Logistics Outsourcing Relationships: Measurement, Antecedents, and Effects of Logistics Outsourcing Performance (Contributions to Management Science) [1 ed.] 3790819166, 9783790819168, 9783790819380

This book uses state-of-the-art scientific methods to reveal that most firms currently do not realize the full potential

185 86 1MB

German Pages 364 [359] Year 2007

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
00.front-matter......Page 2
01.Introduction......Page 15
02.Basic concepts......Page 22
03.Theoretical framework......Page 68
04.Antecedents and effects of logistics outsourcing performance......Page 96
05.Methodology and sample characteristics......Page 157
06.Construct operationalization......Page 192
07.Structural models......Page 248
08.Summary and results......Page 284
09.back-matter......Page 294
Recommend Papers

Logistics Outsourcing Relationships: Measurement, Antecedents, and Effects of Logistics Outsourcing Performance (Contributions to Management Science) [1 ed.]
 3790819166, 9783790819168, 9783790819380

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Logistics Outsourcing Relationships

Contributions to Management Science www.springer.com/series/1505 A. Scholl Balancing and Sequencing of Assembly Lines 1999. ISBN 978-3-7908-1180-3 E. Canestrelli (Ed.) Current Topics in Quantitative Finance 1999. ISBN 978-3-7908-1231-2 W. Bühler/H. Hax/R. Schmidt (Eds.) Empirical Research on the German Capital Market 1999. ISBN 978-3-7908-1193-3 M. Bonilla/T. Casasus/R. Sala (Eds.) Financial Modelling 2000. ISBN 978-3-7908-1282-4 S. Sulzmaier Consumer-Oriented Business Design 2001. ISBN 978-3-7908-1366-1 C. Zopounidis (Ed.) New Trends in Banking Management 2002. ISBN 978-3-7908-1488-0 U. Dorndorf Project Scheduling with Time Windows 2002. ISBN 978-3-7908-1516-0 B. Rapp/P. Jackson (Eds.) Organisation and Work Beyond 2000 2003. ISBN 978-3-7908-1528-3 M. Grossmann Entrepreneurship in Biotechnology 2003. ISBN 978-3-7908-0033-3 H. M. Arnold Technology Shocks 2003. ISBN 978-3-7908-0051-7 T. Ihde Dynamic Alliance Auctions 2004. ISBN 978-3-7908-0098-2 J. Windsperger/G. Cliquet/ G. Hendrikse/M. Tuunanen (Eds.) Economics and Management of Franchising Networks 2004. ISBN 978-3-7908-0202-3 K. Jennewein Intellectual Property Management 2004. ISBN 978-3-7908-0280-1

M. J. Thannhuber The Intelligent Enterprise 2005. ISBN 978-3-7908-1555-9 C. Clarke Automotive Production Systems and Standardisation 2005. ISBN 978-3-7908-1578-8 M. Lütke Entrup Advanced Planning in Fresh Food Industries 2005. ISBN 978-3-7908-1592-4 U. M. Löwer Interorganisational Standards 2006. ISBN 978-3-7908-1653-2 G. Reepmeyer Risk-sharing in the Pharmaceutical Industry 2006. ISBN 978-3-7908-1667-9 E. Kasper Internal Research & Development Markets 2006. ISBN 978-3-7908-1728-7 L. Coleman Why Managers and Companies Take Risks 2006. ISBN 978-3-7908-1695-2 M. A. Bader Intellectual Property Management in R&D Collaborations 2006. ISBN 978-3-7908-1702-7 David L. Cahill Costumer Loyalty in Third Party Logistics Relationships 2007. ISBN 978-3-7908-1903-8 G. Cliquet/G. Hendrikse/ M. Tuunanen/J. Windsperger (Eds.) Economics and Management of Networks 2007. ISBN 978-3-7908-1757-7 Hartmut Hübner The Communicating Company 2007. ISBN 978-3-7908-1928-1 Eike A. Langenberg Guanxi and Business Strategy 2007. ISBN 978-3-7908-1955-7

Jan M. Deepen

Logistics Outsourcing Relationships Measurement, Antecedents, and Effects of Logistics Outsourcing Performance

With 26 Figures and 89 Tables

Physica-Verlag A Springer Company

Series Editors Werner A. Müller Martina Bihn

Author Dr. Jan M. Deepen Ahornallee 14 40468 Düsseldorf Germany [email protected]

Library of Congress Control Number: 2007927314

ISSN 1431-1941 ISBN 978-3-7908-1916-8 Physica-Verlag Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Physica-Verlag. Violations are liable to prosecution under the German Copyright Law. Physica-Verlag is a part of Springer Science+Business Media springer.com © Physica-Verlag Heidelberg 2007 The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Production: LE-TEX Jelonek, Schmidt & V¨ ockler GbR, Leipzig Cover-design: WMX Design GmbH, Heidelberg SPIN 11916413

88/3180YL - 5 4 3 2 1 0

Printed on acid-free paper

To my parents, Marion and Reinhard Deepen

Foreword

This dissertation by Jan Deepen is embedded in a research stream at the Kühne-Center for Logistics Management at WHU – Otto Beisheim School of Management – which explores the outsourcing of logistics services. Conceptually, Deepen is guided by the work of Engelbrecht (2004) which shows that the outsourcing of logistics services positively influences a firm’s logistics performance, which in turn positively influences the overall firm performance. The empirical research also indicated that not the degree of outsourcing is primarily decisive for the former relation, but rather the design of the outsourcing process itself. Within the design of this process, the relationship between logistics service provider and its customer is of particular importance. So far, little to nothing is known on the detailed influence of this relationship on the outsourcing performance. Exactly here is the starting point of this dissertation. It aims at discovering the performance effects that arise from the design of the outsourcing relationship. In the light of the strong effect of outsourcing performance on both logistics- and firm performance this question has not only theoretical, but also practical relevance. In the following, some results of the empirical research conducted by Deepen – which must be considered very advanced both because of its methodology and data quality – shall be highlighted. The complex outsourcing performance model developed by Deepen through extensive conceptual work in very large parts is supported by the empirical data. It consequently allows a very differentiated insight into the dependence between the different factors. Most of the hypothesized relationships are highly significant. Almost two-thirds of the variance of goal achievement is explained through the model, and so is more than one-third of the variance of goal exceedance. This is very encouraging. Viewing the derived hypotheses, it can be observed that the major part finds support. Only three hypotheses had to be rejected, nine could not be tested. These results are evidence of a very profound model formulation. The testing of the relationship between outsourcing performance and logistics performance also provides some very encouraging results. All four hypotheses find support. After all, more than ten percent of the level of logistics services and more than seven percent of the level of logistics costs

VIII

Foreword

are explained. Given the multitude of potential influencing factors of logistics performance, these values must be considered very satisfying. Finally, the findings on the relationship between logistics performance and firm performance are very convincing. Only one hypothesis must be rejected. The research basically confirms the results found in the works of Dehler and Engelbrecht and yet identifies some differences. These primarily concern the effect of the level of logistics costs, which is now shown to positively influence all three dimensions of firm performance. Deepen’s dissertation thus provides very valuable findings, also for a practical context. The robustness of the models against moderating effects of situational variables is another important finding of the study, which also may prove to be a valuable field for further research. Altogether, the dissertation by Deepen is very convincing in many aspects. It is characterized through excellent methodological skills and a precise and consistent theoretical development of the empirical research. The dissertation is built upon a very extensive literature review, which in particular includes the most recent international sources. The study provides important new insights. It consequently does not only advance the research on logistics service providers, but also has substantial practical potential. Therefore, this book will be worth reading for a very broad audience. Vallendar, January 2007 Prof. Dr. Dr. h.c. Jürgen Weber

Preface

It is a common misconception that writing a dissertation largely is a solitary endeavor – in my case it certainly has not been. At the Kühne-Center for Logistics Management at the WHU – Otto Beisheim School of Management – I have found an environment characterized by the interaction of a multitude of people, mutual support for each other, and a liberal and stimulating atmosphere. Creating this environment is the achievement of my doctoral advisor, Prof. Dr. Jürgen Weber. I would like to thank him for his guidance, support, and the freedom and responsibility he gave me during the years we have worked together. I would also like to thank my co-advisor, Prof. Dr. Lutz Kaufmann, who holds the Chair of International Management at the WHU, for providing additional insights and supporting me throughout my dissertation. My dissertation has greatly benefited from the close cooperation with Dr. David Cahill. Right from the beginning, we have been on the same time track with similar topics – while my dissertation took the customer’s view on third party logistics relationships, David’s takes the service provider’s view. This allowed us not only to share the survey data, but also to at least daily hold telephone conferences about the progress made, occurring problems, and potential solutions. This deep cooperation, which started out with two colleagues, ended with two friends. Dr. Carl Marcus Wallenburg, Managing Director of the Kühne-Center for Logistics Management, has also substantially contributed to the success of this dissertation. He has not only always been a very fair critical counterpart in research with brilliant ideas and a sharp mind, but also a good friend. The times together, be it with sports, games, or traveling I will always remember. Research always requires adequate funding, but even more so at a fully private financed institution. In this respect, I would like to thank the sponsors of my work: The Kühne-Foundation, whose support is the basis for all activities at the Kühne-Center for Logistics Management. My appreciation goes to Klaus-Michael Kühne, sole trustor of the foundation, and to its managing director Martin Willhaus. I would also like to thank Dr. Thomas Held, former CEO of Schenker AG, and Dirk Reich, Executive Vice

X

Preface

President Contract Logistics of the Kuehne + Nagel Group, for specifically sponsoring the survey. As I have pointed out above, the environment at the Kühne-Center was critical for my dissertation’s success. This is largely due to the people I met during my time and thanks go to Dr. Christoph Engelbrecht, Dr. Andreas Bacher, Dr. Marcus Groll, Dr. Alexander Schmitt, Nils Daecke, Ulrich Schulze, Dr. Alexandra Matthes, Peter Voss, Dr. Andreas Gebhardt, Wolfdieter Keppler, Serena Trelle, Peter Lukassen, Matthias Mahlendorf, Martina Bender and Christian Busse. Also, my work would have been a lot harder without our administrative assistants Beata Kobylarz, Fotini Noutsia, and Sonja Schmitt. The Kühne-Center is an integral part of the Chair of Controlling and Telecommunication. Located in close proximity, cooperation is close and fruitful. As representatives for the large number of people I have met there, who helped me or simply contributed to making it a great time, I would especially like to thank Dr. Bernd-Oliver Heine, Dr. Eric Zayer, Dr. Mascha Sorg, and Dr. Marius Lissautzki. I am especially grateful to Claudia Warning and Ekin Balik. They always offered me a place to stay when I came to Vallendar in the last phase of my dissertation and thus made my life a lot easier. In the course of the dissertation, I have also conducted research in the USA. The learnings have become important components of my dissertation and subsequent publications. I would like to thank Prof. A. Michael Knemeyer of The Ohio State University and Prof. Thomas J. Goldsby of the University of Kentucky. Tom and Mike have supported my research for years and have become friends during our frequent visits. Finally, I would like to thank the people most important to me: Haike, for always being there for me, for keeping up with my often volatile schedule and for making our time together so wonderful. Lars, my brother, for following my progress from a distance and for motivating me in the right moments. My utmost gratitude goes to my parents, Marion and Reinhard Deepen. They have always encouraged me to find my way and haven given me all the love and support a child can wish for. Without them, I would not be where I am now. To them, I dedicate this book. Düsseldorf, January 2007 Dr. Jan M. Deepen

Contents

1 Introduction............................................................................................. 1 1.1 Motivation ........................................................................................ 1 1.2 Goal .................................................................................................. 4 1.3 Structure............................................................................................ 5 2 Basic concepts.......................................................................................... 9 2.1 Logistics............................................................................................ 9 2.1.1 The nature of logistics ............................................................... 9 2.1.2 Status quo of logistics development ........................................ 16 2.1.3 Performance effects of the different levels of logistics development............................................................................. 17 2.2 Logistics outsourcing...................................................................... 19 2.2.1 Origin and definition ............................................................... 19 2.2.2 Benefits and risks of logistics outsourcing .............................. 21 2.2.3 Markets for and providers of logistics outsourcing ................. 24 2.2.4 Status quo of logistics outsourcing research............................ 29 2.3 Logistics outsourcing relationships ................................................ 33 2.3.1 The terminology of partnerships.............................................. 34 2.3.2 Partnership development ......................................................... 36 2.3.3 Designing logistics outsourcing relationships ......................... 41 2.4 Research model............................................................................... 48 2.4.1 Identification of research needs ............................................... 48 2.4.2 Identification of research questions ......................................... 51 2.4.3 Procedure to answer the research questions ............................ 52 3 Theoretical framework......................................................................... 55 3.1 Theories suited to explain cooperation in logistics relationships ... 55 3.2 Introduction to selected theories..................................................... 56 3.2.1 New institutional economics and transaction cost theory........ 56 3.2.2 Social exchange theory............................................................ 62 3.2.3 Commitment – trust theory...................................................... 69 3.2.4 Contingency approach ............................................................. 72 3.3 Theory integration .......................................................................... 79

XII

Contents

4 Antecedents and effects of logistics outsourcing performance ......... 83 4.1 Performance of logistics outsourcing relationships ........................ 83 4.1.1 Background of logistics outsourcing performance .................. 84 4.1.2 Conceptualization of logistics outsourcing performance......... 85 4.2 Identification of relevant antecedents ............................................. 88 4.2.1 Conceptualization of variables ................................................ 90 4.3 Formulation of a model of logistics outsourcing performance..... 102 4.3.1 Hypotheses on causal linkages between the variables........... 102 4.3.2 Overview of the hypotheses and consequent model.............. 119 4.4 Effects of logistics outsourcing performance ............................... 121 4.4.1 Logistics performance ........................................................... 122 4.4.2 Firm performance .................................................................. 127 4.5 Moderating effects........................................................................ 134 4.5.1 Relevance of adequate contingency variables ....................... 135 4.5.2 Conceptualization of contingency variables.......................... 139 4.5.3 Overview of contingency variables ....................................... 142 5 Methodology and sample characteristics.......................................... 145 5.1 Survey design ............................................................................... 145 5.1.1 Methods for data analysis ...................................................... 146 5.1.2 Method of data collection ...................................................... 147 5.1.3 Questionnaire design and pretest........................................... 149 5.1.4 Data collection....................................................................... 150 5.1.5 Data base, representativeness and potential biases................ 152 5.1.6 Characterization of the participating firms ............................ 153 5.2 Methodological basis for the empirical analysis........................... 155 5.2.1 Basics of measurement models.............................................. 156 5.2.2 Basics of structural models.................................................... 158 5.2.3 Measurement assessment....................................................... 160 5.2.4 Assessment of measurement and structural models .............. 162 5.2.5 Basics for model design and modification ............................ 175 6 Construct operationalization ............................................................. 181 6.1 Antecedents of logistics outsourcing performance....................... 181 6.1.1 Cooperation ........................................................................... 181 6.1.2 Communication ..................................................................... 183 6.1.3 Proactive improvement.......................................................... 186 6.1.4 Trust....................................................................................... 187 6.1.5 Commitment .......................................................................... 190 6.1.6 Functional conflict................................................................. 192 6.1.7 Involvement........................................................................... 195 6.1.8 Opportunism .......................................................................... 197

Contents

XIII

6.1.9 Shared values......................................................................... 199 6.1.10 Openness.............................................................................. 201 6.2 Logistics outsourcing performance............................................... 204 6.2.1 Goal achievement .................................................................. 204 6.2.2 Goal exceedance .................................................................... 207 6.3 Logistics performance .................................................................. 210 6.3.1 Level of logistics services...................................................... 210 6.3.2 Level of logistics costs .......................................................... 213 6.4 Firm performance ......................................................................... 215 6.4.1 Adaptiveness.......................................................................... 216 6.4.2 Market performance .............................................................. 217 6.4.3 Financial performance ........................................................... 219 6.5 Discriminant validity of the operationalized constructs ............... 220 6.5.1 Antecedents and dimensions of logistics outsourcing performance ........................................................................... 220 6.5.2 Logistics outsourcing performance and logistics performance ........................................................................... 223 6.5.3 Logistics performance and firm performance........................ 225 6.6 Contingency factors ...................................................................... 226 6.6.1 External contingency variables.............................................. 226 6.6.2 Internal contingency variables............................................... 231 7 Structural models................................................................................ 237 7.1 Antecedents and dimensions of logistics outsourcing performance.................................................................................. 237 7.1.1 Presentation of the basic model ............................................. 237 7.1.2 Development of a simplified model ...................................... 239 7.1.3 Discussion of the final simplified model ............................... 241 7.2 Effects of logistics outsourcing performance ............................... 247 7.2.1 Logistics outsourcing performance and logistics performance ........................................................................... 248 7.2.2 Logistics performance and firm performance........................ 251 7.3 Contingency variables .................................................................. 257 7.3.1 Moderating effects on the model of logistics outsourcing performance ....................................................... 257 7.3.2 Moderating effects on the model of logistics performance ... 264 7.3.3 Moderating effects on the model of firm performance.......... 266

