Effective Knowledge Transfer in Multinational Corporations - EPDF.TIPS

0 downloads 0 Views 2MB Size Report
2. 2 Knowledge and the MNC. 5. Introduction to knowledge management: ..... impinge on the success of international knowledge transfer, concrete ..... bered that no knowledge transfer is context-free. ...... Wissensmanagement-Schritte zum.
Effective Knowledge Transfer in Multinational Corporations Tina C. Chini

Effective Knowledge Transfer in Multinational Corporations

Effective Knowledge Transfer in Multinational Corporations Tina C. Chini Vienna University of Economics and Business Administration

© Tina C. Chini 2004 Foreword © Bodo B. Schlegelmilch 2004 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The author has asserted her right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2004 by PALGRAVE MACMILLAN Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N.Y. 10010 Companies and representatives throughout the world PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St. Martin’s Press, LLC and of Palgrave Macmillan Ltd. Macmillan® is a registered trademark in the United States, United Kingdom and other countries. Palgrave is a registered trademark in the European Union and other countries. ISBN 1–4039–4220–X This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Chini, Tina C., 1979– Effective knowledge transfer in multinational corporations / Tina C. Chini. p. cm. Includes bibliographical references and index. ISBN 1– 4039–4220–X 1. Knowledge management. 2. International business enterprises. I. Title. HD30.2.C47445 2004 658.4′038—dc22 2004045424 10 9 13 12

8 11

7 10

6 5 09 08

4 07

3 06

2 05

1 04

Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham and Eastbourne

To my mother on her 50th birthday

Contents List of Exhibits

ix

Foreword by Bodo B. Schlegelmilch

xii

Preface

xiv

Acknowledgements

xvi

1 Positioning

1

The knowledge management challenge Aim of the research

2 Knowledge and the MNC

1 2

5

Introduction to knowledge management: a conceptual background The relevance of knowledge management in the MNC

3 Knowledge-Based Determinants of MNC Strategic Configuration

5 19

37

Headquarters–subsidiary relationships Strategic mandates, coordination and control The capability perspective Contingency factors

37 39 49 52

4 A Model of Knowledge Transfer in MNCs

58

Strategic mandate Value of knowledge stock Knowledge transfer capabilities Knowledge transfer effectiveness Organizational distance Cultural distance

5 Research Design and Methodology Research context Data collection Operationalization and measures vii

58 60 61 64 65 66

68 68 70 76

viii Contents

6 Analysis and Results Descriptives of the unit of analysis Analysis of the model’s constructs Hierarchical relationships and culturally close subsidiaries Hierarchical relationships and culturally distant subsidiaries Lateral relationships of subsidiaries

7 Conclusion, Limitations and Implications Conclusion Limitations Implications for future research Managerial implications

81 81 82 107 124 129

138 138 145 146 147

Appendix 1

Specification of the Measurement Model

149

Appendix 2

Model Identification and Assessment

153

Notes and References

157

Bibliography

159

Index

170

List of Exhibits

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.1 3.2 4.1 4.2 4.3 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 6.1 6.2 6.3 6.4

The continuum of data – information – knowledge Tacit and explicit knowledge Definitions of knowledge management The knowledge management value chain A simplified communication model Nine knowledge transfers The knowledge spiral Studies on intra-MNC knowledge transfers Areas of empirical contributions Subsidiaries’ strategic mandate typologies Control mechanisms A model of intra-MNC knowledge transfer Strategic mandates of subsidiaries Overview of hypotheses 1a–1d Industry weights of target and final sample Industries in the final sample Regions in the sample Location of headquarters Location of headquarters in Central/Western Europe Location of subsidiaries Knowledge flows to and from headquarters Knowledge flows to and from the focal subsidiary Employees at units Positions of respondents Strategic mandates of subsidiaries Cross-tabulation: strategic mandate and number of employees 6.5 Strategic positions of headquarters 6.6 Strategic positions of global versus regional headquarters 6.7 Knowledge stock of subsidiaries compared with peer subsidiaries 6.8 Knowledge stock of subsidiaries compared with their headquarters 6.9 Knowledge stock of headquarters compared with subsidiaries 6.10 Hierarchical formal transmission channels ix

7 9 11 13 15 16 18 30 35 41 45 59 59 64 73 73 74 74 75 75 76 76 82 83 84 85 87 87 88 89 89 92

x List of Exhibits

6.11 6.12 6.13 6.14 6.15 6.16 6.17 6.18 6.19 6.20 6.21 6.22 6.23 6.24 6.25 6.26 6.27 6.28 6.29 6.30 6.31 6.32 6.33 6.34 6.35 6.36 6.37 6.38 6.39 6.40 6.41 A1.1 A1.2

Lateral formal transmission channels Informal transmission channels Knowledge management infrastructure Use of knowledge management tools Composition of knowledge transfer processes Knowledge transfer processes Subsidiaries’ perceptions of similarity vis-à-vis headquarters and peer subsidiaries Headquarters’ perceptions of similarity vis-à-vis culturally close and distant subsidiaries Impact of culture on knowledge transfers Subsidiaries’ benefits of knowledge transfers from headquarters Subsidiaries’ benefits of knowledge transfers from culturally close subsidiaries Subsidiaries’ benefits of knowledge transfers from culturally distant subsidiaries Headquarters’ benefits of knowledge transfers from culturally close subsidiaries Headquarters’ benefits of knowledge transfers from culturally distant subsidiaries Satisfaction with knowledge management Creation of a structural equation model Composition of latent variables Formative and reflective variables Selection of variables Model 1 Model 2 Model 3 1st Dataset Global measures of fit of models 1–3 Hypotheses and direct effects Differences between groups 2nd Dataset Hypotheses and direct effects (2nd Dataset) 3rd Dataset Hypotheses and direct effects (3rd Dataset) General assessment of hypotheses Unidimensionality of constructs CFA results: hierarchical relationships

92 93 94 96 97 98 99 100 101 103 104 104 105 105 107 109 110 111 112 113 113 114 114 115 117 120 125 126 129 131 134 149 151

List of Exhibits xi

A1.3 Indicator reliability, factor reliability and average variance explained A2.1 Model descriptives A2.2 Overview of measures of fit A2.3 Suggested assessment of measures of fit

152 154 155 156

Foreword

Doctoral students are usually an enthusiastic and intelligent lot and, in the majority of cases, it is a pleasure to work with them. However, very occasionally there are students who are truly outstanding. In these rare cases, working together is more than a pleasure, it is a privilege. Tina Chini has been such an exceptional student, both in terms of ability and drive. And the outcome in form of this book showcases the quality of her work. Her topic, ‘Effective Knowledge Transfer in Multinational Corporations’ is most relevant for multinational firms. Given that knowledge is now widely regarded as the key determinant of competitiveness, it is vital for companies to understand how knowledge transfer between globally dispersed units is best organized. The problem, however, is that this is not an easy topic to tackle. First, knowledge is a complex construct that is difficult to capture. Second, the data collection invariably has to be conducted in different countries and different cultures. Tina Chini manages to handle these and other methodological challenges very well. She initially points out that not quantity but quality of knowledge transfer is driving the success of multinational corporations. While this appears obvious, it is often ignored in empirical studies that find it easier to assess quantity rather than quality of knowledge flows. In contrast, Tina Chini skilfully integrates both perspectives. Based on a comprehensive literature review, she subsequently develops pertinent research hypotheses and a causal model depicting the complexities of knowledge transfer effectiveness in multinational firms. Next, considerable room is given to provide details of the empirical research design and the development of appropriate measurement scales, including reliability and validity checks. The merits and limitations of the selected analytical technique, namely causal analysis, are also discussed. Together, these insights underline once more the methodological challenges faced by Tina Chini. Her discussion introduces the results and carefully reflects on them in the light of the existing literature. Based on this, Tina Chini explains the diverse implications of her work for the managerial practice of xii

Foreword xiii

multinational companies as well as for theory development. The scope for future research avenues is also outlined. With this book Tina Chini has managed to produce a highly relevant, methodologically sound and comprehensive analysis of knowledge transfer effectiveness in multinational companies. Taken her methodological skills and drive, it is likely that we will see more such high-quality contributions by this researcher. B ODO B. S CHLEGELMILCH Chair of International Marketing and Management Wirtschaftsuniversität Wien (WU Wien)

Preface

Nine out of ten prefaces start by thanking the authors’ family and friends for supporting them during the stressful time of working on their book. Probably this indicates that a particularly hard time is over now and ‘things are getting better’. Having reached this stage myself, I would also like to thank a number of people without whose support this book and the underlying research would not have been possible. Those who have really enabled this research – and will, I hope, also benefit from its outcome at one point or other – are the managers who participated in the study. I hope my lengthy questionnaire and my numerous follow-up calls have not discouraged them to support further empirical studies. Special thanks go to my advisers, Professors Bodo B. Schlegelmilch and Constantine S. Katsikeas, who always had an open ear for my ideas (and complicated models) and shared their experience with me. Next in line are my colleagues, who have gone through similar tortures and were among the few people able to understand what I was doing. Elisabeth deserves to be mentioned separately here as she patiently endured all my beginner’s questions while she was in the final stages of her PhD thesis. I must acknowledge the deepest debt to Nils from the University of Hamburg who has dedicated much time teaching me structural equation modelling. After some more time, however, I reached the point when no one was able to understand what I was really doing and then writing a PhD became a very lonely activity – especially when the most intense phase of writing happens to be in the hottest summer of the century . . . Spending most of the time in a boiling office, I started to divide my friends into two groups: those I would not call any more and those who had to listen to my problems quite frequently. I would like to thank both groups: the first one for still ‘knowing’ me – although some of them might think they were really lucky to be in the first group; the second one, especially Philippa, Denise and Ulli, for lending me an ear and for all the intense discussions we had – obviously not only about my thesis. I must not forget to mention Philipp, who, despite not being in a daily-glass-of-wine reach, supported me a great deal. I will always keep the weekend breakfasts with Theresa in good memory as well as all the nice evenings spent with Kathi and Nina. And I am especially grateful to Tom, who always advocated a xiv

Preface xv

healthy mind–body balance and made me do all kinds of physical exercise from climbing to yoga. I thank you all for being my friends and for supporting me during this stressful time. But, in my case, this does not indicate that the hard times are over. Times might even get harder as our endeavours seem to lead many of us to different places. I could not have published this thesis had it not survived the critical comments of my reviewers Björn, Philippa and my father. At this point I thank Jacky Kippenberger, my editor at Palgrave Macmillan, and copyeditor Keith Povey for helping me to turn my thesis into a book. As far as my own survival is concerned, I owe much to my grandmother, who made sure that my fridge contained some food from time to time, my grandparents and my brother. My last years would not have been so wonderful without Björn. I thank him for sticking with me through all my ups and downs and for constantly challenging me. Most important, not only during the last years but during my entire life, was the support of my parents who – despite not always sharing my views – always helped me to accomplish my goals. This book is dedicated to my mother on her 50th birthday. TINA C. C HINI

Acknowledgements

Exhibit 2.4: KM Value Chain, from From Knowledge Theory to Management Practice: Towards an Integrated Approach by Minsoo Shin, Tony Holden and Ruth A. Schmidt, copyright 2001 Elsevier Ltd. Used by permission of Elsevier Ltd. Exhibit 2.6: The Nine Knowledge Transfers, from A Knowledge-Based Theory of the Firm to Guide in Strategy Formulation by Karl-Eric Sveiby, copyright 2001 Emerald Group Publishing Ltd. Used by permission of Emerald Group Publishing Ltd. Exhibit 2.7: Four Modes of Knowledge Conversion, from The KnowledgeCreating Company: How Japanese Companies Create the Dynamics of Innovation by Ikujiro Nonaka and Hirotaka Takeuchi, copyright 1995 by Oxford University Press, Inc. Used by permission of Oxford University Press, Inc. Exhibit 3.1: Subsidiary Strategy Typologies, from Configurations of Strategy and Structure in Subsidiaries of Multinational Corporations by Julian Birkinshaw and Allen J. Morrison, copyright 1995 by Palgrave Macmillan Ltd. Used by permission of Palgrave Macmillan Ltd. Exhibit 3.2: Classification of Control Mechanisms on Two Dimensions, from Managing the Multinationals by Anne-Wil Harzing, copyright 1999 Edward Elgar Publishing Ltd. Used by permission of Edward Elgar Publishing Ltd. Exhibit 6.26: Structural Model Conceptualization and Exhibit 6.29: Measurement Model Conceptualization, from Introducing LISREL by Adamantios Diamantopoulos and Judy A. Siguaw, copyright 2000 Sage Publications Ltd. Used by permission of Sage Publications Ltd. Every effort has been made to trace all copyright-holders, but if any have been inadvertently overlooked the publishers will be pleased to make the necessary arrangement at the first opportunity.

xvi

1 Positioning

The knowledge management challenge The emerging discussion on knowledge management during the last decade has led to a plethora of theories and models in the business literature. Despite these efforts, most companies do not make use of more than the half of the knowledge available to them (Bullinger, Wörner and Prieto 1998). The major reason given to support that argument is the lack of methods to identify and edit expert knowledge. The few methods used to utilize knowledge that do exist also tend to lack empirical validation (Easterby-Smith and Araujo 1999; Huysman 1999; Lähteenmäki, Toivonen and Mattila 2001). Reviewing the wide body of literature on knowledge management leads to the three main conclusions:

• Literature consists either of studies of organizations’ efforts to implement knowledge management or of detailed descriptions of specific knowledge tools. • Studies tend to be descriptive, largely of a non-empirical nature and tend to address fragmented topic areas rather than the entire process of knowledge transfer. • As the complex process of knowledge transfer is difficult to capture, empirical studies take a very narrow focus. The literature review below will show that the few studies explicitly investigating the intra-MNC knowledge transfer process do not provide a comprehensive model that is based on empirical findings. The notion of knowledge transfer effectiveness, in particular, is hardly ever addressed. 1

2 Effective Knowledge Transfer in MNCs

Many authors have recently outlined the singular importance of knowledge transfer as a challenge in managing knowledge, and the issue becomes especially critical in multinational corporations (MNCs). A major competitive advantage of MNCs is their ability to exploit locally created knowledge worldwide (Kogut and Zander 1995; Nohria and Ghoshal 1997; Gupta and Govindarajan 2000). As MNCs aim to replicate their success across borders, they ‘will need to focus not just on “what” they know, but “how” they gain that knowledge and diffuse it throughout the enterprise’ (Riesenberger 1998, p. 97). Despite the fact that knowledge dissemination across locally dispersed units strongly impacts MNCs’ practices and has thus to be aligned with the companies’ overall strategy, most studies ignore the international dimension and largely fail to research the ultimate effectiveness of intra-MNC knowledge transfers.

Aim of the research The aim of this study is to investigate knowledge transfers taking place between locally dispersed MNC units. In order to exploit the organization’s knowledge stock and support knowledge creation, functional units of MNCs have to share knowledge across organizational entities. This implies that such companies have to be able to transfer knowledge within organizational networks characterized by separation through time, space, culture and language. Viewing knowledge as a corporate resource illustrates the relevance of the international strategy literature: Perlmutter and Heenan (1979), Porter (1980), Bartlett and Ghoshal (1987), Prahalad and Doz (1987), Asakawa (1995) and many others have focused on intra-company transfers and how MNCs attempt to optimize sourcing strategies in terms of location advantages and economies of scale. All of these researchers address the central problem of an organization in a setting of physical separation through time and space, and separation of key members by culture and language. Thus, as Grant (1996, p. 118) puts it: ‘Many current trends in organizational design can be interpreted as attempts to access and integrate the tacit knowledge of organizational members while recognizing the barriers to the transfer of such knowledge.’ It has always to be recognized that the human capability to capture and understand complex facts is rooted in a cultural setting and thus tends to differ across cultural areas. Organizational functions which are dependent on the cultural context and have to cooperate across locally dispersed units lend themselves especially well to an investigation of knowledge transfers in MNCs.

Positioning 3

Beyond the general recognition that cultural differences are likely to impinge on the success of international knowledge transfer, concrete problems emerging in cross-cultural knowledge transfer are rarely addressed in the literature. Current research – discussed in detail in the literature review below – indicates the absence of an unifying framework that could serve as a basis for a research agenda on intra-MNC knowledge transfers. Responding to the need for such a framework, this study aims to develop a conceptual model of knowledge transfer. Based on this model, research hypotheses are advanced and subsequently tested through conducting and analysing the results of a large-scale empirical survey. The study is structured in six main areas.

Knowledge and the MNC First, the general issue of knowledge and knowledge management is presented and theoretical concepts which are relevant for the study highlighted (Chapter 2). Special focus is placed on the conceptualization of the knowledge transfer process and the relevance of knowledge management in MNCs is discussed. The in-depth literature review reflects selected strategic and organizational topics discussed in the research, and a categorization of relevant knowledge transfer studies is given.

Knowledge-based determinants of MNC strategic configuration Then the theoretical building blocks which have a bearing on intra-MNC knowledge transfers are presented and discussed (Chapter 3). Four sections focus on headquarters–subsidiary relationships; strategic mandates, coordination and control; organizational capabilities; and contingency factors.

A model of knowledge transfer in MNCs The research fields discussed in the preceding literature review are drawn together to propose a conceptual model of intra-MNC knowledge transfer (Chapter 4). Based on this model, research hypotheses are developed.

Research design and methodology The methodology section then sheds light on how the survey is conducted (Chapter 5). It explains the general research challenges, the sampling and data collection process and the operationalization of constructs.

4 Effective Knowledge Transfer in MNCs

Analysis and results The analysis (Chapter 6) is split into the investigation of the particular constructs and the structural equation model where the hypotheses are tested.

Conclusion, limitations and implications The conclusion (Chapter 7) summarizes the main findings and tries to integrate them into current research by demonstrating managerial implications and avenues for future research.

2 Knowledge and the MNC

Introduction to knowledge management: a conceptual background The question of how knowledge should best be defined has been the subject of a lively epistemological debate (Shin, Holden and Schmidt 2001). An examination of the various perspectives on the definition of knowledge and their implications for knowledge management forms a useful starting point, enabling researchers and practitioners alike to understand the directions of knowledge management research and the approach taken in this study. First, some concepts of knowledge and how this term is distinguished from data and information are presented. At this stage, it is useful to introduce the dichotomies of tacit and explicit and individual and organizational knowledge. The theoretical approaches to knowledge management are then discussed, starting with an overview of several key definitions. To illustrate that knowledge can be managed successfully only when several processes are leveraged, the concept of the ‘knowledge management value chain’ is introduced. The particular link of the value chain which is the subject of this study is analysed in more detail, and some concepts about how to depict knowledge transfer are presented.

Conceptualizing knowledge Data – information – knowledge Although numerous definitions can be found in the literature, researchers seem to agree on the fact that ‘data, information and knowledge are not interchangeable concepts’ (Davenport and Prusak 1998, p. 1). Since the 5

6 Effective Knowledge Transfer in MNCs

intention is not to give an exhaustive list, only those positions are presented which are regarded as important in understanding the approach to knowledge management in this study.

Data.

Data can be defined as a set of objective facts. They are structured records without information of how to use them in a given context. Data consist of signs and are the raw material to be processed, but they give no hint on how to do so and are thus of limited use (Willke 1998). From the perspective of systems theory, there are no data per se but only data which are constructed by perception. Besides, to be existent data have to be codified – e.g. in numbers, language, or pictures (Willke 1998). In modern organizations, data are usually stored in some sort of information technology (IT) system. As data are only raw material for the creation of information, large quantities of data without any information about their importance or irrelevance can create problems: ‘more data is not always better than less’ (Davenport and Prusak 1998, p. 2f ).

Information.

Information can be defined as data with significance (Kriwet 1997, p. 81; Davenport and Prusak 1998, p. 4). Hence, data considered valuable by a user constitute information. Data in one context may be relevant information in another (Kriwet 1997). Thus, in order to be information, data has to be provided with a meaning which is specific for and dependent on the respective system (Willke 1998).

Knowledge. Knowledge combines various pieces of information with an interpretation and meaning (Nevis, DiBella and Gould 1995; Kriwet 1997). It is created by the target-oriented combination of information, includes a component of subjectivity, insecurities and paradoxes and is subject to ambiguity (Wagner 2000, p. 37). While information derives from data, knowledge derives from information (Davenport and Prusak 1998). While information is a static concept, knowledge is constantly changing. And while information is descriptive and explicit, knowledge includes a normative component and can be explicit or tacit. In line with Plato, Nonaka and Takeuchi (1995) define knowledge as ‘justified true belief’. It is created if somebody makes sense out of a new situation by holding justified beliefs. Sveiby (2001, p. 2), however, defines knowledge as ‘the capacity to act (which may or may not be conscious)’. What all these different approaches seem to have in common is that knowledge is located at the top of the hierarchical structure (Shin,

Knowledge and the MNC 7 Exhibit 2.1

The continuum of data – information – knowledge

Data

Unstructured Isolated Context-independent Low behavioural control Signs Distinction

Information

Knowledge

Structured Embedded Context-dependent High behavioural control Cognitive behavioural patterns Mastery/Capability

Source: Based on Probst, Raub and Romhardt (1999, p. 38).

Holden and Schmidt 2001). Probst, Raub and Romhardt (1999, p. 38) describe the transition from data via information to knowledge as a continuum, an approach especially applicable to an investigation of knowledge transfers in MNCs (Exhibit 2.1). It can be seen as the lowest common denominator, which has to be agreed upon as managers operating in different contexts are likely to hold divergent perceptions of these terms. If information is to be transformed into knowledge, people have to make this transformation happen by comparing the information in different situations, through the combination of various bits of knowledge, conversation with others about the information, or assessing the consequences of the information for decision-taking (Davenport 1998). In organizations, data can be found in records, and information in messages, whereas knowledge is embedded in documents or databases, in organizational processes, routines and norms and is obtained from individuals, groups, or organizational routines either through structured media or through person-to-person contact (Davenport 1998).

Classifications of knowledge On the basis of the distinction between the different cognitive levels of data, information and knowledge, a useful dichotomy seems to be

8 Effective Knowledge Transfer in MNCs

‘knowing what’ (Wissen, savoir-connaître) and ‘knowing how’ (Können, savoir-faire). While the former refers to procedural types of knowledge – i.e. know-how, the latter can be called ‘declarative knowledge’ – i.e. know-what (see also Kogut and Zander 1993; Simonin 1999b; Gupta and Govindarajan 2000; Becerra-Fernandez and Sabherwal 2001). This distinction is vital to the knowledge transfer process as it is decisive for the choice of media and storage devices. It is also important to distinguish between the necessity of ‘higher-order’ knowledge in contrast to mere data for managing certain organizational tasks. Other classifications distinguish between specific and general knowledge or divide knowledge into the three aspects (Hedlund 1994, p. 75):

• Cognitive knowledge • Skills • Knowledge embodied in artifacts, e.g. products. Cognitive knowledge comes in the form of mental constructs. It resides in the minds of people and is also denominated ‘brainware’ (Bennett and Gabriel 1999, p. 216). Skills are competences and capabilities. Shin, Holden and Schmidt (2001), however, suggests that we should distinguish between a school of thought that regards knowledge as an object (cf. Zack 1999; Tenkasi 2000) and another that defines knowledge as a application-related process (cf. Kogut and Zander 1993; McDermott and O’Dell 2001).

Explicit and tacit knowledge Knowledge is generally distinguished along two dimensions which go back to the philosopher Michael Polanyi (1966), who wished to criticize positivist science. The distinction refers to the kind of knowledge, and differentiates between explicit or articulated and tacit or implicit knowledge. Polanyi observed that skills could be exercised without the performer being able to fully account for their cognitive basis. He elaborated a theory according to which all actions included tacit and explicit elements of knowledge and that it was especially hard to articulate the tacit elements, and consequently to pass them on to others. Today, most researchers base their theories on this distinction of tacit and explicit knowledge.

Knowledge and the MNC 9 Exhibit 2.2

Tacit and explicit knowledge

Tacit knowledge (Subjective)

Explicit knowledge (Objective)

Knowledge of experience (body)

Knowledge of rationality (mind)

Simultaneous knowledge (here and now)

Sequential knowledge (there and then)

Analogue knowledge (practice)

Digital knowledge (theory)

Source: Based on Nonaka and Takeuchi (1995, p. 36).

Explicit knowledge. This consists of some systematic language and is codified through words, numbers and codes (Hedlund 1994). This codification makes it amenable to transfer (Riesenberger 1998). Tacit knowledge. This is non-verbalized, intuitive and unarticulated (Hedlund 1994), depends on the experience of the individual, includes beliefs and emotions (Nonaka and Takeuchi 1995; Riesenberger 1998), personal skills and acquired knowledge (Bennett and Gabriel 1999). On the organizational level, tacit knowledge is embedded in organizational routines. Because of its implicit nature it is difficult to formalize and to transfer, but it is precisely this experience-based tacit knowledge which – because of the difficulty in imitating it – creates the basis for a sustainable advantage (Zack 1999). The most prominent model building on the dichotomy of tacit and explicit knowledge is that of Nonaka and Takeuchi (1995). In their work, the authors draw the distinctions shown in Exhibit 2.2. The concept can be criticized by referring back to Polanyi’s (1966) original ideas. In fact, Polanyi states that every piece of knowledge contains both tacit and explicit elements. Researchers have recently recognized this fact and have concluded that the importance for firms lies in their ability to articulate the tacit elements. Hedlund (1994) has emphasized the significance of firms as articulation machines. Based on this insight, Hakanson (2003, p. 9) elaborates on the power and logic of articulation, stating that ‘it seems that most tacit skills of economic interest are at least potentially articulable’. Critics of Nonaka and Takeuchi’s clear distinction also state that the boundary between tacit and explicit knowledge is rather blurred and flexible and that tacit knowledge may be overemphasized (Shin, Holden and Schmidt

10 Effective Knowledge Transfer in MNCs

2001). It is important to note that the value of tacit knowledge does not come from the fact that it cannot be articulated but that it has not been articulated yet.

Organizational and individual knowledge According to the ontological level of the knowledge bearer, Hedlund (1994) distinguishes among the levels of individual, group, organization and inter-organizational domains. Individual knowledge reflects individual experience and constitutes the basis for the development of organizational knowledge. Organizational knowledge is embedded knowledge and comprises belief systems, collective memories, references and values. It ‘resides in the relations between individuals, and is therefore more than the sum of individual knowledge bases’ (Kriwet 1997, p. 83). The inter-organizational domain comprises suppliers and customers (Hedlund 1994). Seen from an even broader perspective, the term ‘social knowledge’ addresses knowledge residing within groups of people. The tension between individual and organizational knowledge is especially critical to the firm as a knowledge integrating institution. As such, knowledge has to be managed as resource.

Theoretical concepts of knowledge management Definitions of knowledge management Apart from any interest in knowledge for economic theory building, there is the question of how to manage knowledge in organizations. A multitude of definitions for knowledge management can be found, and Exhibit 2.3 presents a small choice of how prominent researchers in the field have conceptualized knowledge management. Some authors differentiate between a human resources-oriented approach to knowledge management and a technology-oriented approach ( Jacob and Ebrahimpur 2001). However, recent research has predominantly applied the so-called ‘integrative’ approach, based on a fundamental belief that human and technological components have to be combined in order to reach an optimal result. As the potential of knowledge management to create and maintain a competitive advantage can be realized only by an interplay between technological systems and aligned structures of human communication, the integrative approach is also used in this study. Although the research agenda in knowledge management seems to be rather fragmented and, as Gallupe states (2001, p. 62), there appears to be no systematic framework to provide guidance, some studies have put

Knowledge and the MNC 11 Exhibit 2.3

Definitions of knowledge management

Source

Definition

Birkinshaw (2001, p. 12)

Knowledge management can be seen as a set of techniques and practices that facilitates the flow of knowledge into and within the firm.

Buckley and Carter (1999, p. 82)

Knowledge management contains ‘the internal mechanisms for coordination, that is, for pooling the key information garnered by managers whose task it is to monitor external volatility and discover new opportunities’.

Davenport et al. (2001, p. 117)

Knowledge management is ‘the capability to aggregate, analyze, and use data to make informed decisions that lead to action and generate real business value’.

Demarest (1997, p. 379)

‘knowledge management is the systematic underpinning, observation, instrumentalization, and optimization of the firm’s knowledge economies’.

Leonard-Barton (1995, p. xiii)

‘The primary engine for the creation and growth and of technological capabilities is the development of new products and processes, and it is within this development context that we shall explore knowledge management . . . The management of knowledge, therefore, is a skill, like financial acumen, and managers who understand and develop it will dominate competitively.’

Malhotra (1998, p. 59)

‘Essentially, it embodies organizational processes that seek synergistic combination of data and information processing capacity of information technologies, and the creative and innovative capacity of human beings.’

Stewart et al. (2000, p. 42)

‘The premise is that knowledge assets, like other corporate assets, have to be managed in order to ensure that enterprises derive value from their investment in knowledge assets.’

Tsoukas and Vladimirou (1996, p. 973)

Knowledge management ‘is the dynamic process of turning an unreflective practice into a reflective one by elucidating the rules guiding the activity of the practice, by helping give a particular shape to collective understandings, and by facilitating the emergence of heuristic knowledge’.

an effort into the formulation and categorization of research. Some studies, on the other hand, attempt to explain and theorize the entire knowledge management system and aim to provide an insight into the whole knowledge management process. Such management processes, each including various forms of knowledge, are very complex, and

12 Effective Knowledge Transfer in MNCs

capturing an organization’s entire knowledge management system can be very difficult. Some authors, however, have made an attempt to categorize knowledge management systems. Typical of this genre is the work of Birkinshaw (1999), Hansen, Nohria and Tierney (1999), Hong (1999), McAdam and McCreedy (1999), Zack (1999), Bloodgood and Salisbury (2001), Earl (2001), Gallupe (2001), and Holsapple and Joshi (2001). The way authors attempt to derive a categorization differs considerably. Romhardt (2000), for example, approaches the problem from a systems theory perspective and Tsoukas and Vladimirou (1996) apply a constructionist view. Von Krogh and Venzin (1995) suggest a useful systematization of research streams in knowledge management:

• • • • • • • • •

Knowledge management models Knowledge, conversation, cooperation Measurement and assessment of knowledge Knowledge transfer Knowledge structures Epistemology Knowledge and information technology Knowledge and power Knowledge, networks, and innovation.

Generally, different assumptions are made about the boundaries of knowledge management. They range from comprising only one individual to the integration of business partners and external institutions into the knowledge management system. In this study, knowledge management is understood as an organization-wide activity, and the leveraging of knowledge between different organizational units (i.e. headquarters and subsidiaries) in different countries forms the centre of analysis.

The knowledge management value chain While there is a general debate on whether to conceptualize knowledge as an object or as a process, most researchers seem to agree on the description of knowledge management as a process (Grant 1996). Even so, researchers have presented different links to build the value chain. Hong (1999), for example, divides it into four stages:

• Knowledge acquisition • Information distribution

Knowledge and the MNC 13

• Information interpretation • Organizational memory. A slightly different approach is taken by Hedlund (1994), who distinguishes between the storage, transfer and transformation of knowledge. Shin, Holden and Schmidt (2001) aim to consolidate different contributions (Exhibit 2.4). The first link of the knowledge management value chain in these depictions is the creation of knowledge. Obviously, the prerequisite for knowledge management is the identification of the source, in other words recognizing that unique knowledge exists in an internal source (e.g. an organizational unit) or an external source (e.g. a research institution, a customer, etc.). In the latter case, this link could also be named ‘knowledge acquisition’. It also has to be determined whether the knowledge created and embedded in the context of one country or one subsidiary is of value for the rest of the organization (Kriwet 1997; Gupta and Govindarajan 2000). While deeply ingrained knowledge is primarily relevant to the stakeholders in a specific project and of minor value for other subsidiaries, generalizeable knowledge provides insights with relevance to various subsidiaries. Storage means that knowledge has to be maintained in some kind of individual or organizational memory. This link has to be customized to the type of knowledge, the potential recipient and, thus, the search process. In order to permit storage, not only the existence of such a memory

Exhibit 2.4

The knowledge management value chain Vision and Strategy

KM Value Chain Holzner and Marx (1979) Pentland (1995) Nonaka and Takeuchi (1995) Demarest (1997)

Creation

Storage

Consciousness Construction

Organization

Creation Construction

Identify

Storage

Access

Daal, Hass and Weggeman (1998) Creation Davenport, De Long and Beers (1998) Creation Liebowitz (1999)

Extension

Embodiment Draw-up

Capture

Store

Source: Shin, Holden and Schmidt (2001, p. 341).

Distribution

Application

Transformation

Implementation

Distribution

Application

Dissemination

Application

Dissemination

Use

Dissemination

Apply

Transference

Asset Management

Share

Apply

Evaluate

Sell

14 Effective Knowledge Transfer in MNCs

device but also the willingness of employees to detach the knowledge gained from the individual, is vital. This release of knowledge is likely to have effects on power relations. Power relations may also be a concern where knowledge distribution is concerned. Who should be given access to knowledge, and which knowledge is relevant to whom? Such questions are likely to emerge in the distribution stage; this link of the knowledge management value chain is referred to as distribution, transformation, dissemination, transference, or sharing. But problems often emerge, since different types of knowledge are not easy to distribute – a characteristic Von Hippel (1994) calls ‘stickiness’ (see also Szulanski 1995) – and one has to be aware that transfer always includes some form of transformation as well. It then depends on the type of knowledge to decide which strategy should be applied to manage the transfer process (Davenport 1998). Due to codification, knowledge arriving at the receiver will not be exactly identical with the knowledge of the sender. Another critical point is the choice of transmission media to channel knowledge. The details of knowledge transfer will be discussed in depth in the following chapters. The access and transfer of knowledge does not yet ensure that knowledge will be used. Therefore, an application phase is included in the knowledge management value chain. As the goal of transferring knowledge is to increase value, knowledge transfer is successful only if the newly absorbed knowledge leads to changes in behaviour or the creation of new ideas (Davenport 1998). Knowledge application ultimately seeks to locate the source of competitive advantage (Shin, Holden and Schmidt 2001). Last but not least, it is essential that the knowledge management value chain is strategically driven in order to realize the objectives of an organization. The central assumption is that each knowledge management system has to be intricately interwoven with corporate strategy, structure and processes. Vision and strategy form a ‘control circuit’ for the knowledge management value chain.

The process of knowledge transfer As mentioned above, the process of managing the different types of knowledge is split up into several parts in the knowledge management value chain, and knowledge transfer forms just one stage. The emphasis here is on the ‘distribution’ stage of the knowledge management value chain. It has already been pointed out that the knowledge transfer process is quite complex in itself, and on a broad level, two general approaches

Knowledge and the MNC 15

can be distinguished: the communication model and the knowledge spiral model (Inkpen and Dinur 1998).

Knowledge transfer as a communication model The classical communication model (Shannon and Weaver 1957) depicts messages’ flow from a sender to a recipient. While parts of the message are likely to be transformed or even destroyed by ‘noise’, it is important that the core of the message arrives at the sender. Two critical stages are the encoding phase, when the sender packages the message to fit the media, and the decoding phase, when the receiver has to decipher it again. A simplified model depicting only the building blocks of the Shannon and Weaver (1957) theory is shown in Exhibit 2.5. Szulanski (1996) was among the first to introduce this concept into the knowledge management literature, conceptualizing knowledge transfer as a message transmission from a source to a recipient in a given context. Knowledge transfer is thus seen as a dyadic exchange of knowledge between source and recipient. Inkpen and Dinur (1998) extended this model, and identified four groups of related factors:

• • • •

Source-related factors Recipient-related factors Factors relating to the relationship and distance between the two units Factors related to the nature of the knowledge transferred.

Four stages are then necessary for the transfer process:

• Initiation: transferred knowledge is recognized. • Adaptation: knowledge is changed at the source location to the perceived needs of the recipient. Exhibit 2.5

A simplified communication model

Context

Context

SENDER

RECIPIENT

‘NOISE’ Source: Based on Shannon and Weaver (1957).

16 Effective Knowledge Transfer in MNCs

• Translation: alterations occur at the recipient unit as a part of the general problem solving process of adaptation to the new context.

• Implementation: knowledge is institutionalized to become an integral part of the recipient unit. Although this process is quite easy to understand at first glance, it becomes quite complex at the organizational level when eligible senders and recipients have to be defined. To shed some light on this issue, Sveiby (2001, p. 349) identified nine different knowledge transfers, which are depicted in Exhibit 2.6. For this study, four of these knowledge transfer processes are relevant. It is important to note that a critical feature of modern knowledge management is the time-lag between sender and recipient. While Shannon and Weaver’s (1957) communication model builds on a nearly simultaneous transfer, the knowledge transfer process can be interrupted and knowledge can be stored in the meantime. From a conceptual point of view, however, it is clear that every medium

Exhibit 2.6

Nine knowledge transfers 2. Knowledge transfers from individuals to external structure

3. Knowledge transfers from external structure to individuals 6. Knowledge transfers within external structure

External structure

7. Knowledge transfers from external to internal structure 8. Knowledge transfers from internal to external structure

9. Knowledge transfers within internal structure Source: Sveiby (2001, p. 349).

Individual competence

$

Internal structure

1. Knowledge transfers between individuals

4. Knowledge transfers from individual competence to internal structure

5. Knowledge transfers from internal structure to individual competence

Knowledge and the MNC 17

represents an intermediate storage device, so the approach is not that revolutionary:

• Transfers between individuals: How can we improve the transfer of knowledge1 between people in our organization? This kind of transfer takes place in communication between employees, and here, the simplified model of communication can be applied. Issues such as trust and exposure to different kinds of expertise in the company are important. • Transfers from individual competence to internal structure: How can we improve the conversion of individually held competence to systems, tools and templates? When knowledge of individual competences is stored in repositories, it becomes accessible in the organization’s structure and can be shared by everyone. This definition of ‘transfer’ comprises only the first half of the communication model, namely from the sender to the media. • Transfers from internal structure to individual competence: How can we improve individuals’ competence by using systems, tools and templates? Employees’ capacity to act should be improved by knowledge accessible in the internal system. An important issue at this stage of transfer is the interface between humans and knowledge storage systems. Here, the second half of the communication process is depicted, the transition from the media to the recipient. • Transfers within the internal structure: The strategic question is how the organizations’ systems, tools and processes can be integrated effectively. All transfer processes involving external senders or recipients are excluded from this study. The transfer processes ‘between individuals’ and ‘with the internal structure’ is a necessary phase of the transfer process, but is not highlighted. The first becomes almost irrelevant in global organizations where sender and recipients are characterized by dispersed locations and time-zones and, as stated above, at least for a short time knowledge is stored in a medium. The latter explains only the management of knowledge within the storage system, and thus exceeds the focus of this investigation. Nevertheless, it is important to explain Sveiby’s intentions in order to visualize the complexity of organizational knowledge transfer. The idea of adapting the communication model to explain knowledge transfers not only between individuals but also between entities

18 Effective Knowledge Transfer in MNCs

in global organizations is extremely useful for the development of the conceptual model in this study, and will be taken up later.

The spiral of knowledge Nonaka and Takeuchi (1995) attribute the success of Japanese companies to their effectiveness in creating knowledge. In their pioneering work the authors propose a model of knowledge creation – the spiral of knowledge (Exhibits 2.7). They build on the epistemological dimension of explicit and tacit knowledge and on the ontological dimension which symbolizes the number of people involved in the process. The core assumption of this model is that tacit knowledge has to be mobilized and converted. This means that the model does not only explain knowledge creation but describes processes of transferring knowledge, specifically the so-called ‘conversion’ processes. A spiral is created when the conversion of tacit and explicit knowledge results in higher epistemological and ontological levels. Nonaka and Takeuchi (1995) identify four specific conversion processes:

• Socialization (tacit to tacit): Individuals exchange tacit knowledge without codifying it during the transfer phase, e.g. shared mental models, technical skills. • Externalization (tacit to explicit): In this process, tacit knowledge is made explicit by codifying it in the form of metaphors, analogies, Exhibit 2.7

The knowledge spiral

Tacit knowledge

Tacit knowledge

To

Explicit knowledge

Socialization (Sympathized knowledge)

Externalization (Conceptual knowledge)

Internalization (Operational knowledge)

Combination (Systemic knowledge)

From

Explicit knowledge

Source: Nonaka and Takeuchi (1995, p. 62).

Knowledge and the MNC 19

hypotheses, models, etc. Through such a transformation personal knowledge can be made available on a corporate-wide basis. Externalization is thus the most important process for knowledge creation. • Combination (explicit to explicit): Through combination, concepts are systematized within a knowledge system. Existing elements of knowledge are combined in order to create new explicit knowledge. Several media support combination, e.g. documents, meetings, phone calls. • Internalization (explicit to tacit): The conversion of explicit into tacit knowledge is called ‘internalization’, meaning that incoming knowledge is integrated into an individual’s knowledge base. By combining these two models – the communication model and the knowledge spiral – the four knowledge conversion processes can also be seen as singular transfers between a sender and a recipient. This leads to the conclusion that every sending and every receiving unit has to engage in some of these processes in order to process the inflowing or outflowing knowledge. As described in the communication model, it has to be remembered that no knowledge transfer is context-free. At a later stage, this argument will be recaptured and used to explain knowledge transfers in global organizations.

The relevance of knowledge management in the MNC The outstanding relevance of knowledge management for MNCs is now considered. Theoretical perspectives, such as the conceptualization of the MNC, the growing role of knowledge management in organizational design and the transition from a resource-based view to a knowledge-based view of the firm, form the base of this study. More specifically, knowledge transfers in the MNC are discussed. Having outlined the various levels on which knowledge can be transferred in the MNC – and the focus of this research – the general transferability of knowledge as a resource is briefly addressed. This is followed by an overview and a categorization of the major studies which are relevant for the current research.

