be less inimical to the dissolution of the IJV. This may be attributed to an increase in the awareness of IJV parents

250

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

towards possible ramiﬁcations of cultural distance as they operate in a country where culture is very different. In addition, some managerial practices will also be different as the cultural distance increases. As such, various business routines (e.g., distribution, advertising), which are considered as ‘‘norms’’ in one culture, might be considered impractical and even be censored in another culture (Hofstede, 1980, 1991; Parkhe, 1991). As an example, in a marketing IJV, partner ﬁrms and local entities in the host country might approach marketing issues and problems differently due to dissimilarities in their marketing orientation. However, operating in a culturally similar country, IJV partners might have the opportunity to work with local entities, which have analogous attitudes and similar business routines and therefore they will experience fewer, if any, communication difﬁculties. Obviously, the better IJV partners and local entities understand each other, the lower the likelihood of IJV dissolution since lower cultural distance between the IJV partners and the host country will not only facilitate better communication, but also enable more stable and long-term relationships. Finally, we ﬁnd that there is a signiﬁcant and positive relationship between the market potential and the likelihood of IJV dissolution. This phenomenon may be attributed several reasons. First, parent ﬁrms may wish to reap the beneﬁts of the host country alone, instead of sharing them with a partner. Second, markets with high potential may also have easy-to-handle bureaucratic procedures and incentives, which diminish the investment risk as well as the need for a partner. For example, the IJV between GE Lighting and PT Sinar Angkasa Rangkut of Indonesia formed in June 1994 was terminated in 1995 when GE announced that it obtained 100 percent ownership of the joint venture and the new wholly owned company was named PT GE Lighting Indonesia. John Opie, the president and the chief executive ofﬁcer of GE said that ‘‘Indonesia offers signiﬁcant opportunity, which is important to GR Lighting’s overall Asia-Paciﬁc growth strategy and gaining 100 percent ownership provides us the ﬂexibility necessary to meet the needs of the rapidly growing Indonesian and Asian lighting markets.’’ Therefore, marketing IJVs located at markets with high potential are provided lucrative opportunities such as competitive selling, advertising, and promotion that sometimes partner ﬁrms might choose to terminate these relationships just to get the beneﬁts of these markets individually. The negative, albeit insigniﬁcant, relationship between FDI inﬂow, which may be regarded as a proxy of economic, political, and investment risks, and parents’ propensity to dissolve the IJV supports this argument. Besides, as predicted by TCA, lower levels of such risks decrease

What Causes Break-Ups?

251

the transaction costs, and hence facilitates hierarchical forms instead of arm’s length transactions. Overall, the ﬁndings of this study have speciﬁc methodological vantages as well. First, contrary to most of the studies, which are limited to the IJVs that involve at least one U.S. partner ﬁrm, this study examines these relationships with a more comprehensive dataset, which also includes IJVs formed by non-U.S. companies within the period of 1980–2004. Using only U.S.-based ﬁrms have long been acknowledged as a limitation in the IJV literature since IJV-related phenomena are as relevant to other companies as they are to U.S. companies (e.g., Park & Ungson, 1997; Merchant, 2000). Therefore, we believe that examining IJVs involving non-U.S. ﬁrms enhances our understanding of IJV dissolution. Second, our study is one of the few that report the quadratic effects of some factors on the likelihood of IJV dissolution. Analyzing the quadratic effects helps better capture the dynamics of IJV-related phenomena since it enables to detect any ‘‘ceiling effects’’ and optimal values. Third, as a result of using a vast dataset spanning 21 years and 63 countries, we also take into account the effect of the national culture similarity of partner ﬁrms and host country in addition to inter-partner cultural similarity as an indicator of IJV termination, since we have many IJVs formed by two partners from different countries (e.g., German and French ﬁrms) in another host country (e.g., China). As with any other study, this study has some limitations, which open further avenues of research. In this study, we only focus on marketing IJVs, which imply a more tacit and idiosyncratic knowledge transfer between partner ﬁrms. This aspect contributes to reﬁne the analyses of our hypotheses via utilizing a comprehensive dataset comprising a signiﬁcant portion of IJVs formed within the observation period. However, in the future alternative researches might consider the dynamics of other IJVs formed with different motivations (e.g., manufacturing IJVs or R&D IJVs). Second, using Hofstede’s cultural scores has its own ramiﬁcations. Hofstede’s cultural scores dimensions have not been updated since 1983, which raises serious doubts about their validity in the contemporary context. Moreover, Hofstede’s dimensions are based on the survey of the employees of a single company, which may not represent the characteristics of the entire population. Third, we employ an objective measure of dependent and independent variables since we derive our dataset from secondary sources. However, in the IJV literature, using objective versus subjective measures has always been debated (Park & Ungson, 1997; Parkhe, 1991, 1993). So, future research can examine ‘‘the inner workings,’’

252

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

such as organizational complementarities and managerial basis of the IJVs including strategic ﬁt between partner ﬁrms by using primary data.

REFERENCES Agarwal, J. P. (1980). Determinants of foreign investment: A survey. Review of World Economics, 116(4), 739–773. Agarwal, S., & Ramaswami, S. N. (1992). Choice of foreign market entry mode: Impact of ownership, location and internalization factors. Journal of International Business Studies, 23(1), 1–27. Ahmed, A. A. (1979). Empirical determinants of manufacturing direct foreign investment in developing countries. Economic Development and Cultural Change, 27, 751–767. Anderson, E. (1990). Two ﬁrms, one frontier: On assessing joint venture performance. Sloan Management Review, 31(2), 19–30. Anderson, E., & Gatignon, H. (1986). Modes of foreign entry: A transaction cost analysis and propositions. Journal of International Business and Studies, 17(3), 1–26. Aulakh, P. S., Kotabe, M., & Sahay, A. (1996). Trust and performance in cross-border marketing partnerships: A behavioral approach. Journal of International Business Studies, 27(5), 1005–1032. Auster, E. R. (1986). International corporate linkages: Dynamic forms in changing environments. Academy of Management Review, 8, 567–587. Barkema, H. G., Bell, J. H. J., & Pennings, J. M. (1996). Foreign entry, cultural barriers and learning. Strategic Management Journal, 17, 151–166. Beamish, P. W. (1985). The characteristics of joint ventures in developed and developing countries. Columbia Journal of World Business, 20(3), 13–19. Beamish, P. W. (1993). The characteristics of joint ventures in the people’s republic of China. Journal of International Marketing, 1(2), 29–48. Beamish, P. W., & Berdrow, I. (2003). Learning from IJVs: The unintended outcome. Long Range Planning, 36(3), 285–303. Bleeke, J. A., & Ernst, D. (1993). Collaborating to compete. New York, NY: Wiley. Blodgett, L. L. (1992). Factors in the instability of international joint ventures: An event history analysis. Strategic Management Journal, 13(6), 475–481. Bowman, E. H., & Hurry, D. (1993). Strategy through the options lens: An integrated view of resource investments and the incremental-choice process. Academy of Management Review, 18(4), 760–782. Boyacigiller, N. A. (1990). The role of expatriates in the management of interdependence, complexity, and risk in multinational corporations. Journal of International Business Studies, 21(3), 357–381. Brandenburger, A., & Nalebuff, B. (1996). Co-opetition: A revolution mindset that combines competition and cooperation. New York, NY: Bantam Doubleday Dell. Brouthers, K. (2002). Institutional, cultural and transaction cost inﬂuences on entry mode choice and performance. Journal of International Business Studies, 33(2), 203–221.

What Causes Break-Ups?

253

Calantone, R. J., & Schatzel, K. E. (2000). Strategic foretelling: Communication-based antecedents of a ﬁrm’s propensity to preannounce. Journal of Marketing, 64(1), 17–30. Calantone, R. J., & Zhao, Y. S. (2000). Joint ventures in China: A comparative study of Japanese, Korean, and US partners. Journal of International Marketing, 9(1), 1–23. Caves, R. E. (1996). Multinational enterprise and economic analysis (2nd ed.). New York: Cambridge University Press. Cavusgil, S. T. (1997). Measuring the potential of emerging markets: An indexing approach. Business Horizons, 40(1), 87–91. Cavusgil, S. T., Kiyak, T., & Yeniyurt, S. (2004). Complementary approaches to preliminary foreign market opportunity assessment: Country clustering and country rankings. Industrial Marketing Management, 33(7), 607–617. Child, J., Markoczy, L., & Cheung, T. (1994). Managerial adaptation in Chinese and Hungarian strategic alliances with culturally distinct partners. In: S. Stewart (Ed.), Joint ventures in the people’s republic of China (Vol. 4, pp. 211–231). Greenwich, CT: JAI Press. Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35, 128–152. Contractor, F., & Lorange, P. (1988). Why should ﬁrms cooperate? The strategy and economics for cooperative ventures. In: F. Contractor & P. Lorange (Eds), Cooperative strategies in international business (pp. 3–30). Lexington, MA: D. C. Heath. Coopers and Lybrand. (1986). Collaborative ventures: An emerging phenomenon in information technology. New York, NY: Coopers and Lybrand. Dhanaraj, C., & Beamish, P. W. (2004). Effect of equity ownership on the survival of international joint ventures. Strategic Management Journal, 25(3), 295–302. Ding, D. Z. (1997). Control, conﬂict, and performance: A study of US–Chinese joint ventures. Journal of International Marketing, 5(3), 31–45. Doney, P. M., Cannon, J. P., & Mullen, M. R. (1998). Understanding the inﬂuence of national culture on the development of trust. Academy of Management Review, 23(3), 601–620. Doz, Y. L. (1996). The evolution of cooperation in strategic alliances: Initial conditions or learning processes? Strategic Management Journal, 17(2), 55–83. Dunning, J. H. (1988). The eclectic paradigm of international production: A restatement and some possible extensions. Journal of International Business and Studies, 19, 1–31. Franko, L. G. (1971). Joint venture survival in multinational corporations. New York, NY: Praeger. Geringer, J. M. (1988). Joint venture partner selection: Strategies for developed countries. Westport, CT: Quorum Books. Geringer, J. M., & Hebert, L. (1989). Control and performance of international joint ventures. Journal of International Business Studies, 20(2), 235–254. Geringer, J. M., & Hebert, L. (1991). Measuring performance of international joint ventures. Journal of International Business Studies, 22(2), 249–263. Gomes-Casseres, B. (1987). Joint venture instability: Is it a problem? Columbia Journal of World Business, 22(2), 97–107. Hamel, G., Doz, Y. L., & Prahalad, C. K. (1989). Collaborate with your competitors – And win. Harvard Business Review, 67(1), 133–139. Hanvanich, S., Richards, M., Miller, S. R., & Cavusgil, S. T. (2003). An event study of the effects of partner and location cultural differences in joint ventures. International Business Review, 12(1), 1–16. Harrigan, K. R. (1987). Strategic alliances: Their new role in global competition. Columbia Journal of World Business, 22(2), 67–69.

254

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

Harrigan, K. R. (1988). Strategic alliances and partner asymmetries. In: F. J. Contractor & P. Lorange (Eds), Cooperative strategies in international business. Lexington, MA: Lexington Books. Hastings, R. P. (1986). Supplemental library user’s guide. Version 5 ed. SAS Institute, Cary, NC. Hau, L. N., & Evangelista, F. (2007). Acquiring tacit and explicit marketing knowledge from foreign partners in IJVs. Journal of Business Research, 60, 1152–1165. Hennart, J.-F. (1988). A transaction costs theory of equity joint ventures. Strategic Management Journal, 4, 362–374. Hennart, J.-F. (1991). The transactions cost theory of joint ventures: An empirical study of Japanese subsidiaries in the United States. Management Science, 37, 483–497. Hennart, J.-F., & Zeng, M. (2002). Cross-cultural differences and joint venture longevity. Journal of International Business Studies, 33(4), 699–716. Hill, C., Hwang, P., & Kim, C. (1990). An eclectic theory of the choice of international entry mode. Strategic Management Journal, 11(2), 117–128. Hofstede, G. (1980). Culture’s consequences: International differences in work-related values. London: Sage. Hofstede, G. (1991). Cultures, organizations: Software of the mind. London: McGraw Hill. Hymer, S. H. (1976). The international operations of national ﬁrms: A study of direct foreign investment. Boston, MA: MIT Press. Jarillo, J. C. (1988). On strategic networks. Strategic Management Journal, 9, 31–41. Killing, J. P. (1983). Strategies for joint venture success. New York, NY: Praeger. Kogut, B. (1988). Joint ventures: Theoretical and empirical perspectives. Strategic Management Journal, 9(4), 319–332. Kogut, B. (1989). The stability of joint ventures: Reciprocity and competitive rivalry. Journal of Industrial Economics, 38(2), 183–198. Kogut, B. (1991). Joint ventures and the option to expand and acquire. Management Science, 37(1), 19–32. Kogut, B., & Kulatilaka, N. (2001). Capabilities as real options. Organization Science, 12(6), 744–758. Kogut, B., & Singh, H. (1988). The effect of national culture on the choice of entry model. Journal of International Business Studies, 19, 411–432. Kogut, B., & Zander, U. (1992). Knowledge of the ﬁrm, combinative capabilities, and the replication of technology. Organization Science, 3(3), 383–397. Kwack, S. Y. (1972). A model of US direct investment abroad: A neoclassical approach. Western Economic Journal, 10, 373–383. Lane, H. W., & Beamish, P. W. (1990). Cross-cultural cooperative behavior in joint ventures in LDCs. Management International Review, 30(Special issue), 87–102. Lane, P. J., Salk, J. E., & Lyles, M. A. (2001). Absorptive capacity, learning, and performance in international joint ventures. Strategic Management Journal, 22, 1139–1161. Lele-Pingle, S. (1998). Determinants of duration in marketing versus manufacturing international joint ventures. Ph.D. Dissertation, Purdue University, United States – Indiana. (Publication No. AAT 9921097). Lu, J. W., & Beamish, P. W. (2006). Partnering strategies and performance of SMEs’ international joint ventures. Journal of Business Venturing, 21(4), 461–486. Luo, Y., & Park, S. Ho. (2004). Multiparty cooperation and performance in international equity joint ventures. Journal of International Business Studies, 35, 142–160.

What Causes Break-Ups?

255

Makino, S., Chan, C. M., Isobe, T., & Beamish, P. W. (2007). Intended and unintended termination of international joint ventures. Strategic Management Journal, 28, 1113–1132. McGrath, R. G. (1997). A real options logic for initiating technology positioning investments. Academy of Management Review, 22(4), 974–996. Merchant, H. (2000). Conﬁgurations of international joint ventures. Management International Review, 40(2), 107–140. Merchant, H., & Schendel, D. (2000). How do international joint ventures create shareholder value? Strategic Management Journal, 21(7), 723–777. Mjoen, H., & Tallman, S. (1997). Control and performance in international joint ventures. Organization Science, 8(3), 257–274. Mohr, J., & Spekman, R. (1994). Characteristics of partnership success: Partnership attributes, communication behavior, and conﬂict resolution techniques. Strategic Management Journal, 15, 135–152. Morris, D., & Hergert, M. (1987). Trends in international collaborative agreements. Columbia Journal of World Business, 22(2), 15–21. OECD (Organization for Economic Co-operation and Development). (1983). Investment incentives and disincentives and the international investment process. OECD, Paris. Ohmae, K. (1989). The global logic of strategic alliances. Harvard Business Review, 67(2), 143–155. Park, S. Ho., & Russo, M. V. (1996). When competition eclipses cooperation: An event history analysis of joint venture failure. Management Science, 42(6), 875–890. Park, S. Ho., & Ungson, G. R. (1997). The effect of national culture, organizational complementarity, and economic motivation on joint venture dissolution. Academy of Management Journal, 40(2), 279–307. Park, S. Ho., & Ungson, G. R. (2001). Rethinking the limits of cooperation: A conceptual framework of strategic alliance failure. Organization Science, 12(1), 37–53. Parkhe, A. (1991). Interﬁrm diversity, organizational learning, and longevity. Journal of International Business Studies, 22(4), 579–601. Parkhe, A. (1993). Strategic alliance structuring: A game theoretic and transaction cost examination of interﬁrm cooperation. Academy of Management Journal, 36, 794–829. Porter, M. E. (1987). From competitive strategy to cooperative strategy. Harvard Business Review, 65(3), 43–59. Pothukuchi, V., Damanpour, F., Choi, J., Chen, C. C., & Park, S. Ho. (2002). National and organizational culture differences and international joint venture performance. Journal of International Business Studies, 33(2), 243–265. Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Harvard Business Review, 68(3), 79–87. Rao, A., & Schmidt, S. M. (1998). A behavioral perspective on negotiating international alliance. Journal of International Business Studies, 29, 665–693. Reuer, J. J. (2001). From hybrids to hierarchies: Shareholder wealth effects of joint venture partner buyouts. Strategic Management Journal, 22(1), 27–44. Reuer, J. J., & Koza, M. P. (2000). Asymmetric information and joint venture performance: Theory and evidence for domestic and international joint ventures. Strategic Management Journal, 21(1), 81–88. Reuer, J. J., & Tong, T. W. (2005). Real options in international joint ventures. Journal of Management, 31(3), 403–423.

256

MEHMET BERK TALAY AND M. BILLUR AKDENIZ

Rindﬂeisch, A., & Heide, J. B. (1997). Transaction cost analysis: Past, present, and future applications. Journal of Marketing, 61, 30–54. Ring, P. S., & Van de Ven, A. H. (1994). Developmental processes of cooperative interorganizational relationships. Academy of Management Review, 19, 90–118. Ronen, S., & Shenkar, O. (1985). Clustering countries on attitudinal dimensions: A review and synthesis. Academy of Management Review, 10(3), 435–454. Root, F. R. (1987). Entry strategies for international markets (Revised ed.). New York, NY: Lexington Books. Roth, K., & O’Donnell, S. (1996). Foreign subsidiary compensation strategy: An agency theory perspective. Academy of Management Journal, 39(3), 678–703. Sabi, M. (1988). An application of the theory of foreign direct investment to multinational banking in LDCs. Journal of International Business Studies, 19(3), 433–448. Schaan, J.-L. (1988). How to control joint venture even as a minority partner. Journal of General Management, 14(1), 4–16. Schneider, S. C., & De Meyer, A. (1991). Interpreting and responding to strategic issues: The impact of national culture. Strategic Management Journal, 12, 307–320. Shan, W., & Hamilton, W. (1991). Country-speciﬁc advantage and international cooperation. Strategic Management Journal, 12(6), 419–432. Shenkar, O., & Li, J. (1999). Knowledge search in international cooperative ventures. Organization Science, 10(2), 134–143. Teece, D. J. (1986). Proﬁting from technological innovation: Implications for integration, collaboration, licensing and public policy. Research Policy, 15, 285–305. Tihanyi, L., Grifﬁth, D. A., & Russell, C. J. (2005). The effect of cultural distance on entry mode choice, international diversiﬁcation, and MNE performance: A meta-analysis. Journal of International Business Studies, 36(3), 270–280. Trice, H. M., & Beyer, J. M. (1993). The cultures of work organizations. Englewood Cliffs, NJ: Prentice Hall. Trigeorgis, L. (1996). Real options: Managerial ﬂexibility and strategy in resource allocation. Cambridge, MA: MIT Press. Tsang, E. W. K. (2000). Transaction cost and resource-based explanations of joint ventures: A comparison and synthesis. Organization Studies, 21(1), 215–242. Williamson, O. E. (1991). Strategizing, economizing, and economic organization. Strategic Management Journal, 12(1), 75–94. Wong, Y. Y., Maher, T. E., & Luk, T. K. (2002). The hesitant transfer of strategic management knowledge to international joint ventures in China: Greater willingness seems likely in the future. Management Research News, 25(1), 1–16. Yan, A., & Zeng, M. (1999). International joint venture instability: A critique of previous research. Journal of International Business Studies, 30(2), 397–414. Yip, G. S. (1992). Total global strategy. Englewood Cliffs, NJ: Prentice Hall. Zander, U., & Kogut, B. (1995). Knowledge and the speed of transfer and imitation of organizational capabilities: An empirical test. Organization Science, 6(1), 76–92. Zhao, H., Luo, Y., & Suh, T. (2004). Transaction cost determinants and ownership-based entry mode choice: A meta-analytical review. Journal of International Business Studies, 35(6), 524–544.

GEOGRAPHIC MARKET DIVERSIFICATION: A PREMIUM OR DISCOUNT IN FIRM’S VALUE Z. Seyda Deligonul Geographic diversiﬁcation has been a topic of elusive conclusions in different research streams. International marketing and business studies strongly suggest that there is value in market scope expansion across national boundaries. In contrast, ﬁnance studies doubt the value and typically ﬁnd diversiﬁcation as value-destructive.1 Studies by Lang and Stulz (1994), Berger and Ofek (1995), and Servaes (1996) show unequivocally that diversiﬁed ﬁrms trade at a discount relative to nondiversiﬁed ﬁrms. Their ﬁndings are conﬁrmed to be robust for different countries (Campa & Kedia, 2002; Lins & Servaes, 1999). To add to the pessimism, Errunza, Hogan, and Hung (1999) show that even a private investor can achieve better diversiﬁcation than foreign direct investments, by forming a home-made portfolio from domestically traded foreign securities. There are several varieties of diversiﬁcation: product, market, geography, and various possible combinations of these. Furthermore, diversiﬁcation can be horizontal as in the typical case where technology-transfer is involved or vertical as exempliﬁed by resource-seeking investments. Different levels of combinations are also possible. For example, a ﬁrm can diversify globally with a diversiﬁed line of products or regionalize its single product line in a restricted geography. In our discussion, we examine cross-border geographic market expansion and its possible augmentation by product diversiﬁcation. New Challenges to International Marketing Advances in International Marketing, Volume 20, 257–274 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020013

257

258

Z. SEYDA DELIGONUL

The main question is: why is there controversy regarding the outcomes of diversiﬁcation? The concern is shared by many. Scholars ﬁnd the results ‘conﬂicting’ (Annavarjula & Beldona, 2000), ‘inconclusive’ (Tallman & Li, 1996), ‘mixed’ (Hitt, Hoskisson, & Kim, 1997; Thomas & Eden, 2004), ‘inconsistent’ (Ruigrok & Wagner, 2003), ‘contradictory’ (Contractor, 2007; Ruigrok & Wagner, 2003), ‘conﬂicting’ (Annavarjula & Beldona, 2000), and ‘disappointing’ (Hennart, 2007). This study draws on the extant research to consider possibilities and generates an overview and commentary. Our analysis will not resolve the empirical issue of creating or destroying value by geographic spread, but will consider why there is a strong inconsistency in ﬁndings. The goal is to contribute to better theorizing and design of more robust studies. For a review of diversiﬁcation and its value there are extensive reviews (e.g., Martin & Sayrak, 2003; Contractor, Kundu, & Hsu, 2003; Hitt, Bierman, Uhlenbruck, & Shimizu, 2006; Lu & Beamish, 2004). The rest of the chapter is organized in six sections. The ﬁrst section reports from present literature, the background information regarding the geographic diversiﬁcation with its rewards and burdens. In the next two sections extant ﬁndings are provided. Finally, the author presents the potential challenges of the research in this area and closes the discussion with the pertinent sections on recommendations and conclusions.

DIVERSIFICATION DECISION AND ITS OUTCOMES To explain why international market diversiﬁcation is a viable strategy, a substantial portion of the past literature hinges its conclusions on mainstay perspectives. Some authors utilize internalization theory and transaction cost analysis (e.g., Teece). Others draw from the resource-based explanation of the ﬁrm (e.g., Chang, 1995), institutional theory (e.g., Davis, Desai, & Francis, 2000), organizational learning (Ruigrok & Wagner, 2003); a combination approach (e.g., Madhok, 1997) or eclectic paradigm (Dunning, 1988). These perspectives are widely discussed in the literature. For that reason we present the earlier work only in a brief summary.

Costs of International Market Diversiﬁcation The disadvantages of diversiﬁcation are typically attributed to a long list of reasons. Speciﬁcally, the author’s refer to inefﬁcient allocation of

Geographic Market Diversiﬁcation

259

resources (Lamont, 1997; Rajan, Servaes, & Zingales, 2000), the difﬁculty of designing optimal incentives such as compensation systems (Aron, 1988; Rotemberg & Saloner, 1994), and Harris, Kriebel, and Raviv (1982) underscore the information asymmetries between headquarters and divisional managers. Scharfstein and Stein (2000) discuss incentives for rent-seeking by managers. Jensen (1988) argues for the link between cash payout to shareholders and managers’ value-destroying investments. Denis, Denis, and Sarin (1997) provide empirical evidence of the agency’s costs-of-diversiﬁcation strategy. Given these burdens a ﬁrm will not see any incentive to expand geographically. The diversiﬁcation will handicap the ﬁrm by offsetting costs of acquiring quasi-monopoly power and transferring rents away from focused rivals in the foreign markets. Equally, the ﬁrm will be disadvantaged at home as it accrues additional costs from foreign afﬁliations. Despite this unfavorable logic, many ﬁrms chose to geographically diversify and implicit in this decision are expected advantages from market scope expansion. The potential gains are discussed in the next paragraph.

Gains from International Market Diversiﬁcation The literature favors diversiﬁcation for mostly economic reasons and its potential for arbitrage opportunities. Early studies refer to gains including economies of scale (Chandler, 1977), market power, and increased debtcarrying capacity (Lewellen, 1971). Wernerfelt and Montgomery (1988) underscore ﬁrm-speciﬁc assets that can be transferred and exploited in other markets. Hadlock, Ryngaert, and Thomas (2001) argue for a ﬁrm’s gain from alleviating an adverse selection problem. Schoar (2002) concludes a higher productivity of a diversiﬁed ﬁrm than others within their industry in spite of their discounted status. Also, the diversiﬁcation is recognized as a strategy to enable a ﬁrm to internalize market failures (Khanna & Palepu, 2000), to capitalize on tax beneﬁt possibilities (Errunza & Senbet, 1981), activity sharing, and multipoint competition opportunities (Goerzen & Beamish, 2003; Kogut & Zender, 1993). Firms through diversiﬁcation facilitate a more efﬁcient resource allocation given the beneﬁts of internal capital markets (Stein, 1997). Put together, economic advantage and disadvantage of geographic diversiﬁcation leads to an explanation based on the premise of utilization of excess resources. In economics, excess resources include physical, ﬁnancial, and knowledge-based assets and capabilities; they have zero

260

Z. SEYDA DELIGONUL

or near zero opportunity costs, which creates incentives for the ﬁrm to move them freely. The ﬁrm looks for their alternative utilization through geographic, product, market or country strategies. The goal is to use them with low-risk ends at zero marginal cost. Hence the ﬁrm diversiﬁes. This logic provides an explanation for the diversiﬁcation, but does not explain the valuation discount stemming from geographic spread; on the contrary it only conﬁrms it.2

A GENERAL MODEL OF GEOGRAPHIC DIVERSIFICATION AND PERFORMANCE RELATIONSHIP The clashing ﬁndings about value destruction or premium may stem from the understanding of the performance phenomenon. For that reason, it may be a good attempt to review the extant understanding of the topic. By consolidating the literature, a general model of geographic diversiﬁcation and its performance consequence is given in Fig. 1. This ﬁgure consolidates the frameworks in the extant studies and displays the results in the ﬂavor or structural-equation modeling. Below we discuss the model with its deﬁning characteristics.

Hierarchical Antecedents (Individual to Country Level Factors) One of the characteristics in Fig. 1 and also in the extant research is the offering of multiple idiosyncratic factors as drivers of the diversiﬁcation. From these, the exogenous types are divided between country- and ﬁrmspeciﬁc categories. (The respective bubbles in Fig. 1 are expanded in Fig. 2.) In addition to these factors authors propose a host of additional antecedents for diversiﬁcation. Included in the set are market potential, host-country risk, cultural distance between the home and host markets (country), and institutional (government) transparency, restrictions as to foreign ownership, and intellectual-property protection. Similarly, a rich array of ﬁrm characteristics has been examined. Typically, their discussion is hinged on the ﬁrm’s assets, ﬁrm size, R&D, and marketing expenditures. The level of commitment and various types of experience are also part of this discussion.

261

Geographic Market Diversiﬁcation

Others antecedents ROS

Country factors Institutional factors

ROE External Factors Industry / Country

Industry factors

Performance in Retrospect

ROA

Count Measures Composite Measures

Market scope diversification (depth / scope)

Value outcome Discount / Premium

Elemental Measures Forward looking performance Individual level factors

Treynor

Firm factors Jensen

Group level factors Sharp

Firm level Others (capabilities, commitment etc.)

Fig. 1.

Meta Analytic SEM Model to Summarize Past Findings.

Representing and Measuring the International Market Diversiﬁcation The study attempts to measure market diversiﬁcation both with its depth and scope.3 Many convenient measures are speciﬁed for geographic spread and the extent of asset commitments at business units. Such measures involve a host of proxy variables with the most popular being the number of foreign ofﬁces divided by the total number of ofﬁces (NNS), the number of nations in which a ﬁrm has foreign subsidiaries (NFS), and the total number of subsidiaries abroad. Additionally, more comprehensive indices as well as

262

Z. SEYDA DELIGONUL

1. INDIVIDUAL LEVEL FACTORS 1.1 CEO experience (+) 1.2 CEO compensation (+) 1.3 CEO international experience (+)

TMT International Experience (+) TMT size (+) TMT diversity (+) TMT tenure (+) TMT education (+) TMT age (+)

FIRM FACTORS

2. TEAM LEVEL FACTORS 2.1 2.2 2.3 2.4 2.5 2.6

3 . FIRM LEVEL FACTORS 3.1 Size (+) 3.2 R&D intensity (+) 3.3 Advertising intensity (+) 3.4. Age (+) 3.5 International experience (+) 3.6 Relational capital (+) 3.7 Capital intensity (+) 3.8 Firm human capital (+) 4. INDUSTRY LEVEL FACTORS

5. COUNTRY LEVEL FACTORS 5.1 5.2 5.3 5.4

Host-country risk (–) Favorability of home environment (–) Cultural diversity (–) Favorability of host environment (+)

EXTERNAL FACTORS

4.1 Industry growth (+) 4.2 Industry competition (+) 4.3 Industry profitability (–)

6. INSTITUTIONAL FACTORS 6.1 Favorability of regulative aspects (+) 6.2 Favorability in the cognitive/cultural (+) 6.3 Favorability of normative aspects (+)

Fig. 2. Expanded View of Antecedent Variables Used in the Literature (These are Displayed Only With Main Titles in Fig. 1 for Brevity). Note: In this list the boxes numbered 1,2,3,4,5 are taken from Kirca et al. (forthcoming).

Geographic Market Diversiﬁcation

263

composite indices are offered (e.g., Capone, Hulber, Farley, & Martin, 1988). As an example of the latter kind, the Hirschman index gives more weight and the entropy index gives less weight to the larger businesses. The composite and weighted proxies have received considerable attention with their validity issues in the literature (Hoskisson, Hitt, Johnson, & Moesel, 1993; Boyd, Gove, & Hitt, 2005).

Measuring the Performance Performance is typically measured in two ways. The ﬁrst is based on accounting indicators; these reﬂect the performance in retrospect. The other draws from strategic variety, which represents the forward look in performance. The accounting ratios include proxies such as return on assets, return on equity, and return on sales. The strategic indicators are captured by the use of the Sharp ratio, the Jensen model and others (Fig. 1). Hoskisson et al. (1993) discuss them in great detail. The literature expects that those ﬁrms that chose to create a larger and deeper footprint by geographic diversiﬁcation will submit to extemporaneous forces emanating from the new environment. They will operate under the constraints of the ﬁrm’s realities as they strategize by choice. As indicated in Fig. 1, the result is the ﬁrm’s performance level, which in turn manifests in the market at a premium or discount. This logic is the main framework in the literature and is diagrammed in Fig. 1.

FINDINGS ABOUT GEOGRAPHIC DIVERSIFICATION The evidence from the extant literature suggests that the choice of measures affects the empirical relationship between the market spread and ﬁrm performance (Goerzen & Beamish, 2003; Thomas & Eden, 2004; Allen & Pantzalis, 1996; Table 1). Also, it is apparent from meta studies that geographic market diversiﬁcation exhibits a stronger association with ﬁnancial performance ratios, particularly when revenue and market-share components are involved (Inge Geyskens, Steenkamp, & Kumar, 2006; Kirca, Jayachandran, & Bearden, 2005). This is plausible given that the ﬁrm’s topline provides a more direct assessment of the rewards of international market expansion (Hitt et al., 1997; Lu & Beamish, 2004; Kirca, Deligonul, Perry, & Cavusgil, forthcoming).

264

Z. SEYDA DELIGONUL

Table 1.

Diversiﬁcation and Performance Relationship in the Literaturea.

Study Buckley, Dunning, and Pearce (1978, 1984) Buhner (1987) Fatemi (1984) Qian (1997, 2002) Geringer, Tallman, and Olsen (2000) Riahi-Belkaoui (1998) Ruigrok and Wagner (2003) Grant (1987) Geringer, Beamish, and daCosta (1989) Grant, Jammine, and Thomas (1988) Shaked (1986) Siddharthan and Lall (1982) Haar (1989) Michel and Shaked (1986) Miller and Pras (1980) Sambharya (1995) Tallman and Li (1996) Vernon (1971) Contractor, Kundu, and Hsu (2003) Daniels and Bracker (1989) Gomes and Ramaswamy (1999) Hitt, Hoskisson, and Ireland (1994) Morck and Yeung (1991) Riahl-Belkaoui (1996) Hitt, Hoskisson, and Kim (1997) Ramaswamy (1995) Sullivan (1994) Kohers (1975) Kim and Lyn (1987) Geyikdagi and Geyikdagi (1989) Errunza and Senbet (1981) Brewer (1981) Errunza and Senbet (1984) Gerpott and Jakopin (2005)

Diversiﬁcation Variable

Performance Variable

FSTS

ROA

FSTS FSTS FSTS FSTS FSTS FSTS FSTS FSTS

Market return Risk adjusted ROE ROA, ROE ROA, ROS, Sales growth ROA ROA ROA, ROE, ROS ROA, ROS

FSTS FSTS FSTS FSTS FSTS FSTS FSTS, FATA FSTS, NNS FSTS, FATA, FETE, FETN PCA on FSTS, FETN, FOTO FATA FSTS, FATA, NNS NNS NNS, NFS NNS, E FATA FSTS, FATA, NNS FATA, FETE, FETN, FSTS FDI FBTB FATA, FETE, FSTS FETE AFS, FSTS, NFS, E FSTS, FETN, FATA, FCSC

ROA, ROE, ROS SD of ROE, Beta Sales growth ROA Risk adjusted return SD of ROS ROA, ROS, ROE ROS ROI, ROS ROA, ROS ROA, ROS Cost of sales/total sales, ROA ROA, ROS Market value ROA ROS ROA, ROS, ROA ROA, ROS Cost of capital Tobin’s q, excess market value Beta Excess return Stock return Excess return EBITDA, ROS, ROA

Note: FSTS, foreign sales/total sales; FETN, number of foreign employees/total number of employees; FETE, foreign earnings/total earnings; FBTB, foreign business units/total business units; FOTO, number of foreign ofﬁces/total number of ofﬁces; FATA, foreign assets/total assets; FCTC, foreign subscribers/total subscribers; NNS, number of nations which ﬁrm has foreign subsidiaries; NFS, number of subsidiaries abroad; FDI , foreign direct investment; AFS, absolute foreign sales; E , entropy measure; PCA. principle component analysis. a This table is built on one in Qian, Li, Li, and Qian (2002), where additional information about the details of this research stream is explored.

Geographic Market Diversiﬁcation

265

Additionally, ﬁndings in the literature show weakening of the diversiﬁcation–performance relationship as the cost component gets embedded in the performance measure. Speciﬁcally, proﬁtability ratios for example, will be less strongly associated with the market scope. A plausible discounting of ﬁrm value will then take place due to the costs of running cross-border activities (e.g., governance, transaction, coordination costs). The relationship between international footprints and management effectiveness and efﬁciency will remain strong, although not at the level of revenue-based measures, since efﬁciency and effectiveness measures reﬂect the turnover in terms of revenue- and asset-utilization level (personnel, inventory, and so on). Financial strength is expected to be positively associated with global diversiﬁcation when the market expansion facilitates risk-reduction (Kim, Hwang, & Burgers, 1993) which is consistent under the supposition that the ﬁrm internationalizes its marketing into uncorrelated markets. Furthermore, the scope of diversiﬁcation is strongly associated with performance than with depth. Considering that expanding the potential demand-base for performance provides a better opportunity for the marketing function to appropriate higher beneﬁts than full force investments and considering that the ﬁrms will seek the optimal degree of breadth and depth in internationalization (Allen & Pantzalis, 1996), a larger scope enables a ﬁrm to achieve a better ﬁt with strategic resources in target markets. A better ﬁt in turn, enables the ﬁrm to appropriate value from proprietary assets (Tallman & Li, 1996). Furthermore, a larger scope provides broader choices, while the opposite, higher depth, imputes into strategies deeper commitments, hence inﬂexibilities (Thomas & Eden, 2004). Also, the broader scope of global scale may provide a richer experience in organizational learning, ﬂexibility, and the capacity to engage in multipoint competition than the depth of the multinational footprints (Goerzen & Beamish, 2003; Kogut & Zender, 1993). This is expected given that the studies suggest a stronger positive association of scope with the ﬁrm’s performance than with the depth of the global diversiﬁcation.

EMPIRICAL CHALLENGES AND VALIDITY OF FINDINGS The ﬁndings in the literature stream will have higher meta value4 by improving on the number of issues. The following summarizes some of the weaknesses in this research stream.

266

Z. SEYDA DELIGONUL

It is surprising that all studies use some geographic-sales-share (e.g., foreign-sales-to-total-sales (FSTS) which is the most popular (Table 1)) or some conﬁguration of weighted revenue spread (such as entropy measures) to proxy the market diversiﬁcation level. This choice is pathological when both the antecedent and the consequent contain the same term. Some authors have overlooked this elementary mishap as they correlated the (FSTS) ratio with sales turnover. In that case, both components of the correlation include the same term, revenue, which potentially generates a bogus relationship that appears signiﬁcant. It is surprising that this research stream has not given much attention to problems of spurious linkages between the dependent and independent variables. For instance, such concern is certainly relevant for some of the studies presented in Table 1. The use of only ﬁnancial variables for both the antecedents and consequents redeﬁnes the diversiﬁcation–performance problem. If someone were to look at Table 1 after stripping off the title, that person may never be able to tell that this stream of literature is looking for a relationship between diversiﬁcation and performance. The best guess would be closer to a hypothesis on the sales effectives driven by asset utilization. Such interpretation attests to the conclusion that Table 1 reduces diversiﬁcation phenomena to an investigation of effectiveness of the ﬁnance manager. In that, we face the neglect of non-ﬁnancial dimensions of diversiﬁcation strategy (the cognitive, regulative, normative, social and polity issues). Also lack of indigenous measure may lead to erroneous conclusions about the diversiﬁcation phenomenon. For example, the effect of diversiﬁcation may be modulated by the unrelated ﬁnancials. Selling a building which has nothing to do with diversiﬁcation, will credit the diversiﬁcation performance since it will boost the return on assets by shrinking the asset base. As in this case, the research ﬁndings may be inﬂuenced by reasons other than the intrinsic complexities of the diversiﬁcation phenomenon. This concern may be relevant for a group of studies when they are considered in the context of a diversiﬁcation–performance relationship. Another common practice in this stream of research is to use a combination of indicators. The use of combinatorial techniques to integrate simple measures for the degree of internationalization is geometrically sound, but theoretically ﬂawed. For instance, Contractor et al. (2003) take the linear combination by principle component analysis of the FSTS, number of foreign employees/number of total employees (FETE), and number of foreign ofﬁces/number of total number of ofﬁces (FOTO). This generates a variable, which has no conceptual interpretation, but simply an eigenvalue weighted sum of three fractions. The outcome is a unitless and

Geographic Market Diversiﬁcation

267

also conceptually ambiguous variable, produced by the angles between three vectors (variables). The study intends that it will measure the degree of diversiﬁcation without reasons related to internationalization and without any theory. A better choice would be the incorporating of these three variables as manifest variables in a structured equation model. The use of only ﬁnancial ratios for both the antecedents and consequents is substantially problematic for another reason. The ratios used for the antecedents may correlate with performance outcome, not because of the diversiﬁcation decision, but because of their underlying relationships between intrinsic ﬁnancial decisions typical for the host industry. In the result by selection bias, one can guarantee a preordained conclusion. For example, Kim and Lyn (1987) could possibly change their ﬁndings by opting for a manufacturing context, which show across the board cases including being richer in knowledge assets endowments together with physical-assetrich data units. Selection bias is a true investigative hazard and may be confounding given the earlier conﬁrmation of the link between the capital structure and strategy (cf. Kochhar & Hitt, 1998). Table 1 reveals the researchers’ afﬁnity towards return on assets (ROA) as a performance measure. Almost all investigations use this ratio. In ﬁnancial circles, the ratio of return on assets is considered to be a ﬁnancial management-effectiveness indicator. In that class, however, it is known to be more prone to ﬁnancial manipulation in comparison to similar indicators. Managers can easily improve ROA ratio by adopting a divesting strategy of some of the assets, possibly the non-core assets. (Currently, Kodak case is a good example.5) For its openness to ﬁnancial engineering, ROA’s use without auxiliary indicators will not generate robust results across the ﬁrms. An important feature of the antecedent array is that its content is layered from the individual level to the team and from the team to the ﬁrm. Likewise, at the environment side they extend from industry to country. Both sets of variables are situated to include lower level variables and, in turn, each is embedded within higher levels (Chan, 1998). The current research treats them as independent antecedents (Fig. 1 too shows them as independent) and possibly it concedes to level fallacy where the ﬁndings suffer from validity weaknesses. Also in the hierarchy of embedded variables the interactions between the layers impute distortions in results. In addition to level fallacy, it is common in this stream of research to display aggregation bias. Arregle, He´bert, and Beamish (2006) present an example of this problem along with its consequences in international business research. Brouthers (1998) discusses an example from his study on how the country effect masks the

268

Z. SEYDA DELIGONUL

industry effect when these two effects are treated separately. When they are considered jointly, however, the explanatory power of the model increases 50 per cent. In passing, we may demonstrate some of these problems as an example. For instance, averaging individual performance to obtain team level performance (aggregation) may be invalid in a situation where the star player determines the performance outcome. Additionally, the patterns in each team may show the effect of personal incentives on collective performance, but collectively taken, the conclusion may demonstrate the opposite: personal incentives inhibit team performance (Simpson paradox). The interaction of team and personal effects may mask the motivational studies. All in all, there must be an increased concern regarding modeling, measurement, and analysis issues in hierarchical investigation. To eliminate the possible confounding effects, the antecedent variables that impact the diversiﬁcation decision must be uncorrelated with the ﬁrm’s value. In reality, this requirement rarely holds. For example, industry attractiveness (antecedent) is clearly a factor on the performance outcome (consequent) without the mediation of diversiﬁcation choice. This is an intrinsic problem and a matter of careful research design. For instance, Campa and Kedia (2002) show that in their example, the antecedent variables, which cause diversiﬁcation, directly cause the value destruction. The studies must prevent the blurring of the result when a variable that motivates the diversiﬁcation also drives the performance. Another argument about diversiﬁcation conjures up characteristics that are overlooked but potentially decisive on the ﬁrm’s valuation (premium/ discount). For example, consider a ﬁrm facing technological obsolescences which adversely affects its competitive standing. It is not surprising that this poor performer will trade at a discount relative to other ﬁrms. Plausibly, such a ﬁrm will have a lower opportunity cost to switch its resources between the borders, products, or industries (Campa & Kedia, 2002). Entry to a new foreign market is value destructive in this case. In reality, the reason is technological deterioration, but on the surface the value discount may appear to be a result of expanded market scope. On the other hand, when the opportunity cost of market expansion considered, this ﬁrm’s diversiﬁcation is economically justiﬁed. In another example, Campa and Kedia (2002) consider a ﬁrm with a unique capability. The incomplete information may require a costly search for the best market match at a level that compels the ﬁrm to forego the beneﬁts of specialization (Campa & Kedia, 2002). This ﬁrm, as in Matsusaka’s (2001) model, will draw a diversiﬁcation discount lagging behind the comparable non-diversiﬁed counterpart. The example shows that not taking into account the ﬁrm’s

Geographic Market Diversiﬁcation

269

search, and its potential payback, a ﬁrm faces the diversiﬁcation discount due to involved costs.

RECOMMENDATIONS The evidence is not clear whether the business-spread measures represent the intended concept in a way to lend itself to study the potential relationship between the geographic scope and performance. The basic problem is that business-count methods lack a theory about scope expansion and its effect on corporate functioning. There is no explicit appreciation in the businesscount measure that skills, resources, knowledge assets and so forth may be transferred across geographies to inﬂuence performance. The relatedness between the different environments and business units in different countries are overlooked. Speciﬁcally, these concerns can be addressed by better theorizing. On the operationalization side there is a need for better measurement proxies. In order to protect the analysis from spurious relationships the triangulation of variables is necessary. Some of the older measures fall short of these characteristics, therefore their use requires added caution. Also, the problem of diversiﬁcation and performance relationship should not be displaced by a new problem in the ﬁnancial make-up of a company. The problem has to be investigated in its genuine form by including economic and also regulative, normative, and cultural aspects. Finally, the hierarchical nature of the problem demands the use of such methods as hierarchically structured modeling in order to address the multi-level character of the problem.

CONCLUSION This study provides an overview and commentary about one of the most popular topics in strategic marketing, geographic market diversiﬁcation for international ﬁrms. Despite the wide interest in the topic, the ﬁndings in different studies are deeply divided in their conclusions. The reasons for such disagreement may stem from the speciﬁcs of theorizing and also technical aspects in empirical treatment. In particular, it seems the research cannot clearly deal with ﬁrms that have other reasons to fail (succeed), eventually diversify and face the diversiﬁcation discount (premium). Our discussion shows that, both in conceptualization and also operational

270

Z. SEYDA DELIGONUL

implementation, the empirics of geographic diversiﬁcation is challenging and requires more sophisticated approaches. Given the challenges on whether a ﬁrm performs better or whether it can create value will need a better lens to study, and improved modeling of contingent factors.