XIV

Contents

8 Summary and results.......................................................................... 273 8.1 Main results .................................................................................. 273 8.2 Managerial implications ............................................................... 279 8.3 Recommendations for further research......................................... 281 Appendix: Questionnaire ...................................................................... 283 List of figures.......................................................................................... 299 List of tables ........................................................................................... 301 References............................................................................................... 307

1 Introduction

1.1 Motivation The worldwide usage and importance of logistics outsourcing has grown dramatically over the last decades and will continue to do so. Indeed, a recent study by LANGLEY/DORT/ANG/SYKES (2005) indicates that from total logistics expenditures in 2005, 57% in Western Europe and 44% in the U.S. were directed towards logistics outsourcing. For these figures, they expect to see growth rates of 18% in Western Europe and 16% in the U.S. in the near future (2008-2010). Similarly, usage rates for logistics outsourcing services in the United States have increased from approximately 40 percent of Fortune 500 companies in the early 1990s (LIEB 1992) to approximately 80% in 2004 (LIEB/BENTZ 2004). It has also been estimated that about 40 percent of global logistics is outsourced (WONG/MAHER/NICHOLSON/GURNEY 2000). The continuing globalization of operations will only serve to reinforce this growing reliance on logistics service providers (LSPs) worldwide (ZHU/LEAN/YING 2002). The ability of LSPs to maintain an increasingly relevant role in today’s global supply chains will be largely driven by their continued ability to provide value to their customers. This value arises from both accommodating and exceeding customer service expectations in a more cost effective manner than can be achieved by customers performing the activities themselves. While some companies have made the decision to maintain control of logistics activities, numerous others worldwide have decided to outsource these activities. As recently demonstrated by DAUGHERTY/STANK/ELLINGER (1998), DEHLER (2001), and STANK/GOLDSBY/VICKERY/SAVITSKIE (2003), a company’s logistics performance has an influence on the overall firm performance. ENGELBRECHT (2004) contributed to the discussion by showing that indeed, the outsourcing performance is an important driver of the logistics performance. Thus, clearly understanding the drivers of logistics outsourcing performance is critical knowledge for managers to have in today’s competitive business environment. On this, ENGELBRECHT (2004) found that the degree of logistics outsourcing,

2

1 Introduction

e.g. the extent to which a firm outsources it logistics processes to a logistics service provider, is not the main driver. Instead, it may be assumed that this can be found in the outsourcing arrangement itself. Much remains to be known on how these outsourcing arrangements should be designed. Aside from the technical implementation side, this pertains particularly to the relationship design between the LSP and its customer. While research from relationship marketing supplies a number of very relevant insights (ANDERSON/WEITZ 1989; ANDERSON/ NARUS 1990; MORGAN/HUNT 1994), the adaptation by logistics research has been rather scarce (MOORE 1998; STANK/GOLDSBY/ VICKERY/SAVITSKIE 2003; KNEMEYER/MURPHY 2004). However, the recent findings of logistics outsourcing research focusing stronger on the relationship perspective rather than technical issues are very promising to understand the true drivers of logistics outsourcing performance and its implications. This indicates a particular relevance of further research in the field. As yet, the relevant antecedents of logistics outsourcing performance remain to be identified and must be arranged into a consistent model. Furthermore, these findings must be integrated into the research of DAUGHERTY/STANK/ELLINGER (1998), DEHLER (2001), STANK/ GOLDSBY/VICKERY/SAVITSKIE (2003), and ENGELBRECHT (2004) to obtain a holistic view of the entire logistics performance chain, spanning from the initial outsourcing performance over the overall logistics performance to the firm performance of the customer of the LSP. Before this can be accomplished, a measurement model for logistics outsourcing performance must be developed first. Currently, a large number of different measurement approaches exist, this owing at large to the complexity of the measurement object. Those scales proposed by STANK/GOLDSBY/VICKERY/SAVITSKIE (2003), KNEMEYER/ MURPHY (2004), and ENGELBRECHT (2004) supply detailed and precise insights. All of them are multi-dimensional constructs, in each case focusing on different aspects of outsourcing performance. This makes concentration and consolidation necessary, keeping in mind the different aspects of logistics outsourcing and everyday business reality. The logistics outsourcing decision has been identified as a key consideration in contemporary logistics management (MURPHY/POIST 2000). While its strategic character has been emphasized especially through the insight that it is a facilitator of firm performance, the dominant motivation for logistics outsourcing is still the reduction of logistics costs (ENGELBRECHT 2004). The shorter time horizons these considerations often have compared to attempts to increase the logistics service levels, even though they influence the boundaries of the firm and its competencies, at times lead to the wrong focus in the outsourcing arrangement. So-

1.1 Motivation

3

metimes they even cause their failure. Firms aiming purely at cost reductions will emphasize other aspects in the logistics outsourcing arrangement than those intending to achieve service level increases. This is of particular relevance to the design of logistics outsourcing relationships between LSPs and their customers. Only when the concept of logistics outsourcing performance, its measurement, its effects as well as its antecedents are known and understood, these problems can be overcome in logistics outsourcing practice. To do so, a study conducted with the adequate scientific rigor as demanded by MENTZER/KAHN (1995), proposing distinct hypotheses and then testing them empirically is necessary. It thus is building on the findings of the research conducted in previous years, consolidates it and develops new models that allow further insights for both academic research as well as the logistics practice. Finally, a last deficit of current logistics research must be addressed in this study. The efforts outlined above will facilitate a thorough and general understanding of logistics outsourcing relationships and their effects on logistics outsourcing performance, as well as its effects on logistics- and firm performance. However, the open questions concerning the importance of the role of the context of the firm in the logistics outsourcing context and the general applicability of the performance models remain (CHOW/ HEAVER/HENRIKSSON 1994; PFOHL/ZÖLLNER 1997). Answering these questions, however, is of particular relevance as especially in the logistics outsourcing practice very few standardized outsourcing solutions for complex processes and problems have been developed, thus hindering further cost reductions. To determine what outsourcing strategy has primacy in which context, CHOW/HEAVER/HENRIKSSON (1994, p. 26) suggest the use of contingency models of logistics performance which should include factors such as the environment, the logistical features of the product range or the production technology. While the contingency approach has been criticized in the past for its mechanistic view that there is only one best structural answer to any specific contextual situation and its lack of theory (HAGE 1974; SCHREYÖGG 1980), it can contribute to new insights. If adequately used as a measure to understand the environmental contingency of firms, it can point out that under certain circumstances, differences in organization structure and administrative practices have different implications for a firms’ performance.

4

1 Introduction

1.2 Goal The intention of this study is to contribute to the logistics outsourcing discussion by analyzing the relationships between logistics service providers and their customers. It aims at discovering the performance implications that the design of logistics outsourcing relationships has. A large-scale empirical study was conducted which serves to answer four distinct research questions which were developed to break down the complex problem into several smaller parts that can be individually addressed. The research questions are derived from current research deficits and are in detail discussed in chapter 2.4. In the survey, the relationship between a customer and its most important LSP was analyzed from a perspective of equality between the two parties. Consequently, the term “customer” in this study will, if not specified otherwise, always refer to the customer of the logistics service provider, a common procedure in contemporary logistics outsourcing research (DAUGHERTY/STANK/ELLINGER 1998; STANK/GOLDSBY/VICKERY/SAVITSKIE 2003; KNEMEYER/MURPHY 2005; LANGLEY/DORT/ANG/SYKES 2005). In the following the open questions this study is aiming to address will be briefly described. Its starting point is, as argued above, the relationship between the customer and its LSP. For the customer, a main problem is to identify the factors that influence the relationship. This is a complex task: on the one hand, a multitude of different variables and theoretical approaches exist to choose from. On the other hand, the customer has a particularly strong interest in the implications of the relationship variables for the logistics outsourcing performance, which in turn supposedly influences the firm’s logistics performance. Consequently, the main influencing factors of the relationship must be identified and their effects on the outsourcing performance examined. Additionally, it is of particular interest to see how these different variables might be integrated into one model that comprises the relationship variables as well as their performance effects for the outsourcing of the logistics activities. Here, special consideration must be given to the interdependencies existing between the variables in order to adequately estimate the causal linkages inside the model and to the design of the logistics outsourcing performance construct. The question of the performance effects of the relationship design between customer and LSP is only a strategic one if a connection can be shown between the outsourcing performance and the resulting logistics performance of the customer. The relevance in that case would result from the insight that the design of the outsourcing relationship would have a di-

1.3 Structure

5

rect influence on the logistics performance. This in turn may be of particular relevance if it could be shown to positively affect the overall firm performance. While the recent research as shown in chapter 1.1 suggests that this link indeed exists, it must be tested also for the specific logistics outsourcing performance construct developed in this study. Finally, the question of the validity of the answers to the research questions that deal with the issues presented above must be addressed with respect to the context of the firms. Since the individual context may vary depending on the firms’ specific and individual situations, some of the causal linkages in the models are presumably moderated by those contingency variables. It therefore is necessary to first identify contingency variables with potential moderating effects. In a second step, their influence on the logistics outsourcing relationship model and its performance effects must then be analyzed.

1.3 Structure This study is organized as follows, in general pursuing the framework suggested by MENTZER/KAHN (1995) for theoretically rigorous logistics research. In chapter 1, the motivation for the research is outlined. Building on that, the four research questions that guide this study are presented before the structure of the study is discussed in detail. Chapter 2 introduces the basic concepts of this study. Aside from logistics, its origin, development and the status quo of research, logistics outsourcing and logistics outsourcing relationships are presented. To approach the logistics outsourcing discussion, first its origin and definition is introduced. Then, the benefits and risks of logistics outsourcing are examined and the markets for and the providers of logistics outsourcing services are analyzed, before in chapter 2.2.4 the status quo of logistics outsourcing research is put forward. Finally, chapter 2.3 presents the research on logistics outsourcing relationships. It starts of with an analysis of the terminology of partnerships, before it continues with pointing out why partnerships and their development are of importance in general in buyer-supplier relationships. Then it goes on and discusses what implications this knowledge has for logistics outsourcing relationships in particular, first presenting the status quo of relationship marketing research and then its aspects that have already made their way into current logistics outsourcing relationship marketing. Finally, chapter 2.4 summarizes the findings from the previous chapters and on these grounds identifies different research needs which in chapter 2.4.2 lead to the formulation of four distinct research questions that

6

1 Introduction

will guide further research efforts of the study. There, the procedure to answer these questions is also outlined in detail. Chapter 3 then presents the theoretical framework of the study. Three different theories, transaction cost theory, social exchange theory, and commitment-trust theory, and the contingency approach are all argued to have explanatory value for logistics outsourcing research. Consequently, each of them is presented in detail before an individual examination of its explanatory value is performed. Chapter 3.3 then integrates the theories as an important step for the further analyses. In chapter 4, the antecedents and effects of logistics outsourcing performance are discussed. Initially, the measurement of logistics outsourcing performance is scrutinized and a bi-dimensional construct is conceptualized. Then, the ten most relevant relationship antecedents of logistics outsourcing performance are derived from the literature review and the theories presented in chapter 3. First, they are conceptualized in chapter 4.2.1, before in chapter 4.3.1 hypotheses on the causal linkages between the variables and logistics outsourcing performance as well as amongst themselves are developed. This leads to the development of a comprehensive model of logistics outsourcing performance in chapter 4.3.2. Thereafter, logistics performance and firm performance are conceptualized. Consequently, the effects of outsourcing performance on logistics performance are hypothesized before the effects of the latter on firm performance follow, thus producing a model of the entire logistics performance chain. Chapter 4.5 finally discusses potential moderating effects of internal and external contingency variables on the models developed in the previous chapters. Chapter 5 introduces the methodology of the research and the sample characteristics. At first, it presents the survey design, including covariance structure analysis as the method of choice in this study, the method of data collection, the questionnaire design as well as questions concerning representativeness and potential biases. It then continues with a discussion of the methodological basis for the empirical analysis, including a detailed discussion of fit criteria that serve to asses the quality of both measurement and structural models. At the end of the chapter, the basics for model design and modifications are outlined. The operationalization of the different constructs conceptualized in chapter 4 is performed in chapter 6. Starting with the antecedents of logistics outsourcing performance in chapter 6.1, all constructs of logistics outsourcing-, logistics-, and firm performance are operationalized in detail. After having done this in individual chapters, the discriminant validity of the constructs in the individual models is scrutinized in chapter 6.5, before in chapter 6.6 the external and internal contingency variables are operationalized.

1.3 Structure

7

Chapter 7 introduces the empirical analyses of the structural models. In chapter 7.1, a basic model of the antecedents and dimensions of logistics outsourcing performance as well as a necessary simplification is presented, before the results of the final model are discussed in detail. Similarly, chapter 7.2 presents the models of logistics performance and firm performance and discusses them accordingly, before in chapter 7.3 the moderating effects of the contingency variables on the three performance models are analyzed. The final chapter is chapter 8. It summarizes the main results, develops some managerial implications and outlines the needs for further research that either could not be addressed in this study or have are a direct consequence of this research.

2 Basic concepts

2.1 Logistics 2.1.1 The nature of logistics The concept of logistics in its modern form dates back to the second half to the 20th century (DEHLER 2001, p. 34; WEBER/KUMMER 1998, pp. 1-6). Since then, it has developed into a widely recognized discipline of significant importance to both theory and practice. As WALLENBURG (2004, p. 38) points out, this development is not yet completed, however, and the debate on the true meaning of logistics and its exact specifications is still ongoing: Especially in the logistics industry it becomes apparent that neither a standardized logistics concept nor a consistent notion of logistics exists. While some reduce their understanding to simple transporting-, handling-, and warehousing operations, others view logistics more broadly as a management function. Logistics literature supports this finding of notional heterogeneity with a multitude of different logistics definitions. Especially recognized is the 2005 definition by the Council of Supply Chain Management Professionals (CSCMP 2005, p. 63), where logistics management is seen as part of supply chain management (SCM). It is the part “… that plans implements, and controls the efficient, effective forward and reverse flow and storage of goods, services, and related information between the point of origin and the point of consumption in order to meet customers’ requirements.” This definition directly refers to the importance of economical considerations (efficiency, effectiveness) and at the same time underscores the functional character of logistics. Other definitions such as those of KLAUS (1993, p. 31) and WEBER/KUMMER (1994, p. 21) focus on process- or floworientation and thereby emphasize conceptual components. The apparent differences in the understanding of logistics and the resulting myriad of definitions are a direct consequence from logistics’ nonacademic origin. It rather developed as a practical phenomenon which has gained increasing importance over the past decades for firms in various industries. As WEBER (2002, p. 4) points out, the basic function of logistics

10

2 Basic concepts

is to ensure the availability and supply with required resources, which is based on the flow of materials and goods. However, this concept is not an invention of modern times, but has been of great significance both to public and private enterprises in the form of the transportation, handling, and warehousing of goods. It thereby serves as a foundation for a society based on the division of labor (WALLENBURG 2004, p. 38). Even when taking all the different opinions into account, a consensus exists in logistics research that the central function of logistics is the bridging of space-time disparities concerning goods and materials (WALLENBURG 2004, p. 38). It thereby is an independent function with various possible forms of specialization. Beyond that, logistics is a management concept which, according to PFOHL (1999, pp. V-VI), ought to be anchored throughout the organization with direct consequences for all corporate functions.