The conceptualization of the MNC The overall aim of the theory of the MNC is to explain the level and pattern of the foreign value-added activities of firms (Dunning 1993). In this context, a multitude of definitions of the MNC are given in the international business literature and the question of the decisive criteria

20 Effective Knowledge Transfer in MNCs

of the MNC is still unresolved (see also Rugman and Brain 2003). Almost all authors agree that, in comparison to a domestic company, an MNC has fundamental distinguishing characteristics – with regard to social and cultural norms, government regulations, customer tastes and preferences as well as social and economic structures, the MNC faces various conflicting demands and pressures in its multiple host countries. While domestic companies are confronted with competition within a single market, the MNC is challenged by complex competitive strategies taking place on an international level and requiring extensive coordination efforts. The MNC also manages a complex organizational structure and management system that requires control over its product and its functional and geographic diversity, including linguistic and cultural aspects (Bartlett and Ghoshal 1989). This study follows Bartlett and Ghoshal’s (1989) approach; they classify firms as MNCs if they fulfil two requirements:

• They need to have substantial direct investments in foreign countries, not just an export business. (Vernon 1966 set the – more or less arbitrary – requirement to be engaged in six overseas production operations.) • They need to be engaged in the active management of these offshore assets rather than simply holding them in a passive financial portfolio. Bartlett and Ghoshal regard the second issue as the key differentiating characteristic of an MNC, as it highlights the importance of strategic and organizational integration: ‘What really differentiates the MNC is that it creates an internal organization to carry out key cross-border tasks and transactions internally rather than depending on trade through the open markets’ (Bartlett and Ghoshal 2000, p. 3). This extends to the notion that, in general, knowledge can be transferred more effectively and efficiently through internal than through external market mechanisms (Kogut and Zander 1992). The aim to internalize knowledge flows is thus seen as a primary motivation for foreign direct investment (FDI), which results in a view of the MNC as a repository of valuable knowledge that can be exploited either through the development of new products or through the dissemination of existing products to new locations. The question of how to structure and to manage the relationships between the headquarters and its foreign subsidiaries is one of the core issues when dealing with the study of MNCs (Bartlett and Ghoshal 1989; Birkinshaw et al. 2000; Paterson and Brock 2002). Traditionally, organizational forms of MNCs have been conceptualized in terms of regional scope, product, functional divisions, or matrix structures. By focusing

Knowledge and the MNC 21

on the control of world-wide operations, factors such as the extent and the nature of these operations, the national origin, information-processing capabilities and the mind-set of senior managers can be identified as determinants of the MNC’s choice of a structure (Malnight 2001). Buckley and Casson (1998, p. 22) argue that: ‘Efficient information processing is crucial to cope with the resultant increase in the complexity of decisionmaking. This has important implications for the organizational structure of the MNE.’ Although empirical studies have not yet found clear support for a welldeveloped typology of MNCs, some concurrence can be found in the recent literature about the strategy and structure of MNCs. With regard to the description of strategy, all authors implicitly or explicitly refer to a continuum of integration, coordination, or globalization advantages versus differentiation, responsiveness, or localization advantages (Harzing 2000b). In recent international business research, the conceptualization of the MNC as a learning organization which has continuously to upgrade its competitive advantage has come to the fore. In this context, two schools of thought have developed (Sölvell and Zander 1995):

• First, the Home-Based Model, which builds mainly on the ideas of Porter (1990) and Sölvell, Zander, and Porter (1991), proposes that, in spite of globalization, the MNC makes a clear distinction between core and peripheral activities and remains dependent on local environments. • Second, extending the frameworks of Perlmutter and Heenan (1979) and Bartlett (1984), the Heterarchical Model stresses that ‘traditional headquarter functions are geographically dispersed and none of the dimensions – country, product, or function – is uniformly superordinate in the process of generating new firm-specific knowledge or in strategy formulation and implementation’ (Sölvell and Zander 1995, p. 25). An in-depth discussion of the different models would go beyond the scope of this research (for more insight see Sölvell and Zander 1995), but it has to be emphasized that this study builds on the latter school of thought. Its implications for structuring the MNC will how be discussed. In this study, the firm is conceptualized as a network of units. In this network, units have strategic mandates and thus access and transfer knowledge from different positions (cf. Gupta and Govindarajan 1991; Asakawa 1995; Tsai 2001). Although their network positions differ, the corporate embeddedness (Granovetter 1985) of organizational units in this network provides a basic social context which is common for all units.

22 Effective Knowledge Transfer in MNCs

The Integrated Network Model developed by Bartlett and Ghoshal (1988) ‘models the MNC as a geographically-dispersed set of value-adding activities, each activity of which can be viewed as a semi-autonomous entity, with ownership ties, normative links and certain obligation to head office’ (Paterson and Brock 2002, p. 323). The network approach combines the characteristics of integration, responsiveness and world-wide innovation and learning simultaneously, with some functions coordinated at the global level while others remain local (Harzing 2000a). Due to its ability to deal with complex, large MNCs, the network approach has become one of the preferred designs when conceptualizing MNCs. Bartlett and Ghoshal (1988) propose their theory of the Integrated Network Model and the Transnational Organization as a response to managerial and environmental challenges. Many of these network characteristics have also been emphasized by Hedlund (1994), conceptualizing the MNC as a heterarchy, also called the N-Form, in direct antithesis to the traditional Hierarchy Model of the organization. Other similar models include the Multifocal Organization developed by Doz (1986), the Horizontal Organization developed by White and Poynter (1990) and the Metanational Organization by Doz, Santos and Williamson (2001). Within the Transnational Organization, all subsidiaries and the headquarters are part of an interdependent network. Organizational units are seen as a unique source of skills, capabilities and knowledge. Their contributions to the interdependent network of worldwide operations therefore differ considerably and learning has to take place on a global basis. This implies that headquarters do not necessarily play a dominant role in the organization, and the firm’s ability to react flexibly to fast changing market conditions is increased. The Heterarchy Model of the organization challenges the assumptions of the traditional Hierarchy Model that are unable to reflect the full complexity of the modern MNC and its peripheral operations (Hedlund 1994). In this sense, ‘hierarchy’ can be characterized by prespecified and stable relationships, instrumentality and additive influence of parts, unidirectionality and universality and knowledge and people hierarchies. In contrast, the Heterarchy Model implies many different kinds of subsidiary centres with loose coupling between the units and normative control systems and brings in the necessary multidimensionality of the organizational structure. Analogies to the Transnational Organization include the dispersion of resources, skills and decision making throughout the organization rather than concentration at the top. Activities are simultaneously coordinated along the three primary dimensions – knowledge, action and position of authority – rather than product,

Knowledge and the MNC 23

geography and function. Hedlund (1994) argues that cohesion and protection from anarchy is achieved through normative goal-directed integration comprising shared objectives, knowledge and a common organizational culture as important mechanisms. The increasing complexity of differentiated subunits within conceptualizations of the MNC can be viewed as a response to the multiple, fast-changing environments the MNC faces. Structural variations across geographic units reflect different cultural and economic environments and result in differences in strategic roles and organizational structures. Functional variations reflect differences in strategic objectives. Common structural patterns are thus found among centralized upstream operations like R&D – to take advantage of economies of scale and scope – and decentralized downstream operations like marketing and sales – to respond to differences in national market requirements (Malnight 2001).

Knowledge management and the structure of the MNC It thus becomes evident that knowledge management – specifically knowledge requirements and distribution – influence the structure of the organization (see also Hedlund 1994). According to Galbraith’s (1972) information-processing view of organizational design, and in response to intensifying global competition, increasing the firm’s informationprocessing capacity becomes a necessity if the MNC is to cope with the challenge of growing environmental and organizational complexity (see also Ghoshal, Korine and Szulanski 1994). As already indicated, a critical proposition behind the network approach is that the interdependence of the world-wide units creates additional value through extensive cross-border exchanges. The demand for acquiring, collocating and interpreting environmental and competitive information on a global basis is an important characteristic of the network model. Kogut and Zander (1992) argue that the firm’s combinative capacity reflected in the Integrated Network Model is the primary determinant of the MNC’s superiority vis-à-vis open market transactions. The MNC has the ability to generate additional strategic advantages by combining distributed knowledge, resources and capabilities. Expanding efficiency and scale, accessing specialized and locally embedded resources, enhancing innovation through operations across markets and creating operational flexibility are all aimed to respond to factors outside a firm’s direct control (Kogut and Zander 1992). All these factors refer to the knowledge-based view of the firm (see p. 25) which proposes that the building of world-wide learning capabilities to develop and rapidly

24 Effective Knowledge Transfer in MNCs

diffuse innovations is a vital source of competitive advantage for the MNC (Bartlett and Ghoshal 2000). In this context, Hedlund (1994) presents a coherent argument to explain why, in terms of knowledge management, the Integrated Network Model is superior to the traditional hierarchical and divisional view of the firm: The challenge is not to divide a given task in a way ensuring maximally efficient performance. Rather, it is to position the company so that new tasks can be initiated, often on the basis of separate knowledge pieces from different organizational units. Instead of bringing the information to the given decision point, it becomes a matter of bringing the decision to the knowledge bases. (Hedlund 1994, p. 87) Hedlund emphasizes the importance of integrating mechanisms in order to foster the combination of pieces of knowledge through dialogue and knowledge assimilation. Dialogue and assimilation, as well as the interplay between tacit and explicit knowledge, require shifting groupings of individuals while pooling from a permanent personnel pool – one of the key features of the Transnational Organization. To achieve a sufficient level of dialogue and knowledge assimilation, lateral communication is an important mechanism. In addition, redundancy and repetition of knowledge – formerly regarded as inefficient – are seen as necessary to distribute it throughout the organization. In his research about knowledge management within MNCs, Hedlund (1999, p. 6) refers to two dimensions:

• The knowledge intensity of a firm or activity, which describes the degree to which a firm is dependent on an internal supply of advanced, complex, and recent knowledge. • The knowledge extensity of a firm or activity, meaning the increased dispersion of knowledge in terms of geography, organization (between individual firms and between their departments, subsidiaries, etc.) and substantion (dispersion of knowledge over technical fields and types of knowledge) relevant to the competitive distinctiveness of the firm. The firm, conceptualized as a Heterarchical Transnational Organization, has the ambition to both intensify and extend its knowledge. The key characteristic of the MNC is its ability continuously to combine and recombine knowledge.

Knowledge and the MNC 25

From resource-based view to knowledge-based view ‘Contemporary strategy analysis has seen a shift in emphasis from the structure–conduct–performance [SCP] paradigm which emerged from industrial economics and towards theories which focus on the internal resources of individual firms as a key determinant of competitive advantage’ (Galunic and Rodan 1998, p. 1193). The resource-based view of the firm, first articulated by Wernerfelt (1984), looks at firms in terms of their resources rather than their products and aims to identify strategic options through the exploitation and the development of these resources. Among the most notable works which contributed to the establishment of this theory are Barney (1986), Prahalad and Hamel (1990) and Conner (1991). The term ‘resource’ refers to the firm’s stock of tangible and intangible assets. In the centre of the analysis are resources which are difficult to imitate and constitute the basis for the firm’s competitive advantage. In this context, every firm is regarded as a unique bundle of idiosyncratic resources and capabilities. ‘Capabilities’ or ‘competencies’ represent the organization’s collective capacity for undertaking a specific type of activity’ (Lieberman and Montgomery 1998, p. 1112). Even before the formation of the resource-based view, the international strategy literature had tended to view knowledge as a corporate resource. Perlmutter and Heenan (1979), Porter (1980), Bartlett and Ghoshal (1987), Prahalad and Doz (1987), Asakawa (1995) and many others all focused on intra-company transfers and how MNCs can attempt to optimize sourcing strategies in terms of location advantages and economies of scale. All these researchers address the central problem of organization in a setting of physical separation through time and space and separation of key members by culture and language, calling for a management of dispersed knowledge assets. Thus, as Grant (1996, p. 118) puts it: ‘Many current trends in organizational design can be interpreted as attempts to access and integrate the tacit knowledge of organizational members while recognizing the barriers to the transfer of such knowledge.’ Conner and Prahalad (1996) suggest that performance differences between firms can be traced back primarily to asymmetries in knowledge. Consequently, within the resource-based literature, the view of knowledge as a primary resource has become increasingly common (cf. Grant 1996). The emerging ‘knowledge-based view of the firm’ can be seen as an outgrowth of several streams of research: epistemology, organizational learning, resource-based view of the firm, organizational capabilities and competences, and innovation and new product development (Grant

26 Effective Knowledge Transfer in MNCs

and Baden-Fuller 1995). This view suggests that knowledge is by far the most important resource and that social networks facilitate knowledge sharing within an organization. Knowledge as a resource, however, requires organizational capabilities in order to be productive – an argument that is fully in line with the resource-based view. The knowledge-based theory of the firm is also used to identify circumstances in which collaboration between and within firms is superior to either market or hierarchical governance mechanisms which inefficiently integrate specialized knowledge. This perspective is further developed by Kogut and Zander (1992) and highlighted in the Evolutionary Theory of the firm. Kogut and Zander postulate the inefficiency of markets as means of knowledge transfer and integration, as knowledge – tacit as well as explicit – is highly specific and efficient integration must preserve the efficiencies of specialization in the acquisition and storage of knowledge which is better achieved in a firm. The firm is seen as reservoir of social knowledge that structures cooperative action. There are epistemic communities, in which discourse and coordination play complementary roles in replicating behaviours and exploring new options (Kogut and Zander 1995). Although researchers commonly agree on these characteristics of the knowledge-based view, there is insufficient consensus for it to be recognized as a theory. Grant (1996, p. 110) states that the knowledge-based view concerns issues that go far beyond the traditional concerns of strategic management and extends the knowledge-based view of the firm to:

• Explain the existence of the firm as an institution for the organization of production

• Explore the nature of coordination within the firm • Analyse organizational structure • Determine the boundaries of the firm. Grant (1996) also stresses the discrepancy of organizational and individual knowledge and insists that the roots of the knowledgebased approach have to lie in individual knowledge creation. Subsequently, organizational knowledge has to be created through the firm’s combinative capability.

Knowledge transfers in MNCs Varieties of knowledge transfer in MNCs In the MNC context, various directions and levels of knowledge transfers are possible. Although researchers assign different labels to these transfer

Knowledge and the MNC 27

activities – such as ‘knowledge diffusion’, ‘knowledge sharing’, ‘knowledge dissemination’, ‘transfer of capabilities’, ‘transfer of best practice’ – all transfers of procedural and declarative knowledge will be referred to here as ‘knowledge transfer’. The epistemological concept underlying the specific understanding of knowledge transfer in this research has been presented above (p. 14). The most obvious classification of knowledge transfers in MNCs is the one between external and internal transfers. ‘External transfer’ refers to all sources and recipients of knowledge outside the firm – i.e. other firms, institutions, customers, etc. There is a large body of literature focusing on transfers, strategic alliances between different firms, and (international) joint venture ( JVs). ‘Internal transfers’, however, concern all activities inside the global organization of the MNC. In this respect, transfers between organizational units and within organizational units can be distinguished. As mentioned earlier, this research centres on intra-organizational knowledge transfers between organizational units located in different countries. Following Gupta and Govindarajan (2000), a dyadic and nodal perspective is taken. In the centre of the analysis are hierarchical and lateral flows – i.e. between headquarters and subsidiaries and between subsidiaries and subsidiaries. In the following section, the transferability of knowledge within the MNC will be discussed in relation to the characteristics of knowledge as a resource. A literature review categorizing some relevant studies in the field is then provided.

Transferability of knowledge within the MNC As discussed previously, knowledge transfer cannot mean the ‘transfer’ of one piece of knowledge in terms of delivery from A to B. Knowledge is always context-bound and related to the overall history of the knowledge-holder. It therefore encounters the problem of the ‘hermeneutic circle’: to store knowledge in the organization it has to be abstracted from its context to a certain extent, but as soon as it is taken up by a human being it is linked to a specific context again. Therefore, the common idea of transfer is – if at all – possible only for parts of the original piece of knowledge. (The term ‘piece of knowledge’ is already problematic in itself, as it already demands uncoupling something of a person’s knowledge stock.) What is, of course, critical for the effectiveness of such transfers is that the core message is still conveyed, despite being found in a totally different mantle. The term ‘knowledge transfer’ is thus used throughout this study bearing these significant limitations in mind.

28 Effective Knowledge Transfer in MNCs

These considerations have an important effect on the view of knowledge as a corporate resource. Unlike other resources, it does not diminish in value when shared, but is actually doubled (Sveiby 2001). Although this is true on the organizational level, restrictions may occur at the personal level when power relations are concerned. The notion that ‘knowledge is power’ can have negative effects on the firm if people try to protect their knowledge and are not willing to share it. Foss and Pedersen (2002) conclude that this issue is less important in dynamic contexts, as the likeliness of gaining power through the transfer of knowledge is higher. Although the constraints of employee motivation and their willingness to share knowledge are supposed to play a role in the knowledge transfer process, earlier studies could not provide convincing confirmation (see also Foss and Pedersen 2002; Szulanski 1995). Many researchers also argue that the transfer of knowledge will depend on the type of knowledge transferred. They note that knowledge characteristics differ:

• According to industries (cf. Jacob and Ebrahimpur 2001) and subject areas (cf. Gupta and Govindarajan 1991) – i.e. technological knowledge, marketing knowledge, etc. • According to their level of creation (Foss and Pedersen 2002) – i.e. internal, network, cluster and • According to their degree of tacitness (Kogut and Zander 1993; Simonin 1999b). While the differentiation among industries and subject areas is well reflected in this research, the degree of tacitness is not found as an appropriate criterion for distinction. As argued earlier, the distinction of tacit and explicit knowledge prevalent in the majority of publications seems to be obsolete (see also Hakanson 2003), though it has to be admitted that knowledge differs in composition, which impacts on the ease of transfer. A simple differentiation criterion which seems suitable is knowledge ‘complexity’. This term can be defined as the number of independent routines and resources linked to a particular knowledge asset (see also Zander and Kogut 1995; Simonin 1999b). Grant (1996) assumes that the efficiency of knowledge transfer depends partly upon the aggregation potential of the specific knowledge – aggregation is enhanced when knowledge can be expressed in terms of common language, for example. Idiosyncratic knowledge, which is difficult to

Knowledge and the MNC 29

aggregate at a single place, is therefore hard to transfer. Doz and Santos (1997) address a similar issue when they discuss the packaging of knowledge. They argue that for every piece of knowledge to be transferred there is an appropriate form. A characteristic of knowledge opposed to the view that knowledge is an easily transferable resource is ‘stickiness’. Von Hippel (1994) and Szulanski (1995) researched this phenomenon and state that stickiness is either intrinsic to the knowledge itself or pertinent in the situation. Their argument is specifically aimed at refuting the neoclassical view that knowledge is a public good and transferable without incurring costs. They also emphasize the importance of established linkages between units to enhance knowledge transfer.

State of the art: studies on intra-MNC knowledge transfer Although there is a large number of studies discussing knowledge transfer in general, relatively little attention has been paid to the full scope of knowledge flows that are to be found within a single organizational setting (Teigland, Fey and Birkinshaw 2000). The studies in this field are quite diverse, approaching the issue from various perspectives. Some regard knowledge flows between headquarters and subsidiaries of MNCs as control or administrative mechanisms (cf. Gupta and Govindarajan 1991), others see the optimization of such flows as means for creating competitive advantage (cf. Zander and Kogut 1995; Subramaniam and Venkatraman 2001). In terms of chronological development, the investigation of technology transfers opened the avenue for a more general extension of the topic after the mid-1980s. Generally speaking, while earlier studies concentrated exclusively on technical knowledge, later studies incorporate a more comprehensive field of knowledge – namely managerial and administrative knowledge and marketing knowledge (see also Simonin 1999b). In Exhibit 2.8, the most relevant contributions addressing intraorganizational knowledge transfer are categorized as conceptual or empirical. Only studies either providing a model of intra-MNC knowledge transfer or addressing the transfer process as such are analysed in depth. Requirements or contingencies of knowledge transfer are addressed later (Chapter 3). Works which focus specifically on inter-company transfers and knowledge transfer in JVs, strategic alliances or mergers and acquisitions (M & As) are excluded from this analysis. Exhibit 2.8 shows the most important contributions, which have all been published in major academic journals, in chronological order. The type of

Roles of subsidiaries in the MNC according to knowledge flows Learning in service organizations

Determinants of the speed of knowledge transfer and imitation Conceptual model of knowledge transfer

Procedural and declarative knowledge

Not specified

Procedural knowledge

Instrumental and developmental knowledge

Technology knowledge, procedural

Best practice

Gupta and Govindarajan (1944)

Darr, Argote and Epple (1995)

Zander and Kogut (1995)

Gilbert and Cordey-Hayes (1996)

Grosse (1996)

Szulanski (1996)

Comprehensive taxonomy to barriers to intra-firm knowledge transfer

Technology transfer in service industries

Influence of knowledge flows on control mechanisms

Procedural and declarative knowledge

Gupta and Govindarajan (1991)

Aim

Type of Knowledge

Studies of intra-MNC knowledge transfers

Study

Exhibit 2.8

Quantitative Major barriers are lack of absorptive capacity, causal ambiguity and arduous source–recipient relationship

Quantitative

Acceptance is critical for technological change

Quantitative

Quantitative

Codifiability, teachability and the threat of market preemption are critical

Training and transfer of experts is vital

Quantitative

Quantitative

Differentiated knowledge flow roles are linked to differentiated processes and systems within the MNC Knowledge is found to transfer within the same franchise but not across others

Quantitative

Method

Formal and informal administrative mechanisms reflect knowledge flows

Main Findings

30

Conceptual

Quantitative

Phases of knowledge flows and taxonomy of influencing factors

Strong ties facilitate the transfer of complex knowledge

Identification and transfer of best practices Categorization of knowledge transfers and the influencing factors Effects of the strength of inter-unit ties Knowledge transfer and product development performance MNC–internal knowledge sharing

Not specified

Individual and organizational knowledge

Procedural knowledge

Not specified

Not specified

O’Dell and Grayson (1998)

Von Krogh and Köhne (1998)

Hansen (1999)

Hoopes and Postrel (1999)

Dyer and Nobeoka (2000)

Network-level knowledge sharing processes are vital

‘Glitches’ as a measure of knowledge transfer performance

Best practice transfer as a benchmarking process

Quantitative

Qualitative

Qualitative

Conceptual

Social capital accounts for the superiority of firms over markets to create and transfer knowledge

The role of social capital in relation to intellectual capital within the organization

Intellectual capital

Qualitative

Context similarity and the nature and extent of knowledge transfer mechanism employed are the key success factors

Nahapiet and Ghoshal (1998)

The impact of context on knowledge transfers

Conceptual

Organizational knowledge

Knowledge management gets eventful in different cultures and contexts and has to be packaged differently

Inkpen and Dinur (1998)

Knowledge transfer in geographical dispersion and context differentiation

Not specified

Doz and Santos (1997)

31

Qualitative

Quantitative

Transfers are facilitated through one-company culture Task characteristics influence the effectiveness of knowledge transfer

Knowledge transfer between R&D units Contingency view of knowledge transfer effectiveness Organizational capabilities as preconditions for effective knowledge transfer

R&D know-how

Procedural and declarative knowledge

Not specified

Tacit knowledge

Individual and organizational knowledge

Teigland, Fey and Birkinshaw (2000)

Becerra-Fernandez and Sabherwal (2001)

Gold, Malhotra and Segars (2001)

Subramaniam and Venkatraman (2001)

Sveiby (2001)

Quantitative

Quantitative

Conceptual

Social capital and knowledge integration are key to effective knowledge transfer Differences among countries and rich information processing mechanisms improve product development capabilities Levels of knowledge transfer and their characteristics

Knowledge transfer and transnational product development

Epistemological approach to strategy formulation

Quantitative

Units which are highly embedded and highly integrated transfer most

Reverse knowledge transfers with respect to unit’s embeddedness and integration

Technical knowledge

Hakanson and Nobel (2001)

Quantitative

Different determinants of knowledge inflows and outflows

Determinants of subsidiary knowledge flows

Procedural and declarative knowledge

Method

Gupta and Govindarajan (2000)

Main Findings

Aim

Type of Knowledge

(Continued)

Study

Exhibit 2.8

32

Conceptual

Quantitative

Micro and macro perspective. Differentiation of intra-MNC flows Subsidiaries’ strategic mandate and the headquarters absorptive capacity are critical

Sources of transferable subsidiary knowledge

Network perspective of knowledge transfers within MNCs Headquarters’ benefits of knowledge transfers from subsidiaries

Not specified

Not specified

Procedural and declarative knowledge

Mudambi (2002)

Chini, Ambos and Schlegelmilch (2003)

Knowledge from different sources within the unit has to be transferred differently

Quantitative

Conceptual

Foss and Pedersen (2002)

Differences in cultural pattern and cognitive style moderate effectiveness of knowledge transfers

Transfer of knowledge between different cultures

Quantitative

Organizational knowledge

Units in central network positions create more innovation

Bhagat, et al. (2002)

Impact of a unit’s absorptive capacity and network position on knowledge transfer

Not specified

Tsai (2001)

33

34 Effective Knowledge Transfer in MNCs

knowledge referred to in the research, the overall aim, the main findings and the method used are presented in order to provide an overview. The conceptual studies presented above strengthen the theoretical background of the knowledge-based view of the firm, provide frameworks of how to conceptualize knowledge transfers in MNCs, or focus on the problems and influencing factors in those transfers. Nahapiet and Ghoshal (1998) concentrate on the role of social capital and aim to explain the antecedents of the knowledge-based view from that perspective. Social capital is said to be the reason why firms outperform markets in knowledge transfer. Mudambi (2002) provides a concise literature review and argues in favour of the network perspective. He suggests viewing intra-MNC knowledge transfers from both a macro and a micro perspective. Another strong theoretical contribution is offered by Von Krogh and Köhne (1998), who identify three phases of knowledge transfer: initiation, actual transfer and integration. The authors also describe several factors – such as the nature of knowledge, the interaction of sender and recipient, motivation and corporate and local culture – that have a bearing on the knowledge transfer process. The framework of knowledge transfer outlined by Sveiby (2001) has already been discussed above (p. 16). The issue of knowledge transfer between different contexts in different geographies central to the study in hand has been addressed conceptually by Doz and Santos (1997). They argue that knowledge management becomes ‘eventful’ in the case of geographical dispersion and context differentiation and discuss strategies of packaging knowledge. A cross-cultural perspective is given by Bhagat et al. (2002), who theorize that differences in cultural patterns and cognitive styles moderate the effectiveness of knowledge transfers. Turning to the empirical contributions, researchers concur with Gupta and Govindarajan’s (2000, p. 474) observation that ‘Very little systematic empirical investigation into the determinants of intra-MNC knowledge transfer has so far been attempted.’ Among the notable exceptions are those who have contributed comprehensive empirical studies on intra-organizational knowledge transfer. Despite the overall dearth of empirical contributions, there are some studies that are of particular relevance for the present work. They can be categorized according to the five areas shown in Exhibit 2.9. Hansen (1999) finds that strong ties are especially useful when highly complex knowledge is concerned. In contrast, weak ties prove as efficient means of transfer when knowledge is less complex. Tsai’s (2001) study discusses intra-organizational knowledge transfer from the viewpoint of

Knowledge and the MNC 35 Exhibit 2.9

Areas of empirical contributions

Topic

Study

Ties and the network perspective

Hansen (1999); Tsai (2001)

Subsidiary perspective

Chini, Ambos and Schlegelmilch (2003); Foss and Pedersen (2002); Hakanson and Nobel (2001)

Coordination and control issues

Gupta and Govindarajan (1991, 1994, 2000)

Capabilities and context

Becerra-Fernandez and Sabherwal (2001); Bhagat, et al. (2002); Gold, Malhotra and Segars (2001); Inkpen and Dinur (1998); Subramaniam and Venkatraman (2001)

Transferability of knowledge

Zander and Kogut (1995); Szulanski (1996)

the unit’s centrality in the network and emphasizes its absorptive capacity. As our research includes knowledge transfers from headquarters as well as subsidiaries, it is important to acknowledge the characteristics of reverse knowledge transfers – i.e. from subsidiaries to headquarters – as outlined by Hakanson (2001) and Chini, Ambos and Schlegelmilch (2003) and to differentiate between diverse sources of subsidiary knowledge, as shown in Foss and Pedersen (2002). As mentioned above, the studies of Gupta and Govindarajan (1991, 1994, 2000) are among those which see knowledge flows as control or administrative mechanisms. Some ideas articulated in their (2000) study are applicable in our research. The studies listed as integrating capabilities and context and transferability in Exhibit 2.9 will not be further explained at this point. The discussion on (p. 27) as well as that in Chapter 3 focuses specifically on these issues. It has to be noted that only a few studies explicitly investigate the intra-MNC knowledge transfer process. Although these studies add greatly to our knowledge in this field, most of them take a very narrow perspective and fail to provide an integrative model. Moreover, hardly any study integrates the concept of knowledge transfer effectiveness/ success into their empirical investigation: notable exceptions are BecerraFernandez and Sabherwal (2001) and Gold, Malhotra and Segars (2001). Another attempt to account for this issue empirically was made by Chini, Ambos and Schlegelmilch (2003). As the organizational advantage gained from knowledge transfer is hard to isolate, and therefore hard to

36 Effective Knowledge Transfer in MNCs

measure, most authors concentrate on the incidence of knowledge transfers without distinguishing between effective/successful/beneficial transfers and others. In contrast, our research focuses on the effectiveness of knowledge transfers, assuming that not every transfer of knowledge is effective.

3 Knowledge-Based Determinants of MNC Strategic Configuration

Chapter 3 focuses on the knowledge-based determinants of the MNC’s strategic configuration. As shown in the literature review in Chapter 2, research on knowledge transfers within MNCs is marked by several streams of theory, such as ties and the network perspective, the subsidiary perspective, coordination and control issues, capabilities and context and the transferability of knowledge. Having identified these, the following sections will shed some light on the origins and the importance of these topics and show why they need to be integrated into a comprehensive model of intra-MNC knowledge transfer.

Headquarters–subsidiary relationships Traditionally, the research stream on headquarters–subsidiary relationships focuses on two aspects: the centralization and formalization of decision making and the integration of the portfolio of subsidiaries to maximize their usefulness within the MNC (Paterson and Brock 2002). It has recently been recognized that viewing the global network from the periphery instead of from the centre could add strategic value to managerial decisions. The management of the MNC as a network implies the need to balance local responsiveness and central coordination. However, perceptions are likely to differ as they take their perspective from different parts of the organization. Headquarters–subsidiary relationships can generally be compared to a principal–agent relationship. Nohria and Ghoshal (1994) emphasize that the structure of headquarters–subsidiary relationships has to fit its context. Authors such as White and Poynter (1990) also note that subsidiaries are confronted with different challenges and require different administrative practices. MNCs can no longer rely exclusively on their home base (Doz et al. 1997) – neither in terms of tangible 37

38

Effective Knowledge Transfer in MNCs

resources nor in terms of knowledge. Consequently, the power of subsidiaries vis-à-vis the headquarters increases and an important stream of scholarly research now focuses on the role of subsidiaries in knowledge creation (Holm and Pedersen 2000; Foss and Pedersen 2002). The shift from a hierarchical to a heterarchical conceptualization of the firm has led to a change in perspective from the MNC level to the subsidiary level (Birkinshaw et al. 2000). Moreover, the role of network actors has to be seen from a dynamic viewpoint. First, the configuration of the network is constantly changing: network positions are not stable but reactive to changes in the environmental as well in the firm context. Second, not only headquarters ‘introduce’ strategy, subsidiaries also take strategic decisions (Paterson and Brock 2002). In terms of power relations, a subsidiary is likely to push for more and possibly inappropriate levels of autonomy. Sensitive management styles are required so that the subsidiary can easily see and evaluate its contribution to the overall organizational success (Taggart and Hood 1999). In this context, power can be based on authority as well as on the control of critical resources. Resource dependence between actors is an important base of subsidiary power and autonomy. Subsidiaries’ power is also increasingly based on their role as a source of ideas, skills, knowledge and capabilities. In recent years, a separate research stream has thus focused on the development of subsidiaries (see also Birkinshaw and Hood 1998). Complexity is augmented through the twofold embeddedness of subsidiaries (see also Andersson and Forsgren 2002). A subsidiary’s network consists of relationships with different degrees of embeddedness. On the one hand, units are integrated into their own local context, forming relationships with external partners. On the other hand, they are integrated in the MNC network. Therefore, subsidiaries are facing two separate and sometimes contradictory forces. External embeddedness has to be seen as a source of local knowledge and innovation and can be critical to the MNC’s corporate advantage. Admitting high degrees of external network relationships means that headquarters’ control of the subsidiary, and thus over corporate strategy, decreases (Andersson and Forsgren 1996; Ambos and Reitsperger 2004). Another problem is that headquarters’ knowledge about subsidiaries’ networks is limited and thus controlling the development of subsidiaries becomes a decision made under uncertainty. For this reason, understanding the network can be a source of power independent of the unit’s formal position (Holm, Johanson and Thilenius 1995). It thus becomes evident that the problematic of headquarters–subsidiary relationships has an important bearing on the theory of intra-MNC

Knowledge-Based Determinants of MNC Strategic Configuration

39

knowledge transfers. Not only is the question where knowledge is created important, but also which parts of the organization it can be used in becomes critical. The design of coordination and transmission channels is also likely to be inefficient unless at least some actors possess sufficient knowledge about the network configuration.

Strategic mandates, coordination and control While headquarters–subsidiary relationships have been discussed in the preceding section, the concrete differences between organizational units are now the centre of analysis. As argued above, the MNC can no longer be seen as a centrally managed entity. This implies that the headquarters’ position is likely to differ in the same way as subsidiaries’. While there is a broad stream of literature about the varying strategic mandates of subsidiaries, headquarters’ strategic positions have not been discussed in depth. Despite a lack of classifying frameworks, it is important to note that headquarters also occupy a dynamic network position in the MNC. To ensure that knowledge is sent to the locus in the organization where it can ultimately add to value creation, coordination and control mechanisms constitute important transmission channels for knowledge flows. Having discussed different strategic mandates and the relevance of coordination and control in the MNC for this study, an attempt to align strategic mandates and coordination and control mechanisms is made. In final pages the issue is recaptured, focusing explicitly on the optimization of knowledge flows.

Strategic mandates of subsidiaries Structural variations across geographic units reflect different cultural and economic environments, resulting not only in differences in organizational structures, but also in various strategic roles for subsidiaries (Malnight 2001, p. 1189). Substantial differences across subsidiaries within MNCs exist because of the different strategic contexts in terms of the magnitude and directionality of capital, product and knowledge flows subsidiaries face in the MNC network. While certain subsidiaries engage in intra-corporate transactions, others do not. If they do, then the volume and criticality of these transactions is decisive. For each type of transaction, a subsidiary may engage in high or low levels of transaction inflow as well as high or low levels of transaction outflow. Subsidiaries can also be distinguished as receivers or providers of what is being transferred. As a consequence, the degree of lateral interdependence among subsidiaries also varies across roles (Gupta and Govindarajan 1991).

40

Effective Knowledge Transfer in MNCs

In the literature, the terms subsidiary ‘strategy’, ‘role’, ‘mandate’ or ‘charter’ are often used interchangeably. The utilization of different terminologies is mostly due to the contributor’s research background – organizational behaviour, strategic management, etc. However, one important difference in the conceptualization is whether such mandates are assigned by headquarters or achieved by the subsidiary autonomously. Birkinshaw and Morrison (1995) define ‘role’ as the outcome of a deterministic process whereby the subsidiary fulfils an imposed function. In contrast, ‘strategy’ suggests a higher degree of freedom on the part of subsidiary management to define its own destiny. Although this differentiation is especially critical in subsidiary development, it is not that relevant for the study at hand, because the aim of this study is a survey of the status quo rather than a dynamic view. Throughout this study, the term ‘strategic mandate’ will be used. Several authors (Bartlett and Ghoshal 1986; Gupta and Govindarajan 1991; Martinez and Jarillo 1991) have recognized the different strategic mandates taken by subsidiaries within the MNC, and classified them along certain dimensions. The theoretical foundations and implications presented in literature vary significantly since different criteria were used to categorize the different subsidiaries (Ambos and Reitsperger 2004). Bartlett and Ghoshal (1986) model subsidiary mandates as a function of the significance of the unit’s local environment to the company’s global strategy and its unique resources and capabilities. Gupta and Govindarajan (1991) array subsidiaries according to the extent to which a subsidiary receives knowledge inflows from the rest of the corporation and the extent to which it engages in knowledge outflows to the rest of the corporation. Exhibit 3.1 gives an overview of some subsidiary mandate typologies that have been developed in the literature. Although the underlying rationales vary considerably, Birkinshaw and Morrison (1995) suggest the categorization shown in Exhibit 3.1. Most frameworks take into account the importance of subsidiary responsiveness/autonomy versus integration/interdependence/coordination to characterize subsidiary mandates (Paterson and Brock 2002). However, not all elements of the typologies in Exhibit 3.1. are completely comparable. Ambos (2002) notes that many studies assign the same labels while building on different theoretical bases. To outline the most important characteristics of the typologies which are especially helpful for this research, the approaches of Bartlett and Ghoshal (1986), Jarillo and Martinez (1990), Gupta and Govindarajan (1991), and Birkinshaw and Morrison (1995) were chosen. Although these mandates build on different dimensions, they approximate each other to a certain extent. It has also to be noted that a single subsidiary is likely to have multiple

Knowledge-Based Determinants of MNC Strategic Configuration Exhibit 3.1

41

Subsidiaries’ strategic mandate typologies LOCAL IMPLEMENTER

SPECIALIZED CONTRIBUTOR

WORLD MANDATE

White and Poynter (1984)

Minature Replica

Rationalized Manufacturer

Global Mandate

D’Cruz (1986)

Branch Plant

Globally Rationalized Product Specialist

World Product Mandate

Bartlett and Ghoshal (1986)

Implementer

Contributor

Strategic Leader

Jarillo and Martinez (1990)

Autonomous

Receptive

Active

Gupta and Govindarajan (1991)

Local Innovator, Implementer

Global Innovator

Integrated Player

Integrated

Global Subsidiary Mandate

Roth and Morrison (1992)

Source: Birkinshaw and Morrison (1995, p. 733).

roles at the same time (Birkinshaw and Morrison 1995). Birkinshaw and Morrison’s (1995) research adds that the structural context varies according to the subsidiary’s mandate. A subsidiary’s structural context can be defined as ‘the set of formal and informal management systems that determine the relationship of the subsidiary to its parent and affiliates’ (Birkinshaw and Morrison 1995, p. 730). The structural context is thus consistent with the subsidiary’s strategic objectives and shapes the subsidiary’s mandate. In turn, the subsid-iary’s autonomous actions also shape its structural context. In other words, there are three perspectives for the determination of subsidiary mandates:

• First, headquarters assign a mandate to the subsidiary which is controlled through a variety of formal and informal mechanisms.

• Second, having sufficient freedom, the subsidiary defines its own mandate.

• Third, the local environment determines the subsidiary’s mandate through the influences of the host country characteristics. Some subsidiary mandates which are comparable to a certain extent are now described:

42

Effective Knowledge Transfer in MNCs

The Local Implementer (or Autonomous Subsidiary) This is characterized by limited geographic scope, typically in a single country, and by severely constrained product or value-added scope (Birkinshaw and Morrison 1995). This subsidiary is relatively independent from the rest of the organization since it carries out most of the activities of the value chain ( Jarillo and Martinez 1990). Gupta and Govindarajan (1991) use the term ‘Local Innovator’ for this type of subsidiary, and define it through low knowledge outflows and inflows. The subsidiary has high local responsibility for the creation of relevant know-how in all key functional areas. This knowledge, however, is likely to be idiosyncratic and is of only limited competitive use outside the subsidiary’s host country. In the case of global integration, Local Implementers could also be restricted to fewer value chain activities. Bartlett and Ghoshal’s (1986) as well as Gupta and Govindarajan’s (1991) Implementers correspond to this type. In this context, Local Implementers are highly dependent on the parent and engage heavily in inter-affiliate purchases as they typically specialize in downstream activities such as sales and marketing (Birkinshaw and Morrison 1995). These subsidiaries engage in little knowledge creation on their own and therefore rely on knowledge inflows from headquarters and peer subsidiaries (Gupta and Govindarajan 1991). In this situation, units in less strategically important markets do not need access to critical information; they are aimed at operational execution and at capturing the economies of scale and scope which are crucial to most companies’ global strategies (Bartlett and Ghoshal 1986).

The (Specialized) Contributor (or Receptive Subsidiary) This performs only few activities, typically marketing and sales, and is highly integrated in the rest of the firm (Jarillo and Martinez 1990). High internal product flows and internationally configured value chains are characteristic for this role (Birkinshaw and Morrison 1995). Contributors are mandated to execute strategies and decisions developed elsewhere. However, despite operating in a small or strategically unimportant market, these subsidiaries, which approximate Gupta and Govindarajan’s (1991) Global Innovator, have distinct internal capabilities. They serve as the fountainheads of knowledge for other units and can also be characterized by low knowledge inflows and high knowledge outflows. Their activities are thus tightly coordinated with the activities of other subsidiaries resulting in high levels of interdependence with other affiliates and the parent.

Knowledge-Based Determinants of MNC Strategic Configuration

43

The Black Hole A Black Hole is assumed to offer a potential country-specific advantage to the MNC, but has low firm-specific advantages. Following Birkinshaw and Morrison’s (1995) classification it can be either considered a low-performing World Mandate Subsidiary (see below) or a high-potential Specialized Contributor. As Black Holes are located in important markets, a strong local presence is essential for maintaining the company’s global position. However, a Black Hole is not seen as an acceptable strategic position but a state of transition that has to be overcome (Bartlett and Ghoshal 1986).

Strategic Leader (World Mandate, Active Subsidiary or Integrated Player) This has world-wide or regional responsibility for a product line or entire business or function. Its activities are integrated world-wide, but managed from the subsidiary and not from the headquarters (Birkinshaw and Morrison 1995). An Active Subsidiary performs many activities in close interdependence with the rest of the firm. It constitutes an active node in a tightly knit network (Martinez and Jarillo 1991). Exhibiting high knowledge inflows and outflows, this subsidiary type also has the responsibility for creating knowledge that can be utilized by other subsidiaries. However, it is not self-sufficient in the fulfilment of its own knowledge needs (Gupta and Govindarajan 1991). Following Bartlett and Ghoshal (1986), a Strategic Leader operates in a strategically important market and has high levels of resources and expertise. It thus has no reliance on lateral product flows, external sourcing of raw materials and external selling of products (Birkinshaw and Morrison 1995). This highly competent national subsidiary serves as a partner of headquarters in developing and implementing strategy. It not only has to detect signals of change but also help to analyse the threats and opportunities and develop appropriate responses (Bartlett and Ghoshal 1986).