NOTES 1. The value destruction in this context refers to stock market valuation. 2. I thank the anonymous referee for pointing out the possibility of other explanations in addition to the economics explanation. Here, the selection is based on the reason that this explanation offers an integrative angle putting many other explanations under its umbrella. However, some others such as agency-theory-based motives like managerial hubris are certainly outside the scope of this explanation and also outside the scope of this chapter. 3. The depth of diversiﬁcation refers to level commitment and related arrangements specifying the degree of control. The scope of diversiﬁcation speciﬁes the breath of the strategy in terms of geographic targets and/or products. 4. Meta value here refers to epistemological merit by being consistent across the different investigations. 5. As part of its poor strategic performance at a point of inﬂection in imaging industry Kodak is currently trying to improve its indicators by selling off its ﬁxed assets. This divestiture strategy may be helpful to improve the indicators like ROA but it will not change the fundamentals. Then fundamentals will improve by innovation and by better adaptation to the discontinuity between ﬁlm and digital imaging.

REFERENCES Allen, L., & Pantzalis, P. (1996). Valuation of the operating ﬂexibility of multinational corporations. Journal of International Business Studies, 27(4), 653. Annavarjula, M., & Beldona, S. (2000). Multinationality–performance relationship: A review and reconceptualization. International Journal of Organizational Analysis. Bowling Green, 8(1), 48–67. Aron, D. J. (1988). Ability, moral hazard, ﬁrm size, and diversiﬁcation. The RAND Journal of Economics, 19(1), 72–87. Arregle, J.-L., He´bert, L., & Beamish, P. W. (2006). Mode of international entry: The advantages of multilevel methods. Management International Review, 46(5), 597–620. Berger, P. G., & Ofek, E. (1995). Diversiﬁcation’s effect on ﬁrm value. Journal of Financial Economics, 37(1), 39–65. Boyd, B. K., Gove, S., & Hitt, M. A. (2005). Consequences of measurement problems in strategic management research: The case of Amihud and Lev. Strategic Management Journal, 26, 367–375. Brewer, H. L. (1981). Investor beneﬁts from corporate international diversiﬁcation. Journal of Financial and Quantitative Analysis, 16(1), 113–126.

Geographic Market Diversiﬁcation

271

Brouthers, L. E. (1998). Explaining MNC proﬁtability: Country-speciﬁc, industry-speciﬁc and country industry interactive inﬂuences. Management International Review, 38(4), 345–361. Buckley, P. J., Dunning, J. H., & Pearce, R. B. (1978). The inﬂuence of ﬁrm size, industry, nationality, and degree of multinationality on the growth and proﬁtability of the world’s largest ﬁrms, 1962–1972. Weltwirtschaftliches Archiv, 114(2), 243–257. Buckley, R. J., Dunning, J. H., & Pierce, R. B. (1984). An analysis of growth, and proﬁtability of the world’s largest ﬁrms 1972–1977. Kyklos, 37(1), 3–26. Buhner, R. (1987). Assessing international diversiﬁcation of West German corporations. Strategic Management Journal, 8(1), 25–37. Campa, J. M., & Kedia, S. (2002). Explaining the diversiﬁcation discount. The Journal of Finance, 57(4), 1731–1761. Capone, N., Hulber, J. M., Farley, J. U., & Martin, L. E. (1988). Corporate diversity and economic performance: The impact of market specialization. Strategic Management Journal, 9, 61–74. Chan, D. (1998). Functional relationships among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83, 234–246. Chandler, A. (1977). The visible hand. Cambridge, MA: Belknap Press. Chang, S. J. (1995). International expansion strategy of Japanese ﬁrms: Capability building thorough sequential entry. Academy of Management Journal, 38(2), 383–407. Contractor, F. (2007). 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., Kundu, S. K., & Hsu, C. C. (2003). A three stage theory of international expansion: The link between multinationality and performance in the service sector. Journal of International Business Studies, 34(1), 5–18. Daniels, J. D., & Bracker, J. (1989). Proﬁt performance: Do foreign operations make a difference? Management International Review, 29(1), 46–56. Davis, P. S., Desai, A. B., & Francis, J. D. (2000). Mode of international entry: An isomorphism perspective. Journal of International Business Studies, 31(2), 239–258. Denis, D. J., Denis, D. K., & Sarin, A. (1997). Agency problems, equity ownership, and corporate diversiﬁcation source. The Journal of Finance, 52(1), 135–160. Dunning, J. H. (1988). The eclectic paradigm of international production: A restatement of some possible extensions. Journal of International Business Studies, 19(1), 1–31. Errunza, V., Hogan, K., & Hung, M.-W. (1999). Can the gains from international diversiﬁcation be achieved without trading abroad? The Journal of Finance, 54(6), 2075–2107. Errunza, V. R., & Senbet, L. W. (1981). The effects of international operations on the market value of the ﬁrm: Theory and evidence. Journal of Finance, 36(2), 401–417. Errunza, V. R., & Senbet, L. W. (1984). International corporate diversiﬁcation, market valuation, and size-adjusted evidence. Journal of Finance, 39(3), 727–745. Fatemi, A. M. (1984). Shareholder beneﬁts from corporate international diversiﬁcation. Journal of Finance, 39(5), 1325–1344. Geringer, J. M., Beamish, P. W., & daCosta, R. C. (1989). Diversiﬁcation strategy and internationalization: Implications for MNE performance. Strategic Management Journal, 10(2), 109–119. Geringer, J. M., Taliman, S., & Olsen, D. M. (2000). Product and international diversiﬁcation among Japanese multinational ﬁrms. Strategic Management Journal, 21(1), 51–80.

272

Z. SEYDA DELIGONUL

Gerpot, T. J., & Jakopin, N. M. (2005). The degree of internationalization, and the ﬁnancial performance of European mobile network operators. Telecommunication Policy, 29, 635–661. Geyikdagi, Y. M., & Geyikdagi, N. V. (1989). International diversiﬁcation in Latin America and the industrialized countries. Management International Review, 29(3), 62–71. Goerzen, A., & Beamish, P. W. (2003). Geographic scope and multinational enterprise performance. Strategic Management Journal, 24(13), 1289–1306. Gomes, L., & Ramaswamy, K. (1999). An empirical examination of the form of the relationship between multinationality and performance. Journal of International Business Studies, 30(1), 173–188. Grant, R. M. (1987). Multinationality and performance among British manufacturing companies 1972–84. Journal of International Business Studies, 18(3), 79–89. Grant, R. M., Jarmine, A. P., & Thomas, H. (1988). Diversity, diversiﬁcation, and proﬁtability among British manufacturing companies 1972–84. Academy of Management Journal, 31(4), 771–801. Haar, J. (1989). A comparative analysis of the proﬁtability performance of the largest US, European and Japanese multinational enterprises. Management International Review, 29(3), 5–18. Hadlock, C. J., Ryngaert, M., & Thomas, S. (2001). Corporate structure and equity offerings: Are there beneﬁts to diversiﬁcation? The Journal of Business, 74(4), 613–635. Harris, M., Kriebel, C. H., & Raviv, A. (1982). Asymmetric information incentives and intraﬁrm resource allocation. Management Science, 28(6), 604–620. Hennart, J. F. (2007). The theoretical rationale for a multinationality/performance relationship. Management International Review, 47(3), 423–452. Hitt, M. A., Bierman, L., Uhlenbruck, K., & Shimizu, K. (2006). The importance of resources in the internationalization of professional service ﬁrms: The good, the bad, and the ugly. Academy of Management Journal, 49(6), 1137–1157. Hitt, M. A., Hoskisson, R. E., & Ireland, R. D. (1994). A mid-range theory of the interactive effects of international and product diversiﬁcation on innovation and performance. Journal of Management, 20(2), 297–326. Hitt, M. A., Hoskisson, R. E., & Kim, H. (1997). International diversiﬁcation: Effects on innovation and ﬁrm performance in product-diversiﬁed ﬁrms. Academy of Management Journal, 40(4), 767–798. Hoskisson, R. E., Hitt, M. A., Johnson, R. A., & Moesel, D. D. (1993). Construct validity of an objective (entropy) categorical measure of diversiﬁcation strategy. Strategic Management Journal, 14(3), 215–237. Inge Geyskens, I., Steenkamp, J.-B. E. M., & Kumar, N. (2006). Make, buy, or ally: A transaction cost theory meta-analysis. Academy of Management Journal, 49(3), 519–543. Jensen, M. C. (1988). Takeovers: Their causes and consequences source. The Journal of Economic Perspectives, 2(1), 21–48. Khanna, T., & Palepu, K. (2000). Is group afﬁliation proﬁtable in emerging markets? An analysis of diversiﬁed Indian business groups. The Journal of Finance, 55(2), 867–891. Kim, C. W., Hwang, P., & Burgers, W. P. (1993). Multinationals’ diversiﬁcation and the riskreturn trade-off. Strategic Management Journal, 14(4), 275–286. Kim, W. S., & Lyn, E. O. (1987). Foreign direct investment theories, entry barriers and reverse investments in US manufacturing industries. Journal of International Business Studies, 18(2), 53–67.

Geographic Market Diversiﬁcation

273

Kirca, A., Deligonul, S., Perry, M., & Cavusgil, T. Drivers of ﬁrm multinationality: A metaanalysis. (forthcoming). Kirca, A. H., Jayachandran, S., & Bearden, W. O. (2005). Market orientation: A meta-analytic review and assessment of its antecedents and impact on performance. Journal of Marketing, 69(2), 24–41. Kochhar, R., & Hitt, M. A. (1998). Linking corporate strategy to capital structure: Diversiﬁcation strategy, type and source of ﬁnancing. Strategic Management Journal, 19(Summer), 601–610. Kohers, T. (1975). The effect of multinational corporations on the cost of equity capital of US corporation. Management International Review, 15(2–3), 121–124. Lamont, O. (1997). Cash ﬂow and investment: Evidence from internal capital markets. The Journal of Finance, 52(1), 83–109. Lang, L. H. P., & Stulz, R. M. (1994). Tobin’s q, corporate diversiﬁcation, and ﬁrm performance. The Journal of Political Economy, 102(6), 1248–1280. Lewellen, W. G. (1971). A pure ﬁnancial rationale for the conglomerate merger. The Journal of Finance, 26(2), 521–537. Lins, K., & Servaes, H. (1999). International evidence on the value of corporate diversiﬁcation. The Journal of Finance, 54(6), 2215–2239. Lu, J. W., & Beamish, P. W. (2004). International diversiﬁcation and ﬁrm performance: The S-curve hypothesis. Academy of Management Journal, 47(4), 598–609. Madhok, A. (1997). Cost value and foreign market entry: Transaction and the ﬁrm. Strategic Management Journal, 18(1), 39–61. Martin, J., & Sayrak, A. (2003). Corporate diversiﬁcation and shareholder value. Journal of Corporate Finance, 9, 37–57. Matsusaka, J. G. (2001). Corporate diversiﬁcation, value maximization, and organizational capabilities. The Journal of Business, 74(3), 409–431. Michel, A., & Shaked, I. (1986). Multinational corporations vs domestic corporations: Financial performance and characteristics. Journal of International Business Studies, 17(3), 89–101. Miller, J. C., & Pras, B. (1980). The effects of multinational and export diversiﬁcation on the proﬁt stability of US corporations. Southern Economic Journal, 46(3), 792–805. Morck, R., & Yeung, B. (1991). Why investors value multinationality. Journal of Business, 64(2), 165–187. Qian, G. (1997). Assessing product-market diversiﬁcation of US ﬁrms. Management International Review, 37(2), 127–149. Qian, G. (2002). Multinationality, product diversiﬁcation, and proﬁtability of emerging US small- and medium-sized enterprises. Journal of Business Venturing, 17(6), 611–633. Qian, G., Li, L., Li, J., & Qian, Z. (2002). Regional diversiﬁcation and ﬁrm performance. Journal of International Business Studies, 39, 197–214. Rajan, R., Servaes, H., & Zingales, L. (2000). The cost of diversity: The diversiﬁcation discount and inefﬁcient investment. The Journal of Finance, 55(1), 35–80. Ramaswamy, K. (1995). Multinationality, conﬁguration, and performance: A study of MNEs in the US drug and pharmaceutical sector. Journal of International Management, 1(2), 231–253. Riahi-Belkaoui, A. R. (1998). The effects of the degree of internationalization on ﬁrm performance. International Business Review, 7(3), 315–323. Rotemberg, J. J., & Saloner, G. (1994). Beneﬁts of narrow business strategies. The American Economic Review, 84(5), 1330–1349.

274

Z. SEYDA DELIGONUL

Ruigrok, W., & Wagner, H. (2003). Internationalization and performance: An organizational learning perspective. Management International Review, 43(1), 63–83. Sambharya, R. S. (1995). The combined effect of international diversiﬁcation and product diversiﬁcation strategies on the performance of US based multinational corporations. Management International Review, 35(3), 197–218. Scharfstein, D. S., & Stein, J. C. (2000). The dark side of internal capital markets: Divisional rent-seeking and inefﬁcient investment. The Journal of Finance, 55(6), 2537–2564. Schoar, A. (2002). Effects of corporate diversiﬁcation on productivity. The Journal of Finance, 57(6), 2379–2405. Servaes, H. (1996). The value of diversiﬁcation during the conglomerate merger wave. The Journal of Finance, 51(4), 1201–1225. Shaked, I. (1986). Are multinational corporations safer? Journal of International Business Studies, 17(1), 75–80. Siddharthan, N. S., & Lall, S. (1982). Recent growth of the largest US multinationals. Oxford Bulletin of Economics and Statistics, 44(1), 1–13. Stein, J. C. (1997). Internal capital markets and the competition for corporate resources. The Journal of Finance, 52(1), 111–133. Sullivan, D. (1994). Measuring the degree of internationalization of a ﬁrm. Journal of International Business Studies, 24(2), 325–342. Tallman, S., & Li, J. (1996). Effects of international diversity and product diversity on the performance of multinational ﬁrms. Academy of Management Journal, 39(1), 179–196. Thomas, D. E., & Eden, L. (2004). What is the shape of the multinationality–performance relationship? Multinational Business Review, 12(1), 89–110. Vernon, R. (1971). Sovereignty at bay: The multinational spread of US enterprises. New York: Basic Books. Wernerfelt, B., & Montgomery, C. A. (1988). Tobin’s q and the importance of focus in ﬁrm performance. The American Economic Review, 78(1), 246–250.

PART IV METHODOLOGICAL ADVANCES

THE USE OF PARTIAL LEAST SQUARES PATH MODELING IN INTERNATIONAL MARKETING Jo¨rg Henseler, Christian M. Ringle and Rudolf R. Sinkovics The advent of structural equation modeling (SEM) with latent variables has changed the nature of research in international marketing and management. Researchers acknowledge the possibilities of distinguishing between measurement and structural models and explicitly taking measurement error into account. As Gefen, Straub, and Boudreau (2000, p. 6) point out, ‘‘SEM has become de rigueur in validating instruments and testing linkages between constructs.’’ They furthermore distinguish between two families of SEM techniques: covariance-based techniques, as represented by LISREL, and variance-based techniques, of which partial least squares (PLS) path modeling is the most prominent representative. PLS has been used by a growing number of researchers from various disciplines such as strategic management (e.g., Hulland, 1999), management information systems (e.g., Dibbern, Goles, Hirschheim, & Jayatilaka, 2004), e-business (e.g., Pavlou & Chai, 2002), organizational behavior (e.g., Higgins, Duxbury, & Irving, 1992), marketing (e.g., Reinartz, Krafft, & Hoyer, 2004), and consumer behavior (e.g., Fornell & Robinson, 1983). Since 1987, for instance, more than 20 studies using PLS have been published in ﬁve top-tier marketing journals (Eggert, 2007) – the majority in New Challenges to International Marketing Advances in International Marketing, Volume 20, 277–319 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020014

277

278

JO¨RG HENSELER ET AL.

the last six years. PLS is the method of choice for success factor studies in marketing (Albers, 2009) and for estimating the various national customer satisfaction index models (e.g., Fornell, 1992). The PLS methodology has also achieved an increasingly popular role in empirical research in international marketing, which may represent an appreciation of distinctive methodological features of PLS. As of March 2008, more than 30 articles on international marketing using PLS were published in double-blind reviewed journals. However, these publications show a rather large variability in the way how PLS is applied and how its outcomes are reported. In many instances, the rationale for choosing PLS among a possible set of alternative analytical techniques is not made explicit. Although several articles offer guidance for the use of covariance-based structural equation modeling (CBSEM) in international marketing (e.g., Steenkamp & Baumgartner, 1998; Malhotra, 2001; Iacobucci, Grisaffe, Duhachek, & Marcati, 2003), there are no similar subject-speciﬁc guidelines for the use of PLS. While there are general recommendations for the use of PLS (e.g., Hulland, 1999), the speciﬁc requirements and typical research problems of international marketing have not been addressed yet. Our main aims are to shed light on PLS path modeling as an SEM technique, to reveal the strengths and weaknesses of PLS in general, and to deliver guidelines for its use in international marketing in particular.

1. APPLICATIONS OF PLS PATH MODELING IN INTERNATIONAL MARKETING In order to determine the status quo of PLS path modeling in international marketing research, we conducted an exhaustive literature review. An evaluation of double-blind reviewed journals through important academic publishing databases (e.g., ABI/Inform, Elsevier ScienceDirect, Emerald Insight, Google Scholar, PsycINFO, Swetswise) revealed that more than 30 academic articles in the domain of international marketing (in a broad sense) used PLS path modeling as means of statistical analysis. We assessed what the main motivation for the use of PLS was in respect of each article. Moreover, we checked for applications of PLS in combination with one or more additional methods, and whether the main reason for conducting any additional method(s) was mentioned. Table 1 lists all the identiﬁed academic articles that use PLS as a method of analysis in the context of international marketing. Green and Ryans

Holzmu¨ller and Sto¨ttinger (1996, JIM)

Green and Ryans (1990, JPIM)

Festge and Schwaiger (2007, AIM) Gerpott and Jakopin (2005, SBR) Graham, Mintu, and Rodgers (1994, Mgmt.Sc.)

Calantone, Graham, and Mintu-Wimsatt (1998, IJRM)

Alpert, Kamins, Sakano, Onzo, and Graham (2001, IMR) Birkinshaw, Morrison, and Hulland (1995, SMJ)

Ainuddin, Beamish, Hulland, and Rouse (2007, JWB)

Acedo and Jones (2007, JWB)

Study

Table 1.

‘‘Parameters can be estimated independent of sample size’’ (p. 79), ‘‘PLS avoids parameter estimation biases inherent in regression analysis’’ (p. 80), ‘‘PLS provides the most ﬂexibility regarding measurement of the constructs’’ (p. 80) ‘‘Given that the purpose of this study is to predict [ . . . ], PLS has thus been chosen as the structural equation modeling approach’’ (p. 53), ‘‘[PLS] is more robust with small sample sizes’’ (p. 53), ‘‘the data do not have to be multivariate normal because of the ﬁxed point estimation’’ (p. 53) Not explicitly mentioned

‘‘The PLS technique is justiﬁed where theory is insufﬁciently grounded and the variables or measures do not conform to a rigorously speciﬁed measurement model, or ﬁt a certain distribution’’ (p. 242) ‘‘Use of PLS is especially suited to exploratory studies such as this, where the measures [ . . . ] are new and the relationships [ . . . ] have not been previously tested’’ (p. 56) ‘‘Formative indicators can only be analyzed using partial least squares (PLS), and not by using the more common structural equation technique of LISREL’’ (p. 177–178) ‘‘PLS is most appropriate when sample sizes are small, when assumptions of multivariate normality and interval scaled data cannot be made, and when the researcher is primarily concerned with prediction of the dependent variable’’ (pp. 646–647) ‘‘The PLS parameter estimates better reveal the strength and direction (i.e., positive vs. negative) of the relationships among variables compared to correlation coefﬁcients’’ (p. 28), ‘‘PLS avoids parameters estimation biases common in regression analysis’’ (p. 28) ‘‘The researcher’s focus is placed on the explanation of an endogenous construct’’ (p. 192) Not explicitly mentioned

Motivation for Using PLS Path Modeling

Analysis of subgroups

Multiple regression

N/A

N/A

LISREL

N/A

Logistic regression

N/A

N/A

In order to test path signiﬁcances

Verify model for binary variables N/A

Path signiﬁcances

Group comparison

Multiple regression and Chow test

LISREL

N/A

Group comparison

Reason for Additional Analysis

N/A

t-test

Additional Analysis

Studies Using PLS Path Modeling in International Marketing Research.

Partial Least Squares Path Modeling in International Marketing 279

Mahmood, Bagchi, and Ford (2004, IJEC)

Lee (2000, EJM)

Lee (2001, JBR)

Lee, Yang, and Graham (2006, JIBS)

Johnson, Herrmann, and Gustafsson (2002, JEP) Julien and Ramangalahy (2003, ETP)

Inkpen and Birkenshaw (1994, IBR) Johansson and Yip (1994, SMJ)

Holzmu¨ller and Kasper (1991, MIR)

Study

‘‘PLS avoids many of the restrictive assumptions imposed by other causal models that involve latent variables such as LISREL’’, ‘‘PLS provides measurement assessment’’, ‘‘A jack-knife procedure [ . . . ] generates an approximate t-statistic. This overcomes the disadvantage of the lack of formal signiﬁcance tests for parameters resulting from non-parametric methods’’, ‘‘PLS enables the explicit estimation of the multiple item construct, which affords a comparison of [groups] at the construct level’’ (p. 196) ‘‘The PLS technique imposes minimal demand on measurement scales, sample sizes, and residual distributions. [It] is often used to test and validate exploratory models’’ (p. 20)

‘‘PLS is known to be particularly advantageous in the initial development and assessment phase of theory building,’’ ‘‘the PLS method is [ . . . ] more robust since its does not require either a large sample or normally distributed data’’ (p. 233) ‘‘PLS [ . . . ] is more appropriate for the exploratory nature of [a] study,’’ ‘‘[PLS] allows for formative indicators [ . . . ] and dichotomous constructs’’ (p. 632) ‘‘PLS [ . . . ] can accommodate a small sample size’’ (p. 153)

‘‘Fewer restrictive assumptions,’’ ‘‘ratio between sample size and the number of parameters to be estimated,’’ ‘‘exploratory intention’’ (p. 58) ‘‘All relationships are modeled simultaneously, eliminating concerns about multicollinearity’’ (p. 208) ‘‘Less stringent assumptions about the randomness of the sample and the normality of the distribution of variables’’ (p. 587), ‘‘smaller sample sizes, as each causal subsystem sequence of paths is estimated separately’’ (p. 587) Not explicitly mentioned

Motivation for Using PLS Path Modeling

Table 1. (Continued )

AMOS

Conﬁrmatory factor analysis t-test and ANOVA

Fit statistics

Examine unidimensionality Group comparison

Group comparison

N/A

N/A

t-test and ANOVA

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

Reason for Additional Analysis

Additional Analysis

280 JO¨RG HENSELER ET AL.

Venaik, Midgley, and Devinney (2005, JIBS)

Singh, Fassott, Chao, and Hoffmann (2006a, IMR) Singh, Fassott, Zhao, and Boughton (2006b, JCB) Sto¨ttinger and Holzmu¨ller (2001, MIR) Tsang (2002, SMJ)

Pullman, Granzin, and Olsen (1997, IBR) Pinto, Rodrı´ guez Escudero, and Gutı´ errez Cilla´n (2008, IMM)

Money (2004, JBR) Money and Graham (1999, JIBS) Nakamura, Shaver, and Yeung (1996, IJIO) Nijssen and Douglas (2008, JIM) O’Cass and Fenech (2003, EJM) Pavlou and Chai (2002, JECR)

Mintu-Wimsatt and Graham (2004, JAMS)

N/A N/A

Group comparison

N/A

N/A N/A N/A t-test

N/A

Not explicitly mentioned

‘‘PLS is particularly suitable for data analysis during the early stage of theory development where the theoretical model and its measures are not well formed’’ (p. 841) ‘‘At an early stage of development [ . . . ] the regression based approach of PLS is considered more appropriate than covariance-based methods such as LISREL,’’ applicable when a multivariate normal distribution can not be assured, small sample size in combination with a complex model including second-order constructs, formative indicators (p. 665)

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A Path signiﬁcances group comparison N/A

N/A

N/A

N/A LISREL Regression & Chow Test N/A

N/A

‘‘PLS can deal effectively with formative scales, is distribution free, and is a powerful instrument for analyzing small samples’’ (p. 95) ‘‘[PLS] circumvent[s] the necessity for the multivariate normal assumption’’ (p. 377) ‘‘PLS allows [ . . . ] a simultaneous analysis of both whether the hypothesized relationships at the theoretical level are empirically acceptable, and also how well the measures relate to each construct,’’ ‘‘[t]he ability to include multiple measures for each construct,’’ ‘‘the nature of [ . . . ] measures,’’ ‘‘the small sample size’’ (p. 246) ‘‘The objective of PLS is explanation of the relationships and prediction of the criterion variables of the model’’ (p. 221) ‘‘Avoiding any distributional assumption of the observed variables’’ (p. 160), ‘‘the sample size required in PLS is much smaller’’ (p. 160), ‘‘PLS can handle both types of measurement models, reﬂective and formative’’ (p. 160) ‘‘PLS [ . . . ] uses distribution-free assumptions, [ . . . ] especially when independence of observations is not stipulated’’ (p. 89) Not explicitly mentioned

Not explicitly mentioned

‘‘PLS is a more rigorous approach [ . . . ] compared to correlation and regression analyses’’ (p. 351), ‘‘PLS [ . . . ] avoids small sample size problems,’’ ‘‘PLS minimizes biases associated with [ . . . ] dichotomous and ordinal measures’’ (p. 352) In order to validate regression results N/A

Partial Least Squares Path Modeling in International Marketing 281

282

JO¨RG HENSELER ET AL.

(1990) published the ﬁrst study using PLS in international marketing. Although there were several articles in the 1990s, the popularity of PLS seems to have largely increased in the new millennium. Journals that published PLS applications in the international marketing domain are (journal abbreviation and number of studies in parentheses):

Advances in International Marketing (AIM; 1) Entrepreneurship Theory and Practice (ETP; 1) European Journal of Marketing (EJM; 2) Industrial Marketing Management (IMM; 1) International Business Review (IBR; 2) International Journal of Electronic Commerce (IJEC; 1) International Journal of Industrial Organization (IJIO; 1) International Journal of Research in Marketing (IJRM; 1) International Marketing Review (IMR; 2) Journal of Business Research (JBR; 2) Journal of Consumer Behaviour (JCR; 1) Journal of Economic Psychology (JEP; 1) Journal of Electronic Commerce Research (JECR; 1) Journal of International Business Studies (JIBS; 3) Journal of International Marketing (JIM; 2) Journal of Product Innovation Management (JPIM; 1) Journal of the Academy of Marketing Science (JAMS; 1) Journal of World Business (JWB; 2) Management International Review (MIR; 2) Management Science (Mgmt.Sc.; 1) Schmalenbach Business Review (SBR; 1) Strategic Management Journal (SMJ; 3)

Our ﬁrst analysis of these 33 studies provides some evidence of the increasing popularity of PLS path modeling within the research community. The analysis also points to the authors’ motivation for the use of this particular statistical method. Many researchers argue that the goal of their studies is in line with particular strengths of PLS path modeling. The most important motivations are exploration and prediction, as PLS path modeling is recommended in an early stage of theoretical development in order to test and validate exploratory models. Another powerful feature of PLS path modeling is that it is suitable for prediction-oriented research. Thereby, the methodology assists researchers who focus on the explanation of endogenous constructs. The characteristics of PLS path modeling, which

Partial Least Squares Path Modeling in International Marketing

283

researchers regard as relevant for their studies on international marketing, can be summarized as follows: PLS delivers latent variable scores, i.e. proxies of the constructs, which are measured by one or several indicators (manifest variables). PLS path modeling avoids small sample size problems and can therefore be applied in some situations when other methods cannot. PLS path modeling can estimate very complex models with many latent and manifest variables. PLS path modeling has less stringent assumptions about the distribution of variables and error terms. PLS can handle both reﬂective and formative measurement models. However, there are several arguments regarding the characteristics of PLS path modeling that should be treated with caution: PLS path modeling does not have less stringent assumptions about the representativeness of the sample. Sometimes, PLS path modeling is implicitly presumed to be the only method that can cope with formative indicators (cf. Chin, 1998); however, LISREL also has this ability (cf. Jarvis, MacKenzie, & Podsakoff, 2003). Although comparisons of methods provide some evidence of PLS’ favorable behavior in light of multicollinearity (cf. Gustafsson & Johnson, 2004), PLS does not resist multicollinearity. Since PLS determines measurement models (in Mode B) as well as structural models by means of multiple regressions, PLS estimates can be subject to multicollinearity problems, too. Our review furthermore discloses that some studies use PLS path modeling in combination with other analysis techniques such as t-tests, ANOVA, and CBSEM. Yet, these methods are not always a suitable choice. If, for instance, PLS was selected because of its distribution-free character, it would be inconsequent to introduce distributional assumptions in another analysis such as t-test or ANOVA, or to rely on criteria derived from CBSEM’s chi-square statistic. This ﬁnding provides evidence that either PLS path modeling lacks important features, which makes the use of additional analyses necessary, or that researchers are not aware of the respective extensions of PLS path modeling. In particular, the ﬁndings underpin the strong need for a PLS-based approach to multigroup analysis (MGA) in international marketing in order to compare model parameters across groups such as countries or cultures. The remainder of this chapter is organized as follows: In the following Section 2, we describe the working principle of PLS path modeling and discuss the characteristics of PLS in detail in order to assist research of

284

JO¨RG HENSELER ET AL.

international marketing to distinguish between facts and ﬁction about PLS path modeling. Section 3 addresses the assessment of PLS path modeling results, which represents a key area of concern when applying the methodology. Five issues are covered: Sections 3.1 and 3.2 provide an overview of the use of PLS path modeling for the assessment of reﬂective and formative measurement models, respectively. Section 3.3 presents the available criteria for evaluating the structural model. Section 3.4 presents and discusses the use of bootstrapping for determining conﬁdence intervals and the signiﬁcance of PLS path modeling estimates. As a courtesy to international marketing’s need to compare model parameters across countries, Section 3.5 proposes a new PLS-based approach for group comparisons based on bootstrapping. Finally, Section 4 sums up the key ﬁndings and draws conclusions.

2. PLS PATH MODELING PLS is a family of alternating least squares algorithms, or ‘‘prescriptions,’’ which extend principal component and canonical correlation analysis. The method was designed by Wold (1974, 1982, 1985) for the analysis of high dimensional data in a low-structure environment and has undergone various extensions and modiﬁcations. This chapter builds on Wold’s (1982) basic algorithmic design and some extensions developed by Lohmo¨ller (1989).

2.1. The Nature of PLS Path Models PLS path models are formally deﬁned by two sets of linear equations: the inner model and the outer model. The inner model speciﬁes the relationships between unobserved or latent variables, whereas the outer model speciﬁes the relationships between a latent variable and its observed or manifest variables. The various literatures do not always employ the same terminology. For instance, publications addressing CBSEM (e.g., Rigdon, 1998) often refer to structural models and measurement models or (observed) indicator variables; whereas those focusing on PLS path modeling (e.g., Lohmo¨ller, 1989) use the terms inner model and outer model or manifest variables for similar elements of the causal model. Fig. 1 depicts an example of a PLS path model. In order to simplify the notation of the model and in line with conventional descriptions of PLS, we assume that latent and manifest variables are

285

Partial Least Squares Path Modeling in International Marketing

x11 x12

1

3 1

x13 x21 x22 x23

2

4 2

x24

Inner Model

Outer Model (Here: Formative Mode)

Fig. 1.

x31

31

x32

32

x41

41

x42

42

x43

43

Outer Model (Here: Reflective Mode)

Example of a PLS Path Model.

standardized so that the location parameters can be discarded in the following equations. The inner model for relationships between latent variables can be written as: x ¼ Bx þ z

(1)

where x is the vector of latent variables, B denotes the matrix of coefﬁcients of their relationships, and z represents the inner model residuals. The basic PLS design assumes a recursive inner model that is subject to predictor speciﬁcation. Thus, the inner model constitutes a causal chain system (i.e. with uncorrelated residuals and without correlations between the residual term of a certain endogenous latent variable and its explanatory latent variables). Predictor speciﬁcation reduces Eq. (1) to: ðxjxÞ ¼ Bx

(2)

PLS path modeling includes two different kinds of outer models: reﬂective (Mode A) and formative (Mode B) measurement models. The selection of a certain outer mode is subject to theoretical reasoning (Diamantopoulos & Winklhofer, 2001). The reﬂective mode has causal relationships from the latent variable to the manifest variables in its block. Thus, each manifest variable in a certain measurement model is assumed to be generated as a linear function of its latent variables and the residual e: X x ¼ Lx x þ x

(3)

JO¨RG HENSELER ET AL.

286

where L represents the loading (pattern) coefﬁcients. The outer relationships are also subject to predictor speciﬁcation – implying that there are no correlations between the outer residuals and the latent variable of the same block – that reduces Eq. (3) to: ðX x jxÞ ¼ Lx x

(4)

The formative mode of a measurement model has causal relationships from the manifest variables to the latent variable. For those blocks, the linear relationships are given as follows: x ¼ P x X x þ x

(5)

In the formative mode, predictor speciﬁcation is also in effect, reducing Eq. (5) to: ðxjX x Þ ¼ Px X x

(6)

Moreover, it is important to see that the terms ‘‘formative’’ and ‘‘reﬂective,’’ as well as the connotation which is associated with the classiﬁcation of ‘‘causal’’ and ‘‘effect,’’ point at a difference between the characterization of the latent variable measurement models’ mode. Although a latent variable was originally considered an exact linear combination of its indicators for formative indicator speciﬁcations, a causal indicator speciﬁcation – the original term for these indicators – may be more general in that it holds both in the case of an exact linear combination, as well as when the indicators do not completely determine the latent variable (Bollen, 1989). In this chapter, we consistently use the terms ‘‘formative’’ and ‘‘reﬂective’’ measurement models – in the way they are described, for example, by Jarvis et al. (2003). It must be noted, though, that while we use the terms ‘‘reﬂective’’ and ‘‘formative’’ constructs to refer to latent variables that are measured with reﬂective or formative indicators, ‘‘strictly speaking, it is the (observable) measures (i.e. the indicators) that are being modeled as reﬂective or formative and not the (unobservable) constructs as such’’ (Diamantopoulos, 2006, p. 15). The following section introduces the basic PLS algorithm, which starts with the data matrix of manifest variables and successively computes the latent variable scores and all unknown relationships.

Partial Least Squares Path Modeling in International Marketing

287

2.2. The PLS Path Modeling Algorithm The PLS algorithm is essentially a sequence of regressions in terms of weight vectors. The weight vectors obtained at convergence satisfy ﬁxed point equations (see Dijkstra, 2009, for a general analysis of such equations and ensuing convergence issues). The basic PLS algorithm, as suggested by Lohmo¨ller (1989), includes the following three stages: Stage 1: Iterative estimation of latent variable scores, consisting of a four-step iterative procedure that is repeated until convergence is obtained: (1) (2) (3) (4)

outer approximation of the latent variable scores, estimation of the inner weights, inner approximation of the latent variable scores, and estimation of the outer weights.

Stage 2: Estimation of outer weights/loading and path coefﬁcients. Stage 3: Estimation of location parameters. We draw on Tenenhaus, Esposito Vinzi, Chatelin, and Lauro’s (2005) description of the ﬁrst stage of the PLS path modeling algorithm: Step 1: Outer approximation of the latent variable scores. Outer proxies of the outer latent variables, x^ n , are calculated as linear combinations of their respective indicators. These outer proxies are standardized; i.e. they have a mean of 0 and a standard deviation of 1. The weights of the linear combinations result from step 4 of the previous iteration. When the algorithm is initialized, and no weights are available yet, any arbitrary nontrivial linear combination of indicators can serve as an outer proxy of a latent variable. Step 2: Estimation of the inner weights. Inner weights are calculated for each latent variable in order to reﬂect how strongly the other latent variables are connected to it. There are three schemes available for determining the inner weights. Wold (1982) originally proposed the centroid scheme. Later, Lohmo¨ller (1989) developed the factor weighting and path weighting schemes. The centroid scheme uses the sign of the correlations between a latent variable – or, more precisely, the outer proxy – and its adjacent latent variables; the factor weighting scheme uses the correlations. The path weighting scheme pays tribute to the arrow orientations in the path model. The weights of those latent variables that explain the focal latent variable are set to the regression coefﬁcients stemming from a regression of the focal latent variable (regressant) on its latent repressor variables. The weights of those latent variables, which are explained by the focal latent

288

JO¨RG HENSELER ET AL.

variable, are determined in a similar manner as in the factor weighting scheme. Regardless of the weighting scheme, a weight of zero is assigned to all nonadjacent latent variables. Step 3: Inner approximation of the latent variable scores. Inner proxies of inner the latent variables, x^ n , are calculated as linear combinations of the outer proxies of their respective adjacent latent variables, using the aforedetermined inner weights. Step 4: Estimation of the outer weights. The outer weights are calculated either as the covariances between the inner proxy of each latent variable and its indicators (in Mode A, reﬂective), or as the regression weights resulting from the ordinary least squares regression of the inner proxy of each latent variable on its indicators (in Mode B, formative). These four steps are repeated until the change in outer weights between two iterations drops below a predeﬁned limit. The algorithm terminates after step 1, delivering latent variable scores for all latent variables. Loadings and inner regression coefﬁcients are then calculated in a straightforward way, given the constructed indices and using Eqs. (4) and (5). In order to determine the path coefﬁcients, for each endogenous latent variable a (multiple) linear regression is conducted.

2.3. Methodological Characteristics Methodological literature on PLS path modeling or publications on causal modeling applications that utilize the PLS path modeling approach usually refer to certain advantageous features of this technique (e.g., Fornell & Bookstein, 1982; Jo¨reskog & Wold, 1982; Dijkstra, 1983; Lohmo¨ller, 1989; SchneeweiX, 1991; Falk & Miller, 1992). The popularity of PLS path modeling among scientists and practitioners is rooted in four genuine characteristics: First, instead of solely drawing on the common reﬂective mode, the PLS path modeling algorithm allows the unrestricted computation of cause–effect relationship models that employ both reﬂective and formative measurement models (Diamantopoulos & Winklhofer, 2001). Second, PLS can be used to estimate path models when sample sizes are small (Chin & Newsted, 1999). Third, PLS path models can be very complex (i.e. consist of many latent and manifest variables) without leading to estimation problems (Wold, 1985). PLS path modeling is methodologically advantageous to CBSEM whenever improper or nonconvergent results are likely to occur (i.e. Heywood cases; see Krijnen, Dijkstra, & Gill, 1998).

Partial Least Squares Path Modeling in International Marketing

289

Furthermore, with more complex models, the number of latent and manifest variables may be high in relation to the number of observations. Fourth, PLS path modeling can be used when distributions are highly skewed (Bagozzi, 1994), or the independence of observations is not assured, because, as Fornell (1982, p. 443) has argued, ‘‘there are no distributional requirements.’’ All four characteristics are discussed in detail.

2.3.1. Reﬂective and Formative Measurement Models Structural equation models usually involve latent variables with multiple indicators. The measurement model or outer model speciﬁes the relationship between indicators and latent variables. The direction of path relationships per measurement model and, thus, the causality between the latent variable and its indicators are either described by a reﬂective or a formative mode. The reﬂective measurement model has its roots in classical test theory and psychometrics (Nunnally & Bernstein, 1994). Each indicator represents an error-afﬂicted measurement of the latent variable. The direction of causality is from the construct to the indicators; thus, observed measures are assumed to reﬂect variation in the latent variable. In other words, changes in the construct are expected to be manifested in changes in all of its indicators. Fig. 2 provides an example of a reﬂective measurement model. In some situations, for instance in the early stages of model development, it is appropriate to determine causality from the measures to the construct, rather than vice versa (Blalock, 1971). This kind of situation represents a formative measurement model, which is adequate when a construct is deﬁned as a combination of its indicators. A typical example would be marketing mix elements that are determined by a combination of variables (Fornell, 1982) and, hence, have to be modeled as a (typically linear) 1

Heart Rate x1

2

Lactate Level x2

3

Muscle Proportion x3

Fig. 2.

Cycling Fitness

Example of a Reﬂective Measurement Model.

JO¨RG HENSELER ET AL.

290 Training x1

Nutrition x2

Drug Abuse x3

Fig. 3.

Cycling Fitness

Example of a Formative Measurement Model.

combination of its indicators plus a disturbance term (Diamantopoulos, 2006). Consequently, the latent variable is deﬁned as a weighted score across all representative indicator variables, with each variable embodying an independent dimension in its own right. An increase in the value of one indicator translates into a higher score for the composite variable, regardless of the value of the other indicators. The formative measurement model exhausts the entire domain of the index, meaning that the indicators collectively represent all the relevant dimensions or independent underpinnings of the latent variable. One implication of this direction of causality is that omitting one indicator could omit a unique part of the formative measurement model and change the meaning of the variable (Diamantopoulos & Winklhofer, 2001). Fig. 3 presents an example of a formative measurement model. Although the inclusion of formative measures in CBSEM has been well documented (e.g., Jo¨reskog & Goldberger, 1975; MacCallum & Browne, 1993; Jo¨reskog & So¨rbom, 1996), analysts usually encounter identiﬁcation problems. As a sort of ad-hoc remedy, formative indicators could be modeled in CBSEM by re-specifying the formative indicators as exogenous latent variables with single indicators, ﬁxed unit loadings, and a ﬁxed measurement error (Williams, Edwards, & Vandenberg, 2003). In contrast, similar problems do not arise in PLS path modeling. The PLS path modeling algorithm – as it is presented in Section 2.2 – is equally well suited for SEM with reﬂective and/or formative measurement models. The only problematic issue, however, is connected to manifest variables’ critical level of multicollinearity in formative measurement models. Researchers should carefully decide whether a measurement model shall be formative or reﬂective. Measurement model misspeciﬁcation is an often observed phenomenon. According to Jarvis et al. (2003), 28% of top-level

Partial Least Squares Path Modeling in International Marketing

291

marketing articles use misspeciﬁed measurement models in SEM applications. A substantial number of latent variables in these studies were inappropriately speciﬁed by treating formative measurement models as if they were reﬂective. Misspeciﬁcation of measurement models can bias inner model parameter estimation and lead to incorrect assessments of relationships (Jarvis et al., 2003). The decision to use either formative or reﬂective indicators for a construct should be based on the nature of the causal relationship between the indicators and the latent variables in the measurement model (Bollen, 1989). The most suitable approach to avoid misspeciﬁcation of measurement models in SEM is to consider the conceptual discussion of the differences between formative and reﬂective measurement models by, e.g., Howell, Breivik, and Wilcox (2000a, 2007b); Bagozzi (2007); Bollen (2007); Bollen and Lennox (1991); Diamantopoulos and Winklhofer (2001); Edwards and Bagozzi (2000), as well as the C-OAR-SE procedure (Rossiter, 2002) and the design rules for determining the speciﬁc type of measurement model put forward by Jarvis et al. (2003). Notwithstanding this theoretic foundation, there are only a few endeavors in the academic business literature that stress statistical techniques for the assessment of formative measurement models in PLS path models. For instance, Bollen and Ting’s (1993) conﬁrmatory tetrad analysis (CTA) results are used in SEM (e.g., Bollen & Ting, 2000) to assess whether manifest variables in the measurement model are independent determinants of a latent variable rather than its reﬂections in an effect indicator scale. Gudergan, Ringle, Wende, and Will (2008) show that PLS path modeling assumptions are consistent with the CTA procedure and, thereby, provide a more rigorous foundation for evaluating whether or not empirical data support a reﬂective indicator speciﬁcation rather than a formative indicator speciﬁcation.

2.3.2. Sample Size The sample size argument has its roots in the considerable obstacles faced when conducting CBSEM with small samples. A substantial number of simulation studies on CBSEM compare alternative discrepancy functions and their estimation bias, accuracy, and robustness with respect to sample size. Boomsma and Hoogland (2001), for example, conclude that there are nonconvergence problems and improper CBSEM solutions in small samples (e.g., 200 or fewer cases). These authors provide evidence that CBSEM – depending on the selected discrepancy function and the model complexity – requires several hundred or even thousands of observations.

292

JO¨RG HENSELER ET AL.