Level of logistics knowledge 4

Logistics as supply chain management

3

Logistics as enabler of process orientation within the whole firm

2

Logistics as coordinative function

1

Logistics as functional specialization

Absence of distinctive logistics knowledge Time

Fig. 2-1. The four phases of logistics development (WEBER 2002, p. 5)

This understanding is displayed in considerable manner when research efforts are undertaken to examine and systemize the subject of logistics. Commonly, the systematizations are based on multi-phase concepts derived from empirical research on the development of logistics in a corporate context. Most of these concepts indicate that the development of logistics follows three or four distinct phases (BOWERSOX/DAUGHERTY 1987; WEBER/KUMMER 1998; GÖPFERT 1999; KLAUS 1999a; WEBER 2002), where sometimes the most advanced two phases are viewed as a single phase only. These phases, as indicated in Figure 2-1 (WEBER 2002, p. 5),

2.1 Logistics

11

are determined by the level of logistics knowledge present in a firm and require path dependent development from the lowest to the highest level of logistics knowledge. During the first two phases, efficiency gains of the logistical processes are emphasized, both through specialization and the cross-functional coordination of material flows. After the transition to the third and fourth phases the scope of logistics changes distinctly. It becomes a management function, whose objective is the implementation of a flow- and processorientation throughout the firm, thereby fostering logistical thinking and acting beyond the sole logistics department. However, WEBER (1999, pp. 3-4) points out that even when a firm has reached those higher phases of logistical development, it is important that the functions typical for the lower phases are not neglected. The different phases of logistical development reflect an underlying shift of importance. ALT/SCHMID (2000, p. 80) state that, coming from an emphasis on classical logistical activities such as transportation, handling and warehousing, the flow of information in logistics processes is of increasing concern. While in the early years of logistical development the physical capabilities of a logistics system determined its potential, this has changed until today, where the capabilities of the complementary processes of information exchange are of at least equal importance. In the following chapters, the different phases of logistical development will be shown in greater detail. 2.1.1.1 Logistics as a functional specialization

During the first phase of its development, logistics becomes a specialized function, supplying services and processes required for the efficient flow of materials and goods. These processes mainly include the transportation, handling and warehousing of goods which previously had not been adequately addressed. Historically, the emergence of the first phase of logistical development was caused by a severe change in the market environment in the 1950’s (DEHLER 2001, p. 13). The traditional suppliers’ markets turned into buyers’ markets, requiring new and more sophisticated flows of materials and goods. In contrast to other functions, such as procurement or production, the logistics function back then was underdeveloped and logistics responsibilities were scattered throughout the organization. For this reason, a concentration on the optimization of this function promised broad room for improvement. Through the functional specialization, two separate benefits can be obtained, coming either from the direct optimization of individual processes

12

2 Basic concepts

or from the joint treatment of different processes. WALLENBURG (2004, p. 40) indicates that improvements on the process level can result from experience curve effects or economies of scale. Furthermore efficiency gains can be realized on the planning level through the application of mathematical methods, solving e.g. non-trivial transportation and warehousing problems. Beyond these improvements, optimizing different logistics processes jointly promises great potential that can only be realized if existing interdependencies are taken into account, e.g. when rising costs incurred through higher transportation frequencies are offset by savings through lower inventory levels. On the organizational level, a specialization of the logistics function often leads to the introduction of new departments, combining transporting-, handling- and warehousing functions. At the same time, a functional division can often be observed on a firm-wide basis, created through separate efforts in the areas of procurement-, production-, and distribution-logistics. As a specialized service function, logistics is characterized by the existence of a considerable know-how spread among a clearly definable group of employees. In summary it can be ascertained that the mastering and understanding of the requirements of the first phase of logistical development promises considerable improvements and efficiency gains and simultaneously is the necessary basis for the following phases. 2.1.1.2 Logistics as a coordinative function

After exhausting rationalization potentials during the first phase especially in distribution and transport-intense procurement functions, the focus during the second phase of logistical development is on the coordination of different functions. The efforts concentrate both on the coordination of the flow of materials and goods from source to sink and on expanding the focus towards the entire supply chain, cutting across the boundaries of the firm and comprising customers as well as suppliers (WEBER 1999, pp. 34). Starting point for understanding logistics as a coordinative function was the insufficient consideration of existing interdependencies between different functions of the firm. Facilitated by existing structures, especially procurement, production, and distribution functions were optimized independently. The organizational separation of these functions, however, historically encouraged the development and cultivation of individual interests, obstructing an overall optimization of all processes. But exactly the latter was needed, since the optimization potentials due to specialization for the single functions were already exhausted. Therefore, during the sec-

2.1 Logistics

13

ond phase of logistical development, improvements can be achieved by concentrating on the coordination of the different functions. Examples given by WEBER (2002, p. 11) are the coordination of lot sizes or just-intime supply and production, where the required resources are provided exactly when needed. Resulting from the integrated understanding and planning of the procurement and production functions, cost and performance benefits emerge. The focus thus is on influencing the extent and the structure of the demand for logistical services through appropriate coordination. In doing so, logistics is giving up the former functional separation and rather focuses on integrated processes. This fundamental change in the understanding of logistics causes an increased heterogeneity of the function on the one hand and on the other requires an increased interaction with the responsible management of other functions. The perceived importance of logistics increases during the second phase of logistical development as it is now seen as a means to achieve competitive advantages. The primary concern during this phase is to enable cost leadership –differentiation through performance will be targeted mainly during the following phases. The second phase is building on the know-how of the functional specialization, supplemented by substantial inter-organizational and management knowledge needed for the coordination. Therefore, not only the amplitude of the necessary logistical knowledge increases, but also its depth. 2.1.1.3 Logistics as enabler of process orientation within the firm

The transition towards the third phase of logistical development is characterized by yet another change in the relevance attached to it. Logistics now becomes a management function aiming at implementing the concept of flow orientation inside the entire firm (DEHLER 2001, pp. 16-18). Historically, this development was caused by the changing economic environment. The increasing competitive pressure called for differentiation while simultaneously reducing costs. For doing so, the purely functionally designed structures and systems proved inapt. Yet, by adopting a stronger process orientation when supplying logistical services, complexity reductions could be achieved, thereby better addressing the shifted needs of the markets. Additionally, this can be supported through a stronger emphasis on flexible coordination methods, such as self-coordination of the employees rather than the rigid coordination through plans (KIESER/KUBICEK 1992, pp. 106-117). Because of the transition into a management function, the implementation of flow orientation is not restricted to individual corporate functions.

14

2 Basic concepts

In contrast to the approach of the second phase based on coordination, all logistical structures are generally perceived as being changeable. Thus, when implementing the concept of flow orientation, the original logistical processes transporting, handling, and warehousing lose their exposed significance. Their remaining importance comes from their contribution to the proper functioning of the flow orientation of the firm. With the increasing importance of logistics as a management function, the required logistical knowledge increases as well. At the same time, the broad logistical know-how obtained in the first phases allows a reduction of the distinct specializations on the different logistics functions. Logistical services can now for instance be provided by the same employees responsible for supplying production- or maintenance-activities. In practice, corporate logistics following this understanding are sometimes criticized since they may fail in some of their very basic aspects (WEBER 2002, p. 19): one danger is that with the broader orientation the unique and original logistical skills may suffer. On the other hand, when logistics become a management function it runs the risk of not being adequately anchored in the organization. Consequently, the functional specialization must not necessarily be abandoned when a firm progresses towards the third phase of logistical development. Rather, it is vital to find a compromise which enables and fosters the coexistence of a functional specialization for the supply of logistical services and at the same time anchors the understanding that flow orientation as an important task of the management. 2.1.1.4 Logistics as supply chain management

During the fourth and last phase of logistical development, logistics remains a management function, but extends its scope beyond the boundaries of the firm. Consequently, the concept of process or flow orientation is extended across the supply chain, encompassing now also suppliers and customers, thus ideally spanning from source to sink. Logistics during this phase, now being called supply chain management (SCM), aims at integrating the entire supply chain. This understanding of the concept of supply chain management as a phase of logistical development is not undisputed. As LARSON/ HALLDORSSON (2004, pp. 1-7) point out, in the logistics science community basically four different views of SCM have developed over the years. These include the “traditionalist” view which understands SCM as part of logistics and the “unionist” view which considers logistics as part of SCM. Furthermore, the “re-labeling” perspective believes that what is now SCM was previously logistics. The fourth and “intersectionist” view finally sug-

2.1 Logistics

15

gests that logistics is not the union of logistics, marketing, operations, purchasing etc. but rather includes strategic and integrative elements from all these disciplines. Further insights into the diversity of understandings are given by BECHTEL/JAYARAM (1997) who provide an extensive retrospective review of the literature and research on supply chain management. In the light of this multitude of different understandings it is important to establish that in this work, supply chain management is understood as the most advanced phase of logistical development. While this conflicts with several of the above presented views, it represents the current and widespread understanding of logistics and SCM in Germany. Starting point for the development towards supply chain management was the further increasing demand of firms for more efficiency and effectiveness. Since most of the internal optimization potentials had already been exhausted, only those remained that resulted from the inefficient collaboration between firms being part of the same supply chain. The fact that during this process the individual boundaries of the firm lost part of their former dominant importance was fundamentally enabled by the tremendous progress the information and communication technologies made. Even though supply chains are part of every economy based on the division of labor and therefore have already existed during the other phases of logistical development, it is only during the fourth phase that they obtain a widely recognized importance. Thus, what is new to this phase is the concentration on the supply chain and the introduction of inter-organizational concepts aiming at the realization of optimizing potentials by targeting gains in efficiency and effectiveness. Due to the high complexity of the task and the divergent objective functions the realization of an inter-organizational supply chain management is accompanied by management problems.1 While in partnerships with low intensity the focus is usually only on the adequate supply with information, an increasing intensity requires adjustments in structures and processes as well in order to prepare the former internal structures for the now interorganizational challenges. The management tasks during this phase of logistical development are considerable and complex. Together with the understanding of the need for 1

According to Wallenburg (2004, p. 43) problems arise because firms are usually not only part of one supply chain, but of several. Therefore, firms can be faced with different and incompatible goals for the supply chain management. Additionally, endeavors beneficial for one supply chain can be detrimental for another. Even though, emphasizing the character of a supply chain rather than a supply network is preferable, since the many linkages and connections between the firms would lead to a degree of complexity that would not be manageable.

16

2 Basic concepts

inter-organizational cooperation for supplying goods and services, they are the reason why supply chain management is an own phase of the logistical development. Prerequisite for the implementation of an interorganizational flow orientation not only is the answering to the technological demands, but also the sufficient willingness and capabilities of the participating firms. 2.1.2 Status quo of logistics development After presenting an abstract picture of the empirically observed logistics development in the previous chapters, the present state of this development remains to be shown. DEHLER (2001) and WALLENBURG (2004) both conducted surveys to identify the present state of logistics development in German firms for the years 1999 and 2002. Both surveys used the same scale, which in chapter 6.6.2 of this study is used as an internal contingency factor, while they did not address the same sample of firms and industries. A comparison of the two studies, as WALLENBURG (2004, pp. 4445) shows, has a lot of explanatory potential, because it reveals the status quo of logistics development on the one hand and enables the observation of changes that have occurred in the three year period between the two studies on the other hand. The results reveal a rather traditional understanding of logistics among German firms. In 2002, 68% of the firms are located in one of the first two phases of logistical development. According to their view, the functional specialization or the coordinative function of logistics is of dominant importance. Furthermore, only 12% of the firms see themselves in the fourth phase of logistical development and concentrate on supply chain management and the implementation of inter-organizational flow orientation. These results, however, have their own dynamics and are subject to constant change. Since DEHLER first conducted his survey in 1999, significant advancements can be observed. While in 2002 only 68% of the firms (compared to 82% in 1999) saw themselves in one of the first two phases, the number of firms in the fourth phase almost doubled from 7% to 12% in 2002.

2.1 Logistics

17

Percentage of firms 100

7

90

12

11 80

20

70 60

Phase 4 57 Phase 3

40

82 Phase 2

33

50

Phase 1

40

13

30 20

42

35

22

10

5 8

8 1999 2002 Current logistics development

6

1999 2002 Aspired logistics development

Fig. 2-2. Logistics development in Germany between 1999 and 2002

Furthermore, a significant majority of the firms in 2002 aimed at further developing their logistics. While only 6% intended to remain in the first phase of logistical development, 82% of the firms indicated their ultimate goal to be the fourth phase. This also represents a considerable increase from 1999. Altogether, it can be asserted that a large potential exists among German industrial and retail firms for further development, which is recognized increasingly also by the firms themselves. 2.1.3 Performance effects of the different levels of logistics development As described above, significant advancements in the field of corporate logistics can be observed in recent years. It remains an open question, however, whether or not it is desirable for every individual firm to aim at reaching as high a level of logistics development as possible and to implement logistics as a management function, thereby enabling an interorganizational flow orientation. This will only be the case if it proves that flow orientation is a key performance driver both for logistics and firm performance. DEHLER (2001, pp. 220-226) shows empirically that the higher the flow orientation of a firm, the higher is its logistics performance due to reduced logistics costs and increased levels of logistics service. This is confirmed

18

2 Basic concepts

by ENGELBRECHT (2003, p. 64) in an independent study with a new sample of firms. This finding is of particular relevance, because DEHLER (2001, pp. 233244) also finds that logistics performance directly influences the overall firm performance. As indicated in Figure 2-3, lower logistics costs have a positive direct, and therefore also total, effect on financial performance. However, increased levels of logistics services have a significantly stronger total effect since they affect both the adaptiveness and the market performance of the firm, which in turn both considerably influence the financial performance. This findings again find the empirical support of ENGELBRECHT (2003, pp. 62-62). Logistics Performance

Firm Performance R2: 28.7% Adaptiveness 0.54***

Level of Logistics Services

0.47***

R2: 66.2%

0.47***

Market Performance Level of Logistics Costs

0.50***

R2: 40.6%

0.24***

0.65*** * ** ***

Standardized regression weight with significance level Significant on 10%-level Significant on 5%-level Significant on 1%-level

Financial Performance

Fig. 2-3. Performance effects of logistics (DEHLER 2001, pp. 233-244)

The positive effect of logistics performance on firm performance will in detail be analyzed in chapter 4.4, where also the remaining open questions concerning the influence of both the levels of logistics costs and logistics services on the three dimensions of firm performance will be addressed. Even though they can by no means be considered exhaustive and need some further research, the findings presented above provide insights into the answer to the question whether it is desirable for every individual firm to aim at reaching as high a level of logistics development as possible: even though it may be possible that in individual cases it is not efficient to allocate extensive management capacities to creating flow orientation throughout the firm, flow orientation has in general be shown to positively influence logistics performance. Together with the finding that logistics

2.2 Logistics outsourcing

19

performance is a significant driver of firm performance, the importance of flow orientation as a facilitator of logistics performance is underscored. Consequently, in general firms should aim at reaching as high a level of logistics performance as possible. This points the specific strategic direction for corporate logistics: away from functional oriented optimizations of isolated processes towards a concentration on the entire supply chain and its corresponding flows of material and information.

2.2 Logistics outsourcing After having introduced the different phases of logistics development, the question arises how to organize logistics processes on the level of the individual firm. The options for the firms are to either operate them by themselves or to partially or completely outsource them to a third party in the form of a logistics service provider (LSP), which will be introduced in detail in chapter 2.2.3.2. The following chapter will first highlight the origin of logistics outsourcing and provide a definition, before looking into its benefits and risks. Then, the current scope of logistics outsourcing in practice will be presented before the different kinds of logistics service providers available for outsourcing arrangements are introduced. Finally, an extensive literature review will provide the basis for the identification of research needs which will be addressed in this work. 2.2.1 Origin and definition As chapter 2.1.3 has pointed out, logistics capabilities are an important source of competitive advantage. As described before, the configuration of the individual logistics processes depends largely on the current phase of logistical development. At the same time, the question arises which parties are involved in the formation and realization of the processes. When approaching the concept of logistics outsourcing, RAZZAQUE/ SHENG (1998, p. 89) offer some valuable insights. According to them, a company can basically choose between three different options to handle its logistics activities effectively and efficiently: 1. It can provide the function in-house by making the service 2. It can either set up an own logistics subsidiary or buy a logistics firm 3. It can outsource the service and then buy the service from an external provider.