Coordination and control within the MNC The increased share of sales and profits from overseas and the increased self-sufficiency of subsidiaries is a trigger for engagement in the study of coordination and control mechanisms. In view of the dispersed resources and complex internal flow patterns, the expanding array of coordination and control mechanisms has been an important topic in the MNC literature (Malnight 1996). MNC management faces the question where decisions are made, how operations can be optimized globally and how the country units should

44

Effective Knowledge Transfer in MNCs

report to the headquarters. Even more than the subsidiary’s resource dependence, the design of coordination and control mechanisms can be viewed as a source of headquarters’ control over the subsidiary (Andersson and Forsgren 1996). Control as an integrating mechanism is also aimed at reducing uncertainty and at ensuring that behaviours originating in separate parts of the organization are compatible and support common organizational goals. The fit between the organizational model and the coordination and control mechanisms applied is found to be one of the most influential factors explaining performance differences within MNCs (Harzing 1999). Harzing (1999) provides an in-depth literature review on the classification of coordination and control mechanisms. Definitions of control include two different aspects:

• First, control can be seen as a means to direct behaviour in an organization towards the goals of this organization.

• Second, there is an element of power in this relationship. Although many authors use the terms ‘control’ and ‘coordination’ synonymously (cf. Martinez and Jarillo 1989) coordination can be viewed as the process of ‘integration, harmonization or linking different parts of an organization towards a common goal’ (Harzing 1999, p. 9). In contrast to control, the power element is more implicit. With regard to network relationships in the MNC, coordination and control can be viewed as the way headquarters make sure that subsidiaries behave in concordance with headquarters’ policy. In this study, control is seen as a means to achieve an end called ‘coordination’. To optimize knowledge flows within the MNC, efficient coordination has to ensure that units link up and exchange knowledge through appropriate transmission channels. Most researchers distinguish between formal and informal control mechanisms (Martinez and Jarillo 1989; Gupta and Govindarajan 1991). Gupta and Govindarajan (2000) specifically distinguish between formal integrative mechanisms and corporate socialization mechanisms as transmission channels for knowledge flows. Harzing (1999) classifies diverse control mechanisms prominent in the literature into the four categories shown in Exhibit 3.2. Personalized centralized control describes the decisions taken at the top management level and the personal surveillance of their execution. In contrast, bureaucratic formalized control is an impersonal, indirect mechanism, usually in the form of written rules or manuals. By definition, output control does not focus on behaviour but on outputs. Monitoring

Knowledge-Based Determinants of MNC Strategic Configuration Exhibit 3.2

45

Control mechanisms PERSONAL/CULTURAL (FOUNDED ON SOCIAL INTERACTION)

UNPERSONAL/ BUREAUCRATIC/ (FOUNDED ON INSTRUMENTAL ARTEFACTS)

Direct/Explicit

Personal centralized control

Bureaucratic formalized control

Indirect/ Implicit

Control by socialization and networks

Output control

Source: Harzing (1999, p. 21).

or reporting systems are the means through which financial data, sales figures, etc. are transmitted. The final category, control by socialization and networks, is the most complex. Harzing (1999, p. 22) further distinguishes between:

• Socialization: a control by common organizational culture and shared values

• Informal, lateral and horizontal exchange of information: non-hierarchical control

• Formalized lateral and cross-departmental relations: non-hierarchical but temporarily formalized control. However, these mechanisms cannot be regarded as mutually exclusive. Different control mechanisms are instead implemented simultaneously for different types of subsidiaries, employees and parts of the organization, or for different situations and environments the MNC faces (Ambos 2002; Martinez and Jarillo 1989). Organizational characteristics (size and interdependence) and environmental characteristics (uncertainty and heterogeneity/complexity) are supposed to have the largest influence on the choice of control mechanisms in MNCs (Harzing 1999). Generally, any combinations of these categories are possible, but the literature stresses that some combinations are more probable than others (Harzing 1999). Output control and control by socialization and networks for example, are likely to appear together in situations with high environmental uncertainty, complex technology and limited knowledge of the transformation process. A more recent stream of research finds that strategic mandates and cultural distance have an impact on the choice of coordination and control instruments (Nobel and Birkinshaw 1998; Ambos and Reitsperger 2002). The coordination and control mechanisms used for the different strategic mandates are now discussed.

46

Effective Knowledge Transfer in MNCs

Coordination and control of different strategic mandates The fit between the role of the subsidiary and the type of coordination and control exercised over the subsidiaries proves to be one of the influential factors on the performance of the MNC (Harzing 1999). Research has found evidence for a strong relationship between the role an MNC assigns to a subsidiary and the level of coordination and control that is required (Nobel and Birkinshaw 1998; Ambos and Reitsperger 2002). Specialized Contributors, for example, prove to be more involved in all kinds of control mechanisms than Local Implementers. Local Implementers require less coordination, since they are relatively independent. However, these subsidiaries represent the MNC’s opportunity to capture learning which is the source of its competitive advantage. Bartlett and Ghoshal (1987) propose that, generally, units with implementation responsibility should be managed through tight operating controls with standardized systems used to handle much of the coordination. This corresponds to Birkinshaw and Morrison’s (1995) findings which suggest high levels of bureaucratic formalized control and low strategic autonomy for Local Implementers, respectively. However, in Martinez and Jarillo’s (1989) study, Local Implementers showed the lowest levels of total control compared to the other mandates. In subsidiaries exhibiting high degrees of integration with the rest of the MNC (such as Specialized Contributors and World Mandates) all control mechanisms, especially the subtle ones (control by socialization and networks), are used more extensively (Martinez and Jarillo 1991). The Specialized Contributors seem to be the tightest controlled mandates. They are characterized by high levels of interdependence with peer subsidiaries. Normally, they specialize in a narrow set of value chain activities (Birkinshaw and Morrison 1995), so headquarters have to put local interests above global ones when managing these units. Strategic Leaders must be given freedom to develop responsibility and an entrepreneurial approach although headquarters’ support must still be intensive. For this unit, operating controls may be light and quite routine, but coordination of information and resource flows to and from the unit will probably require intensive involvement of senior management (Bartlett and Ghoshal 1987). The complex local environment accounts for the value of knowledge stock gathered in these subsidiaries; to be able to source and integrate this knowledge, Strategic Leaders demand greater autonomy and less centralization and formalization. Compared to the other mandates, these subsidiaries have the highest levels of strategic autonomy and, thus, the lowest levels of bureaucratic formalized control (Birkinshaw and Morrison 1995). In order to establish control

Knowledge-Based Determinants of MNC Strategic Configuration

47

among Strategic Leaders, the literature suggests the imposition of tight social control (Bartlett and Ghoshal 1986). However, negative arguments against social control also exist. When referring to Strategic Leaders as Centres of Excellence (CoEs), the imposition of a company-centric culture may inhibit the CoE from integrating into the local research community. In this case, the acquisition of new knowledge is inhibited (Egelhoff 1999). Moreover, Hansen (1999) concludes that strong inter-unit ties through tight social control impede the acquisition of new knowledge and increase the risk of redundant knowledge generation in the CoE. Strong social control of off-shore CoEs can even have a negative impact on CoEs’ performance (Ambos and Reitsperger 2004). This shows that the strategic mandate is an important determinant of intra-MNC knowledge transfer. Units have to be managed differently to maximize overall performance and to build the basis for a global network that provides appropriate transmission channels for knowledge flows.

Managing strategic mandates in order to optimize knowledge flows As mentioned earlier, according to Gupta and Govindarajan (1991, 1994, 2000) knowledge flow patterns between organizational units represent a core dimension along which subsidiaries’ strategic contexts can differ. As this is among the main concerns of this research, it is seen as important to address the question of how to manage and control these patterns separately. This study centres on declarative and procedural types of knowledge flows. As far as procedural knowledge is concerned, expertise, skills and capabilities are involved. This type of knowledge can take the form of input processes (e.g. purchasing skills), throughput processes (e.g. product designs, process designs and packaging designs) or output processes (e.g. marketing know-how). Declarative knowledge, such as external market data, is mirrored in the transfer of globally relevant data about key customers, competitors, or suppliers (Gupta and Govindarajan 1991). Three specific major determinants of control mechanisms can be identified that reflect the variations across the subsidiary mandates arising from its position in the MNC’s knowledge flow network (Gupta and Govindarajan 1991):

• Subsidiary differences in lateral interdependence, • Subsidiary differences in global responsibility and authority, • Subsidiary differences in the need for the exercise of autonomous initiative.

48

Effective Knowledge Transfer in MNCs

Lateral interdependencies between subsidiaries can be expected to be a positive function of the extent of both knowledge inflow and knowledge outflow (Gupta and Govindarajan 1994). Administrative mechanisms must be implemented that determine the appropriate degree of information-processing capacity for the subsidiary (Galbraith 1972; Egelhoff 1991). Complex formal integrative mechanisms such as liaison positions enhance the information-processing capacity and are thus expected to be high for subsidiaries playing the role of an Integrated Player (Gupta and Govindarajan 1991). Liaison personnel also serve as an important transmission channel for knowledge flows between organizational units. Another function affecting the information-processing capacity is the intensity of communication linkages in terms of frequency, informality, openness and density, that is generally higher for lateral interdependent subsidiaries such as Integrated Players than for Global Innovators, Implementers and Local Innovators (Gupta and Govindarajan 1991). The perspective and attitude of subsidiary managers also has to be addressed. The greater the degree of lateral interdependence between subsidiaries the greater is the need for global as opposed to local subsidiary management orientation. The ratio of expatriate managers – who are more likely to have a comprehensive understanding of the MNC’s overall global strategy – in contrast to local managers is therefore expected to be high for Integrated Players, medium for Global Innovators and Implementers and low for Local Innovators (Gupta and Govindarajan 1991). Effective management of high lateral interdependence also requires close identification with and commitment to the entire MNC. Subsidiary managers of an Integrated Player have thus to be more corporately socialized than those of subsidiaries mandated as Global Innovators and Implementers – and, of course, much more socialized than those of Local Innovators (Gupta and Govindarajan 1991). With regard to subsidiary differences in global responsibility and authority, mechanisms such as special bonus systems, budget evaluation styles, etc. aim to motivate managers ‘to think of their responsibilities in either global or local terms, as appropriate, and mitigate the emergence of frustration in those contexts where the manager’s responsibility exceeds his or her authority’ (Gupta and Govindarajan 1991, p. 781). Finally, with respect to the variations in the degree of autonomous initiative of subsidiary managers, mechanisms such as decentralization of decision making authority or the size of the subsidiary managers’ bonus relative to their salary have to be tailored to the magnitude and scope of knowledge creation expected from a subsidiary (Gupta and Govindarajan 1991).

Knowledge-Based Determinants of MNC Strategic Configuration

49

From the preceding section, it can be concluded that units are assigned different strategic mandates, and that these mandates require appropriate coordination and control mechanisms in order to reach overall effectiveness in the MNC. In the context of knowledge flows, it is especially important that formal and informal coordination and control mechanisms are in place because they provide the pattern for global exchange. It was also shown that strategic mandates can be based on the distinction of their knowledge flows. Knowledge flows are thus primarily seen as control and administrative mechanisms.

The capability perspective The last two theoretical building blocks have both focused on the organizational unit’s strategic position in the MNC. Even though we recognize the importance of these issues, they cannot account for the entire explanation of knowledge transfer in the MNC. What seems to be equally important is the capability perspective presented here. This perspective stresses the relevance of organizational skills and routines. Very different schools of thought have all come to the conclusion that certain capabilities have an effect on the way knowledge transfer is designed and also on the effectiveness of knowledge transfer. First, some different theoretical approaches to the capability perspective are presented and some distinct knowledge management capabilities are then highlighted.

Approaches to organizational capabilities Drawing on the resource-based and knowledge-based view, a firm is regarded as a unique bundle of idiosyncratic resources and capabilities. Knowledge as a resource needs organizational capabilities to be productive (Grant 1996). Capabilities represent the organization’s collective capacity for undertaking a specific type of activity (Lieberman and Montgomery 1998). The MNC’s capabilities differ in terms of their intrinsic technological opportunities and their correspondence with market opportunities (Zander and Kogut 1995). In this context, the skills and routines developed by the MNC can be viewed as organizational underpinnings. In turn, the creation and maintenance of superior organizational routines is vital for the development and renovation of the firm’s competitive advantage (Kogut and Zander 1995). The most prominent and most widely discussed capability in this respect is the ‘combinative capability’ outlined by Kogut and Zander (1992). It is noticeable that, despite having a different explanation for the existence of the MNC, several research streams stress the importance of

50

Effective Knowledge Transfer in MNCs

this combinative capability. The internalization theorists Buckley and Casson (1976) argue that the very existence of a MNC lies in its ability to internalize externalities by putting together attributes, resources and activities within an internal market at a more efficient rate than markets do. The argument that firms create value through combining dispersed knowledge fits this perspective well, especially if one accepts that markets often fail to transfer this knowledge at a price – a perspective that is proposed by transaction cost economists (Hymer 1976; Teece 1981). Comparing different economic theories of the firm, Chandler (1992) notes the importance of organizational capabilities and finds that learning such capabilities requires trial and error as well as a guided process of learning and experimentation. Following a very different logic, Kogut and Zander (1993) come to a similar, if not even stronger, conclusion regarding the MNC as a knowledge integrating institution. Building an evolutionary theory of the firm, they state that knowledge exists in social relations among cooperating members of a community without fixed boundaries and ground their explanation of the MNC in the firm’s skills and capabilities and their nature. In this school of thought, which is also adopted in this study, the MNC is seen as a social community, whose productive knowledge defines a competitive advantage.

Knowledge management capabilities Although combinative capability has been the centre of discussion – last but not least because of its degree of theoretical abstraction and its deficiency in providing practical guidance – it is not the only capability critical to knowledge transfers in MNCs. In order to capitalize on their asset knowledge, firms have to develop learning capabilities: they need to focus on how to create and diffuse knowledge (Riesenberger 1998). Particularly, if one looks at intra-MNC flows, knowledge transfer has to happen before combinative capability can occur. In order to initially engage in the knowledge transfer process, additional capabilities are necessary. These can also partly be viewed as components of combinative capability. Gold, Malhotra and Segars (2001) find empirical evidence that firms may possess a predisposition for successful knowledge management through the development of key capabilities. Many terms have been used to describe specific knowledge management capabilities (Gold, Malhotra and Segars 2001):

• • • •

Capture, transfer, use (De Long 1997) Acquire, collaborate, integrate, experiment (Leonard-Barton 1995) Create, transfer, assemble, integrate, exploit (Teece 1998) Create, transfer, use (Skryme and Amidon 1997), etc.

Knowledge-Based Determinants of MNC Strategic Configuration

51

At the beginning, units have to seek for new knowledge. Whether the sources are inside or outside the firm is not relevant for this particular research. With regard to the capabilities required, organizational units must be able to ‘sense’ new knowledge in their environment (Ghoshal and Bartlett 1988; Brockhoff 1998). To integrate incoming knowledge, however, another capability is necessary. Cohen and Levinthal (1990) first introduced the construct of absorptive capacity which has recently been applied very frequently in theoretical and empirical studies (for a more detailed discussion see Zahra and George 2002). This construct has also been used to mirror combinative capability, as it is defined as the ability to use prior knowledge to recognize the value of new information, assimilate it and apply it to create new knowledge and capabilities. Assuming that organizational units already possess knowledge, they have to be capable of aggregating it within the unit at an organizational – in contrast to an individual – level. Aggregation at the organizational level implies a certain awareness, at least by the top management team or at functional specialists, that valuable knowledge exists inside the unit. The maintenance and the continuous update of this knowledge stock is important. All these factors can be encapsulated as ‘maintaining valuable knowledge stock’. Of course, this construct also comes close to combinative capability if it is assumed that various pieces of knowledge have to be organized efficiently in order to result in a useful ‘stock’, but aims at totally different issues than absorptive capacity. When it comes to the dissemination of knowledge within the firm, it is argued that certain capabilities are constitutive for the transfer process. Combining the different perspectives highlighted in the literature, the interplay of three mutually reinforcing capabilities seems necessary:

• First, the ability to coordinate with other units at formal and informal levels is critical. This issue has been addressed in the section about coordination and control (p. 43), but it is important to show that these processes can also be seen from a capability perspective. • Second, infrastructure facilitates knowledge transfers and determines how knowledge travels throughout the organization. The importance of technical support has long been stressed in the debate about knowledge management (cf. Hansen, Nohria and Tierney 1999). Infrastructure includes business intelligence, collaboration, distributed learning, knowledge discovery, knowledge mapping, opportunity generation as well as security. But infrastructure alone is not enough. Profound knowledge about how to use different tools has also to exist (Leonard-Barton 1995).

52

Effective Knowledge Transfer in MNCs

Infrastructure also has to be continuously developed – ultimately by its use; long-distance learning tools cannot function without any content, for example. Infrastructure is thus useful only if employees are capable of using it. • Third, knowledge transfer capability is probably most widely recognized in the knowledge management literature, as it concerns the distinct knowledge transfer processes. Depending on the type of knowledge and how it is packaged, different processes have to be applied in order to transfer knowledge either between people or between people and technology (see also Sveiby 2001). As emphasized in the discussion about the transferability of knowledge (p. 27), not every piece of knowledge is available in a form that is amenable to transfer. For example, in some situations brainstorming might be appropriate, while in others entries into repositories of lesson-learnt or databases are possible. To send and to receive knowledge, certain processes have to be applied to transform the knowledge into the appropriate form. The choice of processes depends on the characteristics of the knowledge as well as on the sender, the receiver and their context. Such combination and exchange of knowledge is supported by the presence of social capital (Nahapiet and Ghoshal 1998), which is defined as the sum of actual and potential resources embedded within, available through, and derived from, the network of relationships. This brief compilation of knowledge management capabilities shows that research in this field is rather fragmented. Although authors seem to agree that the development of capabilities is important for effective knowledge transfer, they use different terminologies and hardly ever provide clear definitions of the respective capabilities. The three knowledge process capabilities described above were combined because – according to the literature reviewed – they reflect the most important facets of intra-MNC knowledge transfer.

Contingency factors The conceptualization of knowledge transfers (Chapter 2) builds on Shannon and Weaver’s (1957) communication model. Despite being a very simplified approach, this model lends itself well to facilitate the knowledge transfer process as a dyadic relationship. In this section, the ‘noise’ of the communication model is in the centre of analysis. As transfers between MNC units located in different countries are investigated, some contingency factors are supposed to play an important role

Knowledge-Based Determinants of MNC Strategic Configuration

53

in the transfer process. First, some contingencies which make knowledge management ‘eventful’ are outlined. Then, the two factors – organizational isomorphism and national culture – which are supposed to have the highest impact on the effectiveness of intra-MNC knowledge transfers are highlighted and discussed.

‘Eventful’ knowledge management Doz and Santos (1997) argue that in MNCs, knowledge management becomes ‘eventful’ because of the dispersion in space and time and differentiation of context. Dispersed organizational units – and their knowledge – are embedded in a twofold manner (Granovetter 1985; Andersson and Forsgren 1996). First, local culture highly impacts the way knowledge management is organized. At the same time, MNCs face a pressure to conform to conditions in the local environment and an imperative for consistency within the MNC (Rosenzweig and Singh 1991). Interactions across distance are especially critical because they have to ensure the MNC’s integration and existence as a single entity across cultures (Manev and Stevenson 2001). ‘Thus, organizational and cultural barriers internal to the firm become a prime concern when the firm’s management is seeking the most effective use of its intangible knowledge assets’ (Buckley and Carter 1999, p. 80). The investigation of knowledge transfers between dispersed settings has also led to the recognition that the transfer of knowledge does not imply a ‘full’ replication of knowledge in a new location (cf. Doz and Santos 1997). Indeed, ‘transfer of knowledge is often associated with modification of the existing knowledge to the specific context’ (Foss and Pedersen 2002, p. 54). As far as the advantages and disadvantages of context similarities in knowledge transfers are concerned, two schools of thought have developed (Hansen 1999; Asakawa 2002). Capability-based theories and the product innovation literature hypothesize that a higher degree of interaction between units leads to more familiarity and subsequently to a better understanding of the knowledge transferred (Subramaniam and Venkatraman 2001). In contrast, the network school (cf. Granovetter 1973) argues that units with weak network ties, which are not in regular contact with the rest of the organization, operate in a different context and, thus, are able to introduce new knowledge, and they are viewed as an important source of innovation. But when it comes to transfers of complex knowledge that tend to be characterized by a high degree of tacitness, the instrumental benefits of weak ties are called into question (Hansen 1999). On the one hand, chances to receive more innovative, and

54

Effective Knowledge Transfer in MNCs

potentially more beneficial knowledge, might increase when the source has weak network ties. On the other hand, the likelihood of correctly understanding and subsequently benefiting from such knowledge sources might be lower because of a lack in context similarities.

Institutional isomorphism Functional entities in MNCs are organized differently in terms of organizational structure, corporate culture and core processes. All these components largely determine the effectiveness of knowledge management (Daft and Lengel 1986; Hansen, Nohria and Tierney 1999; Becerra-Fernandez and Sabherwal 2001). Units pursuing different strategic mandates are likely to develop characteristic nodes and processes, and consequently account for boundaries in information processing (Martinez and Jarillo 1989; Gupta and Govindarajan 2000) (see also p. 46). These mandates tend to be manifested in the development of specific characteristics, such as transmission channels, infrastructure and capabilities. Becerra-Fernandez and Sahberwal (2001) found that contingencies related to the task characteristics have an impact on the choice of knowledge transfer processes, and ultimately on the effectiveness of knowledge management. These facets of institutional isomorphism are supposed to have a bearing on the unit’s approach to structure and process knowledge (Asakawa 1995). Institutional isomorphism thus leads to organizational distance between individual organizational units of the MNC. Although we can recognize the innovation potential of loosely coupled networks (Hansen 1999), organizational distance between two organizational units of a MNC is likely to be detrimental to smooth knowledge transfers. Organizational distance is specifically viewed as differences between organizational units in terms of structures, processes and values.

National culture National culture is embedded deeply in everyday life and is relatively impervious to change. When management practices are inconsistent with these deeply held values, employees are likely to feel dissatisfied, distracted, uncomfortable and uncommitted. Congruence between management practices and the characteristics of the national culture therefore produce better performance outcomes (Newman and Nollen 1996). Correctly adapted strategies, such as communication across cultures, are potential sources of competitive advantage in the global business environment. The main strategic challenge is the alignment between key characteristics of the national culture and a firm’s strategy, structure, systems and practices (Griffith and Harvey 2001).

Knowledge-Based Determinants of MNC Strategic Configuration

55

It has to be recognized that knowledge is not culture-free. People in different contexts will attempt to interpret each other’s ideas based on their unique thought worlds. The human capacity to capture and understand complex facts is rooted in cultural settings and, thus, tends to differ across cultural areas. In order to adopt knowledge effectively in a new cultural context, a creative synthesis of different cultural and meaning systems has to be achieved (Tenkasi 2000). The reception of knowledge from another cultural context is likely to be easier when the system of underlying conventions fits the system of meanings of those expected to implement these procedures (Macharzina, Oesterle and Brodel 2001). However, cultural diversity in MNCs has proved to foster creativity and innovation (Gomez-Mejia and Palich 1997). Cultural diversity is thus especially relevant to knowledge creation and long-term perspectives on performance. A lack of context similarities requires knowledge to be transformed so that it conforms to existing cultural expectations (Tenkasi 2000). Beyond the general recognition that cultural differences are likely to impinge on the success of international knowledge transfer, concrete problems emerging in cross-cultural knowledge transfer are hardly ever addressed in the literature, with the notable exception of Bhagat et al. (2002), Subramaniam and Venkatraman (2001) or Doz and Santos (1997). Morosini (1998), for example, has conducted a study on strategy and execution across cultures in global corporate alliances. He concludes that, in order to be transferred successfully, knowledge has to be adapted to the recipient culture’s specifications, and not the other way round. ‘The knowledge thus disseminated can in itself be modified, enriched, and fed back from the local context to the rest of the organization and vice versa’ (Morosini 1998, p. 288). If a dyadic relationship is in the centre of analysis, the concept of cultural distance between two organizational units comes into play. Cultural distance can be defined as the difference between the national cultural characteristics of the home and the host countries (Hennart and Larimo 1998), or as the degree to which the cultural norms in one country are different from those in another country (Kogut and Singh 1988). Originally, this concept was suggested by Johanson and Vahlne (1977), who observed that Swedish firms progressively expand from their home base into countries characterized by less ‘psychic’ distance. Subsequently, this approach became known as the ‘Uppsala Model’. Although cultural distance was first primarily researched in combination with entry strategies – i.e. modes of entry and country risk (Kogut and Singh 1988; Brouthers and Brouthers 2001; Shenkar 2001) – it is

56

Effective Knowledge Transfer in MNCs

now used as a key variable in strategy, management, organization behaviour and human resource management (Shenkar 2001) and is often mentioned in relation to intra-organizational knowledge transfer. Cultural distance has also become an important variable in investigating organizational control systems, particularly when headquarters– subsidiary relationships are the centre of research (Johanson and Vahlne 1977; Kogut and Singh 1988; Nohria and Ghoshal 1994; Roth and O’Donnell 1996). In this context, cultural distance is used from two different angles:

• It can be approached from a transaction cost perspective, arguing that cultural distance increases transaction/coordination costs, as well as from • A resource-based view, regarding the very ability to bridge cultural distance as a unique advantage (Shenkar 2001). Cultural distance also has implications for the micro and the macro level (Manev and Stevenson 2001). At the macro level, higher cultural distance affects the integration of a subsidiary into the network (see also Jemison and Sitkins 1986), whereas at the micro level it may lead to misunderstandings, conflicts, or friction between managers. Smaller cultural distance means a higher degree of similarity in personal background, which facilitates social relationships and the interaction in a dyad, while larger cultural distance mostly hinders. Due to its wide use, implications of this concept can be found in various areas. For this study, the impact of cultural distance on the management of dispersed units and on the transferability of knowledge between cultures, is especially relevant. ‘Cultural distance has largely been taken to represent a hindrance to the performance of the MNE and its affiliates’ (Shenkar 2001, p. 522). Shared attributes often lead to homophily because managers are more likely to establish strong ties with colleagues who have similar attributes, values and perceptions (Manev and Stevenson 2001). Manev and Stevenson (2001) find that cultural distance has a negative impact on the strength of MNC network ties. Similarly, Rosenzweig and Singh (1991) conclude that choice of control mechanisms is dependant on the parent country culture and on the cultural distance between the parent and the subsidiary. If the parent and the subsidiary are from similar cultures, there may be less of a need to impose formal control mechanisms. Cultural distance may also have positive effects, as the firm’s creativity and ability to change is enhanced by more diversity. Especially in

Knowledge-Based Determinants of MNC Strategic Configuration

57

cross-border acquisitions, the target culture may provide a specific set of characteristics which can hardly be replicated in the bidding country (Morosini 1998). Augmentation of cultural distance in MNCs may also thus have positive performance effects. As far as the exchange of knowledge is concerned, Griffith, Zeybek and O’Brien (2001, p. 95) states that: It is important to recognize that the greater the cultural distance (inclusive of national and organizational) between people who are attempting to communicate effectively, and thus the less consistent the communication environment, the less likely there will be sufficient social bonding among individuals to facilitate effective communication. It can be concluded that knowledge management across dispersed units is likely to be influenced by several contingency factors. The phases of decontextualization in the transmitting culture and recontextualization in the recipient culture are very critical for the effectiveness of the transfer process. In particular, organizational and cultural distance is supposed to have a strong impact. However, it is difficult to account for direct effects as, in both cases, greater distance leads to more innovation potential whereas more obstacles are inherent in the transfer process.

4 A Model of Knowledge Transfer in MNCs

According to the literature review, a comprehensive model of crosscultural knowledge transfer within MNCs that can serve as a basis for more systematic empirical work seems to be missing. To respond to this need, a conceptual model of the knowledge transfer process across geographically dispersed units of MNCs is developed. As strategy, organizational structure and functions differ across MNCs, it is suggested that no single best way of transferring knowledge exists. On the contrary, the model proposes that the transfer of knowledge has to correspond to the strategic network position of the organizational unit as well as to the unit’s internal capabilities to manage knowledge. Integrating the literature, Exhibit 4.1 depicts the constructs forming the proposed conceptual model. Each of these constructs is now described in detail and research hypotheses are presented.

Strategic mandate Subsidiaries often vary in the nature of their operations. While some subsidiaries are mandated to contribute to the MNC by generating and disseminating new knowledge, the primary aim of others is to implement or exploit headquarters’ knowledge in the local context (Gupta and Govindarajan 1991; Kuemmerle 1997; Asakawa 2001a; Birkinshaw 2002; Ambos and Schlegelmilch 2004) (Exhibit 4.2). The strategic mandate based on patterns of knowledge inflow and outflow is thus also likely to be a key construct within flows between headquarters and subsidiaries. Based on Gupta and Govindarajan (1991, 1994), four generic subsidiary mandates need to be distinguished: 58

59 Exhibit 4.1

A model of intra-MNC knowledge transfer

Cultural Distance

Strategic Mandate Implementers

Cultural Distance

Local Innovators

Knowledge Transfer Effectiveness

Global Innovators

Knowledge Transfer Capabilities

Integrated Players

Transmission Channels

Satisfaction Perceived Benefit

Infrastructure Process Capabilities

Value of Knowledge Stock Value of Knowledge Stock

Organizational Distance Organizational Distance

Exhibit 4.2

Strategic mandates of subsidiaries

Knowledge Inflows

H i g h L o w

Implementers

Integrated Players

Local Innovators

Global Innovators

Low

High

Knowledge Outflows

60 Effective Knowledge Transfer in MNCs

• Global Innovator (high outflow of knowledge from the subsidiary to the corporation and low inflow from the corporation to the subsidiary)

• Integrated Player (high outflow, high inflow) • Implementer (low outflow, high inflow) and • Local Innovator (low outflow, low inflow). In line with the network approach described in the literature review, headquarters are also likely to differ according to their integration in the global organizational network. Although no labels have so far been assigned for headquarters fulfilling different strategic positions, they should be included in this construct. Thus, we can state a first general hypothesis:

Hypothesis 1 The development of knowledge process capabilities depends on the strategic mandate of the organizational unit.

Value of knowledge stock Organizational units require access to other units’ knowledge and have to possess certain internal capabilities in order to engage in knowledge transfer (Tsai 2001). Thus, having reviewed mandates as the units’ strategic positions in the organization, the focus needs to move to the units’ potentiality to transfer knowledge. The knowledge actually transferred is likely to be influenced by the attractiveness of a unit’s knowledge stock in relation to other units.1 Normally, each organizational unit pursues a dual task: it sends knowledge to others (source unit) and it receives knowledge from others (target unit). The value of a source unit’s knowledge stock strongly determines the unit’s propensity to engage in knowledge transfer: if a unit’s knowledge is not attractive, it will not be asked to share its knowledge. The available knowledge also has to be non-duplicative and useful for other units’ purposes (Gupta and Govindarajan 2000). Thus, a model depicting knowledge transfer within MNCs needs to take account of a unit’s value of knowledge stock relative to headquarters’ and to peer subsidiaries’ stock. Looking at an organizational unit as a recipient of knowledge, its absorptive capacity (Gupta and Govindarajan 2000) is likely to affect

A Model of Knowledge Transfer in MNCs 61

a unit’s ability to handle the incoming knowledge. Cohen and Levinthal (1990) define absorptive capacity as the ability to use prior knowledge to recognize the value of new information, assimilate it and apply it to create new knowledge and capabilities. A valuable knowledge stock is therefore likely to enhance a unit’s absorptive capacity. Prior knowledge (Cohen and Levinthal 1990) and homogeneity of the receiving and sending unit is expected to facilitate the assimilation and exploitation of new knowledge (Gupta and Govindarajan 2000). Foss and Pedersen (2002) also found strong empirical support for the view that the more knowledge the unit creates and absorbs, the more knowledge will be transferred. Thus, we have a second general hypothesis:

Hypothesis 2 A high value of knowledge stock positively affects the development of knowledge transfer capabilities.

Knowledge transfer capabilities In addition to acquisition, creation, utilization and storage, transfer is viewed as a key process in managing corporate knowledge (Marquardt 1996). Free knowledge flow has also been identified as one of the key elements of successful knowledge management (Riesenberger 1998). In order to support the free flow of knowledge, the company has to develop a certain organizational architecture – i.e. cross-functional, flexible structures (Nevis, DiBella and Gould 1995), free flow of communication (Argyris 1994) and a learning culture (Slater and Narver 1995). The actual knowledge transfer process is extremely complex and difficult to capture, since it has both, inter-personal and interorganizational dimensions. Moreover, as pointed out in the literature review (p. 35), transmission channels, infrastructure and processes have to be distinguished. Transmission channels have been identified as key to intra-MNC knowledge transfer (Bartlett and Ghoshal 1989; Gupta and Govindarajan 2000). Gupta and Govindarajan (2000) further differentiate into formal (formal integrative mechanisms) and informal channels (socialization mechanisms). In their empirical study, formal channels show a positive significant influence on the knowledge flow between the subsidiary and

62 Effective Knowledge Transfer in MNCs

the parent corporation, on the one hand, and peer subsidiaries, on the other. When it comes to the influence of informal channels, the results are significant only in the knowledge flow between peer subsidiaries. To analyse knowledge transmission channels, lateral and vertical channels have to be distinguished, on the one hand, and formal and informal channels, on the other. The knowledge management infrastructure of MNCs has to be highly developed in order to maximize the exploitation of resources that are embedded within, available through and derived from a network of units. An important aspect is the technological dimension, which addresses the technology-enabled ties that exist in a firm. Technical infrastructure within the field of a company’s knowledge management includes business intelligence (knowledge regarding competition and the broader economic environment), collaboration of individuals within the company and distributed learning, knowledge discovery (discover internal and/or external knowledge), knowledge mapping (track sources of knowledge), opportunity generation (track knowledge about customers) and security (prevent inappropriate use) (Gold, Malhotra and Segars 2001). Knowledge process capabilities refer to the practices knowledge-sending and knowledge-receiving organizations perform. Following Nonaka and Takeuchi’s (1995) knowledge spiral, four modes of knowledge conversion can be identified:

• Socialization (from individual tacit knowledge to group tacit knowledge)

• Externalization (from tacit knowledge to explicit knowledge) • Internalization (from explicit knowledge to tacit knowledge) • Combination (from separate explicit to systemic explicit knowledge). As these modes cover the possibilities of knowledge conversion between individual and organizational knowledge, they can equally be applied to knowledge transfer between organizational units.2 This view has previously been taken by Doz and Santos (1997), Inkpen and Dinur (1998) and Sveiby (2001), who argue that from an organizational viewpoint, knowledge shared is actually knowledge doubled – i.e. new knowledge is created with every transfer within the organization. If the modes of knowledge conversion are applied to intra-organizational knowledge transfer, externalization occurs before a unit sends knowledge to another. Socialization generally occurs in both organizational units – sender and recipient – as they have to interact mutually. To start

A Model of Knowledge Transfer in MNCs 63

socialization, however, a field of interaction has to be established (Nonaka, Umemoto and Senoo 1996), mostly initiated by the source unit. From a recipient point of view, internalization is used to integrate knowledge into the target unit. Finally, by combination, explicit knowledge is converted into even more complex sets of explicit knowledge (Nonaka and Takeuchi 1995). This mode is more often used by a target unit that has recently received new knowledge and is starting to ‘break it down’ (Nonaka and Takeuchi 1995) into systemic explicit knowledge. In view of the composition of knowledge transfer capabilities, we can refine our general hypothesis as follows.

Hypothesis 1a:

Organizational units characterized by high knowledge outflow and high knowledge inflow (for subsidiaries: Integrated Players) are expected to develop the highest 3 formal and informal transmission channels with peer subsidiaries/headquarters and the highest level of knowledge management infrastructure. Integrated players are expected to develop all knowledge processes at an equal and high level.

Hypothesis 1b: Organizational units characterized by high knowledge outflow and low knowledge inflow (for subsidiaries: Global Innovators) are expected to develop intermediate 4 formal and informal transmission channels with peer subsidiaries/headquarters and an intermediate level of knowledge management infrastructure. Global Innovators are expected to emphasize externalization and socialization. Hypothesis 1c: Organizational units characterized by low knowledge outflow and high knowledge inflow (for subsidiaries: Implementers) are expected to develop intermediate5 formal and informal transmission channels with peer subsidiaries/headquarters and an intermediate level of knowledge management infrastructure. Implementers are expected to emphasize internalization and coordination. Hypothesis 1d: Organizational units characterized by low knowledge outflow and low knowledge inflow (for subsidiaries: Local Innovators) are expected to develop the lowest formal and informal transmission channels with peer subsidiaries/headquarters and a low 6 level of knowledge management infrastructure. Local Innovators are expected to develop all knowledge processes at an equal and low level.

64 Effective Knowledge Transfer in MNCs Exhibit 4.3

Overview of hypotheses 1a–1d

STRATEGIC MANDATE

DEVELOPMENT OF TRANSFER CAPABILITIES

Transmission Channels

Infrastructure

Processes

Integrated Players

Highest

Highest

All processes at a high level

Global Innovators

Intermediate

Intermediate

Externalization, Socialization

Implementers

Intermediate

Intermediate

Internalization, Combination

Lowest

Lowest

All process at a low level

Local Innovators

Exhibit 4.3 gives an overview of hypotheses 1a–1d. From a systemic point of view, the development of these knowledge transfer capabilities – channels, infrastructure and processes – follows a unit’s specific strategic mandate and its ability to transfer knowledge. Knowledge transfer capabilities are mutually reinforcing and have to be coordinated in order to be employed efficiently. Based on these insights we can introduce general hypothesis 3:

Hypothesis 3 Appropriately developed knowledge transfer capabilities (in terms of channels, infrastructure and processes) have a positive impact on the effectiveness of knowledge transfer.

Knowledge transfer effectiveness The aim of knowledge transfer at the recipient unit is to integrate the new knowledge in the unit’s context and to make use of it. Effective utilization refers to the potential to turn knowledge into a competitive advantage-yielding capability (Grant 1996). Effectiveness is generally,

A Model of Knowledge Transfer in MNCs 65

seen as one dimension of performance, besides efficiency and adaptiveness (see also Katsikeas, Leonidou and Morgan 2000). Buckley and Carter (1999) note that an important requirement for effective knowledge transfer is for the source unit to recognize the knowledge requirements of the recipient unit in order to provide what is appropriate, in a form that is appropriate. As human behaviour, knowledge and cognition is guided by the contextual rules and resources resident in social structures and conventions, knowledge has to fit these contextual requirements of the recipient unit. The transfer of knowledge – especially of organizational procedures and management practices – from one cultural context to another is likely to fail unless the governing philosophy – the system of underlying structurational conventions – fits the system of meaning of those expected to implement these procedures from day to day, or unless the organizational routines are transformed so that they conform to existing cultural expectations. (Macharzina, Oesterle and Brodel 2001) The effectiveness of knowledge transfer is thus likely to be influenced by two moderators identified in the literature (Asakawa 1995; Inkpen and Dinur 1998; Buckley and Carter 1999; Simonin 1999b) – organizational distance and cultural distance.

Organizational distance ‘Organizational distance’ refers to differences between organizational units (headquarters–subsidiary, subsidiary–subsidiary) in terms of structures, processes and values. It attempts to capture issues such as differences in approaches towards decision making, for example (Simonin 1999a, p. 473) defines organizational distance as follows: ‘[It] captures the degree of dissimilarity between the partners’ business practices, institutional heritage, and organizational culture.’ Asakawa (1995) suggests that institutional isomorphism has a strong impact on the way local units approach and structure knowledge. With regard to knowledge transfer, it is assumed that organizational distance amplifies ambiguity; a large organizational distance may thus lead to a ‘lack of understanding of the logical linkages between marketing actions and outcomes, inputs and outputs, and causes and effects that characterize a broadly defined marketing-based competency and its transferability’ (Simonin 1999a, p. 467). The following general hypothesis can thus be advanced.

66 Effective Knowledge Transfer in MNCs

Hypothesis 4 The lower the organizational distance between units the higher the effectiveness of knowledge transfer.

Cultural distance ‘Cultural distance’ is the second moderator the literature takes into account. Knowledge and cognition, as human behaviour, is guided by the contextual rules and resources residing in social structures and conventions; transferred knowledge has thus to fit these contextual requirements of the recipient unit. The transfer of knowledge from one cultural context to another is likely to fail if the underlying assumptions are divergent from the system of meaning of those expected to implement the knowledge received. Alternatively, organizational routines have to be transformed so that they can conform to existing cultural expectations (Macharzina, Oesterle and Brodel 2001). Or, as Doz and Santos (1997, p. 23) put it: ‘effective transfer of knowledge is a dialogue between the sender and the receiver about their own contexts and about the object of knowledge.’ Cultural distance may also hamper the identification of market opportunities and understanding of market mechanisms (Simonin 1999b).

Hypothesis 5 The lower the cultural distance between units the higher the effectiveness of knowledge transfer.

Apart from recognizing the role of these moderators, there is little research that addresses the effectiveness of knowledge transfer. Among the noteworthy exceptions are empirical contributions by BecerraFernandez and Sabherwal (2001) and Gold, Malhotra and Segars (2001). Some researchers have already integrated the condition of effectiveness or success into their definition of knowledge transfer (Doz and Santos 1997) – i.e. knowledge transfer as such is always effective transfer (Jensen and Meckling 1995). It is argued, on the other hand, that the effectiveness of knowledge transfer processes depends on the perceived

A Model of Knowledge Transfer in MNCs 67

benefit (Foss and Pedersen 2002, p. 60) and the overall satisfaction with knowledge management (Becerra-Fernandez and Sabherwal 2001). To sum up, a model of knowledge transfer in MNCs which aims to explain the effectiveness of knowledge transfer between organizational units in different countries has been introduced. Five general hypotheses have been developed to describe the impact and the relations of the proposed constructs. As already mentioned, it is important to note that the effectiveness of knowledge transfer between MNC units is contingent on the appropriate development of specific organizational capabilities, developed in response to the unit’s strategic position in the network and the value of its knowledge stock. The effectiveness of knowledge transfer is also likely to be affected by organizational and cultural distance. Chapter 5 will focus on the design and the methodological aspects of the empirical study. Concrete operationalizations of constructs to test the hypotheses developed will then be presented.