In contrast, the sample size can be considerably smaller in PLS path modeling. For example, ‘‘there can be more variables than observations and there may be a small amount of data that are missing completely at random’’ (Tenenhaus et al., 2005, p. 202). Wold (1989) illustrates the low sample size requirement by analyzing a path model based on a data set consisting of 10 observations and 27 manifest variables. A rule of thumb for robust PLS path modeling estimations suggests that the sample size be equal to the larger of the following (Barclay, Higgins, & Thompson, 1995): (1) ten times the number of indicators of the scale with the largest number of formative indicators, or (2) ten times the largest number of structural paths directed at a particular construct in the inner path model. Chin and Newsted (1999) present a Monte Carlo simulation study on PLS with small samples. They ﬁnd that the PLS path modeling approach can provide information about the appropriateness of indicators at sample size as low as 20. This study conﬁrms the consistency at large on loading estimates with increased numbers of observations and numbers of manifest variables per measurement model. As a result of these peculiarities, researchers and practitioners use PLS path modeling, instead of CBSEM, when the sample size is relatively low. However, this persistent belief in publications and research that support the claim that PLS is more efﬁcient at small sample size is inadvertently misleading the research community as it asks for accuracy instead of statistical power. Goodhue, Lewis, and Thompson (2006, p. 9) argue that statistical signiﬁcance is a primary consideration and accuracy a secondary one: ‘‘without statistical signiﬁcance, accuracy contributes no scientiﬁc knowledge.’’ Their ﬁndings suggest that PLS does not have an advantage in terms of detecting statistical signiﬁcance in small sample sizes. Furthermore, Goodhue et al. (2006) ﬁnd no evidence that PLS with bootstrapping provides more statistical power than CBSEM with small sample sizes. The generally accepted ten times rule of thumb for the minimum sample size in PLS analyses can lead to unacceptably low levels of statistical power. It is only in the case of a strong effect size (and high reliability) that rule of thumb may lead to acceptable power. However, the authors provide strong evidence that the ten-times-rule does not take into account effect size, reliability, the number of indicators, or other factors which are known to affect power. Thus, the recommendations on acceptable PLS sample size might be misleading. The choice of an appropriate sample size depends on more than the magnitude of the relationship or the desired level of power. It is evident that ‘‘a researcher must consider the distributional characteristics of the data,

Partial Least Squares Path Modeling in International Marketing

293

potential missing data, the psychometric properties of the variables examined, and the magnitude of the relationships considered before deciding on an appropriate sample size to use or to ensure that a sufﬁcient sample size is actually available to study the phenomena of interest’’ (Marcoulides & Saunders, 2006, p. VI). Marcoulides and Saunders offer other warnings that clearly echo those presented in the CBSEM literature. For instance, SchneeweiX (2001) addresses the magnitude of standard errors in PLS path modeling estimators resulting from not using enough observations (consistency) and indicators for each latent variable (consistency at large). This research provides closed form equations to determine the magnitude of ﬁnite item bias relative to the number of indicators used in a model. SchneeweiX (2001, p. 310) indicates that item bias is generally small when reﬂective measurement models involve many indicators, ‘‘each with a sizable loading and an error which is small and uncorrelated (or only slightly correlated) with other error variables.’’ Goodhue et al. (2006) stress that even though PLS path modeling seems to have no special abilities at small sample size, its performance, in terms of statistical power, is equal to other techniques for normally distributed data. In their view, PLS path modeling is still a convenient and powerful technique that is appropriate for many research situations such as complex research models with sample sizes that would be too small for CBSEM techniques. They therefore note that ‘‘[u]nfortunately PLS does not provide researchers with a magic bullet for achieving adequate statistical power at small sample sizes’’ (Goodhue et al., 2006, p. 10). In a similar vain, Marcoulides and Saunders (2006, p. VIII) state that ‘‘PLS is not a silver bullet to be used with samples of any size!’’ Thus, researchers must ensure that the sample size is large enough to support the conclusions – the PLSrelated rule of thumb might work well in some instances, but in others it might fail miserably. 2.3.3. Model Complexity Some CBSEM discrepancy functions (e.g., GFI and AGFI) decline as model complexity increases (i.e. more observed variables or more constructs), and they may be inappropriate for more complex models (Anderson & Gerbing, 1984). For example, Boomsma and Hoogland (2001) experimentally varied the model complexity (number of estimated parameters) and the number of degrees of freedom and found that the more parameters to be estimated, the more nonconvergence and improper solutions will occur, or – from an information point of view – the larger the estimation requirements, the more information is needed. However, sample size has a primary effect: With

294

JO¨RG HENSELER ET AL.

increasing numbers of observations, there is a decreasing occurrence of nonconvergence and improper solutions. According to these authors’ simulation study results, if the sample size and the factor loadings are about the same, more complications are expected regarding nonconvergence and improper solutions as model complexity increases: ‘‘It was shown that answers to these questions are conditional on data and model characteristics alike. In practice, however, applied researchers often do not know the data and the model characteristics before data collection and analysis. In new areas of applied research, especially when measurement instruments are in a developing stage, little is known about distributional characteristics of observed variables. Also, in phases of model exploration there are uncertainties about the complexity of the ‘‘ﬁnal models,’’ about the number of reliable indicators and the size of factor loadings. It is evident that with better measurements and stronger theoretical foundations of model structures, it becomes much easier to make proper decisions on the choice of estimators and the planning of sample size’’ (Boomsma & Hoogland, 2001, p. 22). This quote stresses the complementary character of PLS in respect of CBSEM in theory testing, especially with respect to the issues of model complexity when the model building proceeds from simple to more complex models. PLS is considered better suited to explain complex relationships (Fornell, 1982; Fornell, Lorange, & Roos, 1990). Wold (1985, pp. 589–590) has therefore stated that ‘‘PLS comes to the fore in larger models, when the importance shifts from individual variables and parameters to packages of variables and aggregate parameters. [...] In large, complex models with latent variables PLS is virtually without competition.’’ Moreover, the PLS algorithm allows for a considerable increase in model complexity and, hence, a noticeable reduction in the distance between subject matter analysis and statistical technique in domains with continuous access to reliable data. 2.3.4. On the Robustness of Parameter Estimates Although several authors argue that the PLS path modeling approach offers certain advantages when compared to covariance-based methods (Chin, 1998; Fornell, 1982), others provide notes of caution on the subject of this discussion (Marcoulides & Saunders, 2006). However, literature on the formal comparison of CBSEM and PLS is rare (cf. Dijkstra, 1983; McDonald, 1996). Only few simulation studies analyze the parameter estimations of both methodologies, thereby uncovering certain peculiarities regarding their behavior in applications. Researchers and practitioners require this information when selecting an appropriate means for estimating a particular SEM with their collected data.

Partial Least Squares Path Modeling in International Marketing

295

A primary Monte Carlo simulation study by Vilares, Almeida, and Coelho (2009) analyzes the effects of two assumptions on the performance of CBSEM and PLS: the symmetry of the distribution and the reﬂective modeling of the indicators. These authors compare both methods’ performance when these assumptions hold and when they are violated; for instance, when the distribution of the observations is skewed and some indicators follow a formative scheme. In their base model (reﬂective model with symmetric data), the quality of the two estimation methods is very similar, especially for the estimation of outer loadings. The Vilares et al. (2009) study sustains the phenomenon of outer loading overestimation by PLS and more conservative outcomes for inner path model relationships, whereas the Maximum Likelihood (ML) method shows exactly the opposite tendency, with overestimation of the path coefﬁcients and a general underestimation of the indicator loadings. However, when a formative latent variable is introduced, the PLS method demonstrates higher robustness compared to CBSEM. The same kind of ﬁnding holds with skewed data results. Accordingly, the authors conclude that in case of skewed data, PLS estimates are better than ML estimates, in terms of both bias and precision. The ML estimators seem to be more sensitive to the various potential deﬁciencies in the data and model speciﬁcation. The research by Vilares et al. (2009) analyzes formative measurement models in that it draws on reﬂective loadings results, thus depicting a situation of measurement model misspeciﬁcation (Jarvis et al., 2003). Another Monte Carlo simulation study by Ringle, Wilson, and Go¨tz (2007) compares the performance of CBSEM and PLS for formative exogenous latent variables in a causal model. Both methodologies provide estimates for the simulated sets of data that are very close to the population parameters when averaged. Although CBSEM estimates in the formative measurement and the structural model signiﬁcantly decrease in accuracy and robustness in respect of nonnormal data, the performance of reﬂective measurement models’ estimations is not seriously affected by these changed data characteristics. Ringle et al. (2007) reveal similar results for PLS path modeling, but the decline in accuracy and robustness is considerably lower compared to CBSEM. On the basis of these ﬁndings, we can conclude that in the normal data scenario, CBSEM provides accurate and robust parameter estimates that are equal or superior in comparison to PLS estimates, no matter what measurement models are used. However, if the premises for the application of CBSEM are violated, such as regarding the required minimum number of observations for robust model estimations or the multivariate normality assumption for some CBSEM discrepancy

JO¨RG HENSELER ET AL.

296

functions, the PLS approach offers robust approximations. It must be noted though that, in general, PLS parameter estimates are less than optimal regarding bias and consistency. The estimates will be asymptotically correct under the condition of consistency at large, i.e. both a large sample size and large numbers of indicators per latent variable (Jo¨reskog & Wold, 1982).

2.3.5. Choice between the Covariance- and Variance-Based SEM Method Even though Wold (1974) developed the ﬁxed-point estimation procedure for causal modeling as an alternative to CBSEM, the covariance-ﬁtting-based SEM procedures (e.g., ML, GLS, ULS) and the variance-based PLS approach are complementary rather than competitive. According to Jo¨reskog (1982, p. 270) ‘‘ML is theory-oriented, and emphasizes the transition from exploratory to conﬁrmatory analysis. PLS is primarily intended for causalpredictive analysis in situations of high complexity but low theoretical information.’’ The philosophical distinction between these approaches is whether to use CBSEM for theory testing and development, or PLS path modeling for predictive applications. As visualized in Fig. 4, in causal modeling situations where prior theory is strong and further testing and development is the goal, CBSEM is the most appropriate statistical methodology. Yet, due to the indeterminacy of factor score estimations, there is a loss of predictive accuracy. This occurrence, of course, is not of concern in theory testing where structural relationships (i.e. parameter estimation) between concepts is of primary concern. Moreover, hypothesis building and the assessment of CBSEM results through global goodness-of-ﬁt

Theory Testing

LISREL PLS

ANN

Prediction

Fig. 4.

Purpose of the Analysis.

Partial Least Squares Path Modeling in International Marketing

297

criteria further emphasizes the theory-testing, rather than theory-building, character of this methodology (Anderson & Gerbing, 1988). In research settings with predictive scope, weak theory, and no need for an understanding of underlying relationships, artiﬁcial neural networks (ANN) may be useful. This approach creates artiﬁcial neurons and their interrelations in the hidden layer that connects the input and output data, in order, for example, to improve predictivity without necessarily creating a model of theoretical meaning. Yet, Hsu, Chen, and Hsieh (2006) ﬁnd that their ANN-based SEM technique is similar to PLS path modeling. Using the iterative estimation technique as described in Section 3.2, the PLS path modeling approach calculates latent variable scores as exact linear combinations of the observed measures. Thereby, the approach avoids the indeterminacy problem and provides an exact deﬁnition of component scores (Fornell, 1982). The PLS approach is adequate for causal modeling applications whose purpose is prediction and/or theory building. Although PLS path modeling can be used for theory conﬁrmation, it assumes that all measured variance is useful for explanation in applications (e.g., Sarkar, Echambadi, Cavusgil, & Aulakh, 2001) and indicates the causal relationships with signiﬁcant effect. Thus, parameter estimates are obtained based on the ability to minimize the residual variances of dependent variables (both latent and observed variables). An assessment of these PLS estimation outcomes – for example, by standard errors that are obtained through bootstrapping (see Section 3.4) – builds on the evaluation of partial path model structures. PLS’ lack of a global optimization function and consequently measures of global goodness of model ﬁt deﬁnitely limits the use of PLS for theory testing. A researcher must arrive at a decision on the causal model analysis with latent variables in order to select an appropriate statistical technique. Instead of using the model to explain the covariation among all indicators, which is the objective of CBSEM, PLS path modeling maximizes the explained variance of all dependent variables and, thus, supports predictionoriented goals. Although the CBSEM and PLS path modeling methodologies differ from a statistical point of view, PLS estimates may represent good proxies of the CBSEM results. If CBSEM premises are violated, such as distributional assumptions, minimum sample size, or maximum model complexity, and related methodological matters arise, such as Heywood cases, an inability to converge to a solution, parameters that are outside reasonable limits, and large standard errors regarding parameter estimates (Rindskopf, 1984), PLS path modeling may represent a reasonable methodological alternative for theory testing.

JO¨RG HENSELER ET AL.

298

3. EVALUATION OF PLS PATH MODEL RESULTS PLS path modeling does not provide any global goodness-of-ﬁt criterion. As a consequence, Chin (1998) has put forward a catalog of criteria to assess partial model structures. A systematic application of these criteria is a two-step process, encompassing (1) the assessment of the outer model and (2) the assessment of the inner model. Fig. 5 depicts this two-step process. At the beginning of the two step process, model assessment focuses on the measurement models. A systematic evaluation of PLS estimates reveals the measurement reliability and validity according to certain criteria that are associated with formative and reﬂective outer model. It only makes sense to evaluate the inner path model estimates when the calculated latent variable scores show evidence of sufﬁcient reliability and validity.

3.1. Assessing Reﬂective Measurement Models Reﬂective measurement models should be assessed with regard to their reliability and validity. Usually, the ﬁrst criterion which is checked is internal consistency reliability. The traditional criterion for internal consistency is Cronbach’s a (Cronbach, 1951), which provides an estimate for the reliability based on the indicator intercorrelations. While Cronbach’s

Outer model assessment

Inner model assessment

Fig. 5.

• Reliability and validity of reflective constructs • Validity of formative constructs

• Variance explanation of endogenous constructs • Effect sizes • Predictive relevance

A Two-Step Process of PLS Path Model Assessment.

Partial Least Squares Path Modeling in International Marketing

299

a assumes that all indicators are equally reliable, PLS prioritizes indicators according to their reliability, resulting in a more reliable composite. As Cronbach’s a tends to provide a severe underestimation of the internal consistency reliability of latent variables in PLS path models, it is more appropriate to apply a different measure, the composite reliability rc (Werts, Linn, & Jo¨reskog, 1974). The composite reliability takes into account that indicators have different loadings, and can be interpreted in the same way as Cronbach’s a. No matter which particular reliability coefﬁcient is used, an internal consistency reliability value above 0.7 in early stages of research and values above 0.8 or 0.9 in more advanced stages of research are regarded as satisfactory (Nunnally & Bernstein, 1994), whereas a value below 0.6 indicates a lack of reliability. As the reliability of indicators varies, the reliability of each indicator should be assessed. Researchers postulate that a latent variable should explain a substantial part of each indicator’s variance (usually at least 50%). Accordingly, the absolute correlations between a construct and each of its manifest variables (i.e. the pﬃﬃﬃﬃﬃﬃ ﬃ absolute standardized outer loadings) should be higher than 0.7 ( 0:5). Moreover, some psychometrists (e.g., Churchill, 1979) recommend eliminating reﬂective indicators from measurement models if their outer standardized loadings are smaller than 0.4. Taking into account PLS’ characteristic of consistency at large, one should be careful when eliminating indicators. Only if an indicator’s reliability is low and eliminating this indicator goes along with a substantial increase of composite reliability, it makes sense to discard this indicator. For the assessment of validity, two validity subtypes are usually examined: the convergent validity and the discriminant validity. Convergent validity signiﬁes that a set of indicators represents one and the same underlying construct, which can be demonstrated through their unidimensionality. Fornell and Larcker (1981) suggest using the average variance extracted (AVE) as a criterion of convergent validity. An AVE value of at least 0.5 indicates sufﬁcient convergent validity, meaning that a latent variable is able to explain more than half of the variance of its indicators on average (e.g., Go¨tz, Liehr-Gobbers, & Krafft, 2009). Discriminant validity is a rather complementary concept: Two conceptually different concepts should exhibit sufﬁcient difference (i.e. the joint set of indicators is expected not to be unidimensional). In PLS path modeling, two measures of discriminant validity have been put forward: The Fornell–Larcker criterion and the crossloadings. The Fornell–Larcker criterion (Fornell & Larcker, 1981) postulates that a latent variable shares more variance with its assigned indicators than with any other latent variable. In statistical terms, the AVE of each latent

JO¨RG HENSELER ET AL.

300

Table 2. Criterion

Description

Composite reliability (rc )

Indicator reliability Average variance extracted (AVE)

Fornell–Larcker criterion

Cross-loadings

Assessing Reﬂective Measurement Models. P P P rc ¼ ð li Þ2 =½ð li Þ2 þ Varði Þ, where li is the outer (component) loading to an indicator, and Varði Þ ¼ 1 l2i in case of standardized indicators. The composite reliability is a measure of internal consistency and must not be lower than 0.6. Absolute standardized outer (component) loadings should be higher than 0.7. P P P AVE ¼ ð l2i Þ=½ l2i þ Varði Þ, where li is the component loading to an indicator and Varði Þ ¼ 1 l2i in case of standardized indicators. The average variance extracted should be higher than 0.5. In order to ensure discriminant validity, the AVE of each latent variable should be higher than the squared correlations with all other latent variables. Thereby, each latent variable shares more variance with its own block of indicators than with another latent variable representing a different block of indicators. Cross-loadings offer another check for discriminant validity. If an indicator has a higher correlation with another latent variable than with its respective latent variable, the appropriateness of the model should be reconsidered.

variable should be greater than the latent variable’s highest squared correlation with any other latent variable. The second criterion of discriminant validity is usually a bit more liberal: The loading of each indicator is expected to be greater than all of its cross-loadings (Chin, 1998; Go¨tz et al., 2009). Although the Fornell–Larcker criterion assesses discriminant validity on the construct level, the cross-loadings allow this kind of evaluation on the indicator level. In summary, a reliable and valid reﬂective measurement of latent variables should meet all the criteria as listed in Table 2. If this is not the case, the researcher may want to exclude single indicators from a speciﬁc measurement model and eventually revise the path model.

3.2. Assessing Formative Measurement Models Bollen (1989) and Bagozzi (1994) emphasize that traditional validity assessments and classical test theory do not apply to manifest variables that are used in formative measurement models and that the concepts of reliability (i.e. internal consistency) and construct validity (i.e. convergent

Partial Least Squares Path Modeling in International Marketing

301

and discriminant validity) are not meaningful when a formative mode is employed. It is the assumption of error-free measures that renders the question of indicator reliability irrelevant. Diamantopoulos (2006, p. 11) concludes in respect of formative measurement that ‘‘reliability becomes an irrelevant criterion for assessing measurement quality.’’ The less important reliability becomes, the more pivotal it is to secure validity. A ﬁrst examination of the validity of formative indicators should use theoretic rationale and expert opinion (Rossiter, 2002). A second assessment of the validity of formative constructs should consist of statistical analyses on two levels: the construct level and the indicator level. At the construct level, the question arises as to whether the formative index indeed carries the intended meaning. A ﬁrst check could be the nomological validity: Does the formative index behave within a net of hypotheses as expected? Those relationships between the formative index and other constructs in the path model that are sufﬁciently referred to in prior research should be strong and signiﬁcant. Moreover, researchers should be concerned about the construct’s error-term n, which represents the part of the construct that is not captured by any indicator. In order to estimate the size of this error, one can assess the external validity by means of regressing the formative index on a reﬂective measure of the same construct. The variance of the error n can be determined as follows: VarðnÞ ¼ 1

g2 relðxÞ

(7)

Here, x stands for the reﬂective measure of the focal construct and g for the correlation between the formative and the reﬂective measure of the same construct, which is equal to the standardized regression coefﬁcient. The ratio of the equation includes a correction of the imperfect reliability of the reﬂective measure. The external validity can ﬁnally be calculated as 1 VarðnÞ. The current literature does not provide any recommendations for thresholds of external validity. We suggest a value of 0.8 as a minimum for external validity, which would mean that the formative index carries about 80% of the intended meaning. However, researchers should always take into account the particularities of their research ﬁelds and eventually ask for a higher external validity. At the indicator level, the question arises as to whether each indicator indeed delivers a contribution to the formative index by carrying the intended meaning. Besides face and content validity, which can both be assessed a priori, some statistical evaluations can be conducted a posteriori.

JO¨RG HENSELER ET AL.

302

There are two cases in which researchers should critically examine whether a particular indicator should enter into the index or not. An indicator can be irrelevant for the construction of the formative index because it either does not have a signiﬁcant impact on the formative index, or because it exhibits high multicollinearity, which could mean that the indicator’s information is redundant. In order to check for the ﬁrst case, the signiﬁcance of the estimated indicator weights can be determined by means of bootstrapping (cf. Davison & Hinkley, 2003; Chin, 1998; Tenenhaus et al., 2005, and see Section 3.4). In order to check for the second case, researchers should assess the degree of multicollinearity among the formative indicators (Diamantopoulos & Winklhofer, 2001; Cassel, Hackl, & Westlund, 2000, Grewal, Cote, & Baumgartner, 2004), for instance, by calculating the variance inﬂation factor (VIF) or the tolerance values. A rule of thumb from econometrics states that VIFs greater than 10 reveal a critical level of multicollinearity. However, any VIF substantially greater than 1 indicates multicollinearity and should alert researchers to the typical problems of multicollinearity. The different steps and criteria for assessing the validity of formative indices are summarized in Table 3. A ﬁnal note of caution: formative indicators should never be discarded simply on the basis of statistical outcomes. Such actions may substantially change the content of the formative index (Jarvis et al., 2003). Thus, the researcher should keep both signiﬁcant and insigniﬁcant formative indicators in the measurement model as long as this is conceptually justiﬁed. Usually, PLS structural model Table 3. Criterion Nomological validity

External validity Signiﬁcance of weights Multicollinearity

Assessing Formative Measurement Models. Description The relationships between the formative index and other constructs in the path model, which are sufﬁciently well known through prior research, should be strong and signiﬁcant. The formative index should explain a big part of the variance of an alternative reﬂective measure of the focal construct. Estimated weights of formative measurement models should be signiﬁcant. Manifest variables in a formative block should be tested for multicollinearity. The variance inﬂation factor (VIF) can be used for such tests. As a rule of thumb, a VIF greater than ten indicates the presence of harmful collinearity. However, any VIF substantially greater than one indicates multicollinearity.

Partial Least Squares Path Modeling in International Marketing

303

estimates hardly alter after performing an elimination of insigniﬁcant or highly collinear formative indicators, providing further support for the decision to retain such indicators in the PLS path model. In the case of multicollinearity, indicator weight estimates can be distorted. This fact requires researchers to be particularly cautious when interpreting indicator weights as a sign of indicator importance. 3.3. Assessing the Structural Model Reliable and valid outer model estimations permit an evaluation of the inner path model estimates. Table 4 provides an overview of the typical criteria. The essential criterion for this assessment is the coefﬁcient of determination (R2 ) of the endogenous latent variables. Chin (1998) describes R2 values of 0.67, 0.33, and 0.19 in PLS path models as substantial, moderate, and weak, respectively. If certain inner path model structures explain a endogenous latent variable by only a few (e.g., one or two) exogenous latent variables, ‘‘moderate’’ R2 may be acceptable. However, if the endogenous latent variable relies on several exogenous latent variables, the R2 value should exhibit at least a substantial level. Lower results, on the contrary, cast Table 4. Criterion R2 of endogenous latent variables Estimates for path coefﬁcients Effect size f 2

Prediction relevance (Q2 and q2 )

Assessing Structural Models. Description

R2 values of 0.67, 0.33, or 0.19 for endogenous latent variables in the inner path model are described as substantial, moderate, or weak by Chin (1998, p. 323). The estimated values for path relationships in the structural model should be evaluated in terms of sign, magnitude, and signiﬁcance (the latter via bootstrapping). f 2 ¼ ðR2included R2excluded Þ=ð1 R2included Þ; values of 0.02, 0.15, and 0.35 can be viewed as a gauge for whether a predictor latent variable has a weak, medium, or large effect at the structural level. The Q2 is calculated based on the blindfolding procedure: P P Q2 ¼ 1 ð D SSED Þ=ð D SSOD Þ. D is the omission distance, SSE is the sum of squares of prediction errors, and SSO is the sum of squares of observations. Q2 -values above zero give evidence that the observed values are well reconstructed and that the model has predictive relevance (Q2 -values below zero indicate a lack of predictive relevance). In correspondence to f 2 , the relative impact of the structural model on the observed measures for latent dependent variables can be assessed: q2 ¼ ðQ2included Q2excluded Þ=ð1 Q2included Þ.

304

JO¨RG HENSELER ET AL.

doubts regarding the theoretical underpinnings and demonstrate that the model is incapable to explain the endogenous latent variable(s). The individual path coefﬁcients of the PLS structural model can be interpreted as standardized beta coefﬁcients of ordinary least squares regressions. Structural paths, whose sign is in keeping with a priori postulated algebraic signs, provide a partial empirical validation of the theoretically assumed relationships between latent variables. Paths that possess an algebraic sign contrary to expectations do not support the a priori formed hypotheses. In order to determine the conﬁdence intervals of the path coefﬁcients and statistical inference, resampling techniques such as bootstrapping should be used (cf. Tenenhaus et al., 2005, and see Section 3.4). Another important evaluation of direct and indirect relationships of the predecessor of a certain endogenous latent variable involves the analysis of mediating (Helm, Eggert, & Garnefeld, 2009) and moderating effects (Henseler & Fassott, 2009). Researchers and practitioners using PLS path modeling should ﬁrst assess their hypothesized path model of direct effects and then conduct additional analyses involving mediating and moderating effects to learn, for instance, more about possible spurious effects or suppressor effects. Albers (2009) proposes a new paradigm for success factor studies in marketing. The signiﬁcance of highly plausible direct inner path model relationships is no longer of interest to researchers and practitioners. Rather, the sum of the direct effect and all indirect effects of a particular latent variable on another (the total effect) should be the subject of evaluation for further interpretation. This new paradigm copes with a frequent observation in PLS path modeling that the standardized inner path model coefﬁcients decline with an increased number of indirect relationships, especially when mediating latent variables have a suppressor effect on the direct path. Consequently, considerable direct relationships may become insigniﬁcant after including additional indirect relationships. In such instances, the total effect should remain at a relatively constant, sizeable level and thus provide more reasonable grounds for conclusions on the inner path model relationships. For each effect in the path model, one can evaluate the effect size by means of Cohen’s (1988) f 2 . The effect size f 2 is calculated as the increase in R2 relative to the proportion of variance of the endogenous latent variable that remains unexplained. According to Cohen (1988), f 2 values of 0.02, 0.15, and 0.35 signify small, medium, and large effects, respectively. Another assessment of the structural model involves the model’s capability to predict. The predominant measure of predictive relevance is

Partial Least Squares Path Modeling in International Marketing

305

Stone-Geisser’s Q2 (Stone, 1974; Geisser, 1975), which can be measured using blindfolding procedures (Tenenhaus et al., 2005). The Stone–Geisser criterion postulates that the model must be able to provide a prediction of the endogenous latent variable’s indicators. The technique represents a synthesis of function ﬁtting and cross-validation. As Chin (1998) points out, ‘‘the prediction of observables or potential observables is of much greater relevance than the estimator of what are often artiﬁcial construct-parameters’’ (p. 320). Wold (1982) argues that the sample reuse technique – especially the blindfolding procedure to obtain the crossvalidated redundancy (instead of the cross-validated communality) – ﬁts the PLS path modeling approach like ‘‘hand in glove.’’ The blindfolding procedure is only applied to endogenous latent variables that have a reﬂective measurement model operationalization. If this value for a certain endogenous latent variable is larger than zero, its explanatory variables provide predictive relevance. In analogy to the effect-size f 2 evaluation, the relative impact of the predictive relevance can be assessed by means of the measure q2 : values of 0.02, 0.15, and 0.35 reveal a small, medium, or large predictive relevance of a certain latent variable, thus explaining the endogenous latent variable under evaluation.

3.4. Bootstrapping The nonparametric bootstrap (Davison & Hinkley, 2003; Efron & Tibshirani, 1993) procedure can be used in PLS path modeling to provide conﬁdence intervals for all parameter estimates, building the basis for statistical inference. In general, the bootstrap technique provides an estimate of the shape, spread, and bias of the sampling distribution of a speciﬁc statistic. Bootstrapping treats the observed sample as if it represents the population. The procedure creates a large, pre-speciﬁed number of bootstrap samples (e.g., 5,000). Each bootstrap sample should have the same number of cases as the original sample. Bootstrap samples are created by randomly drawing cases with replacement from the original sample. PLS estimates the path model for each bootstrap sample. The obtained path model coefﬁcients form a bootstrap distribution, which can be viewed as an approximation of the sampling distribution. The bootstrapping analysis allows for the statistical testing of the hypothesis H 0 : w ¼ 0 (w can be any parameter estimated by PLS) against the alternative hypothesis H 1 : w a 0 at m þ n 2 degrees of freedom (where m is the number of PLS estimates for w in the original sample, which is 1; n is the number of

JO¨RG HENSELER ET AL.

306

bootstrap estimates for w, e.g., 5,000). The PLS results for all bootstrap samples provide the mean value and standard error for each path model coefﬁcient. This information permits a student’s t-test to be performed for the signiﬁcance of path model relationships. Chin (1998) proposes using the following test statistic for PLS: w (8) temp ¼ seðwÞ whereby temp represents the empirical t-value, w the original PLS estimate of a certain path coefﬁcient, and seðwÞ its bootstrapping standard error. The student’s t-distribution table provides the critical t-value at given a-levels and the respective number of degrees of freedom. Instead of just reporting the signiﬁcance of a parameter, it would be more valuable to report the conﬁdence interval. If a conﬁdence interval for an estimated path coefﬁcient w does not include zero, the hypothesis that w equals zero is rejected. Testing with conﬁdence intervals is advantageous, as it provides more information about the estimate of w. Shaffer (1995, p. 575) remarks that ‘‘if the hypothesis is not rejected, the power of the procedure can be gauged by the width of the interval.’’ In order to determine the conﬁdence interval of a model parameter, say a path coefﬁcient, one has to recognize that the bootstrapping procedure generates a distribution T of w. However, signiﬁcance testing needs to account for the distribution of T under H 0 , which is the null distribution of T. This distribution is essential for performing the test since it provides the basis for determining the p-value. If the null distribution is unknown, the asymptotic normality of the test statistic is commonly assumed. Although nonparametric alternatives can be obtained through bootstrapping, this approach overlooks the possibility that the data could be generated under the alternative hypothesis H 1 : y a y0 . This problem can be addressed by examining the statistical correspondence between tests of signiﬁcance and conﬁdence intervals when the null hypothesis concerns a particular parameter value (Gudergan et al., 2008). Bootstrapping conﬁdence intervals are well established (Efron & Tibshirani, 1993; Davison & Hinkley, 2003). If the bootstrap estimates of bias and variance are denoted by bB and vB , then the corresponding approximate 1 a (two-tailed) conﬁdence interval is: 1=2

t bB vB z1a=2

(9)

where bB ¼ w B w, the difference of the mean value of w for all bootstraps and the original PLS estimation of that path coefﬁcient. A null hypothesis

Partial Least Squares Path Modeling in International Marketing

307

H 0 : w ¼ 0 is accepted (or rejected) at a given level a, if the corresponding ð1 aÞ conﬁdence interval includes (or does not include) the parameter value y0 . The bias-corrected bootstrap conﬁdence interval in Eq. (9) provides a basis to account for the aforementioned problem and, thus, can be used as an appropriate means to test the signiﬁcance of PLS-estimated path coefﬁcients (Gudergan et al., 2008). In PLS path modeling, the latent variable scores may have different signs if, for examples, estimation A uses an alternative outer weights initialization of the algorithm (say, þLV 1 ) compared with estimation B (in this case, LV 2 ). Still, the absolute LV-values of both estimations usually represent equivalent solutions, i.e. jþLV 1 j ¼ jLV 2 j (for a discussion on this sign indeterminacy of PLS see Wold, 1985). This sign indeterminacy of PLS latent variable scores may also result in arbitrary sign changes during the bootstrap path model estimations. Such occurrence of sign changes ‘‘pull’’ the mean value of bootstrapping results (e.g., for parameter w) toward zero and, thereby, bias the bootstrap standard error upward. Arbitrary sign changes systematically reduce the value of temp and, thus, the possibility of the rejection of H 0 at a given a level. The standard error of estimates increases dramatically without any real meaning if the sign changes are not properly taken into account (Tenenhaus et al., 2005). Rather than not treating sign changes in the bootstrapping results, we suggest using the individual sign change option. The working principle is as follows: If a PLS path coefﬁcient estimation for a bootstrap subsample shows a different sign compared with the original path model estimation, the procedure reverses the sign of that path coefﬁcient in the bootstrapping subsample. Thus, the signs in the outer and inner models of each resample are made consistent with the signs in the original sample in order to avoid these sign-changerelated problems.

3.5. Group Comparisons and Other Advances in PLS Analyses A threat to the validity of SEM results may lie in the heterogeneity of observations. The level of observed or unobserved heterogeneity can lead to spurious or suppressor effects, and generally to misinterpretations. Typically, researchers in international marketing are well aware of the heterogeneity of observations; that is, different population parameters are likely for different subpopulations such as countries or cultures. Although cross-cultural or cross-national differences are related to observed heterogeneity, there can also be unobserved heterogeneity, which cannot be attributed to any

308

JO¨RG HENSELER ET AL.

predetermined variable. Similar to ignoring observed heterogeneity due to countries or cultures, unobserved heterogeneity represents a serious problem in respect of interpreting PLS results (Ringle, 2006). Our review of PLS path modeling applications in international marketing gives a mixed picture regarding the examination of heterogeneity. Although a signiﬁcant number of studies take observed heterogeneity into account by means of multigroup comparisons, none of the studies accounts for unobserved heterogeneity. However, there are several PLS-based approaches to detect unobserved heterogeneity. Analytical techniques such as ﬁnite mixture partial least squares (FIMIX-PLS; Hahn, Johnson, Herrmann, & Huber, 2002; Ringle, Sarstedt, & Mooi, 2009a; Ringle, Wende, & Will, 2009b) or the recently developed PLS typological alternatives (e.g., Esposito Vinzi, Ringle, Squillacciotti, & Trinchera, 2007; Ringle & Schlittgen, 2007) are becoming mandatory for evaluating PLS path modeling results. Using these techniques, researchers can either conﬁrm that their results are not distorted by unobserved heterogeneity or identify thus far neglected variables that describe the uncovered data segments. As soon as conceptual deliberations or methods such as FIMIX-PLS suggest different models for different subgroups, PLS-based multigroup analysis (MGA) should be conducted. In MGA, a population parameter b is hypothesized to differ for two subpopulations: bð1Þ a bð2Þ . Referring to PLS path modeling, researchers would for instance ask whether one could conclude differences in population parameters from the differences in path coefﬁcients. The primary approach for group comparisons is a t-test, as described by Keil et al. (2000). These authors suggest using the standard errors obtained from bootstrapping as the input for a parametric test. After having exposed the subsamples to separate bootstrap analyses and having made parametric assumptions about the distributions of the parameter standard errors, one can calculate the following statistic for the difference in paths between groups (Keil et al., 2000): bð1Þ bð2Þ t ¼ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃrﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ðnð1Þ 1Þ2 ðnð2Þ 1Þ2 1 1 ð1Þ 2 seðb seðbð2Þ Þ2 Þ þ þ nð1Þ þ nð2Þ 2 nð1Þ þ nð2Þ 2 nð1Þ nð2Þ

(10)

This statistic is asymptotically t-distributed with nð1Þ þ nð2Þ 2 degrees of freedom. The subsample-speciﬁc path coefﬁcients are denoted as b, the sizes of the subsamples as n, and the path coefﬁcient standard errors as resulting from bootstrapping as se.

Partial Least Squares Path Modeling in International Marketing

309

Chin and Dibbern (2009) cast doubt on whether this approach of multigroup comparison with its inherent distributional assumptions ﬁts to PLS path modeling, which is generally regarded as being distribution-free. In line with these doubts, we present an alternative approach to PLS-based group comparisons that does not rely on distributional assumptions, as initially proposed by Henseler (2007). The working principle of the novel PLS-MGA approach is as follows: First, the subsamples to be compared are exposed to separate bootstrap analyses, and the bootstrap outcomes serve as a basis for the hypothesis tests of group differences. Instead of relying on distributional assumptions, the new approach evaluates the observed distribution of the bootstrap outcomes. Given two subsamples with parameter estimates (e.g., a path coefﬁcient), bð1Þ and bð2Þ , the conditional probability Pðbð1Þ 4bð2Þ jbð1Þ bð2Þ Þ has to be determined. Here, bð1Þ and bð2Þ represent the true population parameters of population 1 and 2. A researcher would like to be sure that P is below a speciﬁed a-level before concluding that bð1Þ is greater than bð2Þ . If researchers know the parameter estimates of two subsamples from bootstrapping, they can easily verify how probable a difference in parameters between two subpopulations is and hence test their hypothesis with the following equation: ð1Þ ð2Þ ð1Þ ð2Þ X Y 2b bj 2b þ bi (11) Pðbð1Þ 4bð2Þ jbð1Þ bð2Þ Þ ¼ 1 J2 8j;i ð2Þ In this equation, J denotes the number of bootstrap samples, bð1Þ j and bi ð1Þ ð2Þ the bootstrap parameter estimates, b and b the means of the focal parameters over the bootstrap samples, and Y the unit step function, which has a value of 1 if its argument exceeds 0, otherwise 0. The superscript in parentheses marks the respective group. This equation states that J 2 (i.e. all possible) comparisons of bootstrap parameters have to be made. In fact, this approach can be seen as a Mann–Whitney–Wilcoxon test (Wilcoxon, 1947; Mann & Whitney, 1947) applied to the bootstrap-values corrected for the original parameter values. PLS-MGA does not require any distributional assumptions and is simple to apply by using the bootstrap outputs that are generated by the prevailing PLS implementations such as SmartPLS (Ringle, Wende, & Will, 2005). The way of determining the probability whether a population parameter differs across two sub-populations is unique to this new approach. Researchers can easily conduct the ﬁnal calculations with available spreadsheet software such as MS Excel (a sample spreadsheet can be obtained from the ﬁrst author).

310

JO¨RG HENSELER ET AL.

Besides the aforementioned methodological advances for addressing the requirement to capture unobserved heterogeneity, recent developments in the evaluation of PLS path modeling results focus on the problem of measurement model misspeciﬁcation. Bucic and Gudergan (2004) use a PLS application to demonstrate the signiﬁcant consequences of misspeciﬁcation for path model estimates. The seven criteria that have been put forward by Jarvis et al. (2003) permit an a priori examination of the outer mode. In addition, a conﬁrmatory tetrad measurement model analysis of PLS (CTA-PLS; Gudergan et al., 2008) provides a statistical test to evaluate the outer measurement model a posteriori. Statistical support is there for specifying not only the measurement model but also the structural model. Research by Marcoulides (2003) and Wilson, Callaghan, Ringle, and Henseler (2007) addresses the inner path model speciﬁcation search to uncover superior structures and more appropriate inner relationship directions. Additional advances on the subject of evaluating path modeling results involve the development of global goodness-of-ﬁt criteria. The PLS goodness-of-ﬁt proposal by Tenenhaus et al. (2005) is the geometric mean of the average communalities (outer measurement model) and the average R2 of endogenous latent variables, and is normed between 0 and 1, where a higher value represents better path model estimations. However, while this criterion has only been applied in few applications (e.g., Esposito Vinzi et al., 2007) and has not been systematically analyzed in simulation studies, it is restricted to reﬂective outer models and may be subject to systematic improvement by means of path model modiﬁcations/manipulations. Future developments in this line of research may combine PLS and ML CBSEM global goodness-of-ﬁt criteria. This procedure is well established in CBSEM for the asymptotic distribution-free estimators (ADF), which—like PLS— does not follow distributional assumptions that allow the global goodnessof-ﬁt to be computed. Consequently, researchers and practitioners should note the substantial link between CBSEM and PLS path modeling for which the literature has called over the last decades.

4. SUMMARY AND CONCLUSION In international marketing research, both CBSEM and PLS provide a powerful framework for estimating causal models with latent variables and systems of simultaneous equations with measurement errors. CBSEM and PLS path modeling constitute two complementary, yet distinctive,

Partial Least Squares Path Modeling in International Marketing

311

statistical techniques for estimating parameters of conceptual models. A critical review of the PLS application in international marketing reveals that this methodology has increased in popularity, especially for multigroup analyses of PLS results for different nations. Our paper illustrates the PLS path modeling methodology and typical criteria for the assessment of results. Moreover, we introduce several advances in evaluating PLS model estimations, such as a novel approach for multigroup comparison (PLS-MGA). PLS is based on least squares estimation with the primary objective being to maximize the explanation of variance in a structural equation model’s dependent constructs. Jo¨reskog and Wold (1982, p. 270) suggest that ‘‘PLS is primarily intended for causal-predictive analysis in situations of high complexity but low theoretical information’’. In contrast, the primary measures used in CBSEM are overall goodness-of-ﬁt measures that assess how well the hypothesized model ﬁts the observed data. The model estimation is theory-oriented and emphasizes the conﬁrmatory, rather than the exploratory, analysis. Consequently, in international marketing research, CBSEM should be used either to empirically conﬁrm a system of hypotheses that underlie a causal model or to test and compare results for alternative theoretically established causal models. The prediction-oriented PLS method, on the contrary, does not require strong theory and can be used as a theory-building method (Gefen et al., 2000). PLS offers excellent capabilities for work with small samples and formative measurement, as the methodology is sufﬁcient for most success factor (cause indicator) analyses in international marketing research. Moreover, complementary analytical techniques such as FIMIX-PLS for uncovering unobserved heterogeneity or multigroup PLS analysis may provide further differentiated path modeling results that allow for more precise interpretation. The review of PLS studies in international marketing reveals that few studies conduct multigroup analyses to identify differences in path coefﬁcients across subgroups. No study has addressed the potentially serious problems of ignoring unobserved heterogeneity (Ringle, 2006). Owing to the severity of the consequences of neglecting unobserved heterogeneity, such additional analyses should become a standard means of evaluating PLS path modeling results, especially in international marketing research. A ﬁnal concern refers to the choice of SEM method. There may be situations in which CBSEM is preferable, in other situations PLS may be preferable. Moreover, there may be situations where using CBSEM is desirable but unobtainable, for example, due to violations in some key CBSEM assumptions (e.g., regarding sample size, distribution, and model

JO¨RG HENSELER ET AL.

312

identiﬁcation); in such cases, PLS may provide a realistic alternative to CBSEM. Some authors also critically point out that both CBSEM and PLS follow the classical test theory paradigm, while especially for international marketing applications alternative methodologies following the probabilistic test theory may be advisable (Ewing, Salzberger, & Sinkovics, 2005). All in all, the ﬁndings of our paper are deemed to help researchers in international marketing to adopt a more holistic perspective and to make informed decisions about the SEM method based on the nature and key objectives of their study.

REFERENCES Acedo, F. J., & Jones, M. V. (2007). Speed of internationalization and entrepreneurial cognition: Insights and a comparison between international new ventures, exporters and domestic ﬁrms. Journal of World Business, 42(3), 236–252. Ainuddin, R. A., Beamish, P. W., Hulland, J. S., & Rouse, M. J. (2007). Resource attributes and ﬁrm performance in international joint ventures. Journal of World Business, 42(1), 47–60. Albers, S. (2009). PLS and success factor studies in marketing. In: V. Esposito Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods, and applications. Berlin: Springer (in print). Alpert, F., Kamins, M., Sakano, T., Onzo, N., & Graham, J. (2001). Retail buyer beliefs, attitude, and behavior toward pioneer and me-too follower brands. International Marketing Review, 18(2), 160–187. Anderson, J. C., & Gerbing, D. W. (1984). The effect of sampling error on convergence, improper solutions, and goodness of ﬁt indices for maximum likelihood conﬁrmatory factor analysis. Psychometrika, 49(2), 155–173. Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411–423. Bagozzi, R. P. (1994). In: Principles of marketing research (pp. 317–385). Oxford: Blackwell. Bagozzi, R. P. (2007). On the meaning of formative measurement and how it differs from reﬂective measurement: Comment on Howell, Breivik, and Wilcox (2007). Psychological Methods, 12(2), 229–237. Barclay, D. W., Higgins, C., & Thompson, R. (1995). The partial least squares approach to causal modeling: Personal computer adoption and use as illustration. Technology Studies, 2(2), 285–309. Birkinshaw, J., Morrison, A., & Hulland, J. (1995). Structural and competitive determinants of a global integration strategy. Strategic Management Journal, 16(8), 637–655. Blalock, H. M. (1971). Causal models in the social sciences. Chicago, IL: Aldine-Atherton. Bollen, K. A. (1989). Structural equations with latent variables. New York, NY: Wiley. Bollen, K. A. (2007). Interpretational confounding is due to misspeciﬁcation, not to type of indicator: Comment on Howell, Breivik, and Wilcox (2007). Psychological Methods, 12(2), 219–228.

Partial Least Squares Path Modeling in International Marketing

313

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. (1993). Conﬁrmatory tetrad analysis. Sociological Methodology, 23, 147–175. Bollen, K. A., & Ting, K. (2000). A tetrad test for causal indicators. Psychological Methods, 5(1), 3–22. Boomsma, A., & Hoogland, J. J. (2001). The robustness of LISREL modeling revisited. In: R. Cudeck, S. du Toit & D. So¨rbom (Eds), Structural equation modeling: Present and future (pp. 139–168). Chicago, IL: Scientiﬁc Software International. Bucic, T., & Gudergan, S. P. (2004). The impact of organizational settings on creativity and learning in alliances. Management, 7(3), 257–273. Calantone, R. J., Graham, J. L., & Mintu-Wimsatt, A. (1998). Problem-solving approach in an international context: Antecedents and outcomes. International Journal of Research in Marketing, 15(1), 19–35. Cassel, C. M., Hackl, P., & Westlund, A. H. (2000). On measurement of intangible assets: A study of robustness of partial least squares. Total Quality Management, 11(7), 897–907. Chin, W. W. (1998). The partial least squares approach to structural equation modeling. In: G. A. Marcoulides (Ed.), Modern Methods for Business Research (pp. 295–358). Mahwah, NJ: Lawrence Erlbaum Associates. Chin, W. W., & Dibbern, J. (2009). A permutation based procedure for multi-group (PLS) analysis: Results of tests of differences on simulated data. In: V. Esposito Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods and applications in marketing and related ﬁelds. Springer (In press). 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. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Cronbach, L. J. (1951). Coefﬁcient alpha and the internal structure of tests. Psychometrika, 16(3), 297–334. Davison, A. C., & Hinkley, D. V. (2003). Bootstrap methods and their application (2nd ed.). New York, NY: Cambridge University Press. Diamantopoulos, A. (2006). The error term in formative measurement models: Interpretation and modeling implications. Journal of Modelling in Management, 1(1), 7–17. Diamantopoulos, A., & Winklhofer, H. (2001). Index construction with formative indicators: An alternative to scale development. Journal of Marketing Research, 38(2), 269–277. Dibbern, J., Goles, T., Hirschheim, R. A., & Jayatilaka, B. (2004). Information systems outsourcing: A survey and analysis of the literature. The Data Base for Advances in Information Systems, 35(4), 6–102. Dijkstra, T. K. (1983). Some comments on maximum likelihood and partial least squares methods. Journal of Econometrics, 22(1/2), 67–90. Dijkstra, T. K. (2009). Latent variables and indices: Herman Wold’s basic design and partial least squares. In: V. Esposito Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods, and applications. Berlin: Springer (in print).