20

2 Basic concepts

The issue of outsourcing logistics services has received widespread attention over the last 15 years (BOWERSOX/DAUGHERTY/DROGE/ROGERS/ WARDLOW 1989; SHEFFI 1990; BARDI/TRACEY 1991; COOPER 1993; VIRUM 1993; MURPHY/DALEY 1994; RAO/YOUNG 1994; BAGCHI/VIRUM 1996; LIEB/RANDALL 1996; TATE 1996; SINK/LANGLEY 1997; RAZZAQUE/SHENG 1998). In the early discussion, different views of the meaning of logistics outsourcing became apparent. LIEB/MILLEN/VAN WASSENHOVE (1993) suggested that outsourcing, third-party logistics and contract logistics generally mean the same thing. BRADLEY (1994a) pointed out that service providers must offer at least two services that are bundled and combined, with a single point of accountability using distinct information systems which are dedicated to and integral to the logistics process. This is contrary to the view of LIEB/MILLEN/VAN WASSENHOVE (1993, p. 35) who note that outsourcing “may be narrow in scope” and can also be limited to only one type of service such as warehousing. After the initial dissension on the scope required to justify the use of the term “logistics outsourcing” more general definitions have been accepted. LAMBERT/EMMELHAINZ/GARDNER (1999, p. 165) state that logistics outsourcing is “the use of a third-party provider for all or part of an organization’s logistics operations” and add that its utilization by the firms is increasing. RABINOVICH/WINDLE/DRESNER/CORSI (1999, p. 353) define logistics outsourcing relationships even more broadly as “long and shortterm contracts or alliances between manufacturing and service firms and third party logistics providers”. For this work, logistics outsourcing will be understood in line with the definition provided by LAMBERT/EMMELHAINZ/GARDNER (1999), while the focus will be on the contract logistics described by RABINOVICH/WINDLE/DRESNER/CORSI (1999). The outsourcing trend has been continuously growing over the last years. It has been following the changes that have also been inducing the four phases of logistical development as presented in chapter 2.2.1. According to different authors such as TRUNICK (1992), SHEFFI (1990), FOSTER/MULLER (1990), BYRNE (1993) and RAO/YOUNG/NOVICK (1993) another important driving force behind this has been the increasing globalization of business. The continuously growing global markets and the accompanying sourcing of parts and materials from other countries has increased the demands on the logistics function (BOVET 1991; COOPER 1993; FAWCETT/BIROU/COFIELD TAYLOR 1993; MCCABE (1990); WHYBARK 1990) and led to more complex supply chains (BRADLEY 1994, p. 49). The lack of specific knowledge and suitable infrastructure in the targeted markets forced firms to turn to the competence of logistics service providers. In recent years, the outsourcing trend has gained even more momentum as the consensus in firms formed that the utilization of a logis-

2.2 Logistics outsourcing

21

tics service provider generally can reduce the cost of logistics processes and increase their quality (ELLRAM/COOPER 1990, pp. 2-9; LAMBERT/ EMMELHAINZ/GARDNER 1996, pp. 2-5; DEEPEN 2003, p. 121). Logistics service providers (LSP) suitable for providing these services today exist in abundance, reacting to the ever increasing demands of the customers and the subsequently developing markets. Due to the fact that a number of firms do not view logistics as a core competency or even if they do, are willing to outsource them to a third party (DEEPEN 2003, pp. 140141), outsourcing has become a relevant option. However, since the needs differ in every individual case, WALLENBURG (2004, p. 46) argues that every firm must answer two important question before actually outsourcing: 1. Which part of logistics shall be outsourced? 2. Who shall provide the service? The first question is very fundamental in nature and has a strategic character, while the second primarily calls for the identification of the adequate requirements in order to select the appropriate LSP. Due to the importance of the first question, it will be analyzed in greater detail in chapter 2.2.2 which addresses the advantages and disadvantages of logistics outsourcing. To allow insights towards the second question on who shall provide the service, chapter 2.2.3 will first offer an overview of the current scope and development of logistics outsourcing before introducing and discussing the relevant types of logistics service providers (LSPs). Finally, chapter 2.2.4 will conclude the remarks on logistics outsourcing with an overview of the current status quo of logistics outsourcing research. 2.2.2 Benefits and risks of logistics outsourcing Essential for answering the question regarding the optimal outsourcing scope are the resources of the respective firm and alongside the trade-off between consequential advantages and disadvantages. This will vary according to the individual firms’ perception of the benefits and risks associated with the particular outsourcing arrangement. Although they are inherently different, some aspects commonly associated with logistics outsourcing shall be presented in the following chapters. 2.2.2.1 Advantages of logistics outsourcing

The most frequently mentioned benefit of outsourcing is the reduction of the firm’s logistics costs (CAVINATO 1989, p. 14; BARDI/TRACEY 1991,

22

2 Basic concepts

pp. 15-21; LALONDE/MALTZ 1992, p.3; LIEB 1992, p. 37; BROWNE/ ALLEN 2001, p. 259). This can become manifest in several different ways: LIEB (1992, p. 29) and BRADLEY (1994a) point out that logistics service providers can be more efficient than a manufacturer, because logistics is their core business. Hence, specialization effects and the proper utilization of core competencies lead to lower production costs (BRETZKE 1993b, p. 38). Furthermore, inefficiencies which have not become apparent as long as the service was produced in-house and therefore was not subject to competition are eliminated (WALLENBURG 2004, p. 47). Lower production costs can also be achieved through economies of scale and scope resulting from the larger volumes of similar or equal logistics services a LSP produces (SCHÄFER-KUNZ/TEWALD 1998, pp. 30-36) and through the higher utilization ratio of the assets employed (BRETZKE 1993a, p. 10). Furthermore, logistics service providers can balance varying demand patterns better than a single manufacturing firm by diversifying their customer portfolios (DEEPEN 2003, p. 130) and reduce labor costs by benefiting from lower wage levels compared to those in manufacturing industries (STÖLZLE 1996, p. 124). Logistics outsourcing also directly affects the cost position of a firm due to a reduced need for capital investments. FOSTER/MULLER (1990, pp. 3032), RICHARDSON (1992, p. 22) and RICHARDSON (1995, p. 61) point out that investments in facilities can be reduced while SHEFFI (1990, pp. 2739), LACITY/WILLCOCKS/FEENY (1995, pp. 84-90) and RICHARDSON (1995, p. 61) state that costly information technology expenditures can be saved when outsourced to a logistics service provider. Beyond that, logistics outsourcing also allows for a decrease of the workforce and the associated investments (FOSTER/MULLER 1990, pp. 30-32; RICHARDSON 1992; RICHARDSON 1995), which proves valuable especially in countries with tightly regulated and inflexible labor markets such as Germany and France (SCHERER 2004). The effects mentioned above stemming from the reduction of capital investments ideally allow a firm to source only the required logistics services and to thus convert the formerly fixed costs of the logistics capacities into variable costs (KUMMER 1993, p. 29; RICHARDSON 1993). Besides all theses different potentials of cost reduction, however, logistics outsourcing has some further benefits for the firm. As HESKETT (1977, p. 85) states with respect to the firms’ overall logistics processes: “management cannot measure the importance of logistics in terms of costs alone”. Especially in recent years the realization has spread among firms that outsourcing logistics can also lead to improvements in logistics performance that in-house could not be achieved. Among these improvements are the following:

2.2 Logistics outsourcing

23

As a result of outsourcing, the expertise, technology, and infrastructure of the LSP can be utilized (BROWNE/ALLEN 2001, pp. 259-260). This can lead to a higher logistics performance in multiple dimensions. LALONDE/MALTZ (1992, p. 3) identify higher quality, better service, optimized asset use, and increased flexibility. Multiple authors go into further detail, such as RICHARDSON (1995, p. 61) who mentions faster transit times, less damage, and improved on-time delivery. The increased flexibility is a major benefit for firms. It allows firms to become more responsive as the needs of the market or customers change, as the LSP contributes by supplying its know-how and existing resources (BROWNE/ALLEN 2001, pp. 259-260; WALLENBURG 2004, p. 47). At the same time, the firm is enabled to concentrate on its own core business and its core competencies. This is particularly significant with respect to the core competence debate suggesting that due to limited internal resources and a growing complexity of the market competitive advantage cannot be attained in all areas simultaneously and focusing is necessary (PENROSE 1959; WERNERFELT 1984; PRAHALAD/HAMEL 1990; AMIT/ SCHOEMAKER 1993). Outsourcing logistics to a service provider allows for this concentration on core competencies, reduces the complexity of the firms’ business processes (KUMMER 1993, p. 29), and consequently facilitates sustainable competitive advantage. Furthermore, outsourcing reduces both the strategic and the operative risk of the firm. The strategic risk in the form of investment decisions in assets is outsourced, as well as operative risks, e.g. missed deadlines, unexpectedly surging costs or quality problems in the logistics processes, which all now have to be borne by the LSP. Another factor whose importance varies according to the corporate context and the business environment is mentioned by LYNCH (2000b, pp. 911), who points out that labor considerations must not be neglected when making the outsourcing decision. Problems with the workforce, originating e.g. from a high rate of unionization (USA) or particular labor agreements concerning wages (Germany) can be passed on to the LSP. 2.2.2.2 Disadvantages of logistics outsourcing

After the initial outsourcing debate had a rather euphoric notion, realization came over the years that outsourcing is accompanied by some disadvantages and risks (LYNCH/IMADA/BOOKBINDER 1994, p. 103; MCIVOR 2000, pp. 22-23; WENTWORTH 2003, pp. 57-58). One of the most commonly cited risks is the loss of control (BARDI/TRACEY 1991; LYNCH/IMADA/BOOKBINDER 1994, p. 103) and WENTWORTH 2003, pp. 57-58), paired with the dependence on an LSP of-

24

2 Basic concepts

ten accompanying the relationship. The firm must rely on the LSP to fulfill the service as agreed upon in the contract, but then depends on the LSP as the very source for the data it needs for judging whether the levels of quality and service have been achieved or not (WENTWORTH 2003, pp. 57). The same holds true for the LSP’s truthful declaration of the costs incurred when rendering the logistics service, which frequently is the base for the price charged to the firm. This effect is aggravated in the case that a firm outsources the entire logistics function, thereby losing its internal logistics skills and hence its capabilities to judge the outsourcing performance. That can be the origin for opportunistic behavior on the side of the LSP. If the firm wants to limit the potential for opportunistic behavior, it must install control mechanisms. As argued also in chapter 3.2.1, these will produce transaction costs such as bargaining and control costs, which must be added to the overall cost when making the outsourcing decision. It has been pointed out in the previous chapter that outsourcing can reduce the complexity of business processes, enabling the firm to concentrate on its core business. It must be noted, however, that in the relationship with the LSP coordination efforts between the parties are necessary, adding some other form of complexity (WALLENBURG 2004, p. 48), which, depending on the context of the relationship, could turn into a serious obstacle en route to successful outsourcing. The importance of considering the context of the firm and its implications will be further discussed in chapter 3.2.4. Other authors point to the complexity of outsourcing projects as one immanent and significant disadvantage. According to MCIVOR (2000, pp. 24-26), the strategic dimension of outsourcing projects is often neglected, leading to sub-optimal results based on the short term reasons of cost reduction and capacity issues. He concludes that problems frequently occur because complex issues, such as a formal outsourcing process, an adequate cost analysis and a thorough definition of the own core business have not been paid sufficient attention. 2.2.3 Markets for and providers of logistics outsourcing The previous two chapters have introduced some of the complexity which characterizes the recent logistics outsourcing debate. It must be observed that the outlined advantages and disadvantages alone do not allow a general statement that logistics outsourcing is always beneficial for the firm – in fact, a detailed analysis is needed for every single outsourcing decision. Logistics outsourcing, however, today is an important phenomenon within the business world. Following WALLENBURG (2004, p. 48), the ef-

2.2 Logistics outsourcing

25

ficiency hypothesis can therefore be assumed, according to which the sheer and lasting existence of the phenomenon suggests that it is beneficial for the firms, who, if not beneficial, simply would no longer outsource. In order to grasp the magnitude of logistics outsourcing in practice, the following two chapters will provide a short overview of the growing market and its future prospects as well as an introduction of the different kinds of logistics service providers available for logistics outsourcing. 2.2.3.1 Current scope and development of logistics outsourcing

Logistics services carry significant economic importance due to their immense volume. KLAUS (2003, pp. 59-69) estimates the European logistics market with all activities, ranging from simple mail logistics over complex contract logistics to bulk logistics and land-, air-, and sea transport, for the years 2001/2002 at 585 billion €. Accounting for the biggest share is the German market with a volume of an estimated 150 billion €. These absolute numbers are difficult to verify and other studies operate with more conservative numbers. DATAMONITOR (2000, p. 14) include in-house as well as outsourced logistics and estimate the German market for logistics in 1999 at 36 billion €, while the German Society for Transportation Economics and Logistics (GVB 2000, pp. 32-35) is more in line with KLAUS and determines the size of the market at 121 billion € in 1999. The tendency towards outsourcing thus is very strong and still growing. While KLAUS (2003, p. 59) estimates the average degree of outsourcing to be 45%, other studies provide a broader range. BAUMGARTEN/THOMS (2002, p. 15) find that the outsourcing degree in the three German key industries automotive, consumer goods, and retail ranges between 40% and 50% with the consumer goods industry leading the way. This is an considerable increase from a previous study conducted two years earlier by BAUMGARTEN/WALTER (2000, p. 13) who found that the average over all industries to be 28%. This trend will continue for the coming years. ENGELBRECHT (2003, p. 59) finds in his study that 60% of German logistics executives want to expand their outsourcing activities. This is consistent to the findings of LANGLEY/ALLEN/DALE (2004, p. 9) who state that current logistics expenditures directed towards outsourcing in Western Europe are 61% in 2004 and will grow to 68% in the period between 2007 and 2009. The analog numbers for the important logistics market in the U.S. project an increase from 44% in 2004 to 49% between 2007 and 2009. These number are further confirmed by the 2005 survey of LANGLEY/DORT/ANG/SYKES (2005, p. 12) which uses a different sample of firms. They find that from total logistics expenditures in Western Europe, 57% were directed towards

26

2 Basic concepts

outsourcing in 2005 and will further grow to 67% in the near future (20082010). Again, the numbers for the U.S. are slightly lower at 44% in 2005 and 51% for 2008-2010. The previously cited numbers point to an important trend. According to LANGLEY/ALLEN/DALE (2004), the projected increase in logistics outsourcing has been slowing down over the last years. While in 2002 it was 40% per year, it decreased to 14% in 2003 and in 2004 was at just over 11%. This could either mean that firms take a more cautious approach towards involvement with logistics service providers in general or that the growth rate for the extent of use of logistics services provided by third parties is stabilizing (LANGLEY/ALLEN/DALE 2004, p. 9). When looking at what logistics services firms outsource, it becomes evident that the activities most frequently outsourced are operational in nature, such as warehousing, transportation, handling, or customs clearance, while activities more strategic in nature or activities that are directly customer-related have a lesser propensity to be outsourced (LANGLEY/ALLEN/ DALE 2004, pp. 9-10). Recent developments suggest that this is slowly changing. According to BAUMGARTEN/WALTER (2000, pp. 22-29), an increasing outsourcing of more complex activities can be observed, especially in previously neglected areas close to production and with a distinct emphasis on the already outsourcing dominated procurement and distribution activities. The estimates provided above demonstrate the significance of the market for third party logistics services. Since the market has seen strong development over the last two decades, the potential for further increase is mainly limited to areas which so far have not been focused, such as activities close to production processes and complex contract logistics services. It can be expected that the market is going to eventually stabilize on a high level as recent research from LANGLEY/ALLEN/DALE (2004, p. 9) indicates. This might be due to the fact that the overall market for logistics, in-house as well as outsourced, is growing at a fairly slow rate, in line with the respective domestic GDP, while the outsourcing trend has been expanding rapidly. Since complete outsourcing of all logistics tasks does not make sense in most cases (WALLENBURG 2004, p. 49), the degree of outsourcing can be expected to ultimately level off. This will mean that the strong competition among logistics service providers, which could already be observed in times of strong market growth, will even intensify over the coming years. As this and the preceding chapters have shown, logistics service providers play a decisive role logistics outsourcing context, since they ultimately enable logistics outsourcing. Therefore, the different kinds of LSPs will be introduced and discussed in detail in the following chapter.