5 Research Design and Methodology

Chapter 5 presents the method applied in the empirical study. First, some general issues about the research context and possible challenges of knowledge management research in MNCs are discussed. Based on this discussion, it is demonstrated that some of these issues are addressed in the choice of the unit of analysis and the questionnaire design. Two standardized questionnaires are generated, one for headquarters and one for subsidiaries, and pre-test interviews were helpful in refining the instrument. The second section describes the steps in data collection, starting with the choice of the target sample and the sampling process. Informant selection is then described and the final sample presented. The operationalization of constructs is finally outlined and measures used in the study discussed.

Research context Challenges in knowledge management research in MNCs Knowledge transfers between organizational units of MNCs provide the empirical setting for this research. To survey knowledge flows between MNC units, it is necessary to investigate multiple network nodes, and globally dispersed units lend themselves especially well to the investigation of locally adapted as well as globally standardized knowledge. The complex requirements of empirical multi-unit MNC research, however, also account for the scarcity of empirical work in the field of knowledge transfer. When multiple nodes are included, power relations between headquarters and subsidiaries, cultural distance and organizational structure play a critical role, which can hardly be analysed in isolation. Objective measures are also hard to find on the nature, quantity, or quality 68

Research Design and Methodology 69

of knowledge. Researchers largely have thus to rely on subjective measures. Scholars have long since pointed out that headquarters’ and subsidiary perceptions might differ. Reviewing the literature on embeddedness of units (Granovetter 1985; Andersson and Forsgren 1996, 2002), there is reason to believe that this is also the case for knowledge transfer. Researchers address the problem of capturing perception gaps between headquarters and subsidiaries by investigating corresponding relations in a multi-unit perspective. The lack of objective measures and the need to control for perception gaps are the main reasons which suggest the value of a perceptional approach. The literature (cf. Birkinshaw et al. 2000; Asakawa 2001a) suggests that we should distinguish dyad headquarters– subsidiary relationships from subsidiary–subsidiary relationships. This study focuses on dyadic perceptions and distinguishes between knowledge flows that occur laterally among subsidiaries and hierarchical flows between headquarters and subsidiaries.

Unit of analysis The unit of analysis in this study is the organizational unit of a MNC – i.e. headquarters or a subsidiary – and its knowledge transfers. Relationships and knowledge transfers between several units will be analysed. The aim was that one sampled MNC should provide five respondents – one in headquarters and four subsidiary managers in different countries. In limiting the study to one functional area in the MNC – i.e. marketing – it was sought to reduce complexity to a manageable level. Earlier research convincingly noted that different functions will differ in terms of knowledge processed and management approach (Reger 1997; Harzing 1999; Ambos 2002). The presence of a primary up-stream activity like research and development (R&D) and manufacturing or of a primary down-stream activity like marketing and sales is expected to moderate effects in the analysis of knowledge flows (Gupta and Govindarajan 2000). Focusing on the marketing function in headquarters and subsidiaries seems appropriate in terms of the primary research aim – the investigation of knowledge transfers – because coordination and integration of world-wide marketing knowledge is generally perceived as particularly important (Buckley and Carter 1999; Simonin 1999b). Marketing is also assigned a high level of competence in subsidiaries (Foss and Pedersen 2002), and thus marketing knowledge is also likely to be created in a decentralized fashion. Despite the problems associated with the key informant approach, such as measurement error and common method variance (see also Harzing 1999, p. 183), this approach was chosen for the survey. Owing to the complexity of the research design and the problems of response in this

70

Effective Knowledge Transfer in MNCs

research field, it was unlikely that enough multiple responses would be received; however, as a several units are involved in the survey, the consistency of answers throughout a company can be analysed.1

Questionnaire design and pre-test Measures of all constructs of the model were developed, based on an in-depth review of the literature, and some validated scales were adopted from major earlier studies. Subsequently, a preliminary questionnaire was composed: as headquarters’ and subsidiaries’ managers were addressed, two slightly different versions of the questionnaire were created. The questionnaire was carefully designed to be simple to complete. It was structured in ten subsections, each highlighting the question of the respective section. To enable thematic orientation, the research topics were used as headlines, e.g. Management Systems. The questionnaire predominantly consists of 7-point Likert scales; respondents had to answer two questions by filling in numbers from 1 to 7 and one open question where they had to indicate the locations of subsidiaries they were in contact with. The last page included demographics questions about the unit’s address, its set up, size and position in the market. To evaluate content validity of the measures selected, the questionnaires were pre-tested by four potential respondents in headquarters and subsidiaries. These interviews gave further insights into knowledge management practices and helped to refine the research instrument. In the pre-test interviews, some questions arose about the meaning of different knowledge management tools. Some managers had problems in answering questions, as they did not know some of the ‘highly sophisticated’ knowledge management tools, such as ‘groupware’. In these cases, examples were added in order to foster a better understanding. Another problem was the instruction to refer to other subsidiaries. Managers found it difficult to refer to culturally close and distant subsidiaries, indicated as subsidiary A and B, and to come back to these categories several times. An additional issue was the length of the questionnaire. Pre-tests took about one hour, including additional questions, clarifications and discussions. It was assumed that managers answering the questionnaire on their own would take about 15–20 minutes. In response to the feedback received from these interviews, the questionnaires were modified to create a final version.

Data collection Target sample The European Top 500 2 served as a sample frame for the empirical investigation. Companies were selected on the basis of their turnover,

Research Design and Methodology 71

their degree of internationalization and their industrial affiliation. The research plan involved data collection at two levels – headquarters and subsidiaries. Earlier research indicated that dyadic relationships are notoriously difficult to sample because of the need to include pairs in the survey (Asakawa 2001a; Ambos 2002). Furthermore, to address network effects adequately among headquarters and subsidiaries as well as subsidiaries and subsidiaries, only MNCs with more than six overseas units (Vernon 1966) were targeted. Because of resource limitations, an initial target sample of sixty MNCs was set. Given the European Top 500 as a sample frame, the sample size of each industry was calculated on the basis of its contribution to gross domestic product (GDP) and the industries were weighted according to their economic input to the European Union’s GDP in 2000. As industry structures highly influence the organization of MNCs, and subsequently their knowledge transfer practices, it was found important to control for industry-specific results. According to Jacob and Ebrahimpur (2001, p. 15) ‘organizational culture and the nature of the industry in question are important determinants of the value of tacit knowledge’. Industries were classified based on NAICS (North American Industry Classification System) codes, and industry strata were set. A more detailed differentiation of industries was not conducted, as the sample size was not sufficient to allow a meaningful comparison. Given these methodological considerations, the initial sample frame was screened and firms qualifying – i.e. operating at least six overseas subsidiaries (Vernon 1966) – were contacted in descending order according to their revenues. Whenever a company declined to cooperate in the survey, the next largest company in terms of turnover was approached. After this sampling procedure, sixty MNCs agreed to cooperate, resulting in a sample of 240 subsidiaries and sixty headquarters, i.e. 300 units.

Informant selection Informants were selected based on the following procedure and questionnaires were subsequently mailed to them: 1 Telephone contacts with headquarters or subsidiaries located in Germanspeaking countries to secure support of research. 2 Headquarters providing contact persons in four subsidiaries in different countries or agreeing to distribute the questionnaire themselves to four subsidiaries. 3 Mailing of questionnaire.

72

Effective Knowledge Transfer in MNCs

Informant contacts were first established by telephone. Starting with the largest corporations, senior managers (usually in charge of marketing) were contacted and asked to cooperate. Research shows that pre-contacts usually increase response rates, response speed, and data quality (Harzing 2000a). Headquarters were asked to coordinate the participation of their units because they are supposed to have more influence on the international coordination necessary. They were asked to give contact persons in marketing positions in four subsidiaries abroad or distribute the questionnaire directly to their colleagues. After carefully weighing the advantages of other methods (cf. Harzing 1997; Salzberger, Sinkovics and Schlegelmilch 1999), it was decided to conduct the survey by mail. To enhance response rates, the possibility of sending back-up questionnaires by e-mail was considered as an option because of the speed of transfer to remote subsidiaries. Several follow-up calls and e-mails were needed in order to collect the questionnaires. As some managers asked to receive the questionnaire by e-mail, some electronic versions were mailed and returned by fax. By the end of October 2002, data collection was finished.

Final sample The final sample consisted of 162 MNCs units belonging to forty-five companies. Thirty-eight headquarters and 124 subsidiaries participated. Despite their initial agreement, fifteen MNCs did not provide any data at all; eighteen companies returned the required five questionnaires, providing data on the headquarters and on four subsidiaries. Two MNCs even involved six units. In six cases, four questionnaires were completed, seven MNCs returned three, six gave information about two units and in three MNCs it was possible to survey only one unit. The sample represents leading MNCs of diverse industries, such as manufacturing (56 per cent), finance and insurance (21 per cent) and other services (11 per cent), including consulting companies, for example. Exhibits 5.1 and 5.2 depict the industry partitions of the target sample and the final sample. 70 per cent (114 units) of all surveyed companies were among the industry’s top five performers, and 102 of those market leaders had held their position for more than five years. Surveyed companies operated in twenty-nine countries in different geographic regions (Exhibit 5.3). Austria, Belgium, France, Germany, Ireland, the Netherlands, Switzerland and the United Kingdom are categorized as Central/Western Europe; Italy and Spain as Southern Europe; Norway and Sweden as Northern Europe; Bulgaria, Croatia, the Czech Republic, Hungary, Poland, Romania, Russia, Slovakia and Slovenia as

Research Design and Methodology 73 Exhibit 5.1

Industry weights of target and final sample

Industry weighted as contribution to EU GDP (2000) Agriculture, Forestry, Fishing

Target sample (%)

Final sample (%)

Utilities

2.52 3.21

5.6 0

Construction

8.03

0

Manufacturing

22.78

56.2

Trade (wholesale + retail)

14.06

1.9

Transportation and Warehousing + Information

8.15

4.3

Finance, Insurance

7.80

21.0

Real Estate, Rental and Leasing Accommodation and Food Services Other Services

0

3.96

0

14.57

11 100

SUM

Exhibit 5.2

14.92

100

Industries in the final sample

Other Services 11.1%

Agriculture/Forestry 5.6%

Finance/Insurance 21.0%

Transportation 4.3%

Manufacturing 56.2%

Trade 1.9%

Central and Eastern Europe (CEE); China, Japan and Singapore as South East Asia. The majority of headquarters were located in Central/Western Europe (Exhibit 5.4), followed by CEE countries and Northern Europe. Out of the thirty-three headquarters located in Central/Western Europe, 26 were situated in Austria (Exhibit 5.5), and the response rate of Austrian companies was significantly higher. As noted by Harzing

74

Effective Knowledge Transfer in MNCs

Exhibit 5.3

Regions in the sample

CEE 30.2% Northern Europe 2.5% Southern Europe 4.9%

Australia 0.6% North America 3.1% Asia 3.1%

Central/Western Europe 55.6%

Exhibit 5.4

Location of headquarters

CEE 5.3% Northern Europe 2.6% Southern Europe 5.3%

Central / Western Europe 86.8%

(1999), Austrian managers felt a stronger affiliation to a study conducted by a local research institution, and contacts made in previous projects could be used. Although the Austrian bias of the sample has to be borne in mind, there are no indicators that country of origin effects played an important role in the study. Out of the twenty-six Austrian

Research Design and Methodology 75 Exhibit 5.5

Location of headquarters in Central/Western Europe

Belgium 3.0% United Kingdom 3.0% Switzerland 6.1% Germany 9.1%

Austria 78.8%

Exhibit 5.6

Location of subsidiaries

CEE 37.9%

Australia 0.8% North America 4.0%

Northern Europe 2.4%

Asia 4.0%

Southern Europe 4.8%

Centr. / West. Europe 46.0%

headquarters, ten were regional headquarters, mostly responsible for the CEE region. In total, twenty-four global and fourteen regional headquarters were represented in the sample. Most subsidiaries are also found in Central/Western Europe and CEE countries (Exhibit 5.6). As headquarters were free to choose the

76

Effective Knowledge Transfer in MNCs

subsidiaries participating in the study, they were influenced by geographic proximity. Moreover, many Central/Western European headquarters directly controlled CEE subsidiaries; fifty-seven subsidiaries responded that they had been created as acquisition or merger, fortyfive subsidiaries were set up as a greenfield operation. The remaining seventeen did not provide information about their set-up mode.

Operationalization and measures Strategic mandate In order to measure the strategic mandate of the unit, Gupta and Govindarajan (1994) have developed a useful instrument. The mandate, they argue, is actually defined by the intensity and the direction of knowledge flows. In this measure, different areas of data (i.e. market data on customers, market data on competitors) and procedural knowledge (i.e. marketing know-how, distribution know-how, technology know-how and purchasing know-how) essential to marketing departments are addressed. Exhibits 5.7 and 5.8 depict the knowledge flows which this study surveyed.

Exhibit 5.7

Knowledge flows to and from headquarters

OUTFLOW

Subsidiary A

Headquarters

Subsidiary B

INFLOW Exhibit 5.8

Knowledge flows to and from the focal subsidiary

OUTFLOW

Headquarters

Focal Subsidiary

INFLOW

Peer subsidiary

Research Design and Methodology 77

Following Gupta and Govindarajan (1991, 1994), the subsidiary mandate is operationalized as a product of knowledge outflows and knowledge inflows. As shown in Exhibits 5.7 and 5.8, respondents were asked to indicate the extent of knowledge flow from/to their subsidiary/headquarters on a Likert-type scale from 1 (not at all) to 7 (a very great deal).3 Knowledge flows were assessed on a six-item instrument. Consistent with Gupta and Govindarajan (1991, 1994), the resulting four types are termed ‘Global Innovators’, ‘Integrated Players’, ‘Implementers’ and ‘Local Innovators’.

Value of knowledge stock As mentioned above, a unit’s potentiality to transfer knowledge is determined by the value of its knowledge stock. This construct accounts for the unit’s ability to send knowledge, as well as its ability to process incoming knowledge, also called absorptive capacity. The theoretical assumptions of Gupta and Govindarajan (2000) provide guidance in the identification of a source unit’s value of knowledge stock. The authors suggest that a knowledge stock which is non-duplicative and relevant for the rest of the global network is an indication of a unit’s high value of knowledge stock. Gupta and Govindarajan (2000) operationalize this construct in terms of mode of entry, subsidiary size and the relative economic level of the host country. Despite their theoretical guidance, it was decided to develop a new scale which explicitly aims at surveying the unit’s perception of its own knowledge stock in relation to other units deemed to be useful. Again, a Likert-type scale was used (1 = much lower; 7 = much higher) and managers were asked: Generally, compared to others (all your subsidiaries), your subsidiary’s (headquarters’) knowledge stock in the following area is . . . For practical reasons, respondents should refer to the same types of knowledge (know-how) as explored by the ‘strategic mandate measure’.

Knowledge transfer capabilities Three different scales, which have been used in major empirical studies, are proposed in order to capture the three dimensions of knowledge transfer capabilities referred to above. The existence and richness of transmission channels scale was adopted from Gupta and Govindarajan (2000). Here, the use of liaison personnel, temporary task forces and permanent teams as lateral and hierarchical coordination instruments was explored on a scale from 1 (very infrequently) to 7 (very frequently). For the use of informal coordination

78

Effective Knowledge Transfer in MNCs

channels, working experience in other units, the use of mentors and the participation in executive development programmes involving employees from different subsidiaries were taken as proxies. For these questions, binominal answers (‘yes’ or ‘no’) were provided. The scale for knowledge management infrastructure stems from Gold, Malhotra and Segars (2001). Respondents were asked to assess how developed several aspects of their unit’s infrastructure are. While this scale primarily captured infrastructure-related practices, the next scale included different knowledge management tools in order to identify the four knowledge conversion processes suggested by Nonaka and Takeuchi (1995). Becerra-Fernandez and Sabherwal (2001) developed empirical measures for evaluating the extent to which each of the four knowledge process capabilities – externalization, internalization, combination and socialization – are used. As the scale includes different knowledge management tools, the question is formulated as follows: Please indicate how frequently each of the following knowledge management processes and tools is used in your company: If one of these does not exist in your company, please choose ‘Not applicable.’ Possible answers range from 1 (very infrequently) to 7 (very frequently), ‘not applicable’ is an additional option if the tools do not exist in the company.

Organizational distance Organizational distance has been measured in two ways. As a first approach, a means of self-perception of the differences between the own unit and the headquarters/other subsidiaries were chosen. For the former, it was decided to adapt the scale used by Simonin (1999a) and to choose the following questions: Generally, business practices and operational mechanisms are very similar. Generally, corporate culture and management style are very similar. Again, respondents were asked to provide information about agreement (1 = strongly disagree; 7 = strongly agree) with this statement, relating to hierarchical and lateral relationships versus culturally close and distant units. However, the measure proposed by Simonin (1999b) focuses only on selected aspects of organizational practices and the resulting distance

Research Design and Methodology 79

between units. It was thus decided to implement another, more sophisticated, organizational distance measure. Several strategic tools were integrated into the questionnaire in order to measure the company’s ‘strategic value disciplines’ based on Treacy and Wiersema (1995). For this purpose, a scale was developed based on the detailed specifications in the literature. Variables explored the orientation of a unit’s organization. Answers from headquarters and subsidiaries were compared and the distance calculated according to their divergent ratings. As different units of one company were surveyed, a comparison of each unit’s perceptions, and thus a relative measure of organizational distance, became possible.

Cultural distance To include cultural distance as a construct, several questions exploring whether cultural distance was explicitly perceived to be a barrier to knowledge transfer were formulated, based on the discussion in the literature: It seems to me that national culture influences the way of doing business heavily. Many misunderstandings and cultural conflicts emerge from knowledge transfer between units in different countries. Generally, language differences are major obstacles in communicating with and understanding subsidiaries. Questions integrated into measures of the strategic mandate, transfer capabilities and perceived benefit were aimed at exploring whether there was less transfer to culturally distant units within the company. Shenkar (2001) published an interesting review about measuring cultural distance which is essential to this approach. Kogut and Singh’s (1988, p. 422) formulas were also used in order to measure country culture differences ‘objectively’. Based on Hofstede’s (1980) indices, a composite index was formed which captured the deviation along each of the four cultural dimensions.

Knowledge transfer effectiveness In line with the other parts of the model, a self-perception measure was proposed, which aimed to capture the perceived benefits originating from knowledge inflows. This approach can be seen as an ex post analysis of knowledge transfer, assuming that successful transfer of knowledge will

80

Effective Knowledge Transfer in MNCs

lead to perceived benefits. A scale was developed in order to measure the perceived benefits (1 = not at all; 7 = a very great deal) stemming from the different areas of data/know-how mentioned above. At the same time, respondents were asked where this knowledge inflow came from, in order to survey the quality of knowledge transfer from the different organizational units. On a broader scale, the individual satisfaction with processes related to knowledge transfer on different organizational levels was captured (Becerra-Fernandez and Sabherwal 2001). Three organizational levels of knowledge management were addressed: the employee’s own context, the unit in question and the global organization. Possible answers ranged from 1 (strongly disagree) to 7 (strongly agree).

6 Analysis and Results

The analysis follows through the calculation of empirical results and discusses the implications of the findings. First, some descriptives about the unit of analysis are presented. Second, each of the model’s theoretical constructs are discussed and descriptive statistics and bivariate tests generated in order to provide an overview about the units’ characteristics and their interdependencies. Third, the constructs are integrated into the model and the hypotheses are tested using structural equation modelling. As data can be combined to three different sets, the three models are analysed and differences outlined.

Descriptives of the unit of analysis As outlined above (Chapter 5), the 162 surveyed units belong to forty-five MNCs that are mostly market leaders. They operate in twenty-nine countries and in ten geographic regions. In the sample there is a preponderance of firms located in Central and Western Europe – on the headquarters as well as on the subsidiary level. MNCs operate in diverse industries. However, the leading industry groups, manufacturing and finance and insurance constitute 77 per cent of the sample. To learn more about the characteristics of the headquarters and subsidiaries surveyed, the units’ size and their set-up are analysed, and an overview of the respondents’ position is given. The unit’s size is measured by their number of employees. On average, the headquarters and the subsidiaries in the sample employ 654 employees (Exhibit 6.1). Separating headquarters and subsidiaries, it becomes evident that headquarters tend to be more personnel-intensive. The average number of 81

82

Effective Knowledge Transfer in MNCs

Exhibit 6.1

Employees at units

> 2,500 8.5% 2,000–2,499 0.7% 1,500–1,999 1.4% 1,000–1,499 4.2% 500–999 12.7%

< 249 50.7%

250–499 21.8%

employees in headquarters is 1,019, whereas the average subsidiary employs 638 people. As mentioned earlier, fourteen headquarters fulfil the task of a regional headquarters, and twenty-four are the global centre of operations. In the case of subsidiaries, it is important to know how they have been founded as the mode of set-up often accounts for the extent of integration in the global network and the importance of the unit as a source or a recipient of knowledge. Among the subsidiaries surveyed, 44 per cent have been set up as a greenfield, 56 per cent as a merger or acquisition. The identification of respondents is vital for the validity of the data. As mentioned above, vice-presidents of marketing and top marketing managers were targeted in this study. At the end of the questionnaire, respondents were asked to fill in their position. Grouping together similar titles, Exhibit 6.2 lists the positions participants indicated in the survey. The key informant approach was successfully accomplished and data are assumed to represent the view of the companies’ top management.

Analysis of the model’s constructs In this section, the different constructs just discussed are analysed in turn. First, descriptives are generated to get an overview. Second, if possible,

Analysis and Results 83 Exhibit 6.2 Positions of respondents POSITION CEO/General Manager

NO. OF RESPONDENTS 53

Board Member

8

Vice-President/Marketing

18

Country Manager

8

Head of Marketing Group/Department

13

Advertising & PR Management

5

Business Development Manager

2

Marketing Manager

35

Communication/Information Service/Knowledge Manager

10

Controller

2

Executive Assistant

1

comparisons between different groups (i.e. headquarters and subsidiaries, culturally close and culturally distant units) are calculated. As some constructs are measured in different ways, the various results are outlined. In some cases, different operationalizations are used to perform robustness checks. The results and their impact on current research are then discussed.

Antecedents to knowledge transfer The subsidiary’s strategic mandate and the unit’s value of knowledge stock have been theoretically identified as antecedents in the model of intra-MNC knowledge transfer. The analysis of the strategic mandate will be conducted here at the subsidiary level and the characteristics that support knowledge transfer investigated separately on the headquarters and the subsidiary level.

Strategic mandate To identify strategic mandates, the analysis was conducted at the subsidiary level; 124 subsidiary cases were included. As a first step, the means of the different knowledge flows were calculated. On a scale from 1 (not at all) to 7 (a very great deal), the means of hierarchical flows (inflows from headquarters and outflows to headquarters) were 3.27 and 2.25, and for lateral flows (inflows from peer subsidiaries

84

Effective Knowledge Transfer in MNCs

and outflows to peer subsidiaries) means of 2.20 and 3.10 were generated. This result showed that inflows from headquarters had the highest value, slightly higher than inflows from subsidiaries and outflows to headquarters. Interestingly, outflows to subsidiaries ranked second. Next, subsidiary mandates were calculated, following Gupta and Govindarajan (2000). Taking the overall knowledge inflow and outflow over all knowledge types, the four subsidiary mandates discussed previously were derived by using median splits along these two composite measures. Looking at the sum of inflows and outflows, means and medians lay very close (inflow mean = 3.64; outflow mean = 3.58; inflow median =3.58; outflow median = 3.54). Calculation results indicated that thirty-eight of the subsidiaries surveyed could be called ‘Global Innovators’, twenty-nine ‘Local Innovators’, nine ‘Implementers’ and seven ‘Integrated Players’ (Exhibit 6.3). Cases containing one or more missing values in one of the original scales were excluded from the analysis. This resulted in forty-one missing values after the sequence of calculations. The predominance of ‘Integrated Players’ and ‘Local Innovators’ in the sample suggested that sixty-seven of the subsidiaries surveyed (81 per cent) were characterized by equal knowledge inflows and outflows. Only a minority (19 per cent) were characterized by a divergent level of knowledge flows. Integrated Players by definition contribute extensively to the knowledge base of the firm. However, other than global innovators they are not self-sufficient, but are dependent on knowledge residing elsewhere in the firm. Gupta and Govindarajan (1991) define Local Innovators as units dealing with idiosyncratic knowledge resources. Exhibit 6.3

Strategic mandates of subsidiaries

Knowledge Inflows

H i g h

Implementers 9 (11%)

Integrated Players 38 (46%)

L o w

Local Innovators 29 (35%)

Global Innovators 7 (8%)

Low

High

Knowledge Outflows

Analysis and Results 85

In order to check for the robustness of the construct, K-means cluster analysis was used. A two-cluster solution was generated which confirmed the median-split calculation. The question then arose whether the typology of strategic mandates represented groups of subsidiaries which differed considerably from each other. To investigate differences between the four mandates, cross-tabulations were generated for the size of the company, referring to the number of employees and the subsidiaries’ mode of set-up (Exhibit 6.4). No significant differences between the units’ sizes were found. Most subsidiaries lay in the smallest segments. While 52 per cent of Integrated Players had fewer than 249 employees, 18 per cent had 500–999 and 18 per cent 1,000–1,499. Local Innovators were similarly distributed, with 50 per cent below 249 employees, 26 per cent 500 –999 and 12 per cent 1,000 –1,499. As far as the mode of set-up is concerned, 39 per cent of Integrated Players were established as acquisition or merger, 61 per cent as a greenfield. Local Innovators, however, showed an inverse pattern – 67 per cent had been set up as an acquisition or merger, 33 per cent as a greenfield. However, there was evidence that the influence of the mode of set-up was not persistent, but the integration of units originally set up as a merger or acquisition changed over time (Hakanson and Nobel 2000). At this stage, one might wonder if the prevalence of only two strategic mandates – one highly embedded and one rather independent – was a unique characteristic of the sample. From a theoretical point of view there was no such evidence. Possibly, the fact that headquarters had to Exhibit 6.4

Cross-tabulation: strategic mandate and number of employees MANDATE * No. of employees Crosstabulation

Count No. of employees

MANDATE

Total

< 249

250 –499

500– 999

1000 –1499

> 2500

Total

Implementer

3

4

Global Innovator

7

Local Innovator

13

6

3

1

3

26

Integrated Player

17

6

6

3

1

33

40

16

9

4

4

73

7 7

86

Effective Knowledge Transfer in MNCs

nominate the units to participate in this survey influenced this result. Headquarters were asked to involve one subsidiary perceived as culturally close and one perceived as culturally distant in the study. A reason could be that managers perceived subsidiaries as culturally distant that were the least integrated into the global communication and control processes and largely relied on their own knowledge (Local Innovators). However, a correlation analysis of cultural distance according to Kogut and Singh’s (1988) distance measure and organizational distance 1 with the two mandates reveals that the correlations were not significant. Another possible explanation could be that headquarters tended to involve subsidiaries that they were regularly in contact with, on the one hand (Integrated Players), and rather autonomous subsidiaries (Local Innovators), on the other. To sum up the analysis of subsidiaries’ strategic mandates, the distinction of four strategic mandates does not make much sense in this case. Only two types, Integrated Players and Local Innovators, prevail. One reason could be that the typology suggested by Gupta and Govindarajan (2000), based on knowledge flows, is too simplistic, another that the random choice of subsidiaries participating in the survey led to biased results:

• A better classification approach might be the differentiation between global and local mandates. Birkinshaw and Morrison (1995), for example, adopt the term ‘world mandate’ for subsidiaries playing a role as a strategic leader. As opposed to locally centred subsidiaries, a subsidiary assigned a world product mandate engages in specialization and has decision making authority over a broad array of interconnected functions. Contrasting highly integrated to rather autonomous subsidiaries might lead to more convincing results. • To account for possible problems in the measure, the strategic mandates will not be included in the structural equation model. Instead, the original scales measuring lateral and hierarchical knowledge flows will be used. As mentioned earlier, headquarters are also likely to differ in their strategic position in the global organizational network. Applying the same analysis as in the subsidiary cases but, of course, including only hierarchical flows, the following characteristics were found: the means ranked 3.7 for inflows and 4.6 for outflows. Medians lay at 4.0 for inflows and 4.5 for outflows. For headquarters, a more balanced distribution can be seen (Exhibit 6.5). Among headquarters that provided data, most were characterized by high knowledge outflows and high knowledge inflows, fulfilling a traditional central position in the organization. It is important to notice that

Analysis and Results 87 Exhibit 6.5

Strategic positions of headquarters

Knowledge Inflows

H i g h

L o w

Exhibit 6.6

7 (25%)

11 (40%)

7 (25%)

3 (10%)

Low

High

Knowledge Outflows

Strategic positions of global versus regional headquarters TYPE

MANDATE

Total

Global HQ

Regional HQ

Total

IN high/OUT low

6

1

7

IN low/OUT high

2

1

3

IN low/OUT low

4

3

7

IN high/OUT high

7

4

11

19

9

28

the smallest group engaged in high knowledge outflows and low inflows. This position can be explained by a type of headquarters that sends many directions and advice to subsidiaries but does not get much knowledge back. It is possible that headquarters implemented mere output control mechanisms and were not engaged into more qualitative control. When this compilation was split between global and regional headquarters, it became evident that a larger percentage of global headquarters received high inflows whereas regional headquarters showed a similar pattern as subsidiaries’ knowledge flows. Exhibit 6.6 refers to the twentyeight headquarters that provided sufficient information to calculate their strategic mandate.

88

Effective Knowledge Transfer in MNCs

Value of knowledge stock While the strategic mandate aims to map the unit strategically, the value of knowledge stock aims to explain the unit’s potentiality to engage in knowledge transfer. The unit’s value of knowledge stock was analysed, first on the subsidiary level and subsequently the headquarters level. The results were also calculated using the operationalization of Gupta and Govindarajan (2000). As far as the subsidiaries’ knowledge stock compared to their peer subsidiaries was concerned, the majority of units perceived themselves to have a higher value of knowledge stock than their peers. The value of knowledge stock was measured on a scale from 1 (much lower) to 7 (much higher). The midpoint is thus located at 3.5. All means higher than 3.5 pointed towards a higher value of knowledge stock compared to others. As shown in Exhibit 6.7, the most valuable area seems to be ‘market data on customers’, followed by ‘marketing know-how’. ‘Purchasing know-how’ has the lowest mean. The knowledge stock of subsidiaries vis-à-vis their headquarters in the different areas is displayed in Exhibit 6.8. As Exhibit 6.8 shows, subsidiaries generally perceived their own knowledge stock as more valuable than headquarters’. All means were higher than 3.5 – i.e. the majority of subsidiaries perceived their knowledge stock in the respective areas to be higher than the knowledge stock of headquarters. Most valuable appear to be ‘market data on customers’, followed by ‘market data on competitors’. ‘Technology know-how’ ranked last. But even here, the mean indicated that most subsidiaries’ perceived their knowledge stock as higher than that of their headquarters. Exhibit 6.7

Knowledge stock of subsidiaries compared with peer subsidiaries

No. of companies

70 market data on customers

60 50

market data on competitors

40

marketing know-how

30

distribution know-how

20

technology know-how

10

purchasing know-how

0 1 2 3 Much lower

4

5 6 7 Much higher

Analysis and Results 89

No. of companies

Exhibit 6.8

Knowledge stock of subsidiaries compared with their headquarters

50 45 40 35 30 25 20 15 10 5 0

market data on customers market data on competitors marketing know-how distribution know-how technology know-how purchasing know-how 1 2 3 Much lower

Exhibit 6.9

4

5 6 7 Much higher

Knowledge stock of headquarters compared with subsidiaries

No. of companies

16 market data on customers

14 12

market data on competitors

10 8

marketing know-how

6

distribution know-how

4

technology know-how

2

purchasing know-how

0 1

2

3

Much higher

4

5

6

7

Much lower

When it came to headquarters, ‘technology know-how’, ‘purchasing know-how’ and ‘marketing know-how’ were the areas where headquarters perceived their knowledge stock to be far more valuable than the knowledge stock of subsidiaries (Exhibit 6.9). But, in parallel to subsidiaries, no area was found where headquarters perceived their knowledge stock lower compared to that of subsidiaries. What is evident from these comparisons is that all units somehow overestimate their own knowledge stock. It is remarkable that in the comparison of headquarters with subsidiaries, the highest means were reached, followed by subsidiaries’ estimations vis-à-vis headquarters. The lowest means were reached in the comparison of peer subsidiaries.

90

Effective Knowledge Transfer in MNCs

In sum, centralized functions such as technology and purchasing seem to lie with headquarters. This confirms the presumption made earlier that sampled headquarters mostly fulfil a traditional central function. Local market analysis is mostly done by the subsidiaries. This results in high values for ‘market data on customers’ and ‘market data on competitors’. The results for marketing know-how are ambiguous. This can be explained by a division of strategic tasks in the headquarters and local adaptation in subsidiaries. To compare the results of these newly developed measures with the original operationalization of Gupta and Govindarajan (2000), the mode of entry, subsidiary size and the relative economic level of the host country were measured. The underlying rationale is that units possess a valuable knowledge stock if their knowledge is non-duplicative and relevant for other units. Thus, the authors hypothesized that the highest value of knowledge stock will reside in units which are product of an acquisition/merger, which are relatively large and which are situated in a host country that is economically more advanced than the home country. To calculate the value of a unit’s knowledge stock, a formula assigning equal weight to the three variables is suggested:2 Relative size + Mode of set-up + Relative level of development = Value of knowledge stock To leverage differences between MNCs of different sizes, the unit’s size is taken relative to all surveyed units. A value of 1 is assigned to the units’ set-up as acquisitions or mergers and 0 to greenfield operations. If the relative level of economic development of the host country is higher than the home country’s, a value of 1 is entered. An equal level of development equals 0.5 and a lower level 0. Thus, the maximum to be reached for the value of knowledge stock is 3, the minimum 0. For the analysis, ninety-four subsidiary cases are valid because of missing values. The mean ranks at 0.85 and 75 per cent of subsidiaries have a level of knowledge stock lower then 1.50 according to this calculation. Gupta and Govindarajan (2000) only tested knowledge flows on the hierarchical level and did not take inter-subsidiary relationships into account. Lateral relationships will thus also be excluded from this analysis. Nevertheless, it was found important to assess headquarters’ value of knowledge stock, too. Obviously, the mode of set-up becomes irrelevant in the investigation of headquarters and the economic level of development cannot be assessed vis-à-vis all subsidiaries. For this purpose the same formula was applied without the variable ‘Mode of set-up’. Instead

Analysis and Results 91

of the relative level of economic development the absolute development was taken, assigning 1 to highly developed and 0 to emerging economies. To achieve comparable values, headquarters’ results were standardized on the 0–3 level. The results indicated that headquarters’ value of knowledge stock reached higher overall values with a mean of 1.68. 25 per cent of headquarters ranked lower than 1.52 and 50 per cent were distributed between 1.52 and 1.83. To compare the results, the two measures of the units’ value of knowledge stock were standardized using z-scores. A correlation analysis revealed a Pearson correlation coefficient of 0.067 that is not significant. It can thus be concluded that the two measures do not generate the same output. Overall, the perceptional measure reached a higher mean (1.83) compared to the (1.05) of Gupta and Govindarajan’s (2000) measure. The divergence of the two measures is most likely due to the different measurement approaches. First, variables are not comparable and aim at different aspects – i.e. the perceived knowledge stock versus a derived measure to investigate non-duplicative characteristics and relevance of knowledge. Second, the new scale builds on respondents’ perceptions, whereas the Gupta and Govindarajan’s (2000) measure builds on factual data. The obvious disadvantage of the perceptional measure is the tendency to overestimate the unit’s own knowledge stock and the existence of perception gaps. As mentioned above, all units assess their own value of knowledge stock higher than others. Nevertheless, Gupta and Govindarajan’s (2000) measure seems very problematic; although, their explanations about the influence of relative size, mode of set-up and relative level of economic development appear admissible, the concepts seems to be too general. Given that other measures in the survey relating to the extent of knowledge transfers are also perceptional, the perceptional measure of knowledge stock is favoured for this study.

Knowledge transfer capabilities Three different scales, which have been used to capture the three dimensions of knowledge transfer capabilities – transmission channels, knowledge transfer infrastructure and knowledge transfer processes – are now the subject of analysis.

Transmission channels The formal knowledge transmission channels, operationalized as coordination instruments between units – liaision personnel, temporary task forces and permanent teams – are examined for all surveyed units,

92

Effective Knowledge Transfer in MNCs

headquarters and subsidiaries. Judging from the means, liaision personnel were used most frequently for coordination between headquarters and subsidiaries. Independent sample t-tests revealed that there were no significant differences between headquarters and subsidiaries as far as hierarchical formal transmission channels were concerned (Exhibit 6.10). 3 Between peer subsidiaries, the same mean ranking for formal coordination instruments was observed: liaision personnel, followed by temporary task forces and permanent teams. The means for lateral transmission channels reached slightly lower values (Exhibit 6.11). When it came to informal channels, 29 per cent of subsidiary managers had working experience in headquarters and 22 per cent in another subsidiary

No. of companies

Exhibit 6.10

Hierarchical formal transmission channels

45 40 35

liaision personnel – hierarchical

30 25 20 15

permanent teams – hierarchical temporary task forces – hierarchical

10 5 0 1

Exhibit 6.11

2

3 4 5 Frequency

6

7

Lateral formal transmission channels

No. of companies

35 30 liaision personnel – lateral

25 20

permanent teams – lateral

15

temporary task forces – lateral

10 5 0 1

2

3

4 5 Frequency

6

7

Percentage of Headquarters Managers

Working Experience at Another Subsidiary

22

Working Experience at Culturally Distant Subsidiaries 11

Working Experience at Headquarters

29

Working Experience at Culturally Close Subsidiaries

30

Informal transmission channels

Percentage of Subsidiary Managers

Exhibit 6.12

41

Mentor in Culturally Close Subsidiary

45

Mentor in Headquarters

24

Mentor in Culturally Distant Subsidiary

18

Mentor in a Subsidiary

74

Participated in Executive Development Programmes with Several Subsidiaries

57

Participated in Executive Development Programmes with Several Subsidiaries

93

94

Effective Knowledge Transfer in MNCs

(Exhibit 6.12). More mentors were situated in headquarters. 30 per cent of headquarters managers were sent to culturally close subsidiaries to get work experience and only 11 per cent to culturally distant subsidiaries. Their mentors were rather found in culturally close subsidiaries too. In sum, slightly more headquarters (65 per cent) than subsidiary managers (63 per cent) had a mentor. The participation in executive development programmes where participants from several subsidiaries are involved was higher on the side of headquarters managers (74 per cent) than on the side of the subsidiary managers (57 per cent).

Knowledge management infrastructure When all surveyed units were analysed, the results in Exhibit 6.13 were generated for knowledge management infrastructure. To give more details on their means and medians, descriptives are added. Knowledge management technology that allowed people to collaborate inside the unit ranked among the most frequently used infrastructure in companies, followed by technology that allowed collaboration with people outside the unit. Another important component of knowledge transfer infrastructure was instruments to retrieve and use knowledge

Exhibit 6.13

Knowledge management infrastructure

Categorize product knowledge Categorize process knowledge Monitor practices Unit-internal collaboration Collaboration across units Single learning Multiple learning Search for new knowledge Map location of knowledge Product /process knowledge Market/compet. knowledge Generate new ideas 1.0

2.0

3.0

4.0 Frequency

5.0

6.0

7.0

Analysis and Results 95

about products and processes, and about markets and competition. This point is closely linked to technology to search for new knowledge, which constitutes another important facet of the company’s infrastructure. It can be concluded that collaboration tools (inside the unit and outside) are most important. Communication in the company seems to be the first priority. Closely behind communication tools ranked various technologies to retrieve, use and search for knowledge. Technologies to monitor competition and business partners fell in the same category. The lowest values were assigned to infrastructure that allowed learning as a group from multiple sources or at multiple points in time, and technology to map the location of specific types of knowledge. These findings pointed towards a low development of sophisticated learning and training infrastructure and the lack of directories or navigation instruments within the knowledge management systems. When the responses of headquarters and subsidiary managers were analysed in isolation, no major differences between the groups could be found in the results of independent sample t-tests.4

Knowledge transfer processes The measure of knowledge transfer capabilities included the usage frequency of different tools (Exhibit 6.14). 162 managers were asked to indicate on a scale from 1 (very infrequently) to 7 (very frequently), how often they used the different instruments. The result of this analysis is shown below. First, the use of the different tools is described and discussed, and the four knowledge conversion processes suggested by Nonaka and Takeuchi (1995) then are explored. The most frequently utilized knowledge management tools were face-to-face meetings, learning-by-doing and on-the-job training – all knowledge management tools closely tied to personal contacts. These were followed by case-based problem solving technology, databases, and intra-/internet. A separate analysis of headquarters and subsidiary cases using independent sample t-tests does reveal some differences. 5 For the tools capture and transfer of experts’ knowledge, decision support systems, employee rotation and projects across subsidiaries significant differences were found. In these cases headquarters’ means ranked higher than subsidiaries’. As the mean differences were very small (max. 0.92 for employee rotation), no important implications are expected. To give an example, Exhibit 6.15 presents the use of face-to-face meetings, which was the most prominent tool used by managers; fortyseven managers indicated that they used this tool very frequently.