314

JO¨RG HENSELER ET AL.

Edwards, J. R., & Bagozzi, R. P. (2000). On the nature and direction of relationships between constructs and measures. Psychological Methods, 5(2), 155–174. Efron, B., & Tibshirani, R. J. (1993). An introduction to the bootstrap. New York, NY: Chapman Hall. Eggert, A. (2007). Getting your PLS research published: A personal and interpersonal perspective. In: Proceedings of the academy of marketing science 2007 world marketing congress, July 11–14, Verona, Italy. Esposito Vinzi, V., Ringle, C. M., Squillacciotti, S., & Trinchera, L. (2007). Capturing and treating unobserved heterogeneity by response based segmentation in PLS path modeling: A comparison of alternative methods by computational experiments. ESSEC Research Center Working Paper 07019, ESSEC Business School Paris-Singapore, Paris. Ewing, M. T., Salzberger, T., & Sinkovics, R. R. (2005). An alternate approach to assessing cross-cultural measurement equivalence in advertising research. Journal of Advertising, 34(1), 17–36. Falk, R. F., & Miller, N. B. (1992). A primer for soft modeling. Akron, OH: University of Akron Press. Festge, F., & Schwaiger, M. (2007). The drivers of customer satisfaction with industrial goods: An international study. In: C. R. Taylor & D.-H. Lee (Eds), Advances in international marketing (Vol. 18, pp. 179–207). Amsterdam: Elsevier. Fornell, C. (1982). A second generation of multivariate analysis: An overview. In: C. Fornell (Ed.), A second generation of multivariate analysis (Vol. 1, pp. 1–21). New York, NY: Praeger. Fornell, C. (1992). A national customer satisfaction barometer: The Swedish experience. Journal of Marketing, 56(1), 6–21. Fornell, C., & Bookstein, F. (1982). A comparative analysis of two structural equation models: LISREL and PLS applied to market data. In: C. Fornell (Ed.), A second generation of multivariate analysis (Vol. 1, pp. 289–323). New York, NY: Praeger. Fornell, C., & Larcker, D. F. (1981). Structural equation models with unobservable variables and measurement error: Algebra and statistics. Journal of Marketing Research, 18(3), 328–388. Fornell, C., Lorange, P., & Roos, J. (1990). The cooperative venture formation process: A latent variable structural modeling approach. Management Science, 36(10), 1246–1255. Fornell, C., & Robinson, W. T. (1983). Industrial organization and consumer satisfaction/ dissatisfaction. Journal of Consumer Research, 9(4), 403–412. Gefen, D., Straub, D. W., & Boudreau, M.-C. (2000). Structural equation modelling and regression: Guidelines for research practice. Communications of the Association for Information Systems, 4(7), 1–78. Geisser, S. (1975). A predictive approach to the random effect model. Biometrika, 61(1), 101–107. Gerpott, T., & Jakopin, N. (2005). International marketing standardization and ﬁnancial performance of mobile network operators – An empirical analysis. Schmalenbach Business Review, 57(3), 198–228. Goodhue, D., Lewis, W., & Thompson, R. (2006). PLS, small sample size, and statistical power in MIS research. In: HICSS ‘06: Proceedings of the 39th annual Hawaii international conference on system sciences, IEEE Computer Society, Washington, DC, pp. 202.2, CD-ROM, 10 pages. Go¨tz, O., Liehr-Gobbers, K., & Krafft, M. (2009). Evaluation of structural equation models using the partial least squares (PLS) approach. In: V. Esposito Vinzi, W. W. Chin,

Partial Least Squares Path Modeling in International Marketing

315

J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods, and applications. Berlin: Springer (in print). Graham, J. L., Mintu, A. T., & Rodgers, W. (1994). Explorations of negotiation behaviors in ten foreign cultures using a model developed in the United States. Management Science, 40(1), 72–95. Green, D. H., & Ryans, A. B. (1990). Entry strategies and market performance: Causal modeling of a business simulation. Journal of Product Innovation Management, 7(1), 45–58. Grewal, R., Cote, J. A., & Baumgartner, H. (2004). Multicollinearity and measurement error in structural equation models: Implications for theory testing. Marketing Science, 23(4), 519–529. Gudergan, S. P., Ringle, C. M., Wende, S., & Will, A. (2008). Conﬁrmatory tetrad analysis for evaluating the mode of measurement models in PLS path modeling. Journal of Business Research, (in print). Gustafsson, A., & Johnson, M. D. (2004). Determining attribute importance in a service satisfaction model. Journal of Service Research, 7(2), 124–141. Hahn, C., Johnson, M. D., Herrmann, A., & Huber, F. (2002). Capturing customer heterogeneity using a ﬁnite mixture PLS approach. Schmalenbach Business Review, 54(3), 243–269. Helm, S., Eggert, A., & Garnefeld, I. (2009). Modelling the impact of corporate reputation on customer satisfaction and loyalty using PLS. In: V. Esposito Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods, and applications. Berlin: Springer (in print). Henseler, J. (2007). A new and simple approach to multi-group analysis in partial least squares path modeling. In: H. Martens, T. Næs & M. Martens (Eds), PLS’07 international symposium on PLS and related methods – causalities explored by indirect observation (pp. 104–107). Norway: Matforsk, A˚s. Henseler, J., & Fassott, G. (2009). Testing moderating effects in PLS path models: An illustration of available procedures. In: V. Esposito Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods, and applications. Berlin: Springer (in print). Higgins, C. A., Duxbury, L. E., & Irving, R. H. (1992). Work-family conﬂict in the dual-career family. Organizational Behavior and Human Decision Processes, 51(1), 51–75. Holzmu¨ller, H. H., & Kasper, H. (1991). On a theory of export performance: Personal and organizational determinants of export trade activities observed in small and mediumsized ﬁrms. Management International Review, 31(Special Issue), 45–70. Holzmu¨ller, H. H., & Sto¨ttinger, B. (1996). Structural modeling of success factors in exporting: Cross-validation and further development of an export performance model. Journal of International Marketing, 4(2), 29–55. Howell, R. D., Breivik, E., & Wilcox, J. B. (2007a). Is formative measurement really measurement? Reply to Bollen (2007) and Bagozzi (2007). Psychological Methods, 12(2), 238–245. Howell, R. D., Breivik, E., & Wilcox, J. B. (2007b). Reconsidering formative measurement. Psychological Methods, 12(2), 205–218. Hsu, H.-H., Chen, W.-H., & Hsieh, M.-J. (2006). Robustness testing of PLS, LISREL, EQS and ANN-based SEM for measuring customer satisfaction. Total Quality Management and Business Excellence, 17(3), 355–372. Hulland, J. S. (1999). Use of partial least squares (PLS) in strategic management research: A review of four recent studies. Strategic Management Journal, 20(4), 195–204.

316

JO¨RG HENSELER ET AL.

Iacobucci, D., Grisaffe, D., Duhachek, A., & Marcati, A. (2003). FAC-SEM: A methodology for modeling factorial structural equations models, applied to cross-cultural and crossindustry drivers of customer evaluations. Journal of Service Research, 6(1), 3–23. Inkpen, A. C., & Birkenshaw, J. (1994). International joint ventures and performance: An interorganizational perspective. International Business Review, 3(3), 201–217. 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. Johansson, J. K., & Yip, G. S. (1994). Exploiting globalization potential: U.S. and Japanese strategies. Strategic Management Journal, 15(8), 579–601. Johnson, M. D., Herrmann, A., & Gustafsson, A. (2002). Comparing customer satisfaction across industries and countries. Journal of Economic Psychology, 23(6), 749–769. Jo¨reskog, K. G. (1982). The LISREL approach to causal model-building in the social sciences. In: K. G. Jo¨reskog & H. O. Wold (Eds), Systems under indirect observation, Part I (pp. 81–100). Amsterdam: North-Holland. Jo¨reskog, K. G., & Goldberger, A. S. (1975). Estimation of a model with multiple indicators and multiple causes of a single latent variable. Journal of the American Statistical Association, 70(351), 631–639. Jo¨reskog, K. G., & So¨rbom, D. (1996). LISREL 8: Structural equation modeling with the SIMPLIS command language. Chicago, IL: Scientiﬁc Software International. Jo¨reskog, K. G., & Wold, H. O. (Eds). (1982). The ML and PLS technique for modelling with latent variables: Historical and comparative aspects. In: Systems under indirect oberservation, Part I (pp. 263–270). Amsterdam: North-Holland. Julien, P. A., & Ramangalahy, C. (2003). Competitive strategy and performance of exporting SMEs: An empirical investigation of the impact of their export information search and competencies. Entrepreneurship Theory and Practice, 27(3), 227–245. Keil, M., Tan, B. C., Wei, K.-K., Saarinen, T., Tuunainen, V., & Wassenaar, A. (2000). A crosscultural study on escalation of commitment behavior in software projects. Management Information Systems Quarterly, 24(2), 299–325. Krijnen, W. P., Dijkstra, T. K., & Gill, R. D. (1998). Conditions for factor (in) determinacy in factor analysis. Psychometrika, 63(4), 359–367. Lee, D. Y. (2000). Retail bargaining behaviour of American and Chinese customers. European Journal of Marketing, 34(1/2), 190–206. Lee, D. Y. (2001). Power, conﬂict and satisfaction in IJV supplier – Chinese distributor channels. Journal of Business Research, 52(2), 149–160. Lee, K.-H., Yang, G., & Graham, J. L. (2006). Tension and trust in international business negotiation: American executives negotiating with Chinese executives. Journal of International Business Studies, 37(5), 623–641. Lohmo¨ller, J.-B. (1989). Latent variable path modeling with partial least squares. Heidelberg: Physica. MacCallum, R. C., & Browne, M. W. (1993). The use of causal indicators in covariance structure models: Some practical issues. Psychological Bulletin, 114(3), 533–541. Mahmood, M. A., Bagchi, K., & Ford, T. C. (2004). On-line shopping behavior: Cross-country empirical research. International Journal of Electronic Commerce, 9(1), 9–30. Malhotra, N. K. (2001). Guest editorial: Cross-cultural marketing research in the twenty-ﬁrst century. International Marketing Review, 18(3), 230–234.

Partial Least Squares Path Modeling in International Marketing

317

Mann, H. B., & Whitney, D. R. (1947). On a test of whether one of two random variables is stochastically larger than the other. Annals of Mathematical Statistics, 18(1), 50–60. Marcoulides, G. A. (2003). PLS model speciﬁcation searches using optimization algorithms. In: M. Vilares, M. Tenenhaus, P. S. Coelho, V. Esposito Vinzi & A. Morineau (Eds), PLS and related methods: Proceedings of the PLS’03 international symposium (pp. 75–86). Paris: Decisia. Marcoulides, G. A., & Saunders, C. (2006). PLS: A silver bullet? Management Information Systems Quarterly, 30(2), III–IX. McDonald, R. P. (1996). Path analysis with composite variables. Multivariate Behavioral Research, 31(2), 239–270. Mintu-Wimsatt, A., & Graham, J. L. (2004). Testing a negotiation model on Canadian anglophone and Mexican exporters. Journal of the Academy of Marketing Science, 32(3), 345–356. Money, R. B. (2004). Word-of-mouth promotion and switching behavior in Japanese and American business-to-business service clients. Journal of Business Research, 57(3), 297–305. Money, R. B., & Graham, J. L. (1999). Salesperson performance, pay, and job satisfaction: Tests of a model using data collected in the United States and Japan. Journal of International Business Studies, 30(1), 149–151. Nakamura, M., Shaver, J. M., & Yeung, B. (1996). An empirical investigation of joint venture dynamics: Evidence from US–Japan joint ventures. International Journal of Industrial Organization, 14(4), 521–541. Nijssen, E. J., & Douglas, S. P. (2008). Consumer world-mindedness, social-mindedness and store image. Journal of International Marketing, 16(3), 84–107. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York, NY: McGraw-Hill. O’Cass, A., & Fenech, T. (2003). Web retailing adoption: Exploring the nature of internet users web retailing behaviour. Journal of Retailing and Consumer Services, 10(2), 81–94. Pavlou, P. A., & Chai, L. (2002). What drives electronic commerce across cultures? A crosscultural empirical investigation of the theory of planned behaviour. Journal of Electronic Commerce Research, 3(4), 240–253. Pinto, J. R., Rodrı´ guez Escudero, A. I., & Gutı´ errez Cilla´n, J. (2008). Order, positioning, scope and outcomes of market entry. Industrial Marketing Management, 37(2), 154–166. Pullman, M. E., Granzin, K. L., & Olsen, J. E. (1997). The efﬁcacy of cognition-and emotionbased ‘‘buy domestic’’ appeals: Conceptualization, empirical test, and managerial implications. International Business Review, 6(3), 209–231. 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. 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). Structural equation models. Sociological Methods and Research, 13(1), 109–119. Ringle, C. M. (2006). Segmentation for path models and unobserved heterogeneity: The ﬁnite mixture partial least squares approach. Research Papers on Marketing and Retailing 35, University of Hamburg, Hamburg.

318

JO¨RG HENSELER ET AL.

Ringle, C. M., Sarstedt, M., & Mooi, E. A. (2009a). Response-based segmentation using ﬁnite mixture partial least squares: Theoretical foundations and an application to American customer satisfaction index data. Annals of Information Systems, in print. Ringle, C. M., & Schlittgen, R. (2007). A genetic algorithm segmentation approach for uncovering and separating groups of data in PLS path modeling. In: H. Martens, T. Næs & M. Martens (Eds), PLS’07 International symposium on PLS and related methods – causalities explored by indirect observation (pp. 75–78). Norway: Matforsk, A˚s. Ringle, C. M., Wende, S., & Will, A. (2005). SmartPLS 2.0 M3. Available at http:// www.smartpls.de Ringle, C. M., Wende, S., & Will, A. (2009b). The ﬁnite mixture partial least squares approach: Methodology and application. In: V. Esposito Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods, and applications. Berlin: Springer (in print). Ringle, C. M., Wilson, B., & Go¨tz, O. (2007). A Monte Carlo robustness study on formative measurement model speciﬁcation in CBSEM and PLS. In: H. Martens, T. Næs & M. Martens (Eds), PLS’07 international symposium on PLS and related methods – Causalities explored by indirect observation (pp. 108–111). Norway: Matforsk, A˚s. Rossiter, J. R. (2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19(4), 305–335. Sarkar, M. B., Echambadi, R., Cavusgil, S. T., & Aulakh, P. S. (2001). The inﬂuence of complementarity, compatibility, and relationship capital on alliance performance. Journal of the Academy of Marketing Science, 29(4), 358–373. SchneeweiX, H. (1991). Models with latent variables: LISREL versus PLS. Statistica Neerlandica, 45(2), 145–157. SchneeweiX, H. (2001). Consistency at large in models with latent variables. In: K. Haagen, D. J. Bartholomew & M. Deistler (Eds), Statistical modelling and latent variables (pp. 299–320). Amsterdam: North-Holland. Shaffer, J. P. (1995). Multiple hypothesis testing. Annual Review of Psychology, 46, 561–584. Singh, N., Fassott, G., Chao, M. C., & Hoffmann, J. A. (2006a). Understanding international web site usage: A cross-national study of German, Brazilian, and Taiwanese online consumers. International Marketing Review, 23(1), 83–97. Singh, N., Fassott, G., Zhao, H., & Boughton, P. (2006b). A cross-cultural analysis of German, Chinese and Indian consumers perception of web site adaptation. Journal of Consumer Behaviour, 5(1), 56–68. Steenkamp, J.-B., & Baumgartner, H. (1998). Assessing measurement invariance in crossnational consumer research. Journal of Consumer Research, 25(1), 78–90. Stone, M. (1974). Cross-validatory choice and assessment of statistical predictions. Journal of the Royal Statistical Society, 36, 111–147. Sto¨ttinger, B., & Holzmu¨ller, H. (2001). Cross national stability of an export performance model: A comparative study of Austria and the US. Management International Review, 41(1), 7–28. Tenenhaus, M., Esposito Vinzi, V., Chatelin, Y.-M., & Lauro, C. (2005). PLS path modeling. Computational Statistics & Data Analysis, 48(1), 159–205. Tsang, E. (2002). Acquiring knowledge by foreign partners from international joint ventures in a transition economy: Learning-by-doing and learning myopia. Strategic Management Journal, 23(9), 835–854.

Partial Least Squares Path Modeling in International Marketing

319

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. Vilares, M. J., Almeida, M. H., & Coelho, P. S. (2009). Comparison of likelihood and PLS estimators for structural equation modeling: A simulation with customer satisfaction data. In: V. Esposito Vinzi, W. W. Chin, J. Henseler & H. Wang (Eds), Handbook of partial least squares: Concepts, methods, and applications. Berlin: Springer (in print). Werts, C. E., Linn, R. L., & Jo¨reskog, K. G. (1974). Intraclass reliability estimates: Testing structural assumptions. Educational and Psychological Measurement, 34(1), 25–33. Wilcoxon, F. (1947). Probability tables for individual comparisons by ranking methods. Biometrics, 3(3), 119–122. Williams, L. J., Edwards, J. R., & Vandenberg, R. J. (2003). Recent advances in causal modeling methods for organizational and management research. Journal of Management, 29(6), 903–936. Wilson, B., Callaghan, W., Ringle, C. M., & Henseler, J. (2007). Exploring causal path directionality for a marketing model using Cohen’s path method. In: H. Martens, T. Næs & M. Martens (Eds), PLS’07 international symposium on PLS and related methods – Causalities explored by indirect observation (pp. 57–61). Norway: Matforsk, A˚s. Wold, H. O. (1974). Causal ﬂows with latent variables: Partings of the ways in the light of NIPALS modeling. European Economic Review, 5(1), 67–86. Wold, H. O. (1982). Soft modeling: The basic design and some extensions. In: K. G. Jo¨reskog & H. O. Wold (Eds), Systems under indirect observations, Part II (pp. 1–54). Amsterdam: North-Holland. Wold, H. O. (1985). Partial least squares. In: S. Kotz & N. L. Johnson (Eds), Encyclopedia of statistical sciences (Vol. 6, pp. 581–591). New York, NY: Wiley. Wold, H. O. (1989). Introduction to the second generation of multivariate analysis. In: H. O. Wold (Ed.), Theoretical empiricism: A general rationale for scientiﬁc model-building (pp. VIII–XL). New York, NY: Paragon House.

ADVANCING THE UNDERSTANDING OF CONSTRUCT VALIDITY AND CROSS-NATIONAL COMPARABILITY: ILLUSTRATED BY A FIVE-COUNTRY STUDY OF CORPORATE EXPORT INFORMATION USAGE Thomas Salzberger, Hartmut H. Holzmu¨ller and Anne Souchon As far as the measurement of latent variables is concerned, research in international business incurs all methodological problems and is susceptible to all pitfalls which generally characterize research in the social sciences. However, international research, by its very nature, attracts further intricacies as a consequence of crossing cultural borders (e.g. Winter & Prohaska, 1983; Craig & Douglas, 2000, 2002; Okazaki & Mueller, 2007; Zhang, Beatty, & Walsh, 2008). The huge body of literature concerned with cross-cultural1 research demonstrates how complicated comparative research actually is. However, even within a culturally homogeneous population, the measurement of latent variables represents a tremendous New Challenges to International Marketing Advances in International Marketing, Volume 20, 321–360 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020015

321

322

THOMAS SALZBERGER ET AL.

challenge, which we often do not awake to because of the apparent ease with which we measure highly abstract concepts in marketing and business research. The measurement of latent variables by means of manifest indicators involves the hypothesis that the latent variable really exists and that it does so in the same way for all respondents in the population (Michell, 1997, 1999). Furthermore, the manifest indicators have to operate in the same way for all individual respondents to justify any sort of aggregate statistics and comparisons across individuals. In other words, the meaning of the latent variable and the metric of the measures we estimate have to be invariant. If the respondents all have the same cultural background, we usually take invariance for granted. However, sociodemographic variables such as age or gender may very well have an impact on the way items in a questionnaire are interpreted. In fact, the very meaning of the latent variable may differ, e.g. between men and women, or young and old. In principle, all personal characteristics we can conceive of may affect the measurement of latent variables. Currently, measurement equivalence is hardly ever addressed in this case. In part, this is due to the predominant measurement theory in marketing, i.e. classical test theory (CTT; Lord & Novick, 1968). CTT is mainly concerned with distributions and aggregate statistics. Key concepts such as reliability (Traub, 1994) are deﬁned in terms of variances. The quality of an item is assessed and judged on the basis of its correlation with the latent variable or the total score. The evaluation of validity also depends heavily on evidence based on correlations. Consequently, we focus on the behavior of items and measures within the given population rather than questioning the homogeneity of the population. Although reliability refers to a population, the estimate from a particular sample does not transcend that sample. This is certainly a major weakness of CTT getting in the way of properly investigating invariance. The focus on the population level also implies a rather weak foundation of measures at the individual level. Although invariance may at least be a plausible assumption in intracultural studies, in cross-cultural research full invariance of manifest items and, consequently, of the measures of the latent variable, is rather unlikely. It is a problem that has to be tackled explicitly (Salzberger, Sinkovics, & Schlegelmilch, 1999). In the following, we will illustrate how the Rasch model (Rasch, 1960) helps investigate measurement equivalence. Since the property of invariance lies at the very heart of the Rasch model, crosscultural invariance does not involve extra complexity in the analysis over and above scale assessment in culturally homogenous studies. Insofar any sort of research involving the measurement of latent variables may beneﬁt

Understanding of Construct Validity and Cross-National Comparability

323

from contributions aimed at solving problems of equivalence testing in international business research. The higher likelihood of invariant item meaning and item functioning in culturally homogenous settings does not discharge us from empirically addressing this issue. Appropriate psychometric data analyses, which assess the invariance of items and establish a common metric across cultures, are important and indispensable. Without doubt equally important is the investigation of sources and factors responsible for a potential lack of equivalence. In this respect, Bauer (1989, 1995) offers a comprehensive framework of factors that need to be considered in order to design a measurement instrument and to devise the process of data collection that ultimately lead to comparable measures. Bauer’s scheme emphasizes the interrelationships between various building blocks of cross-culturally equivalent research. Important concepts were also suggested by Poortinga (1975), Adler (1977, 1983), Berry (1980), Douglas and Craig (1983), Sekaran (1983), Malpass and Poortinga (1986), Toyne and Walters (1989), Usunier (1996), Holzmu¨ller (1995), van de Vijver and Leung (1997), and Cavusgil and Das (1997). Although a discussion of these concepts is beyond the scope of this contribution, we strongly emphasize the importance of addressing these qualitative issues whenever cross-cultural research is undertaken. Psychometric analyses aim at disclosing problems in the data, e.g. a lack of measurement invariance. They do not represent a cure for bad data. This is particularly true for the Rasch model, which requires data to withstand the model’s properties.

COMPARABILITY OF MEASURES ACROSS NATIONAL BORDERS Measures are comparable if and only if measurement equivalence has been demonstrated. Although comparability and equivalence of measures are sometimes used interchangeably, we advocate a subtle but important difference in meaning. Comparability implies that measures from one group can be compared with measures from another group. It is a property of the measures, which is given or not. In particular, comparability presumes valid measures within each group compared. Measurement equivalence, by contrast, refers to the way measures are derived and estimated. It is intrinsically tied to the underlying theory of measurement. Thus, measurement equivalence cannot be dealt with in isolation. Its assessment has to be incorporated into the theoretical framework of measurement. Measurement

324

THOMAS SALZBERGER ET AL.

equivalence is closely connected to construct validity for it refers to the way manifest indicators are related to the latent variable, within a particular culture and across different cultures. From this it follows that equivalence cannot, or should not, be treated as a separate issue but as a constitutive element of validity. A discussion of measurement equivalence without addressing validity would be incomplete. In terms of the measurement theory referred to in marketing and business research, CTT (Lord & Novick, 1968) is still the predominant paradigm, whereas item response theory (IRT; Lord, 1980; Hulin, Drasgow, & Parsons, 1983; Hambleton & Swaminathan, 1985; Hambleton, Swaminathan, & Rogers, 1991; van der Linden & Hambleton, 1997; Embretson & Reise, 2000) has been slowly attracting more interest in the past one or two decades. Obviously, it does not make sense to assess the properties of a measurement instrument based on CTT and then assess measurement equivalence based on IRT, and vice versa. The choice of the measurement model has to be motivated by more fundamental considerations than the assessment of measurement equivalence. When Schaffer and Riordan (2003) identify two best practices for assessing equivalence (one is based on covariance structure analysis, i.e. multigroup CFA, the other is IRT), they actually acknowledge two different paradigms for tackling the problem of measuring latent variables. However, it is true that IRT much more easily accommodates the assessment of measurement equivalence, whereas in the CFA approach an extension has to be made which is (apparently) irrelevant in intra-cultural studies.

Measurement Equivalence and Validity As argued earlier, measurement equivalence and validity are closely related. Since validity is one of the most complicated, controversial, and contentious concepts in research in the social sciences (see Borsboom, Mellenbergh, & Van Heerden, 2004), measurement equivalence is bound to be a disputative issue as well. The commonly accepted state-of-the-art understanding of validity in marketing is the threefold view of validity comprising three different ‘‘types’’ of validity (e.g. Churchill & Iacobucci, 2005). Content validity refers to the question whether the domain of the characteristic is captured by the scale. Construct validity is concerned with whether the manifest indicators work the way the substantive theory of the construct suggests. Finally, we are interested in high criterion-related validity, e.g. how well an export information usage scale explains variations in export

Understanding of Construct Validity and Cross-National Comparability

325

performance. In terms of the actual assessment of validity, these types differ signiﬁcantly. Content validity is typically evaluated qualitatively. It is strongly associated with the substantive theory that has given rise to the postulation of the latent construct. By contrast, construct validity is typically a quantitative issue allowing for a more rigorous investigation. It is closely connected to the measurement theory. In CTT, ﬁt of the data to a factor analytic model would be important evidence of construct validity, whereas in IRT the ﬁt of the data to the particular IRT model is deemed necessary. Finally, criterion-based validity actually requires a sound operationalization. In other words, measurement has to be substantiated before measures can be related to other variables and other measures. For this reason, criterion-related evidence as justiﬁcation of validity has been criticized (Kane, 2001). Borsboom (2005) suggests concentrating on construct validity as the core of the very concept of validity. Sijtsma (2006), however, advocates nomological validity, whereas Rossiter (2002) in a recent, and controversially discussed, contribution emphasized content validity. Measurement equivalence relates to construct validity as well as content validity. Since full invariance of all manifest indicators is the exception rather than the rule, a limited, sometimes very small, set of invariant items is typically considered sufﬁcient to establish a common metric across cultures and thereby justifying measurement equivalence (Steenkamp & Baumgartner, 1998). However, the minimal requirements from a statistical point of view (e.g. in the multigroup CFA, the identiﬁcation of one invariant item in addition to the one deﬁning the metric of the latent variable is technically sufﬁcient, see Steenkamp & Baumgartner, 1998) have to be distinguished from qualitative considerations of what the latent variable actually means in different cultures (Salzberger et al., 1999). In terms of the required number of invariant items, there can be no rule applicable under all circumstances. It strongly depends on theoretical considerations as to why particular items work differently. Consequently, the assessment of measurement equivalence should never be regarded as a purely quantitative exercise no matter what measurement theory is referred to. Measurement equivalence is also a qualitative issue. Finally, there is no point in ‘‘enforcing’’ comparability by discarding the majority of the items during the assessment of measurement equivalence, which work very well within cultures. In this case, very little (i.e. the technical comparability) would actually be won but much (in terms of content validity) would be lost, a situation that is not dissimilar to the attenuation paradox (Loevinger, 1954). A rigorous procedure of testing measurement equivalence may or may not

326

THOMAS SALZBERGER ET AL.

provide evidence of comparability of a latent variable’s quantiﬁcation across cultures but it will always yield valuable qualitative insight.

Methodological Approaches to Assess Cross-National Equivalence As outlined earlier, the choice of the methodological approach toward the problem of measurement equivalence cannot be seen in isolation. Depending on the measurement theory, there is a different best practice approach. In CTT, the multigroup CFA approach is the most powerful way to assess equivalence, whereas in IRT the investigation of differential item functioning (DIF) addresses the issue of equivalence. Classical Approaches Since the vast majority of quantitative studies in international marketing are based on CTT, regarding measurement equivalence, approaches embedded in CTT are most popular. Early attempts to meet the potential threat of comparing measures that are not comparable have concentrated on the concept of reliability (Davis, Douglas, & Silk, 1981). Since reliability depends as much on the variances of the measures (i.e. the true score variances) as it depends on error variances, equal reliability is neither necessary nor sufﬁcient evidence of measurement equivalence. With the rise of factor analysis and, subsequently, the accelerated diffusion of structural equation modeling, tests for equivalence have improved signiﬁcantly. In particular, the multigroup CFA allows for estimating the measurement model simultaneously for multiple groups while imposing equality constraints on selected parameters (Meredith 1964, 1993; Bollen 1989; Horn & McArdle 1992). Mullen (1995) and Singh (1995) suggested equality constraints on the salient factor loadings across cultures (non-salient loadings are ﬁxed to zero anyway), thereby testing for metric equivalence. Steenkamp and Baumgartner (1998) further advanced the approach by pointing out that comparisons of means of latent variables across groups require equal item intercept parameters in addition to equal loadings. The consideration of item intercepts represents an extension of the CFA which is hardly ever considered in intra-cultural studies. However, the multigroup CFA with item intercepts would lend itself perfectly as a way to test measurement equivalence based on, e.g. gender or any other observed person factor. In marketing, Steenkamp and Baumgartner’s (1998) seminal contribution is still an excellent introduction into multigroup CFA.

Understanding of Construct Validity and Cross-National Comparability

327

The multigroup CFA-based approach has some noteworthy limitations. Any shortcoming of CTT/CFA in general (e.g. the sample dependence of factor loadings) evidently affects the multigroup CFA approach. Particularly relevant to measurement equivalence is the effect of the mismatch of continuous and unbounded latent variable measures on the one hand and discrete and bounded manifest indicator scores on the other. Therefore, item intercepts and factor loadings are not independent (see Salzberger et al., 1999). Proﬁle analysis (Morris & Pavett, 1992; Pavett & Morris, 1995) is another approach that compares item scores from different groups graphically. It aims at identifying additive biases by checking whether lines connecting the item means within a group are parallel. Mullen (1995, p. 579) notes that ‘‘even if the lines in the graph are parallel, the differences in means may not all be due to response bias.’’ Mullen further concludes that ‘‘there has been no diagnostic method available to assess cross-national differences in scalar equivalence with conﬁdence’’ before the introduction to multigroup CFA. From a more fundamental point of view, the comparison of item raw scores, which are not interval scaled and certainly not linear, is problematic anyway. Item Response Theory IRT is a collective term for many different models designed to explain measurement. All IRT models primarily refer to the individual response rather than aggregate statistics. In the simplest case, the Rasch model (Rasch, 1960), the dichotomous item, the probability of agreement is modeled as a logistic function2 of a person parameter (the measure we are ultimately interested in) and an item parameter. The latter is the counterpart to the person measure. In achievement testing, this parameter is easily interpreted as the item difﬁculty, whereas in attitude measurement its meaning is less straightforward. On an abstract level, the item parameter captures the amount of the property the item represents and which is to be measured. Consequently, with attitudinal constructs, a larger item parameter implies a stronger attitude, i.e. its endorsability is smaller compared to an item which is ‘‘easier to agree with.’’ Hence Ewing, Salzberger, and Sinkovics (2005) suggest using the term endorsability. Alternatively the item parameter could be referred to as item location. Thus, there is nothing that rules out the application of IRT models in attitude measurement. It should be noted that the item location parameter, or endorsability, has a counterpart in CTT/CFA, namely the item intercept parameter (Wright, 1996), which is modeled in multigroup CFA. In contrast

328

THOMAS SALZBERGER ET AL.

to CFA, the item location parameter in IRT does not depend on the sample. Furthermore, in IRT, item and person parameters are placed onto the same latent dimensions, i.e. they share a common metric and are directly comparable. ebv di 1 þ ebv di

(1)

eai ðbv di Þ 1 þ eai ðbv di Þ

(2)

Pðavi ¼ 1Þ ¼

Pðavi ¼ 1Þ ¼

There is, however, a qualiﬁcation on the claim of sample independence of the item parameter, which is related to the type of IRT model. A fundamental distinction exists between models providing for equal item discrimination [e.g. the dichotomous Rasch model by Rasch (1960), see Eq. (1) with bv being the location parameter of respondent v, di being the location of item i and avi being the response by person v to item i] and those allowing for variation in item discrimination [e.g. the Birnbaum model (Birnbaum, 1968) for dichotomous responses, see Eq. (2) with an additional parameter ai modeling the item’s discrimination]. Therefore, the Birnbaum model, or the two-parameter logistic model, features an item discrimination parameter, whereas in the Rasch model the item discrimination is ﬁxed at 1 for all items. This distinction generalizes to polytomous models, which are much more suitable in marketing research for most of our measurement instruments feature multicategorical response scales. All models with equal item discrimination are summarized as Rasch models (Fischer & Molenaar, 1995). Models with equal item discrimination have several elegant and advantageous properties. Since person and item parameters are linked additively, the model complies with additive conjoint measurement (Luce & Tukey, 1964) and satisﬁes cancellation conditions as set out by axiomatic measurement theory (see Karabatsos, 2001, pp. 398ff. for details). Item and person raw scores are sufﬁcient statistics allowing for separability of parameter estimates by conditional maximum likelihood estimation (Molenaar, 1995) or pairwise parameter estimation (Andrich & Luo, 2003). A more tangible consequence of these technical properties is the mutual invariance of item and person parameters. Rasch (1961) referred to this characteristic of the model as speciﬁc objectivity. Objectivity implies that item locations do not depend on the respondents used to estimate them and, vice versa, person location parameters are independent of the speciﬁc

Understanding of Construct Validity and Cross-National Comparability

329

items which are instrumental in their estimation. Thus, there are no distributional assumptions involved in parameter estimation in contrast to the strong, and in many cases unrealistic, assumption of normally distributed respondents in CFA and more general IRT models. There are two important restrictions of objectivity. First, objectivity is conﬁned to a frame of reference; hence, it is called speciﬁc objectivity rather than general objectivity. The frame of reference comprises all factors that are relevant in the measurement process, most noteworthy the population of respondents and the conditions of data collection. Second, speciﬁc objectivity, or invariance, is a requirement. It is neither an untested assumption nor a property that would be independent of data. From this it follows that invariance has to be tested empirically. Item location parameter estimates from different samples of respondents have to be equal up to random variation. In fact, it is a very powerful way to test the ﬁt of the data to the model. All factors, internal ones such as the grouping of low versus high raw scores, and external characteristics such as sociodemographic or psychographic variables qualify as criteria to test for invariance. The need to test for cross-cultural or cross-national equivalence ﬁts nicely in this framework of assessing invariance and represents no extra burden in terms of testing the measurement model. Despite the favorable properties following from equal item discrimination, the latter might be seen as overly restrictive. More general IRT models [see Singh (2004) for an introduction in a marketing context and De Jong, Steenkamp, and Fox (2007) for a recent application in cross-cultural measurement] are certainly more ﬂexible in this respect. However, although variation in item discrimination allows for better describing data, we no longer test the structural requirements of quantity in the data. The Rasch model does not aim at best accounting for given data. It rather sets out the structure in the data which indicates quantitative person and item location parameters. Thus, the Rasch model addresses what Michell (1999) calls the ‘‘scientiﬁc task of measurement,’’ i.e. the provision of evidence of the existence of a quantitative latent variable. Neglecting the scientiﬁc task ‘‘sounds a warning to us about the nature of measurement in the social sciences’’ (Balnaves & Caputi, 2001, p. 51). Finally, it should be noted that Churchill (1979) originally asked for items representing an ‘‘equal amount of the core of the construct.’’ The parallel model in CTT/CFA comes much closer to the Rasch model compared to the congeneric model which prevails in contemporary quantitative research in marketing. Thus, in the following we concentrate on the Rasch model as a suitable measurement model and its strong potential to test for measurement equivalence.

330

THOMAS SALZBERGER ET AL.

RASCH MODEL FOR MEASUREMENT Two factors have made the Rasch model more amenable to applications in marketing research. First, the dichotomous model (Rasch, 1960) has been generalized to the polytomous response format (Andrich, 1978, 1988; Masters, 1982). Second, accessible, user-friendly software has become available (e.g. RUMM2020 by Andrich, Sheridan, & Luo, 2003a; WINSTEPS by Linacre, 2007; or ConQuest by Wu, Adams, & Wilson, 1997). For users of the open-source statistical software R, a special package for the estimation of Rasch models has been provided by Mair and Hatzinger (2007). The importance of convenient software for the acceptance and dissemination of statistical and psychometric models cannot be underestimated. Without LISREL (Jo¨reskog & So¨rbom, 2003), EQS (Bentler, 2003), AMOS (Arbuckle, 2005), or MPLUS (Muthen & Muthen, 2007), the widespread application of structural equation models (SEM) and models beyond SEM would be inconceivable. In marketing applications, we typically use multicategorical response scales. Dichotomization of originally polytomous data is disadvantageous because of the loss of information. Even more problematic are theoretical concerns as ﬁt to a polytomous Rasch model mathematically rules out ﬁt of dichotomized data to a dichotomous Rasch model (see Andrich, 1995a, 1995b, on this issue). As outlined earlier, in the dichotomous Rasch model, there is only one item location parameter.3 If a particular person matches the item location exactly, then the probability of agreement amounts to 0.5 and equals the probability of disagreement. Respondents located further down the scale have a smaller chance to respond positively, whereas those located further up the scale have a higher probability of agreement than disagreement. Consequently, the item location parameter can be thought of as a threshold between disagreement and agreement as the most likely response. This principle generalizes to polytomous items. With k categories, we need to specify k – 1 threshold parameters. Consequently, the ﬁrst threshold marks the point on the latent continuum where the ﬁrst and the second category are equally likely. The mean of all threshold parameters stands for the overall location of the item. In the Partial Credit Model (Masters, 1982; Andrich, 1988), the thresholds are modeled for each item individually, i.e. each item has its own threshold structure. In this case, items with different numbers of categories can be mixed. Alternatively, the Rating Scale Model (Andrich, 1978) speciﬁes a common threshold structure across all items. Then the

Understanding of Construct Validity and Cross-National Comparability

331

relative distances between thresholds are the same across items while the overall item locations vary, of course. Since the special properties of the Rasch model apply to both polytomous models, there is no need to constrain the threshold structure across items. Eq. (3) (Andrich, 1988, p. 366) states the probabilities for all response categories as a function of the model parameters bv (person location parameter), di (item overall location, mean of thresholds), and tij (threshold j for item i). The denominator g is the sum of all numerators. The numerator for the ﬁrst category (scored 0) is 1. Fig. 1 displays the category characteristic curves of a ﬁve-category item. Each curve describes the probability of a particular response as a function of the person location. m P Pðavi ¼ xjbv ; tij ; j ¼ 1 . . . m; 0oxomÞ ¼

e

tij

þ xðbv di Þ

j¼1

g

(3)

It is important to notice that in the Rasch model, the thresholds only model the transition point (equal probability) of two adjacent categories (e.g. between categories 2 and 3) but not between all categories below or above that point (e.g. probability to choose category 0, 1, or 2 versus category 3 or 4). The latter are referred to as Thurstone thresholds (Linacre, 1998), whereas in the polytomous Rasch model, Rasch–Andrich thresholds

Fig. 1.

Category Characteristic Curves for a Polytomous Item Under the Rasch Model.

332

THOMAS SALZBERGER ET AL.

(Linacre, 2001) are modeled. This implies that empirical estimates of category thresholds can take any order. If their order does not correspond to the order of categories, the thresholds are said to be disordered. Although the threshold order is not an explicit requirement in the Rasch model (the data may still ﬁt the model), disordered thresholds are often seen as problematic as they provide evidence that the presumed order of the response scale might not hold (see Andrich, 1995a, 1995b). Disordered thresholds imply a lack of discrimination between two categories. There are some typical situations leading to disordered thresholds. First, too many categories may ask too much of the respondents, who might not be able or willing to distinguish between all categories offered. Second, oddly verbalized categories may cause problems. The widely used option of neither agree nor disagree is particularly prone to malfunctioning. It attracts not only respondents in the middle of the item but also the undecided or respondents to whom the item is not applicable. Third, negatively worded items often result into threshold disorder as it is even more difﬁcult to discriminate between categories in this case. Positively worded items usually sustain more response categories than negatively worded indicators. Regardless of the actual reason of disordered thresholds, in the end, the researcher has to decide how to react. Ideally, a change of the item wording or the number of categories should be considered. With existing data, a post-hoc collapsing of adjacent categories is a manageable remedy. It should be noted that in this case no loss of information occurs since the threshold reversal indicates that the distinction between the categories that are collapsed is not informative at all.

Applying the Rasch Model A Rasch analysis of a measurement instrument requires several steps. Each set of items indicating a unidimensional latent variable has to be treated separately. Although there is a meaningful order in which these steps should be taken, problems in the data may require an iterative approach and returning to an earlier step might be necessary at any time. Insofar, the ﬂow chart in Fig. 2 is an idealized scheme. The analysis should start with an investigation of the threshold order since these not only contradict the presumed order of the response options but also often cause item misﬁt. Thus, in an iterative process, categories with reversed threshold parameters should be collapsed until all thresholds are properly ordered.

333

Understanding of Construct Validity and Cross-National Comparability

Within-culture data sets

Pooled data

Cross-cultural data set

#1: Examination of thresholds

Reversed

Collapse categories/ split item * A - Rating scale functioning analysis

#1: Examination of thresholds

Reversed

Collapse categories *

Properly ordered remaining items

#2: Examination of item fit

Misfit

Number of cultures?

=1

Discard item *

remaining items

Properly ordered

#2: Examination of item fit

Misfit

Discard item *

>1

Fit

Fit

#2a: Test of DIF

DIF

Split item *

Input for cross-cultural analysis

No DIF Split item *

D - Supplementary analyses B – Item compatibility analysis

Number of cultures?

>1

#3: Test of DIF

DIF

Split item *

No DIF

=1

C – DIF analysis

#4: Estimation of final indicator and respondent locations

#4a: Investigation of person fit

#5: Mean comparison across cultures

* Whenever item thresholds are reversed, items show misfit or display DIF, at any one time only one action (collapsing categories, splitting or discarding item) should be undertaken due to the potential interaction of threshold reversals, misfit and/or DIF. to : see annotations to this chart

Cross-culturally equivalent scale

Fig. 2.

Rasch Analysis Flow Chart for Cross-Cultural Research.

334

THOMAS SALZBERGER ET AL.

ANNOTATIONS TO THE RASCH ANALYSIS FLOW CHART FOR CROSS-CULTURAL RESEARCH In the analysis of the pooled data set, culture has to be included as a person factor. Separate analyses of the scale for each culture (hereafter referred to as supplementary analyses) should be carried out simultaneously as they can provide input for subsequent decisions on category collapsing and/or item splitting. See 2 . 2 The supplementary analyses are carried out separately for each culture. The sequence of steps undertaken mirrors the course of the main analysis of the pooled data except for being one-group analyses. After examining the thresholds (see also 3 ), item ﬁt is assessed (compare 5 ). The results (which items ﬁt, which don’t; which items display disordered thresholds, which don’t) provide a guideline for actions during the main analysis of the pooled data. 3 Step 1 of the main analysis (rating scale functioning analysis). The analysis starts with the examination of the threshold ordering. Neither can item ﬁt be assessed meaningfully nor can the results be interpreted appropriately unless all thresholds are in proper order. 4 If reversed thresholds occur, a combination of category collapsing and item splitting should be considered as a post hoc remedy. Which speciﬁc action is suitable can be derived from insight gained by the supplementary analyses. If, e.g. the thresholds of one item are disordered in the same way across all cultures, a common rescoring of the item will very likely be sufﬁcient. However, a different functioning of the item response scale in the cultures involved will require item splitting in order to accommodate cultural idiosyncrasies. After collapsing categories, tantamount to a rescoring of the item responses, the thresholds have to be re-examined. Since malfunctioning of items compromises the estimation of model parameters, at any one time, only one action should be undertaken. Items with most severely disordered thresholds, e.g. those for which only the extreme categories seem to work, should be targeted ﬁrst. Any post-hoc provisions should be considered preliminary. Only new data, gathered by applying altered response scale formats, can conﬁrm proper response scales. However, since a 1

Understanding of Construct Validity and Cross-National Comparability

5

6

7

8

9

10

335

non-uniform format across all items is prone to confuse respondents, a common response scale is advisable. Step 2 of the analysis (item compatibility analysis). The second important element of the analysis is the assessment of the compatibility of the items to the Rasch model. This is done by an item-ﬁt analysis. Again, only one item, starting with the one displaying the worst ﬁt, should be considered at one time. In a one-group analysis, a misﬁtting item has to be discarded. In a multicultural setting, however, item misﬁt may be the result of misﬁt in some but not all cultures, or it may be due to DIF. That’s why, an item should only be discarded if no further splitting of the item is possible (see 7 ). In case of misﬁt, a test of DIF is advisable. If no DIF prevails, the item should be split up based on input from the supplementary analyses (see 2 ). In this case, misﬁt in some of the cultures is expected. If the item displays DIF, item splitting has to be considered, as well. However, DIF may be the sole reason of misﬁt in this case. Thus, the DIF analysis should be carried out, even though the consequences, i.e. item splitting, are the same. But the rationale might be different. Step 3 of the analysis (DIF analysis). The investigation of DIF is a central element in any cross-cultural study. Severe DIF usually becomes manifest in item misﬁt. Therefore, DIF is also a matter of investigation at the second step of the analysis (compare 7 ). Since items can be affected by minor, though signiﬁcant, DIF even when ﬁtting the model over-all, tests of DIF should be carried out at this stage. Item splitting is necessary whenever DIF occurs. Items displaying no sign of DIF and items conﬁned to one culture (as a result of item splitting) directly enter the ﬁnal phase during which the ﬁnal indicator and respondent locations are estimated. Step 4 (estimation of model parameters) and step 5 of the analysis (mean comparison). After scale puriﬁcation, the cross-culturally valid and comparable indicator locations as well as the respondent locations can be estimated. Based on these interval scaled parameters, mean comparisons across cultures can validly be carried out at step 5 (see also 10 ). Step 4a of the analysis (person ﬁt investigation).