2.2 Logistics outsourcing

27

2.2.3.2 The role of logistics service providers in logistics outsourcing

As mentioned above, logistics outsourcing is enabled by a broad range of logistics service providers. In line with BHATNAGAR/SOHAL/MILLEN (1999, p. 570), in this work the term logistics service provider (LSP) will refer to an outside provider employed by firms to perform some or all of its logistics activities. LSPs differ considerably in the scope and depths of services offered. Due to their special relevance for the topic, the most important forms of LSPs will be presented in this chapter. The market for logistics services can be segmented along the lines of the four phases of logistics development introduced in chapter 2.2.1. The services offered by the LSPs are adapted to the needs of the respective customers. They range from a narrow spectrum, mainly consisting of warehousing and transportation services, for customers of the primal phases to integrated service portfolios including a multitude of different services for the advanced phases. Altogether, five kinds of LSPs can be distinguished: carriers, freight forwarders, courier & express & parcel/postal providers (CEP), third party / contract LSPs (3PLs) and fourth party LSP (4PLs). According to SCHMITT (2006, pp. 33-38), the above mentioned logistics service providers can be hierarchically classified depending on their service portfolio. Carriers typically own logistics assets and concentrate mainly on supplying transportation services. They are mostly confined to either road, sea, air or rail transportation and only in few cases also offer combinations of these services. They receive their orders either directly from the customer or through a freight forwarder and with their service portfolio cater to the needs of traditional logistics of the first phase of logistics development. With increasing sophistication of logistics processes, freight forwarders address the growing needs of the customers by offering coordinating functions and intermediating services. They bundle transportations services, offer warehousing and in increasingly also supply a combination of the two. While the focus of the freight forwarders’ services is still on providing physical processes, they also carry out additional services such as transportation planning and management including providing the associated information systems and also sometimes act as carriers by using own asset for transportation or warehousing. Overall, freight forwarders in their coordinating function address the needs of firms which are located in the second phase of logistics. The third phase of logistics development requires logistics to enable inter-organizational flow and process orientation and therefore demands comprehensive logistics solutions. During this phase, solution providers in the form of CEP and 3PLs depending on the needs of the customers be-

28

2 Basic concepts

come increasingly important. For customers aiming at end customer distribution services, the CEP offer integrated services that ensure the distribution of small units to any destination, often with time-critical shipments. 3PLs or contract LSPs focus on business customers and provide service packages that carried out on a longer term contractual basis. The solutions commonly include several services, such as warehousing, pick&pack or order handling. Increasingly, 3PLs also provide more customized services integrating into the customers value chain, such as fleet management, order handling, complaints management, or assembly services. For parts of the services offered that the 3PL could not provide alone, due to a lack of own assets, frequently carriers or freight forwarders are employed. The above mentioned logistics service providers, concentrating on complex, long term contract based logistics solutions, are in the focus of this work. Since the terminology varies considerably even beyond 3PL and contract LSP, they will in the following be called logistics service providers (LSPs) as both terms meet the requirements of the definition presented above provided by BHATNAGAR/SOHAL/MILLEN (1999, p. 570) following which an LSP is employed by “an outside company to perform some or all of the firm’s logistics activities”. Consistent with BERGLUND/VAN LAARHOVEN/SHARMAN/WANDEL (1999, p. 59), LSPs in the context of third party logistics offer “[…] activities […] consisting of at least management and execution of transportation and warehousing […]. In addition, other activities can be included […]. Also, […] the contract [is required] to contain some management, analytical or design activities, and the length of the cooperation between shipper and provider […] [must] be at least one year […]”. All relationships between LSPs and their customers analyzed in the latter must fall under this definition in order to adequately distinguish between advanced logistics outsourcing relationships and traditional “arm’s lengths” sourcing of transportation and/or warehousing. Offering services beyond those mentioned above are the 4PL-Providers. Even though this term is still utilized inconsistently, SCHMITT (2006, p. 36) argues that following a widespread common understanding, 4PL refers to a logistics service provider which serves as an intermediary and general contractor for inter-organizational supply chains without supplying any physical process by itself. Rather, it employs carriers, CEPs, or other LSPs for the physical processes and concentrates on planning, conceptualizing and managing the supply chain. It therefore is virtually operating without physical assets and therefore supposedly is neutral.

2.2 Logistics outsourcing

29

2.2.4 Status quo of logistics outsourcing research After the market of logistics services and the different kinds of logistics service providers have been introduced, in the following the status quo of logistics outsourcing research will be presented. As the previous chapters have shown, the scale and scope of logistics outsourcing has been steadily growing over the last decades, creating markets of significant sizes and consequently creating opportunities for firms which find themselves under ever increasing pressure to lower costs and increase logistics performance. An end of the trend is not in sight. Therefore, a high relevance of research in logistics outsourcing and its success factors can be assumed. It therefore does not come as a surprise that MENTZER/KAHN (1995, p. 242) find “that logistics research is a growing and viable research area” by investigating the development of articles published in the Journal of Business Logistics between its inauguration in 1978 and 1993. However, they formulate substantial criticism on the quality of the overall empirical logistics research conducted in these early years. According to MENTZER/KAHN (1995, pp. 240-244) more than 53% of the existing research is normative in nature and does not have empirical content. 36% of the research includes case studies only and a mere 4% involves hypothesis testing. According to MENTZER/KAHN (1995) this is urgently needed for subsequent theory development and testing and is the main driver of successful logistics research. They conclude that as of 1995, the few studies that have included hypothesis testing procedures have mostly not stood up to scientific rigor. Commonly, they fail to provide information on the reliability and validity of the data while at the same time the statistical methods being utilized are limited in most cases to regression or correlation analysis. According to ENGELBRECHT (2004, pp. 28-29), the situation has not changed significantly until 2002. Therefore, the argument of MENTZER/KAHN (1995, p. 244) still prevails which demands a more rigorous and scientific approach including more hypotheses: “If the discipline is to become more theoretically rigorous, it must progress through the framework and thus pursue more hypothesis testing studies. While many researchers would prefer to pursue only exploratory research because of the flexibility in topic selection, ease of data analysis, and less meticulous rigor, a maturing scientific discipline mandates a shift toward greater hypothesis testing, more rigorous date analysis, and standard discussions of validity and reliability”. According to WALLENBURG (2004, p. 52), logistics research has always shown a strong practice-orientation. In earlier years the focus was on functional specialization, targeting almost exclusively individual competencies

30

2 Basic concepts

and local optimization potentials. Most commonly, the research reflected the view and interests of a single firm only. In recent years, the focus has shifted towards the entire supply chain and factors of collaboration that enable the optimization of entire process-chains rather than isolated processes. Overall, logistics research has traditionally been confronted with numerous operative and practical problems (WALLENBURG 2004, p. 52), requiring quick and efficient solutions. By adhering to this need, it remained rather conceptual and – in the few empirical works – mostly qualitative. Thus, it does not come as a surprise that logistics research has also recently been criticized for its lack of quantitative empirical research (GARVER/ MENTZER 1999, pp. 33-34; CORSTEN 2003, pp. 49-51). While the shortcomings portrayed above are valid for the broad field of contemporary logistics research and the studies explicitly targeting logistics outsourcing are still scarce, a slow change towards more research in that area and a more rigorous scientific approach as demanded by MENTZER/KAHN can be observed. Since the research until 1999 has already been summarized in great detail by several articles (MURPHY/POIST 1998; RAZZAQUE/SHENG 1998; MURPHY/POIST 2000), the following section will concentrate on the evolving research published since then (Table 2-1). Table 2-1. Key logistics outsourcing related research since 1999 Author(s) BHATNAGAR/SOHAL/MILLEN (1999) BOYSON/CORSI/DRESNER/ RABINOVICH (1999)

Key logistics outsourcing related finding Cost savings, customer satisfaction, and flexibility are the main reasons for outsourcing Logistics outsourcing is effective in enabling firms to achieve competitive advantage, improve their customer service levels, and reduce their overall logistics costs VAN LAARHOVEN/BERGSome outsourcing relationships are perceived LUND/PETERS (2000) by the shipper as more successful than others. Some success factors are: well defined requirements, procedures, systems, and close relationships BAUMGARTEN/WALTER (2000) Logistics outsourcing is a major trend driven by the firms’ desire to reduce costs and to increase flexibility and service levels BAUMGARTEN/THOMS (2002) The degree of logistics outsourcing in German firms is high and still increasing DEEPEN (2003) Reducing logistics costs and increasing logistics performance are main drivers of logistics outsourcing – contextual factors such as asset

2.2 Logistics outsourcing

31

specificity determine the respective advantages and disadvantages KNEMEYER/CORSI/MURPHY Exploratory findings show that developing (2003) closer relationships between shipper and 3PL create increased costs, but at the same time promise benefits LANGLEY/ALLEN/COLOMBO The marketplace for 3PL services continues to (2003) change – both users and providers are becoming more capable while the expectations of each other are rising. While more productive and meaningful 3PL customer-supplier relationships evolve, a gap exists between what customers receive and what they expect to receive STANK/GOLDSBY/VICKERY/ Relational performance by LSPs is the most SAVITSKIE (2003) important factor engendering customer satisfaction LANGLEY/ALLEN/DALE (2004) While 3PL users generally feel that their relationships with 3PL providers are successful, they are able to find areas of improvement, such as implementing capable IT, instituting effective management and relationship processes and integrating services and technologies globally. Customer demands for performance and sophistication are accelerating KNEMEYER/MURPHY (2004) Various relationship marketing dimensions (trust, communication, opportunistic behavior etc.) influence the buyers perception of 3PL performance ENGELBRECHT (2004) Outsourcing performance is explained by the degree of outsourcing only to a limited extent. The implementation of the outsourcing project has a significantly higher explanatory value STRAUBE/PFOHL/GÜNTHER/ Logistics providers must cope with growing DANGELMAIER (2005) uncertainty, increasing cost pressure, and more complex value chains and at the same time fulfill increasingly changing customer demands KNEMEYER/MURPHY (2005) Relationship characteristics such as communication with the 3PL provider have a more profound impact on logistics outsourcing relationship outcomes than customer attributes such as firm size, number of functions outsourced, and the number of 3PL relationships LANGLEY/DORT/ANG/SYKES 3PL users continue to view a collaborative (2005) partnership approach with their 3PL providers as key to improving the user-company 3PL-

32

2 Basic concepts performance. Different to past surveys, pricing has become the most important attribute in selecting a 3PL provider

The studies between 1999 and parts of 2003 (BHATNAGAR/SOHAL/ MILLEN 1999; BOYSON/CORSI/DRESNER/RABINOVICH 1999; VAN LAARHOVEN/BERGLUND/PETERS 2000; BAUMGARTEN/WALTER 2000; BAUMGARTEN/THOMS 2002; DEEPEN 2003) offer very valuable insights into logistics outsourcing, but must in large parts be subject to the same criticism as voiced by MENTZER/KAHN (1995, pp. 240-244) . They are largely descriptive in nature, the only exception being BOYSON/CORSI/DRESNER/ RABINOVICH (1999) who also perform an ANOVA analysis for selected questions. None of the studies involve hypothesis testing procedures. Findings include that the understanding of the meaning of logistics outsourcing for achieving competitive advantage is increasing, that cost reductions and logistics performance increases are the most significant motivators for the outsourcing decision, and the considerable importance of the relationship formation for the project’s success. The most serious shortcoming of the research until 2003 is the lacking proof for the performance effects of logistics outsourcing. The descriptive and yet in parts normative studies presented above have failed to show with the adequate scientific rigor that logistics outsourcing does increase logistics performance and at the same time have not been able to identify the success factors for logistics outsourcing with advanced statistical methods. This is starting to change in 2003 with the works of KNEMEYER/CORSI/ MURPHY (2003), STANK/GOLDSBY/VICKERY/SAVITSKIE (2003), KNEMEYER/MURPHY (2004) and ENGELBRECHT (2004). ENGELBRECHT (2004, pp. 244-250) shows in a partial model, using structural equation modelling (SEM) to test his hypotheses, that the degree of outsourcing can explain 8% of the logistics cost position of a firm. The hypothesized direct effect of the degree of outsourcing on the level of logistics services turns out to be non-significant. These findings indicate for the first time on a high statistical level that the descriptive studies of the past, which have normatively assumed the performance effect of logistics outsourcing, were right in their assumption, even though their research procedure may have been lacking the adequate scientific rigor. Nevertheless, the explanatory value of the degree of outsourcing is quite low. It must therefore be assumed, that the true drivers behind logistics outsourcing performance as yet remain to be identified. An important contribution to this discussion is made by KNEMEYER/CORSI/MURPHY (2003). They employ SEM to show that the

2.3 Logistics outsourcing relationships

33

benefits of developing closer relationships between customer and the LSP can justify the accompanying increasing costs. This is of particular interest since STANK/GOLDSBY/VICKERY/SAVITSKIE (2003, pp. 41-45) show that relational performance of the LSP is the single most important factor in obtaining customer satisfaction, which in turn can be understood as an expression of the achievement of the goals previously set for the outsourcing project. In a recent study, KNEMEYER/MURPHY (2004, pp. 45-46) furthermore demonstrate by using SEM that various relationship marketing dimensions, such as trust, communication and opportunistic behavior, influence the buyers perception of the logistics service providers’ performance and thus are relevant factors influencing the outsourcing performance. The findings presented above suggest that the degree of outsourcing alone cannot explain the performance effects of logistics outsourcing. The main driver must lie somewhere else. Recent studies have proposed that the formation of the relationship between the buyer and the LSP is one of the main drivers (STANK/GOLDSBY/VICKERY/SAVITSKIE 2003; KNEMEYER/MURPHY 2004; LANGLEY/DORT/ANG/SYKES 2005; KNEMEYER/ MURPHY 2005). However, as yet no model has been developed that would encompass all relevant dimensions and factors of relationships. Furthermore, it remains to be shown which performance effects the individual constructs of the model would have and which dependencies inside the model would exist. All these different findings are needed to further advance research in this field and to develop feasible recommendations for everyday management. A last open question of utmost importance is the role of the context of the firm in the logistics outsourcing context (CHOW/HEAVER/HENRIKSSON 1994; PFOHL/ZÖLLNER 1997). As already stated in chapter 2.2.3, it is virtually impossible to make a generalized and universally valid recommendation for the outsourcing decision. In fact, it is crucial to first investigate the context of the firm in order to determine what is outsourced to whom. To determine what outsourcing strategy has primacy in which context, CHOW/HEAVER/HENRIKSSON (1994, p. 26) suggest the use of contingency models of logistics performance which should include factors such as the environment, the product line or the production technology.

2.3 Logistics outsourcing relationships Chapter 2.2.4 has shown that the performance of logistics outsourcing projects cannot be explained by the extent of outsourced services only. Other performance drivers, such as the right choice of the outsourced services,

34

2 Basic concepts

the implementation process or the design of the outsourcing relationship between the buyer and the LSP, have also been identified as relevant. LAMBERT/EMMELHAINZ/GARDNER (1999, pp.165-166) state that in outsourcing arrangements, firms enter into long-term relationships with LSPs, often also called partnerships. The hope is that by joining forces both firms will improve efficiency, profitability and customer service. These relationships can, if they turn out successful, give both parties a competitive advantage in the marketplace (TATE 1996, pp. 7-13). According to LAMBERT/EMMELHAINZ/GARDNER (1999, pp. 165-166), not all of these so-called partnerships are successful. They fail for various reasons, such as unrealistic expectations relating to the structure of the outcomes of the relationship. While several models exist that support management during the decision for or against outsourcing and sometimes even also specify criteria for the selection of the LSP, as yet no model has been proposed that can provide a systematic method for identifying the most appropriate type of outsourcing relationship. However, according to LAMBERT/EMMELHAINZ/ GARDNER (1999, p. 166), after the decision to outsource has been made, the key issue remaining is how the relationship should be designed. In order to address the complex issue of logistics outsourcing relationships, the next chapters will sequentially present the most important aspects, thereby following the structure outlined below: At first, chapter 2.3.1 will introduce the terminology of partnerships before chapter 2.3.2 will analyze partnership development. To do so, chapter 2.3.2.1 will reflect on the importance and the effects of partnership development, before chapter 2.3.2.2 will introduce the current research on the process of partnership development. After thus having established the need for sustainable relationships also in the logistics outsourcing context, chapter 2.3.2 will examine how these relationships between logistics service providers and their customers should be designed. This requires the presentation of the most relevant relationship marketing models in chapter 2.3.3.1, before finally in chapter 2.3.3.2 the aspects of relationship marketing in the current logistics outsourcing relationship research can be presented. Combined, the information of chapter 2.3 will thus enable the understanding of the status quo of logistics outsourcing relationship research and the identification of consequential research needs. 2.3.1 The terminology of partnerships To describe the phenomenon of inter-organizational relationship between a buyer and a supplier of a good or service, a multitude of different terms are being used. This can be viewed as a reflection of the diversity of inter-