96

Effective Knowledge Transfer in MNCs

Exhibit 6.14 Use of knowledge management tools

Capture and transfer Desision support sys. Learning by observ. Chat groups Employee rotation Subsidiaries project Analogies/metaphors Team collaboration Databases Web-based data ‘Yellow pages’ Apprentices/mentors Brainstorming camps Best practices Intra- and internet Learning-by-doing On-the-job training Problem solving tech. Face-to-face meeting 2

3

4

5

6

7

Frequency

Nobody indicated that face-to-face meetings were used very infrequently in his/her company. Another pattern is shown by chat groups/web-based discussion groups. Thirty managers indicated they use this tool very infrequently; only two managers used this tool very frequently. This phenomenon could be due to the fact that this tool requires a substantial amount of infrastructure investment and employee’s IT skills. Although these results seem to be unspectacular, it is important to recognize that a rather balanced use of knowledge management tools can be observed. Despite the prevalence of ‘conservative’ tools, such as databases or face-to-face meetings, which existed long before the trend towards a knowledge economy developed, nearly every company has established sophisticated tools such as web-based groupware or decision support systems. But the infrequent use of these tools demonstrates that most technology-intensive instruments are not yet accepted by employees. The four knowledge conversion processes proposed by Nonaka and Takeuchi (1995) are the next subject of analysis. Based on the categorizations developed by Becerra-Fernandez and Sabherwal (2001), the

Analysis and Results 97 Exhibit 6.15 Composition of knowledge transfer processes Capture and transfer of experts’ knowledge Decision support systems Problem solving technology

EXTERNALIZATION

Analogies and metaphors Pointers to expertise Chat groups/web-based discussion groups Team collaboration tools Best practices and lessons learned Databases

COMBINATION

Web-based access to data Intra- and internet pages Apprentices and mentors Brainstorming camps

SOCIALIZATION

Employee rotation Subsidiaries projects Learning-by-doing On-the-job training

INTERNALIZATION

Learning by observation Face-to-face meetings

different items were aggregated to result in ‘externalization’, ‘combination’, ‘socialization’, and ‘internalization’, as in Exhibit 6.15. As seen in Exhibit 6.16, internalization was the knowledge transfer process mostly used by the 162 units surveyed. Second ranked combination, followed by socialization and externalization. Again, independent sample t-tests6 were used to calculate differences between headquarters and subsidiaries. Results showed, that only socialization differed significantly between the groups, where the headquarters’ mean ranked slightly higher.

98

Effective Knowledge Transfer in MNCs

Exhibit 6.16 Knowledge transfer processes

6.0

5.5

Mean

5.0

4.5

4.0

3.5

3.0 Externalization Combination

Socialization Internalization

Earlier, it was hypothesized that units engaging in high outflow of knowledge would emphasize externalization and socialization processes, and units generating high inflows would make more use of internalization and combination. From a descriptive point of view, it can be concluded that units engage more in knowledge absorbing activities, where they have to internalize knowledge by transforming it from explicit into tacit knowledge and to combine separate explicit to become systemic explicit knowledge. This finding does not seem to be in line with the strategic mandate of the units surveyed discussed earlier. There, the majority of units were found to exhibit rather low knowledge inflows. What can be concluded at this stage is that units seem to use the processes internalization and combination more often. An explanation could be that the processes of a recipient unit are more demanding, or more timeintensive, so that units have to apply them more to integrate incoming knowledge. Alternatively, units might find it easier to cope with those processes. To shed some light on these issues, the concrete hypotheses will be tested later in the structural equation model. What has to be noted at this stage, however, is the problematic nature of the construct ‘strategic mandate’ in relation to knowledge transfer capabilities. One-way analysis of variance (ANOVA)7 was used

Analysis and Results 99

to test for differences between the four subsidiary groups in the development of knowledge transfer capabilities, and no significant results were found in any of the categories. Thus, as already stated, only knowledge inflows and outflows of units’ representing the units integration in the global network will be subject of the analysis in the structural equation model. The specific mandates of the units analysed are apparently not valuable for this particular sample.

Context factors Organizational distance Organizational distance was measured in two ways. As a first approach, the perceived similarity of business practices and operational mechanisms and of corporate culture and management style were explored. Again, this was done on the headquarters’ as well as on the subsidiary’s level. Subsidiary managers had to rate the two types of similarity vis-à-vis headquarters and vis-à-vis peer subsidiaries, resulting in the results shown in Exhibit 6.17. Business practices and operational mechanisms between the focal subsidiary and peer subsidiaries reached the highest value, meaning that they are perceived to be very similar on the subsidiary level. However, Exhibit 6.17 Subsidiaries’ perceptions of similarity vis-à-vis headquarters and peer subsidiaries

4.9

4.8

Mean

4.7

4.6

4.5

4.4

4.3 Practices HQ Practices subs

Style HQ

Style subs

100 Effective Knowledge Transfer in MNCs

corporate culture and management style seemed to be more similar between headquarters and the focal subsidiary. Generally, it should be noted that these mean ratings were all relatively close to each other, within a range of 4.3 to 4.9 on a scale from 1 to 7. Headquarters’ managers had to assess the similarity of business practices and operational and corporate culture and management style compared to culturally close and culturally distant subsidiaries (Exhibit 6.18). The pattern in Exhibit 6.18 clearly shows a much broader range of results (from 4.0 to 5.5) and a higher similarity between headquarters and culturally close subsidiaries with regard to both categories – business practices and operational mechanisms and corporate culture and management style. Summing up, subsidiaries indicated that they perceived higher similarity to their peers with regard to business practices and operational mechanisms, and higher similarity to headquarters with regard to corporate culture and management style. Headquarters found culturally close subsidiaries more similar (in terms of organizational distance) than culturally distant subsidiaries, in respect of both business practices and operational mechanisms and corporate culture and management style. Organizational distance was also calculated as a relative measure between headquarters and subsidiaries. For this purpose the strategic Exhibit 6.18 Headquarters’ perceptions of similarity vis-à-vis culturally close and distant subsidiaries

6.0

5.5

Mean

5.0

4.5

4.0

3.5 Practices close subs Style close subs Practices distant subs Style distant subs

Analysis and Results 101

orientation scales of headquarters and every subsidiary were compared and a distance measure, following the guidelines of Kogut and Singh’s (1988) cultural distance index, was generated. As a result, a mean of 1.9 was found among 118 cases. The highest value reached was 9.78, but 75 per cent of all cases lay below 2.22. To compare the measures, a Pearson correlation analysis was performed. To standardize data, the perceptional measure of organizational distance was generated as mean of both variables. The perceptional measure and the relative organizational distance index were entered as z-scores. However, the relative organizational distance measure did not correlate significantly with the aggregated perceptional measure. A reason for this divergence could be that the relative measure is, of course, based on a multitude of function-centred variables and does not build on general impression.

Cultural distance The question of cultural distance has been addressed in various ways (Exhibit 6.19). First, explicitly aiming at cultural problems occurring in knowledge management, three issues were surveyed on a scale from 1 to 7:

Exhibit 6.19 Impact of culture on knowledge transfers

5.5

5.0

Mean

4.5

4.0

3.5

3.0 National culture Misunderstandings Language differences

102 Effective Knowledge Transfer in MNCs

• The influence of national culture on the way of doing business • Misunderstandings and cultural conflicts emerging from knowledge transfers between different countries

• Language differences as obstacles in communication. The results were quite straightforward. According to the managers surveyed, national culture heavily influenced the way of doing business, but they did not believe that misunderstandings and cultural conflicts impeded the knowledge transfer between units in different countries. Language differences also did not seem to be major obstacles. These results might be due to respondents’ backgrounds. They all held top management positions and were, at least to some extent, involved in international operations. As the analysis of informal communication channels has shown, many respondents had gained working experience abroad and were sensitized to cultural differences. Second, Kogut and Singh (1988) indices were used to calculate the cultural distance. Only hierarchical relationships between headquarters and subsidiaries were analysed. The highest distance reached was 4.96, the mean scores 1.68. Kogut and Singh’s (1988) cultural distance index correlates significantly with the aggregated measure of the impact of culture on knowledge transfers. Astonishingly, the correlation was negative. Units which were more culturally distant, according to the Kogut and Singh’s definition, perceived less impact of cultural differences on their knowledge transfers. Although the operationalization of cultural distance adopted from Kogut and Singh’s (1988) Hofstede-based formula is commonly used in the business literature, this approach has received much criticism during recent years. Shenkar (2001) identifies deficient hidden conceptual and methodological assumptions in the cultural distance construct. Some refer directly to Hofstede’s concept of culture (cf. Au 2000), such as spatial and corporate homogeneity, or consider that the concept of long-term orientation (Hofstede and Bond 1988), which is especially relevant for research in Asian countries, has never been incorporated into the formula. It is has also been questioned, whether the different indices should be equally weighted or that the use of the concept of distance is inaccurate, as the definition implies symmetry and linearity (cf. Black and Mendenhall 1991). Alternatives to the widely used formula might be a more comprehensive view of national culture, including cognitive measures. To avoid such problems, suggested measures aimed explicitly at respondents’ subjective perceptions of cultural distance are used in the following analysis.

Analysis and Results 103

Third, the differences in knowledge management between culturally close and culturally distant subsidiaries are outlined in the discussion of the constructs ‘transmission channels’, ‘organizational distance’ and ‘benefit of knowledge transfers’. As can be seen from the detailed analyses, the well-known phenomenon that interaction with culturally close units seems to be easier generally holds true.

Knowledge transfer effectiveness Knowledge transfer effectiveness was measured by two scales – the benefit from knowledge transfers and the satisfaction with knowledge management.

Benefit from knowledge transfers On the subsidiary level, benefits from knowledge transfers from three sources were assessed (Exhibit 6.20):

• Headquarters • Culturally close subsidiaries • Culturally distant subsidiaries. As in other constructs, six different areas of data and know-how were surveyed. Technology and marketing know-how represented the most beneficial knowledge flows from headquarters (Exhibit 6.21). Distribution, purchasing and marketing data on competitors were found at a quite low level, the mean ranking being around 3.5–4 (on a scale from 1 to 7). Market data on customers was regarded as the least valuable knowledge inflows.

No. of companies

Exhibit 6.20

Subsidiaries’ benefits of knowledge transfers from headquarters

45 40 35 30 25 20 15 10 5 0

market data on customers market data on competitors marketing know-how distribution know-how technology know-how purchasing know-how 1

2

3

4 5 Benefit

6

7

104 Effective Knowledge Transfer in MNCs Exhibit 6.21 Subsidiaries’ benefits of knowledge transfers from culturally close subsidiaries

No. of companies

25 20

market data on customers

15

market data on competitors marketing know-how

10

distribution know-how technology know-how

5

purchasing know-how 0 1

2

3

4

5

6

7

Benefit

No. of companies

Exhibit 6.22 Subsidiaries’ benefits of knowledge transfers from culturally distant subsidiaries

50 45 40 35 30 25 20 15 10 5 0

market data on customers market data on competitors marketing know-how distribution know-how technology know-how purchasing know-how 1

2

3

4

5

6

7

Benefit

Market data on customers was the only area of knowledge transfers from culturally close subsidiaries whose mean ranked higher than those from headquarters (Exhibit 6.22). Marketing and technology know-how notably ranked significantly lower from this perspective. When knowledge transfers from culturally distant subsidiaries were compared to the previous exhibits, the pattern looked quite different and means did not reach the same level. The most beneficial transfer was perceived to be marketing know-how but the mean scored only at 2.5. The ranking was similar to the benefits from headquarters, but on a lower level.

Analysis and Results 105

At the headquarters level, information about the benefits from knowledge transfers from culturally close subsidiaries and culturally distant subsidiaries were provided. The results shown in Exhibit 6.23 and 6.24 were found. Interestingly, market data on customers and competitors were among the most beneficial categories. This pattern was very clear when inflows from culturally distant subsidiaries were considered. From culturally close subsidiaries, marketing know-how ranks second. To sum up, these findings point towards the fact that transfers from different units were perceived differently in terms of their benefits to local operations. On a general level, most beneficial to subsidiaries seemed

No. of companies

Exhibit 6.23 Headquarters’ benefits of knowledge transfers from culturally close subsidiaries 20 18 16 14 12 10 8 6 4 2 0

market data on customers market data on competitors marketing know-how distribution know-how technology know-how purchasing know-how

1

2

3

4 5 Benefit

6

7

Exhibit 6.24 Headquarters’ benefits of knowledge transfers from culturally distant subsidiaries

No. of companies

12 market data on customers

10 8

market data on competitors

6

marketing know-how

4

distribution know-how technology know-how

2

purchasing know-how 0 1

2

3

4 Benefit

5

6

7

106 Effective Knowledge Transfer in MNCs

to be transfers from headquarters. Only transfers of market data on customers from culturally close subsidiaries outpaced benefits from headquarters. Transfers from culturally distant subsidiaries were least beneficial, but showed the same pattern as benefits from headquarters. Astonishingly headquarters did not benefit most from higher-order knowledge such as know-how, but market data about competitors. One reason might be that the cost-benefit relation was perceived to be higher when no complicated processes of decoding and adaptation were involved. Another reason could be the aggressive competitive behaviour of the companies in the sample, which consisted largely of industry leaders. It is also feasible that subsidiaries were not given enough freedom to develop valuable knowledge – or, at least, that perceived competence varied across subsidiaries. This could explain why marketing know-how from culturally close subsidiaries was quite important to headquarters. What does not seem to add much value, neither to headquarters nor to subsidiaries, was purchasing know-how.

Satisfaction with knowledge management The other scale sought to measure knowledge transfer effectiveness was the satisfaction with knowledge management (Exhibit 6.25). The overall satisfaction with knowledge management was indicated by the availability of knowledge, the knowledge sharing and the improvement of effectiveness through available knowledge on a personal level, a unit level and a company-wide level. Two additional variables explored the clarity of the knowledge transferred and the speed of knowledge transfer. Exhibit 6.25 shows the means found for these items, by including all respondents, headquarters and subsidiary managers. Results showed that the improvement of effectiveness achieved the highest values in all categories – the personal, the unit and the whole organization. The general pattern showed that satisfaction with knowledge management was higher in the personal areas and declined when the unit level and the organizational level was considered. Overall, all means ranked above 3.5, which indicated a positive rating in all categories. It can thus be concluded that managers tended to be satisfied with knowledge management in their companies. Most satisfaction was found in the personal area. This could be explained by the higher autonomy in this field and the possibility to adjust knowledge management to personal needs. Least satisfaction was perceived in organization-wide knowledge sharing and in overall satisfaction with knowledge management in the global organization.

Analysis and Results 107 Exhibit 6.25 Satisfaction with knowledge management

Available (my tasks) Effective (my tasks) Satisfied (person) Available (unit) Sharing in unit Effective (unit) Satisfied (unit) Available (org) Sharing (org) Effective (org) Satisfied (org) Clarity Speed 0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Agreement Source: Diamantopoulos and Siguaw (2000, p. 15).

A separate analysis for headquarters and subsidiaries lead to very similar results. Using t-tests, no major differences could be detected.8

Hierarchical relationships and culturally close subsidiaries The following section, which constitutes the major part of the analysis, is dedicated to structural equation modelling. A short introduction to structural equation modelling is given in order to explain the aims of this method, and why it was chosen. This section elaborates on the theoretical composition of such models in general and on the construction of the ‘Model of Intra-MNC Knowledge Transfer’ in particular. Having chosen the optimal measurement model, the structural equation analysis is performed. Hypotheses tests are presented and the results discussed. Different sets of data which either refer to hierarchical, (between headquarters and subsidiaries), or lateral (between peer subsidiaries) relationships will be included. To contrast them with each other, further analyses are run and the models’ results are compared.

108 Effective Knowledge Transfer in MNCs

Structural equation modelling ‘Structural equation modelling (SEM) is a statistical methodology that takes a confirmatory (i.e. hypothesis-testing) approach to the analysis of a structural theory bearing on some phenomenon’ (Byrne 2001, p. 3). SEM thus falls into the category of multivariate methods, and typically conveys causal processes. The processes under study are represented by a series of structural (i.e. regression) equations. Covariance matrices depict the associations of observed variables, which lead to the explanation of relations of a smaller number of underlying constructs. Causal analysis enables the modelling and the estimation of complex structures of dependence simultaneously, which is especially useful in the behavioural sciences where researchers are often interested in studying theoretical constructs that cannot be observed directly (Byrne 2001, p. 4). This requirement is often recognized in the knowledge management literature, where constructs tend to be complex and, by definition, not directly observable. Many authors in the field have used SEM (cf. Simonin 1999b; Becerra-Fernandez and Sabherwal 2001; Gold, Malhotra and Segars 2001). A major characteristic of causal analysis is the differentiation of observable (manifest) and latent variables. The latter are more complex constructs which cannot be observed directly. Latent variables are commonly called factors and observed (or manifest) variables are named indicators. Moreover, it is helpful to distinguish between exogenous latent variables and endogenous latent variables. ‘Exogenous latent variables are synonymous with independent variables; they “cause” fluctuations in the values of other latent variables in the model . . . Endogenous latent variables are synonymous with dependent variables and, as such, are influenced by the exogenous variables in the model, either directly or indirectly’ (Byrne 2001, p. 5). Exhibit 6.26 provides guidance on how to construct a structural model. The structure of hypotheses includes endogenous and exogenous variables. The unit’s strategic mandate, its ability to transfer knowledge, and cultural and organizational distance are exogenous variables, whereas knowledge transfer capabilities and the effectiveness of knowledge transfer are endogenous. The development of the structural model follows the discussion of hypotheses and the underlying arguments (see Chapter 4). A model that specifies direction of cause from one direction only, as in this case, is termed recursive. One that allows for reciprocal feedback effects is called nonrecursive (Diamantopoulos and Siguaw 2000, p. 18). Exhibit 6.27 indicates the endogenous and exogenous factors in the recursive structural model. Exogenous factors are represented by shaded ellipses, endogenous factors by white ellipses.

109 Exhibit 6.26

Creation of a structural equation model

START

Past theories, evidence, experience, exploratory research

Yes

Irrelevant constructs included?

No

Identification of constructs (latent variables for inclusion in the model)

Designation of latent variables as exogenous or endogenous

No

Consider development of alternative models

No

Can each variable be clearly designated as exogenous or endogenous?

Irrelevant relationships included?

Yes

No

No

Yes Ordering of endogenous variables

Is order clear? Yes

Yes

Relevant constructs omitted?

Specification of expected relationships for each endogenous variable (including zero relationships)

No

No

Yes

Relevant relationships omitted?

No Final model for testing

Source: Diamantopoulos and Siguaw (2000, p. 15).

110 Effective Knowledge Transfer in MNCs Exhibit 6.27 Composition of latent variables

Strategic Mandate exogenous

Cultural Distance exogenous Knowledge Transfer Effectiveness endogenous

Value of Knowledge Stock exogenous

Knowledge Transfer Capabilities endogenous

Organizational Distance exogenous

Owing to the high theoretical requirements of the conceptualization of knowledge flows the model can become quite complex. A second level of latent variables is included, as some comstructs consist of multiple facets. This is shown in Exhibit 6.28. In contrast to the analysis of the model’s constructs only metric scales can be used for SEM. To allow for a more differentiated measurement, the original scales of knowledge inflows and outflows are used instead of the calculated strategic mandate. Moreover, the perceptional variables are used for organizational and cultural distance. As a result, all scales entered in the model are 7-point Likert scales. When it comes to construct-to-indicator relationships, recent research distinguishes between formative and reflective indicators (Diamantopoulos and Siguaw 2000, p. 14). A construct with reflective indicators means that the indicators are expressed as a function of the construct: in this case, the latent variable is thought to ‘cause’ observed indicators. A formative construct, however, is expressed as a function of those indicators: indicators are thought to ‘cause’ the latent variable. A characteristic of constructs with formative indicators is that the indicators of the same construct ‘can have positive, negative, or no correlation’ with one another (Bollen and Lennox 1991, p. 307). Validation of constructs with formative indicators rests mainly on the thoroughness with which the construct domain is tapped (i.e. content validity) (Johnson et al. 1996).

Analysis and Results 111

In Exhibit 6.28, reflective variables are labelled ‘R’. To describe the procedure of structural equation modelling, studies generally distinguish between the structural model and the measurement model (cf. Homburg and Pflesser 1999b; Byrne 2001). The structural model depicts the hypothesized relations between the latent variables – i.e. the constructs. The measurement model, however, relates the indicator variables to the factor variables. Exhibit 6.29 shows a decision tree for the construction of the measurement model. More detail on the adoption of existent scales and the development of new scales for the hypothesized model is provided on p. 76. It is assumed that each indicator represents an erroneous measurement of one, or several, latent variables (Homburg and Pflesser 1999b, p. 641). As it is highly unlikely that a perfect fit will be achieved, the model-fitting process can be summarized as follows (Byrne 2001, p. 7): Data = Model + Residual

Exhibit 6.28 Formative and reflective variables Cultural Distance exogenous Strategic Mandate exogenous Knowledge Outflow Knowledge Inflow

Cultural Distance

Knowledge Transfer Capabilities endogenous T1: Formal Channels

Value of Knowledge Stock exogenous Knowledge Stock

T2: Infrastructure R T3: Process Capabilities R

Organizational Distance Organizational Distance exogenous

Knowledge Transfer Effectiveness endogenous Satisfaction R Perceived Benefit

112 Effective Knowledge Transfer in MNCs Exhibit 6.29

Selection of variables

START

Past methodologies, scales, exploratory research

Are measures available for the latent variables in the model?

No

Consider development of new measure(s)

Yes Identify potentially relevant manifest variables to act as indicators for latent variables

Are multiple measures available for all constructs?

Yes

No

No

Selection of manifest variables for inclusion in measurement model

Is use of a single measure realistic?

Yes

Source: Diamantopoulos and Siguaw (2000, p. 17).

The residual represents the discrepancy between the observed data in the sample and the hypothesized model. In order to be of scientific use, a model has to be overidentified – i.e. the number of data points has to exceed the number of parameters to be estimated (Byrne 2001, p. 35). Of course, due the large number of indicators and the complex structure of relationships, an estimation of the original model including all indicators would lead to an underidentified

Analysis and Results 113

model. It was therefore decided to split the model into three partial models and reduce the number of included indicators. Exhibits 6.30 – 6.32 depict the specifications of these structural models. 9 As an extensive amount of variables were gathered, including interdependencies of various subsidiaries and headquarters at hierarchical and lateral levels, choices on which variables are most suitable for SEM had to be made. It was thus decided to include either hierarchical or lateral relationships and to limit the analysis to the relations to culturally close or culturally distant subsidiaries. As shown in Exhibit 6.33, the first set of data to be analysed was composed of headquarters relating to culturally close subsidiaries, and subsidiaries relating to their headquarters. Exhibit 6.30 Model 1

Strategic Mandate exogenous Outflow Inflow

T1: Formal Channels T2: Infrastructure R

Value of Knowledge Stock exogenous

Knowledge Transfer Capabilities endogenous

T3: Process Capabilities R

Knowledge Stock

Exhibit 6.31 Model 2

T1: Formal Channels T2: Infrastructure R T3: Process Capabilities R Knowledge Transfer Capabilities exogenous

Satisfaction R

Perceived Benefit

Knowledge Transfer Effectiveness endogenous

114 Effective Knowledge Transfer in MNCs Exhibit 6.32

Model 3

Cultural Distance exogenous Cultural Distance Satisfaction R

Organizational Distance

Perceived Benefit

Knowledge Transfer Effectiveness endogenous

Organizational Distance exogenous Exhibit 6.33 1st Dataset

1st DATASET

38 Headquarters

2nd DATASET

3rd DATASET

38 Headquarters 124 Subsidiaries

124 Culturally Close Subsidiaries

124 Culturally Distant Subsidiaries

Assessment of the model After the specification of the model (see also Appendix 1, p. 149) and the parameter estimation, the model has to be assessed (see Appendix 2, p. 153). If this step leads to negative assessment, a modification of the model is necessary. In case of positive assessment, a detailed interpretation of results must follow. Exhibit 6.34 shows that models 1–3 generally reached good measures of fit and could be assessed as reliable and valid at this stage. All NFI and CFI values between 0.918 and 0.961 were excellent as they were close to 1.00. Chi-square/df scores were above 2.5 in all cases

Analysis and Results 115 Exhibit 6.34 Global measures of fit of models 1–3 Model

Chi2

df

p-value

RMSEA

Chi2/df

NFI

CFI

Model 1

805.04

269

0.000

0.111

2.993

0.918

0.943

Model 2

768.66

205

0.000

0.131

3.749

0.919

0.939

Model 3

220.76

75

0.000

0.110

2.944

0.961

0.974

and reached even 3.749 in model 2, which could be seen as a very good reproduction of reality. RMSEA, however, was slightly too high in all cases, between 0.11 and 0.13. Ideal values would be smaller than 0.05 or lie between 0.08 and 0.10. The explanation could be that aggregated and reduced factor variables did not allow a close approximation of reality. To test the significance of hypotheses, one-tailed t-tests were used. The AMOS software module provides these results as critical ratios (c.r.). For models with more than 200 degrees of freedom, a minimum value of 2.345 is requested to meet α = 0.01 significance, 1.658 for α = 0.05 and 1,286 for α = 0.1. For more than 70 degrees of freedom, cut-off values of 2.381 (α = 0.01) 1.667 (α = 0.05) and 1.294 (α = 0.1) are recommended (Backhaus et al. 1996, p. 573). Based on these recommendations, models 1–3 were tested. The results showed that most hypotheses were significant at the 0.05 level. An overview of all hypotheses presented in the three models is now provided and their theoretical direction – i.e. positive or negative – indicated. The direct effects drawn from the structural equation models’ output symbolize the causal relationships. Unstandardized results are dependent on the unit of measurement of the observed variable and are thus difficult to interpret. Therefore, standardized values are typically examined (Byrne 2001, p. 89). As outlined above, some hypotheses did not reach significant levels, and some effects shown by the models were contradictory to the hypothese (Exhibit 6.35). The results are then discussed in detail and possible implications outlined.

Discussion of Results

Hypothesis 1 The development of knowledge process capabilities depends on the strategic mandate of the organizational unit.

116 Effective Knowledge Transfer in MNCs

Hypothesis 1a:

Organizational units characterized by high knowledge outflow and high knowledge inflow (for subsidiaries: Integrated Players) are expected to develop the highest 10 formal and informal transmission channels with peer subsidiaries/headquarters and the highest level of knowledge management infrastructure. Integrated players are expected to develop all knowledge processes at an equal and high level.

Hypothesis 1b: Organizational units characterized by high knowledge outflow and low knowledge inflow (for subsidiaries: Global Innovators) are expected to develop intermediate11 formal and informal transmission channels with peer subsidiaries/headquarters and an intermediate level of knowledge management infrastructure. Global Innovators are expected to emphasize externalization and socialization. Hypothesis 1c: Organizational units characterized by low knowledge outflow and high knowledge inflow (for subsidiaries: Implementers) are expected to develop intermediate12 formal and informal transmission channels with peer subsidiaries/headquarters and an intermediate level of knowledge management infrastructure. Implementers are expected to emphasize internalization and coordination. Hypothesis 1d: Organizational units characterized by low knowledge outflow and low knowledge inflow (for subsidiaries: Local Innovators) are expected to develop the lowest formal and informal transmission channels with peer subsidiaries/ headquarters and a low 13 level of knowledge management infrastructure. Local Innovators are expected to develop all knowledge processes at an equal and low level.

It was hypothesized that the strategic position in the organizational network would act as an antecedent to the development of knowledge transfer capabilities. As already discussed (p. 85), two strategic subsidiary mandates prevailed in the sample, Integrated Players and Local Innovators. At this stage, the impact of the units’ strategic mandate, as determined by existent knowledge flows, was tested in the structural equation model. This was also done for headquarters. To achieve comparable results for headquarters and subsidiaries, only hierarchical knowledge flows were included.

117 Exhibit 6.35

Hypotheses and direct effects

Hypotheses

Model 1 OUT---> T1 High levels of knowledge outflow positively affect the development of formal transmission channels

Direct Effect

Hypothesis accepted

0.976***



OUT ---> T2 High levels of knowledge outflow positively affect the development of knowledge management infrastructure

0.922***



OUT ---> T3 High levels of knowledge inflow positively affect the development of knowledge transfer processes

0.918***



IN ---> T1 High levels of knowledge inflow positively affect the development of formal transmission channels

not sig.

X

IN ---> T2 High levels of knowledge inflow positively affect the development of knowledge management infrastructure

–0.176***

X

IN ---> T3 High levels of knowledge inflow positively affect the development of knowledge transfer processes

–0.221***

X

ABIL1---> T1 A unit’s high value of knowledge stock positively affects the development of formal transmission channels

not sig.

X

ABIL1 ---> T2 A unit’s high value of knowledge stock positively affects the development of knowledge management infrastructure

–0.345***

X

ABIL1 ---> T3 A unit’s high value of knowledge stock positively affects the development of knowledge transfer processes

–0.330***

X

0.511**



–0.626**

X

Model 2 T1 ---> EFF1 A high development of formal transmission channels positively affects the perceived benefit of knowledge transfers T1 ---> EFF2 A high development of formal transmission channels positively affects the satisfaction with knowledge management

118 Exhibit 6.35

(continued)

Hypotheses

Direct Effect

Hypothesis accepted

T2 ---> EFF1 A high development of knowledge management infrastructure positively affects the perceived benefit of knowledge transfers

0.147**



T2 ---> EFF2 A high development of knowledge management infrastructure positively affects the satisfaction with knowledge management

0.452***



T3 ---> EFF1 A high development of knowledge management processes positively affects the perceived benefit of knowledge transfers

0.847***



T3 ---> EFF2 A high development of knowledge management processes positively affects the satisfaction with knowledge management

0.636***



0.967***



ORG ---> EFF2 High organizational similarity between the units – i.e. low organizational distance – positively affects the satisfaction with knowledge management

0.583***



CULT---> EFF1 High cultural similarity between the units – i.e. low cultural distance – positively affects the perceived benefit of knowledge transfers

not sig.

X

CULT ---> EFF2 High cultural similarity between the units – i.e. low cultural distance – positively affects the satisfaction with knowledge management

not sig.

X

Model 3 ORG ---> EFF1 High organizational similarity between the units – i.e. low organizational distance – positively affects the perceived benefit of knowledge transfers

Notes: *** significant at the 0.01 level (one-tailed). ** significant at the 0.05 level (one-tailed). * significant at the 0.10 level (one-tailed). Hypothesis accepted √. Hypothesis rejected X.

Analysis and Results 119

Looking at results, units characterized by high levels of outflow developed all knowledge transfer capabilities at a high level. The strongest impact was found on knowledge management infrastructure (0.922), but processes were also causally related to knowledge outflows by a regression coefficient of 0.918. The lowest, but still significantly positive, impact was seen on the development of formal transmission channels. Another picture was shown by units exhibiting high levels of knowledge inflows. While the impact on the development of formal transmission channels was not significant, the other two hypotheses reached significant levels but showed a negative instead of the hypothesized positive relationship. However, the negative empirical relationships did not reach higher regression scores than 0.221 and thus had only a low impact. Consequently, all hypotheses concerning the impact of high knowledge inflows on the development of knowledge transfer capabilities had to be rejected. The conclusion drawn from these findings was that units exhibiting high knowledge inflows did not develop their knowledge transfer capabilities at a high level. With regard to units that engaged heavily in knowledge outflows, the hypotheses could be confirmed. Thus, units with high knowledge outflow and high knowledge inflow did not necessarily develop the highest level of knowledge transfer capabilities. It could also be that units characterized by high knowledge outflows and low inflows reached equal levels of knowledge transfer capabilities. However, the measure of knowledge inflow might be inflated by perception differences and overestimation, resulting in inflow levels that were proportionally larger than the degree of transfer capabilities development. In this context, common method variance could play a role, as both categories were surveyed in one questionnaire. Additionally, those units characterized by high levels of knowledge inflow might not be cognizant of the channels, infrastructure and processes they used. To test whether differences between four groups (high inflow – high outflow; high inflow – low outflow; low inflow – high outflow; low inflow – low outflow), created through median splits in inflows and outflows, were significant, one-way ANOVA was used. Only variables where significant differences on the 0.1 level were found are reported in Exhibit 6.36. As shown above, only six variables exhibited significant differences between two of the four groups, and the mean difference was negligibly low. It can thus be stated that no important differences between the groups existed with regard to the characteristics of knowledge transfer capabilities.

120 Effective Knowledge Transfer in MNCs Exhibit 6.36

Differences between groups

VARIABLE

SIGNIFICANT DIFFERENCES BETWEEN GROUPS

Sig.

Infrastructure Categorize Product Knowledge

Low Inflow – Low Outflow AND Low Inflow – High Outflow

0.098

Categorize Product Knowledge

Low Inflow – Low Outflow AND High Inflow – High Outflow

0.032

Multiple Learning from Single Source and Time

Low Inflow – Low Outflow AND Low Inflow – High Outflow

0.098

Multiple Learning from Single Source and Time

Low Inflow – Low Outflow AND High Inflow – High Outflow

0.003

Map Location of Specific Type of Knowledge

Low Inflow – Low Outflow AND High Inflow – High Outflow

0.000

Retrieve and Use Knowledge About Products and Processes

Low Inflow – Low Outflow AND High Inflow – High Outflow

0.018

Analogies and Metaphors

Low Inflow – Low Outflow AND High Inflow – High Outflow

0.025

Brainstorming Camps

Low Inflow – Low Outflow AND High Inflow – High Outflow

0.087

Process

Summing up, there is empirical evidence that units characterized by high levels of knowledge outflow tended to highly develop all three knowledge transfer processes. Units heavily engaging in outflows did put efforts into the development of knowledge transfer capabilities to make sure that the knowledge they sent reached the receiver. On the other side, units did not seem to mind how they got the knowledge in – i.e. units characterized by high inflows did not develop high levels of knowledge transfer capabilities. It could be assumed that the development

Analysis and Results 121

of transfer capabilities was a task of those units which were in the sending position; this phenomenon could be interpreted as ‘ignorant’ behaviour of the recipient. However, when four groups were created, differences were significant for only six variables of knowledge management infrastructure and knowledge processes.

Hypothesis 2 A high value of knowledge stock positively affects the development of knowledge transfer capabilities.

Additional antecedents to the development of knowledge transfer capabilities were hypothesized to be units’ characteristics that supported the knowledge transfer, namely the value of knowledge. A high value of knowledge stock was seen as a unit’s ability to send knowledge, hypothesized to enhance the development of knowledge transfer capabilities. But results showed that all three hypotheses concerning the development of knowledge transfer capabilities had to be rejected. The impact on the development of formal transmission channels was insignificant. Although the two other hypotheses reached significance, the empirical direction, which was negative, did not correspond to the predictions. The explanation of these facts could lie in motivational aspects. Although the findings of earlier studies (Szulanski 1995; Foss and Pedersen 2002) concluded that motivational issues were not relevant and could be ignored, model results could be explained by integrating such points. Possibly, a unit which possessed a high value of knowledge stock that was non-duplicative and relevant to others did not want to give that asset away. Consequently, it did not engage in the development of transfer capabilities. As it is not possible to include such measures into the study ex post, a strong recommendation for further research is to account for such motivational factors. Another reason could be the lack of coordination in the MNC. Units did not share knowledge because they simply did not know that others had valuable knowledge which could be applicable in their own context. As shown in the analysis of the model’s constructs (p. 88) units tended to evaluate their own knowledge higher than others’. This could lead to a self-reliant attitude. If there seemed to be no valuable knowledge ‘out there’ it was not worth developing any transmission channels.

122 Effective Knowledge Transfer in MNCs

Hypothesis 3 Appropriately developed knowledge transfer capabilities (in terms of channels, infrastructure and processes) have a positive impact on the effectiveness of knowledge transfer. The impact of the three knowledge transfer capabilities on two components of knowledge transfer effectiveness was tested. As described earlier, knowledge transfer effectiveness was represented by the perceived benefit of knowledge transfers and the satisfaction with knowledge management. In this model, only the effectiveness of hierarchical transfers was explored. All hypotheses concerning the perceived benefit of knowledge transfers could be confirmed. The impact of the development of formal transmission channels, knowledge management infrastructure and knowledge processes was positive and highly, significant reaching values of 0.511, 0.147 and 0.847. Thus, the development of knowledge management infrastructure had only a small effect, while formal transmission channels and processes reached impact levels which were among the highest throughout the partial models. The development of formal transmission channels had a clearly positive impact on the perceived benefit, but satisfaction with these tools was significantly negative (−0.626). At first glance, it seemed paradoxical that an item had a positive effect on the perceived benefit but a negative impact on satisfaction, but a simple explanation is at hand. Formal transmission channels are often implemented in a centralized and strict manner and fulfil coordination and control functions. As control is often seen as coercive, satisfaction is rather low. However, formal transmission channels seem to be effective as they clearly augment the perceived benefit. Thus, they might be enforced, but they work efficiently. It has to be speculated why the development of infrastructure has only low effects on the perceived benefit of knowledge transfers. One reason could be that this category largely consists of technical facilities that link directorates across the firm. Often they are complex to understand and it is difficult for employees to get acquainted with such technologies. Maybe, employees take the cost of learning how to use these tools into account in their assessment. A further explanation could be that the beneficial effects of knowledge management infrastructure are not yet recognized. Theoretically, it could also be that the existence of infrastructure in the company is seen as a given prerequisite and the impact on

Analysis and Results 123

the effectiveness of knowledge transfers is not assigned adequate value. But this is seen as rather unlikely because, according to frequencies and descriptives, the technical state of development is still low in the majority of companies. The impact of knowledge management infrastructure on satisfaction was slightly higher than on the perceived benefit. This could be explained by the fact that infrastructure also comprises learning and training facilities which are often seen as employees’ fringe benefits. Moreover, as the analysis of constructs revealed, managers are least satisfied with knowledge sharing on the organization-wide level. Knowledge management infrastructure aims to enhance processes at this level. As outlined earlier, knowledge management processes refer to the practices of knowledge-sending and a knowledge-receiving unit. The high impact of knowledge management processes on perceived benefit could be attributed to the fact that they are tied to people and sometimes contain highly personalized elements. Other than mere infrastructure, which exists in a company twenty-four hours a day, these processes have to be applied in the right case at the right time. Moreover, as knowledge management processes always involve the sender and or the receiver, or both simultaneously, they are perceived to be crucial for the actual encoding and decoding processes. Failures at this stage could lead to the loss of knowledge and inhibit knowledge transfer. If applied adequately, knowledge management processes improve availability and accessibility, and make knowledge easy to decode. However, if these processes do not work adequately, recapturing of knowledge might become a time-intensive and strenuous task. The effect on satisfaction linked to the development of knowledge processes was also significantly positive (0.636). Again, as employees engage personally in these processes cooperative behaviour and the exercise of processes are strongly linked to the perceived outcome and associated with opportunity cost. As knowledge sharing is perceived as a timeconsuming task, people could as well dedicate their time to other duties – which often result in more direct and better visible outcomes. It might also be because of the more direct personal involvement that satisfaction with knowledge management is augmented.

Hypotheses 4 and 5 The lower the organizational and cultural distance between units the higher the effectiveness of knowledge transfer.

124 Effective Knowledge Transfer in MNCs

The effectiveness of knowledge transfer was hypothesized to be moderated by the organizational and cultural distance between the sending and the receiving unit. The impact of organizational similarity (the counterpart of organizational distance) on the perceived benefit of knowledge transfers was significantly positive, scoring 0.967. Satisfaction was equally positively influenced, ranking 0.583. This points towards the fact that similarity in organizational structure and style enhances the connectivity and knowledge can thus be easily transferred. Such effects have often been suggested in the literature and many studies have come to similar conclusions. This notion can be confirmed by this sample. For cultural distance, on the contrary, all results were insignificant. This may be attributed to the problematical nature of the construct in the measurement model. Such a tendency could already be assumed, looking at the descriptives of that construct. Although the impact of cultural differences has long been highlighted in the knowledge management literature (Barkema and Vermeulen 1997; Tenkasi 2000; Bhagat et al. 2002), managers surveyed did not see cultural distance as an impediment to effective knowledge transfer. The issues most often cited in the literature – language problems and misunderstandings based on cultural artifacts – did not seem to be an issue – at least at the top management level. However, results could be different if employees characterized by lower international exposure or a lower level of education were surveyed. From the above, it could also be concluded that in MNCs corporate culture becomes more important than country culture. Many authors (Adler 1983; Nohria and Ghoshal 1994) have pointed towards the fact that in large companies represented by many diverse national cultures, corporate culture has a considerable cohesive power which may also act as coordination and control instrument. It could be possible that, especially at the top management level, the power of corporate culture overrides the conflicting effects of different origins and socializations.

Hierarchical relationships and culturally distant subsidiaries As mentioned earlier, the models analyzed in the previous section include headquarters’ and subsidiaries’ data. Headquarters’ managers answered questions referring to two different subsidiaries, one being culturally close, one culturally distant. Subsidiary managers referred to hierarchical (i.e. headquarters) and lateral (i.e. another subsidiary) relationships. In the models discussed so far, the data on headquarters referring to culturally

Analysis and Results 125 Exhibit 6.37 2nd Dataset

1st DATASET

38 Headquarters

2nd DATASET

3rd DATASET

38 Headquarters 124 Subsidiaries

124 Culturally Close Subsidiaries

124 Culturally Distant Subsidiaries

close subsidiaries and subsidiaries referring to their headquarters was merged. Cross-validation was done by comparing the results of a second and a third set of data, as shown in Exhibit 6.37. While the second dataset is a slightly modified version of the first, including headquarters which relate to culturally distant subsidiaries and subsidiaries to headquarters, the third dataset includes subsidiary data only. In this case, subsidiary managers refered to their lateral relations with other subsidiaries. Thus, the third dataset has a smaller sample size, including only 124 cases. The analysis of the two additional sets of data is now performed. As the procedure has already been described in the previous section, the explanations are limited to the major steps. As the same model was used with slightly modified data, results are very similar to the first analysis. When it comes to the testing of hypotheses, more differences can be observed. In contrast to the first dataset, all hypotheses derived from knowledge inflow are insignificant. The hypothesis ‘A high development of formal transmission channels affects the satisfaction with knowledge management positively’ (T1 ------> EFF2), which is also rejected in the first analysis, has now turned insignificant (Exhibit 6.38).

Discussion of results It has to be emphasized that the 124 cases of the first analysis (i.e. subsidiaries that relate to their headquarters) are also included in the second

126 Exhibit 6.38

Hypotheses and direct effects (2nd Dataset)

HYPOTHESES

Model 1 OUT---> T1 High levels of knowledge outflow positively affect the development of formal transmission channels

DIRECT EFFECT

HYPOTHESIS ACCEPTED

0.992**



OUT ---> T2 High levels of knowledge outflow positively affect the development of knowledge management infrastructure

0.964***



OUT ---> T3 High levels of knowledge outflow positively affect the development of knowledge transfer processes

0.964***



IN ---> T1 High levels of knowledge inflow positively affect the development of formal transmission channels

not sig.

X

IN ---> T2 High levels of knowledge inflow positively affect the development of knowledge management infrastructure

not sig.

X

IN ---> T3 High levels of knowledge inflow positively affect the development of knowledge transfer processes

not sig.

X

ABIL1---> T1 A unit’s high value of knowledge stock positively affects the development of formal transmission channels

not sig.