336

THOMAS SALZBERGER ET AL.

The second step of the scale analysis pertains to item ﬁt (see Andrich, Sheridan, & Luo, 2003b for the statistics described). One way to look at item ﬁt is the comparison of expected responses (the expected item score depending on the person location) and actual responses. The differences of what is the observed minus what is expected are assessed in score groups of respondents and eventually lead to an approximately w2 distributed ﬁt statistic. This test of ﬁt can be carried out for every individual item but can also be summed over all items to get a summary test of ﬁt for the whole instrument. Another item ﬁt statistic is based on each respondent’s residual resulting in an approximately z-distributed statistic (ﬁt residual) indicating whether the indicator is discriminating as expected (ﬁt residual between 2.5 and þ2.5), whether it is overdiscriminating (ﬁt residual less than 2.5) or underdiscriminating (ﬁt residual greater than þ2.5). The third step should investigate invariance empirically. If an item lacks invariance, it shows DIF. DIF may be the sole reason for item misﬁt as assessed on the previous step but it should also be tested when items ﬁt properly. As this test is of utmost importance in checking for measurement equivalence, it will be discussed in more detail in the subsequent chapter. Although the distribution of respondents has no impact on the item estimates as such due to speciﬁc objectivity, precision and ﬁt assessment do depend on a sufﬁcient matching between respondents and items. This issue is referred to as targeting. The person separation index (PSI; Andrich, 1982), which is deﬁned analogously to classical reliability, is helpful in this respect. Unlike reliability in CTT, the PSI in Rasch modeling is not as crucial. However, a low PSI is of concern for various reasons. Insufﬁcient targeting may result in a low PSI, compromising the power of the test of ﬁt as standard errors for item as well as person estimates increase. Poor targeting means that a higher number of items is required to achieve the same precision that is provided by a properly targeted instrument. So, from an economical point of view, a high PSI is desirable. If the targeting is satisfactory and item ﬁt statistics are inconspicuous (i.e. non-signiﬁcant), a low PSI may indicate multidimensionality (values less than about 0.6 almost certainly imply severe problems in the data).

Using the Rasch Model to Assess Measurement Equivalence The application of the Rasch-model to cross-cultural or cross-national data is straightforward in theory but might get complicated in practice, particularly if many cultures are involved and thresholds are partly

Understanding of Construct Validity and Cross-National Comparability

337

disordered. As a general recommendation, the number of response categories should be kept rather small (no speciﬁc number can be given, though, as it depends on the particular countries as well as on the construct and respondent characteristics such as their involvement) and the verbalization of the categories should be conﬁned to the extreme categories as translations of response category qualiﬁers rarely correspond perfectly resulting in different threshold structures and possibly threshold disorder. The analysis should start with an investigation of item ﬁt within each country, in particular if there are more than two groups involved (see supplementary analyses in Fig. 2). These analyses point out which items do not work properly in which countries and help identify items that function differently. The main analysis refers to the pooled data set, and all items are tested for their model compatibility. At step 1, as in any analysis of polytomous data, the threshold order should be scrutinized. Disordered thresholds can be resolved by scoring adjacent categories identically. As soon as all thresholds are ordered, the ﬁt of the items is assessed (step 2). Within one culture, misﬁtting items are discarded. In cross-cultural studies, such a conclusion might be premature. The item might not work in one culture (causing overall misﬁt) but function perfectly in another. Alternatively, the item may work differently in the groups analyzed. In other words, the indicator has a culture-speciﬁc location. Then, the item is said to display DIF. In any case, a DIF-analysis has to be carried out (step 2a). Strictly speaking, speciﬁc objectivity calls for excluding items which do not have the same location in all countries. The reason is that the interpretation of the scale depends on the cultural afﬁliation of the respondent. Therefore, the measures for respondents with equal raw scores but from different cultures are not the same. Desirable as complete speciﬁc objectivity may be, in cross-cultural research, it is highly unlikely that all items are functioning the same way across cultural borders. Excluding deviating items would result in to a small number of items representing a serious threat to criterion-based validity. Fortunately, the trade-off between criterion-based validity and speciﬁc objectivity can be accommodated for by relaxing the latter. If an item is affected by DIF, it can be split into two or more ‘‘versions’’ for analytical purposes. Each version is associated with one or more cultures. This allows for culture-speciﬁc estimation of the indicator location. This procedure does not require the item to ﬁt in each culture under scrutiny. If an item fails to work properly in one culture, it can easily be discarded for that group but retained for all other cultures.

338

THOMAS SALZBERGER ET AL.

Accounting for cultural idiosyncrasies in the way described parallels the concept of partial invariance outlined by Steenkamp and Baumgartner (1998) in the context of CTT-based assessment of equivalence. The crucial point is that common items, which function equivalently across countries, i.e. identical location and identical threshold distances, deﬁne the scale. These items can be referred to as (at least derived) etics compared to emic items that are linked to the common scale. Invariant items act as a bridge between countries. It is clear intuitively that the more items are invariant the more stable the bridge becomes and the easier it is to interpret the construct cross-nationally. There are different approaches to test for DIF. In this study, we make use of an ANOVA-based method implemented in RUMM 2020 (a windowsbased Rasch program by Andrich et al., 2003a). To this end, the respondents are ﬁrst grouped into classes of similar location along the continuum like in the general test of ﬁt. The factor possibly leading to DIF, i.e. country in our case, is another criterion to deﬁne the groups in the ANOVA. The class interval is one main effect which is of no particular interest in this case as it is just another general test of ﬁt. So, misﬁtting items typically exhibit a signiﬁcant class interval main effect. More interesting is the DIF factor representing the other main effect. If there is no DIF, the mean residuals between countries should be the same (i.e. zero) within each class interval. If, however, the item is consistently easier to endorse in one country compared to another, then the mean residual in the former group is positive, whereas it is negative for the latter resulting in a signiﬁcant main effect for the DIF factor. This sort of DIF is referred to as uniform DIF. The empirical item response functions (IRF; the mean item score for respondents from a particular country) are still parallel. A different type of DIF is given, if an item is easier (or harder) to agree with for some country compared to another at the bottom end of the scale, whereas the reverse is true for the top end of the scale. In this case, the empirical IRFs intersect. In the ANOVA, this sort of DIF is captured by the interaction term of class interval and DIF factor. It goes without saying that items displaying DIF need special attention. Deleting DIF items in cross-cultural studies would be the most rigorous procedure, but it may result in a very small number of items questioning the whole measurement process. In fact, with many countries, not a single item may be equivalent across all countries. That’s why items affected by DIF should better be retained provided they ﬁt within each culture for which they are used. This can be done by splitting the item into virtual versions of the same items for each country or a group of countries. For the other countries,

Understanding of Construct Validity and Cross-National Comparability

339

the item is scored missing. Input from the supplementary analyses carried out before should help identify countries for which common item estimates are reasonable. This procedure is most likely to succeed in case of uniform DIF. Non-uniform DIF, by contrast, implies more severe problems as the item misﬁts in at least one culture but may generally be a problematic item. If all items ﬁt and no indication of DIF prevails, the ﬁnal estimation of item and respondent locations can be carried out. Finally, mean comparisons can be conducted by referring to the interval-scaled measures which are comparable across all cultures that have been investigated. The question how many items should be invariant in order to establish a common metric across countries cannot be answered absolutely. Technically two invariant items provide sufﬁcient empirical evidence to justify a common metric across two countries. However, from a qualitative point of view, a larger number is certainly desirable. When accounting for DIF by splitting the item, we always incur the risk of capitalizing on chance. This is particularly true if there is no plausible explanation why DIF occurs. Ideally, a scale development process should include expectations in terms of DIF from the outset. From a puristic point of view, accounting for DIF violates speciﬁc objectivity (see Molenaar, 1997). It is no longer sufﬁcient to know the raw score to estimate a person measure but we need to take the country into account. Once again, a plausible theoretical explanation of DIF seems to be a reasonable compromise in this respect. Kreiner and Christensen (2007) speak of essential objectivity in this context. In case of more than two countries, no item has to be completely invariant. If a set of items is invariant for countries A and B, and another set of items is invariant for countries B and C, then the links between A and B, and B and C also established a link between A and C. Hence, keeping item splitting to an absolute minimum strengthens overall measurement equivalence. An additional examination of person ﬁt is advisable in conjunction with stage 4. Although respondent characteristics theoretically do not bear on indicator characteristics, this principle does not apply in case of misﬁtting persons. However, compared to the number of items, the number of persons is usually large, and consequently, person misﬁt has a smaller impact on the item estimates than item misﬁt has on the respondents estimates. There is no straightforward remedy for person misﬁt. Excluding deviant respondents can change aggregate statistics, like the group mean, and should therefore be considered with caution. The quantitative analysis alone cannot provide a deﬁnite answer but only points out where problems lie. Qualitative considerations should be made to trace the reasons for misﬁtting persons.

340

THOMAS SALZBERGER ET AL.

If systematic person misﬁt occurs, it might be caused by DIF due to variables other than culture. Alternatively, the frame of reference of the instrument, i.e. the set of admissible respondents allowing for speciﬁcally objective measurement, requires re-deﬁnition.

EMPIRICAL ILLUSTRATION: EXPORT INFORMATION USAGE The main purpose of the empirical example is the demonstration how the analysis based on the Rasch model helps derive measures of export information usage across cultures. As more and more industries are interrelated internationally, for many companies export performance is a crucial factor. Companies typically aim at compensating uncertainty and a lack of information that is almost inevitably connected to exporting by collecting information (Katsikeas, 1994; Leonidou, 1995). Less attention has been paid to the problem of how information is processed within the company. Companies gather information – and never use it in decision making (Knott & Wildavsky, 1981). Companies make decisions – and collect information afterward in order to justify the decision (March & Shapira, 1982). Thus, the key to success is not mere information collection but the way companies use information effectively (Hart & Diamantopoulos, 1993). Diamantopoulos and Souchon (1998, 1999) suggested a measurement instrument (comprising 28 items in total) designed to assess export information utilization as a multidimensional construct to distinguish between three distinct sorts of information usage (Souchon & Diamantopoulos, 1996). First, instrumental use of information (covered by nine items) refers to situations in which information is collected for a speciﬁc purpose and then actually used to make a decision (Deshpande´ & Zaltman, 1982). Second, conceptual use of information (measured by eight indicators) covers indirect application of information. In this case, information broadens the managerial knowledge base in general. There is no clear relationship between a piece of information and a particular project or decision, however. Both, instrumental and conceptual use of information are ‘‘reasonable’’ ways of dealing with information positively related with corporate success. By contrast, the third type of information use, symbolic use (Knorr, 1977), refers to fundamentally different patterns of information usage. Information is said to be used symbolically, if information is distorted to support the decision maker’s opinion or selectively used to justify previously made decisions.

Understanding of Construct Validity and Cross-National Comparability

341

Finally, symbolic use occurs when information is collected merely to maintain good relationships with information providers. Symbolic use is operationalized by 11 items. This study aims at assessing the validity of the multidimensional construct of export information usage across country and language borders. In particular, the cross-national equivalence of the items in the UK, the USA, New Zealand, Austria, and Germany is examined by the Rasch model. The scale is applied to three different modes of information acquisition (for details, refer to Souchon & Diamantopoulos, 1999). Export marketing research (EMR) is a formalized, planned, and more or less objective process of collecting information. Export market intelligence (EMI), by contrast, covers the constant ﬂow of incidental information. It is less systematic and more subjective. Finally, export assistance (EA) refers to institutional information provided mostly by ofﬁcial bodies. It is assumed that the company’s propensity to use information in a certain way may depend on the source. In other words, a company may use information from formal EMR more strongly instrumentally than information gathered from EMI. However, for a different company, the opposite could be true. The data underlying the empirical example has been provided by Salzberger et al. (2002). The measures were developed by Souchon & Diamantopoulos (1996, 1999). The cross-cultural study was conducted in ﬁve countries, in the UK (n ¼ 198), the US (n ¼ 161), New Zealand (n ¼ 239), Austria (n ¼ 220), and Germany (n ¼ 171). The countries were selected on comprehensive theoretical considerations (see Diamantopoulos, Souchon, Durden, Axinn, & Holzmu¨ller, 2003). In our context, it is interesting to investigate to what extent culturally relatively close countries such as the UK and New Zealand safeguard measurement equivalence and how different languages bear on item invariance. Undoubtedly, the range of countries does not come anywhere near a comprehensive coverage of world cultures. However, in this contribution, we focus on methodological issues. In fact, vastly different cultures would not issue a real challenge to a method to reveal measurement invariance. By contrast, comparatively subtle differences between closely related cultures, like the UK and New Zealand or Austria and Germany, are more suitable as they put the sensitivity of the method to the test. The questionnaire was initially developed in the UK and has been translated into German deploying appropriate forward–backward procedures. A verbally comparatively labeled, ﬁve-point ordered response scale was provided for each item. The total population was restricted to manufacturing companies engaged in exporting. As the sampling frame data from Dun and Bradstreet has been used.

342

THOMAS SALZBERGER ET AL.

The data set has been scrutinized repeatedly in terms of cross-cultural invariance (Diamantopoulos et al., 2003; Souchon et al., 2003). However, measurement equivalence was based exclusively on traditional approaches. Souchon et al. (2003) investigated the scale reliability coefﬁcients as suggested by Davis et al. (1981). A multigroup CFA was not carried out because of the large number of missing values as many companies did not respond to all items applied to all three information sources (Souchon et al., 2003). Diamantopoulos et al. (2003) also referred to reliability coefﬁcients but mainly relied on proﬁle analysis (see Mullen, 1995) as a means of testing metric equivalence. No lack of invariance was identiﬁed. This is to some extent a surprising ﬁnding as one would certainly expect some degree of non-invariance with ﬁve countries involved. From a Rasch point of view, these procedures are unsatisfactory as no rigorous testing for item ﬁt and item invariance has been carried out. Moreover, mean comparisons across countries require scalar invariance, i.e. absence of additive biases or, at least, biases have to be taken into account. In the following, we will investigate measurement equivalence based on the polytomous Rasch model. All analyses have been carried out using RUMM 2020 (Andrich et al., 2003a). As there are three different styles of informational usage, namely instrumental, conceptual, and symbolic use, and three sources of information, namely EMI, EMR, and EA, nine applications would have to be tested. Since we focus on the demonstration of the analysis, we decided to concentrate on the instrumental use of information collected through EMI. This source is widely used and, consequently, the sample size is the largest. Subsequently, we will brieﬂy deal with the stability of the scale across sources of information.

RESULTS Instrumental Use of Export Marketing Intelligence At ﬁrst, the scales have been tested within each country sample in order to identify problematic items and threshold disordering. The most striking result at this stage was that, in general, German companies tended to use the answer scale most often in an unanticipated way leading to reversed thresholds requiring rescoring in the sense of collapsing adjacent categories (Andrich, 1995a, 1995b). Subsequently, the data sets have been pooled in order to assess DIF. Items displaying DIF were split up to allow for country-speciﬁc estimates. For the sake of parsimony, all efforts have been

Understanding of Construct Validity and Cross-National Comparability

343

undertaken to specify as few as possible location parameters for one and the same item. This means that, e.g. one item parameter was estimated for three countries and one parameter for the remaining two countries. For instrumental use of information, only one of nine items (item /uioutdaS: information loses its value over time) showed misﬁt in every culture and has therefore been discarded entirely. It is quite plausible that the mere fact that the temporal value of information is limited does not necessarily mean that information is used for a speciﬁc problem. Five of the remaining eight items are fully invariant across countries except Germany (invariant items are indicated by shaded cells in Table 1 covering more than one culture). Interestingly, these items explicitly pertain to information collection behavior and usage (/uispeciS: information is actively sought . . . in response to a speciﬁc decision; /uipartiS: information . . . is used speciﬁcally; /uindeciS: no export decision would be made without information . . . ; /uinouseS: information . . . not used (reverse coded); /uitransS: information . . . is translated into . . . practical action). These items represent a strong link between countries and constitute a solid basis of cross-national measurement equivalence. In contrast, the items showing DIF are related rather indirectly to information usage in that they address perceived consequences such as /uiaccurS: decisions based on information are more accurate than intuitive ones or /uiconﬁS: our conﬁdence in making export decisions is increased. For both items the location parameters for UK companies are signiﬁcantly smaller. This means that, conditioned on the same level of instrumental usage, UK companies claim to have more conﬁdence in their decision-making behavior, and they consider their decisions more accurate. Even when there is no invariant item across all ﬁve countries, Germany can be linked to the common scale because of sharing two items that are invariant across all countries except the UK and one item that has the same location in Germany and Austria (/uidiffrS: decision would be different without information of this type). For two other items, the overall item location of Germany is similar to the other countries, but the scoring format differs due to initially reversed thresholds in Germany. However, the link between Germany and the other countries is not as strong as it is between the remaining countries. Table 1 contains the results of the Rasch analysis of the items indicating instrumental usage of information across all ﬁve countries. The ﬁrst value at the top in each cell represents the indicator location estimate, e.g. the location of the item coded /uispeciS is 0.072. In terms of all items in the scale, this item has an about average endorsability. The locations stated in Table 1 are means of all threshold

Information collected by this method is often used speciﬁcally to make a particular export decision

Decisions based on information collected in this way are more accurate than wholly intuitive ones

Our conﬁdence in making 1.181 export decisions is increased (SE ¼ 0.122; as a result of information w2 ¼ 18.95, collected in this way df ¼ 9, p ¼ .03)

uiparti

uiaccur

uiconﬁ

–1.107 (SE ¼ 0.113; w2 ¼ 11.74, df ¼ 9, p ¼ .23)

Information is actively sought out from this method in response to a speciﬁc decision at hand

uispeci

misﬁt

UK (n ¼ 187)

Information gathered in this way for a speciﬁc problem loses its value over time

Item Content (Abbreviated)

2

misﬁt

USA (n ¼ 141)

misﬁt

NZL (n ¼ 224)

GER (n ¼ 166)

0.405 (SE ¼ 0.13; w2 ¼ 2.16, df ¼ 9, p ¼ .99)

0.251 (SE ¼ 0.13; w2 ¼ 5.58, df ¼ 9, p ¼ .78)

misﬁt

0.445 (SE ¼ 0.063; w2 ¼ 32.13, df ¼ 9, p ¼ .0001)

0.005 (SE ¼ 0.055; w2 ¼ 8.86, df ¼ 9, p ¼ .45)

0.135 (SE ¼ 0.049; w2 ¼ 8.67, df ¼ 9, p ¼ .47)

0.072a (SE ¼ 0.046; w ¼ 11.11, df ¼ 9, p ¼ .27)

misﬁt

AT (n ¼ 218)

Item Location Estimates for Instrumental Information Usage (Export Marketing Intelligence).

uioutda

Item Code

Table 1.

344 THOMAS SALZBERGER ET AL.

The majority of information gathered by this method is not used. (reverse coded)

Information collected in this way is translated into signiﬁcant practical action

uinouse

uitrans

1.29

0.785

w2 ¼ 194.81 (df ¼ 141), p ¼ .002

1.27

0.04 (SE ¼ 0.057; w2 ¼ 17.29, df ¼ 9, p ¼ .04)

0.221 (SE ¼ 0.05; w2 ¼ 17.11, df ¼ 9, p ¼ .05)

1.29

0.074 (SE ¼ 0.059; w2 ¼ 8.89, df ¼ 9, p ¼ .45)

1.018 (SE ¼ 0.056; w2 ¼ 10.54, df ¼ 9, p ¼ .31)

0.421 (SE ¼ 0.073; w2 ¼ 6.10, df ¼ 9, p ¼ .73)

1.48

0.809 (SE ¼ 0.11; w2 ¼ 12.7, df ¼ 8, p ¼ .12)

0.386 (SE ¼ 0.13; w2 ¼ 11.8, df ¼ 8, p ¼ .16)

0.874 (SE ¼ 0.13; w2 ¼ 11.25, df ¼ 8, p ¼ .19)

0.421

a The parameters can roughly be interpreted as follows: a parameter of 0.072 means that a company with a score of 1, e.g., has a probability of p ¼ exp (10.072)/[1þexp (10.072)] ¼ .72 for endorsement while a parameter of 0.251 means that the probability amounts to .68. However, for actual category probabilities, the threshold parameters (not presented here) have to be taken into account.

Person separation index

Model ﬁt:

0.85

No export decision would be made without information collected by this method

uindeci

Country mean ANOVA: p o.0001 UK vs. AT, USA, NZL, GER

Without information gathered 0.074 by this method, decisions made would be very different

uidiffr

Understanding of Construct Validity and Cross-National Comparability 345

346

THOMAS SALZBERGER ET AL.

parameters. In parenthesis, the standard error of the location estimate is presented, as well as the w2 ﬁt statistic with its degrees of freedom and the associated p-value. At the bottom of Table 1, overall ﬁt statistics and the country means are given. The overall model ﬁt is marginal, but given the large number of individual tests of item ﬁt, it can be considered satisfactory. The same applies to the PSI. Turning to the scores of companies, the mean of companies from the US, New Zealand, and Austria are virtually identical. The mean of Germany is slightly higher but the difference is not signiﬁcant. By contrast, UK companies have a signiﬁcantly (p o.0001) lower tendency to use information gathered from EMI in an instrumental way. As the Rasch analysis has identiﬁed one misﬁtting item and for all remaining items at least some degree of DIF, the comparison of the results based on Rasch measures with an analysis of variance applied to raw scores is of interest (see Table 2). The raw scores have been computed by adding all items except the item /uioutdaS, which misﬁts in the Rasch analysis and also shows a consistently low loading in factor analytic analyses. The original scoring function has been employed. Consequently, the ‘‘naı¨ ve’’ raw score differs from the raw score as the sufﬁcient statistic in the Rasch model. Furthermore, the country means reported by Diamantopoulos et al. (2003) are considered. It should be noted that Diamantopoulos et al. (2003) as well as Souchon et al. (2003), in their analyses, did not differentiate between instrumental and conceptual use of information because of a lack of discriminant validity according to standard factor analytic procedures. In other words, items from both scales were analyzed conjointly. Still, the results from the Rasch analysis and the measures by Diamantopoulos et al. (2003) are, in principle, comparable since both claim to measure one and the same latent variable. If the latent variable comprising instrumental and conceptual use of information is different from instrumental use, only then instrumental and conceptual use items could not have been mixed in the ﬁrst place. The differences between the measures are striking. In the Rasch analysis, UK companies score lowest and German companies highest. However, there are no signiﬁcant differences between Germany and the US, Austria, and New Zealand. For the last three countries, the means of the Rasch measures are virtually identical. The raw scores provide a similar picture, but the UK signiﬁcantly differs only from Germany. The same applies to companies from the US. Thus, the Rasch measures indicate the corporate usage of export information much more precisely demonstrating vividly that conclusions based on simple raw scores are inappropriate.

347

Understanding of Construct Validity and Cross-National Comparability

Table 2. Sort of Measures

Analysis of variance Countriesb

UK (n ¼ 187) USA (n ¼ 141) AT (n ¼ 218) NZL (n ¼ 224) GER (n ¼ 166) Signiﬁcance within subgroup

Mean Comparisons of Instrumental Usage of Export Information. Rasch Measures

Raw Scores

Raw Meansa (Diamantopoulos et al., 2003, p. 29)

F ¼ 7.33, df1 ¼ 4, df2 ¼ 931, p ¼ .00001

F ¼ 5.72, df1 ¼ 4, df2 ¼ 931, p ¼ .0001

ANCOVA F ¼ 18.075, df ¼ 4, po.0001

Homogeneous subgroup 1

Homogeneous subgroup 2

0.85

p ¼ 1.0

1.27 1.29 1.29 1.48 p ¼ .57

Homogeneous subgroup 1 27.87 28.36 28.89 28.90 p ¼ .14

Homogeneous subgroup 2

28.89 28.90 29.66 p ¼ .43

3.43 3.45 3.49 3.74 3.58 n.a.c

a

In the study by Diamantopoulos et al. (2003), a common scale was derived for instrumental and conceptual use of information. b AT stands for Austria, NZL for New Zealand, GER for Germany. c In the study by Diamantopoulos et al. (2003), no multiple group comparisons are reported.

This is due to three reasons. First, the Rasch analysis performs a non-linear transformation of Rasch raw scores to measures taking the boundaries of item responses into account. Second, the Rasch raw scores differ from ordinary raw scores by the adjusted scoring functions that were applied as a consequence of the investigation of the threshold ordering. Third, the different functioning of the items for different countries, assessed by the DIF analysis, has also an impact on the transformation of raw scores to measures. Although the last issue could, in principle, be accounted for by invariance testing based on traditional measurement procedures (multigroup CFA), the other two procedures are unique to Rasch analysis. The raw means in Diamantopoulos et al. (2003) fail to reveal the large difference between the UK, and the US and Austria, respectively, for which highly similar means are reported. By contrast, New Zealand scores highest while, according to the Rasch analysis, there is no real difference to the US and Austria.

348

THOMAS SALZBERGER ET AL.

Cross-Source Stability Since the items have been applied to other sources of information, namely EMR and EA, the scale has also been analyzed referring to these information sources. EMR as a source of information appears to entail a more consistent interpretation of the indicators across cultures. However, sample sizes are considerably lower, ranging from n ¼ 73 (US) to n ¼ 132 (New Zealand) leading to much less power in detecting lack of item invariance. The same applies to information based on EA where the sample sizes lie in the same range. In principle, data from all three information sources could be analyzed simultaneously. First, a more stringent model, a so-called multifacet model (Linacre, 1989), could be set up according to which a common structure of the items prevails over all three sources (i.e. there is only one item location parameter per item) with an additional information source parameter accounting for mean differences in the item locations between sources. However, there is no theoretical basis for this assumption to be viable. We expect some similarity in scale usage across information sources but such a rigorous degree of stability is very unlikely. Moreover, the dependence of the degree of instrumental usage of information on its source between companies is not expected to be identical. A multifacet model would assume a constant difference between information sources. In fact, all attempts to estimate such a model failed. Second, we could conjointly analyze the complete data by stacking response data from all three sources. Technically, the information source would become a person factor, for which a DIF analysis could be carried out. In contrast to the multifacet model, for each company multiple location parameters would be estimated (provided the company supplied data for more than one source of information). Since companies appear up to three times in the data, there is some dependency in the response patterns which may cause problems. Consequently, results should be interpreted with caution. This sort of analysis identiﬁes DIF for one variable only, the item /uindeciS: no export decision would be made without information collected by this method. It is easier to endorse the question for information from marketing intelligence than for information from marketing research, for which the item, however, shows unsatisfactory ﬁt. Given the comparatively small sample sizes for EMR and EA, we also address the problem by looking at the correlations of item location estimates from separate calibrations within each source (see Table 3 for results).

349

Understanding of Construct Validity and Cross-National Comparability

Table 3. Country

Correlations of Item Locations across Sources of Information, Instrumental Use of Information. Type of Use: Instrumental Use

United Kingdom Source: Source: Austria Source: Source: United States Source: Source: New Zealand Source: Source: Germany Source: Source:

export export export export export export export export export export

marketing assistance marketing assistance marketing assistance marketing assistance marketing assistance

Source: Export Source: Export Assistance Marketing Research

intelligence

.90

intelligence

.93

intelligence

.89

intelligence

.89

intelligence

.62

.88 .96 .89 .97 .73 .96 .73 .96 .83 .85

For Austria and the UK, all correlations are very high, ranging from .88 to .97. For the remaining countries, some correlations clearly fall short of this level. This, however, is exclusively due to the variable no export decision would be made without information collected by this method. After excluding this variable from the analysis, all correlations are greater than 0.85, demonstrating a strong stability of the scale across information sources. This is an important ﬁnding because otherwise the construct of information processing could not be conceptualized in the same way.

Correlations of Respondent (Company) Locations Although the stability of the scales across sources of information is a problem of item location invariance, the dependence of information usage on the speciﬁc source is a question of correlations between respondent locations. Corporate information processing behavior may be independent of the source of information, i.e. companies use information instrumentally to a certain degree across sources. Then person locations as measures of instrumental use based on different sources of information should be highly correlated. Alternatively, companies may show speciﬁc patterns of information usage depending on the source, i.e. some companies use information from EMR more instrumentally than information from EA,

350

THOMAS SALZBERGER ET AL.

whereas the opposite is true for other companies. In this case, the person locations should be uncorrelated. Table 4 displays the correlations for instrumental use of information. They are in the range of (non-signiﬁcant) 0.07–0.58. This suggests that instrumental use of information does not resemble a ‘‘corporate trait,’’ implying information is used in the same way irrespective of its source. Since correlations are insensitive to mean differences, high correlations only indicate similar patterns across countries but do not imply equal means. Mean differences between sources have therefore been tested by analyses of variance for repeated measures. For instrumental use, throughout samples from all countries information collected through EMI is most strongly used in an instrumental way. With the exception of New Zealand, EMR and EA yield information that is used instrumentally to the same extent. New Zealand companies use marketing research–based information more instrumentally than information from EA. In summary, instrumental use of information depends on the company and the source of information though there are systematic differences between sources. Surprisingly, it is the less systematic EMI as a source of information that yields market knowledge that is used most intensively in a rationale, i.e. instrumental way.

Table 4. Country

United Kingdom Austria United States New Zealand Germany

Correlations of Company Measures across Sources of Information, Instrumental Use of Information. Instrumental Use of

Export Export Export Export Export Export Export Export Export Export

marketing marketing marketing marketing marketing marketing marketing marketing marketing marketing

* A correlation signiﬁcant at a ¼ 0.05. ** A correlation signiﬁcant at a ¼ 0.01.

research assistance research assistance research assistance research assistance research assistance

Instrumental Use of Export assistance

Export marketing intelligence

.54**

.16* .26** .31** .42** .23* .22* .07 .27** .41** .30**

.34** .58** .41** .36**

Understanding of Construct Validity and Cross-National Comparability

351

Other Types of Information Usage Besides instrumental use of information, conceptual and symbolic use have been referred to in the study. The items intended to measure conceptual use consistently fail to meet minimum requirements of ﬁt to the Rasch model. By contrast, for EMI as the source of information, all items in the scale for symbolic use ﬁt fairly well.4 However, only two items are invariant across all ﬁve countries, four items are invariant for the Anglo-Saxon countries and equivalent between Austria and Germany but not across the different languages. This suggests there being a stronger language-related DIF for items of symbolic use. From a theoretical point of view, this is not unreasonable as the exact meaning of these items is much harder to maintain when translating the instrument. There might also be a different propensity to admit that information is used in a non-reasonable way between the Germanic and the Anglo-Saxon countries. Fortunately, the two fully invariant items are relatively neutral in this respect providing conﬁdence that a comparable scale metric has been deﬁned correctly. As Diamantopoulos et al. (2003) did not detect any violations of item invariance and, consequently, no corrective action was undertaken, differences in the mean country scores are to be expected. In fact, there is some agreement as both analyses identify Austrian companies as most reluctant to use information in a symbolic way whereas US companies score highest. However, a closer look reveals noteworthy nonconformity as far as distances between Germanic and Anglo-Saxon countries are concerned. For instance, Diamantopoulos et al. (2003) report an equal distance between the UK and New Zealand on the one hand and the UK and Austria on the other. The Rasch analysis, accounting for DIF in the usage of the items, shows that the distance between UK and Austria is 2.98 times as large as that between UK and New Zealand. Similarly, the distance between the UK and the US is 6.5 times the distance between the UK and Germany using raw scores while according to the Rasch analysis the distance between the UK and the US is actually smaller with the mean difference between the UK and Germany being 2.89 times as large. Thus, the non-invariance of many items is all but negligible.

DISCUSSION OF CONCLUSIONS AND IMPLICATIONS In international research in marketing, it is now standard to address the problem of measurement equivalence one way or another. However, the

352

THOMAS SALZBERGER ET AL.

complexity of item invariance requires sophisticated approaches such as multigroup CFA or DIF-testing in Rasch analysis. Conventional methods such as the investigation of reliability coefﬁcients or the proﬁle analysis may point at severe problems in the data, but they suffer from, at least, three major shortcomings. First, there are serious theoretical limitations. Reliability informs us of the amount of error given the true variance in the data. The latter need not be invariant across countries, and additive biases will not affect reliability directly anyway. Second, they are not sensitive enough to provide positive evidence of measurement equivalence. Consequently, these methods entail a conﬁrmative bias. Third, no corrective actions are advised in order to account for incomplete measurement equivalence if a lack of invariance is indicated. The empirical study conﬁrms these concerns. Neither a thorough analysis of reliability coefﬁcients nor a proﬁle analysis could reveal limitations in terms of cross-country measurement invariance. The Rasch model not only uncovers DIF but also provides for appropriate adjustments. Comparisons of Rasch measures and raw scores across countries demonstrate that the impact of DIF is substantial. Hence, one should not expect selective DIF to cancel out when measures are estimated from multiple item scales. Although, in principle, DIF for one item may work in the opposite direction compared to DIF for another item, such ‘‘auto-corrective’’ circumstances must not be presumed without empirical evidence. Furthermore, the study conﬁrms once again that the Rasch model is applicable to attitudinal constructs. Although there are no theoretical concerns anyway, marketing scholars have been reluctant to submit their non-achievement constructs to Rasch analysis, or to more general IRT models. In marketing, the Rasch model has been applied to constructs as diverse as brand image (Mazanec, 1979), brand attitude (Bechtel and Wiley, 1983, Bechtel, 1985), consumer acquisition patterns for durable goods (Soutar, Bell, & Wallis, 1990), ownership patterns for durable goods and ﬁnancial assets (Soutar & Cornish-Ward, 1997), people’s leisure activities (Soutar & Ryan, 1999), perception of advertisements (Ewing et al., 2005), technophobia (Salzberger & Sinkovics, 2006), cognitive dissonance (Koller & Salzberger, 2007), affective response to consumption (Ganglmair & Lawson, 2003a, 2003b; Ganglmair-Wooliscroft, 2007) and audience engrossment (Scott, Harris, & Craig-Lees, 2007; Scott & Salzberger, 2008). This study demonstrates that the Rasch model can also be applied to variables at the ﬁrm level rather than at the level of individual traits. Strictly speaking, the main conclusions are not conﬁned to cross-national or cross-cultural research, although measurement equivalence and appropriate

Understanding of Construct Validity and Cross-National Comparability

353

ways to test for and deal with DIF are particularly important in studies involving this sort of heterogeneity. However, even in intra-cultural measurement, DIF may be an issue and a more thorough testing for invariance is advisable. It is to be hoped that methodological contributions in international marketing will eventually have a repercussion on measurement in marketing in general. Compared to multigroup CFA, the Rasch model offers some fundamental advantages, which are attractive from both a theoretical and a practical perspective. In terms of the theoretical underpinning of measurement in the social science, we should keep in mind that the measurement of latent variables implies a strong hypothesis, namely that a latent variable really exists and that manifest indicators are suitable to quantify that variable. The Rasch model meets the requirements that follow from this in a much more rigorous way than CFA does. The non-linearity of bounded item raw scores is taken into account by estimating linear, interval-scaled person and item measures in the Rasch model. The item measures, i.e. their locations, inform us about the structure of the construct in a way that goes considerably beyond what CFA tells us. The structure of the threshold estimates helps us identify malfunctioning response scales advising us to apply a proper scoring scheme and improve response scales in future research. Particularly convenient for international research is the ease with which measurement equivalence can be assessed. Since invariance is a fundamental property that is to be investigated in any Rasch analysis, international research does not imply increased complexity. Finally, the powerful way of establishing indirect links between countries allows us to compare measures across groups even without a single item that is invariant for all countries or the two countries in question. This is particularly appealing in the case of many cultures or very different cultures. Finally, more trustworthy measures of latent constructs also increase our conﬁdence in tests of structural hypotheses and the conclusions drawn. This promises a favorable effect on the application of scientiﬁc ﬁndings to practical marketing problems. In particular, the consideration of each individual item response pattern in the Rasch model allows for a considerably enhanced assessment of measures at the individual level, i.e. the ﬁrm level in the present context. This applies to the precision of a measure as expressed by its standard error, as well as its foundation due to the availability of person ﬁt statistics (Hulin et al., 1983; Smith, 1986; Klauer, 1995; Molenaar & Hoijtink, 1996; Meijer, 2001; von Davier & Molenaar, 2003).

354

THOMAS SALZBERGER ET AL.

At present, we are far from exploiting the potential that lies in the Rasch model as far as measurement of latent variables in international marketing, and in marketing in general, is concerned.

NOTES 1. In this contribution, the terms cross-cultural and cross-national are used virtually interchangeably, since we focus on the methodological consequences of measurement across multiple groups. It goes without saying that we strongly advocate a sound operationalization of culture and that we are aware of the problems of using country as a proxy for culture. A deep understanding of culture facilitates theoretically well-founded expectations as to how the measurement instrument would behave. 2. Originally, IRT models employed the cumulative normal distribution (normal ogive models). Since the cumulative logistic function deviates only insigniﬁcantly (if a normalizing constant is used) from the cumulative normal but is much easier to handle, modern IRT models are based on the logistic function. 3. The Rasch model, as discussed in this contribution, is conﬁned to unidimensional measurement. Although multidimensional Rasch models do exist, applications of Rasch models typically involve unidimensional scales. This does not mean that multidimensional constructs could not be measured on a Rasch basis. We simply need separate unidimensional models for each dimension, and no item can be indicative of more than one dimension. 4. Detailed results (item location estimates, ﬁt statistics, etc.) are available from the ﬁrst author on request.

REFERENCES Adler, L. L. (Ed.). (1977). Issues in cross-cultural research. Annals of the New York Academy of Sciences, New York. Adler, N. J. (1983). A typology of management studies involving culture. Journal of International Business Studies (Fall), 29–47. Andrich, D. (1978). A rating formulation for ordered response categories. Psychometrika, 43(4), 561–573. Andrich, D. (1982). An index of person separation in latent trait theory, the traditional KR-20 Index, and the Guttman scale response pattern. Education Research and Perspectives, 9(1), 95–104. Andrich, D. (1988). A general form of Rasch’s extended logistic model for partial credit scoring. Applied Measurement in Education, 1(4), 363–378. Andrich, D. (1995a). Models for measurement, precision and the non-dichotomization of graded responses. Psychometrika, 60(1), 7–26. Andrich, D. (1995b). Further remarks on the non-dichotomization of graded response. Psychometrika, 60(1), 37–46.

Understanding of Construct Validity and Cross-National Comparability

355

Andrich, D., & Luo, G. (2003). Conditional pairwise estimation in the Rasch model for ordered response categories using principal components. Journal of Applied Measurement, 4(3), 205–221. Andrich, D., Sheridan, B. S., & Luo, G. (2003a). RUMM (Rasch unidimensional measurement models) 2020t version 4.0 for Windowst upgrade version 4600.0159, Murdoch/Perth. Western Australia: RUMM Laboratory Pty Ltd. Andrich, D., Sheridan, B. S., & Luo, G. (2003b). Displaying the Rumm2020 Analysis. RUMM Laboratory Working Paper, Perth, Western Australia. Arbuckle, J. L. (2005). AMOS 6.0 [computer software]. Spring House, PA: Amos Development Corporation. Balnaves, M., & Caputi, P. (2001). Introduction to quantitative research methods, an investigative approach. London: Sage. Bauer, E. (1989). U¨bersetzungsprobleme und u¨bersetzungsmethoden bei einer multinationalen marketingforschung [problems of and methods for translation in multinational research]. GfK Jahrbuch der Absatz- und Verbrauchsforschung, 2, 174–205. Bauer, E. (1995). Internationale marketingforschung [international marketing research]. Mu¨nchen: Oldenbourg. Bechtel, G. G. (1985). Generalizing the Rasch model for consumer rating scales. Marketing Science, 4(1), 66–77. Bechtel, G. G., & Wiley, J. B. (1983). Probabilistic measurement of attributes: A logit analysis by generalized least squares. Marketing Science, 2(4), 389–405. Bentler, P. M. (2003). EQS 6.1 for Windows [computer software]. Encino, CA: Multivariate Software. Berry, J. W. (1980). Introduction to methodology. In: H. C. Triandis & J. W. Berry (Eds), Handbook of cross-cultural psychology. Vol. 2. Methodology (pp. 1–28). Boston, MA: Allyn and Bacon. Birnbaum, A. (1968). Some latent trait models and their use in inferring an examinee’s ability. In: F. M. Lord & M. R. Novick (Eds), Statistical theories of mental test scores. Reading, MA: Addison-Wesley, Chapters 17–20. Bollen, K. A. (1989). Structural equations with latent variables. New York, NY: Wiley. Borsboom, D. (2005). Measuring the mind: Conceptual issues in contemporary psychometrics. Cambridge: Cambridge University Press. Borsboom, D., Mellenbergh, G. J., & Van Heerden, J. (2004). The concept of validity. Psychological Review, 111, 1061–1071. Cavusgil, S. T., & Das, A. (1997). Methodological issues in empirical cross-cultural research: A survey of the management literature and a framework. Management International Review, 37(1), 71–96. Churchill, G. A. (1979). A paradigm for developing better measures of marketing constructs. Journal of Marketing Research, XVI(February), 64–73. Churchill, G. A., & Iacobucci, D. (2005). Marketing research, methodological foundations (9th ed.). Mason, OH: South-Western. Craig, C. S., & Douglas, S. P. (2002). Conducting international marketing research in the twenty-ﬁrst century. International Marketing Review, 18(1), 80–90. Craig, S. C., & Douglas, S. P. (2000). International marketing research (2nd ed.). Chichester: Wiley. Davis, H. L., Douglas, S. P., & Silk, A. J. (1981). Measure unreliability: A hidden threat to cross-national marketing research? Journal of Marketing, 45(Spring), 98–109.

356

THOMAS SALZBERGER ET AL.

De Jong, M. G., Steenkamp, J.-B. E. M., & Fox, J.-P. (2007). Relaxing measurement invariance in cross-national consumer research using a hierarchical IRT model. Journal of Consumer Research, 34(2), 260–278. Deshpande´, R., & Zaltman, G. (1982). Factors affecting the use of market research information: A path analysis. Journal of Marketing Research, 19, 14–31. Diamantopoulos, A., & Souchon, A. L. (1998). Information utilisation by exporting ﬁrms: Conceptualisation, measurement, and impact on export performance. In: S. Urban & C. Nanopoulos (Eds), Information and management: Utilization of technology – structural and cultural impact (pp. 111–140). Wiesbaden: Gabler. Diamantopoulos, A., & Souchon, A. L. (1999). Measuring export information use: Scale development and validation. Journal of Business Research, 46(1), 1–14. Diamantopoulos, A., Souchon, A. L., Durden, G. R., Axinn, C. N., & Holzmu¨ller, H. H. (2003). Towards an understanding of cross-national similarities and differences in export information utilization: A perceptual mapping approach. International Marketing Review, 20(1), 17–43. Douglas, S. P., & Craig, S. C. (1983). International marketing research. Englewood Cliffs: Prentice-Hall. Embretson, S. E., & Reise, S. P. (2000). Item response theory for psychologists. Mahwah, NJ: Lawrence Erlbaum Associates. Ewing, M., Salzberger, T., & Sinkovics, R. (2005). An alternate approach to assessing crosscultural measurement equivalence in advertising research. Journal of Advertising, 34(1), 17–36. Fischer, G. H., & Molenaar, I. W. (Eds). (1995). Rasch models, foundations recent developments, and applications. New York, NY: Springer. Ganglmair, A., & Lawson, R. (2003a). Advantages of Rasch modelling for the development of a scale to measure affective response to consumption. European Advances in Consumer Research, 6, 162–168. Ganglmair, A., & Lawson, R. (2003b). Measuring affective response to consumption using Rasch modelling. Journal of Customer Satisfaction, Dissatisfaction and Complaining Behavior, 16, 198–221. Ganglmair-Wooliscroft, A. (2007). A comparison of affective response to consumption in two contexts. der markt, 46(1þ2), 50–60. Hambleton, R. K., & Swaminathan, H. (1985). Item response theory: Principles and applications. Boston, MA: Kluwer Nijhoff Publishing. Hambleton, R. K., Swaminathan, H., & Rogers, H. J. (1991). Fundamentals of item response theory. Newbury Park, CA: Sage. Hart, S., & Diamantopoulos, A. (1993). Marketing research activity and company performance: Evidence from manufacturing industry. European Journal of Marketing, 27(5), 54–72. Holzmu¨ller, H. H. (1995). Konzeptionelle und methodische Probleme in der interkulturellen management- und marketingforschung [conceptual and methodological problems in crosscultural management and marketing research]. Stuttgart: Scha¨ffer Poeschel. Horn, J. L., & McArdle, J. J. (1992). A practical and theoretical guide to measurement invariance in aging research. Experimental Aging Research, 18(3), 117–144. Hulin, C. L., Drasgow, F., & Parsons, C. (1983). Item response theory: Applications to psychological measurement. Homewood, IL: Dow & Jones Irwin. Jo¨reskog, K. G., & So¨rbom, D. (2003). LISREL 8.54 [computer software]. Lincolnwood, IL: Scientiﬁc Software International.