2.3 Logistics outsourcing relationships

35

organizational relationships in practice (BALLING 1997, p. 12). Among those terms are competitive collaboration, coalition, working partnership, collaboration, partnership, and cooperation. To characterize the horizontal cooperation between competitors, HAMEL/DOZ/PRAHALAD (1989, p. 133) introduce the term “competitive collaboration”. Manifestations of this can be joint ventures, product licensing, cooperative research or outsourcing agreements. While this, according to HAMEL/DOZ/PRAHALAD (1989), is a strategy that can strengthen both parties against outsiders, it has triggered unease among some partners due to the long-term consequences which might leave one side weaker than the other. In an early publication, PORTER/FULLER (1986, p. 315) use the term “coalition” to describe long-term alliances between firms that cooperate in certain areas without actually merging the businesses. This can include joint ventures, licensing contracts, supply contracts, distribution contracts or various different agreements. An expression describing the vertical cooperation between two parties is introduced by ANDERSON/NARUS (1990, p. 42). They state that with the increase of long-term commitments between buyers and suppliers, “working partnerships” aiming at mutual success are developing. Those they define as “the extent to which there is mutual recognition and understanding that the success of each firm depends in part on the other firm, with each firm consequently taking actions so as to provide a coordinated effort focused on jointly satisfying the requirements of the customer marketplace”. Another term used frequently especially in the supply chain management context is collaboration. According to STANK/KELLER/DAUGHERTY (2001, p. 31) it is a process of decision making among interdependent parties, involving joint ownership of decisions and collective responsibility for the outcomes. SCHRAGE (1990) views it as an affective, volitional, and mutual shared process in which two or more departments work together. Furthermore, they have a common vision, exhibit mutual understanding, share resources and achieve collective goals. All terms presented above reflect the idea that cooperative actions are needed to achieve the desired goals which prompted the outsourcing decision and resulted in the relationship between the customer and the LSP. This thought is particularly dominant in the term “partnership” which is very present in the discussion of logistics relationships (STUART 1993; GARDNER/COOPER/NOORDEWIER 1994; KANTER 1994; ELLRAM/HENDRICK 1995; GENTRY 1996; TATE 1996; LAMBERT/EMMELHAINZ/ GARDNER 1999). Again, various definitions exist. LALONDE/COOPER (1989, p. 6) define a logistics partnership as “a relationship between two entities in the logistics channel that entails a sharing of benefits and bur-

36

2 Basic concepts

dens over some agreed upon time horizon”. Several other definitions include further key characteristics such as information sharing, shared risks and rewards, long-term focus, joint activities and the concept of trust (DWYER/SCHURR/OH 1987; NOORDEWIER/JOHN/NEVIN 1990; GATTORNA 1991; ELLRAM 1995a). According to LAMBERT/EMMELHAINZ/GARDNER (1999, p. 166), all these definitions are incomplete. They address only some aspects of the partnership and do not adequately emphasize the need for customization of the relationship. LAMBERT/EMMELHAINZ/GARDNER (1996b, p. 28) therefore introduce an own definition which views a partnership as “a tailored business relationship based upon mutual trust, openness, shared risk, and shared rewards that yields a competitive advantage, resulting in business performance greater than would be achieved by the firms individually”. The above definition will be adopted as it is very comprehensive, addressing both behavioral and monetary aspects, and thus is displaying exactly the understanding of partnerships in this study. This is especially important with respect to the later development of a logistics outsourcing relationship model in chapter 4, which will contain a number of characteristics that originate from this understanding of partnerships. The following chapters will in detail analyze the development of such partnerships. Special attention will first be given to the importance and the effects of partnership development, before in a second step the status quo of research on the process of partnership development will be examined. 2.3.2 Partnership development 2.3.2.1 Importance and effects of partnership development

According to ELLRAM (1992), relationships in a supply chain can be simple if they involve the purchase of commodities or complex if they involve specialty products obtainable only from a limited number of suppliers or if they require specialized assets to produce. The same is true for relationships between a customer and a LSP, whose product range can span from commodity to highly customized services. Before the understanding for the potential of supply chains spread quite recently, the predominant form of business relationships between firms was mainly adversarial in nature with individual firms seeking to achieve cost reductions or profit improvements at the expense of their customers and/or suppliers (DUFFY/FEARNE 2004, p. 57). These forms of interorganizational relationships today are commonly referred to as “arm’s-length” relationships (BOYLE/DWYER/ROBICHEAUX/SIMPSON 1992, p 463; SHEM-

2.3 Logistics outsourcing relationships

37

WELL/CRONIN/BULLARD 1994, p. 57; DYER/CHO/CHU 1998, pp. 57-58; HOYT/HUQ 2000, p. 750; SKJOETT-LARSEN/THERNOE/ANDERSEN 2003, p. 531). Gradually, this perception has been changing. This was caused by the increasing comprehension that the majority of business relations are not characterized through discrete transactions anymore, but rather through relational exchange. According to DWYER/SCHURR/OH (1987, p.12) the idea of a discrete transaction is the foundation on which the concepts of relationships are built. Following MACNEIL (1980, p. 60), the archetype of a discrete transaction is represented by an exchange were money is on the one side and an easily measured commodity is on the other. He states: „Discreteness is the separating of a transaction from all else between the participants at the same time and before and after. Its ideal, never achieved in life, occurs when there is nothing in between the parties, never has been, and never will be”. The concept described in this definition excludes all relational elements and therefore will occur only very rarely in business relations. Discrete transactions are an important prerequisite for arm’s length relationships. Only if all “going concern” is excluded from the relationship, can individual firms perfectly maximize their own profits at the expense of customers and/or suppliers without future consequences in mind. As soon as repeating contacts between firms are occurring, some form of relationship between them will emerge and pure arm’s length relationships based on individually benefiting from adversarial behavior towards the other party will no longer be optimal. With the increasing popularity of the supply chain management concept, the former arm’s length relationships gradually take on characteristics of relational exchange relationships (HOYT/HUQ 2000, p. 750). These are characterized by the fact that they are viewed by the partners both in terms of their history and their anticipated future (MACNEIL 1978; MACNEIL 1980). Furthermore, they involve implicit and explicit assumptions, trust and planning (DWYER/SCHURR/OH 1987, p. 12) as well as relational factors such as collaboration and information sharing (HOYT/HUQ 2000, p. 750). According to DWYER/SCHURR/OH (1987, p. 12), relational exchange participants can be expected to derive complex, personal, non-economic satisfaction and engage in social exchange. These relational exchanges constitute the basis for partnerships in the logistics context as described in chapter 2.3.3. A major driver for the increase of relational exchanges and more partnership were the benefits expected by the customers. Through their working together with LSPs, they could jointly strive for individual or mutual goals with the implicit or explicit desire to achieve benefits that would not

38

2 Basic concepts

be obtainable without the two parties working together (SCHERMERHORN 1975, p. 847; STERN/REVE 1980, p. 57; BROWN 1981, p. 7). This will be called in the following “cooperative behavior”. In the logistics context, its benefits include profit enhancement, process improvements, or increased competitive advantage (LAMBERT/EMMELHAINZ/ GARDNER 1999, p. 170). Opposing theses benefits are the dangers and pitfalls of partnering which according to LAMBERT/EMMELHAINZ/GARDNER (1999, p. 166) have received far too little attention. LIEB/RANDALL (1996, p. 311) show in multiple surveys that the three most common concerns of customers for entering logistics partnerships are the potential loss of direct control of logistics activities, uncertainty about the service levels to be provided by the outside company and questions of the true costs of using a LSP. ACKERMAN (1996, p.35) identifies several reasons why logistics partnerships may fail. These include the buyer and the seller not having a realistic understanding about the job to be done, over-promising of the seller or the inability to deliver on the promise, deliberate attempts of the customer’s management to make the partnership fail or unprofitability for the seller, and subsequent service failures. Further factors leading to partnership failure are pointed out by ELLRAM (1995b, pp. 41-42). In a survey of 80 pairs of buyers and suppliers who mutually agreed that they are strategic partners she found that the most prominent factors include poor communication, lack of top management support for the partnership, lack of trust, lack of total quality commitment by the supplier, poor upfront planning or the lack of shared goals, and a missing strategic direction of the partnership. For the most part, these causes for partnership failure fall into two categories suggested by STUART/MCCUTCHEON (1995, pp. 5-6). They identify perceptual mismatches over the appropriate degree of partnering and failures in the building process of the relationship as central factors of partnership failure. 2.3.2.2 Current research on the process of partnership development

After the preceding chapter has pointed out the importance of partnerships and their effects for the participating firms, in the following the status quo of research on the process of partnership development will be introduced. This process has been subject to substantial research. GARDNER/COOPER/ NOORDEWIER (1994, p. 137) developed a five stage strategic model of the partnership building process. It includes choosing a partnership strategy in the first step, then choosing one or more partners, designing the partnership in the third and evaluating it in the fourth step and finally evaluating the partnership strategy. ELLRAM (1995a, p. 12) presents a different and yet quite similar five stage process for the development and evolution of

2.3 Logistics outsourcing relationships

39

purchasing partnerships. It includes a preliminary phase, the identification of potential partners, the screening and selecting of the former, the establishment of relationships in the fourth phase and finally the evaluation of the relationship. As specific selection criteria for potential partners she identifies among others cultural compatibility and long term stability of the partner, both in strategic and in financial terms. The main factors determining the degree of a partnership that should be aimed for by the parties involved are highlighted by STUART (1993, p. 27), who identifies the level of committed resources, the potential for productivity improvements and competitive advantage, the level of joint problem solving, and the sharing of benefits as important. A more complex model for logistics alliances is presented by BAGCHI/VIRUM (1996, pp. 98-106). It basically comprises three phases, consisting of a total of 22 individual steps. First, the need for an alliance is identified, then it is planed and managed, and in a third phase, operations must be measured and controlled. According to BAGCHI/VIRUM (1996, p. 95) “in a logistics alliance, the parties ideally consider each other as partners”. LAMBERT/EMMELHAINZ/GARDNER (1999, p. 167) acknowledge these models because they help managers to determine whether an outsourcing arrangement is needed and identify criteria for the supplier selection process. They criticize, however, that while these models are good tools for assessing the appropriateness of outsourcing they do not go far enough in addressing issues once the decision is made. Furthermore, they point out that the models offer no guidance about the type of the relationship to choose, for instance an arm’s length contract or a partnership and at the same time assume that all partnerships are the same. They continue by observing that while managers mostly decide correctly that outsourcing is needed, they often make incorrect decisions on the specific type of relationship needed for their purpose. This leads to failure of the partnerships not due to the inappropriateness of outsourcing, but due to the managers from both sides not agreeing on the type of arrangement to be used. Consequently, LAMBERT/EMMELHAINZ/GARDNER (1999, p. 167-170) develop an own partnership development and implementation model to address the shortcomings of the previous research. It includes three major elements in the form of drivers, facilitators, and management components. Drivers are compelling reasons to partner which consist of strategic benefits that result from strengthening a relationship. They might include asset or cost efficiencies, enhanced customer service, marketing advantage as well as profit growth and stability. For a partnership to succeed, it must have sufficient drivers which also do not need to be the same for each party.

40

2 Basic concepts

Facilitators provide a supportive environment for growth and maintenance of a relationship. The four primary facilitators in a relationship are corporate compatibility, similar managerial philosophy and techniques, mutuality, and symmetry. Facilitators reflect the degree of compatibility between the partners and thereby indicate the likelihood of partnership success. This according to LAMBERT/EMMELHAINZ/GARDNER (1999) follows the logic, that the more compatible the firms are, the better are the chances for partnership success. LAMBERT/EMMELHAINZ/GARDNER (1999, p. 170) state that while drivers and facilitators determine the most appropriate degree of integration, whether that integration is achieved depends on management components. These are joint activities and processes used to build and sustain a partnership. They include planning, joint operating controls, communications, risk and reward sharing, trust and commitment, contract style, scope, and financial investment. Once it is determined that a particular degree of partnership is wanted, the two parties should jointly plan how to implement it within each organization. The partnership is then tailored to the degree of integration by using varying levels of each of the management components. The model of LAMBERT/EMMELHAINZ/GARDNER (1999) is suitable to address the issue of partnership configuration once the outsourcing decision has been made. It supplies valuable insights on the design of individual relationships and thereby helps to answer the question how close a relationship should be in order to maximise the net benefits for both parties. Unfortunately, the causal linkages between the different management components and therefore of the factors influencing the performance of the partnership are not determined. Furthermore, the model is only normative in its conclusion that relationships designed to fit the context of the firms are more successful. Empirical evidence still needs to be produced. Therefore, a causal model would be desirable that takes up the ideas of LAMBERT/EMMELHAINZ/GARDNER (1999) and makes a deterministic proposition on the importance of the single factors and their true performance effects. This demand is consistent with that of KNEMEYER/MURPHY (2004, p. 38) who state that the potential causal linkages between relationship dimensions such as trust and communication and successful third party logistics arrangements have to be examined. As the following chapter will show, some researchers have taken up the idea behind the model of LAMBERT/GARDNER/EMMELHAINZ and developed first insights into the possible causal linkages between relationship dimensions that could influence the performance of logistics partnership arrangements.

2.3 Logistics outsourcing relationships

41

2.3.3 Designing logistics outsourcing relationships The previous chapters have indicated the importance of partnership development and its effects as well as introduced different models developed to support the process of partnership development. A further important part in developing successful partnerships is to understand how relationships between the different parties involved ought to be designed. According to SHETH/PARVATIYAR (1995, p. 399), some social psychologists suggest that competition is inherently destructive and mutual cooperation is inherently more productive. While few authors such as PACHÉ (1998, p. 308) doubt this, suggesting that the adversarial approach to logistics relationships may offer some advantages, because e.g. cooperation may slow down contractors’ improvements, a broad majority of scholars embraces the idea of the positive effects of cooperation on relationships. This thought is a dominant factor in relationship marketing which is a cornerstone of the further analysis. Research on relationship marketing was initiated and significantly advanced through representatives of the Nordic School of Services (GRÖNROOS/GUMMESSON 1985; GRÖNROOS 1989; GRÖNROOS 1990; GRÖNROOS 1994; GUMMESSON 1987; GUMMESSON 1997). According to MORGAN/HUNT (1994, p. 22), relationship marketing refers to “all marketing activities directed toward establishing, developing, and maintaining successful relational exchanges”. As CRAVENS (1995, pp. 48-57) points out, relational marketing as opposed to transactional marketing, focuses on long term associations of buyers and sellers. This is of particular importance, since KALWANI/NARAYANDAS (1995, p. 14) state that “supplier firms in long term relationships with select [sic!] customers are able to retain or even improve their profitability levels more than firms that employ a transactional approach to servicing customers”. This finding receives initial support in the logistics context through the research on logistics relationships by LAMBERT/EMMELHAINZ/ GARDNER (1999) and KNEMEYER/MURPHY (2004). Relationship marketing may prove to be a useful tool for understanding the importance of partnerships in logistics outsourcing with the accompanying performance effects on the one hand and the design of these relationships on the other. As STOCK (1997, p. 528) puts it, a potential application of relationship marketing variables such as commitment, trust, communication or cooperation could facilitate a better understanding of logistics relationships. The next chapter will therefore introduce selected relationship marketing models that offer insights into the adequate design of relationships. After that, chapter 2.3.3.2 will first discuss which aspects of relationship mar-

42

2 Basic concepts

keting so far have been addressed in the logistics outsourcing relationship research literature and then identify remaining research needs. 2.3.3.1 Relationship marketing models

Much has been written on the effect relationship marketing variables have on the design of relationships. Of particular importance are the empirical works that with adequate scientific rigor test hypotheses on the causal linkages that influence buyer-seller relationships in general and that of the customer and the LSP in particular, because they allow to gain deeper insights on how relationships between different parties should be designed. Through the early days of work in this field, this issue has primarily been addressed on the general level. Works on buyer-seller relationships include those of ANDERSON/NARUS (1984), ANDERSON/LODISH/WEITZ (1987), ANDERSON/WEITZ (1989), HEIDE/JOHN (1990), NOORDEWIER/ JOHN/NEVIN (1990), HALLÉN/JOHANSON/SEYED-MOHAMED (1991), ANDERSON/WEITZ (1992), HEIDE/JOHN (1992), HAN/WILSON/DANT (1993), GANESAN (1994), MORGAN/HUNT (1994), WEITZ/JAP (1995), and WILSON (1995). Some of the most notable contributions have come from ANDERSON/WEITZ (1989), ANDERSON/NARUS (1990), and MORGAN/ HUNT (1994). Not only have they been on the forefront of relationship marketing research, but they also introduced new and important views on the issue of how relationships ought to be designed. By doing so, they introduced and structured a number of relationship marketing variables that have later become important factors in logistics relationship research. ANDERSON/WEITZ (1989, p. 311) examine the determinants of continuity in conventional industrial channel dyads. They formulate a model in whose center are trust, the degree of communication, and power imbalance. Further constructs include goal congruence, cultural similarity, stakes, age of the relationship, and the perceived continuity of the relationship. The findings put forward by ANDERSON/WEITZ (1989, p. 319) show that trust has an import impact on the stability of dyads and is influenced among other factors by the degree of communication and the goal congruence of the two parties. Using a similar approach, ANDERSON/NARUS (1990, p. 44) develop a model of distributor firm and manufacturer firm working partnerships. They theorize that the communication between parties positively influences the level of trust, which in turn leads to increased cooperation. This cooperation, which ANDERSON/NARUS (1990, p. 45) define as “similar or complementary coordinated actions taken by firms in interdependent relationships to achieve mutual outcomes with expected reciprocation over