X

ABIL1 ---> T2 A unit’s high value of knowledge stock positively affects the development of knowledge management infrastructure

–0.251***

X

ABIL1 ---> T3 A unit’s high value of knowledge stock positively affects the development of knowledge transfer processes

–0.236***

X

0.729**



Model 2 T1 ---> EFF1 A high development of formal transmission channels positively affects the perceived benefit of knowledge transfers

127

T1 ---> EFF2 A high development of formal transmission channels positively affects the satisfaction with knowledge management

not sig.

X

T2 ---> EFF1 A high development of knowledge management infrastructure positively affects the perceived benefit of knowledge transfers

0.164**



T2 ---> EFF2 A high development of knowledge management infrastructure positively affects the satisfaction with knowledge management

0.400***



T3 ---> EFF1 A high development of knowledge management processes positively affects the perceived benefit of knowledge transfers

0.664***



T3 ---> EFF2 A high development of knowledge management processes positively affects the satisfaction with knowledge management positively.

0.816***



0.999***



ORG ---> EFF2 High organizational similarity between the units – i.e. low organizational distance – positively affects the satisfaction with knowledge management

0.397**



CULT---> EFF1 High cultural similarity between the units – i.e. low cultural distance – positively affects the perceived benefit of knowledge transfers

not sig.

X

CULT ---> EFF2 High cultural similarity between the units – i.e. low cultural distance – positively affects the satisfaction with knowledge management

not sig.

X

Model 3 ORG ---> EFF1 High organizational similarity between the units – i.e. low organizational distance – positively affects the perceived benefit of knowledge transfers

Notes: *** significant at the 0.01 level (one-tailed). ** significant at the 0.05 level (one-tailed). * significant at the 0.10 level (one-tailed). Hypothesis accepted √. Hypothesis rejected X.

128 Effective Knowledge Transfer in MNCs

dataset. Essentially, only forty-eight cases vary compared to the initial analysis. Some interesting divergences were nevertheless found which account for the difference between headquarters’ relations to culturally close and culturally distant subsidiaries. As far as the hypotheses tested in model 1 are concerned, results corresponded to those of the first dataset. The impact of knowledge outflows on the development of all knowledge transfer capabilities was significantly positive. In all these cases, the effects scored slightly higher than in the first dataset, but generated only a negligible difference. The hypotheses regarding the effects of knowledge inflows were all rejected, though all respective results have turned insignificant. The unit’s ability to transfer knowledge – i.e. its knowledge stock – showed negative effects on the development of knowledge management infrastructure and processes – instead of positive – and no statistically significant impact on the development of formal transmission channels was found. Again these results reflected the above findings, but the negative empirical direction was higher for the second dataset. Model 2 essentially showed the same pattern as before. The impact of formal transmission channels on the perceived benefit of knowledge transfers reached a coefficient of 0.729, which was higher than in the previous analysis. This suggests that in a sample where headquarters and relations to their culturally distant (instead of culturally close), units are included the development of formal transmission channels has an even higher effect on the perceived benefit. In this case, however, the relation to the satisfaction with knowledge management turned statistically insignificant. While the effect of a sophisticated knowledge management infrastructure on the perceived benefits stayed comparatively low, the impact of knowledge management processes reached slightly lower levels for this data. The development of formal transmission channels thus has a higher impact on the perceived benefit than the development of knowledge management processes in this sample. From these results, it can be concluded that, when relations to culturally distant units are surveyed, formal transmission channels work more effectively, as they account for tighter control and can be communicated easily throughout the organization. Nevertheless, the impact of knowledge management processes was still considerable, reaching a coefficient of 0.664. The impact on satisfaction achieved by the development of knowledge management processes was the highest, generating a coefficient of 0.816. But the development of infrastructure also led to more satisfaction with

Analysis and Results 129

knowledge management in the company. In this case, the value of 0.400 was about the same as in the first analysis. As expected, model 3 led to very similar results. The impact of organizational similarity on the perceived benefit of knowledge transfers was slightly higher, whereas the impact on satisfaction ranked slightly lower. The hypotheses that cultural distance reduces the effectiveness of knowledge transfer again turned insignificant.

Lateral relationships of subsidiaries As a next step, the analysis was performed with a dataset that differed more from the first one. Only subsidiaries were included in the models and were tested with regard to their lateral relationships with their peer subsidiaries. As a consequence, the sample size was reduced to 124 cases as shown in Exhibit 6.39. Overall, a good model fit was achieved by all partial models. The pattern generated by the measures of fit, again, corresponded to the previous analyses. In this respect, the subsidiary dataset including lateral relationships seemed to fit as well as the ‘hierarchical’ datasets. Next, the hypotheses were tested. Owing the lower sample size of 124, different cut-off values were used to test for significance. For a 0.01 significance level, a minimum value of 2.358 had to be reached, for 0.05 a value of 1.658, and for 0.10 a value of 1.289 was required (Backhaus et al. 1996, p. 573). In contrast to the first and second datasets, only two Exhibit 6.39

3rd Dataset

1st DATASET

38 Headquarters

2nd DATASET

3rd DATASET

38 Headquarters 124 Subsidiaries

124 Culturally Close Subsidiaries

124 Culturally Distant Subsidiaries

130 Effective Knowledge Transfer in MNCs

hypotheses turned insignificant, ‘High value of knowledge stock affects the development of formal transmission channels positively’ (ABIL1 -------------> T1) and ‘Cultural distance between units affects the perceived benefit of knowledge transfers negatively’ (CULT -------------> EFF1). As both did not reached significance in the above analysis, these problematical cases were already expected. Important findings were that the same hypotheses that were rejected earlier showed the same directions as in the ‘hierarchical samples’. More inflow from other subsidiaries does not seem to support the development of knowledge transfer channels; a higher level of knowledge stock does not enhance development of these channels either. From this perspective, the same tendencies can be found in all three datasets. However, it is important to note that a significant negative impact of cultural distance on knowledge transfer effectiveness was found. This hypothesis never reached significant levels before. To illustrate these findings, Exhibit 6.40 shows which hypotheses have to be rejected.

Discussion of results Obviously, the third dataset contained some new information. Instead of focusing on hierarchical relationships, lateral knowledge flows between subsidiaries were now in the centre of study. Accordingly, some notable differences were found compared to the two prior analyses. Generally, model 1 produced the same results. The hypotheses concerning high knowledge outflow were all confirmed. Coefficients were significantly positive and ranked slightly lower, but still reached levels of 0.807–0.848. All other hypotheses had to be rejected. The only difference with regard to the first dataset was that high knowledge inflows had a significant but negative effect on the development of formal transmission channels. (Results were insignificant earlier.) An important issue, however, arose in model 2. The effect of formal transmission channels on the perceived benefit of knowledge flows, which was significantly positive in the previous analysis and reached coefficients of 0.729 and 0.511 respectively, now showed a negative empirical relationship. Therefore, the hypothesis that ‘A high development of formal transmission channels affects the perceived benefit of knowledge transfers positively’ had to be rejected. Even more amazing, the relation of the development of knowledge management infrastructure and the satisfaction, which had always been rejected before, was significant, generating a coefficient of 0.366. This results in the notion that the development of formal transmission channels does not affect the perceived benefit of knowledge transfers positively, but has a positive impact on the satisfaction

131 Exhibit 6.40

Hypotheses and direct effects (3rd Dataset)

HYPOTHESES

DIRECT EFFECT

HYPOTHESIS ACCEPTED

0.807***



OUT ---> T2 High levels of knowledge outflow positively affect the development of knowledge management infrastructure

0.848***



OUT ---> T3 High levels of knowledge outflow positively affect the development of knowledge transfer processes

0.827***



IN ---> T1 High levels of knowledge inflow positively affect the development of formal transmission channels

–0.590***

X

IN ---> T2 High levels of knowledge inflow positively affect the development of knowledge management infrastructure

–0.518***

X

IN ---> T3 High levels of knowledge inflow positively affect the development of knowledge transfer processes

–0.542***

X

ABIL1---> T1 A unit’s high value of knowledge stock positively affects the development of formal transmission channels

not sig.

X

ABIL1 ---> T2 A unit’s high value of knowledge stock positively affects the development of knowledge management infrastructure

–0.114*

X

ABIL1 ---> T3 A unit’s high value of knowledge stock positively affects the development of knowledge transfer processes

–0.151*

X

–0.886**

X

0.336*



Model 1 OUT---> T1 High levels of knowledge outflow positively affect the development of formal transmission channels

Model 2 T1 ---> EFF1 A high development of formal transmission channels positively affects the perceived benefit of knowledge transfers T1 ---> EFF2 A high development of formal transmission channels positively affects the satisfaction with knowledge management

132 Exhibit 6.40

(continued)

HYPOTHESES

DIRECT EFFECT

T2 ---> EFF1 A high development of knowledge management infrastructure positively affects the perceived benefit of knowledge transfers

0.264**



T2 ---> EFF2 A high development of knowledge management infrastructure positively affects the satisfaction with knowledge management

0.436***



T3 ---> EFF1 A high development of knowledge management processes positively affects the perceived benefit of knowledge transfers

0.381**



T3 ---> EFF2 A high development of knowledge management processes positively affects the satisfaction with knowledge management

0.822**



0.980***



ORG ---> EFF2 High organizational similarity between the units – i.e. low organizational distance – positively affects the satisfaction with knowledge management

0.391*



CULT---> EFF1 High cultural similarity between the units – i.e. low cultural distance – positively affects the perceived benefit of knowledge transfers

not sig.

X

CULT ---> EFF2 High cultural similarity between the units – i.e. low cultural distance – positively affects the satisfaction with knowledge management

–0.921**



Model 3 ORG ---> EFF1 High organizational similarity between the units – i.e. low organizational distance – positively affects the perceived benefit of knowledge transfers

Notes: *** significant at the 0.01 level (one-tailed). ** significant at the 0.05 level (one-tailed). * significant at the 0.10 level (one-tailed). Hypothesis accepted √. Hypothesis rejected X.

HYPOTHESIS ACCEPTED

Analysis and Results 133

with knowledge management. Such a finding supports the explanation of the previous cases that headquarters tend to enforce the establishment of formal channels and use them as control instrument – a fact that does not exist in the lateral view. In spite of this, formal transmission channels give subsidiaries the opportunity to build lateral ties that do not necessarily have to use the headquarters as a hub. This results in higher collaboration and the exchange of knowledge. Such autonomy-granting factors could improve the overall satisfaction with knowledge management in the company. When it came to the effect of knowledge management infrastructure, the impact on the perceived benefit was slightly higher than previously, but still low (0.264). Still, the results for the satisfaction were comparable. So was the effect of the development of knowledge management processes on satisfaction. The impact on the perceived benefit, however, was considerably lower – 0.381 instead of 0.847 and 0.664. It could be concluded that the effective exercise of knowledge management processes required close cooperation. Assuming that subsidiaries interact less on the lateral level than on the hierarchical level, the declining order of coefficients from culturally close and culturally distant to lateral flows could be explained. Knowledge management processes thus have the potential to lead to an increase in perceived benefit, but only when they are executed in a relatively close and controlled environment. Model 3 generated approximately the same values for the impact of organizational similarity on knowledge transfer effectiveness as the previous analyses – 0.980 for perceived benefit and 0.391 for satisfaction. Astonishingly, the negative impact of cultural distance was confirmed by this dataset. While the effect on the perceived benefit stayed insignificant, a highly negative impact on the satisfaction with knowledge management could be confirmed (−0.921). It could be that in hierarchical relationships, the roles are clearly defined and less cultural problems emerge because corporate culture and the experience of operations provide clear guidelines. When it comes to lateral relationships, it may not be clear to other subsidiaries which position in the network is assigned to their counterpart and cultural issues might play a more important role – or, at least, are thought to account for misunderstandings and failures in knowledge transfer. To sum up, Exhibit 6.41 shows an overall assessment of the hypotheses by the different sets of data.

134 Exhibit 6.41 General assessment of hypotheses Dataset 1

Dataset 2

Dataset 3

Headquarters

Headquarters

Subsidiaries

Close Subs

Distant Subs

Subsidiaries

Model 1 OUT---> T1 High levels of knowledge outflow positively affect the development of formal transmission channels



















X

X

X

X

X

X

X

X

X

OUT ---> T2 High levels of knowledge outflow positively affect the development of knowledge management infrastructure OUT ---> T3 High levels of knowledge outflow positively affect the development of knowledge transfer processes IN ---> T1 High levels of knowledge inflow positively affect the development of formal transmission channels IN ---> T2 High levels of knowledge inflow positively affect the development of knowledge management infrastructure IN ---> T3 High levels of knowledge inflow positively affect the development of knowledge transfer processes

135

ABIL1---> T1 A unit’s high value of knowledge stock positively affects the development of formal transmission channels

X

X

X

X

X

X

X

X

X





X

X

X



ABIL1 ---> T2 A unit’s high value of knowledge stock positively affects the development of knowledge management infrastructure

ABIL1 ---> T3 A unit’s high value of knowledge stock positively affects the development of knowledge transfer processes

Model 2 T1 ---> EFF1 A high development of formal transmission channels positively affects the perceived benefit of knowledge transfers T1 ---> EFF2 A high development of formal transmission channels positively affects the satisfaction with knowledge management

136 Exhibit 6.41 (continued) Dataset 1

Dataset 2

Dataset 3

Headquarters

Headquarters

Subsidiaries

Close Subs

Distant Subs

Subsidiaries

T2 ---> EFF1 A high development of knowledge management infrastructure positively affects the perceived benefit of knowledge transfers

























T2 ---> EFF2 A high development of knowledge management infrastructure positively affects the satisfaction with knowledge management T3 ---> EFF1 A high development of knowledge management processes positively affects the perceived benefit of knowledge transfers T3 ---> EFF2 A high development of knowledge management processes positively affects the satisfaction with knowledge management

137

Model 3 High organizational similarity between the units – i.e. low organizational distance – positively affects the perceived benefit of knowledge transfers













X

X

X

X

X



ORG ---> EFF2 High organizational similarity between the units – i.e. low organizational distance – positively affects the satisfaction with knowledge management CULT---> EFF1 High cultural similarity between the units – i.e. low cultural distance – positively affects the perceived benefit of knowledge transfers CULT ---> EFF2 High cultural similarity between the units – i.e. low cultural distance – positively affects the satisfaction with knowledge management

Notes: Hypothesis accepted √. Hypothesis rejected X.

7 Conclusion, Limitations and Implications

Conclusion This study has investigated knowledge transfers between dispersed MNC units. The aim was to shed more light on the effectiveness of these transfers. As shown in the literature review, the importance of knowledge management in the MNC is indisputable and management of dispersed knowledge is not only quite important but critically affects organizational design as well as the MNC’s strategic configuration. Many authors have recently addressed the topic of knowledge management in various areas, but only a few have systematically investigated intra-MNC knowledge transfers. Even though the works cited in the ‘state-of-the-art’ review add much to our knowledge, it was also shown that research is largely fragmented and takes different perspectives. Transfer is mostly seen as beneficial per se and the effectiveness of cost-intensive intra-MNC knowledge transfers is hardly ever questioned. In an attempt to combine some of these perspectives while concentrating on the effectiveness of knowledge transfer, a comprehensive model of knowledge transfer within MNCs was developed. It was argued that effective knowledge transfer is contingent on the development of several knowledge transfer capabilities which, in turn, are developed in response to the unit’s strategic position in the network and the unit’s value of knowledge stock. Moreover, two contingency factors, organizational and cultural distance, were seen as decisive moderators to effective transfer. Pursuing a nodal level of analysis, the determinants of effective knowledge transfer were tested empirically. The study focused on lateral and hierarchical knowledge transfers of 162 MNC units (thirty-eight headquarters and 124 subsidiaries). While most of the measures applied were drawn from major earlier studies, the remaining were developed 138

Conclusion, Limitations and Implications 139

specifically for this survey. They reached a high level of quality in terms of reliability and validity and were found appropriate to be entered into the structural equation model. The results of the general models were affirmative, supporting many of the hypotheses proposed. The results can be extended to the research fields of organizational networks, coordination mechanisms, capabilities, knowledge management tools and infrastructure and context. Nine of the major findings are that:

• The network perspective is applicable to the sampled firms. • The majority of units engage either in low inflows and low outflows or high flows in both directions.

• The formal transmission channels used in hierarchical relations add • • • • • •

considerably to beneficial knowledge transfer, but they do not have a positive impact on the satisfaction with knowledge management. In lateral relationships, formal channels have a negative impact on the benefit of knowledge transfers but a positive impact on satisfaction. While units showing high levels of knowledge outflows were found to develop knowledge management capabilities at a high level, units exhibiting a high level of inflow did not show a positive development. No evidence was found that a high value of knowledge stock supports the development of transfer capabilities. ‘Conservative’ knowledge management tools such as databases or face-to-face meetings are dominant. Culture was found to be insignificant in all datasets including hierarchical relationships, but showed a positive impact in the lateral sample. Organizational similarity has a highly positive impact on knowledge transfer effectiveness.

The issues mentioned above will now be discussed in more detail.

The network perspective The theoretical basis of this study is the network model of the MNC. According to this perspective, the strategic tasks of organizational units should differ according to their relative level of specialization. Headquarters are not necessarily assigned a central, hierarchical position; all subsidiaries and the headquarters are part of an interdependent network. This implies that headquarters do not have to play a dominant role in the organization and thus the firm’s ability to react flexibly to fast changing market conditions is increased. Among the sampled firms, centralized functions, such as technology and purchasing, seem to lie with the headquarters whereas local market analysis is largely done by

140 Effective Knowledge Transfer in MNCs

subsidiaries. This points towards a rather traditional, hierarchical position of headquarters. As far as knowledge flows are concerned, most headquarters are characterized by high knowledge outflows and high knowledge inflows. Only the smallest group engages in high knowledge outflows and low knowledge inflows – which reflects a traditional position. The underlying rationale is that firms usually expanded by so-called ‘forward’ transfers – i.e. from the centre to the periphery – and were not supplied with much knowledge from their subsidiaries (Vernon 1966). Traditionally, the flows back to the headquarters included only standardized reports or financial data. Only recently has the phenomenon of transfers of knowledge from the subsidiary to the headquarters, so-called ‘reverse flows’, been investigated (Hakanson and Nobel 2000, 2001; Frost 2001; Chini, Ambos and Schlegelmilch 2003; Zhou and Frost 2003). In this research, the headquarters surveyed seem to receive a lot of these ‘reverse’ flows and also benefit from them. Compared to regional headquarters, a larger percentage of global headquarters receives high inflows. Asakawa and Lehrer (2003), however, found that regional headquarters manage local innovation relays and are thought to be more involved in ‘reverse’ transfers than global headquarters. Overall, regional headquarters in this sample follow the same pattern as most other units – exhibiting either high inflows and high outflows or low inflows and low outflows. In the literature, the role of regional headquarters has been investigated, if at all, from the perspective of coordination and control than from a knowledge transfer point of view. Only recently, has research on regional headquarters focused on the identification and mobilization of knowledge at that level (Asakawa and Lehrer 2003). Regional offices are supposed to function as innovation relays, linking local knowledge to the MNCs’ global operations. The few regional headquarters surveyed in this study predominantly engage either in high knowledge flows in both directions or show a low level of inflows and outflows. Whereas the high inflow/high outflow type can be seen as such a relay of innovation, the existence of the other type cannot be explained. Arbitrary clustering of different national units may have led to such patterns; alternatively, the sampled firms are showing only ‘trace elements’ of a network organization, but have not yet reached the fully developed form (see also Wolf 1997). As far as the differences between individual subsidiaries were concerned, it became obvious that not all of them were assigned the same strategic mandate. However, in contrast to the four mandates suggested by Gupta and Govindarajan (1991), only two different types prevailed.

Conclusion, Limitations and Implications 141

Subsidiaries seemed to be characterized either by high inflows and high outflows (Integrated Players) or low inflows and low outflows (Local Innovators). The most interesting finding in this regard was that knowledge inflows and outflows seemed to develop simultaneously. Another remarkable aspect was that high knowledge outflows had a positive effect on the development of transfer capabilities, whereas high levels of inflow showed no significant effect on capability development. In sum, the network perspective provides a useful theoretical framework for analysing this sample. Because of the lack of a dynamic perspective, the network cannot be investigated in depth, but it is important to show that the average company follows a heterarchical rather than a hierarchical model. It seems that firms are developing towards networks but in many cases the headquarters is still the centre of gravity. To investigate this issue in more depth, longitudinal studies may be needed.

Coordination mechanisms As far as coordination between units is concerned, formal and informal mechanisms, which serve as knowledge transmission channels, have been surveyed. In both lateral and vertical relations, liaison personnel are most frequently employed, followed by temporary task forces and permanent teams. This accounts for the fact that connectivity between units is reinforced rather by integrative mechanisms than on a temporary basis or in the form of specialized teams. Among informal channels, executive development programmes across subsidiaries and the use of mentors dominate. Gupta and Govindarajan (2000) found that units tightly controlled in the form of formal and informal socialization generate high knowledge outflows. This study suggests that the effectiveness of such flows can be called into question. As shown in the results section, the formal transmission channels used in hierarchical relations add considerably to beneficial knowledge transfer. But they do not have a positive impact on satisfaction with knowledge management. However, when culturally close subsidiaries are included in the analysis instead of culturally distant subsidiaries, formal channels add more to the benefit, though still without a positive influence on satisfaction. This can be explained by the fact that coordination and control mechanisms are seen as coercive as they are imposed by headquarters for control reasons in hierarchical relationships. However, they seem to work effectively, as they contribute significantly to the perceived benefit. The hierarchical point of view thus confirms Gupta and Govindarajan’s (2000) findings that knowledge flows through formal channels are beneficial, but they do not augment the satisfaction with knowledge transfers.

142 Effective Knowledge Transfer in MNCs

On the subsidiary level, where Gupta and Govindarajan (2000) also found support for their hypothesis that tightly controlled units exhibit high knowledge outflows, this research generates different results. Astonishingly, the impact totally changes when lateral relationships are considered. Formal channels then have a negative impact on the benefit of knowledge transfers but a positive one on satisfaction. In this case, the lateral ties seem to be voluntarily built and not enforced by headquarters. Lacking a direct link, such relationships cannot be managed from headquarters. Therefore more autonomy is granted, augmenting satisfaction. The other side of the coin is that the perceived benefit from knowledge transfers is not coupled to lateral formal coordination. No positive relationship could be found. More research applying more complex multirespondent designs and accounting for perception gaps might be helpful to investigate these issues further. As increased interaction between subsidiaries can be seen as a way of building social capital, this finding does not support the theory of Nahapiet and Ghoshal (1998). The authors suggest that social capital enhances the firm’s combinative capability and ultimately adds to value creation. Gold, Malhotra and Segars (2001) also focus on the role of social capital and find that technology, structure and culture form a definitional basis for the theoretical framework of social capital and positively impact key aspects of organizational effectiveness. In the study at hand, such creation of social capital was found to have a positive impact on satisfaction with knowledge management. However, the effect on additional benefit to operations – which is comparable to value creation – could not be confirmed.

Capabilities Several knowledge management studies have researched the transfer of capabilities – e.g. production capabilities – (Zander and Kogut 1995; Mowery, Oxley and Silverman 1996; Subramaniam and Venkatraman 2001). To avoid confusion, it needs to be emphasized once again that in this study the specific capabilities to transfer knowledge were at the centre of analysis. This approach is in line with Gold, Malhotra and Segars (2001, p. 206) whose results suggest that theories of knowledge capabilities provide a rich resource for developing empirically based studies. The authors also stress that capabilities can provide a useful benchmark for knowledge management within the firm, but are complex not only in definition but also in operationalization. Notwithstanding this complexity, integrating the concept of mutually reinforcing capabilities into the model of knowledge transfer is a vital task. The development of

Conclusion, Limitations and Implications 143

transmission channels, infrastructure and processes was used as an additive construct to mirror knowledge transfer capabilities. While a high value of knowledge stock residing in a unit was found to have a positive influence on knowledge outflows in Gupta and Govindarajan’s (2000) study, it does not support the development of knowledge transfer capabilities in this study. It has to be recognized that the underlying operationalization is different. While Gupta and Govindarajan (2000) use secondary data – namely mode of entry and subsidiary size – a perceptional measure was applied in this study. The hypothesis that a high value of knowledge stock – which symbolizes the units’ potentiality to engage in knowledge transfer – positively affects the development of knowledge transfer capabilities was clearly rejected. This is a rather surprising fact that might be explained only by motivational factors. To the author’s best knowledge, however, motivational issues have not been found to play a dominant role in any major studies (Szulanski 1996; Gupta and Govindarajan 2000; Foss and Pedersen 2002). From a network perspective, some interesting results were received. While units showing high levels of knowledge outflows were found to develop knowledge management capabilities at a high level, units exhibiting high inflow did not show a positive development. This might be explained by the fact that the ‘receivers’ of knowledge are not cognizant of the channels, the infrastructure nor the processes they use. Further investigation of these phenomena must be left to future research.

Knowledge management tools and infrastructure The most utilized knowledge management tools are face-to-face meetings, learning-by-doing and on-the-job training – all knowledge management tools closely tied to personal contacts. Overall results indicate a balanced use of ‘codified’ and ‘personalized’ (Hansen, Nohria and Tierney 1999) knowledge management tools. Despite the prevalence of ‘conservative’ tools, such as databases or face-to-face meetings, which existed long before the trend towards a knowledge economy was established, nearly every company has established sophisticated tools such as web-based groupware or decision support systems. But the infrequent use of these tools demonstrates that most technology-intensive instruments are not yet accepted by employees. No matter what relationships are included into the analysis, there is evidence that infrastructure adds to overall knowledge transfer effectiveness. However, the impact of highly developed infrastructure on the perceived benefit is lower than the effect of the other knowledge transfer capabilities, but has a high bearing on satisfaction with knowledge management.

144 Effective Knowledge Transfer in MNCs

One possible explanation is that infrastructure, comprising features such as learning tools, is seen as a fringe benefit or a corporate incentive. On the other hand, its small impact on perceived benefits can be attributed only to lack of skills and resistance to technical features. The importance of the other two capabilities, formal transmission channels and processes, may outweigh the impact of infrastructure on effectiveness.

Context Empirical observations suggest a consistent pattern of results for contingency factors. As in many other studies, context dissimilarities were expected to be obstacles to effective knowledge transfer – despite recognizing the innovation potential of loose ties (see Chapter 3). The gap between theory and empirical findings in this area could also be confirmed by this study. Generally, theoretical studies tend to argue that cultural differences impact the process of knowledge transfer. However, many studies clearly failed to find empirical support (Lyles and Salk 1996; Simonin 1999b; Zhou and Frost 2003), while others confirmed their hypotheses (cf. Kogut and Singh 1988; Ambos 2002). This study ranks among the first category, which is well represented in the research on international knowledge transfers. In his study about the transfer of marketing know-how across strategic alliances Simonin (1999b), for example, did not confirm the hypothesis that cultural distance impedes knowledge transfer. Lyles and Salk (1996) did not find any significant effects of cultural misunderstandings/differences and knowledge acquisition. Zhou and Frost (2003) had to reject their hypothesis that cultural distance alters reverse knowledge transfers. A way out of this dilemma is proposed by Subramaniam and Venkatraman (2001), who expect to get more insight when focusing on tacit differences among countries – despite the intrinsic difficulties in doing so. What could also explain these non-findings is that contrary forces are at work: On the one hand, cultural differences cause misunderstandings and complicate the knowledge transfer while, on the other, new knowledge – perceived as beneficial – is likely to originate in unfamiliar contexts. The results found in this study are somewhat unique. Culture was found to be insignificant in all datasets including hierarchical relationships, but cultural distance had a negative impact on knowledge transfer effectiveness in the lateral sample. However, the result is significant only for satisfaction with knowledge management. This means that cultural distance has a negative effect on satisfaction only when knowledge transfers between peer subsidiaries are concerned. One could go as far as

Conclusion, Limitations and Implications 145

to assume that lateral exchange is per se more spontaneous and informal and, thus, involves more personal relations and social capital, where cultural differences are addressed directly. However, to draw firm conclusions, more empirical work is needed. As far as organizational isomorphism is concerned, the results illustrate that organizational similarity highly impacts knowledge transfer effectiveness. It seems that in this matter – which is less researched than cultural distance in the field of knowledge management – not much empirical support has been found either. Simonin (1999b) for example, did not find an impact of organizational distance and concludes that (marketing) functions are relatively homogeneous across contexts. Teigland, Fey and Birkinshaw (2000) conducted a qualitative study about knowledge dissemination in R&D and found that a one-company culture is a key facilitator for knowledge transfer. The conclusion on contingency factors’ impact on knowledge transfer effectiveness is that the importance of organizational similarities prevails in MNCs which are trying to become a one-company culture (Teigland, Fey and Birkinshaw 2000) or even a ‘knowledge culture’ (De Long 1997; Hauschild, Licht and Stein 2001).

Effectiveness The main rationale underlying this study is that not only the existence of knowledge transfers but also the effectiveness of these transfers is critical for the continuous (re-)creation of the MNC’s competitive advantage. It was convincingly shown that effective transfer is, indeed, contingent on many factors – predominantly on the development of knowledge transfer capabilities and on organizational similarity. By far the highest impact among the capabilities across all models was reached by the knowledge process capabilities. As effectiveness is difficult to measure directly, perceived benefit and satisfaction were used as proxies. As illustrated, these two constructs produce divergent results and it has to be concluded that different factors influenced perceived benefit and satisfaction. This is one of the major findings of this study, as the distinction between perceived benefit and satisfaction will allow managers to adopt appropriate strategic actions.

Limitations As with all social science models aiming to depict complex relationships, some important limitations need to be considered. One corollary already

146 Effective Knowledge Transfer in MNCs

mentioned earlier is that of a stable perspective. To survey network relationships, the development of knowledge transfers and their integration over time a dynamic perspective would have been advantageous. However, such a research design is difficult when a large sample of MNCs is targeted. One-firm studies (cf. Dyer and Nobeoka 2000) are better suited to investigate such network characteristics but generalizability of results is limited. Alternatively, longitudinal studies could shed more light on these issues. The main constraint in this specific case was the time frame for the research. Several studies solve that problem by focusing on secondary data – primarily patent citations (Cantwell and Janne 1999; Frost 2001; Zhou and Frost 2003). But given the methodological and conceptual limitations of patent citations (see also Criscuolo 2003), it was considered important to gather primary data. Another issue is that findings are based on self-reported data, entailing potential respondent bias or general method variance. However, multiple assessments (e.g. reliability, exploratory factor analysis, confirmatory factor analysis) reported good properties. This supports the validity of the results. Moreover, like most social science models, that developed in this study excludes some potentially important factors. Although several theoretical perspectives were addressed, including the main measures and controlling for side effects, other forces – not included in this model – might be at work. It would have gone beyond the scope of this study to include diverse sources of knowledge (Foss and Pedersen 2002) as well as its re-use and application. Although only a limited spectrum was investigated, a unit’s external knowledge access is also characterized by its network position (Tsai 2001) and thus was surveyed indirectly. Last but not least, this study pioneered the exploration of knowledge transfer effectiveness where the development of a viable measure is critical. It was shown that perceived benefit and satisfaction are two facets of effectiveness, which are sometimes influenced by different factors. Despite these limitations, the study makes some valuable contributions to practice and identifies some potentially important directions for future research.

Implications for future research Given the usual limitations of perceptual measures, objective measures on knowledge transfer effectiveness would be useful in future research; a multifacet construct probably is needed. The measures developed in this research exhibit good quality of reliability and validity and should provide a useful tool for further investigation into this topic.

Conclusion, Limitations and Implications 147

A promising avenue for further research could also centre more on the motivations of knowledge transfers and on the actual outcomes of effective transfer. It would be valuable to find out in which situations knowledge is transferred – e.g. if the unit decides to distribute the knowledge or if other units request this knowledge. In contrast to findings of earlier research, motivational factors and protectiveness could play a role. On the other side of the spectrum, the integration of knowledge is vital. Some studies, especially from the international JV literature (cf. Lyles and Salk 1996; Griffith, Zeybek and O’Brien 2001), have already provided some useful insights in this regard. However, researchers have to go further and question the overall contribution of knowledge flows to the firm’s operations. At the moment, authors unanimously call for an increase of knowledge flows (the reasons are discussed in this study), but a critical and systematic investigation of the cost-benefit relation of such an increase would be welcome. Acknowledging the methodological limitations of research questions concerning the ultimate benefit of knowledge flows, this study has made only a small first step developing a critical view on knowledge transfers by arguing that effectiveness – not mere transfer – is the decisive dimension. More light needs also to be shed on the issue of context similarities. Although the topic is present in almost every study in the international arena, researchers find divergent results without a clear trend in either direction. Of course the conceptualization of culture is an unsolved problem in social science, the issue often being naively circumvented by using Hofstede-based measures. Looking at other disciplines, such as cultural studies, could add to a more realistic approach – which will undoubtedly be more complex and, thus, harder to integrate. The specific suggestion of this research is to approach the topic of knowledge flows across different contexts from two perspectives. First, with regard to novelty, innovation potential and usefulness. Secondly, centring on the transmission problems which arise from de- and re-contextualization.

Managerial implications Although the results of this study cannot address all the potential obstacles that managers may face in their quest to create knowledgebased organizations, it implies that certain factors have an impact on the effectiveness of knowledge transfers, and might help managers to cope with this topic. Clearly, the results suggest that managers must first assess the unit’s strategic position in the network. Less important, however, is the unit’s

148 Effective Knowledge Transfer in MNCs

value of knowledge stock. It would thus be advisable to stay well informed about the network configuration of the firm and to acquire as much network knowledge as possible. Knowledge about the network configuration is a potential source of power. If the unit’s value of knowledge stock has no effect on the development of transfer capabilities it suggests that units holding attractive knowledge are not likely to distribute this knowledge generously. Managers have then to search for this knowledge actively in order to make it available for the whole organization. The empirical observations suggest that satisfaction with knowledge management and the perceived benefit from knowledge transfers are only partly influenced by the same factors. While formal coordination mechanisms clearly augment the benefit from inflows, they do not increase satisfaction. In contrast, highly developed infrastructure rather increases satisfaction but has only a small impact on perceived benefit. In this respect, this study provides some insight and potential means to influence these dimensions. Knowledge process capabilities are responsible for the highest increase in benefit. Although a systematic development of these capabilities is suggested, such a step has to be undertaken cautiously. All knowledge transfer capabilities can be regarded as mutually reinforcing, and neglecting one is likely to alter the whole mechanism. A detailed balancing of specific capabilities would go beyond the scope of this research, however. What seems to be a particularly relevant finding for managers in charge of international transfers is the effect of organizational and cultural dissimilarities. This study clearly suggests that in hierarchical relationships, organizational homogeneity is vital. However, no impact of culture could be found. Advice to managers would be that continuous investment in a one-company culture pays off in knowledge transfer effectiveness. When focusing on lateral relationships between subsidiaries, national culture does have a detrimental effect; probably more ‘mentality-building’ has to be carried out in this field.

Appendix 1 Specification of the Measurement Model

The specification of the measurement model includes the choice of variables included in the model. Latent and manifest variables have to meet several prerequisites, such as unidimensionality and reliability, to achieve a good model fit in the final calculations. These criteria are now explained and the exclusion of several items described for the dataset ‘Hierarchical Relationships and Culturally Distant Subsidiaries’.

Unidimensionality of the constructs Prior to the assessment of the measurement model’s reliability and validity, the constructs’ unidimensional characteristics have to be confirmed. Unidimensionality of a measurement model is given if different sets of indictors identify one factor, preferably the underlying construct (Andres 2003, p. 182). If a measure is unidimensional, then all the items measure the same trait. However, the amount of error of measurement in those items may vary (Anderson and Gerbing 1982). Exhibit A1.1 shows the variables that were eventually included in the models. The CFA (confirmatory factor analysis) procedure basically follows the same model as causal analysis but includes only a single construct in place of the whole model. The aim is a close reproduction of the empirical matrix of covariances.

Exhibit A1.1 Unidimensionality of constructs CONSTRUCT/ INDICATOR

CRONBACH ALPHA

Out: Strategic Mandate out_h_ma out_h_di out_h_te out_h_pu

0.8775

In: Strategic Mandate In_h_ma In_h_di In_h_te In_h_pu

0.8542

ITEM-TO-TOTAL CORRELATION

FACTOR LOADING

VEA 73.15%

0.7365 0.7712 0.7095 0.7227

0.881 0.860 0.844 0.836

0.7195 0.7133 0.6610 0.6898

0.853 0.848 0.829 0.806

69.60%

149

150 Exhibit A1.1

(continued)

CONSTRUCT/ INDICATOR

CRONBACH ALPHA

Abil: Knowledge Stock sto_ma sto_di sto_te sto_pu

0.8176

T1: Formal Channels co_li co_tf co_pt

0.5197

T2: Infrastructure infra_10 infra_9 infra_11 infra_6 infra_8 infra_12

0.8700

T3: Processes cap_1 cap_2 cap_3 cap_4 cap_5 cap_6 cap_7 cap_8 cap_9 cap_10 cap_11 cap_12 cap_13 cap_14 cap_15 cap_16 cap_17 cap_18 cap_19

ITEM-TO-TOTAL FACTOR CORRELATION LOADING

VEA 64.67%

0.6368 0.5960 0.6388 0.6808

0.834 0.806 0.800 0.776

0.3539 0.4426 0.211

0.827 0.772 0.523

0.7779 0.7326 0.6986 0.5875 0.6220 0.5871

0.866 0.831 0.806 0.710 0.747 0.710

51.79%

60.91%

0.840 0.3074 0.4880 0.3291 0.5844 0.6228 0.5101 0.4101 0.3725 0.4894 0.5305 0.6140 0.4769 0.3588 0.5526 0.3564 0.3265 0.2542 0.1687 0.3373

C: Cultural Distance cult_3 cult_2 cult_1

0.679

Eff1: Benefit ben_h_di ben_h_ma

0.8368

60.90% 0.5152 0.4986 0.4655

0.688 0.651 0.591

0.7214 0.6967

0.859 0.844

67.22%

Appendix 1 151 ben_h_pu ben_h_te Eff2: Satisfaction sato_2 sato_4 sat_1 sati_3 sato_1

0.6357 0.6182

0.795 0.779

0.8040 0.7602 0.6920 0.6948 0.7323

0.885 0.856 0.802 0.804 0.833

0.8923

69.981%

Exhibit A1.2 CFA results: hierarchical relationships Modell Infrastructure (T2) Processes (T3) Satisfaction (Eff2)

Chi2

df

70.50 9 317.09 146 15.85 5

p-Value

RMSEA

Chi2/df

NFI

CFI

0.000 0.000 0.000

0.150 0.085 0.116

4.611 2.172 3.170

0.985 0.955 0.993

0.988 0.975 0.995

CFA is used to test the following constructs of the model which have been introduced as reflective variables:

• Knowledge Transfer Infrastructure (T2) • Knowledge Processes (T3) • Satisfaction (Eff2) As Exhibit A1.2 shows, results are good and the constructs can be included without further modification.

Reliability of the measurement model Having investigated the unidimensionality of constructs for a suitable measurement model, the reliabilty of the underlying variables has to be examined. Indicator reliabilty (IR) tells us how much of the variance is explained by the underlying factor. To assess the reliability of a factor, factor reliability (FR) and average variance extracted (AVE) are used. Indicator reliability is calculated on a normed [0;1] interval and is referred to as Squared Multiple Correlation in the AMOS module. The desirable cut-off values depend on the size of the sample. A sample size of n < 100 demands values of 0.6–0.9 while samples of n 100 < n < 400 require values of 0.4–0.6 (Andres 2003, p. 192). The average variance explained (AVE) indicates how much of the total variance is explained by the construct (Fornell and Larcker 1981). Values higher than 0.5 are required in this respect. As factor reliability and average variance explained are not calculated by the AMOS module, the formulas underlying the calculations are shown in Exhibit A1.3 and the results afterwards.

152 Exhibit A1.3

Indicator reliability, factor reliability and average variance explained

IR T2: Infrastructure infra_6 infra_8 infra_9 infra_10 infra_11 infra_12 T3: Processes Ext Int Soc Com Eff2: Satisfaction sato_2 sato_4 sat_1 sati_3 sato_1

0.415 0.441 0.652 0.759 0.555 0.375

0.278 0.183 0.618 0.571

0.785 0.704 0.504 0.532 0.618

Factor Score Weights

0.082 0.098 0.197 0.336 0.154 0.087

0.084 0.069 0.213 0.183

0.339 0.215 0.11 0.137 0.166

^2

S.E.

0.006724 0.009604 0.038809 0.112896 0.023716 0.007569

0.105 0.093 0.084

0.007056 0.004761 0.045369 0.033489

0.114921 0.046225 0.0121 0.018769 0.027556

FR

AVE

0.306

0.669

0.070

0.200

0.419

0.754

0.08 0.089

0.242 0.496 0.469

0.076 0.083 0.071 0.075

Appendix 2 Model Identification and Assessment

Model identification The refined measurements are now entered into the structural equation analyses. To estimate parameters, the maximum likelihood (ML) method is used. The latter is based on the following four assumptions (Byrne 2001, p. 70):

• • • •

The sample is very large (asymptotic) The distribution of the observed variables is multivariate normal The hypothesized model is valid The scale of the observed variables is continuous (recent discussions also admit Likert-type scales).

‘If a unique solution for the values of the structural equation parameters can be found, the model is considered to be identified. As a consequence, the parameters are considered to be estimable and the model therefore testable’ (Byrne 2001, p. 35). The identification of a model depends on whether a matrix of covariances defines the collectivity of the estimated parameter unambiguously or whether there exist other matrices of covariances which lead to the same parameter estimation. The identification of a model can be solved through a nonlinear equation model with q*(q + 1)/2 equations and t variables (q being the number of indicators and t the number of parameters to be estimated). The number of parameters to be estimated (t) must not be higher than the number of empirical variances and covariances (q*(q + 1)/2). The difference of these tells us the model’s degrees of freedom (Homburg and Pflesser 1999b, p. 645). The aim of SEM is to obtain an overidentified model – i.e. a model in which the number of estimable parameters is less than the number of data points (variances, covariances of the observed variables). In case of a just-identified model, there is a one-to-one correspondence between the data and the structural parameters, in case of an underidentified model, the number of estimable parameter exceeds the number of data points. Thus, only overidentified models show positive degrees of freedom and allow for a rejection of the model (Byrne 2001, p. 35). Models 1–3 show the descriptive information in Exhibit A2.1.