Understanding of Construct Validity and Cross-National Comparability

357

Kane, M. T. (2001). Current concerns in validity theory. Journal of Educational Measurement, 38, 319–342. Karabatsos, G. (2001). The Rasch model, additive conjoint measurement, and new models of probabilistic measurement theory. Journal of Applied Measurement, 2(4), 389–423. Katsikeas, C. S. (1994). Perceived export problems and export involvement: The case of Greek exporting manufacturers. Journal of Global Marketing, 7(4), 29–57. Klauer, K. C. (1995). The assessment of person ﬁt. In: G. H. Fischer & I. W. Molenaar (Eds), Rasch models: Foundations, recent developments and applications (pp. 97–110). New York, NY: Springer Verlag. Knorr, K. D. (1977). Policymakers’ use of social science knowledge: Symbolic or instrumental? In: C. H. Weiss (Ed.), Using social research in public policy making. Lexington, MA: Lexington Books/D.C. Heath and Company. Knott, J., & Wildavsky, A. (1981). If dissemination is the solution, what is the problem? In: R. F. Rich (Ed.), The knowledge cycle (pp. 99–136). Beverly Hills: Sage. Koller, M., & Salzberger, T. (2007). Cognitive dissonance as a relevant construct throughout the decision-making and consumption process – An empirical investigation related to a package tour. Journal of Customer Behaviour, 6(3), 217–227. Kreiner, S., & Christensen, K. B. (2007). Validity and objectivity in health-related scales: Analysis by graphical loglinear Rasch models. In: M. von Davier & C. Carstensen (Eds), Multivariate and mixture distribution Rasch models. Extensions and applications (pp. 329–346). New York: Springer. Leonidou, L. C. (1995). Export barriers: Non-exporters’ perceptions. International Marketing Review, 12(1), 4–25. Linacre, J. M. (1989). Many-facet Rasch measurement. Chicago, IL: MESA Press. Linacre, J. M. (1998). Thurstone thresholds and the Rasch model. Rasch Measurement Transactions, 12(2), 634–635. Linacre, J. M. (2001). Category, step and threshold: Deﬁnitions and disordering. Rasch Measurement Transactions, 15(1), 794. Linacre, J. M. (2007). Winsteps 3.64.1, Rasch Model Computer Program, Chicago, IL, 1991–2007. Loevinger, J. (1954). The attenuation paradox in test theory. Psychological Bulletin, 51, 493–504. Lord, F. M. (1980). Applications of item response theory to practical testing problems. Hillsdale, NJ: Lawrence Erlbaum Associate. Lord, F. M., & Novick, M. R. (1968). Statistical theories of mental test scores. Reading, MA: Addison-Wesley. Luce, R. D., & Tukey, J. W. (1964). Simultaneous conjoint measurement: A new type of fundamental measurement. Journal of Mathematical Psychology, 1, 1–27. Mair, P., & Hatzinger, R. (2007). Extended Rasch modeling: The eRm package for the application of IRT models in R, Research Report Series/Department of Statistics and Mathematics, Nr. 47, January 2007, Department of Statistics and Mathematics, WU Wien. Malpass, R. S., & Poortinga, Y. H. (1986). Strategies for design and analysis. In: W. J. Lonner & J. W. Berry (Eds), Field methods in cross-cultural research, cross-cultural research and methodology series (Vol. 8, pp. 47–83). Beverly Hills, CA: Sage. March, J. G., & Shapira, J. (1982). Behavioral decision theory and organizational decision theory. In: G. R. Ungson & D. N. Braunstein (Eds), Decision-making: An interdisciplinary inquiry. Boston, MA: Kent.

358

THOMAS SALZBERGER ET AL.

Masters, G. N. (1982). A Rasch model for partial credit scoring. Psychometrika, 47(2), 149–174. Mazanec, J. (1979). Probabilistische meXverfahren in der marketingforschung: Ein empirischer anwendungsversuch zur planung absatzwirtschaftlicher strategien des imagetransfers, [probabilistic measurement models in marketing research: an empirical application to image transfer in marketing planning]. Marketing-Zeitschrift fu¨r Forschung und Praxis, 3, 174–186. Meijer, R. R. (2001). Winsteps diagnosing item score patterns using IRT based person-ﬁt statistics. Research Report. Enschede: Faculty of Educational Science and Technology, University of Twente, The Netherlands Meredith, W. (1964). Notes on factorial invariance. Psychometrika, 29(2), 177–185. Meredith, W. (1993). Measurement invariance, factor analysis and factorial invariance. Psychometrika, 58(4), 525–543. Michell, J. (1997). Quantitative science and the deﬁnition of measurement in psychology. British Journal of Psychology, 88, 355–383. Michell, J. (1999). Measurement in psychology – A critical history of a methodological concept. Cambridge: Cambridge University Press. Molenaar, I. W. (1995). Estimation of item parameters. In: G. H. Fischer & I. W. Molenaar (Eds), Rasch models, foundations recent developments, and applications (pp. 39–51). New York, NY: Springer. Molenaar, I. W. (1997). Lenient or strict application of IRT with an eye on the practical consequences. In: J. Rost & R. Langenheine (Eds), Applications of latent trait and latent class models in the social sciences (pp. 38–49). Mu¨nster: Waxmann. Molenaar, I. W., & Hoijtink, H. (1996). Person-ﬁt and the Rasch model, with an application to knowledge of logical quantors. Applied Measurement in Education, 9(1), 27–45. Morris, T., & Pavett, C. M. (1992). Management style and productivity in two cultures. Journal of International Business Studies, 23(1), 169–179. Mullen, M. R. (1995). Diagnosing measurement equivalence in cross-national research. Journal of International Business Studies, 26(3), 573–596. Muthen, L. K. & Muthen, B. O. (2007). Mplus 5.1 [computer software]. Okazaki, S., & Mueller, B. (2007). Cross-cultural advertising research: Where we have been and where we need to go. International Marketing Review, 24(5), 499–518. Pavett, C., & Morris, T. (1995). Management styles within a multinational corporation: A ﬁve country comparative study. Human Relations, 48(10), 1171–1191. Poortinga, Y. H. (1975). Some implications of three different approaches to intercultural comparison. In: J. W. Berry & W. J. Lonner (Eds), Applied cross-cultural psychology (pp. 329–332). Amsterdam: Swets and Zeitlinger. Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests (1980 Reprint of the original publication in 1960 by the Danish Institute for Educational Research). Chicago, IL: MESA Press. Rasch, G. (1961). On general laws and the meaning of measurement in psychology. Berkeley symposium on mathematical statistics and theory of probability (Vol. IV). Berkeley: University of California Press. Rossiter, J. (2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19, 305–335. Salzberger, T., Holzmu¨ller, H. H., Souchon, A. L., Durden, G. R., Diamantopoulos, A., & Axinn, C. N. (2002). An Investigation of Construct Validity and Cross-National Comparability-A Five-Country Study of Export Information Usage in the UK, Austria,

Understanding of Construct Validity and Cross-National Comparability

359

Germany, the United States, and New Zealand. Proceedings of the 31st EMAC conference, Braga, Portugal, May 28–31, 2002. Salzberger, T., & Sinkovics, R. (2006). Reconsidering the problem of data equivalence in international marketing research – Contrasting approaches based on CFA and the Rasch model for measurement. International Marketing Review, 23(4), 390–417. Salzberger, T., Sinkovics, R., & Schlegelmilch, B. B. (1999). Data equivalence in cross-cultural research: A comparison of classical test theory and latent trait theory based approaches. Australasian Marketing Journal, 7(2), 23–38. Schaffer, B. S., & Riordan, C. M. (2003). A review of cross-cultural methodologies for organizational research: A best-practices approach. Organizational Research Methods, 6(2), 169–215. Scott, J., Harris, J., & Craig-Lees, M. (2007). Reﬁning audience engrossment: A comparison of the usefulness of rasch modelling and factor analysis in scale development. In: Proceedings of the 2007 Australia and New Zealand Marketing Academy Conference (ANZMAC), The University of Otago, Dunedin, New Zealand. Scott, J., & Salzberger, T. (2008). Investigating the threshold ordering of the audience engrossment scale using the polytomous Rasch model. 3rd International Conference of Measurement in Health, Education, Psychology and Marketing, January 2008. Perth: University of Western Australia (UWA). Sekaran, U. (1983). Methodological and theoretical issues and advancements in cross-cultural research. Journal of International Business Studies (Fall), 61–73. Sijtsma, K. (2006). Psychometrics in psychological research: Role model or partner in science. Psychometrika, 71(3), 451–455. Singh, J. (1995). Measurement issues in cross-national research. Journal of International Business Studies, 26(3), 597–619. Singh, J. (2004). Tackling measurement problems with item response theory: Principles, characteristics, and assessment, with an illustrative example. Journal of Business Research, 57(2), 184–208. Smith, R. M. (1986). Person ﬁt in the Rasch model. Educational and Psychological Measurement, 46(2), 359–372. Souchon, A. L., & Diamantopoulos, A. (1996). A conceptual framework of export marketing information use: Key issues and research propositions. Journal of International Marketing, 4(3), 49–71. Souchon, A. L., & Diamantopoulos, A. (1999). Export information acquisition modes: Measure development and validation. International Marketing Review, 16(2), 143–168. Souchon, A. L., Diamantopoulos, A., Holzmu¨ller, H. H., Axinn, C. N., Sinkula, J. M., Simmet, H., & Durden, G. R. (2003). Export information use: A ﬁve-country investigation of key determinants. Journal of International Marketing, 11(3), 106–127. Soutar, G. N., Bell, R., & Wallis, Y. (1990). Consumer acquisition patterns for durable goods: A Rasch analysis. Asia Paciﬁc International Journal of Marketing, 2(1), 31–39. Soutar, G. N., & Cornish-Ward, S. (1997). Ownership patterns for durable goods and ﬁnancial assets: A Rasch analysis. Applied Economics, 29, 903–911. Soutar, G. N., & Ryan, M. (1999). People’s leisure activities: A logistic modelling approach. In: Proceedings of the 1999 Australia and New Zealand Marketing Academy Conference (ANZMAC), The University of New South Wales, Sydney, Australia. Steenkamp, J.-B. E. M., & Baumgartner, H. (1998). Assessing measurement invariance in crossnational consumer research. Journal of Consumer Research, 25, 78–90.

360

THOMAS SALZBERGER ET AL.

Toyne, B., & Walters, P. G. P. (1989). Global marketing management: A strategic perspective (2nd ed.). Boston, MA: Allyn and Bacon. Traub, R. E. (1994). Reliability for the social sciences, theory and applications. SAGE measurement methods for the social sciences series (Vol. 3). Thousand Oaks: Sage. Usunier, J.-C. (1996). Marketing across cultures (2nd ed.). London: Prentice Hall. van de Vijver, F. J. R., & Leung, K. (1997). Methods and data analysis for cross-cultural research. Thousand Oaks: Sage. van der Linden, W. J., & Hambleton, R. (Eds). (1997). Handbook of modern item response theory. New York, NY: Springer. von Davier, M., & Molenaar, I. W. (2003). A person-ﬁt index for polytomous Rasch models, latent class models, and their mixture generalizations. Psychometrika, 68(2), 213–228. Winter, L. G., & Prohaska, C. R. (1983). Methodological problems in the comparative analysis of international marketing systems. Academy of Marketing Science, 11(4), 417–442. Wright, B. D. (1996). Comparing Rasch measurement and factor analysis. Structural Equation Modeling, 3(1), 3–24. Wu, M. L., Adams, R. J., & Wilson, M. R. (1997). ConQuest: Generalised item response modelling software [computer program]. Hawthorn, Victoria: Australian Council for Educational Research. Zhang, J., Beatty, S. E., & Walsh, G. (2008). Review and future directions of cross-cultural consumer services research. Journal of Business Research, 61(3), 211–224.

THE DYNAMICS OF TECHNOLOGICAL READINESS IN MARKETING UNITS: WHY CROSS-CULTURAL EXAMINATION IS NECESSARY R. Glenn Richey, Daniel G. Bachrach, Michael G. Harvey and Hui Wang Research on interorganizational relationships in marketing is increasingly focusing on the ability of marketing managers to effectively adopt and/or implement new technology with relational partners (Anderson & Tushman, 1990; Bharadwaj, 2000; Bower & Christensen, 1995; Calantone, Grifﬁth, & Yalcinkaya, 2006; Charitou & Markides, 2003; Cooper & Zmud, 1990; Johnson & Selnes, 2004; Leonidou, Katsikeas, & Hadjimarcou, 2002; Pﬂugheoft, Ramamurthy, Sooﬁ, Yasai-Ardekani, & Zahedi, 2003). Increasingly, marketing managers must consider both simple and advanced technological tools to improve information sharing systems based on productivity and proﬁtability of an entire marketing channel or supply chain. Marketing managers need to assess technology as a tool for improving overall performance and lowering operating and transaction costs in the channel of distribution (Brynjolfsson & Hitt, 1996; Devaraj & Kohli, 2000;

New Challenges to International Marketing Advances in International Marketing, Volume 20, 361–388 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020016

361

362

R. GLENN RICHEY ET AL.

Gove, Sirmon, & Hitt, 2003; Howell & Higgins, 1990; Lee & Bose, 2002; Song & Xie, 2000). As such, technological innovations such as Automatic Replenishment Systems, Enterprise Resource Planning Systems, RFID reading and tracking systems, etc. are being implemented with the hope of advancing the speed and accuracy of joint work in channel systems (Clemons & Gu, 2003; Kumar & Zhao, 1999; Song & Xie, 2000) and ultimately marketing performance (Sanayei & Jamshidian, 2004; Wright, 2002). However, empirical results regarding the manifest utilities of technological innovations have been mixed and marketing managers have thus been hesitant in adopting new technologies (Iansiti, 1998; Globerman, 1975; Henderson & Clark, 1990). Recent research in marketing decision-making has focused on the centrality of technological readiness for effective implementation and use of technology (Froehle, 2006; Parasuraman, 2000; Parasuraman & Colby, 2001), which has been argued to play a pivotal role in the development of competitive advantage (Bahouth, 1994; Colby & Parasuraman, 2003). Parasuraman (2000) deﬁned technological readiness as ‘‘people’s propensity to embrace and use new technologies for accomplishing goals . . . The construct can be viewed as an overall state-of-mind resulting from a gestalt of mental enablers and inhibitors that collectively determine a person’s (group’s or ﬁrm’s) predisposition to use new technologies’’ (p. 308). However, despite the apparent signiﬁcance of technological readiness (Widgery, Tubbs, & Nicholson, 2004), little is known regarding its management, antecedents, or contextual moderators in a marketing context. This research examines two drivers of marketing work unit technological readiness, as well as a potential contextual moderator of its incidence. Speciﬁcally, this research examines the role of work unit performance and task interdependence such as found in channels of distribution or global strategic alliance relationships (Thompson, 1967) across two distinct cultural context in the United States and the People’s Republic of China (PRC) (Hofstede, 1980). Consistent with recent research examining perceptions of the utility of information management systems (Bendoly, Bachrach, Wang, & Zhang, 2006), the current analysis supports striking differences in marketing managers’ technological readiness that appear dependent on the interaction between culture and task interdependence, key to decision-making as found in channel relationships and strategic alliances (Bhagat, Kedia, Harveston, & Triandis, 2002; Kotabe, Martin, & Domoto, 2003; Tse, Lee, Vertinsky, & Wehrung, 1988). We note speciﬁc variables that key decision makers must account for when selecting the ‘‘appropriate’’ work unit(s) for technology implementation in a marketing context.

The Dynamics of Technological Readiness in Marketing Units

363

THEORETICAL FOUNDATION The Resource-Based View of the Firm Technology is often deﬁned as a valuable ﬁrm resource particularly relative to marketing functions in the organization (Barua, Kriebel, & Mukhopadhyay, 1995; Bharadwaj, 2000; Christensen, Johnson, & Rigby, 2002). When resources are customized to match a marketing strategy, they become ﬁrm speciﬁc, and thus central to ﬁrm performance. Marketing employees also may be deﬁned as an essential resource, in the context of key marketing activities that need to be accomplished (Barney, 1986, 2001; Coff, 2002; Dess & Picken, 1999). Because both employees and technology play key roles, we ground this study in the resource-based view (RBV) of the ﬁrm (Barney, 1991, 2001; Peteraf, 1993; Wernerfelt, 1984). The RBV has a historic foundation in the study of internal human interactions impacting ﬁrm performance (Penrose, 1955, 1959). The Penrose focus was revived by Wernerfelt (1984), who expressed the importance of analyzing ﬁrms from the ‘‘resource perspective’’ and criticized extant research as a ‘‘black box’’ economic view largely ignoring the human dimension. Current conceptualizations of the RBV assume the following: asset/resource heterogeneity (ﬁrms must have resources that differ from other ﬁrms), imperfect mobility of assets (ﬁrm assets must not be able to move easily between ﬁrms), ex post limits on competition, and ex ante limits on competition (environmental temporal limitations exist on competitive resource position and valuation) (Peteraf, 1993). These RBV assumptions allow for the development of a framework for resource outcome valuation that compares complex resource ‘‘bundles’’ (e.g., a skilled marketing unit vs. unskilled marketing unit; a highly taskinterdependent marketing unit vs. a highly task-independent marketing unit). This comparison is based on resource dimensions of convertibility, rarity, imitability, and substitutability (Srivastava, Shervani, & Fahey, 1998). Effective work units (particularly in a marketing context) and their ﬁrm-speciﬁc abilities (e.g., technological readiness) are argued to be valuable, rare, inimitable, and organizationally speciﬁc resources when managed correctly. This is similar to the resource-based prescriptions drawn by international marketing researchers examining learning exploration and exploitation (O¨zsomer & Genctu¨rk, 2003). The goal of RBV (Barney, 1991, 2001) is the creation of rents for the ﬁrm (Peteraf, 1993). In a business-to-business context, exchange value is created as a product or service moves down the supply chain (Bowman, 2001;

364

R. GLENN RICHEY ET AL.

Morash & Lynch, 2002). Under conditions of effective deployment, (e.g., assignment of specially trained or selected marketing expertise to a speciﬁc set of marketing tasks), a ﬁrm may generate superior value (Priem & Butler, 2001a, 2001b). Inputs include raw factors (e.g., people, technological hardware) and assets (e.g., technological abilities, managerial abilities). Raw factors are generic resources that can be easily acquired. Assets are more complicated, and thus are the key to the development of a competitive advantage. When applied by key decision makers to an operational context, assets are transformed into capabilities (Grifﬁth & Harvey, 2001). Assets are both tangible (visible) and intangible (invisible) resources controlled by the ﬁrm (Bogaert, Maetens, & van Cauwenbergh, 1994). Marketers often deploy assets such as employees, technological ‘‘knowhow,’’ brand equity, patents, and speciﬁc knowledge (Schultze, 1994). Assets become capabilities in combination with matched organizational processes (cf. Day, 1994). These marketing utility ‘‘bundles’’ include skills and knowledge that create ﬁrm-speciﬁc asset resource combinations (Amit & Shoemaker, 1993; Coff, 2002; Dess & Picken, 1999). When ﬁrms develop the ability to adjust capabilities to match the competitive environment, they become dynamic (Srivastava et al., 1998; Grifﬁth & Harvey, 2001; Helfat & Peteraf, 2003). Thus, when a capability is matched to a cultural context, a task-structure context, and a performance context, it becomes a dynamic capability (Day, 1994). The managerial goal from the RBV model is the development of core competencies for leveraging a ﬁrm’s strategic marketing weapons and tactics towards sustainable competitive advantage. As such, a discrete technology or skill is not a core competency. Rather, a core competency is a set of skills that inﬂuence a set of capabilities. Core competencies are deﬁned by Prahalad and Hamel (1990) as ‘‘the collective learning in the organization, especially how to coordinate diverse production skills and integrate multiple streams of technologies’’ (p. 79). When capabilities, such as superior unit performance and/or high task interdependence (Wageman, 1995) are achieved in combination, a higher order capability may be formed and thus become a core competency. For a ﬁrm to develop a core competency in ‘‘technological readiness’’ there must be a matched combination of managerial marketing skills and assets to be found in the marketing area of the ﬁrm (Coff, 2002). When culturally matched, a dynamic core competency may help ﬁrms develop marketing performance advantages and/or superior rents (Day, 1994). Technological readiness helps predict the beneﬁts of adopting a technology at a dynamic level (Brooke, 1997). If a marketing work unit is

The Dynamics of Technological Readiness in Marketing Units

365

ready to utilize a technology, this becomes a deployable resource of the ﬁrm, and, if matched appropriately, ultimately a dynamic capability when ﬁt to the environment. If the marketing unit is not ready for implementation, the asset may be wasted. If ﬁrms adopt technologies for which units are not ready, this compromises effective asset management and ﬁrm performance (Day, 1994; Dierickx & Cool, 1989; Drury & Farhoomand, 1999). Thus, we argue that readiness for implementation and use of a technological resource may be a better predictor of marketing performance than adoption (Gopalakrishnan & Damanpour, 2000).

CONCEPTUAL DEVELOPMENT Readiness for Technological Implementation As early as 1974, researchers noted the importance of operational and strategic preparedness or readiness (Delbecq, 1974; Froehle, 2006). Despite the growth of readiness research in strategy, marketing, and public policy, existing research on managing technology generally ignores ‘‘readiness.’’ Research typically falls into two broad categories: adoption of technology (Lippman & McCardle, 1991; Lee, 1985) and management (or manipulation) of technology (Banker & Kauffman, 2004; Dolinsky, Vollmann, & Maggard, 1990; Gruber & Niles, 1975; Levi & Nault, 2004). While central to the success of marketing decision-making, too often technology is adopted without decision makers attending to the readiness of the marketing unit(s) exposed to the technology. Ultimately, researchers and key decision makers may be putting the ‘‘cart before the horse,’’ by expecting adoption per se to drive performance or use of the technology (Christensen et al., 2002; Christianson, 2003; Damanpour, 1991; Davis, 1989). We suggest that if key decision makers do not know the competencies of the marketing employees manipulating the technology, especially in cross-cultural contexts, it is impossible to realize full effectiveness in adoption, implementation, and/or market dissemination decisions. The concept of technological readiness is different from adoption. It focuses on marketers’ readiness to utilize technology, and the application of technology to improve marketing performance (Damanpour, 1987; Jassawala & Shashittal, 1998; Parasuraman, 2000). This is an important issue for marketing managers to consider as they attempt to match technologies to work unit and business needs. Often, unit-level technological application decisions take place after adoption. For instance, SAP was adopted by a number of ﬁrms in the

366

R. GLENN RICHEY ET AL.

1990s in both marketing and supply chain units. Following adoption, many of those units and (in turn) ﬁrms struggled with implementation (Liberatore & Breem, 1997; Garvin & March, 1997; Ross, 1998; Tornatsky & Klein, 1982). Thus, our chief research question is: can marketing managers improve technological readiness, given an a priori knowledge of work unit characteristics? The characteristics we focus on in this study are task interdependence (such as found in channel and strategic alliance relationships), past performance, and the moderating role of national culture. The technology itself need not be sophisticated for technological readiness to matter (e.g., purchasing on the Internet from e-tailers). With this in mind operationally, recent research has adapted the readiness construct to ﬁt ﬁrm level readiness in the study of business partners and in business units for application in marketing strategy and international business (Richey, Daugherty, & Roath, 2007). Both technological optimism and technological innovativeness are strong positive sub-dimensions of technological readiness that may contribute to ﬁrm-to-ﬁrm contact quality, order release ﬂexibility, information quality, order requisitioning efﬁciency, accuracy and condition, timeliness, overall customer satisfaction, and ﬁnancial and market-based performance (Bucher, Birkenmeier, Brodbeck, & Escher, 2003).

Direct Effects of Decision Makers: Task Interdependence It is argued that marketing unit task interdependence may play a role in unit-level technological readiness in channels of distribution or global strategic alliances. Task interdependence has several different conceptualizations. Some marketing researchers (e.g., Johnson & Johnson, 1989; Wageman, 1995) distinguish task from resource interdependence (e.g., Thompson, 1967; Van de Ven & Ferry, 1980), while others focus on the processes by which inputs are combined to create outputs. Early on Thompson (1967) viewed interdependence as a characteristic of work tied to a speciﬁc technology while Shea and Guzzo (1989) viewed interdependence as an attribute of employee behavior in project completion. For the purposes of our study, task interdependence is the extent to which marketing managers and employees depend on other members of their unit to carry out their work (Brass, 1985; Kiggundu, 1983; Van der Vegt, Emans, & Van de Vliert, 2001; Van der Vegt & Janssen, 2003). Task interdependence has been shown to increase communication, helping, and information sharing of tacit knowledge (Crawford & Haaland, 1972; Johnson, 1973) that is vital to the unit and inter-unit’s understanding

The Dynamics of Technological Readiness in Marketing Units

367

of new technology. Task interdependence also inﬂuences group cooperation, which contributes to group success when inter-unit cooperation is essential to successfully complete interrelated tasks (Feldman, 1984; Kiesler & Kiesler, 1970; Shaw, 1981). Task interdependence makes mutual helping and other cooperative behaviors, such as learning, important in the completion of tasks (Organ, 1988; Wageman, 1995). Task interdependence also inﬂuences feelings of responsibility for work outcomes (Pearce & Gregersen, 1991; Van der Vegt, Emans, & Van de Vliert, 1998), team identiﬁcation (Van der Vegt, Van de Vliert, & Oosterhof, 2003), and ﬁrm innovativeness (Van der Vegt & Janssen, 2003). Thus, task interdependence may result in the development of norms of marketing cooperation, including coordination and learning that increases marketing employees’ preparedness for changes in operating norms important to issues in global channels of distribution (Edmondson, Winslow, Bohmer, & Pisano, 2003; Feldman, 1984; Thomas, 1957). Past research has supported the fact that work units and marketing managers can differ highly in their readiness for the implementation of technology (Agarwal & Prasad, 1999). Work units that are optimistic about technology express a belief that it offers increased control, ﬂexibility, and efﬁciency (Richey et al., 2007). Cooperative, interdependent, helpful employees are likely to see technology as a tool to both break down bottlenecks and take advantage of ﬁrm and/or market opportunities. Since task interdependence inﬂuences cooperation, learning, innovativeness, and team identiﬁcation it is predicted that: H1. Task interdependence is positively associated with technological readiness. Unit Performance Marketing managers in global supply chain settings continually seek ways to increase unit and inter-unit performance (Brandyberry, Rai, & White, 1999). Yet, research on the topic of productivity and proﬁtability related to technological innovation in work units is mixed (Bharadwaj, 2000; Brynjolfsson & Hitt, 1996; Mata, Fuerst, & Barney, 1995). Analyses often result in insigniﬁcant or negative correlations, possibly caused by issues including but not limited to unit learning errors (Sawhney & Prandelli, 2000). Increasing customer pressure to dispense quality products to the right markets, in the right shape, at the right time complicates the management of these units (Levitt, Thomsen, Christiansen, & Kunz, 1999). As such, complexity drives marketing managers to rely increasingly on the marketing

368

R. GLENN RICHEY ET AL.

unit itself to take the lead and deliver on promises to channel members as well as customers (Richey & Bachrach, 2004). Deductively, better past performance may be a signal to marketing decision makers of higher levels of technological readiness. Marketing units with a history of superior past performance likely possess ﬂexibility with resource tools and agility with the work process. Superior marketing performers may also possess the conﬁdence to implement new technologies and the innovativeness to make technology work properly (Damanpour, 1987; Damanpour, Szabat, & Evan, 1989; Davis, 1989; Dewar & Dutton, 1986). Unit performance acts as a signal for marketing employees within the unit, and inﬂuences how they respond to challenges and change (Coff, 2002; Dyer & Singh, 1998). High performing marketing units develop a culture of performance (Gotwon & DiTomaso, 1992; Reid & Hubbell, 2005), and seek out tools to help them maintain their status over other units (Gaertner & Insko, 2000; Huddy, 2004). Further it is also reasonable to expect that marketing managers and employees in high performing units will have higher motivation and ability than comparable marketing managers and employees in lower performing units (Locke, Mento, & Katcher, 1978). Thus, performance signals, out-group competitiveness, and motivation and ability in high performing marketing units are all stronger than in lower performing units. Given this discussion, it is suggested that: H2. Unit performance is positively associated with technological readiness. Interaction Effects of National Culture: Individualism/Collectivism The lack of inclusion of culture in many studies of technology management and implementation in a marketing context is a noteworthy limitation, given its deep contextual role (Hampden-Turner & Trompenaars, 1998; Tse et al., 1988). Hofstede’s (1980) research has identiﬁed a dimension of culture which we argue may be central in understanding the impact of task interdependence and unit performance on technological readiness: Individualism– Collectivism (IC). This continuum of belongingness reﬂects the level of social interconnectedness among members of a society and tends to vary in systematic ways across societies and individuals within those societies (cf. Hofstede, 1980; Earley & Gibson, 1998). Collectivism shapes the extent to which individuals view and identify themselves through membership in in-groups (Triandis, 1988), which tends

The Dynamics of Technological Readiness in Marketing Units

369

to have systematic effects on social behavior within group settings (Earley, 1989, 1993; Tajfel & Turner, 1986). Some cultures, such as that found in the PRC, develop citizens who are decidedly collectivistic in orientation and others, such as that found in the United States develop citizens who are primarily individualistic (Hofstede, 1980). Within collectivistic contexts, when an out-group member gains in-group status (e.g., trust) this member achieves acceptance and is expected to abide by group norms (Gudykunst, 1983). Once an out-group member is accepted, the relationship becomes more intimate than typical relationships formed within individualistic contexts (Grifﬁth, Hu, & Ryans, 2000; Gudykunst, 1983). An important attribute of contexts characterized by a collectivistic culture is that individuals within those contexts view and identify themselves less as individuals and more as group or unit members (Earley, 1989, 1993; Tajfel & Turner, 1986; Triandis, 1988). The deﬁning distinction between contexts that vary along the IC continuum, and individuals within those contexts whose behaviors are inﬂuenced by the dominant social character, is that, for individualists personal interests are more important than the interests of the (in)group, as are the attainment of individual (as opposed to group) goals. In contrast, for collectivists the interests of the group are paramount (Triandis, 1989; Yamaguchi, 1994), and as noted by Earley (1989), ‘‘a driving force within a collectivistic culture is cooperation so as to attain group goals and safeguard welfare’’ (p. 567). Empirical research, dealing with the moderating effects of IC on social loaﬁng on one end of the IC continuum (Earley, 1993), and the effects of IC on citizenship behaviors on the other end of the performance continuum (cf. Moorman & Blakely, 1995), support these characterizations of individuals within contexts dominated by individualistic or collectivistic cultures. Individuals in contexts-characterized principally by collectivistic cultural values are more likely to expect cooperative behaviors from individuals in their in-unit (cf. Earley, 1989; Lam, Hui, & Law, 1999) than individuals in more individualistic contexts. There is a higher level of social interconnectedness amongst in-unit members in collectivistic contexts than in individualistic contexts (Hofstede, 1980; Earley & Gibson, 1998) as well as a greater emphasis on interpersonal harmony (Leung, Au, Fernadez-Dols, & Iwawaki, 1992). Collectivists tend to expect in-unit members to be focused on the accomplishment of group goals (Triandis, 1989; Yamaguchi, 1994), to have feelings of mutual dependence (Triandis, 1989), and feel bound to one another by common goals, interests, and mutual commitment (Earley & Gibson, 1998). Although there is some empirical evidence that the nature of

370

R. GLENN RICHEY ET AL.

interdependence relationships amongst tasks may affect self-deﬁnitions in the short-term (Breer & Locke, 1965) these effects appear to be temporally bounded (Wagner, VanDyne, LePine, & Hollenbeck, 1997). Thus, although past unit performance is expected to inﬂuence technological readiness, its effects should depend on the cultural context within which readiness is measured. Speciﬁcally, as noted earlier, there are differences in the importance placed on the achievement of group goals (i.e., unit performance) in individualistic vs. collectivistic cultures. Further, in collectivistic contexts, although there is a premium placed on goals among members of the ingroup, no such premium exists in ‘‘out-group’’ contexts, where there is a signiﬁcantly higher level of mistrust and social distance (Earley, 1993). Although in the United States it is expected that past unit performance will inﬂuence technological readiness (irrespective of the level of task interdependence in the unit), the effects of past unit performance on technological readiness in the PRC are expected to depend on the level of task interdependence within the unit. In the United States, within a single work unit, the idea of ‘‘in-group’’ vs. ‘‘out-group’’ is not an explicit cultural phenomenon. Although inter-group identiﬁcation plays a role in self-construal in the United States (Huddy, 2004), this phenomenon exists between groups. However, in the PRC, even within a single work unit, the idea of ‘‘in-group’’ vs. ‘‘out-group’’ has a fundamental inﬂuence on how unit members deﬁne their social relationships, their performance expectations, and loyalties. We argue that among evaluators in the PRC, performance will have the same positive effect on technological readiness as it is predicted to have in the United States, in units characterized by high levels of task interdependence. However, in units characterized by low levels of task interdependence, unit members will not characterize their functionally independent coworkers as members of their ‘‘in-group,’’ and so their perceptions of technological readiness will be less subject to the inﬂuence of performance feedback. Based on the following arguments, the following is hypothesized: H3. The positive relationship between unit performance and technological readiness is similar for evaluators from the United States and PRC in conditions characterized by high task interdependence. However, the relationship between unit performance and technological readiness is mitigated for evaluators in the PRC in conditions characterized by low task interdependence.

The Dynamics of Technological Readiness in Marketing Units

371

RESEARCH METHODOLOGY Design and Participants Following Chen, Chen, and Meindl’s (1998) caution regarding appropriate designs when IC is employed as an exogenous factor, we used a 2 2 2 ‘‘factorial experimental design, with IC having the status of an independent variable, crossed with a second [and third] independent variable’’ (p. 299) in this case, with task interdependence (high vs. low) and unit-level performance (high vs. low) which were manipulated factors. In the U. S. sample, participants were 209 graduate and undergraduate business students, 21.91 years of age (SD ¼ 2.48); 52% were male, 96% were born in the United States, and English was the ﬁrst language of 97%. Participants had 3.53 years of undergraduate education (SD ¼ .68), 1.21 years of full time work experience (SD ¼ 2.27), 1.65 years with their immediately previous organizations (SD ¼ 2.21), and 1 year in their most recent positions (SD ¼ 1.43). Participants had been direct supervisors of 2.51 (SD ¼ 5.70) and indirect supervisors of 4.34 employees (SD ¼ 13.64). In the sample from the PRC, participants were 174 graduate and undergraduate business students, 21.35 years of age (SD ¼ 2.64); 47% were male, 98% were born in the PRC, Chinese was the ﬁrst language of 97%. Participants had 3.25 years of undergraduate education (SD ¼ 1.69), three-quarter years of full time work experience (SD ¼ 2.27), 1.14 years with their immediately previous organizations (SD ¼ 3.61), and .47 years in their most recent positions (SD ¼ 1.55). Participants had been direct supervisors of 2.94 (SD ¼ 6.09) and indirect supervisors of 15.25 employees (SD ¼ 48.53). It should be noted that no response bias was found across groups. A student sample and lab setting was selected to heavily control external biasing conditions. This is consistent with many recent top studies examining a cross-cultural issues (Allen, Takeda, & White, 2005; Chan, Li, Diehl, & Terlutter, 2007; de Pillis, Kernochan, Meilich, Prosser, & Whiting, 2008; Ford, Nonis, & Hudson, 2005; Gefen, Geri, & Paravastu, 2007; Jung & Sung, 2008; Martz Jr., Biscaccianti, Neil, & Williams, 2005; Ross, Broyles, & Leingpibul, 2008; Shiloh, Gu¨venc- , & O¨nkal 2007; Wright & Drewery, 2006). Similar to these studies, we were interested in isolating a basic employee process. The use of a student sample establishes the presence of this process in an isolated setting. Moreover, the results we obtain from this sample establish the presence of a basic process that will be present in all people at some level. If we are able to ﬁnd the process in a student sample with limited professional experience, this will further

372

R. GLENN RICHEY ET AL.

reinforce the issue of its foundational characteristics. Once we have established the presence of this basic process in an examination using a student sample, researchers may then move our test to an organizational context where one may more systematically examine processes within the organization (including characteristics of the sample itself ), that affect the process we observe in the current study. With this in mind – and identify any potential confounds to the study design – we include an analysis of the demographic characteristics of our samples. Such an analysis should determine whether the study variables had systematic effects on the measured variable in the study. Consistent with recent cross-cultural experimental research in the PRC and the United States exploring basic process perceptions with applicability to organizational settings (e.g., Bachrach, Wang, Bendoly, & Zhang, 2007), the current study made use of two equivalent student samples in the PRC and the United States. Consistent with the protocol reported in previous experimental research in the domain (Bendoly et al., 2006), the two samples used in the current study were matched with respect to a range of demographic characteristics to facilitate cross-cultural comparisons. Given the basic-process character of the relationships we examine, and the goal of the current study to develop structurally derived causal inferences derivative of these conceptually exogenous structural factors, the use of student samples is widely recognized in the experimental psychology domain as appropriate to the stated objectives of current study (Berkowitz & Donnerstein, 1982; Campbell & Stanley, 1963; Gordon, Slade, & Schmitt, 1986; Greenberg, 1987).

Procedure In both the United States and the PRC, subjects participated in a behavioral lab where they were randomly assigned to an interdependence (high or low) and a performance (high or low) condition taking the form of a written case (Bachrach, Powell, Bendoly, & Richey, 2006). Once assigned, participants were asked not to communicate with one another, and were read a set of instructions wherein they were informed they were participating in a decision-making task, the goal of which was to examine the technological readiness of employees in production settings. Participants were asked to indicate the extent to which employees in the production unit described in the case demonstrated the attitudes described later. Following appropriate forward and back translation methodology, participants were given

The Dynamics of Technological Readiness in Marketing Units

373

a chance to ask questions, were exposed to the manipulations, provided ratings, were debriefed, and released (see appendix).

RESULTS We used Van der Vegt and Janssen’s (2003) ﬁve-item scale to measure task interdependence. Examples of these items include: Employees in this unit ‘‘ . . . need information and advice from their colleagues to perform their jobs well,’’ and ‘‘need to collaborate with colleagues to perform their jobs well.’’ Previous research using this scale has shown it to have acceptable reliability (a ¼ .85) (Bachrach et al., 2006). To determine the manipulation’s effectiveness, an independent-samples t-test was conducted. Results indicated that interdependence reported for the high (6.15, SD ¼ .85) and low (4.32, SD ¼ 1.78) conditions was signiﬁcantly different [t (382) ¼ 13.01, po.001] and in the expected direction. The ﬁve-item scale used by Bachrach et al. was used to measure overall performance. Examples of items used to measure overall performance in this study included, ‘‘This unit is indispensable to the organization,’’ and ‘‘This unit is extremely valuable to the organization.’’ The reliability of this scale in the current study was (a ¼ .89). Results from an independent-samples t-test indicated that the high (5.40, SD ¼ 1.28) and low (3.97, SD ¼ 1.49) performance conditions were signiﬁcantly different [t (381) ¼ 10.01, po.001] and in the expected direction. IC was measured with Wagner and Moch’s (1986) ﬁve-item values scale, which measures speciﬁc prescriptions for the behavior of work group members. Examples of these items include: Employees in this unit ‘‘ . . . should realize that they are sometimes going to have to make sacriﬁces for the sake of the work group as a whole,’’ and ‘‘ . . . should do their best to cooperate with each other instead of trying to work things out on their own.’’ To determine whether there were differences in the IC values across the two groups, an independent-samples t-test was conducted. Results indicated that the IC values reported for the United States (3.88, SD ¼ 1.02) and the PRC samples (5.04, SD ¼ 1.15) were signiﬁcantly different t (377) ¼ 10.29, po.001 and in the expected direction. Subjects completed these checks immediately following exposure to the manipulations. Construct Means, Standard Deviations, Reliabilities, and Correlations Means, standard deviations, reliability estimates, and correlations are reported in Table 1.

374

R. GLENN RICHEY ET AL.

Table 1. Variable

Means, Standard Deviations, Internal Consistency Reliabilities, and Correlations. Mean

SD

1

2

3

4

5.28 4.67 4.41 4.44

1.65 1.56 1.23 1.02

(.89/.80) .01 .17** .09

(.89/.80) .04 .31**

(.84/.71) .09

(.84/.74)

Interdependence Unit performance IC Values Technology readiness

Note: Reliability estimates appear in parenthesis along the diagonal. The ﬁrst entry inside the parenthesis is Cronbach’s index of internal consistency (a) and the second entry is Fornell and Larcker’s (1981) r. ** po.01, one-tailed.

Table 2. Means, Standard Deviations, and Ranges of Technological Readiness Across Eight Conditions. Minimum

Maximum

Mean

SD

U.S.

PRC

U.S.

PRC

U.S.

PRC

U.S.

PRC

High performance condition High interdependence Technological readiness Low interdependence Technological readiness

1.71

3.00

5.71

6.86

4.79

4.73

.89

.98

2.71

1.71

7.00

6.43

4.71

4.65

.88

.94

Low performance condition High interdependence Technological readiness Low interdependence Technological readiness

1.71

1.57

7.00

5.86

4.11

4.21

.97

1.05

1.71

2.29

5.71

6.43

3.69

4.71

.93

1.03

Hypothesis Tests for the Effects of Interdependence, Performance, and Culture on Technological Readiness The means, standard deviations, and ranges of technological readiness across the eight experimental conditions are presented in Table 2, and the technological readiness means are plotted in Fig. 1. As one can see in Fig. 1, the pattern of interaction between unit performance and task interdependence on the technological readiness ratings provided by evaluators in this study appear to be different across the two culture conditions. The most notable difference between these two conditions appears to be a substantially stronger main effect of unit performance on perceptions of technological

The Dynamics of Technological Readiness in Marketing Units

375

Technological Readiness in the PRC Ratings of Employee Technological Readiness

5.5

5 4.73 4.71

4.65

4.5 4.21 4

3.5

3 Low Unit Performance

High Unit Performance

Unit Performance High Task Interdependence

Ratings of Employee Technological Readiness

5.5

Low Task Interdependence

Technological Readiness in the U.S.

5 4.79 4.71 4.5 4.11 4

3.69 3.5

3 Low Unit Performance High Unit Performance Unit Performance High Task Interdependence

Low Task Interdependence

Fig. 1. Technological Readiness in the U.S. and the PRC Across Unit-Level Performance and Task Interdependence Conditions – Manipulation Checks.

376

R. GLENN RICHEY ET AL.

readiness in the United States across the low task interdependence conditions. Unit-level performance appears to have very similar effects on technological readiness at high levels of task interdependence across the two culture conditions. In order to test H13, analyses of variance were executed to examine the main effects of interdependence and performance on ratings of technological readiness. Consistent with the visual data reported in Fig. 1, and in support of H1, these analyses revealed signiﬁcant main effects of performance for technological readiness [F (1, 357) ¼ 24.90, po.001]. However, these analyses failed to reveal a signiﬁcant main effect of task interdependence, [F (1, 357) ¼ .20, n.s.] on technological readiness, providing no support for H2. We continued this analysis in order to test the three-way interaction effects proposed in H3. In support of H3, these analyses revealed signiﬁcant interaction effects between task interdependence, performance, and culture on technological readiness, [F (1, 357) ¼ 8.53, po.01].

DISCUSSION AND MANAGERIAL IMPLICATIONS The results from this research indicate signiﬁcant complexity in marketing decision-making concerning technology and human resources both domestically and abroad in marketing operating systems. The ﬁndings presented here may exert limitations on our ability to generalize all industries, due to the sample and controlled environment. Yet, it does provide a strong empirical and theoretical basis for summary conclusions and extended research in (global) resource-based analysis in a marketing management context. This study embraces the importance of technological readiness in marketing as a driver of operational efﬁciency as supported by recent research. Moreover, this study examines the inﬂuence of task interdependence and unit-level performance on unit technological readiness. In doing so we offer a suggested experimental design methodological approach that future researchers should employ to isolate individual behavioral conditions in future cross-cultural research. This approach should assist in reducing the large amount of environmental and response error present in may crosscultural studies – especially if the researchers are looking for the existence of three-way interaction effects. Summarizing our results, we suggest that marketing managers should proceed with care when using extant knowledge about task interdependence, past performance, and technological readiness in an environment similar to global channels of distribution. Overall, we show that past performance has

The Dynamics of Technological Readiness in Marketing Units

377

a direct and signiﬁcant positive impact on a work unit’s technological readiness. Conversely, we ﬁnd that no signiﬁcant impact is revealed for task interdependence. Highly successful marketing units are likely to be more ready for the implementation and use of a new technology than less successful units. Weaker marketing managers will be less optimistic and innovative about technology, making implementation more difﬁcult. Decision makers looking to adopt and integrate technology into extant processes should match these initiatives with high performing units. The past performance–technological readiness relationship is important for domestic marketing decision makers attempting to match work units with technologically oriented tasks. However, it is also important in this context to understand the role of culture. Decision makers should not discount the impact of national culture in the management of work units. Depending on a cultural context, managers must consider past performance and task interdependence to evaluate unit technological readiness. The results from our study suggest that culture plays a moderating role on the relationship between work unit performance and technological readiness. Marketing managers hoping to effectively implement new technology must evaluate both past performance and task interdependence. In the U.S. cultural context, more of each is better, such that higher performing units that are more task interdependent are likely to be more optimistic and innovative in their readiness to use technology. On the contrary, lower performing units, with fewer interdependent tasks, are likely to be less ready for implementation and use of a technology. This linear effect should be easy for marketing managers in the United States to understand and manipulate to their advantage. Yet, the effect takes on an entirely different proﬁle in emerging economies with collectivist cultures like that found in the PRC. This discovery can complicate global expansion of channels of distribution market entry decisions/strategies and the management of culturally different marketing employees (cf. Burgel & Murray, 2000). The growing interest in the emerging Chinese economy was a primary driver of our research (Calantone et al., 2006). As marketers attempt to develop strategies in the PRC, and other emerging markets, they must account for the dynamic impact culture has on global managerial decisionmaking concerning technological implementation. In our comparison of the United States and the PRC, we ﬁnd one such important example in that the linear ‘‘higher is better’’ result for marketing units in the United States becomes a complex crossover effect for marketing units in the PRC. Surprisingly, this research reveals that Chinese units with weak past performance, that are less task interdependent, will be superior in terms of

378

R. GLENN RICHEY ET AL.

their technological readiness relative to units with similar past performance that are more task interdependent. This result may be a product of the inﬂuences within collectivist contexts. Unlike the United States, for weaker units, less task interdependence is preferred when targeting new technologies. In such instances, low task interdependence may allow for the development of a group leader or technological champion that can assist in the learning of how to use the technology (Beath, 1991; DeGregori, 2005). Conversely, units that are weak performers, but highly task interdependent, are likely to be less ready for technological implementation and use than others. This very well could be the product of groupthink (Janis & Mann, 1977) causing the unit to stay mired in a stage of weak readiness and unable to learn at a sufﬁcient rate. In the future, researchers should look to the moderating impact of uncertainty avoidance in an attempt to further explain these behaviors.