2.3 Logistics outsourcing relationships

43

time” in their model is one of the key antecedents of satisfaction. Other factors employed in the model include relative dependence, conflict, functionality of conflict and the influence both firms have on one another. After conducting dyadic survey research with a total of 504 firms and 1365 informants, ANDERSON/NARUS (1990, pp. 48-53) showed the effect of the constructs from both the perspectives of the distributing and the manufacturing firms. Among the most noteworthy results are that from the distributors point of view communication directly influences cooperation, which in turn has a positive effect on both the trust between the parties and the functionality of the conflicts. The relative dependence has only a secondary importance in the model, merely affecting the functionality of the conflict indirectly via the influence over the partner firm. Similar and yet more detailed results are obtained by ANDERSON/ NARUS (1990, p. 53) in their second model on working partnerships which analyzes the perspective of manufacturing firms. Here, they can show that communication positively influences cooperation and trust at the same time. The hypothesized causal linkage where increased trust leads to higher cooperation is reversed – empirically, increased cooperation leads to higher levels of trust. Trust in turn significantly decreases the levels of conflict in the relationship and increases the satisfaction, which decreases as conflicts increase. Again, relative dependence has a comparatively unimportant role in the model, affecting the levels of conflict indirectly via the influence by the partner firm. The most influential work of the past decade concerning relationship marketing is the article of MORGAN/HUNT (1994). They place relationship commitment and trust in the center of their model and explicitly exclude power which until then had seen strong utilization in relationship research. MORGAN/HUNT (1994, p. 22) theorize that “the presence of relationship commitment and trust is central to successful relationship marketing, not power and its ability to ‘condition others’”. This constitutes a paradigm shift which has since its publication strongly influenced relationship research. In their model, MORGAN/HUNT (1994) propose several precursors of relationship commitment and trust in the form of shared values, communication, opportunistic behavior, relationship benefits, and relationship termination costs. Commitment and trust in turn impact the five outcome dimensions acquiescence, propensity to leave, cooperation, functional conflict and uncertainty. Empirically, MORGAN/HUNT (1994, p. 30) can show that with few exceptions, the proposed causal linkages prevail. The most notable findings include that trust is influenced by shared values, communication, and opportunistic behavior while it in turn influences cooperation, functional con-

44

2 Basic concepts

flict, uncertainty, and relationship commitment. Commitment is further influenced by shared values, relationship benefits and relationship termination costs and leads to acquiescence, cooperation and reduces the propensity to leave. The work of MORGAN/HUNT (1994) was of particular importance to relationship marketing research, as it gave the discipline further structure as well as a new direction at the same time. So far, however, it must be noted that even though a process of harmonization concerning the factors utilized in relationship marketing research has set in, no common understanding concerning their exact selection exist as yet. As the three studies introduced in this chapter show, a multitude of relationship factors are used to analyze the issue. WILSON (1995, p. 337) presented an extended list of 13 relationship variables including commitment, trust, cooperation, mutual goals, interdependence or power imbalance, performance satisfaction, alternatives, adaptation, non-retrievable investments, shared technology, summative constructs, structural bonds, and social bonds. Even though almost a decade has passed since the article’s publication, these factors remain part of the set that contemporary research draws from. While the multitude of works in relationship marketing have without doubt lead to a better understanding of how relationships should be designed, one important point of criticism must be mentioned. As HEIDE/STUMP (1995, p. 58) note, the present knowledge of buyer-supplier relationships is limited, since one particular unanswered question pertains to the performance effects implications of relationship formation. They further state that while the theoretical models are based on the assumption that “relationships are established in order to enhance some aspect of performance”, no statistically sound proof has been put forward. In the years since the publication of the article by HEIDE/STUMP (1995), some additional, yet not fundamentally new research, has been produced. Consequently, the issue will be addressed as a particularly relevant research need in chapter 2.4.1. 2.3.3.2 Aspects of relationship marketing in current logistics outsourcing relationship research

As this chapter will show, logistics research has only slowly and reluctantly taken up the models proposed in the relationship marketing literature and adapted them in parts for their own purposes. Yet, a number of open questions that need to be addressed in further research still remain. The first to explicitly examine trust and commitment in logistics alliances in a hypotheses driven model was MOORE (1998, pp. 25-26).

2.3 Logistics outsourcing relationships

45

Previously, a number of logistics researchers had made contributions that led the way for the integration of relationship marketing into the logistics relationship research. LALONDE/COOPER (1989, p. 6) observed that buyers and third party logistics providers in most business sectors have shifted from transaction-driven to contract-driven relationships. MOORE (1998, p. 25) interprets that it appears that some relationships over time evolve into partnerships or alliances as mutual trust develops between the customer and the LSP. This, he concludes, suggests that the parties involved are making commitments to establish closer and longer-term relationships. BOWERSOX/DAUGHERTY/DROGE/ROGERS/WARDLOW (1989) describe logistics alliances as long-term cooperative relationships in which both parties seek to establish a jointly rewarding exchange in which mutual trust and both parties’ commitment, including the sharing of risks and rewards, are of particular importance. In a further article, BOWERSOX (1990) adds that apart from commitment and trust, especially cooperation, dependency, and the sharing of risks and rewards are vital for logistics relationships. He furthermore remarks that the absence of trust can be detrimental to the relationship. A similar argumentation can be found in GARDNER/COOPER (1988) who assert that logistics relationships extend beyond transactional exchanges and typically evolve over a long period of time. They also point out the importance of cooperation. In long-term relationships, the parties are motivated to use cooperation to manage risks and uncertainty in order to reap the benefits normally attained through vertical integration. ELLRAM/COOPER (1990, p. 4) state that successful logistics alliances have a long-term orientation that requires trust, loyalty and the sharing of information, risks and rewards since only by sharing benefits and burdens parties are managing risks and develop trust and loyalty. As the research presented above demonstrates, the parties involved in buyer-supplier relationships are invoked to move from purely short-term and opportunistic positions towards long-term partnerships. These demands, however, have yet again a normative character. Until it can be proven with the adequate scientific rigor – as requested by MENTZER/ KAHN (1995, p. 244), employing hypotheses driven models and empirical research – the issue cannot be considered resolved. This is especially true for the implied performance effects of partnership formation. A first significant step as mentioned above is taken by MOORE (1998, pp. 24-37). In an effort to combine the findings of the relationship marketing literature with the individual characteristics proposed by logistics research, he developed a first causal model. In the model, MOORE includes 9 factors which either apply to the buyer or the third party. In detail, these

46

2 Basic concepts

are the third party’s information exchange, relationship commitment and equity as well as the buyers relationship benefits, trust in the third party, relationship conflict, risk of third party opportunism, relationship effectiveness, and relationship commitment. MOORE (1998, p. 32) shows that 11 of the proposed 23 hypotheses do not find empirical support and have to be rejected. Among those that prevail are that the third party’s relationship commitment influences the buyer’s relationship commitment positively, while it is influenced negatively by the buyer’s relationship conflict. Furthermore, MOORE (1998) can show that the third party’s information exchange reduces the relationship conflict while only the third party’s equity enhances the buyer’s trust in the third party. Increasing trust, in turn, positively influences the buyer’s relationship effectiveness which is also negatively affected by relationship conflict. While these findings suggest that both trust and commitment are important factors in relationships between customers and LSPs, according to MOORE (1998, p. 33) they seem to suggest that in terms of logistics alliance success, the variables commitment and relationship effectiveness are influenced stronger through negative outcomes initiated by conflict rather than through positive outcomes associated with trust. While the research of MOORE (1998) is an important contribution to logistics relationship research, a number of questions remain to be answered. Additional to the fact that several reasonable hypotheses suggested by previous research could not be supported, the connection between trust, commitment, and logistics alliance success, which he tries to establish remains normative. The model does not contain constructs adequate for measuring the actual alliance success and therefore must refrain from drawing conclusions on this issue. The parallel works of KNEMEYER/MURPHY (2004) and ENGELBRECHT (2004) have since overcome this deficit and are the first to empirically prove the connection between relationship design and logistics outsourcing performance. KNEMEYER/MURPHY (2004, pp. 38-39) explicitly refer to and build on the research conducted by MOORE (1998) and point out that the potential causal linkages between relationship marketing dimensions such as trust and communication and successful 3PL arrangements remain to be shown. In their research, KNEMEYER/MURPHY (2004, p. 40) assign trust a central role which is affecting all three performance dimension in the form of asset reduction performance, channel performance, and operations performance. Trust in turn is directly positively influenced by the four factors specific investments, prior satisfaction, 3PL reputation, and communication as well as negatively effected by opportunistic behavior.

2.3 Logistics outsourcing relationships

47

Empirically testing their model with a total of 388 responses, KNEMEYER/MURPHY (2004, p. 44) are able to show that all previously proposed hypotheses are supported. Here, trust – along with communication – indeed has a positive influence on both channel performance and operations performance. This fact indicates that the so often normatively assumed causal linkage between relationship marketing variables and partnership success exists. The model furthermore provides support for the hypotheses that communication is positively affecting trust which in turn is linked to a positive asset reduction performance. Additionally, all antecedents of trust have the proposed effects. The work by KNEMEYER/MURPHY (2004) must therefore be considered another very important step in logistics relationship research. While they only test a limited number of antecedents for logistics outsourcing performance, the insights they offer for designing successful logistics outsourcing relationships are far-reaching. It is particularly interesting to note that the parallel work by ENGELBRECHT (2004, pp. 272-277) addresses the same issue from a slightly different perspective. In an effort to explain how the implementation of an outsourcing project influences the outsourcing performance, ENGELBRECHT (2004) proposes a model that can explain with a squared multiple correlation (R2) of 40% a significant share of logistics outsourcing performance. While chapter 5.2 will in detail discuss the meaning of different statistical measures, it must be noted here that a R2 of 0.4 or 40% can be considered relatively high for a partial model which does not utilize most of the traditional relationship variables such as commitment and trust. Instead, the model does contain conflict and communication, and employs the experience of the customer with outsourcing projects, the involvement of the LSP in the outsourcing process, the duration of the implementation period and the improvement efforts displayed by the LSP. Empirically, ENGELBRECHT (2004, p. 276) can provide support for his hypotheses that the performance of logistics outsourcing relationships is influenced negatively by the level of conflicts in the relationship and positively by the improvement efforts of the LSP. He furthermore shows that higher levels of improvement efforts of the LSP lead to more communication, which is also strongly positively affected by the involvement of the LSP in the outsourcing process. Even though the model of ENGELBRECHT (2004) has an exploratory character and does not employ many of the traditional variables of relational marketing, it holds two important insights. The first is that the implementation of logistics outsourcing projects indeed has a significant influence on the logistics outsourcing performance. The second is that additional to the constructs from relationship marketing whose causal link-

48

2 Basic concepts

ages to outsourcing performance have already been shown, other constructs such as improvement efforts of the LSP seem to have a considerable influence as well. The preceding chapters have in detail analyzed the importance and effects of partnerships, have introduced models for the process of partnership development and have given insights into the designing of logistics outsourcing relationships. For this, the status quo of relationship marketing research and its consequences for logistics outsourcing research were presented. During the discussion, a number of research needs and open questions were identified. The following chapter presents a summary of those, identifies the most important research needs and on these grounds develops research questions that will guide the further research in this work.

2.4 Research model 2.4.1 Identification of research needs Logistics research has been receiving increasing academic attention for the past two decades, triggered by the substantial changes in the worldwide logistics markets and the accompanying need for a better and abstract understanding of the underlying mechanisms that drive logistics performance and enable firms to gain competitive advantages. Following this development the topic of logistics outsourcing has been of particular importance as an ever increasing number of firms decided to utilize logistics service providers to deliver what in most cases was not seen as a core competence of the respective firms. Consequently, an intensive debate developed over possible benefits and risks associated with logistics outsourcing (LIEB/MILLEN/VAN WASSENHOVE 1993; MCGINNIS/KOCHUNNY/ACKERMAN 1995). Unfortunately, the logistics outsourcing debate for the most part neglected the importance of the relationship between customer and logistics service provider and thus cannot answer the question which performance implications relationship formation has for the outsourcing arrangement. Only recently works by MOORE (1998), KNEMEYER/MURPHY (2004) and ENGELBRECHT (2004) have started to address this issue. The majority of firms today are outsourcing at least parts of their logistics activities to third parties. While those firms that currently do not outsource broadly indicate that they are planning to do so, outsourcing firms show considerable interest to expand their involvement with third parties as argued in chapter 2.2.3.1. As this process evolves, relationships between customers and LSPs are becoming closer and increasingly long-term oriented (LAMBERT/EMMELHAINZ/GARDNER 1999; HOYT/HUQ 2000).

2.4 Research model

49

Knowing about this development, the question arises which factors truly are important for the successful realization of outsourcing projects and how they must be addressed to optimize outsourcing performance. One of the most fundamental issues to be studied in this context is the formation and design of the relationships between the parties involved in logistics outsourcing. As argued in chapter 2.3.3.2, the logistics literature so far has not adequately dealt with this problem. One explanation for this is that logistics outsourcing decisions presently are predominantly motivated by the potential to cut costs while the significantly stronger effect to increase the firms logistics performance through outsourcing as demonstrated by DEHLER (2001, pp. 233-244) and ENGELBRECHT (2004, pp. 63-64) is largely ignored. Support for the finding of adequate ways for the designing of relationships comes from research in the field of relationship marketing which explicitly focuses on this issue. The absence of specific logistics oriented research on this topic of course is only relevant if the general results of relationship marketing are not easily transferable to the relationships between customers and LSPs. That the outsourcing of logistics activities is a complex matter has been established above – the specificity of the processes, the long-term orientation of the contracts as well as the dependence of the customer, resulting from the sometimes large investments involved and the meaning of logistics performance for customer satisfaction all contribute to this. A general and uncritical transfer of relationship marketing findings to logistics research therefore is not possible. Instead, the applicability of key relationship marketing concepts for the logistics context must be examined in detail and therefore be one significant part of future logistics research. While the works by MOORE (1998), KNEMEYER/MURPHY (2004) and ENGELBRECHT (2004) as well as numerous articles from relationship marketing (ANDERSON/NARUS 1984; ANDERSON/LODISH/WEITZ 1987; ANDERSON/WEITZ 1989; HEIDE/JOHN 1990; NOORDEWIER/JOHN/NEVIN 1990; HALLÉN/JOHANSON/SEYED-MOHAMED 1991; ANDERSON/WEITZ 1992; HEIDE/JOHN 1992; HAN/WILSON/DANT 1993; GANESAN 1994; MORGAN/HUNT 1994; WEITZ/JAP 1995; WILSON 1995) have begun to investigate the link between the different variables influencing relationships, much remains to be learned. One of the most important points is to reveal which factors indeed are relevant in an outsourcing relationship. Additionally, the question must be addressed to which extent – if any – performance effects can be traced back to the way the relationship between customer and LSP is designed. So far, no approach exists that would combine all relevant factors of relationship formation and the subsequential per-