Assessment of the model After the specification of the model and the parameter estimation, the model has to be assessed. If this step leads to negative assessment, a modification of the model is necessary. In case of positive assessment, a detailed interpretation of

153

154 Appendix 2 Exhibit A2.1 Model descriptives MODEL 1 Number of variables in the model Number of observed variables Number of unobserved variables Number of exogenous variables Number of endogenous variables Sample size Number of distinct sample moments Number of distinct parameters to be estimated Degrees of freedom

MODEL 2 MODEL 3

56 25 31 28 28 162 350 81

49 22 27 25 24 162 275 70

32 14 18 16 16 162 119 44

269

205

75

results has to follow. According to Homburg and Baumgartner (1995), the following five steps need to be taken to assess a model: 1. Assessment of formal aspects: strange estimates and high standard errors point towards a misspecification of the model. 2. Assessment of the measurement model’s goodness of fit: local measures of fit. 3. Assessment of the whole model’s goodness of fit: global measures of fit. 4. Assessment of the structural equation model’s goodness of fit. 5. Cross-validation/comparison with alternative model structures: does the model explain the structures – i.e. the covariances of observed variables – of a second set of data: Step 1 refers to the fact that indicators for unidentified models are high standard errors and incomprehensible or strange estimates (Homburg and Pflesser 1999b, p. 645). Parameter estimates should exhibit the correct sign and size, and be consistent with the underlying theory. ‘Examples of parameters exhibiting unreasonable estimates are correlations > 1.00, negative variances, and covariance or correlation matrices that are not positive definite’ (Byrne 2001). Moreover, standard errors which are excessively large or small indicate a poor model fit. For the statistical significance of parameter estimates, the AMOS Module provides the critical ratio (c.r.), which represents the parameter estimate divided by its standard error. ‘Based on a level of 0.05, the test statistic needs to be > 1.96 before the hypothesis (that the estimate equals 0.0) can be rejected. Nonsignificant parameters, with the exception of error variances, can be considered unimportant to the model’ (Byrne 2001, p. 76). Steps 2 and 3 refer to the more specific model assessment which is based on several measures of fit. Generally, global and local measures of fit have to be distinguished. While global measures of fit assess the whole model, local measures of fit relate to specific parts of the model. Based on Homburg and Pflesser (1999b, p. 648), Exhibit A2.2 gives an overview. As shown in Exhibit A2.2, local measures of fit relate either to the measurement model or to the structural model. The reliability of indicators (IR) and factors (FR) was discussed earlier in the course of CFA and found adequate for both

Appendix 2 155 Exhibit A2.2 Overview of measures of fit

Measures of Fit Global Measures of Fit StandAlone Measures of Fit InferentialStatistic Measures of Fit

Local Measures of Fit

Incremental Measures of Fit

Descriptive Measures of Fit

accounting for degrees of freedom GFI

not accounting for degrees of freedom – AGFI – Chi-Square/df

Measures of the Structural Model

Measures of the Measurement Model

Squared multiple correlation

– IR – FR – AVE – t-value of factor loading

accounting for degrees of freedom

not accounting for degrees of freedom

NFI

CFI

Source: Based on Homburg and Pflesser (1999b).

measures. To assess particular equations of the structural model, squared multiple correlation is measured. On the construct level, the average amount of variance extracted is critical. All these values are normed on the [0;1] interval, where high values point towards a high quality of measurement (Homburg and Pflesser 1999b, p. 649). ‘Values less than 0.05 are indicative of good fit, between 0.05 and under 0.08 of reasonable fit, between 0.08 and 0.10 of mediocre fit and > 0.10 of poor fit’ (Diamantopoulos and Siguaw 2000, p. 85). Moreover, each factor loading’s significance can be tested in a t-test. Given a significance level of 0.01 for a one-tailed solution, t-values should meet a minimum value of 2.345 (Backhaus et al. 1996, p. 573). Global measures of fit are based on the comparison of the empirical matrix of covariances and the matrix of covariances produced by the model. Stand-alone and incremental measures of fit can be distinguished. The latter assess the relevant model on the basis of a null model. Values close to 1.0 point towards a good model fit. In these models, GFI (goodness of fit index) and AGFI (adjusted goodness of fit index) are not calculated because cases with ‘missing values’ are included in the analysis (Arbuckle and Wothke 1995, p. 331). Both relevant measures used in the analysis, normed fit index (NFI) and comparative fit index (CFI), compare the chi-square values of the model and the null model in different ways. Stand-alone measures can be seen from inferential statistics’ and from a descriptive point of view. Chi-square test statistics test the absolute ‘reproduction’ of reality through the model. ‘The more the implied and sample covariances differ, the bigger the chi-square statistic, and the stronger the evidence against the null hypothesis’ (Arbuckle and Wothke 1995, p. 97). This can be seen ambiguously, as a good approximation of reality is more important than an

156 Appendix 2 Exhibit A2.3 Suggested assessment of measures of fit Measures of Fit Measurement Model Indicator Reliability (IR) Factor Reliability (FR) Average Variance Extracted (AVE)

≥ 0.4 ≥ 0.6 ≥ 0.5

Whole Model RMSEA Chi-Square/df NFI CFI

≤ 0.05 (≥ 0.08; ≤ 0.10) ≥ 2.5 ≥ 0.9 ≥ 0.9

exact reproduction (Homburg and Pflesser 1999a, p. 427). Therefore, RMSEA (root mean squared error of approximation) is often seen as a better measure. For both measures, a low value is ideal. RMSEA values less than 0.05 indicate good fit, and values of high as 0.08 represent reasonable errors of approximation in the population. RMSEA values ranging from 0.08 to 0.10 indicate mediocre fit and those greater than 0.10 poor fit (Buckley and Casson 1976; Byrne 2001). Descriptive measures of fit, which also account for degrees of freedom are – given equal approximation – more meaningful as models with fewer parameters would be assessed worse (Homburg and Pflesser 1999a, p.428). Many authors stress that critical cut-off values of measures of fit are problematic in the sense that they should not be regarded as isolated (cf. Homburg and Pflesser 1999b; Diamantopoulos and Siguaw 2000). The sample size and the model’s complexity in particular influence the value of fit measures. A basic guidance is given by recommending the values in Exhibit A2.3 (Homburg and Pflesser 1999b, p. 651; Byrne 2001, p. 85).

Notes and References 2

Knowledge and the MNC

1 Sveiby calls it ‘competence’ in this context.

4

A Model of Knowledge Transfers in MNCs

1 Szulanski (1996) finds that the major barriers to internal knowledge transfer are not motivational factors but knowledge-related factors, such as the recipient’s lack of absorptive capacity. Based on these findings and the subsequent discussion of Foss and Pedersen (2002), motivational factors are excluded from this model of knowledge transfer in order to limit complexity. 2 A more detailed argument was presented in the conceptualization of the knowledge transfer process (p. 14). 3 Higher than all other strategic mandates. 4 Higher than Local Innovators but less than Integrated Players. 5 Higher than Local Innovators but less than Integrated Players. 6 Lower than all other strategic mandates.

5

Research Design and Methodology

1 A single-informant approach is a common weakness in international business research, which is shared with prominent studies such as Ghoshal and Nohria (1989) or Harzing (1999). 2 http://www.top500.de/g0030800.htm, 2003/06/27. 3 As only the strategic mandates of subsidiaries are involved, headquarters’ data would not be necessary for this construct. Nevertheless, it was decided to gather this data because the intensity and the direction of knowledge flows are vital for this study and will be analysed in the structural equation model.

6

Analyses and Results

1 The detailed characteristics of cultural and organizational distance are discussed later (p. 99). 2 Gupta and Govindarajan (2000) entered the variables into regression together with several other variables. Thus, the original method cannot be replicated. 3 The assumption of normal distribution is met by all variables. 4 The assumption of normal distribution is met by all variables. 5 The assumption of normal distribution is met by all variables. 6 The assumption of normal distribution is met by all variables. 7 The assumption of normal distribution is met by all variables. 8 The assumption of normal distribution is met by all variables.

157

158 Notes and References 9 The calculation of one whole model was possible only by aggregating all scales to the level of indicators. The summated-scales model’s measures of fit are similar to what is reported in the following chapters. However, factor reliability and average variance explained reach only low levels. 10 Higher than all other strategic mandates. 11 Higher than Local Innovators but less than Integrated Players. 12 Higher than Local Innovators but less than Integrated Players. 13 Lower than all other strategic mandates.

Bibliography

Adler, N. J. (1983), ‘A typology of management studies involving culture’, Journal of International Business Studies, 14(2), 29–47. Ambos, B. (2002), Internationales Forschungs- und Entwicklungsmanagement: Strategische Mandate, Koordination und Erfolg ausländischer Tochtergesellschaften, Wiesbaden: Gabler. Ambos, B. and W. D. Reitsperger (2002), ‘Governing knowledge processes in MNCs: the case of German R&D units abroad’, 28th EIBA Annual Conference, Athens, 8–10 December. Ambos, B. and W. Reitsperger (2004), ‘Offshore centers of excellence: social control and success’, Management International Review, 44(1). Ambos, B. and B. B. Schlegelmilch (2004), ‘The use of international R&D teams – an empirical investigation of selected contingency factors’, Journal of World Business, 39(2). Anderson, J. C. and D. W. Gerbing (1982), ‘Some methods for respecifying measurement models to obtain unidimensional constuct measurement’, Journal of Marketing Research, 19, 453–60. Andersson, U. and M. Forsgren (1996), ‘Subsidiary embeddedness and control in the multinational corporation’, International Business Review, 5(5), 319–34. Andersson, U. and M. Forsgren (2002), ‘The strategic impact of external networks: subsidiary performance and competence development in the multinational corporation’, Strategic Management Journal, 23(11), 979–96. Andres, N. (2003), Ausstrahlungseffekte beim Co-Branding – Eine verhaltenstheoretischempirische Analyse am Beispiel eines realen Produkttests kurzlebiger Konsumgüter, Hamburg: Dr. Kovac. Arbuckle, J. L. and W. Wothke (1995), Amos 4.0 User’s Guide, Chicago: SmallWaters. Argyris, C. (1994), ‘Good communication that blocks learning’, Harvard Business Review, July – August, 77–104. Asakawa, K. (1995), ‘Managing knowledge conversion processes across boarders: toward a framework of international knowledge management’, INSEAD Working Paper Series, 95/91/OB, Fontainbleau. Asakawa, K. (2001a), ‘Evolving headquarters–subsidiary dynamics in international R&D: the case of Japanese multinationals’, R&D Management, 31(1), 1–14. Asakawa, K. (2001b), ‘Organizational tension in international R&D management: the case of Japanese firms’, Research Policy, 30, 735–57. Asakawa, K. (2002), ‘Exploring the mode of cross-border knowledge mobilization by multinational corporations: an eclectic approach’, Strategic Management Society Annual Meetings, Paris. Asakawa, K. and M. Lehrer (2003), ‘Managing local knowledge assets globally: the role of regional innovation relays’, Journal of World Business, 38(1), 31–42. Au, K. Y. (2000), ‘Intra-cultural variation as another construct of international management: a study based on secondary data of 42 countries’, Journal of International Management, 6, 217–38.

159

160 Bibliography Backhaus, K., B. Erichson, W. Plinke and R. Weiber (1996), Multivariate Analysemethoden, Berlin: Springer. Barkema, H. G. and F. Vermeulen (1997), ‘What differences in the cultural backgrounds of partners are detrimental for international joint ventures?’, Journal of International Business Studies, 28(4), 845–64. Barney, J. B. (1986), ‘Types of competition and the theory of strategy’, Academy of Management Review, 11(4), 791–801. Bartlett, C. A. (1984), Organization and Control for Global Enterprises: Influences, Characteristics and Guidelines, Boston: Harvard Business School Press. Bartlett, C. A. and S. Ghoshal (1986), ‘Tap your subsidiaries for global reach’, Harvard Business Review, November–December, 87–94. Bartlett, C. A. and S. Ghoshal (1987), ‘Managing across boarders: new organizational responses’, Sloan Management Review, 29(1), 43–53. Bartlett, C. A. and S. Ghoshal (1988), ‘Organizing for worldwide effectiveness: the transnational solution’, California Management Review, 41(1), 54–74. Bartlett, C. A. and S. Ghoshal (1989), Managing Across Boarders: The Transnational Solution, Boston: Harvard Business School Press. Bartlett, C. A. and S. Ghoshal (2000), Transnational Management, 3rd edn. Boston: McGraw-Hill. Becerra-Fernandez, I. and R. Sabherwal (2001), ‘Organizational knowledge management: a contingency perspective’, Journal of Management Information Systems, 18(1), 23–55. Bennett, R. and H. Gabriel (1999), ‘Organizational factors and knowledge management within large marketing departments: an empirical study’, Journal of Knowledge Management, 3(3), 212–25. Bhagat, R. S., B. L. Keida, P. D. Harveston and H. C. Triandis (2002), ‘Cultural variations in the cross-border transfer of organizational knowledge: an integrated framework’, Academy of Management Review, 27(2), 204–21. Birkinshaw, J. (1999), ‘Acquiring intellect: managing the integration of knowledgeintensive acquisitions’, Business Horizons, May–June, 33–40. Birkinshaw, J. (2001), ‘Why is knowledge management so difficult?’, Business Strategy Review, 12(1), 11–18. Birkinshaw, J. (2002), ‘Managing internal R&D networks in global firms’, Long Range Planning, 35(3), 245–68. Birkinshaw, J., U. Holm, P. Thilenius and N. Arvidsson (2000), ‘Consequences of perception gaps in the headquarters–subsidiary relationship’, International Business Review, 9(3), 321–44. Birkinshaw, J. and N. Hood (1998), Multinational Corporate Evolution and Subsidiary Development, London: Macmillan. Birkinshaw, J. and A. J. Morrison (1995), ‘Configurations of strategy and structure in subsidiaries of multinational corporations’, Journal of International Business Studies, 26(4), 729–53. Black, S. J. and M. Mendenhall (1991), ‘The U-curve adjustment hypothesis revisited: a review and theoretical framework’, Journal of International Business Studies, 22(2), 225–47. Bloodgood, J. M. and D. Salisbury (2001), ‘Understanding the influence of organizational change strategies on information technology and knowledge management strategies’, Decision Support Systems, 31, 55–69.

Bibliography

161

Bollen, K. and R. Lennox (1991), ‘Conventional wisdom on measurement: a structural equation perspective’, Psychological Bulletin, 110(2), 305–14. Brockhoff, K. (1998), Internationalization of Research and Development, Berlin: Springer. Brouthers, K. D. and L. E. Brouthers (2001), ‘Explaining the national cultural distance paradox’, Journal of International Business Studies, 32(1), 177–89. Buckley, P. J. and M. J. Carter (1999), ‘Managing cross-border complementary knowledge’, International Studies of Management & Organization, 29(1), 80–104. Buckley, P. J. and M. C. Casson (1976), The Future of the Multinational Enterprise, London: Macmillan. Buckley, P. J. and M. C. Casson (1998), ‘Models of the Multinational Enterprise’, Journal of International Business Studies, 29(1), 21–44. Bullinger, H.-J., K. Wörner, and J. Prieto (1998), ‘Wissensmanagement – Modelle und Strategien für die Praxis’, in A. Zeller (ed.) Wissensmanagement-Schritte zum intelligenten Unternehmen, Berlin: Springer. Byrne, B. B. (2001), Structural Equation Modeling with AMOS, Mahwah, NJ: Lawrence Erlbaum Associates. Cantwell, J. and O. Janne (1999), ‘Technological globalization and innovative centers: the role of corporate technological leadership and location hierarchy’, Research Policy, 28, 119–20. Chandler, A. (1992), ‘What is a firm? A historical perspective’, European Economic Review, 36(2/3), 483–92. Chini, T. C., B. Ambos and B. B. Schlegelmilch (2003), ‘Benefiting from subsidiary knowledge: an empirical investigation of reverse knowledge transfers’, AIB Annual Meeting, Monterey, California, 5–8 July. Cohen, W. M. and D. A. Levinthal (1990), ‘Absorptive capacity: a new perspective on learning and innovation’, Administrative Science Quarterly, 35(1), 128–52. Conner, K. R. (1991), ‘A historical comparison of resource-based theory and five schools of thought within industrial organization economics: do we have a new theory of the firm?’, Journal of Management, 17, 121–54. Conner, K. R. and C. K. Prahalad (1996), ‘A resource-based theory of the firm: knowledge versus opportunism’, Journal of Management, 17(1), 121–54. Criscuolo, P. (2003), ‘Reverse technology transfer: a patent citation analysis of the European chemical and pharmaceutical sectors’, AIB Annual Meeting, Monterey, California, 5–8 July. Daal, B. V., M. Hass and M. Weggeman (1998), ‘The knowledge matrix: a participatory method for individual knowledge gap determination’, Knowledge and Process Management, 5(4), 255–63. Daft, R. L. and R. H. Lengel (1986), ‘Organizational information requirements, media richness and structural design’, Management Science, 32(5), 554–71. Darr, E. D., L. Argote and D. Epple (1995), ‘The acquisition, transfer, and depreciation of knowledge in service organizations: productivity in franchises’, Management Science, 41(11), 1750–62. Davenport, T. H. (1998), Working Knowledge, Boston: Harvard Business School Press. Davenport, T. H., D. W. De Long and M. C. Beers (1998), ‘Successful knowledge management projects’, Sloan Management Review, Winter, 43–57. Davenport, T. H., J. G. Harris, D. W. De Long and A. L. Jacobson (2001), ‘Data to knowledge to results: building analytic capability’, California Management Review, 43(2), 117–38.

162 Bibliography Davenport, T. H. and L. Prusak (1998), Working Knowledge, Cambridge, MA: Harvard Business School Press. D’Cruz, J. R. (1986), ‘Strategic management of subsidiaries’, in H. Etemad and L. S. Dulude (eds), Managing the multinational subsidiary, London: Croom Helm. De Long, D. W. (1997), ‘Building the knowledge-based organization: how culture drives knowledge behaviors’, Ernst & Young Center for Business Innovation, Working Paper, Boston. Demarest, M. (1997), ‘Understanding knowledge management’, Long Range Planning, 30(3), 374–84. Diamantopoulos, A. and J. A. Siguaw (2000), Introducing LISREL, London: Sage. Doz, Y. (1986), Strategic Management in Multinational Corporations, Oxford: Pergamon Press. Doz, Y., K. Asakawa, J. F. P. Santos and P. J. Williamson (1997), ‘The metanational corporation’, INSEAD Working Paper Series, 97/60/SM, Fontainebleau. Doz, Y. and J. F. P. Santos (1997), ‘On the management of knowledge: from the transparency of collocation and co-setting to the quandary of dispersion and differentiation’, INSEAD Working Paper Series, 97/119/SM, Fontainbleau. Doz, Y., J. F. P. Santos and P. J. Williamson (2001), From Global to Metanational, Boston: Harvard Business School Press. Dunning, J. H. (1993), The Globalization of Business: The Challenge of the 1990s, London: Routledge. Dyer, J. H. and K. Nobeoka (2000), ‘Creating and managing a high-performance knowledge-sharing network: the Toyota case’, Strategic Management Journal, 21, 245–367. Earl, M. (2001), ‘Knowledge management strategies: towards a taxonomy’, Journal of Management Information Systems, 18(1), 215–33. Easterby-Smith, M. and L. Araujo (1999), ‘Organizational learning: current debates and opportunities’, in John Burgoyne, (ed.), Organizational Learning and the Learning Organization, London: Sage. Egelhoff, W. G. (1991), ‘Information-processing theory and the multinational enterprise’, Journal of International Business Studies, 22(3), 341–68. Egelhoff, W. G. (1999), ‘Organizational equilibrium and organizational change: two different perspectives of the multinational enterprise’, Journal of International Management, 5, 15–33. Fornell, C. and D. F. Larcker (1981), ‘Structural equation models with unobservable variables and measurement error: algebra and statistics’, Journal of Marketing Research, 18, 382–88. Foss, N. J. and T. Pedersen (2002), ‘Transferring knowledge in MNCs: the role of sources of subsidiary knowledge in organizational context’, Journal of International Management, 8(1), 49–67. Frost, T. S. (2001), ‘The geographic sources of foreign subsidiaries’ innovations’, Strategic Management Journal, 22, 101–23. Galbraith, J. R. (1972), ‘Organizational design: an information processing view’, in P. R. Laurence (ed.), Organization, planning and cases, Homeword, IL. Irwin. Gallupe, B. (2001), ‘Knowledge management systems: surveying the landscape’, International Journal of Managment Reviews, 3(1), 61–77. Galunic, C. D. and S. Rodan (1998), ‘Resource recombinations in the firm: knowledge structures and the potential for Schumpeterian innovation’, Strategic Management Journal, 19, 1193–201.

Bibliography

163

Ghoshal, S. and C. A. Bartlett (1988), ‘Creation, adoption, and diffusion of innovations by subsidiaries of MNCs’, Journal of International Business Studies, 9(3), 365–88. Ghoshal, S., H. Korine, and G. Szulanski (1994), ‘Interunit communication in multinational corporations’, Management Science, 40(1), 96–110. Ghoshal, S. and N. Nohria (1989), ‘Internal differentiation within multinational corporations’, Strategic Management Journal, 10, 323–37. Gilbert, M. and M. Cordey-Hayes (1996), ‘Understanding the process of knowledge transfer to achieve successful technological innovation’, Technovation, 16(6), 301–12. Gold, A. H., A. Malhotra and A. H. Segars (2001), ‘Knowledge management: an organizational capabilities perspective’, Journal of Management Information Systems, 18(1), 185–214. Gomez-Mejia, L. R. and L. E. Palich (1997), ‘Cultural diversity and the performance of multinational firms’, Journal of International Business Studies, 28(2), 309–35. Granovetter, M. (1973), ‘The strength of weak ties’, American Journal of Sociology, 6, 1360–80. Granovetter, M. (1985), ‘Economic action and social structure: the problem of embeddedness’, American Journal of Sociology, 91(3), 481–510. Grant, R. M. (1996), ‘Towards a knowledge-based theory of the firm’, Strategic Management Journal, 17, 109–22. Grant, R. M. and C. Baden-Fuller (1995), ‘A knowledge-based theory of inter-firm collaboration’, Academy of Management Journal (Best Paper Proceedings), 17–22. Griffith, D. A. and M. G. Harvey (2001), ‘Executive insights: an intercultural communication model for use in global interorganizational networks’, Journal of International Marketing, 9(3), 87–103. Griffith, D. A., A. Y. Zeybek and M. O’Brien (2001), ‘Knowledge transfer as a means for relationships development: a Kazakhstan–foreign international joint venture illustration’, Journal of International Marketing, 9(2), 1–18. Grosse, R. (1996), ‘International technology transfer in services’, Journal of International Business Studies, 27(4), 781–800. Gupta, A. K. and V. Govindarajan (1991), ‘Knowledge flows and the structure of control within multinational corporations’, Academy of Management Review, 16(4), 768–92. Gupta, A. K. and V. Govindarajan (1994), ‘Organizing for knowledge flows within MNCs’, International Business Review, 3(4), 443–57. Gupta, A. K. and V. Govindarajan (2000), ‘Knowledge flows within multinational corporations’, Strategic Management Journal, 21, 473–96. Hakanson, L. (2003), ‘Creating knowledge. The power and logic of articulation’, Working Paper, Copenhagen Business School, Copenhagen. Hakanson, L. and R. Nobel (2000), ‘Technology characteristics and reverse knoweldge transfer’, Management International Review, 40 (Special Issue), 29–48. Hakanson, L. and R. Nobel (2001), ‘Organizational characteristics and reverse knowledge transfer’, Management International Review, 41(4), 395–420. Hansen, M. T. (1999), ‘The search–transfer problem’, Administrative Science Quarterly, 44, 82–111. Hansen, M. T., N. Nohria, and T. Tierney (1999), ‘What’s your strategy for managing knowledge?’, Harvard Business Review, March–April, 106–16.

164 Bibliography Harzing, A.-W. (1997), ‘Response rates in international mail surveys: results of a 22 country study’, International Business Review, 6(6), 641–65. Harzing, A.-W. (1999), Managing the Multinational, Northhampton VT Edward Elgar. Harzing, A.-W. (2000a), ‘Cross-national industrial mail surveys: why do response rates differ between countries?’, Industrial Marketing Management, 29(3), 243–354. Harzing, A.-W. (2000b), ‘An empirical analysis and extension of the Bartlett and Ghoshal typology of multinational companies’, Journal of International Business Studies, 31(1), 101–20. Hauschild, S., T. Licht and W. Stein (2001), ‘Creating a knowledge culture’, The McKinsey Quarterly, 1, 74–81. Hedlund, G. (1994), ‘A model of knowledge management and the N-form corporation’, Strategic Management Journal, 15, 73–90. Hedlund, G. (1999), ‘The intensity and extensity of knowledge and the multinational corporation as a nearly recomposable system (NRS)’, Management International Review, 39 (Special Issue), 5–44. Hennart, J.-F. and J. Larimo (1998), ‘The impact of culture on the strategy of multinational enterprise: does national origin affect ownership decisions?’, Journal of International Business Studies, 29(3), 515–38. Hofstede, G. (1980). Culture’s consequences, London: Sage. Hofstede, G. and M. H. Bond (1988), ‘The Confucius connection: from cultural roots to economic growth’, Organizational Dynamics, 16(4), 4–32. Holm, U., J. Johanson, and P. Thilenius (1995), ‘Headquarters’ knowledge of subsidiary network contexts in the multinational corporation’, International Studies of Management & Organization, 25(1–2), 97–119. Holm, U. and T. Pedersen (2000), The Emergence and Impact of MNC Centers of Excellence: A Subsidiary Perspective, London: Macmillan. Holsapple, C. W. and K. D. Joshi (2001), ‘Organizational knowledge resources’, Decision Support Systems, 31, 39–54. Holzner, B. and J. Marx (1979), The knowledge application system in society, Boston: Allyn and Bacon. Homburg, C. and H. Baumgartner (1995), ‘Beurteilung von Kausalmodellen’, Zeitschrift für Forschung und Praxis, 3, 162–76. Homburg, C. and C. Pflesser (1999a), ‘Konfirmatorische Faktoranalyse’, in C. Homburg (ed.), Marktforschung, Wiesbaden: Gabler. Homburg, C. and C. Pflesser (1999b), ‘Strukturgleichungsmodelle mit latenten Variablen: Kausalanalyse’, in C. Homburg (ed.), Marktforschung, Wiesbaden: Gabler. Hong, J. (1999), ‘Structuring for organizational learning’, The Learning Organization, 6(4), 173–85. Hoopes, D. G. and S. Postrel (1999), ‘Shared knowledge, “glitches”, and product development performance’, Strategic Management Journal, 20, 837–65. Huysman, M. (1999), ‘Balancing biases: a critical review of the literature on organizational learning’, John Burgoyne, (ed.), Organizational Learning and the Learning Organization, London: Sage. Hymer, S. H. (1976), The International Operations of National Firms: A Study of Direct Foreign Investment, Cambridge, MA: MIT Press. Inkpen, A. C. and A. Dinur (1998), ‘The transfer and management of knowledge in the multinational corporation: considering context’, Carnegie Bosch Institute Working Paper.

Bibliography

165

Jacob, M. and G. Ebrahimpur (2001), ‘Experience vs. expertise. The role of implicit understandings of knowledge in determining the nature of kowledge transfer in two companies’, Journal of Intellectual Capital, 2(1), 74–89. Jarillo, J. C. and J. I. Martinez (1990), ‘Different roles for subsidiaries: the case of multinational corporations in Spain’, Strategic Management Journal, 11(7), 501–12. Jemison, D. B. and S. B. Sitkins (1986), ‘Corporate acquisitions: a process perspective’, Academy of Management Review, 11(1), 145–63. Jensen, M. C. and W. H. Meckling (1995), ‘Specific and general knowledge, and organizational structure’, Journal of Applied Corporate Finance, 8(2), 4–18. Johnson, J. L., J. B. Cullen, T. Sakano and H. Takenouchi (1996), ‘Setting the stage for trust and strategic integration in Japanese–US cooperative alliances’, Journal of International Business Studies, 27(5), 981–1004. Johanson, J. L. and J.-E. Vahlne (1977), ‘The internationalization process of the firm – a model of knowledge development and increasing foreign market commitments’, Journal of International Business Studies, 8(1), 23–33. Katsikeas, C. S., L. C. Leonidou and N. A. Morgan (2000), ‘Firm-level export performance assessment: review, evaluation, and development’, Academy of Marketing Science Journal, 28(4), 493–511. Kogut, B. and H. Singh (1988), ‘The effect of national culture on the choice of entry mode’, Journal of International Business Studies, Fall, 411–32. Kogut, B. and U. Zander (1992), ‘Knowledge of the firm, combinative capabilities, and the replication of technology’, Organization Science, 3, 383–97. Kogut, B. and U. Zander (1993), ‘Knowledge of the firm and the evolutionary theory of the multinational corporation’, Journal of International Business Studies, 24(4), 625–45. Kogut, B. and U. Zander (1995), ‘Knowledge, market failiure and multinational enterprise: a reply’, Journal of International Business Studies, 26(2), 417–24. Kriwet, C. K. (1997), Inter- and Intraorganizational Knowledge Transfer, Universität St. Gallen: Bamberg. Krogh, G. von and M. Köhne (1998), ‘Der Wissenstransfer in Unternehmen: Phasen des Wissenstransfers und wichtige Einflussfaktoren’, Die Unternehmung, 52(5/6), 235–52. Krogh, G. von and M. Venzin (1995), ‘Anhaltende Wettbewerbsvorteile durch Wissensmanagement’, Die Unternehmung 52(6), 417–36. Kuemmerle, W. (1997), ‘Building effective R&D capabilities abroad’, Harvard Business Review (March–April), 61–9. Lähteenmäki, S., J. Toivonen and M. Mattila (2001), ‘Critical aspects of organisational learning research and proposals for its measurment’, British Journal of Management, 12, 113–29. Leonard-Barton, D. (1995), Wellsprings of Knowledge, Boston: Harvard Business School Press. Lieberman, M. B. and D. B. Montgomery (1998), ‘First-mover (dis)advantages: retrospective and link with the resource-based view’, Strategic Management Journal, 19, 1111–25. Liebowitz, J. (1999), ‘Key ingredients to the success of an organization’s knowledge management strategy’, Knowledge and Process Management, 6(1), 37–40. Lyles, M. A. and J. E. Salk (1996), ‘Knowledge acquisition from foreign partners in international joint ventures’, Journal of International Business Studies, 27(5), 877–903.

166 Bibliography Macharzina, K., M.-J. Oesterle, and D. Brodel (2001), ‘Learning in multinationals’, in J. Child (ed.), Handbook of Organizational Learning and Knowledge, Oxford: Oxford University Press. Malhotra, Y. (1998), ‘Tools@work: deciphering the knowledge managment hype’, Journal for Quality and Participation, 21(4), 58–60. Malnight, T. W. (1996), ‘The transition from decentralized to network-based MNC structures: an evolutionary perspective’, Journal of International Business Studies, 27(1), 43–65. Malnight, T. W. (2001), ‘Emerging structural patterns within multinational corporations: toward process-based structures’, Academy of Management Journal, 44(6), 1187–1208. Manev, I. M. and W. B. Stevenson (2001), ‘Nationality, cultural distance, and expatriate status: effects on the managerial network in a multinational enterprise’, Journal of International Business Studies, 32(2), 285–303. Marquardt, M. (1996), Building the Learning Organization, New York: McGraw-Hill. Martinez, J. I. and J. C. Jarillo (1989), ‘The evolution of research on coordination mechanisms in multinational corporations’, Journal of International Business Studies, 20(3), 489–514. Martinez, J. I. and J. C. Jarillo (1991), ‘Coordination demands of international strategies’, Journal of International Business Studies, 22(3), 429–44. McAdam, R. M. and S. McCreedy (1999), ‘A critical review of knowledge management models’, The Learning Organization, 6(3), 91–100. McDermott, R. and C. O’Dell (2001), ‘Overcoming cultural barriers to sharing knowledge’, Journal of Knowledge Management, 5(1), 76–85. Morosini, P. (1998), Managing Cultural Differences, Oxford: Pergamon. Mowery, D. C., J. E. Oxley and B. S. Silverman (1996), ‘Strategic alliances and interfirm knowledge transfer’, Strategic Management Journal, 17 (Winter Special Issue), 77–91. Mudambi, R. (2002), ‘Knowledge management within multinational firms’, Journal of International Management, 8(1), 1. Nahapiet, J. and S. Ghoshal (1998), ‘Social capital, intellectual capital, and the organizational advantage’, Academy of Management Review, 23(2), 242–66. Nevis, E. C., A. J. DiBella and J. M. Gould (1995), ‘Understanding organizations as learning systems’, Sloan Management Review, 36(2), 73–87. Newman, K. L. and S. d. Nollen (1996), ‘Culture and congruence: the fit between management practices and national culture’, Journal of International Business Studies, 27(4), 753–80. Nobel, R. and J. Birkinshaw (1998), ‘Innovation in multinational corporations: control and communication pattern in international R&D operations’, Strategic Management Journal, 19, 479–96. Nohria, N. and S. Ghoshal (1994), ‘Differentiated fit and shared values: alternatives for managing headquarters–subsidiary relationships’, Strategic Management Journal, 15, 491–502. Nohria, N. and S. Ghoshal (1997), The Differentiated Network: Organizing Multinational Corporations for Value Creation. San Francisco: Jossey Bass. Nonaka, I. and H. Takeuchi (1995), The Knowledge Creating Company: How Japanese Companies Create the Dynamics of Innovation, New York: Oxford University Press.

Bibliography

167

Nonaka, I., K. Umemoto and D. Senoo (1996), ‘From information processing to knowledge creation: a paradigm shift in business management’, Technology in Society, 18(2), 203–18. O’Dell, C. and J. C. Grayson (1998), ‘If only we knew what we know: identification and transfer of internal best practices’, California Management Review, 40(3), 154–74. Paterson, S. L. and D. M. Brock (2002), ‘The development of subsidiary–management research: review and theoretical analysis’, International Business Review, 11(2), 139–63. Pentland, B. T. (1995), ‘Information systems and organizational learning: the social epistemology of organizational knowledge systems’, Accounting Management and Information Technologies, 5(1), 1–21. Perlmutter, H. V. and D. A. Heenan (1979), Multinational Organization Development Reading, MA: Addison-Wesley. Polanyi, M. (1966), The Tacit Dimension, London: Routledge. Porter, M. E. (1980), Competitive Strategy, New York: Free Press. Porter, M. E. (1990), The Competitive Advantage of Nations, New York: Free Press. Prahalad, C. K. and Y. L. Doz (1987), The Multinational Mission: Balancing Local Demands and Global Vision, New York: Free Press. Prahalad, C. K. and G. Hamel (1990), ‘The core competence of the corporation’, Harvard Business Review, 68(3), 95–117. Probst, G., S. Raub and K. Romhardt (1999), Wissen managen–wie Unternehmen ihre wertvollste Resource optimal nutzen, Wiesbaden: Gabler. Reger, G. (1997), Koordination und strategisches Management internationaler Innovationsprozesse, Heidelberg: Physica. Riesenberger, J. R. (1998), ‘Knowledge – the source of sustainable competitive advantage’, Journal of International Marketing, 6(3), 94–107. Romhardt, K. (2000), ‘Interventionen in die organisatorische Wissensbasis zwischen Theorien und Praxis Wechen Beitrag kann ein Systemtheorie leisten?’, Université de Genève (HEC), Geneva. Rosenzweig, P. M. and J. V. Singh (1991), ‘Organizational environments and the multinational enterprise’, Academy of Management Review, 16(2), 340–61. Roth, K. and A. J. Morrison (1992). ‘Implementing global strategy: characteristics of global subsidiary mandates’, Journal of International Business Studies, 23(4), 175–96. Roth, K. and S. O’Donnell (1996), ‘Foreign subsidiary compensation strategy: an agency theory perspective’, Academy of Management Journal, 39(3), 678–703. Rugman, A. M. and C. Brain (2003), ‘Multinational enterprises are regional not global’, Multinational Business Review, 11(1), 3–13. Salzberger, T., R. Sinkovics and B. B. Schlegelmilch (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. Shannon, C. E. and W. Weaver (1957), Mathematical Theory of Communication, Chicago: University of Illinois Press. Shenkar, O. (2001), ‘Cultural distance revisited: towards a more rigorous conceptualization and measurement of cultural differences’, Journal of International Business Studies, 32(2), 519–35. Shin, M., T. Holden and R. A. Schmidt (2001), ‘From knowledge theory to management practice: towards an integrated approach’, Information Processing & Management, 37(2), 335–55.

168 Bibliography 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 Studies, 30(3), 463–90. Skryme, D. J. and D. M. Amidon (1997), Creating the Knowledge-Based Business. Business intelligence/Ernst & Young survey of the state of practice, Business Intelligence Ltd. Slater, S. F. and J. C. Narver (1995), ‘Market orientation and the learning organization’, Journal of Marketing, 59, 63–74. Sölvell, O. and I. Zander (1995), ‘Organization of the dynamic multinational enterprise: the home-based and the heterarchical MNE’, International Studies of Management & Organization, 25(1), 17–29. Sölvell, O., I. Zander and M. E. Porter (1991), Advantage Sweden, Stockholm: Norstedts. Stewart, K. A., R. Baskerville, V. C. Storey, A. Raven and C. Long (2000), ‘Confronting the assumptions underlying the management of knowledge and investigating knowledge management’, The Data Base for Advances in Informations Systems, 31(4), 41–53. Subramaniam, M. and N. Venkatraman (2001), ‘Determinants of transnational new product development capability: testing the influence of transferring and deploying tacit overseas knowledge’, Strategic Management Journal, 22, 359–78. Sveiby, K.-E. (2001), ‘A knowledge-based theory of the firm to guide in strategy formulation’, Journal of Intellectual Capital, 2(4), 344–58. Szulanski, G. (1995), ‘Unpacking stickiness: an empirical investigation of the barriers to transfer best practice inside the firm’, Academy of Management Journal, (Special Volume, Best Paper Proceeding), 437–41. Szulanski, G. (1996), ‘Exploring internal stickiness: impediments to the transfer of best practice within the firm’, Strategic Management Journal, 17(Special Issue), 27–43. Taggart, J. and N. Hood (1999), ‘Determinants of autonomy in multinational corporation subsidiaries’, European Management Journal, 17(2), 226–37. Teece, D. J. (1981), ‘The market for know-how and the efficient international transfer of technology’, Annals, AAPSS, 458, 81–96. Teece, D. J. (1998), ‘Capturing value from knowledge assets: the new economy, markets for know-how and intangible assets’, California Management Review, 40(3), 55–79. Teigland, R., C. F. Fey and J. Birkinshaw (2000), ‘Knowledge dissemination in global R&D operations: an empirical study of multinationals in the high technology electronics industry’, Management International Review, 40(1), 49–77. Tenkasi, R. V. (2000), ‘The dynamics of cultural knowledge and learning in creating viable theories of global change and action’, Organization Development Journal, 18(2), 74–90. Treacy, M. and F. Wiersema (1995), Discipline of Market Leaders: Choose your Customers, Narrow your Focus, Dominate your Market, Reading: Addison-Wesley. Tsai, W. (2001), ‘Knowledge transfer in intraorganizational networks: effects of network position and absorptive capacity on business unit innovation and performance’, Academy of Management Journal, 44(5), 996–1004.

Bibliography

169

Tsoukas, H. and E. Vladimirou (1996), ‘The firm as a distributed knowledge system: a constructionist approach’, Strategic Management Journal, 17 (Winter Special Issue), 11–25. Vernon, R. (1966), ‘International investment and international trade in the product cycle’, Quarterly Journal of Economics, 80, 190–207. Von Hippel, E. (1994), ‘“Sticky information” and the locus of problem solving: implications for innovation’, Management Science, 40(4), 429–39. Wagner, P. (2000), ‘Die unsichtbare Ressource: Wissen, Vorausssetzungen fürs Gelingen’, Unternehmensentwicklung, March–April 2000, 6. Wernerfelt, B. (1984), ‘A resource-based view of the firm’, Strategic Management Journal, 5(2), 171–80. White, R. E. and T. A. Poynter (1990), ‘Organizing for world-wide advantage’, in Gunnar Hedlund (ed.), Managing the Global Firm, London: Routledge. Willke, H. (1998), Systemisches Wissensmanagement, Stuttgart: Lucius & Lucius. Wolf, J. (1997), ‘From “starworks” to networks and heterarchies? Theoretical rational and empirical evidence of HRM organization in large mulitnational corporations’, Management International Review, 37(1), 145–69. Zack, M. H. (1999), ‘Managing codified knowledge’, Sloan Management Review, Summer, 45–58. Zahra, S. A. and G. George (2002), ‘Absorptive capacity: a review, reconceptualization, and extension’, Academy of Management Review, 27(2), 185–203. Zander, U. and B. Kogut (1995), ‘Knowledge and the speed of the transfer and imitation of organizational capabilities: an empirical test’, Organization Science, 6(1), 76–92. Zhou, C. and T. Frost (2003), ‘Centrifugal forces, R&D co-practice, and “reverse knowledge flows” in multinational firms’, AIB Annual Meeting, Monterey, California, 5–8 July.