CONCLUSION This study develops an economic super power vs. emerging market grounding for future research in the management of marketing functions. Overall, this study uncovers a very dynamic trend in the management of resources and speciﬁcally global human resources and taps students who will be the managerial leadership of tomorrow. When crossing borders, marketing managers must examine both unit task-interdependence and past performance if they hope to match technologically ready units to technological implementation initiatives. This study provides an explanation for why technological readiness may differ when moving from the United States to a collectivist emerging market – or more speciﬁcally – to the potential future master of the world economy, the PRC. Global ﬁrms hoping to integrate existing technology into new cultures and new technology into existing marketing processes in collectivist emerging markets will need to adapt their resource decision-making strategy. The ﬁndings of this study suggest that, if marketing managers use a strategy of implementing technology through their most task-interdependent units in collectivist cultures, as they would in the United States, the initiative will likely fail. Marketing managers should be encouraged to target less task-independent units in collectivist contexts where a leader or champion can emerge to drive technological readiness across the unit (Beath, 1991). As the competitive contexts within which organizations initiate the implementation of technology increasingly demand the ﬂexibility implied by

The Dynamics of Technological Readiness in Marketing Units

379

self-managing teams, understanding the implications of task interdependence in this equation is essential (Langfred, 2005). Recognition that the implications of task interdependence for readiness to utilize technology may be different for teams in different cultural contexts facilitates the effective management of culturally distinct work groups. As the use of autonomous work groups across cultural boundaries increases, systematic understanding of the interactions between the variables examined in the current study can only increase the potency of allocations of scarce resources.

REFERENCES Agarwal, R., & Prasad, J. (1999). Are individual differences germane to the acceptance of new information technologies? Decision Sciences, 30(2), 361–391. Allen, R. S., Takeda, M., & White, C. S. (2005). Cross-cultural equity sensitivity: A test of differences between the United States and Japan. Journal of Managerial Psychology, 20(8), 641–662. Amit, R., & Shoemaker, P. J. (1993). Strategic assets and organizational rent. Strategic Management Journal, 14(1), 33–46. Anderson, P., & Tushman, M. (1990). Technological discontinuities and dominant designs: A cyclical model of technological change. Administrative Science Quarterly, 35(6), 604–633. Bachrach, D. G., Powell, B. C., Bendoly, E., & Richey, R. G. (2006). Organizational citizenship behavior and performance evaluations: Exploring the impact of task interdependence. Journal of Applied Psychology, 91(1), 193–201. Bachrach, D. G., Wang, H., Bendoly, E., & Zhang, S. (2007). Importance of organizational citizenship behavior for overall performance evaluation: Comparing the role of task interdependence in China and the USA. Management and Organization Review, 3, 255–276. Bahouth, S. B. (1994). Technological readiness as a business strategy. Industrial Management and Data Systems, 94(8), 8–12. Banker, R. D., & Kauffman, R. J. (2004). The evolution of research on information systems: A ﬁftieth-year survey of the literature in management science. Management Science, 50(3), 281–298. Barney, J. B. (1986). Organizational culture: Can it be a source of sustained competitive advantage? Academy of Management Review, 11(3), 656–665. Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99–120. Barney, J. B. (2001). Resource-based theories of competitive advantage: A ten-year retrospective on the resource-based view. Journal of Management, 27(6), 643–650. Barua, A., Kriebel, C. H., & Mukhopadhyay, T. (1995). Information technologies and business value: An analytic and empirical investigation. Information Systems Research, 6(1), 3–23. Beath, C. M. (1991). Supporting the information technology champion. MIS Quarterly, 15(3), 355–372. Bendoly, E., Bachrach, D. G., Wang, H., & Zhang, S. (2006). ERP in the minds of supervisors: Joint roles of task interdependence and cultural norms. International Journal of Operations and Production Management, 26(5), 558–578.

380

R. GLENN RICHEY ET AL.

Berkowitz, L., & Donnerstein, E. (1982). External validity is more than skin deep: Some answers to criticism of laboratory experiments. American Psychologist, 37, 245–257. Bhagat, R. S., Kedia, B. L., Harveston, P. D., & Triandis, H. C. (2002). Cultural variations in the cross-border transfer of organizational knowledge: An integrative framework. Academy of Management Review, 27(2), 204–219. Bharadwaj, A. S. (2000). A resource-based perspective on information technology capability and ﬁrm performance: An empirical investigation. MIS Quarterly, 24(1), 169–196. Bogaert, I., Maetens, R., & van Cauwenbergh, A. (1994). Strategy as a situational puzzle: The ﬁt of components. In: G. Hamel & A. Heene (Eds), Competency based competition. Chichester, UK: Wiley. Bower, J. L., & Christensen, C. M. (1995). Disruptive technologies: Catching the wave. Harvard Business Review, 73(1), 43–53. Bowman, C. (2001). Value in the resource-based view of the ﬁrm: A contribution to the debate. The Academy of Management Review, 26(4), 501–502. Brandyberry, A., Rai, A., & White, G. P. (1999). Intermediate performance impacts of advanced manufacturing technology systems: An empirical investigation. Decision Sciences, 30(4), 993–1020. Brass, D. J. (1985). Technology and the structuring of jobs: Employee satisfaction, performance, and inﬂuence. Organizational Behavior and Human Decision Processes, 35(2), 216–240. Breer, P. E., & Locke, E. A. (1965). Task experience: A source of attitudes. Homewood, IL: Dorsey Press. Brooke, C. (1997). A modiﬁed framework for the evaluation of technological choice. In: T. McMaster, E. Mumford, E. B. Swanson, B. Warboys & D. Wastell (Eds), Facilitating technology transfer through partnership. London, UK: Chapman and Hall Publishers. Brynjolfsson, E., & Hitt, L. M. (1996). Paradox lost? Firm level evidence on the returns to information systems spending. Management Science, 42(4), 541–558. Bucher, P., Birkenmeier, B., Brodbeck, H., & Escher, J.-P. (2003). Management principles for evaluating and introducing disruptive technologies: The case of nanotechnology in Switzerland. R and D Management, 33(2), 149–163. Burgel, O., & Murray, G. C. (2000). The international market entry choices of start-up companies in high-technology industries. Journal of International Marketing, 8(2), 33–62. Calantone, R. J., Grifﬁth, D. A., & Yalcinkaya, G. (2006). An empirical examination of a technology adoption model for the context of China. Journal of International Marketing, 14(4), 1–27. Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi experimental designs for research. Chicago: Rand McNally. Chan, K., Li, L., Diehl, S., & Terlutter, R. (2007). Consumers’ response to offensive advertising: A cross cultural study. International Marketing Review, 24(5), 606–615. Charitou, C. D., & Markides, C. C. (2003). Responses to disruptive strategic innovation. MIT Sloan Management Review, 44(2), 55–63. Chen, C. C., Chen, X.-P., & Meindl, J. R. (1998). How can cooperation be fostered: The cultural effects of individualism–collectivism. Academy of Management Review, 23(2), 285–304. Christensen, C. M., Johnson, M. W., & Rigby, D. K. (2002). Foundations for growth. MIT Sloan Management Review, 43(3), 22–31. Christianson, C. (2003). Strategy and innovation: Breakthrough insight and ideas for driving growth. Harvard Business School Publishing and Insight, 1(2), 1–4.

The Dynamics of Technological Readiness in Marketing Units

381

Clemons, E. K., & Gu, B. (2003). Justifying contingent information technology investments: Balancing the need for speed of action with certainty before action. Journal of Management Information Systems, 20(2), 11–48. Coff, R. W. (2002). Human capital, shared expertise, and the likelihood of impasse in corporate acquisitions. Journal of Management, 28(1), 107–128. Colby, C. L., & Parasuraman, A. (2003). Technology still matters. Marketing Management, 12(4), 28–33. Cooper, R. B., & Zmud, R. M. (1990). Information technology implementation research: A technological diffusion approach. Management Science, 36(2), 123–139. Crawford, J. L., & Haaland, G. A. (1972). Predecisional information-seeking and subsequent conformity in the social inﬂuence process. Journal of Personality and Social Psychology, 23(1), 112–119. Damanpour, F. (1987). The adoption of technological, administrative and ancillary innovations: Impact of organization factors. Journal of Management, 13(4), 675–688. Damanpour, F. (1991). Organizational innovation: A meta-analysis of effects of determinants and moderators. Academy of Management Journal, 17(7), 555–590. Damanpour, F., Szabat, K. A., & Evan, W. M. (1989). The relationship between types of innovation and organizational performance. Journal of Management Studies, 26(6), 587–601. Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340. Day, G. S. (1994). The capabilities of market-driven organizations. Journal of Marketing, 58(4), 37–52. DeGregori, T. R. (2005). Frontier technologies for emerging economies: The entrepreneur as science and technology champion. International Journal of Entrepreneurship and Innovation Management, 5(5), 508–514. Delbecq, A. L. (1974). Contextual variables affecting decision-making in program planning. Decision Sciences, 5(4), 726–742. de Pillis, E., Kernochan, R., Meilich, O., Prosser, E., & Whiting, V. (2008). Are managerial gender stereotypes universal? The case of Hawai’i. Cross Cultural Management, 15(1), 94–106. Dess, G. G., & Picken, J. C. (1999). Beyond productivity: How leading companies achieve superior performance by leveraging their human capital. New York, NY: AMACOM. Devaraj, S., & Kohli, R. (2000). Information technology payoff in the health-care industry: A longitudinal study. Journal of Management Information Systems, 16(4), 41–67. Dewar, R. D., & Dutton, J. E. (1986). The adoption of radical and incremental innovations: An experimental analysis. Management Science, 32(11), 1422–1433. Dierickx, I., & Cool, K. (1989). Asset stock accumulation and sustainable competitive advantage. Management Science, 35(6), 1504–1511. Dolinsky, L. R., Vollmann, T. E., & Maggard, M. J. (1990). Adjusting replenishment orders to reﬂect learning in a material requirements planning environment. Management Science, 36(12), 1532–1547. Drury, D. H., & Farhoomand, A. (1999). Innovation diffusion and implementation. International Journal of Innovation Management, 3(2), 133–157. Dyer, J. H., & Singh, H. (1998). The relational view: Cooperative strategy and sources of interorganizational competitive advantage. Academy of Management Review, 23(4), 660–679. Earley, P. C. (1989). Social loaﬁng and collectivism. Administrative Science Quarterly, 34(4), 565–581.

382

R. GLENN RICHEY ET AL.

Earley, P. C. (1993). East meets west meets mideast: Further explorations of collectivistic and individualistic work groups. Academy of Management Journal, 36(2), 319–348. Earley, P. C., & Gibson, C. B. (1998). Taking stock in our progress on individualism–collectivism: 100 years of solidarity and community. Journal of Management, 24(3), 265–304. Edmondson, A. C., Winslow, A. B., Bohmer, R. M. J., & Pisano, G. P. (2003). Learning how and learning what: Effects of tacit and codiﬁed knowledge on performance improvement following technology adoption. Decision Sciences, 34(2), 197–223. Feldman, D. C. (1984). The development and enforcement of group norms. Academy of Management Review, 9(1), 47–53. Ford, C. W., Nonis, S. A., & Hudson, G. I. (2005). A cross-cultural comparison of value systems and consumer ethics. Cross Cultural Management, 12(4), 36–51. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. Froehle, C. M. (2006). Service personnel, technology, and their interaction in inﬂuencing customer satisfaction. Decision Sciences, 37(1), 5–38. Gaertner, L., & Insko, C. A. (2000). Intergroup discrimination in the minimal group paradigm: Categorization, reciprocation, or fear? Journal of Personality and Social Psychology, 79(1), 77–94. Garvin, D. A., & March, A. (1997). A note on knowledge management. Harvard Business School Online Case Series, 1–20. Gefen, D., Geri, N., & Paravastu, N. (2007). Viva la diffe´rence: The cross-culture differences within us. International Journal of E-Collaboration, 3(3), 1–14. Globerman, S. (1975). Technological diffusion in the Canadian tool and die industry. Review of Economics and Statistics, 57(4), 428–434. Gopalakrishnan, S., & Damanpour, F. (2000). The impact of organizational context on innovation adoption in commercial banks. IEEE Transactions on Engineering Management, 47(1), 14–25. Gordon, M. E., Allen Slade, L., & Schmitt, N. (1986). The ‘Science of the Sophomore’ revisited: From conjecture to empiricism. Academy of Management Review, 11(1), 191–207. Gotwon, G. G., & DiTomaso, N. (1992). Predicting corporate performance from organization culture. Journal of Management Studies, 29(6), 783–798. Gove, S., Sirmon, D. G., & Hitt, M. A. (2003). Relative resource advantages: The effect of resources and resource management on organizational performance. Paper presented at the annual Strategic Management Society Conference, Baltimore. Greenberg, J. (1987). The college sophomore as guinea pig: Setting the record straight. Academy of Management Review, 12, 157–159. Grifﬁth, D. A., & Harvey, M. G. (2001). A resource perspective of global dynamic capabilities. Journal of International Business Studies, 32(3), 597–606. Grifﬁth, D. A., Hu, M. Y., & Ryans, J. K., Jr. (2000). Process standardization across intra- and inter-cultural relationships. Journal of International Business Studies, 31(2), 303–324. Gruber, W. H., & Niles, J. S. (1975). The science–technology–utilization relationship in management. Management Science, 21(8), 956–963. Gudykunst, W. B. (1983). Intercultural communication theory. Beverly Hills, CA: Sage. Hampden-Turner, C., & Trompenaars, F. (1998). Riding the waves of culture (2nd ed.). New York, NY: McGraw Hill. Helfat, C. E., & Peteraf, M. A. (2003). The dynamic resource-based view: Capability lifecycles. Strategic Management Journal, 24(10), 997–1010.

The Dynamics of Technological Readiness in Marketing Units

383

Henderson, R. M., & Clark, K. B. (1990). Architectural innovation: The reconﬁguration of existing product technologies and the failure of established ﬁrms. Administrative Science Quarterly, 35(1), 9–30. Hofstede, G. (1980). Culture’s consequences: International differences in work-related values. Beverly Hills, CA: Sage. Howell, J. M., & Higgins, C. A. (1990). Champions of technological innovation. Administrative Science Quarterly, 35(2), 317–341. Huddy, L. (2004). Contrasting theoretical approaches to inter-group relations. Political Psychology, 25(6), 947–967. Iansiti, M. (1998). Technology integration: Making critical choices in a dynamic world. Boston, MA: Harvard Business School Press. Janis, I. L., & Mann, L. (1977). Decision-making: A psychological analysis of conﬂict, choice, and commitment. New York, NY: Free Press. Jassawala, A. R., & Shashittal, H. C. (1998). Accelerating technology transfer: Thinking about organizational paranoia. Journal of Engineering and Technology Management, 15(1), 153–157. Johnson, D. W. (1973). Communication in conﬂict situations: A critical review of the research. International Journal of Group Tensions, 3(1), 46–67. Johnson, D. W., & Johnson, R. T. (1989). Cooperation and competition: Theory and research. Edina, MN: Interaction Book Co. Johnson, M. D., & Selnes, F. (2004). Customer portfolio management: Toward a dynamic theory of exchange relationships. Journal of Marketing, 68(2), 1–17. Jung, J., & Sung,, E. (2008). Consumer-based brand equity; Comparisons among Americans and South Koreans in the USA and South Koreans in Korea. Journal of Fashion Marketing and Management, 12(1), 24–36. Kiesler, C. A., & Kiesler, S. B. (1970). Conformity. Reading, MA: Addison-Wesley. Kiggundu, M. N. (1983). Task interdependence and job design: Test of a theory. Organizational Behavior and Human Performance, 31(2), 145–172. Kotabe, M., Martin, X., & Domoto, H. (2003). Gaining from vertical partnerships: Knowledge transfer, relationship duration and supplier performance improvement in the U.S. and Japanese auto industries. Strategic Management Journal, 24(4), 293–316. Kumar, A., & Leon Zhao, J. (1999). Dynamic routing and operational controls in workﬂow management systems. Management Science, 45(2), 253–272. Lam, S. S. K., Hui, C., & Law, K. S. (1999). Organizational citizenship behavior: Comparing perspectives of supervisors and subordinates across four international samples. Journal of Applied Psychology, 84(4), 594–601. Langfred, C. W. (2005). Autonomy and performance in teams: The multilevel moderating effect of task interdependence. Journal of Management, 31(4), 513–529. Lee, J., & Bose, U. (2002). Operational linkage between diverse dimensions of information technology investments and multifaceted aspects of a ﬁrm’s economic performance. Journal of Information Technology, 17(3), 119–131. Lee, T. K. (1985). On the joint decisions of R and D and technology adoption. Management Science, 31(8), 959–969. Leonidou, L. C., Katsikeas, C. S., & Hadjimarcou, J. (2002). Executive insights: Building successful export business relationships: A behavioral perspective. Journal of International Marketing, 10(3), 96–115.

384

R. GLENN RICHEY ET AL.

Leung, K., Au, Y.-F., Fernadez-Dols, J. M., & Iwawaki, S. (1992). Preferences for methods of conﬂict processing in two collectivistic cultures. International Journal of Psychology, 27(2), 195–209. Levi, M. D., & Nault, B. R. (2004). Converting technology to mitigate environmental damage. Management Science, 50(8), 1015–1030. Levitt, R. E., Thomsen, J., Christiansen, T. R., & Kunz, J. C. (1999). Simulating project work processes and organizations: Toward a micro-contingency theory of organizational design. Management Science, 45(11), 1479–1495. Liberatore, M. J., & Breem, D. (1997). Adoption and implementation of digital-imaging technology in the banking and insurance industries. IEEE Transactions on Engineering Management, 44(4), 367–377. Lippman, S. A., & McCardle, K. F. (1991). Uncertain search: A model of search among technologies of uncertain values. Management Science, 37(11), 1474–1490. Locke, E. A., Mento, A. J., & Katcher, B. L. (1978). The interaction of ability and motivation in performance: An exploration of the meaning of moderators. Personnel Psychology, 31(2), 269–291. Martz, W. M. B., Jr., Biscaccianti, A., Neil, T. C., & Williams, R. J. (2005). A multi culture perception of the entrepreneurial lifestyle. Journal of Enterprising Culture, 13(4), 359–370. Mata, F. J., Fuerst, W. L., & Barney, J. B. (1995). Information technology and sustained competitive advantage: A resource-based analysis. MIS Quarterly, 19(4), 487–505. Moorman, R. H., & Blakely, G. L. (1995). Individualism–collectivism as an individual difference predictor of organizational citizenship behavior. Journal of Organizational Behavior, 16(2), 127–142. Morash, E. A., & Lynch, D. F. (2002). Public policy and global supply chain capabilities and performance: A resource-based view. Journal of International Marketing, 10(1), 25–51. Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome. Lexington, MA: Lexington Books. O¨zsomer, A., & Genctu¨rk, E. (2003). A resource-based model of market learning in the subsidiary: The capabilities of exploration and exploitation. Journal of International Marketing, 11(3), 1–29. Parasuraman, A. (2000). Technology Readiness Index (TRI). A multiple-item scale to measure readiness to embrace new technologies. Journal of Service Research, 2(4), 307–320. Parasuraman, A., & Colby, C. S. (2001). Techno-ready marketing. New York, NY: The Free Press. Pearce, J. L., & Gregersen, H. B. (1991). Task interdependence and extra-role behavior: A test of the mediating effects of felt responsibility. Journal of Applied Psychology, 76(6), 838–844. Penrose, E. T. (1955). Limits to the growth and size of ﬁrms. The American Economic Review, 45(2), 531–543. Penrose, E. T. (1959). The theory of the growth of the ﬁrm. New York, NY: Wiley. Peteraf, M. A. (1993). The cornerstones of competitive advantage: A resource-based view. Strategic Management Journal, 14(3), 179–191. Pﬂugheoft, K. A., Ramamurthy, K., Sooﬁ, E. S., Yasai-Ardekani, M., & Zahedi, F. (2003). Multiple conceptualizations of small business web use and beneﬁt. Decision Sciences, 34(3), 467–512. Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Harvard Business Review, 68(3), 79–91.

The Dynamics of Technological Readiness in Marketing Units

385

Priem, R. L., & Butler, J. E. (2001a). Is the resource-based ‘View’ a useful perspective for strategic management research? The Academy of Management Review, 26(1), 22–40. Priem, R. L., & Butler, J. E. (2001b). Tautology in the resource-based view and the implications of externally determined resource value: Further comments. The Academy of Management Review, 26(1), 57–66. Reid, J., & Hubbell, V. (2005). Creating a performance culture. Ivey Business Journal, 69(4), 1–4. Richey, R. G., & Bachrach, D. G. (2004). A pragmatic approach to quality improvement. Supply Chain Management Review, 8(3), 24–31. Richey, R. G., Daugherty, P. J., & Roath, A. S. (2007). Technological readiness and complementarity: Dynamic capabilities impacting logistics service competency and performance. Journal of Business Logistics, 28(2), 10–21. Ross, J. W. (1998). The ERP revolution: Surviving versus thriving. Sloan School of Management, MIT: Center for Information Systems Research. Ross, R. H., Broyles, S. A., & Leingpibul, T. (2008). Alternative measures of satisfaction in cross-cultural settings. The Journal of Product and Brand Management, 17(2), 82–95. Sanayei, A., & Jamshidian, M. (2004). Information technology and software export. Journal of International Marketing and Marketing Research, 29(1), 13–20. Sawhney, M., & Prandelli, E. (2000). Communities of creation: Managing distributed innovation in turbulent markets. California Management Review, 42(4), 24–54. Schultze, W. S. (1994). The two schools of thought in resource based theory: Deﬁnitions and implications for research. In: P. Srivastava, A. S. Huff & J. E. Dutton (Eds), Advances in strategic management (p. 10A). Greenwich, CT: JAI Press. Shaw, M. E. (1981). Group dynamics: The psychology of small group behavior. New York, NY: Harper. Shea, G. P., & Guzzo, R. A. (1989). Groups as human resources. In: G. R. Ferris & K. M. Rowlands (Eds), Research in personnel and human resources management (Vol. 5, pp. 323–356). Greenwich, CT: JAI Press. Shiloh, S., Gu¨venc- , G., & O¨nkal, D. (2007). Cognitive and emotional representations of terror attacks: A cross-cultural exploration. Risk Analysis, 27(2), 397–408. Song, M., & Xie, J. (2000). Does innovativeness moderate the relationship between crossfunctional integration and product performance? Journal of International Marketing, 8(4), 61–89. Srivastava, R. K., Shervani, T. A., & Fahey, L. (1998). Market-based assets and shareholder value: A framework for analysis. Journal of Marketing, 61(1), 2–17. Tajfel, H., & Turner, J. (1986). The social identity theory of intergroup behavior. In: S. Worchel & W. G. Austin (Eds), Psychology of intergroup relations. Chicago, IL: Nelson-Hall. Thomas, E. J. (1957). Effects of facilitative role interdependence on group functioning. Human Relations, 10(1), 347–366. Thompson, J. D. (1967). Organizations in action. New York, NY: McGraw Hill. Tornatsky, L., & Klein, K. (1982). Innovation characteristics and adoption–implementation: A meta-analysis of ﬁndings. IEEE Transactions of Engineering Management, 29(1), 28–45. Triandis, H. C. (1988). Collectivism v. individualism: A reconceptualization of a basic concept cross-cultural social psychology. In: G. K. Vermam & C. Bagley (Eds), Cross-cultural studies of personality, attitudes, and cognition. New York: St. Martin’s Press. Triandis, H. C. (1989). Self and social behavior in differing cultural contexts. Psychological Review, 96(1), 269–289.

386

R. GLENN RICHEY ET AL.

Tse, D. K., Lee, K.-h., Vertinsky, I., & Wehrung, D. A. (1988). Does culture matter? A crosscultural study of executives’ choice, decisiveness, and risk adjustment in international marketing. Journal of Marketing, 52(3), 81–95. Van de Ven, A. H., & Ferry, D. L. (1980). Measuring and assessing organizations. New York, NY: Wiley. Van der Vegt, G. S., Emans, B. J. M., & Van de Vliert, E. (1998). Motivating effects of task and outcome interdependence in work teams. Group and Organization Management, 23(2), 124–143. Van der Vegt, G. S., Emans, B. J. M., & Van de Vliert, E. (2001). Patterns of interdependence in work teams: A two-level investigation of the relations with job and team satisfaction. Personnel Psychology, 54(1), 51–69. Van der Vegt, G. S., & Janssen, O. (2003). Joint impact of interdependence and group diversity on innovation. Journal of Management, 29(5), 729–751. Van der Vegt, G. S., Van de Vliert, E., & Oosterhof, A. (2003). Information dissimilarity and organizational citizenship behavior: The role of intra-team interdependence and team identiﬁcation. Academy of Management Journal, 46(6), 715–727. Wageman, R. (1995). Interdependence and group effectiveness. Administrative Science Quarterly, 40(1), 145–180. Wagner, J. A., III, & Moch, M. K. (1986). Individualism–collectivism: Concept and measure. Group and Organization Studies, 11(3), 280–304. Wagner, J. A., III, VanDyne, L., LePine, J. A., & Hollenbeck, J. R. (1997). Studies of individualism-collectivism: Effects on team member performance on solitary and conjunctive tasks. Unpublished paper. Wernerfelt, B. (1984). A resource-based view of the ﬁrm. Strategic Management Journal, 5(2), 171–180. Widgery, R., Tubbs, S. L., & Nicholson, D. (2004). Using the technological readiness audit for strategic competitive advantage. Journal of American Academy of Business, 4(1/2), 23–27. Wright, A. (2002). Technology as an enabler of the global branding of retail ﬁnancial services. Journal of International Marketing, 10(2), 83–98. Wright, N. S., & Drewery, G. P. (2006). Forming cohesion in culturally heterogeneous teams: Differences in Japanese, Paciﬁc Islander and Anglo experiences. Cross Cultural Management, 13(1), 43–54. Yamaguchi, S. (1994). Collectivism among the Japanese: A perspective from the self. In: U. Kim, H. C. Triandis, C. Kagitcibasi, C. Choi & G. Yoon (Eds), Individualism and collectivism: Theory, method, and application. Newbury Park, CA: Sage.

APPENDIX. INTERDEPENDENCE AND PERFORMANCE MANIPULATIONS (SEE BACHRACH ET AL., 2006) High Interdependence A production unit in a small factory produces several cellular phone models for Nokia. The 10 members of the production unit rotate responsibility for

The Dynamics of Technological Readiness in Marketing Units

387

each of 10 sequential operations. Factory policy allows no buffer stocks of work-in-process to be maintained between operations. Although the unit’s members are required to work 8 h each day, they must work together to determine work schedule because of the buffer-stock policy. Workers at this factory have to communicate information regarding color, stock, and equipment to each other because the phone that they make leaves the factory in ﬁve colors and cabinet styles with 10 functionality packages. Because half of the workers at this factory share equipment, they have to coordinate their use of production facilities. Over the past several years, both the design of the product and the production process has changed extensively.

Low Interdependence A production unit in a small factory produces a single cellular phone model for Nokia. Each of the 10 members of the production unit is responsible for one of 10 sequential operations. Factory policy allows large buffer stocks of work-in-process to be maintained between operations. Although all 10 individuals are required to work 8 h each day, they can begin and end their activities according to a ﬂexible work schedule because of the buffer-stock policy. Workers at this factory do not have to communicate information regarding color, stock, or equipment to each other because the phone that they make leaves the factory in only one color, cabinet style, and functionality package. All of the workers at this factory have their own equipment, and so do not have to coordinate their use of production facilities. Over the last several years, neither the product design nor the production process has changed.

High Unit-Level Performance Last year, this production unit was the most productive unit in its region, producing at a rate 25% higher than the next closest factory. Of the 15 other factories producing similar products, this unit had the lowest materials cost percentage (waste), highest operating efﬁciency, fewest customer complaints, highest customer satisfaction, and fewest returns (added to the end of the interdependence manipulation).

388

R. GLENN RICHEY ET AL.

Low Unit-Level Performance Last year, this production unit was the least productive unit in its region, producing at a rate 25% lower than the next closest factory. Of the 15 other factories producing similar products, this unit had the highest materials cost percentage (waste), lowest operating efﬁciency, most customer complaints, lowest customer satisfaction, and most returns (added to the end of the interdependence manipulation).

METHODOLOGICAL RIGOR OF ORDINARY LEAST SQUARES REGRESSION STUDIES IN MAINSTREAM MARKETING: AN ASSESSMENT OF RESEARCH REPORTED IN THE JOURNAL OF MARKETING Z. Seyda Deligonul, Brian R. Chabowski, Steven H. Seggie, Shichun Xu and S. Tamer Cavusgil Marketing, as a ﬁeld of inquiry and practice, has gained increasing importance since the beginning of the 20th Century (Bartels, 1951) and the Journal of Marketing (JM) has been the most consistently inﬂuential marketing journal in the ﬁeld. Since its inception in 1936, JM has progressed toward publishing more scholarly works (Kerin, 1996), as the ﬁeld matured in its scientiﬁc approach (e.g., Bartels, 1951; Hunt, 1976, 2002; Sheth, Gardner, & Garrett, 1988; Sheth & Sisodia, 1999; Taylor, 1965). This emphasis on marketing as a science ensures that rigor in marketing studies remains central for the entire scholarly community from contributors and gatekeepers to journal editors and scholars. But, if marketing knowledge has New Challenges to International Marketing Advances in International Marketing, Volume 20, 389–418 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1474-7979/doi:10.1108/S1474-7979(2009)0000020017

389

390

Z. SEYDA DELIGONUL ET AL.

advanced, how well is this progress reﬂected in research methodology? How competent have marketing scholars been in their use of methodological tools? In the past, the issue of methodological rigor has been investigated, in part, by capturing a snapshot of rigor in the application of a single tool (McCloskey & Ziliak, 1996; Ziliak & McCloskey, 2004). Others have offered critiques on the use of speciﬁc techniques. Examples of the latter include structural equation modeling (SEM) (Jarvis, MacKenzie, & Podsakoff, 2003; Shook, Ketchen, Hult, & Kacmar, 2004) and cluster analysis (Ketchen & Shook, 1996). Such efforts are indispensable if the ﬁeld is to enhance its overall value. Knowledge consolidation may be inhibited if the results are based on poor use of investigative means. This study seeks to explore the status of methodological rigor in marketing studies. In an effort to investigate rigor in the use of a very common tool employed by marketing researchers, we analyzed all studies that were published in JM between 1990 and 2005 using Ordinary Least Squares (OLS) regression. The choice of JM is driven by its generally accepted position as the premium mainstream marketing journal (Hult, Neese, & Bashaw, 1997; Mort, McColl-Kennedy, Kiel, & Soutar, 2004; Theoharakis & Hirst, 2002). As such, we would expect manuscripts published in JM to be examples of best practices in their use of methodology. The choice of OLS is driven by its relative prominence as a trackable application in the most widely recognized, mainstream outlet in the ﬁeld of marketing. In addition, OLS has been a nucleus for developing other fundamental techniques. Between 1990 and 2005, there were 147 OLS regression equations in 83 separate studies published in the JM. Over the same time period, 53 articles used SEM, and 23 studies employed logistic regression. With its frequent and consistent application, OLS provides us with the opportunity to explore a number of epistemological issues about its application by marketing scholars. This study intends to make several contributions. First, we aim to address the issue of rigor as a disciplinary ideal. Speciﬁcally, we investigate the strengths and weaknesses in the preparation, analysis, and reporting of empirical research in the use of OLS regressions. We focus on variations and deﬁciencies that may have escaped the attention of gatekeepers in the review process. The objective here is not to critique any individual study but to be able to comment on the degree to which marketing scholars, collectively, demonstrate ‘‘proper’’ application of a fundamental research technique. Second, we examine emergent patterns in the application of this research tool. Included in this analysis is the issue of whether the application of OLS

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

391

has improved over time. Third, we attempt to characterize the nature of consensus over the use of OLS. This is important because the extent to which consensus prevails with regard to conduct – agreement on the means of scholarship – signals the degree of maturity of a ﬁeld. Fourth, with this study, we aim to offer insights for gatekeepers and editors of marketing journals to ensure a high level of rigor in studies that employ OLS. Finally, this study can serve as reference point if investigations of this nature are to be conducted periodically over time. The remainder of this chapter is organized as follows. We begin by discussing the analytical method we employed to investigate each OLS study. Next, we present the results with a focus on general reporting, sample size, testing rigor, and overall interpretation. We then offer post-hoc analyses and offer selected prescriptions.

DESCRIPTION OF THE METHOD Scope of the Study We searched JM from 1990 to 2005 to identify all studies that employed OLS regression.1 The search resulted in 83 articles with 147 OLS regressions. Many authors speciﬁcally state that they used OLS and these were promptly included in the sample. Other authors acknowledged regression as a methodological procedure without explicitly specifying an estimation technique. To inquire whether OLS was used, we communicated with all of these authors. In all, 51 authors were contacted, with 44 responding. Of the 44 that responded 43 had used OLS; the only exception was subsequently excluded from the analysis. The remaining seven articles were checked once more, and consensus was reached that OLS had most likely been used. Therefore, they were also included in the study.

Assessment Format The overall analytical approach follows similar analyses of methodological rigor in marketing and related ﬁelds (Jarvis et al., 2003; McCloskey & Ziliak, 1996). Some assessment criteria were adapted to address issues speciﬁc to OLS applications, while additional criteria were developed also to accomplish this task. These assessment criteria were then presented to four separate methodological experts in OLS. These experts offered various

392

Z. SEYDA DELIGONUL ET AL.

reﬁnements before a ﬁnal set of 24 criteria were developed. These are presented in Table 1. The assessment criteria encompass issues from the application of the technique to producing useful results with full power. Brieﬂy, the investigation explored: the type of the regression (Criteria 1, or C1); distributional assumption (C10); problems of ﬁt (C6, C9); sample and sample size (C7); and testing, power, and reporting (C3, C4, C8). Additional criteria were directed to issues of: multicollinearity (C11); relevance and practicality (C20); needs for standardization and units of measurement (C2, C21–C24); and moderation analysis (C13–C18). Some criteria addressed objective issues (C1, C5, C7–C10, C12), while others required subjective assessment.

Description of the Assessment Process Three marketing scholars with an extensive knowledge of the OLS technique served as raters. Each rater ﬁrst read each article to gain a comprehensive understanding of the research issues addressed in the study. Then, the researchers independently coded each article against the assessment criteria. Next, all three met to compare and ﬁnalize the decisions on each article. A fourth rater was brought in to resolve any disagreements. The inter-rater agreement was calculated to be 92%. This is comparable to similar studies (Ketchen & Shook, 1996). Finally, statistical analyses were performed to delineate the signiﬁcance of and patterns in the ﬁndings. Table 2 reveals the percent of regression analyses that meet each of the 24 assessment criteria.

Unit of Analysis and Coding As there were often more than one OLS regression used in each of the 83 JM articles (and each may warrant different judgments to assessment criteria), each of the 147 regression analyses is coded separately. As an example, Kotabe (1990) used more than one regression. However, only one of them included interaction effects, which renders C17 not applicable to other regressions in the same article. Therefore, this regression was coded separately from other regressions in that study. The complete set of 83 studies included in the analysis is provided in Table 3.

Criterion 8: For null hypotheses not rejected, do the authors state the power of the test?

Criterion 7: If the sample size is large, do authors acknowledge that signiﬁcance may be an artifact of the large sample size?

Criterion 6: Based on a combination of the number of independent variables and sample size, are readers assured that the authors are not guilty of overﬁtting?

Criterion 4: Does the paper report all ‘‘t’’ statistics, standard errors, or p-values? Criterion 5: In a multiple regression, did the authors report adjusted R2?

Criterion 2: Are the ﬁrst and second moments for all regression variables included? Criterion 3: Does the paper report ‘‘f’’ statistics or p-values for the general linear relation?

Remarks

Marketing scholars have reported R2 and/or adjusted R2. Traditionally R2 was reported. However, we know that, as a measure, it is sensitive to the number of independent variables. R2 increases as independent variables are added even if the variable is not meaningful (Hair et al., 1998; Kutner et al., 2004). Therefore, researchers have proposed the use of adjusted R2 as it adjusts for the number of independent variables. This is another issue of sample size. One of the problems with a small sample is the danger of overﬁtting (Hair et al., 1998). Based on previous research, ten data points per independent variable is sufﬁcient for us to be conﬁdent that there is no overﬁtting. While a large sample size is an advantage for statistical power considerations, very large samples are overly sensitive to tests of statistical signiﬁcance may be over sensitive. For very large sample sizes, we suggest more than 1,000 observations (Hair et al., 1998). A very large sample may lead to a test ﬁnding all relationships to be statistically signiﬁcant. Generally, the null hypothesis in marketing research is that of no relationship, with the stated hypothesis being the alternate. It may be the case that in instances where the null hypothesis is not rejected, the regression model may not have had sufﬁcient statistical power.

Clearly there is more than one type of regression and we felt that in the interest of full reporting, marketing scholars should clearly state which type of regression they use. If we are using OLS to explain marketing phenomena then the reporting of means and standard deviations is essential. It is not enough for us to provide information on the individual parameters without also providing fundamental information regarding the signiﬁcance of the overall model. It is a standard expectation for at least one of these to be reported.

Assessment Criteria Used in Study.

Criterion 1: Do the authors state what type of regression they used?

Criteria

Table 1. Methodological Rigor of OLS Regression Studies in Mainstream Marketing 393

Criterion 13: For articles using split sample to test moderation effects, are the authors clear on the criterion used to split the sample?

Criterion 12: If the data covers a span of time, have the authors dealt with longitudinal considerations?

Criterion 11: If theoretically supported hypotheses are not supported statistically, have tests been conducted for multicollinearity?

Criterion 9: Do the authors provide evidence of the model’s goodness of ﬁt? If not, do they at least indicate that they performed such tests and were satisﬁed with the results? Criterion 10: If using a small sample and relying on t-, F-, or w2 tests, have the authors established that the residuals are normally distributed, or at least close to normally distributed?

Criteria

Remarks There are various goodness-of-ﬁt tests that can and should be employed by the researcher. These can include (but are not limited to) tests of linearity, homoscedasticity, and residual diagnostics. One of the assumptions of OLS is that the residuals are normally distributed (Kennedy, 2003; Kutner et al., 2004). If we have a large enough sample this is not an issue. However, with a small sample, there are some concerns. The deﬁnition of a small sample size is arbitrary in many respects and there is disagreement among researchers. It has even been stated that less than 20 data observations per independent variable constitutes a small sample (Hair et al., 1998). However, this is quite constricting. Thus, we deﬁned a small sample size as a sample with less than ﬁve data points per independent variable. An assumption of OLS is that there is no multicollinearity in the model. High levels of multicollinearity lead to estimation problems including parameter estimates with the incorrect sign among other issues (Kennedy, 2003; Kutner et al., 2004). One reason for theoretically supported hypotheses lacking support may be the existence of multicollinearity. One point to note here is that a correlation matrix does not show the existence (or lack) of multicollinearity. It is able to show the correlation between two variables but cannot detect the situation where three variables are collinear with no two variables exhibiting signs of collinearity (Kennedy, 2003). Thus, we require the use of variance inﬂation factors or some other test beyond the correlation matrix to detect the existence of multicollinearity (Kutner et al., 2004). Included in this are (among other factors) adjustments for inﬂation where necessary, testing residuals for autocorrelation, and adjusting for population size. Authors must establish how the subgroups in the moderation analysis are determined. Providing enough information and justiﬁcation on the cutoff value (median, mean, etc.) has both theoretical and practical implications.

Table 1. (Continued ) 394 Z. SEYDA DELIGONUL ET AL.

Criterion 19: Do the authors either refrain from making assertions about main effects when they have a signiﬁcant interaction or provide additional evidence of main effects?

Criterion 18: When moderation analyses are conducted on categorical interaction variables, do the authors establish that there are equal error variances across subgroups?

Criterion 17: When moderation analyses are conducted on categorical interaction variables, do the authors establish that there are equal sample sizes across levels?

Criterion 16: For articles using interaction to test moderation effects, do the authors take into consideration any possible reduced statistical power?

Criterion 15: For articles using split sample to test moderation effects and to claim moderating effects, do the authors conduct a formal Chow test or tests of similar effect?

Criterion 14: For regressions using split sample by median to test moderation effects, do the authors establish that the groups are statistically different on means of the moderating variable?

When two subgroups in the moderation analysis are split at the median, the means of the moderating variables may not signiﬁcantly differ across the two groups due to the distribution properties of the pooled group. As a consequence, the moderation analysis result may not be reliable. To establish the existence of the moderating effect, it is essential to prove that the slopes of the regressions are statistically different across different subgroups. Stating a moderation effect based on the presence of different coefﬁcients across two groups is not sufﬁcient. A formal test such as Chow test is required. When moderated multiple regression suffers from the limitation of low statistical power Type II error rates can increase. This leads researchers to conclude a non-existent interaction effects when there actually is such an interaction effect. Therefore, it is important to take into consideration the power of the statistical test when discussing ﬁndings of this kind. In moderated multiple regression analysis, an unequal subgroup sample size will be inadequate to detect some moderating effects. Typically, this limitation cannot be overcome by using a large sample (StoneRomero, Alliger, & Aguinis, 1994; Aguinis & Stone-Romero, 1997). When the assumption of equal error variances across groups is violated, it has been shown that tests of the differences in the regression slopes cannot be clearly interpreted. As a result, the F test could be either too conservative or liberal (Stone & Hollenbeck, 1989). Researchers often avoid the use of multiplicative interactive terms due to problems of interpretation (Friedrich, 1982). In fact, those that use multiplicative interactions must be careful in interpretation (Friedrich, 1982; Kutner et al., 2004). When there is a signiﬁcant interaction term, the meaning of the main effect is no longer the same as if no interaction effect was detected. The main effect is now dependent on the level of the other, multiplicative interaction predictor variable(s) in the interaction (Kutner et al., 2004).

Methodological Rigor of OLS Regression Studies in Mainstream Marketing 395

Criterion 24: Do the authors avoid making assertions of practical signiﬁcance based on coefﬁcients that are not statistically signiﬁcant?

Criterion 23: Are the authors clear on and able to express what constitutes practical – rather than statistical – signiﬁcance?

Criterion 22: If standardized coefﬁcients are reported, do the authors interpret them within the model?

Criterion 21: Do the authors report standardized coefﬁcients within the model when necessary?

Criterion 20: Are the coefﬁcients of the model managerially interpretable?

Criteria

Remarks Claims have been made that much marketing research has sacriﬁced relevance at the expense of rigor (Sheth et al., 1988). Therefore, it is incumbent upon researchers to use coefﬁcients that can be interpreted by managers in a meaningful way. We have argued here that treating non-metric scales such as a Likert scale in the same fashion as metric scales is not a valid practice as the interpretation of the coefﬁcients based on non-metric scales are not easily interpretable. Some marketing scholars report unstandardized coefﬁcients, others report standardized coefﬁcients, and still more report both. If coefﬁcients are in different units, then it is essential to use standardized coefﬁcients to enable comparison. Standardized coefﬁcients allow for comparison of coefﬁcients within a model. Therefore, one would expect reporting standardized coefﬁcients would interpret them within the context of the overall model. The argument has been made that a statistically signiﬁcant coefﬁcient is not necessarily practically signiﬁcant. In interpreting results, author should bear in mind that a statistically signiﬁcant coefﬁcient only indicates the inﬂuence on the dependent by one unit increase in the independent variable as different from zero. This does not necessarily mean that the effect is practically signiﬁcant. Even though statistical signiﬁcance does not necessarily mean practical signiﬁcance, we cannot infer that insigniﬁcant coefﬁcients imply practical signiﬁcance.

Table 1. (Continued )

396 Z. SEYDA DELIGONUL ET AL.

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

397

Table 2. Results of Consensus Evaluations across Assessment Criteria. Assessment Criteria Do the Authors:

C1. Specify what type of regression was used? C2. Report the ﬁrst and second moments for all regression variables included? C3. Report ‘‘F ’’ statistics or p-values for the general linear relation? C4. Report all ‘‘t’’ statistics, standard errors or p-values? C5. Report adjusted R2 in multiple regressions? C6. Use enough sample size to avoid overﬁtting problem? C7. Acknowledge possibility of sample size artifact when n is greater than 1,000? C8. State the power of the test? C9. Provide evidence of the model’s goodness-of-ﬁt? C10. Establish residual normality when sample size is small? C11. Conduct tests or control for the problem of multicollinearity? C12. Properly address longitudinal considerations? C13. Clarify splitting criteria when using split sample to test moderation? C14. Establish signiﬁcant difference of means on moderating variable when split sample by median? C15. Conduct a Chow test to establish the difference of slope of regression? C16. Take into consideration possible reduced statistical power when using interaction terms to test moderation? C17. When using interaction terms involving categorical variables, establish the equality of sample size across levels? C18. When using interaction terms involving categorical variables, establish the equality of error variances across levels? C19. Refrain from making assertions about main effects when they have a signiﬁcant interaction effect or provide additional evidence of the main effects?