Index

Notes: ch=chapter; e =exhibit; n=note; bold=extended discussion or heading/word emphasized in main text.

absorptive capacity (Cohen and Levinthal) 33e, 51, 60–1, 77, 157(n1 to ch4) Active Subsidiary 43 adjusted goodness of fit index (AGFI) 155e, 155 Adler, N. J. 124 agriculture, forestry, fishing 73e Ambos, B. 33e, 35e, 35, 40, 144 AMOS software module 115, 151, 154 Analysis of Variance (ANOVA) 98, 119, 157(n7) Andersson, U. 38 Argote, L. 30e Arvidsson, N. 160 Asakawa, K. 21, 25, 65, 69, 140 Asia 74e, 75e, 102 Au, K. Y. 102 Australia 74e, 75e Austria 72, 73–5 autonomy 38, 46, 133, 142 average variance extracted (AVE) 151, 152e, 155e, 156e Barney, J. B. 25 Bartlett, C. A. 2, 20, 21, 22, 25, 40, 41e, 42, 43, 46 Baskerville, R. 161 Baumgartner, H. 154 Becerra-Fernandez, I. 8, 32e, 35e, 35, 54, 66, 78, 96, 108 Beers, M. C. 13e behaviour 44, 66, 123 Belgium 72, 75e best practice 27, 30e, 31e, 96e Bhagat, R. S., et al. (2002) 33e, 34, 55, 160 Harveston, P. D. 160 Keida, B. L. 160 Triandis, H. C. 160

Birkinshaw, J. 32e, 38, 40, 41e, 41, 43, 46, 86, 145 categorization of knowledge management systems 12 definition of knowledge management 11e Birkinshaw, J., et al. 69, 160 Arvidsson, N. 160 Holm, U. 160 Thilenius, P. 160 Black, S. J. 102 ‘Black Hole’ 43 Bloodgood, J. M. 12 Brain, C. 20 Buckley, P. J. 11e, 21, 50, 65 Bulgaria 72 business intelligence 51, 62 business practice 78, 99, 99e, 100, 100e Byrne, B. B. 111, 156 capabilities 22, 23, 25, 32e, 35e, 37, 40, 47, 53, 54 combinative 49, 50, 51 human 2 knowledge management 50–2, 139 learning 50 organizational 3, 49–50, 67 see also knowledge transfer capabilities capture and transfer of experts’ knowledge 95, 96e, 97e Carter, M. J. 11e, 65 Casson, M. C. 21, 50 centralization 37, 139 Centres of Excellence (CoEs) 47 CFA (confirmatory factor analysis) 149, 151e, 154 Chandler, A. 50 chat groups 96e, 96, 97e Chi squares 114–15, 115e, 155e, 156e China 73 Chini, T. C. 33e, 35e, 35

170

Index 171 cognition 66 ‘cognitive knowledge’ 8 Cohen, W. M. 51, 61 combination (knowledge conversion process) 18e, 19, 62, 63, 64e, 78, 97–8 common method variance 69, 119 communication 86 in the company 95 informal channels 102 lateral 24 communication model (Shannon and Weaver, 1957) 15–18 ‘noise’ 15e, 52 companies/firms domestic (contrasted with MNCs) 20 evolutionary theory 50 comparative fit index (CFI) 114, 115e, 155e, 155, 156e competence 17, 157(n1 to ch2) competitive advantage 2, 9, 10, 14, 24, 25, 29, 46, 54, 64, 145 competitors 76, 95 see also market data on competitors Conner, K. R. 25 construction 73e constructionism 12 consulting companies 72 context 15, 19, 27, 31e, 34, 35e, 37, 38, 68–70, 144–5, 147 structural 41 context factors 99–103 cultural distance 101–3 organizational distance 99–101 context similarities 147 contingency factors 3, 52–7, 145 ‘eventful’ knowledge management 53–4 institutional isomorphism 53, 54, 65 national culture 53, 54–7 control 3, 35e, 35, 37, 43–5, 46–7, 86, 122, 124, 140 bureaucratic formalized 44, 45e, 46 formal and informal 44, 49 means to achieve an end called ‘coordination’ 44 mechanisms (four categories) 44, 45e personalized centralized 44, 45e by socialization and networks 45e, 45, 46 two aspects 44

control instruments/mechanisms 39, 49, 133, 139, 141–2 determinants 47 formal 92, 148 coordination 3, 21, 35e, 37, 39, 43–5, 46–7, 51, 64, 92, 122, 124, 140 formal 49 informal 49, 77–8 ‘hypothesis 1c’ 63, 116 lacking 121 lateral and hierarchical 77 Cordey-Hayes, M. 30e covariance 108, 149, 153, 154 creativity 55, 56 Criscuolo, P. 146 critical ratios (c.r.) 115, 154 Croatia 72 Cronbach alpha 149–51e cross-border exchanges 20, 23 cultural distance 3, 45, 55–6, 57, 59e, 66, 67, 68, 79, 86, 93e, 94, 100, 100e, 101–3, 108, 110e, 110, 111e, 114e, 124–9, 133, 138, 139, 144, 145, 148, 150e, 157(n1 to ch6) hidden assumptions hypothesis 66, 102, 118e, 123–4, 127e, 132e, 137e impact on knowledge transfer 101e, 102 micro and macro level 56 resource-based view 56 second dataset 127e subjective perceptions 102 third dataset 130, 131e transaction cost perspective 56 cultural distance index (Kogut and Singh) 101 cultural mandates 45 culture 2, 31e, 33e, 34, 139, 142, 147 company-centric 47 corporate 78, 99, 100, 124, 133 local 53 national/country 53, 54–7, 101e, 102, 124, 148 one-company 145, 148 operational 100 organizational 71 customers 13, 62, 76 see also market data on customers Czech Republic 72 D’Cruz, J. R. 41e Daal, B. V. 13e

172 Index Darr, E. D. 30e data 5, 7e, 7, 11e, 103 codification 6 definition 6 different areas 76 primary 146 secondary 146 data collection 3, 68, 70–6 controlling for industry-specific results 71 final sample 72–6, 73e informant selection 71–2 target sample 68, 70–1, 73e two levels (headquarters and subsidiaries) 71 databases 7, 95, 96e, 96, 97e, 139, 143 Davenport, T. H. 13e Davenport, T. H., et al. 11e, 161 De Long, D. W. 161 Harris, J. G. 161 Jacobson, A. L. 161 De Long, D. W. 13e, 161 decision support systems 95, 96e, 96, 97e, 143 decision-making 21, 22, 48, 65 formalization 37 Demarest, M. 11e, 13e df (degrees of freedom) scores 114–15, 115e, 155e, 156e Diamantopoulos, A. 109n, 112n Dinur, A. 15, 31e, 35e, 62 distribution know-how 88e, 89e, 103, 103e, 104e, 105, 105e documents 7, 19 Doz, Y. L. 22, 25, 29, 31e, 34, 53, 55, 62, 66 Dyer, J. H. 31e, 146 Earl, M. 12 Ebrahimpur, G. 28, 71 economies of scale 2, 23, 25, 42 embeddedness 32e external 38 twofold (subsidiaries) 38 empiricism 1, 21, 34, 35, 35e, 50, 51, 61, 66, 68, 70, 78, 81, 120, 121, 128, 130, 138, 142, 144, 145, 148, 149, 153, 155 employee rotation 95, 96e, 97e employees 17, 28, 78, 80, 81–2, 82e, 85, 122, 123, 124, 143 environment (business) 40, 41, 46

epistemic communities 26 epistemology 12, 18, 27, 32e Epple, D. 30e Europe Central and Eastern (CEE) 73, 74e, 75e, 75–6 Central/Western 72, 73, 74e, 75e, 75–6, 81 Northern 72, 73, 74e, 75e Southern 72, 74e, 75e executive development programmes 78, 93e, 141 externalization (knowledge conversion process) 18–19, 18e, 62, 64e, 78, 97, 97e, 98e, 98 ‘hypothesis 1b’ 63, 116 face-to-face meetings 19, 95, 96e, 96, 97e, 139, 143 factor reliability (FR) 151, 152e, 154, 155e, 156e Fey, C. F. 32e, 145 field of interaction 63 finance and insurance 72, 73e, 81 foreign direct investment (FDI) 20 Forsgren, M. 38 Foss, N. J. 28, 33e, 35e, 35, 61, 157(n1 to ch4) France 72 Frost, T. 144 Galbraith, J. R. 23 Gallupe, B. 10, 12 George, G. 51 German-speaking countries 71 Germany 72, 75e Ghoshal, S. 2, 20, 22, 23, 25, 31e, 34, 37, 40, 41e, 42, 43, 46, 124, 142, 157(n1 to ch5) Gilbert, M. 30e global exchange 49 Global Innovators 41e, 42, 48, 59e, 60, 64e, 77, 84, 84e, 85e ‘hypothesis 1b’ 63, 116, 157(n4 to ch4), 158(n11) globalization 21 Gold, A. H. 32e, 35e, 35, 50, 66, 78, 108, 142 goodness of fit index (GFI) 155e, 155 Govindarajan, V. 8, 21, 27, 28, 29, 30e, 32e, 34, 35e, 35, 40, 41e, 42, 44, 47, 61, 76–7, 84, 86, 88, 90, 91, 140–1, 142, 143, 157(n2 to ch6)

Index 173 Granovetter, M. 53 Grant, R. M. 2, 25, 28 Grayson, J. C. 31e Griffith, D. A. 57, 147 Grosse, R. 30e groupware 70, 96, 143 Gupta, A. K. 8, 21, 27, 28, 29, 30e, 32e, 34, 35e, 35, 40, 41e, 42, 44, 47, 61, 76–7, 84, 86, 88, 90, 91, 140–1, 142, 143, 157(n2 to ch6) Hakanson, L. 9, 28, 32e, 35e, 35 Hamel, G. 25 Hansen, M. T. 12, 31e, 34, 35e, 47, 51 Harris, J. G. 161 Harveston, P. D. 160 Harzing, A.-W. 44, 45, 69, 72, 157(n1 to ch5) Hass, M. 13e headquarters 20, 22, 29, 33e, 39–44, 46, 58, 60, 68, 69, 72, 79, 83–4, 85–6, 88, 89, 92, 95, 97, 106, 107, 113, 114e, 116, 138, 139–40, 141, 142 benefit from knowledge transfers 103 benefits of knowledge transfers from culturally close subsidiaries 105, 105e benefits of knowledge transfers from culturally distant subsidiaries 105, 105e functions 21 global 87e, 87, 140 hierarchical knowledge flows 27 informal transmission channels 93e, 94 knowledge flows 76e, 157(n3 to ch5) location 73–6 managers 70 perceptions of similarity vis-à-vis culturally close and distant subsidiaries 100e personnel-intensive 81–2 regional 87e, 87, 140 strategic orientation scales 100–1 strategic positions 87e strategic tasks 90 value of knowledge stock 90–1 headquarters–subsidiary relationships 37–9, 56, 61–2, 63, 65, 78, 102 central coordination 37 culturally close subsidiaries 107–24, 134–7e

culturally distant subsidiaries 124–9, 134–7e data 107 dyadic 69, 71 ‘hypotheses 1a–1d’ 63 local responsiveness 37 principal–agent relationship 37 results 125–9, 134–7e second dataset 124–9 see also subsidiary–subsidiary relationships Hedlund, G. 9, 10, 13, 22, 23, 24 Heenan, D. A. 2, 21, 25 ‘hermeneutic circle’ 27 Hierarchical Relationships and Culturally Distant Subsidiaries (dataset) 149 Hofstede, G. 79, 102 Hofstede-based measures 147 Holden, T. 8, 13, 13e Holm, U. 160 Holsapple, C. W. 12 Holzner, B. 13e Homburg, C. 111, 154, 156 homophily 56 Hong, J. 12 Hood, N. 38 Hoopes, D. G. 31e Horizontal Organization (White and Poynter) 22 host country (economic level) 77, 90–1 Hungary 72 imitation 30e Implementers see Local Implementers indicator reliability (IR) 151, 152e, 154, 155e, 156e individuals 16e, 17, 18 industries 71, 72, 73e information 5, 7e, 7, 11e, 46 definition 6 information distribution 12 information exchange 45 information interpretation 13 information processing 21 information technology (IT) 6, 11e information-processing capacity 48 infrastructure 51, 54, 59e, 113e reflective variable 111e Inkpen, A. C. 15, 31e, 35e, 62 innovation 22, 57, 140, 147

174 Index institutional isomorphism/ organizational isomorphism 53, 54, 65, 145 Integrated Network Model (Bartlett and Ghoshal) 22, 23, 24 Integrated Players 41e, 43, 48, 59e, 60, 64e, 77, 84–6, 141 ‘hypothesis 1a’ 63, 116, 157(n3–5 to ch4), 158(n10–12) integrating mechanisms (Hedlund) 24 integration 21, 22, 34, 37, 146 intellectual capital 31e interactions across distance 53 internalization (knowledge conversion process) 18e, 19, 50, 62, 64e, 78, 97–8 ‘hypothesis 1c’ 63, 116 internet 95, 96e, 97e interviews 68 intranet 95, 96e, 97e Ireland 72 Italy 72 Jacob, M. 28, 71 Jacobson, A. L. 161 Japan 73 Jarillo, J. C. 40, 41e, 44, 46 Jemison, D. B. 56 Johanson, J. L. 55 joint ventures ( JVs) 27, 29 Joshi, K. D. 12 K-means cluster analysis 85 Katsikeas, C. S. 65 Keida, B. L. 160 key informant approach 69, 82, 157(n1 to ch5) know-how 103, 106 knowing how 8 knowing what 8 knowledge characteristics ‘context-bound’ 27 ‘application-related process’ 8 ‘complexity’ 28 ‘has to be adapted to recipient culture’s specifications’ 55 ‘not culture-free’ 55 classification/types 7–8, 30–3e cognitive 8 declarative 27, 30e, 32e, 33e, 47

developmental 30e explicit 5, 6, 8–10, 18, 18e, 19, 24, 26, 28, 62, 98 explicit (separate) 62 explicit (systemic) 62, 63, 98 highly complex 34 idiosyncratic 28–9 individual 5, 10, 31e, 32e, 62 instrumental 30e internal or external 13 managerial and administrative 29 new 51, 58, 62, 94e, 95, 144 organizational 5, 10, 31e, 32e, 33e, 62 procedural 27, 30e, 31e, 32e, 33e, 47, 76 social 10 tacit 5, 6, 8–10, 18, 18e, 24, 26, 28, 32e, 51, 62, 71, 98, 144 technical 29, 32e general acquisition 12, 26, 47, 144 aggregation 28, 51 application 13e, 14 asymmetries 25 availability 106 codification 14 conceptualization 5–10 data–information–knowledge continuum 7e definitions 5, 6–7 diffusion/dissemination 13e, 14, 27, 50, 51 dispersed assets 25 distribution 13e, 14 identification and mobilization 140 identification of source 13 neoclassical view 29 ‘object’ 8, 12 process 12 ‘resource’ 10 sources (subsidiary companies) 35 ‘stickiness’ 29 storage 13–14, 13e, 17, 26 subjective measures 69 transferability 27–9, 35e, 37 transformation 13, 14 type transferred 28 use 13e, 14 knowledge conversion processes 18–19, 78, 96–8

Index 175 knowledge creation 13, 13e, 18, 39, 42, 43, 50, 55 individual 26 role of subsidiaries 38, 39 knowledge culture 145 knowledge discovery 51, 62 knowledge extensity 24 knowledge flows 44, 77, 87, 110, 140 as control or administrative mechanisms 35 hierarchical 83, 86, 116 intensity and direction 76 internalization 20 lateral 83–4, 86 optimization 47–9 overall contribution to firm’s operations 147 perceived benefit 130 knowledge inflows 32e, 40, 42, 43, 48, 58, 59e, 60, 76e, 77, 84, 84e, 86, 87e, 87, 98, 99, 103, 105, 110, 111e, 113e, 125, 128, 139, 140, 141, 143 benefit from 148 first dataset 130 ‘hypotheses’ 63, 116, 117e, 119, 120e, 126e, 131e, 134e perceived benefits 79 second dataset 126e third dataset 130, 131e knowledge integration 32e, 51, 147 knowledge intensity 24 knowledge management 3, 57, 61, 101, 103, 129, 138, 145, 148 capabilities 50–2, 139 challenge 1–2 conceptual background 5–19 definitions 10–12 effectiveness 54 ‘eventful’ 53–4 human resources-oriented approach 10 hypothesis 117e, 118e, 127e, 131e, 132e, 135e, 136e, 137e ‘integrative’ approach 10 ‘organization-wide activity’ 12 organizational levels 80 overall satisfaction 67 perceived benefit 66–7, 117e relevance 19–36 satisfaction with 106–7, 107e, 117e, 118e, 122, 125, 127e, 130, 131–2e, 133, 135e, 136e, 137e, 139, 141–2, 143–4, 157(n8)

second dataset 127e structure of MNC 23–4 technology-oriented approach 10 theoretical concepts 10–19 third dataset 131e value chain 12–14 vision and strategy 13e, 14 knowledge management infrastructure 78, 94–5, 117e–18e, 119, 120e, 122, 128, 130, 133, 143–4, 148, 157(n4 to ch6) ‘hypotheses’ 63, 64e, 116, 117e, 118e, 119, 126e, 127e, 131e, 132e, 134e, 135e, 136e impact on satisfaction 123 means and medians 94 perceived benefit of knowledge transfers 122–3 second dataset 126e, 127e third dataset 131e, 132e variables 120e knowledge management processes 123, 128, 133 effect on satisfaction 123 encoding and decoding 123 hypothesis accepted 118e, 127e, 132e, 136e perceived benefit 123 second dataset 127e third dataset 132e knowledge management systems 12 knowledge management tools 70, 78, 95, 139, 143–4 ‘conservative’ 139, 143 use 96e knowledge mapping 51, 62 knowledge outflows 32e, 40, 42, 43, 48, 58, 59e, 60, 76e, 77, 84, 84e, 86, 87e, 87, 99, 110, 111e, 113e, 119, 120e, 120, 130, 139–43 hypothesis 63, 116, 117e, 126e, 134e second dataset 126e third dataset 131e ‘knowledge is power’ 28 knowledge process capabilities 59e, 62, 78, 113e, 145, 148 hypothesis 1 60, 115 reflective variable 111e, 113e knowledge processes 11e, 94e, 95, 120e, 122, 150e, 151e, 151, 152e input, throughout, output 47

176 Index knowledge sharing 13e, 14, 27, 123 internal 31e willingness 28 knowledge stock 128, 150e maintaining valuable 51 perceived 91 see also value of knowledge stock knowledge transfer 13, 14, 51, 52, 68, 139, 148 antecedents 83–91 barriers 30e, 157(n1 to ch4) beneficial 141 benefit 103, 139 between individuals 16e, 17 comprehensive model 37 conceptual studies 29, 31e, 32e, 33e, 34 contextualization problems 147 decoding and adaptation 106 decontextualization and recontextualization 57 dyadic relationship 52, 55 effectiveness 32e, 35–6, 49 empirical studies 29 ex post analysis 79–80 external 27 findings 30–3e ‘forward’ 140 headquarters–subsidiary 3 hierarchical 27, 138, 139, 141 hypotheses accepted 117e, 118e, 122, 126e, 127e, 132e, 136e, 137e hypothesis rejected 118e, 127e, 130, 131e, 132e, 135e, 137e improvement of effectiveness 106 from individual competence to internal structure 16e, 17 individual satisfaction 80 integrative model 35 internal 27 from internal structure to individual competence 16e, 17 intra-MNC 29, 30–3e intra-organizational 27, 29 lateral 27, 138, 139, 141, 142, 145, 148 methodology 30–3e micro perspective 34 MNCs 26–36 model 58–67, 157 motivations 147 nine types (Sveiby) 16e

not context-free 19 organizational and cultural barriers 53 phases 34 potentiality to engage in 88 qualitative studies 31e, 32e quality 80 quantitative studies 30e, 31e, 32e, 33e reverse 35 ‘reverse flows’ 140 speed 30e, 106 state-of-the-art (literature review) 29–36, 138 technological dimension 62 terminology 27 third dataset 130 varieties 16e, 26–7 within internal structure 16e, 17 see also perceived benefit; satisfaction knowledge transfer capabilities 27, 52, 59e, 61–4, 64e, 77–8, 79, 91–9, 108, 110e, 111e, 113e, 116, 119–21, 128, 138, 139, 141, 142–3, 144, 145, 148, 157 (n2–6 to ch4) further research 121 hypothesis 61, 64, 117e, 119, 121, 122 knowledge management infrastructure 94–5, 157(n4 to ch6) knowledge transfer processes 95–9, 157(n5–7) transmission channels 91–4 knowledge transfer effectiveness 59e, 64–7, 79–80, 103–7, 108, 110e, 111e, 129, 138, 139, 143, 144, 145, 146, 147, 148 aim of research 2–3 benefit from knowledge transfers 103–6 hypotheses 64, 66, 122–4 outcomes 147 perceived benefit 111e, 113e, 114e positioning 1–4 satisfaction 106–7, 107e, 111e, 113e, 114e third dataset 130 knowledge transfer infrastructure 143, 150e, 151e, 151, 152e reflective variable 111e, 113e

Index 177 knowledge transfer in MNCs: conceptual model 3, 58–67, 157 analysis and results 4, 81–137, 157–8 avenues for future research 4, 141, 142, 143, 145, 146–7 conclusion 4, 138–45 constructs 82–107 context 144–5 cultural distance (hypothesis 5) 66 data collection 3, 68, 70–6 descriptives of unit of analysis 81–2 gap between theory and empirical findings 144 hierarchical relationships and culturally close subsidiaries 107–24, 134–7e hierarchical relationships and culturally distant subsidiaries 124–9, 134–7e knowledge transfer capabilities (hypothesis 3) 61–4, 122, 157(n2–6 to ch4) knowledge transfer effectiveness (hypotheses 4–5) 64–7 lateral relationships 129–33, 134–7e limitations 145–6 managerial implications 4, 147–8 organizational distance (hypothesis 4) 65–6 position of respondents 83e results 115–37 stable perspective 146 strategic mandate (hypothesis 1) 58–60, 63, 64e, 115, 116 value of knowledge stock (hypothesis 2) 60–1, 157(n1 to ch4) see also Model of Intra-MNC Knowledge Transfer; research design and methodology; structural equation modelling knowledge transfer processes 14–19, 52, 61, 95–9, 120, 121, 122, 143, 144, 157(n5–7 to ch6) adaptation 15 communication model 15–18 composition 97e conceptualization 3, 157(n2 to ch4) context 15 decoding/encoding 15 four stages 15–16 hypotheses 116, 117e, 126e, 131e, 134e, 135e

implementation 16 initiation 15 media 8 ‘noise’ 15e, 52 second dataset 126e, 127e sender–recipient 15, 15e, 16, 17 spiral of knowledge 18–19, 62 storage devices 8 third dataset 131e, 132e time-lag 16 translation 16 variables 120e knowledge transmission channels 39, 47, 48, 54, 59e, 77, 91–4, 103, 122, 143 corporate socialization 44 formal 61–2, 92, 92e, 113e, 117e, 119, 121, 122, 125, 126–7e, 128, 130, 131e, 133, 134e, 135e, 139, 141, 142, 144, 150e formal integrative mechanisms 44 hierarchical/vertical 62, 92, 92e ‘hypotheses 1a–1d’ 63, 64e, 116 hypothesis accepted 117e, 119, 126e, 131e, 134e hypothesis rejected 117e, 119, 126–7e, 130, 131e, 134–5e informal 61, 62, 63, 93e, 116, 141 lateral 62, 92, 92e second dataset 126e third dataset 130, 131e Kogut, B. 8, 23, 25, 28, 29, 30e, 35e, 49, 50, 79, 86, 101, 102, 144 Köhne, M. 31e, 34 Korine, H. 23 Krogh, G. von 12, 31e, 34 language 2, 79, 101e, 102, 124 systematic 9 learning 22, 46, 51, 94e, 95, 120e, 123 learning culture 61 learning-by-doing 95, 96e, 97e, 143 Lehrer, M. 140 Leonard-Barton, D. 11e Leonidou, L. C. 65 Levinthal, D. A. 51, 61 liaison personnel 48, 77, 91, 92, 92e, 141 Liebowitz, J. 13e Likert-type scales 70, 77, 79, 80, 110, 153 literature business 1, 102 control mechanisms 45

178 Index literature – continued coordination and control 47 cross-cultural knowledge transfer 55 embeddedness of MNC units 69 international business 19, 157 (n1 to ch5) international JV 147 international strategy 2, 25 knowledge integration 147 knowledge management 1–2, 15, 52, 108, 124 knowledge transfer 29–36 multinational corporations 21, 43 product innovation 53 role of regional headquarters 140 strategic mandates (of subsidiaries) 39–40 strategic value disciplines 79 Local Implementers (autonomous subsidiaries) 41e, 42, 46, 48, 59e, 60, 64e, 77, 84, 84e, 85e ‘hypothesis 1c’ 63, 116, 157(n5 to ch4), 158(n12) Local Innovators 41e, 48, 59e, 60, 64e, 77, 84, 84e, 85, 85e, 86, 141 ‘hypothesis 1d’ 63, 116, 157 (n4–6 to ch4), 158(n11–13) location 2, 17, 25, 70 Long, C. 161 long-term orientation (concept) 102 longitudinal studies 146 Lyles, M. A. 144, 147 mail 71, 72 Malhotra, A. 32e, 35e, 35, 50, 66, 78, 108, 142 Malhotra, Y. 11e management 43 active 20 management practices 54 management style 38, 78, 99, 99e, 100, 100e management systems 41 managers 56, 70, 77, 96, 102, 106, 123 expatriate 48 headquarters 86, 93e, 94, 95, 100, 124 implications of research findings 147–8 major finding of study 145 subsidiary companies 48, 93e, 94, 95, 99, 124 working experience 93e, 94

Manev, I. M. 56 manuals 44 manufacturing 69, 72, 73e, 81 market analysis 90, 139 market data on competitors/ customers 88–90, 103–6 market knowledge 95 market mechanisms 66 market opportunities 66 marketing 23, 28, 29, 42, 69, 72, 76, 145 marketing know-how 88–90, 103, 103e, 104e, 104, 105, 105e, 106, 144 marketing managers 83e markets 50 inefficiency as means of knowledge transfer 26 national 23 Martinez, J. I. 40, 41e, 44, 46 Marx, J. 13e maximum likelihood (ML) method 153 McAdam, R. M. 12 McCreedy, S. 12 McDermott, R. 8 measurement error 69 measurement models 156e measures of fit 154, 155e, 155, 156 descriptive 156 incremental 155 stand-alone 155 suggested assessment 156e Mendenhall, M. 102 mentors 78, 93e, 94, 96e, 141 mergers and acquisitions (M & A) 29, 82, 85, 90 message transmission 15 Metanational Organization (Doz, Santos and Williamson) 22 misunderstandings (cultural) 101e, 102, 124, 133, 144 MNCs see multinational corporations mode of entry 77, 90, 143 mode of set-up 82, 90, 91 Model of Intra-MNC Knowledge Transfer 107 monitoring 44 Morgan, N. A. 65 Morosini, P. 55 Morrison, A. J. 40, 41e, 41, 43, 46, 86 motivation 28, 121, 143, 147, 157(n1 to ch4) Mudambi, R. 33e, 34

Index 179 Multifocal Organization (Doz) 22 multinational corporations (MNCs) 2, 81 ability continuously to combine and recombine knowledge 24 Austrian 73–4 business environment 23 combinative capability 142 competitive information 23 conceputalization 19–23 coordination and control 43–5 European Top 500 70–1, 157(n2 to ch5) evolutionary theory 26 foreign added-value activities 19 functional areas 69 heterarchical model 21, 22, 141 hierarchy model 22, 24 home-based model 21 integrated network model (Bartlett and Ghoshal) 22, 23, 24 integrative mechanisms 141 Japanese 18 knowledge and 3, 5–36 knowledge transfers 26–36 knowledge-based view 19, 23, 25–6, 34, 49 ‘knowledge-integrating institution’ 50 learning organization 21, 23–4 modes of entry 55 N-Form 22 network perspective 139–41 ‘network of units’ 21–2 relevance of knowledge management 19–36 requirements (Bartlett and Ghoshal) 20 resource-based view 19, 25–6, 49 role assigned to subsidiaries 46 shift from hierarchical to heterarchical conceptualization 37–8 ‘social community’ 50 strategic and organizational integration 20 structure 23–4 see also organizational units multinational corporations: knowledge-based determinants of strategic configuration 3, 37–57 capability perspective 49–52 contingency factors 52–7

coordination and control (different strategic mandates) 46–7 coordination and control (within MNC) 43–5 headquarters–subsidiary relationships 37–9 knowledge flow optimization 47–9 strategic mandates 39–43, 46–7, 47–9 Nahapiet, J. 31e, 34, 142 Netherlands 72 network effects 71 network knowledge 148 network perspective 139–41, 143, 146 networks 21–2, 31e, 33e, 34, 35e, 35, 37, 38, 44, 45e, 45, 46, 47, 53, 56, 58, 60, 133 global 86, 99 interdependent 22 strategic position in 116 Nobel, R. 32e, 35e Nobeoka, K. 31e, 146 Nohria, N. 12, 51, 37, 124, 157(n1 to ch5) Nonaka, I. 6, 9e, 9, 13e, 18, 18e, 62, 78, 95, 96 normed fit index (NFI) 114, 115e, 155e, 155, 156e norms 7 North America 74e, 75e North American Industry Classification System (NAICS) codes 71 Norway 72 objectivity 6 O’Brien, M. 57, 147 O’Dell, C. 8, 31e one-firm studies 146 operational distance 78–9 self-perception 78 operational mechanisms 99, 100 opportunity cost 123 opportunity generation 51, 62 organizational capabilities 3, 49–50, 67 effectiveness 142 networks 139 memory 13 processes 7 skills and routines 49 structure 68 see also institutional isomorphism

180 Index organizational distance 54, 59e, 65–6, 67, 86, 99–101, 103, 108, 110e, 110, 111e, 114e, 133, 138, 139, 145, 148, 157(n1 to ch6) hypothesis 66, 118e, 123–4, 127e, 132e, 137e second dataset 127e third dataset 131e organizational units (of MNCs) 63, 68, 80, 108 dual task 60 extent of integration in global network 82 hypotheses 60, 115, 116 orientation 79 potentiality to transfer knowledge 60 size (number of employees) 81–2, 82e, 85, 85e ‘source’ and ‘target’ 60 strategic position, 67, 147–8 unit of analysis 69–70, 157(n1 to ch5) working experience in other, 78 output control 45e, 45 outputs 44 p-value 115e patent citations 146 Pearson correlation coefficient 91, 101 Pedersen, T. 28, 33e, 35e, 35, 61, 157(n1 to ch4) peer subsidiaries see subsidiary–subsidiary relationships Pentland, B. T. 13e people/humans 17, 52 perceived benefit 103–6, 113–14e, 123–4, 129, 130, 139, 143, 144, 146, 150–1e hypotheses 117e, 118e, 122, 126e, 127e, 130, 131e, 132e, 135e, 136e, 137e proxy for knowledge transfer effectiveness 145 second dataset 126e, 127e third dataset 130, 131e, 132e perception 6, 79 cultural distance 102 headquarters–subsidiary relationships 37 perception gaps 69, 91 perception measure 101 performance 44, 54, 55, 57, 65 Perlmutter, H. V. 2, 21, 25

permanent teams 77, 91, 92, 92e personal contacts 95, 143 Pflesser, C. 111, 154, 156 ‘piece of knowledge’ 27 Plato 6 Poland 72 Polanyi, M. 8, 9 Porter, M. E. 2, 21, 25 Postrel, S. 31e power 14, 28, 38, 68, 148 Poynter, T. A. 22, 37, 41e Prahalad, C. K. 25 Probst, G. 7e, 7 product development 31e, 32e flows 43 knowledge 120e products 11e, 42, 43, 86, 94e, 95 ‘psychic’ distance 55 purchasing 42, 76, 139 purchasing know-how 88–90, 103, 103e, 104e, 105e, 106 questionnaires 68, 71, 79, 119 design and pre-test 70 pre-contacts (effect on response rate) 72 response rate 73–4 Raub, S. 7e, 7 Raven, A. 161 real estate 73e regression equations 108 research design and methodology 3, 68–80, 157 challenges in knowledge management research in MNCs 68–9 countries 72–6, 81 country of origin effects (insignificant) 74 data collection 3, 68, 70–6 final sample 68 operationalization and measures 68, 76–80 questionnaire design and pre-test 70 research context 68–70 sampling process 3, 68 unit of analysis 69–70 research and development 23, 32e, 69, 145 resource dependence 38, 44 resource flows 46 resources 22, 23, 25, 40, 49

Index 181 responsiveness 21, 22 ‘role’ 40 Romania 72 Romhardt, K. 7e, 7, 12 root mean squared error of approximation (RMSEA) 115, 156e Rosenzweig, P. M. 56 Roth, K. 41e routines 7, 9, 49 Rugman, A. M. 20 Russia 72

115e,

Sabherwal, R. 8, 32e, 35e, 35, 54, 66, 78, 96, 108 sales 42, 43, 69 Salisbury, D. 12 Salk, J. E. 144, 147 Salzberger, T. 72 Santos, J. F. P. 22, 29, 31e, 34, 53, 55, 62, 66 satisfaction 106–7, 107e, 122, 124, 125, 128–33, 139, 141–6, 148, 151e, 151, 152e hypothesis 117e, 118e, 127e, 131–2e, 135–7e proxy for knowledge transfer effectiveness 145 reflective variable 111e, 113e, 114e second dataset 127e third dataset 131–2e Schlegelmilch, B. B. 33e, 35e, 35, 72 Schmidt, R. A. 8, 13, 13e security 51, 62 Segars, A. H. 32e, 35e, 35, 50, 66, 78, 108, 142 service industries 30e service sector 72, 73e Shannon, C. E. 15, 15e, 16, 52 Shenkar, O. 79, 102 Shin, M. 8, 13, 13e Siguaw, J. A. 109n, 112n Simonin, B. L. 8, 28, 29, 65, 78, 108, 144, 145 Singapore 73 Singh, H. 79, 86, 101, 102, 144 Singh, J. V. 56 Sinkovics, R. 72 Sitkins, S. B. 56 skills 8, 22, 47, 49, 50 lacking 144 Slovakia 72 Slovenia 72

social capital 31e, 32e, 34, 142, 145 ‘social knowledge’ 10 socialization (knowledge conversion process) 18, 18e, 45e, 45, 46, 48, 61, 62–3, 64e, 78, 97, 97e, 98e, 98, 124 formal and informal 141 ‘hypothesis 1b’ 63, 116 Sölvell, O. 21 ‘South East Asia’ 73 space 2 Spain 72 Specialized Contributors (receptive subsidiaries) 41e, 42, 43, 46 spiral of knowledge 18–19, 62 Stevenson, W. B. 56 Stewart, K. A., et al. (2000) 11e, 161 Baskerville, R. 161 Long, C. 161 Raven, A. 161 Storey, V. C. 161 Storey, V. C. 161 strategic alliances 27, 144 Strategic Leaders 41e, 43, 46, 47 strategic mandates (of subsidiaries) 3, 21, 39–43, 45, 58–60, 64, 83–7, 98, 99, 108, 110e, 110, 111e, 113, 113e, 116, 140, 149e, 157(n3, n6 to ch4), 158(n10, n13) coordination and control 46–7 cross-tabulation with number of employees 85, 85e embedded 85 ‘global’ and ‘local’ 86 hypothesis one 60 independent 85 means and medians 84–5 measurement 76–7, 79, 157(n3 to ch5) medians 86 nature of operations 58 optimization of knowledge flows 47–9 perspectives 41 subsidiaries 39–43 ‘world’ 86 see also Global Innovators; Integrated Players; Local Implementers; Local Innovators strategic value disciplines (Treacy and Wiersema) 79 strategy 40

182 Index structural equation modelling (SEM) 4, 81, 86, 98–9, 107–24, 139, 157(n3 to ch5), 158(n9–13) assessment of model 114–15, 153–6 average variance extracted (AVE) 151, 152e, 155e, 156e causal relationships 115 construct-to-indicator relationships 110 constructs 109e, 111, 112e, 149, 149e–51e content validity 110 ‘data = model + residual’ 111 dataset one 114e, 125 dataset two 125e, 125 dataset three 125, 129e datasets 107, 139, 144 degrees of freedom 153, 154e, 155e, 156 description 108–13, 158(n9) descriptives 154e design 108, 109e factor reliability (FR) 151, 152e, 154, 155e, 156e formative variables 111e measures of fit 154, 155e, 155 ‘has to be overidentified’ 112 hierarchical (headquarters– subsidiary) data 107 hierarchical relationships and culturally close subsidiaries 107–24, 158(n9–13) hypotheses 115 ‘hypothesis-testing approach’ 108 identification 153 indicator reliability (IR) 151, 152e, 154, 155e, 156e indicators (formative and reflective) 110 just-identified 153 latent variables (factors) 108, 109e, 110e, 110, 111, 149 latent variables (endogenous/ exogenous) 108, 109e, 154e lateral (peer subsidiaries) data 107 maximum likelihood (ML) method 153 multivariate methods 108 null hypothesis 155 observable (manifest) variables (indicators) 108, 110, 115, 149, 149e–51e, 153, 154e, 158(n9) overidentified 153

parameter estimation 114, 153, 154e, 154 recursive and non-recursive 108 reflective variables 111, 111e rejection 153 reliability of measurement model 151–2 results 115–24 sample size 154e, 156 selection of variables 112e specification of model 114, 149–52 Squared Multiple Correlation 151 structural model v. measurement model 111 three partial models 113 unidimensionality of constructs 149–51 unstandardized results 115 variables 151, 154e structural equation modelling (SEM): Model One 81, 113e, 114, 115, 117e descriptives 154e first dataset 134–5e global measures of fit 115e second dataset 126e, 128, 134–5e third dataset 130, 131e, 134–5e structural equation modelling (SEM): Model Two 81, 113e, 114, 115, 117e–18e descriptives 154e first dataset 135–6e global measures of fit 115e second dataset 126–7e, 128–9, 135–6e third dataset 130, 131–2e, 133, 135–6e structural equation modelling (SEM): Model Three 81, 114e, 114, 115, 118e descriptives 154e first dataset 135–6e global measures of fit 115e second dataset 127e, 129, 136–7e third dataset 132e, 133, 136–7e structure–conduct–performance (SCP) paradigm 25 Subramaniam, M. 29, 32e, 35e, 55, 144 subsidiaries 13, 20, 22, 24, 29, 30e, 33e, 35e, 35, 41, 68, 69, 72, 79, 81–90, 92, 95, 97, 138, 139–40, 141, 142 autonomous initiative 47, 48 benefit from knowledge transfers 103

Index 183 benefits of knowledge transfers from culturally close subsidiaries 104e, 104, 106 benefits of knowledge transfers from culturally distant subsidiaries 104e, 104, 106 benefits of knowledge transfers from headquarters 103e, 104, 105–6 culturally close 100, 100e, 103, 114e, 107–24 culturally distant 100, 100e, 103, 114e, 124–9 focal 76e, 99, 100 global responsibility and authority 47, 48 greenfield investment 82, 85, 90 hierarchical relationships 107–29 ‘hypotheses 1a–1d’ 63, 116 informal transmission channels 93e, 94 integration/interdependence/ coordination 40 knowledge flows 76e knowledge stock 88e lateral interdependence 39, 47, 48 lateral knowledge flows 27 local adaptation 90 location 75e, 75–6 managers 70 mode of set-up 85 perceptions of similarity vis-à-vis headquarters and peer subsidiaries 99e responsiveness/autonomy 40 role 38, 39, 46 self-sufficiency 43 size 77, 82, 90, 91, 143 strategic mandates 39–43 strategic orientation scales 100–1 twofold embeddedness 38 via merger or acquisition 82, 85 see also headquarters–subsidiary relationships; strategic mandates subsidiary groups 99 subsidiary–subsidiary relationships (‘lateral relationships’/‘peer subsidiaries’) 46, 62, 65, 71, 83–4, 89, 90, 99, 113, 125e, 125, 144–5 ‘hypotheses 1a–1d’ 63 knowledge flows 76e knowledge stock 88e results 130–3, 134–7e

third dataset 129–3, 134–7e transmission channels 92 see also headquarters–subsidiary relationships summated-scales model 158(n9) Sveiby, K-E. 6, 16, 16e, 17, 32e, 34, 52, 62, 157(n1 to ch2) Sweden 72 Switzerland 72, 75e systems theory 6, 12 Szulanski, G. 15, 23, 28, 29, 30e, 35e, 157(n1 to ch4) t-tests 95, 97, 107, 115, 155 Takeuchi, H. 6, 9e, 9, 13e, 18, 18e, 62, 78, 95, 96 technology 28, 30e, 45, 52, 76, 139, 142 to monitor competition and business partners 95 problem-solving 95, 96e, 97e to retrieve, use and search for knowledge 95 technology know-how 88, 88e, 89e, 89, 90, 103, 103e, 104e, 104, 105e technology transfer 29, 30e Teigland, R. 32e, 145 telephone contacts 19, 71, 72 temporary task forces 77, 91, 92, 92e Tenkasi, R. V. 8 Thilenius, P. 160 Tierney, T. 12, 51 ties and network perspective 34, 35e, 37 time 2, 17, 120e, 123, 146 top/senior management 46, 51, 82, 102, 124 trade (wholesale and retail) 73e training 123 on-the-job 95, 96e, 97e, 143 transaction costs 50, 56 transmission channels see knowledge transmission channels transnational organizations 22, 24 heterarchical 24 transportation 73e Treacy, M. 79 Triandis, H. C. 160 Tsai, W. 21, 33e, 34, 35e Tsoukas, H. 11e, 12 uncertainty 38, 44, 45 United Kingdom 72, 75e Uppsala Model 55

184 Index Vahlne, J.-E. 55 value added/added value 19, 42 value chain 42, 46 ‘control circuit’ 14 four stages (Hong) 12–13 knowledge management 12–14 strategically driven 14 value creation 39, 142 value of knowledge stock 59e, 60–1, 67, 77, 83, 88–91, 110e, 111e, 113e, 138, 143, 148, 157(n1 to ch4) further research 121 headquarters compared with subsidiaries 89e hypothesis 61, 117e, 121, 126e, 131e, 135e perception 77 second dataset 126e subsidiaries compared with headquarters 89e subsidiaries compared with peer subsidiaries 88e third dataset 130, 131e variables 90, 157(n2 to ch6)

variances 153, 154, 155 Venkatraman, N. 29, 32e, 35e, 55, 144 Venzin, M. 12 Vernon, R. 20 Vladimirou, E. 11e, 12 Von Hippel, E. 29 Weaver, W. 15, 15e, 16, 52 Weggeman, M. 13e Wernerfelt, B. 25 White, R. E. 22, 37, 41e Wiersema, F. 79 Williamson, P. J. 22 Wolf, J. 140 World Mandate 41e, 43, 46 Zack, M. H. 8, 12 Zahra, S. A. 51 Zander, I. 21 Zander, U. 8, 23, 25, 28, 29, 30e, 35e, 49, 50 Zeybek, A. Y. 57, 147 Zhou, C. 144