Total Number of Articles for which the Question is Applicable

Percent of Articles Meeting the Criterion

83 83

37.3 49.4

83

62.7

83

97.6

81

53.1

82

87.8

12

0

46 83

10.9 7.2

2

0

45

51.1

12

66.7

7

100.0

6

33.3

7

28.6

26

15.4

12

8.3

12

0

31

41.9

398

Z. SEYDA DELIGONUL ET AL.

Table 2. (Continued ) Assessment Criteria Do the Authors:

C20. Use coefﬁcients that are managerially interpretable? C21. Report standardized coefﬁcients within the model when necessary? C22. Interpret the standardized coefﬁcients within the model? C23. Clarify and express what constitutes practical (rather than statistical) signiﬁcance? C24. Avoid asserting practical signiﬁcance when coefﬁcients are statistically insigniﬁcant?

Total Number of Articles for which the Question is Applicable

Percent of Articles Meeting the Criterion

82

19.5

21

61.9

42

28.6

81

28.4

79

100.0

KEY ISSUES IN THE APPLICATION OF OLS REGRESSION On the basis of previous assessments in the literature, we focus on ﬁve key issues in the use of OLS regression (Mason & Perreault, 1991; McCloskey & Ziliak, 1996): (1) unambiguous reporting of model ﬁt and hypothesis testing; (2) sample size and power; (3) use of diagnostic tests; (4) moderation analysis; and (5) interpretation of results. All require analysis of both subjective and objective criteria.

Unambiguous Reporting of Model Fit and Hypothesis Testing Certain minimum requirements exist with regard to reporting that are essential in regression studies. We contend that these minimum requirements include: estimation technique, descriptive statistics, model goodnessof-ﬁt (via F-statistic and adjusted R2), and individual signiﬁcance statistics for coefﬁcients. Only 37% of the articles reveal the estimation procedure used (C1). Though the JM authors state that their study was based on regression, precise speciﬁcation of the estimation method was neglected in two-thirds of the analyses. JM authors chose not to be generous in the provision of descriptive statistics, with only 49% of articles providing basic descriptive

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

Table 3. List of Journal of Marketing Articles Studied. 1990 Aaker and Keller Noordewier, John, and Nevin Kotabe Narver and Slater 1991 Walters Takada and Jain Bolton and Drew Tanner, Hunt, and Eppright 1992 Goolsby and Hunt Oliva, Oliver, and MacMillan 1993 Bucklin and Sengupta Jaworski, Stathakopoulos, and Krishnan Jaworski and Kohli Helsen, Jedidi, and DeSarbo Moorman, Deshpande, and Zaltman Cravens, Ingram, LaForge, and Young 1994 Heide Ganesan Slater and Narver Kohli and Jaworski 1995 Venkatesh, Kohli, and Zaltman Gundlach, Achrol, and Mentzer Anderson and Robertson Robertson, Eliashberg, and Rymon Agrawal and Kamakura Bhattacharya, Rao, and Glynn 1996 Challagalla and Shervani Polly, Siddarth, Siegel, Haddix, Merritt, Giovino, and Eriksen Narasimhan, Neslin, and Sen Zeithaml, Berry, and Parasuraman Urbany, Kalapurakal, and Dickson Ramaswami Frazier and Lassar

1997 Song and Parry Eliashberg and Shugan Smith and Cooper-Martin 1998 Mittal, Ross, and Baldasare Murry and Heide Henderson and Cote Sparks and Hunt Moorman and Miner Money, Gilly, and Graham 1999 Homburg, Workman, and Krohmer Capron and Hulland Kraft Weiss, Anderson, and MacInnis 2000 Reinartz and Kumar Sethi Sivadas and Dwyer Matsuno and Mentzer Lehmann and Weinberg Brown, Dev, and Lee 2001 Wathne, Biong, and Heide Sarin and Mahajan Antia and Frazier Hewett and Bearden 2002 Geyskens, Gielens, and Dekimpe Atuahene-Gima and Li Balasubramanian and Cole Joseph and Richardson Goldenberg, Libai, and Muller McAlexander, Schouten, and Koening Srinivasan, Lilien, and Rangaswamy 2003 Basuroy, Chatterjee, and Ravid Ailawadi, Lehmann, and Neslin Bendapudi and Leone Heide

399

400

Z. SEYDA DELIGONUL ET AL.

Table 3. (Continued ) Maxham and Netemeyer Sprott, Manning, and Miyazaki Vorhies and Morgan

Atuahene-Gima and Murray Joshi and Sharma Rao, Agarwal, and Dahlhoff

2004 Inman, Shankar, and Ferraro Wathne and Heide Johnson, Sohi, and Grewal Klein, Smith, and John Lichtenstein, Drumright, and Braig

2005 Fox and Hoch Vorhies and Morgan Klein and Ahluwalia Wernerfelt Kirca, Jayachandran, and Bearden

statistics (C2). Space limitations may have impacted the exclusion of ﬁrst and second moments (mean and standard deviation) for variables included in regression equations. However, inclusion of these key indicators is not only good practice for intuitive assessment of models employing OLS, but it is also important in its replication – an important characteristic in scientiﬁc inquiry. Goodness-of-ﬁt in the form of the F-statistic was reported in 63% of the articles (C3). However, only 53% report adjusted R2 values (C5). As compared to our ﬁndings concerning goodness-of-ﬁt measures, JM authors seem to pay greater attention to statistics related to hypothesis testing. Almost all articles (98%) report t-statistics, p-values, or standard errors, thus allowing for interpretation of statistical signiﬁcance (C4). This evidence may lead to the conclusion that researchers were more focused on individual hypotheses than on overall models. This may explain the greater frequency in reporting of t-statistics as compared to the reporting of F-statistics. This is most likely not a reﬂection of hypotheses becoming the centerpiece of empirical research in marketing, but rather a focus to establish the relevance of hypotheses. Scholars should be aware that the basis of any hypothesis testing has to be the existence of a strong model reﬂecting the underlying theory. Without a strong model, only a set of conjectures are stated in a bundle of hypotheses.

Sample Size and Power Two principal issues were addressed with respect to samples: the actual size of the sample and the power of the test. Sample size is one of the most important elements within a researcher’s control (Hair, Anderson, Tatham, & Black, 1998). Indeed, ‘‘too large’’ or ‘‘too small’’ of a sample constitute

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

401

potential problems for the researcher. On the one hand, researchers aim to avoid too large of a sample in their statistical testing so that statistical signiﬁcance does not become an automatic outcome of resultant variance reduction. On the other hand, scholars attempt to work with a sizable sample in order not to degrade the estimation and testing methods while they improve the model’s overall statistical power.2 Our analysis reveals that most JM authors (88%) protected their study from overﬁtting by using a sample size at least ten times as large as the number of independent variables (C6). Only two featured sample sizes that are ‘‘too small’’ (less than 5 data points per independent variable). At the other extreme, 12 articles had very large samples (greater than 1,000). Neither of the two articles with a small sample expressed concern about the quality of results, and none of the 12 OLS articles with a very large sample acknowledged that their ﬁndings might be an artifact of the sample size. Of the articles with null hypotheses not rejected, only 11% mentioned the power of the test (C8). To draw attention to the problem, Wathne, Biong, and Heide (2001) caution readers that, ‘‘Although the relative importance measures are unaffected by sample size, careful attention should be paid to the various signiﬁcance tests, because a lack of hypothesis support may be attributable to a lack of statistical power’’ (p. 62). Without pointing out the potential inﬂuence of statistical power in results section of the model’s hypothesis testing, readers may reach an erroneous conclusion with respect to the practical signiﬁcance of the ﬁndings. There is no easy resolution to the sample size selection and power issue. The issue is a consideration of a trade-off and not a problem of using a ‘‘small’’ or ‘‘large’’ sample size per se. Such a balance needs to be addressed either by interpretation of the results or by adding more statistical tests to demonstrate the robustness of the ﬁndings. For a small sample size, this means that the authors must be cautious about the residuals of the regression and the automatic assumption of distributional characteristics such as normality. In addition, they should acknowledge that they might be overﬁtting the data. More data points imply more information. The advantage of a large sample should not blind authors. In the reﬂexive penchant for abundance, the investigation must also recognize the resultant defects. A good study is cognizant of ill consequences from every favorable element so that the resultant view will enable readers to judge each trade-off as an imperfect decision choice. The issue of not specifying the power of the test is noteworthy. Lack of power may limit the ability of tests to ﬁnd statistical signiﬁcance and may

402

Z. SEYDA DELIGONUL ET AL.

explain why theoretically supported alternate hypotheses are not ‘‘accepted.’’ The fact that only 11% of articles with unsupported hypotheses discuss the topic of statistical power suggests that the authors may not have tested for power. If, in fact, those authors did calculate statistical power and overlooked the issues related to power, then the possibility that power-loss could be a potential reason as to why signiﬁcance was not found in the study. These ﬁndings with respect to sample size and power implementation produce two warnings. First, for every beneﬁt in a research process, there is a cost involved. For example, when an extraordinarily large sample is employed, the price is the almost automatic rejection of the null hypothesis. Researchers must adopt a dialectical approach and view every decision as a double-sided choice. Second, results concerning the power and sample size criteria employed are revealing about the use of OLS in the marketing literature. Their effective levels are deemed critically important based on careful consideration of the underlying scientiﬁc (not statistical) goals of the study. Statistical considerations typically should be used to identify a plan that is effective in meeting scientiﬁc goals.

Use of Diagnostic Tests The Gauss–Markov theorem outlines the ﬁve main assumptions underlying OLS as follows: (1) the parameters of the model are linear; (2) the sample used is randomly drawn; (3) the expected value of the overall error term is zero; (4) there are no exact linear relationships among the independent variables; and (5) the variance of the overall error term is the same regardless of the different combinations of independent variables possible (Wooldridge, 2003). These assumptions necessitate the use of certain diagnostic tests to conﬁrm that the assumptions hold. However, when the JM studies are analyzed with respect to the use of diagnostics to fulﬁll the assumptions of parameter linearity and homoscedasticity, we ﬁnd that only 7% of the articles provided examples of the appropriate diagnostic tests (C9). We then examined JM articles with hypotheses lacking empirical support to discover whether tests of multicollinearity were conducted.3 Only 51% of the applicable articles addressed multicollinearity in an appropriate manner (C11). This typically would be accomplished through the use of variance inﬂation factors or the tolerance values.

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

403

Though some researchers resorted to the use of the correlation matrix for this purpose, the correlation matrix does not show the existence (or lack) of multicollinearity. Such an examination only exhibits the correlation between any two variables, and cannot detect the collinearity of three or more variables in the absence of indications showing a lack of collinearity between two variables (Kennedy, 2003). Finally, for those studies that make use of longitudinal data, we explored whether appropriate diagnostic measures are taken. The results show that of the 12 articles that employ time-series data, 67% took into consideration issues pertaining to sequential dependence (C12). Multicollinearity is a major reason for failing to provide support for theoretically supported hypotheses. Potentially, it may inﬂate the standard errors (Mason & Perreault, 1991) in OLS regression which, in turn, may cause a coefﬁcient to lack statistical signiﬁcance. In addition, the presence of multicollinearity may lead to coefﬁcients in a direction opposite of established theory. In either form, clearly multicollinearity can distort results. Therefore, it is incumbent upon scholars to test for the existence of multicollinearity using either variance inﬂation factors or tolerance values. We caution that the research process must be perceived as a two-level process. At the ﬁrst level, data analysis techniques are performed, and at the second level, they are doubted and critiqued. Researchers may be too quick to assume robustness without the use of suitable tools to tighten their analysis. The studies surveyed typically do not discuss techniques such as a residual scatter-plot analysis to test for non-linearity of the model parameters (Hair et al., 1998). Similarly, neither the Goldfeld–Quandt test nor the Breusch–Pagan test is employed widely for variance homogeneity (Kennedy, 2003; Kutner, Nachtsheim, & Neter, 2004; Wooldridge, 2003). A good study should be made to deliberate effort toward dialectical thinking and a two-tier analysis in order to avoid these types of problems.

Moderation Analysis OLS regression is also a powerful and dominant tool to detect effects in which a moderator variable interacts with independent variables to explain the variance in the dependent variable (Stone-Romero & Anderson, 1994). Two general approaches are popular in testing for moderation effects using OLS regression. The ﬁrst involves splitting the sample observations into subgroups according to the score of the moderator variable. Another approach employs product terms in the regression and establishes

404

Z. SEYDA DELIGONUL ET AL.

moderation effects by examining the statistical signiﬁcance of the product term’s coefﬁcient. Of the studies that used the split-sample approach, all clearly stated their splitting criteria (100%). Of the six articles that used the median-split approach to test for moderation, only 33% established the signiﬁcance of the means in the subgroups along the moderating variable scores (C14). In such cases, a formal Chow test is typically required to establish the statistical signiﬁcance of the slope parameters of the different subgroups. Of the seven articles that used the split-sample approach, 29% conducted a Chow (or similar) test to claim the existence of the moderating effects (C15). Detecting moderation effects by employing product terms in the regression equations has its own problems. The most notable one is the loss of statistical power in which case a high Type II error rate will occur (Aguinis, 2002). Of the 26 articles that employed a product term to analyze the moderation effect, only 15% explicitly considered the power function of the analysis (C16). When the product term involves categorical variables, the equality of sample sizes and error variances across the different levels of the moderating variable is an important requirement to ensure the power of the test. Of the 12 articles with a product term involving categorical variables, only one provided sufﬁcient information on the equality of the sample size across subgroups and no articles checked for the equality of the error variances (C18). To build theories, most marketing scholars adopt a vantage point close to the contingency school. The contingency school is a class of theories that contend signiﬁcance of situational factors in moderating the relationships between variables. In other words, there is no one universal relationship between A and B, but the relationship is shaped under the effect(s) of C and other effects. Such an approach to address contextual complexity is a suitable method for a scientiﬁc discipline. Techniques to implement this view, however, are deeply problematic in practice. It seems that marketing scholars are on the right track with their choice of the contingent perspective as part of their conceptualization.

Accurate Interpretation of Results Interpretation issues are addressed in two ways: those with methodological considerations and those with practical considerations. On the methodological side, we examined the interpretation of results involving interactions, and the reporting of standardized versus unstandardized coefﬁcients. On the

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

405

practical side, we were concerned with analyzing how managerially interpretable the coefﬁcients are, and whether or not the authors were clear as to the differences between practical and statistical signiﬁcance. In the study, 31 articles had a statistically signiﬁcant interaction effect. However, in only 42% of these cases did the author(s) refrain from making assertions about main effects only when they have a statistically signiﬁcant interaction, or provide additional evidence of the main effects (C19). Also, 62% of the regressions reported standardized coefﬁcients to compare the relative importance of independent variables (C21). However, of the 42 articles in which standardized coefﬁcients were provided, only 29% were actually used for comparison. In addition, the analysis shows that only 20% of the OLS regressions had a set of managerially interpretable coefﬁcients (i.e., as those variables that were objectively measured; that is, not through the use of a Likert scale), while another 41% of the OLS regressions had some managerially interpretable and some non-interpretable coefﬁcients. Finally, the results reveal that only 28% of the articles in the study clearly differentiate between practical and statistical signiﬁcance (C23). The careful interpretation of interaction effects is crucial to ensure that proper conclusions may be reached from the analysis. If an interaction effect is found to be statistically signiﬁcant, then the meaning of the main effect is now in the context of the interaction. That is to say, the main effect is dependent upon the level of the other variable in the interaction (Kutner et al., 2004). Thus, in the presence of statistically signiﬁcant interactions, it is typically not appropriate to interpret the related main effects without referring to the interaction effect. Rather, the complete inﬂuence of an independent variable must include both the main effect as well as the interaction effect (Hair et al., 1998). The second issue is that of standardized versus unstandardized coefﬁcients. When authors try to compare the relative importance of the independent variables in explaining the variation of the dependent variable, an unstandardized coefﬁcient is problematic unless all independent variables are measured in the same units. Authors that use standardized coefﬁcients render the scale of individual predictor variables irrelevant by standardizing the coefﬁcients based on each variable’s standard deviation. Thus, independent variables may be compared without regard for the units originally employed at data entry (Wooldridge, 2003). The ﬁrst of the practical issues that we analyzed related to how managerially interpretable the coefﬁcients were. We believe that the issue of managerial interpretability is crucial to bridge the gap between scholars and practitioners. Often, research in marketing appears to be overrun with

406

Z. SEYDA DELIGONUL ET AL.

abstract constructs. As a result, the applicability of the models to the marketing practitioner may be limited. The last issue of interpretability that we examined is the issue of statistical signiﬁcance versus practical signiﬁcance (McCloskey & Ziliak, 1996). This concern stems from researchers perhaps being overly reliant upon statistical signiﬁcance in deriving their conclusions without devoting enough attention to the practical signiﬁcance of the coefﬁcient. Statistical signiﬁcance tells us whether a coefﬁcient’s distance from zero is probabilistically sizable in comparison to the variance in the sample. The potential ﬂaw of relying only on statistical signiﬁcance is that the researcher can arrive at insigniﬁcance or signiﬁcance by, for example, selecting an appropriate sample size. With a large sample, the variance will be small, forcing the test results toward signiﬁcance. Practical signiﬁcance, on the other hand, examines how substantial the magnitudes are from a practical point of view. It considers if these magnitudes matter in terms of outcomes of pertinent decisions. The current study suggests a certain amount of ambiguity with regard to this issue.

POST-HOC ANALYSIS We analyzed the data for trends in the use of OLS regression over time using Poisson means over four distinct time periods4 (Tarone, 1982). In testing, our motivation for employing Tarone’s approach is two-fold: (1) the data are historical in nature and (2) we cannot presume distributional characteristics and form of trend for our compiled information. This approach is considered robust given that it is known to be asymptotically optimal against any smooth monotone alternative (Tarone, 1982). The test statistic Y2 is computed for each criterion (see Tarone, 1982, for the exact steps to follow; Aczel & Sounderpandian, 2002). Given that the statistic is asymptotically distributed as a w2 random variable with one degree of freedom (Tarone, 1982, p. 458), we tabulated the results in Table 4. Also presented are some special cases, which result from the nature of the data. Tests with sparse count data often lead to inconclusive statistics. Due to a lack of observations, Y2 statistic was not computable for C7, C10, and C18. In addition, Y2 statistic calculations are inconclusive resulting from only afﬁrmation of speciﬁc criteria (C13 and C24) with little variation across the time surveyed. As can be seen in Table 4, only two upward trends were identiﬁed in our study: C2 and C12. Thus, we can conclude that the use of ﬁrst and second

C1. Do the authors state what type of regression was used? C2. Are the ﬁrst and second moments for all regression variables included? C3. Does the paper report ‘‘F ’’ statistics or p-values for the general linear relation? C4. Does the paper report all ‘‘t’’ statistics, standard errors, or p-values? C5. In a multiple regression, did the authors report adjusted R2? C6. Based on a combination of the number of independent variables and sample size, are readers assured that the authors are not guilty of overﬁtting? C7. If the sample size is large, do authors acknowledge that signiﬁcance may be an artifact of large sample size? C8. For null hypotheses not rejected, do the authors state the power of the test? C9. Do the authors provide evidence of the model’s goodness-of-ﬁt? If not, do they at least indicate that they performed such tests and were satisﬁed with the results? C10. If using a small sample and relying on t-, F-, or w2 tests, have the authors established that the residuals are normally distributed or at least close to normally distributed?

Criteria

Time Dimension Analysis.

2/13 (15.4%) 2/20 (10.0%)

0/0

0/0

0/1 (0%)

0/2 (0%) 1/9 (11.1%) 2/16 (12.5%)

10/20 (50.0%) 11/20 (55.0%) 17/20 (85.0%) 20/20 (100%) 10/20 (50.0%) 19/20 (95.0%)

1994–1997

7/16 (43.8%) 3/16 (18.8%) 5/16 (31.3%) 16/16 (100%) 7/15 (46.7%) 15/16 (93.8%)

1990–1993

0/1 (0%)

1/13 (7.7%) 2/20 (10.0%)

0/2 (0%)

4/20 (20.0%) 12/20 (60.0%) 14/20 (70.0%) 19/20 (95.0%) 11/20 (55.0%) 16/19 (84.2%)

1998–2001

0/1 (0%)

1/11 (9.1%) 0/27 (0%)

0/7 (0%)

10/27 (37.0%) 15/27 (55.6%) 16/27 (59.3%) 26/27 (96.3%) 15/26 (57.7%) 22/27 (84.2%)

2002–2005

‘‘Yes’’ Count/Total Number of Articles (Where Applicable) Period Covered

Table 4.

Upward trend No trend No trend No trend No trend

6.68 3.46 2.08 2.91 0.90

No trend

1.20

No value

No trend

0.17

No value

No trend

Trend

0.06

Signiﬁcance Test (Y2)

Methodological Rigor of OLS Regression Studies in Mainstream Marketing 407

C11. If theoretically supported hypotheses are not supported statistically, have tests been conducted for multicollinearity? C12. If the data covers a span of time, have the authors dealt with longitudinal considerations? C13. For articles using split sample to test moderation effects, are the authors clear on criterion used to split the sample? C14. For regressions using split sample by median to test moderation effects, do the authors establish that the groups are statistically different on means of the moderating variable? C15. For articles using split sample to test moderation effects and to claim moderating effects, do the authors conduct a formal Chow test or tests of similar effect? C16. For articles using interaction to test moderation effects, do the authors take into consideration any possible reduced power? C17. When moderation analyses are conducted on categorical interaction variables, do the authors establish that there are equal sample sizes across levels?

Criteria

0/1 (0%)

0/5 (0%)

1/1 (100%)

0/1 (0%)

1/1 (100%)

0/3 (0%)

3/9 (33.3%)

1990–1993

0/2 (0%)

2/5 (40.0%)

0/3 (0%)

2/3 (66.7%)

3/3 (100%)

1/2 (50.0%)

6/14 (42.9%)

1994–1997

0/4 (0%)

1/7 (14.3%)

1/3 (33.3%)

0/2 (0%)

3/3 (100%)

1/1 (100%)

6/12 (50.0%)

1998–2001

1/5 (20.0%)

1/9 (11.1%)

0/0

0/0

0/0

6/6 (100%)

8/10 (80.0%)

2002–2005

‘‘Yes’’ Count/Total Number of Articles (Where Applicable) Period Covered

Table 4. (Continued )

1.33

0.05

No trend

No trend

No trend

No trend

o0.01

0.18

No trend

Upward trend

No trend

Trend

0.21

5.39

1.00

Signiﬁcance Test (Y2)

408 Z. SEYDA DELIGONUL ET AL.

C18. When moderation analyses are conducted on categorical interaction variables, do the authors establish that there are equal error variances across subgroups? C19. Do the authors either refrain from making assertions about main effects when they have a signiﬁcant interaction or provide additional evidence of main effects? C20. Are the coefﬁcients of the model managerially interpretable? C21. Do the authors report standardized coefﬁcients within the model when necessary? C22. If standardized coefﬁcients are reported, do the authors interpret them within the model? C23. Are the authors clear on and able to express what constitutes practical – rather than statistical –signiﬁcance? C24. Do the authors avoid making assertions of practical signiﬁcance based on coefﬁcients that are not statistically signiﬁcant? 2/9 (22.2%) 6/19 (31.6%)

5/9 (55.6%) 6/16 (37.5%) 19/19 (100%)

3/20 (15.0%) 4/7 (57.1%)

4/16 (25.0%) 5/6 (83.3%)

15/15 (100%)

4/5 (80.0%)

0/2 (0%)

4/5 (80.0%)

0/1 (0%)

18/18 (100%)

4/19 (21.1%)

3/11 (27.3%)

2/19 (10.5%) 2/4 (50.0%)

2/8 (25.0%)

0/4 (0%)

27/27 (100%)

7/27 (25.9%)

2/13 (15.4%)

7/27 (25.9%) 2/4 (50.0%)

3/13 (23.1%)

0/5 (0%)

3.1

0.01

No trend

No trend

No trend

No trend

0.94

1.07

No trend

No trend

0.80

2.52

No value

Methodological Rigor of OLS Regression Studies in Mainstream Marketing 409

410

Z. SEYDA DELIGONUL ET AL.

moments (mean and standard deviation, respectively) for all regression variables (C2) has improved over the 15-year time period. Also, progress has been made in the application of longitudinal studies (C12). However, while no trend was found for the rest of the criteria used in our assessment of the marketing literature, it is important to note that tests for multicollinearity (C11) have improved from 33% to 80% over the time surveyed. This indicates an increased awareness concerning the required diagnostics associated with implementing OLS regression models. Also, promising results may be discussed for criteria already in high compliance (e.g., above 80%) with the parameters of this study: C4, C6, C13, and C24. Though we did not ﬁnd an upward trend with these criteria, we can infer that proper implementation has been conducted when reporting the statistical signiﬁcance of coefﬁcients in a model (C4), ﬁtting data (C6), providing clear reasons for testing moderation by split sample (C13), and avoiding practical inferences based on statistically insigniﬁcant coefﬁcients (C24). Three criteria have shown a marked decline: discussion concerning main effects in interaction models (C19), reporting standardized coefﬁcients when necessary (C21), and interpretation of standardized coefﬁcients (C22). Even though the Y2 statistic was not found to be signiﬁcant, the decline observed for these two criteria suggest that particular improvement is needed in these criteria for future research.

PRESCRIPTIVE SUGGESTIONS The present analysis is based on what has appeared in the published articles of the JM. It is possible that many cases discussed above conducted certain tests. However, due to space limitations, these methodological details may have been excluded. Nevertheless, we propose a set of guidelines for assessing OLS regression analysis. Though not completely comprehensive, these suggestions address topics of highest priority for future studies employing OLS (Hair et al., 1998; Kennedy, 1998; Kutner et al., 2004; Mason & Perreault, 1991; Pindyck & Rubinfeld, 1998; Wooldridge, 2003). While improvement has been noted on some of these topics, a comprehensive set of steps is included for reference. These steps may be categorized into ﬁve separate groups addressing particular topics related to the use of regression: (a) general reporting of empirical contents; (b) sample size; (c) diagnostic tests and remedies; (d) moderation; and (e) interpretation.

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

411

Reporting The distribution of variables included in an OLS model is generally described by their means and standard deviations (Pindyck & Rubinfeld, 1998). Since each of these two descriptive statistics is a representation of the sample derived for a particular study, their inclusion is important for understanding the underlying distribution properties of any variable included in the model (Pindyck & Rubinfeld, 1998). Therefore, each of these descriptive statistics must be included to assess the average of the outcomes obtained and dispersion of both independent and dependent variables in an OLS model. Reporting the t-statistic of an independent variable is equally important when employing OLS as an estimation technique. Since it notes the statistical signiﬁcance of a single parameter within an estimated model, this statistic is necessary to determine whether a particular hypothesis is supported by the estimated model (Pindyck & Rubinfeld, 1998; Wooldridge, 2003). Generally accepted signiﬁcance levels are 0.01 and 0.05 (Pindyck & Rubinfeld, 1998; Wooldridge, 2003), though less stringent models include 0.10, as well. Pindyck and Rubinfeld (1998) indicate that including exact p-values (e.g., 0.07) in a model may be more useful to the researcher than indications of statistical signiﬁcance at particular levels (e.g., 0.01 or 0.05). Whether the statistical signiﬁcance of a parameter is reported by p-value, standard error, or t-statistic, is not paramount. However, the researcher must give some indication of a parameter’s statistical signiﬁcance to assist with the interpretation of an OLS model. Likewise, reporting the F-statistic is equally essential with an OLS regression. Since it indicates a general statistical signiﬁcance test of the overall model, the F-statistic (or, its p-value) is typically included together with the R2 or adjusted R2 (Wooldridge, 2003). Thus, by displaying this statistic, researchers are better able to assess the overall signiﬁcance of the model itself. While the R2 of an OLS equation should always be included (Wooldridge, 2003), there are limitations to using only R2 as the coefﬁcient of determination (Kutner et al., 2004; Pindyck & Rubinfeld, 1998). Adjusted R2 should not necessarily be used to replace the R2 statistic (Wooldridge, 2003), but an important reason to use the adjusted R2 statistic with an OLS model is that – contrary to R2 – it penalizes OLS models employing more independent variables (Hair et al., 1998; Kennedy, 1998; Kutner et al., 2004; Pindyck & Rubinfeld, 1998; Wooldridge, 2003). Since marketing researchers typically provide multiple variations of independent variables and compare

412

Z. SEYDA DELIGONUL ET AL.

OLS models as a result, the inclusion of both a model’s R2 and adjusted R2 would allow researchers to more directly evaluate separate models.

Sample Size Recommendations regarding sample size address issues with small samples, large samples, and the power of tests that may be performed on a model. Considered one of the most important elements within a researcher’s control (Hair et al., 1998), the role of a project’s sample size may have important inﬂuence on the ﬁndings discovered. We recommend that researchers clearly report sample size for their tests. This is especially important when authors use different sample sizes in the same study to test more than one regression equation. Simply reporting the overall sample size is not sufﬁcient for proper interpretation of results. In addition, future researchers should provide further evidence of normal (or, near-normal) distribution of the data used or its residuals (Hair et al., 1998; Kutner et al., 2004; Pindyck & Rubinfeld, 1998).

Diagnostic Tests We also propose that authors need to address the diagnosis issue of an OLS model for a model’s lack of linearity, heteroscedasticity, multicollinearity, and autocorrelation (Hair et al., 1998; Kennedy, 1998; Kutner et al., 2004; Pindyck & Rubinfeld, 1998; Wooldridge, 2003). This topic is covered extensively in the traditional marketing methods literature. As such, we will not consider the topic in great detail here.

Moderation When a moderation effect is tested by means of a split sample, authors should clearly elaborate on the criteria used to create the subgroups and preferably provide some theoretical justiﬁcation for the choice. As mentioned earlier, a median split and a mean split approach may well result in both different subgroup sample sizes as well as different moderation effect conclusions. When split by median is employed, it is essential to test for the signiﬁcance of the means of the moderating variables across groups so to ensure enough power to detect the moderating effect.

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

413

When product terms are used in a moderated multiple regression, authors must check the equality of both the sample sizes and the equality of error variances across the subgroups so to ensure the statistical power of the study. When a moderating effect fails to emerge through this approach, caution should be added in the model’s interpretation.

Interpretation Particular interpretation of OLS models may be categorized into two distinct topics concerning standardized coefﬁcients and interaction effects. Important with the use of different units among independent variables, standardized coefﬁcients allow for the description of relative importance of the independent variables included (Hair et al., 1998; Pindyck & Rubinfeld, 1998). By rendering the scale of individual predictor variables irrelevant with standardized coefﬁcients, independent variables may be compared more deﬁnitively (Wooldridge, 2003). In addition, the interpretation of interaction effects in OLS models is important for researchers. Once an interaction term is introduced that originally had only main effects, the original independent variable(s) cannot be interpreted alone (Pindyck & Rubinfeld, 1998). Rather, the complete inﬂuence of an independent variable must include both the main effect as well as the interaction effect (Hair et al., 1998).

CONCLUSION We conclude by commenting on four main topics. First, we offer some thoughts about the general status of OLS-related research. Second, we brieﬂy discuss improvements in the application of OLS in marketing research. Third, we introduce some possible areas for improvement. Finally, we provide some concluding thoughts with regard to OLS in marketing research.

General Status Overall, scholars in marketing tend to use OLS competently. In fact, the present results closely follow those from other ﬁelds – speciﬁcally economics (McCloskey & Ziliak, 1996). While the time period of this study differs,

414

Z. SEYDA DELIGONUL ET AL.

a comparison of related survey items shows remarkable similarity in results. For example, akin to marketing scholars, economists report power in less than one-sixth of the sample. In addition, the distinction of statistical and practical signiﬁcance is recognized by one-third of the studies in both ﬁelds. Also, articles in economics report descriptive statistics in one-third of the sample. Such resemblance between these ﬁelds may be attributed to the origin of the two disciplines, the background of each ﬁeld’s scholars, or a similar level of rigor inherent in each ﬁeld. These parallels are curious observations which deserve further investigation.

Usage Transformation OLS as a methodological tool appears to be used more effectively in marketing. The reporting of interpretative aspects has evolved over the years particularly with regard to basic considerations. Over the time period studied (1990–2005), an upward trend was noted for the application of multicollinearity tests with unsupported hypotheses (C11) in OLS studies. In addition, considerations pertaining to longitudinal relationships (C12) tended to have an upward trend. Such interest and improvement in time series studies may be a result of having access to more sophisticated statistical techniques and of improved data gathering.

Improvement Opportunities The remaining criteria showing considerably less change and may be grouped in three different categories: (a) fundamental reporting concerns; (b) moderation study topics; and (c) various OLS model interpretation issues. First, less improvement has been made on a wide array of criteria ranging from the clear statement of the type of regression used (C1) to an indication of the OLS model’s power when theoretically posited hypotheses are not supported (C8). In addition to these two particular criteria, general improvement is called for in the indication of the statistical signiﬁcance of general linear relationships (C3), of a model’s adjusted R2 (C5), that signiﬁcance may be a result of large sample size (C8), that goodness-of-ﬁt tests were performed (C9), and that residuals are normally distributed (C10). In fact, with nearly 98% of articles utilizing this criterion, the reporting of coefﬁcient statistical signiﬁcance (C4) has remained quite high. Also, we observed that almost 88% of the marketing studies avoided overﬁtting.

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

415

Taken together, these two ﬁndings in themselves indicate marketing scholars seem to be applying OLS reasonably well. With respect to studies employing moderation, an equally compelling issue relates to the use of precise information concerning split samples (C13). This indicates a fundamental understanding of OLS. However, we offer caution on several issues: a requirement is that groups resulting from split samples have statistically different means (C14); a Chow or similar statistical test has been performed to establish that the slopes of the groups are different (C15); reduced power has been considered in contexts using moderation (C16); and equal sample sizes (C17) and error variances (C18) are established with categorical interaction variables. By considering and improving on these issues, marketing scholars will be better equipped to enhance the rigor of OLS studies employing moderation analyses. The third category addressing various OLS model interpretation issues has one criterion that has been implemented well thus far. The avoidance of assertions concerning practical signiﬁcance with statistically insigniﬁcance coefﬁcients (C24) indicates that there is a ﬁrm understanding of the basis for discussion in OLS studies: statistical signiﬁcance. However, we call on authors to make adjustments in their application of OLS in two areas: (1) by including managerially interpretable coefﬁcients (C20) and (2) being much more clear on the relevance of statistical and practical signiﬁcance (C23). In general, this call follows recent discussion in the marketing literature to structure future studies to be more relevant to practitioners.

Final Thoughts From the Kuhnian perspective, normal science (with marketing as an exemplar) is a science where knowledge gradually accumulates as scholars generate solutions to extant puzzles. Theories respond to peculiarities of a current dilemma posed by previous and current studies. However, as time progresses, scientiﬁc inquiry focuses on anomalies which end up becoming mismatches between the current best models available and tangible observation. The tensions that arise from this failure to predict results become equally potent in driving new adjustments to the dominant views of the ﬁeld. Although the nature of knowledge advancement in normal science has long been under scrutiny, techniques that drive the progress of science have received undeservedly less attention. As knowledge accumulates in a normal science, what happens to the techniques of inquiry? Over time, do they

416

Z. SEYDA DELIGONUL ET AL.

improve, worsen, or remain the same? This study argues and demonstrates that similar progress occurs in research methodologies employed by marketing scholars. In their quest for scholarship, investigators start from generic techniques. A tool such as OLS regression is sufﬁcient for most early empirical work. Further theorizing and resulting anomalies in knowledge compel scholars to adjust the modus operandi. Thus, in time, adaptations and alterations from the core method drive contingent questions and emergent methods with increased rigor. For example, SEM has become an attempt to resolve the quest for causality, logistic regression for scale problems with the dependent variable, and partial least squares regression for particular cases of small sample sizes and non-linearity. Simply put, derivations from a generic technique such as OLS multiply with the needs of the ﬁeld. In a discipline such as marketing, the tools available to scholars continuously improve and develop to ﬁt the demands of scholarship. However, as newer tools such as SEM appear, the generic version does not vanish. Based on this study, OLS remains as a pervasive tool with staying power in the ﬁeld of marketing. Moreover, its use tends to be more selective to target its best application. OLS will continue to occupy an important place in empirical work. With ongoing reﬁnement of its use, adherence to the prescriptive guidelines offered here, and the precision of its application, the marketing discipline will continue to beneﬁt from its use.

NOTES 1. We chose not to include other forms of estimation methods (e.g., maximum likelihood estimation, generalized least squares, etc.) partly due to the lack of frequency of use and to avoid comparison difﬁculties. 2. The power of a test is deﬁned as the probability of rejecting the null hypothesis given that the null hypothesis is false. If the power of a test is .9 then the likelihood of choosing H1 (alternative hypothesis), given that H0 (null hypothesis) is false is .9. A lack of power in OLS testing may hamper selecting the alternative hypothesis when the alternative should be selected. 3. We treated each OLS article as one, assuming that any test of multicollinearity would be conducted for all variables in the equation. 4. Testing of sparse data for trends is most challenging. In our case, it was appropriate to use a special test, which has its origins in Cochran’s (1954) work about trends in Poisson count processes. When sequential data need to be analyzed, this test deals with the hypothesis of ‘‘no change in the mean’’ over multiple time windows. To derive a statistic, the test assumes each count is distributed as a Poisson random variable with mean l that conforms to any twice-differentiable

Methodological Rigor of OLS Regression Studies in Mainstream Marketing

417

monotone-function over a closed interval. In other words, the test recognizes any trend that can be expressed as a function of a linear function. By requiring twice differentiable, the analysis excludes some cases such as a step function, but includes a wide variety of forms including curves. The use of this test in our study is easily justiﬁable because the test is designed for count processes and is relatively effective with sparse data sets. In addition, the test is known to be asymptotically optimal against any smooth monotone alternative (Tarone, 1982).

REFERENCES Aczel, A. D., & Sounderpandian, J. (2002). Complete business statistics. New York, NY: McGraw-Hill/Irwin. Aguinis, H. (2002). Estimation of interaction effects in organization studies. Organizational Research Methods, 5(3), 207–211. Aguinis, H., & Stone-Romero, E. F. (1997). Methodological artifacts in moderated multiple regression and their effects on statistical power. Journal of Applied Psychology, 82, 192–206. Bartels, R. (1951). Can marketing be a science? Journal of Marketing, 15(03), 319–328. Cochran, W. G. (1954). The combination of estimates from different experiments. Biometrics, 10, 101–129. Friedrich, R. J. (1982). In defense of multiplicative terms in multiple regression equations. American Journal of Political Science, 26(4), 797–833. Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall. Hult, G. T. M., Neese, W. T., & Bashaw, R. E. (1997). Faculty perceptions of marketing journals. Journal of Marketing Education, 19(1), 37–52. Hunt, S. D. (1976). The nature and scope of marketing. Journal of Marketing, 40(03), 17–28. Hunt, S. D. (2002). Foundations of marketing theory: Toward a general theory of marketing. Armonk, NY: M.E. Sharpe. 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. Kennedy, P. (1998). A guide to econometrics. Cambridge, MA: The MIT Press. Kennedy, P. (2003). A guide to econometrics. Cambridge, MA: The MIT Press. Kerin, R. A. (Ed.) (1996). In pursuit of an ideal: The editorial and literary history of the Journal of Marketing. Journal of Marketing, 60(1), 1–13. Ketchen, D. J., & Shook, C. L. (1996). The application of cluster analysis in strategic management research: An analysis and critique. Strategic Management Journal, 17(6), 441–458. Kotabe, M. (1990). Corporate product policy and innovative behavior of European and Japanese multinationals: An empirical investigation. Journal of Marketing, 54(2), 19–33. Kutner, M. H., Nachtsheim, C. J., & Neter, J. (2004). Applied linear regression models. Boston, MA: McGraw-Hill/Irwin. Mason, C. H., & Perreault, W. D., Jr. (1991). Collinearity, power, and interpretation of multiple regression analysis. Journal of Marketing Research, 28(3), 268–280. McCloskey, D. N., & Ziliak, S. T. (1996). The standard error of regressions. Journal of Economic Literature, 34(1), 97.

418

Z. SEYDA DELIGONUL ET AL.

Mort, G. S., McColl-Kennedy, J. R., Kiel, G., & Soutar, G. N. (2004). Perceptions of marketing journals by senior academics in Australia and New Zealand. Australasian Marketing Journal, 12(2), 51. Pindyck, R. S., & Rubinfeld, D. L. (1998). Econometric models and economic forecasts. Boston, MA: McGraw-Hill/Irwin. Sheth, J. N., Gardner, D. M., & Garrett, D. E. (1988). Marketing theory: Evolution and evaluation. New York, NY: Wiley. Sheth, J. N., & Sisodia, R. S. (1999). Revisiting marketing’s lawlike generalizations. Journal of the Academy of Marketing Science, 27(1), 71–87. Shook, C. L., Ketchen, D. J., Hult, G. T. M., & Kacmar, K. M. (2004). An assessment of the use structural equation modeling in strategic management research. Strategic Management Journal, 25(4), 397–404. Stone, E. F., & Hollenbeck, J. R. (1989). Clarifying some controversial issues surrounding statistica. Journal of Applied Psychology, 74(1), 3–10. Stone-Romero, E. F., Alliger, G. M., & Aguinis, H. (1994). Type II error problems in the use of moderated multiple regression for the detection of moderating effects of dichotomous variables. Journal of Management, 20(1), 167–178. Stone-Romero, E. F., & Anderson, L. E. (1994). Relative power of moderated multiple regression and the comparison of subgroup correlation coefﬁcients for detecting moderating effects. Journal of Applied Psychology, 79, 354–359. Tarone, R. E. (1982). The use of historical control information in testing for a trend in Poisson means. Biometrics, 38(2), 457–462. Taylor, W. J. (1965). Is marketing a science? Revisited. Journal of Marketing, 29(03), 49–53. Theoharakis, V., & Hirst, A. (2002). Perceptual differences of marketing journals: A worldwide perspective. Marketing Letters, 13(4), 389. Wathne, K. H., Biong, H., & Heide, J. B. (2001). Choice of supplier in embedded markets: Relationship and marketing program effects. Journal of Marketing, 65, 54. Wooldridge, J. M. (2003). Introductory econometrics: A modern approach. Cincinnati, OH: Thomson South-Western. Ziliak, S. T., & McCloskey, D. N. (2004). Size matters: The standard error of regressions in the American Economic Review. Journal of Socio-Economics, 33(5), 527–554.

New Challenges to International Marketing (Advances in International Marketing, Volume 20)

International Marketing: Emerging Markets (Advances in International Marketing, Volume 21)

New Directions in International Advertising Research (Advances in International Marketing)

Introduction to International Marketing (Marketing in Action)

International Marketing

International Marketing , Eighth Edition

International Marketing Research

International Marketing (15th Edition)

International Marketing Research

Advances in International Accounting, Volume 20

Challenges in Relationship Marketing

Advances in International Accounting, Volume 20

Relationship Between Exporters and Their Foreign Sales and Marketing Intermediaries (Advances in International Marketing, Volume 16)

International Marketing: Strategy and Theory

Advances in Electronic Marketing

Advances in Electronic Marketing

Reassessing the Internationalization of the Firm (Advances in International Marketing)

International Marketing Review, Volume 20, Number 3, 2003, International business negotiations

Advances in International Accounting, Volume 18 (Advances in International Accounting)

Advances in International Accounting, Volume 17 (Advances in International Accounting)

Advances in International Accounting, Volume 16 (Advances in International Accounting)

Advances in International Accounting, Volume 16 (Advances in International Accounting)

Advances in International Accounting, Volume 18 (Advances in International Accounting)

Advances in International Accounting, Volume 17 (Advances in International Accounting)

Marketing and Football: an international perspective (Sports Marketing)

International Perspectives, Volume 20 (Advances in Learning and Behavioral Disabilities)

International Marketing Research, Volume 17: Opportunities and Challenges in the 21st Century

International Marketing Research, Volume 17: Opportunities and Challenges in the 21st Century

The New Marketing

Marketing

Marketing

New Challenges to International Marketing (Advances in International Marketing, Volume 20)

International Marketing: Emerging Markets (Advances in International Marketing, Volume 21)

New Directions in International Advertising Research (Advances in International Marketing)

Introduction to International Marketing (Marketing in Action)

International Marketing

International Marketing , Eighth Edition

International Marketing Research

International Marketing (15th Edition)

International Marketing Research

Advances in International Accounting, Volume 20

Challenges in Relationship Marketing

Advances in International Accounting, Volume 20

Relationship Between Exporters and Their Foreign Sales and Marketing Intermediaries (Advances in International Marketing, Volume 16)

International Marketing: Strategy and Theory

Advances in Electronic Marketing

Advances in Electronic Marketing

Reassessing the Internationalization of the Firm (Advances in International Marketing)

International Marketing Review, Volume 20, Number 3, 2003, International business negotiations

Advances in International Accounting, Volume 18 (Advances in International Accounting)

Advances in International Accounting, Volume 17 (Advances in International Accounting)

Advances in International Accounting, Volume 16 (Advances in International Accounting)

Advances in International Accounting, Volume 16 (Advances in International Accounting)

Advances in International Accounting, Volume 18 (Advances in International Accounting)

Advances in International Accounting, Volume 17 (Advances in International Accounting)

Marketing and Football: an international perspective (Sports Marketing)

International Perspectives, Volume 20 (Advances in Learning and Behavioral Disabilities)

International Marketing Research, Volume 17: Opportunities and Challenges in the 21st Century

International Marketing Research, Volume 17: Opportunities and Challenges in the 21st Century

The New Marketing

Marketing

Marketing

Recommend Documents