50

2 Basic concepts

formance effects into a closed theoretical model with forecasting value for both logistics theory and practice. As shown in chapters 2.3.3.1 and 2.3.3.2, a number of researchers such as ANDERSON/WEITZ (1989), ANDERSON/NARUS (1990), MORGAN/HUNT (1994), MOORE (1998), KNEMEYER/MURPHY (2004) and ENGELBRECHT (2004) have already addressed the issue of relationship formation. It must be noted, however, that the theoretic foundation of the causal linkages between the different factors in some of these works is rather thin and does not meet the requirements concerning adequate scientific rigor as formulated by MENTZER/KAHN (1995). A future model encompassing the relevant factors for logistics relationship formation must therefore rather be derived from a solid theoretical basis. A good example is set by WALLENBURG (2004, pp. 89-122) while other works, such as those of CRUTCHFIELD (2001) and RANAWEERA/NEELY (2003), to the contrary argue virtually without a foundation in theory and instead concentrate on variables that have been previously used in other research contexts. While this of course is also an important part of research and model design, assuming causal linkages between variables solely on this basis without the adequate theoretical foundation must leave open the questions whether an observed interrelation truly is a causal linkage or merely a correlation. Overcoming these deficits and thus pursuing a rigorous scientific approach is a key goal of this work. A major shortcoming of the current relationship literature is the lack of context-related research. While numerous studies have been investigating the buyer-supplier relationship by interviewing individual firms or dyads, the resulting models have never been tested for potential moderating effects of contingency variables. This implicitly assumes that within the respective studies, all analyzed relationships are similar and comparable. However, this disregards the possibility that the various surveyed relationships may be subject to entirely different contingencies, which potentially could lead to significantly different results if modeled accordingly. Possible implications that could result e.g. from either the concentration on too few industries or the inclusion of too many industries in the sample might therefore remain unnoticed, even though it could be expected that the causal linkages between the relationship variables differ significantly. This could for instance be the case between consumer and industrial goods industries or between the suppliers of commodity goods and those of highly specialized high-tech equipment. An example may be the central relationship variable “trust” which might be of utmost importance in highly customized relationships with a high dependence of the buyer on the timely and error-free delivery of goods, while in a relationship that only

2.4 Research model

51

involves the transportation of an industrial commodity, it may only play a minor role. As this chapter has shown, a number of deficits exist in current logistics outsourcing relationship research. These deficits must be addressed, in parts by adapting existing relationship marketing literature, in parts by developing new approaches which as yet have not been considered in logistics research. 2.4.2 Identification of research questions The intention of this work is to analyze the relationships between customers and logistics service providers in order to discover the performance implications that the design of logistics outsourcing relationships have. To do so, it is vital to break down the complex problem into several smaller parts that can be addressed through a few distinct research questions. This will increase transparency on the one hand and reduce the complexity of the task on the other. When adopting the perspective of the customer, the main problem is to identify the factors that influence the relationship with the LSP. This is a complex task: on the one hand, a multitude of different variables and theoretical approaches exist to choose from. On the other hand, the customer has a particularly strong interest in the implications of the relationship variables for the logistics outsourcing performance, which in turn may influence the firm’s logistics performance. The first research question therefore is: R1 :

Which are the main influencing factors of an outsourcing relationship between a customer and a LSP and what influence do they have on the outsourcing performance?

Additionally, it is of particular interest to see how these different variables might be integrated into one model that comprises the relationship constructs as well as their performance effects for the outsourcing of the logistics activities. Here, special consideration must be given to the interdependencies existing between the variables in order to adequately estimate the causal linkages inside the model and to the design of the logistics outsourcing performance construct. The corresponding research question is the following:

52

2 Basic concepts

R2 :

How are the influencing factors of outsourcing relationships and their performance effects to be integrated into one model and which interdependencies must be considered between them?

The question of the performance effects of the relationship design between customer and LSP is only a strategic one if a connection can be shown between the outsourcing performance and the resulting logistics performance of the customer. The relevance in that case would result from the insight that the design of the outsourcing relationship would have a direct influence on the logistics performance. This in turn may be of particular relevance if it could be shown to positively affect the overall firm performance. The third research question therefore reads: R3 :

Which influence does the outsourcing performance have on the logistics performance and on the firm performance?

Finally, the question of the validity of the answers to research questions 1-3 must be addressed with respect to the context of the firms. Since the individual context may vary depending on the firms’ specific and individual situations, some of the causal linkages in the models are presumably moderated by those contingency variables. It therefore is necessary to first identify contingency variables with potential moderating effects. In a second step, their influence on the logistics outsourcing relationship model and its performance effects must be analyzed. The fourth research question thus is the following: R4 :

Which contingency variables for the logistics outsourcing context can be identified and which moderating effects can be observed in the outsourcing performance model and its causal linkages to logistics and firm performance?

2.4.3 Procedure to answer the research questions MENTZER/KAHN (1995, p. 231) state that “research is an intricate and rigorous process that should not be taken lightly nor pursued in an unstructured manner”. The research questions as proposed in chapter 2.4.2 aid this process by lending structure and enabling first insights to be gained through conceptual thoughts based on solid theory selection. To comprehensively answer the research questions, however, the theoretical thoughts must be challenged by reality on the basis of empirical

2.4 Research model

53

data material and an adequate methodology, performed with substantial scientific rigor. Since numerous different approaches regarding the design of the research process are imaginable, the procedure chosen for this study will be outlined in the following. It is based on the framework suggested by MENTZER/KAHN (1995, pp. 233-240) for logistics research as presented in Figure 2-4, because it constitutes a research procedure that corresponds to the established and high standards of other academic disciplines and fields of research and therefore ensures the demanded scientific rigor. Idea generation

Literature review

Observation

Substantive justification

Theory

Hypothesis

Constructs

Methodology

Measures

Analysis

Conclusion

Fig. 2-4. Framework of logistics research (MENTZER/KAHN 1995, p. 234)

In a first step, the idea for the research has been generated through extensive literature review and practical observations. This promoted the substantive justification which led to the development of research questions and thus provides the impetus for the research endeavor. In the next step, the theoretical foundation for the research will be built, employing a pluralistic approach to adequately address the complexity of the research questions. Utilizing the generated findings from theory and the existing empirical research, the variables of the outsourcing relationship model will then be conceptualized to subsequently derive hypotheses on their causal linkages and performance effects.

54

2 Basic concepts

Building both on the theory and on the hypotheses, constructs will be developed that via a multitude of different measures operationalize the research questions. After that, as postulated by MENTZER/KAHN (1995, pp. 237-240), a methodology suited to answer the research questions will be chosen. For this particular research, it will be a large empirical survey that will be conducted in order to gather the database needed for the calculation of the causal linkages in the model via structural equation modeling. After that, the data will be analyzed with scrutiny to ensure that all research questions formulated in chapter 2.4.2 will be addressed and conclusions can be drawn.

3 Theoretical framework

3.1 Theories suited to explain cooperation in logistics relationships So far, no single theory exists that would be suitable to propose the optimal way to manage relationships in general – and between customers and LSPs in particular – in order to maximize performance outcomes. As WALLENBURG (2004, pp. 61-62) puts it, the issue of relationship formation would be simple under the assumption that firms act purely economical and are not characterized by bounded rationality. Perfect contracts between buyers and sellers could be agreed upon after customers would have chosen the service provider purely on criteria of efficiency and effectiveness. Central for the choice of the customer would only be the costs and the corresponding service levels as well as the competence of the service provider, which is the one critical determinant for the development of costs and service levels in the future. Relationship factors, such as trust, commitment or communication would not matter in this world of perfect contracts in which purely rational decisions would determine whether a service is produced in-house, outsourced to a service provider or if the service provider must be changed in order to reduce costs or increase the service level. Just the introduction of bounded rationality adds a degree of complexity to relationship management that severely complicates the issue. It therefore does not come as a surprise that no single theory as yet has been put forward that would be widely accepted to explain how relationships should be formed under which circumstances to maximize its outcome. However, as described in the previous chapter, numerous authors have addressed relationship formation and management. The theories employed range from neo-classical approaches over industrial economics, new micro-economic theories (transaction cost analysis and principal agent theory) and game theory to social exchange theory, the commitment-trust theory, the contingency approach and the resource-based view. Even though several remarkable works have originated by using one theory only, recent years have seen an increase in research employing a

56

3 Theoretical framework

combination of theories (HEIDE 1994; HOYT/HUQ 2000; SKJOETT-LARSEN 2000; ENGELBRECHT 2004; WALLENBURG 2004). Only this plurality seems to adequately address the issue of relationship management. It must be noted, however, that for a significant explanation the theories must be combined in a complementary, not in a competing way. In order to fully embrace the problem, the following chapters will introduce a selection of theories potentially suited if combined to explain how relationships between customers and LSPs should be designed to maximize performance outcomes and to adequately address the largely neglected issue which contingency factors might influence the explanatory power of the resulting model. While several theories are suited in general to explain the phenomenon of relationship formation, the theories employed in this study will be specifically selected to answer the research questions introduced in chapter 2.4.2. As will be shown, social exchange theory and commitment-trust theory are best suited to formulate the outsourcing relationship model requested by research question 1 and 2, while transaction cost theory and contingency theory are appropriate to describe the general conditions and business environment that determine the contingency factors demanded by research question 4. Since all of these theories have been widely discussed, their description in the following chapters will be deliberately limited to the most important aspects. More room will be provided for the presentation of the contribution the theories can make to logistics relationship research.

3.2 Introduction to selected theories 3.2.1 New institutional economics and transaction cost theory According to WILLIAMSON (1981, p. 552), “a transaction occurs when a good or service is transferred across a technologically separable interface”. Transaction cost theory builds on the work by COASE (1937), who identified several limitations to the neoclassical paradigm for understanding relationships between firms. Central to neoclassical economics is the concept of a single product firm, operating in a perfectly competitive industry with multiple competitors all producing the same product under the same conditions and all facing the same market demand curve (HOBBS 1996, pp. 15-16). The standard neoclassical transaction involves the exchange of a homogenous product without quality variations. Consequently, no costs are involved in the transaction since measuring the value of the product is not necessary – if products were to exhibit differences in qual-

3.2 Introduction to selected theories

57

ity, they are regarded as distinct products serving different markets. In neoclassical theory, economic agents are assumed to possess perfect information, hence there is no uncertainty regarding prices, product characteristics or the behavior of competitors and trading partners. Opportunism therefore cannot exist. The approach as put forward by COASE (1937) recognized that there are costs to using the market mechanism. These include the costs of discovering the appropriate price, the cost for negotiating contracts for each exchange transaction and the costs incurred for specifying the details of a transaction in contracts which later were termed “transaction costs” (ARROW 1970, p. 67-81). COASE (1937, p. 395) argued that “a firm will tend to expand until the costs of organising an extra transaction within the firm become equal to the costs of carrying out the same transaction by means of an exchange on the open market or the costs of organising in another firm”. He thereby provided a rationale for the existence of firms which was based on the costs of carrying out a transaction. While these thoughts did not have a major impact on economic thought for decades, in the 1970s the interest in transaction costs gradually increased. Groundbreaking work in the development of the theory of transaction costs was carried out by WILLIAMSON (1975) and WILLIAMSON (1979). Gradually, several other theories based on the concept of transaction costs emerged. These include alongside transaction cost theory (WILLIAMSON 1979) also the property-rights school (ALCHIAN 1965; DEMSETZ 1967; ALCHIAN/DEMSETZ 1972) and agency theory (JENSEN/ MECKLING 1976). Since the transaction cost theory is especially suited to explain the efficiency of different forms of organization and coordination, the following chapter will present the theory in greater detail. 3.2.1.1 Basic concepts of transaction cost theory

The object of transaction cost theory is to explain the existence of different firms and to analyze the optimality of specific coordination mechanisms depending on transaction characteristics. This allows insights into the question why specific activities are carried out internally or are bought from the market (WILLIAMSON 1975; WILLIAMSON 1985). Whenever transactions between parties are attempted or realized, a number of costs arise due to imperfections of the economic system. These transaction costs are defined by WILLIAMSON (1981, pp. 552-553) as the “competitive costs of planning, adapting, and monitoring task completion under alternative governance structures”. These can be decomposed into four separate costs

58

3 Theoretical framework

(WILLIAMSON 1985; NORTH 1990; HENNART 1993; DYER 1997): search costs, contracting costs, monitoring costs and enforcement costs. Search costs include the costs of gathering information to identify and evaluate potential trading partners and in essence consist of travel and communication expenses as well as costs for advisory services. Contracting costs refer to the costs associated with negotiating and writing an agreement. This particularly applies to the fees for legal counseling and the costs that internal coordination generates. Monitoring costs mean those costs associated with monitoring the agreement to ensure that each party fulfills the predetermined set of obligations. Enforcement costs are those costs that are incurred when a party must ex post bargain with or sanction a trading partner that does not perform according to the agreement. Transaction cost theory rests on two basic behavioral assumptions about the transaction partners involved: bounded rationality and opportunism. Bounded rationality can be the result from insufficient information, limits in management perception or the limited capacity to process information. The assumption is that the actors try to act rationally, but due to intellectual restrictions or the inability to adequately communicate their opinions and knowledge to others they are limited in their rational behavior. This is particularly relevant in situations of high uncertainty or complexity, because it further increases the pressure on the actors. Resulting from the assumption of bounded rationality, perfect contracts are impossible as neither can all possible options be identified nor can all future situations be considered. Moreover, once a contract has been accepted by both sides, it will be impossible to perfectly control the results. These factors enable the second assumption of opportunistic behavior. This, defined by WILLIAMSON (1985, p. 47) as “self-interest seeking with guile”, for at least one transaction partner leads to suboptimal results and must therefore be limited as far as possible. According to WILLIAMSON (1985, p. 2) transactions can be characterized by the three critical dimensions frequency, uncertainty and asset specificity. The frequency of transactions has implication for the amortization of transaction-specific investments. More frequent transactions reward long-term relations between parties through shorter amortization periods and lower average costs per transaction through the realization of economies of scale (PICOT 1991, p. 347; ANTLITZ 1999, p. 92). Therefore, increasing frequency in transactions should foster hierarchic coordination. A similar argumentation is used for uncertainty. The higher the uncertainty about the future, the more difficult is it to draw up contracts and the higher contracting and monitoring costs will be, which translates into a steeper gradient of the respective curves as presented in Figure 3-1. It can thus be argued that with higher levels of uncertainty transaction costs will

3.2 Introduction to selected theories

59

rise and consequently purchasing of activities on the market is less attractive. According to WILLIAMSON (1990, p. 142), the most important factor influencing the level of transaction costs is the asset specificity. He defines: “asset specificity has reference to the degree to which an asset can be redeployed to alternative uses and by alternative users without sacrifice of productive value”. The higher the “sacrifice of productive value” of an asset when used in a different transaction, the higher is the asset specificity and the more attractive is a hierarchic coordination of the activity. Depending on the individual characteristics of these three dimensions, the most appropriate coordination mechanism for a transaction can be determined. Available are the three forms of coordination market, hierarchy and cooperation, which are displayed as different curves in Figure 3-1. Transaction costs

Market

Cooperation

Hierarchy

Transaction dimensions (asset specificity, uncertainty, frequency)

Fig. 3-1. The optimal form of organization depending on transaction dimensions

Generally speaking, for a transaction with low asset specificity under low uncertainty, the market is the optimal form of coordination, followed by cooperation and hierarchy. It must be noted that with increasing transactions costs, the costs for using the market as form of coordination increase quicker than those of a cooperation and hierarchy. Therefore, depending on the characteristics of a transaction, the optimal form of coordination can be determined. Transaction cost theory furthermore has many different facets and applications that could be elaborated on in greater detail. Since it has already been discussed widely in the literature (WILLIAMSON 1975; WILLIAMSON 1985; RINDFLEISCH/HEIDE 1997), its introduction will be limited to the comments presented above which are sufficient for the purpose of this

60

3 Theoretical framework

work. The upcoming chapter will make further efforts to point out its explanatory value to logistics relationship research. 3.2.1.2 Explanatory value for logistics relationship research

Transaction cost theory views firms and markets as alternative forms for coordinating a transaction and suggests that the exchange governance is driven by the firms’ desire to minimize the direct and opportunity costs of exchange (RINDFLEISCH/HEIDE 1997, p. 31). Thus, transaction cost theory allows valuable insights on the choice of the form of coordination depending on the characteristics of a transaction and suggests circumstances under which long-term relationships are preferable over other forms of coordination for a transaction. However, according to LAMBE/WITTMANN/SPEKMANN (2001, pp. 2-3) one of transaction cost theory’s basic premises is “that the risk of partner opportunism limits the effectiveness of relational governance in exchange relationships”. This is in strong contrast to the empirical findings of several researchers who have shown that indeed relational control in the form of norms or personal relations is often an effective means of governance (ANDERSON/NARUS 1984; ANDERSON/NARUS 1990; DWYER/SCHURR/OH 1987; MORGAN/HUNT 1994; WILSON 1995). Additionally, doubt has been cast on transaction cost theory’s basic assumption of universal opportunism, especially in relational exchange (HEIDE/JOHN 1992; MORGAN/HUNT 1994). It must therefore be ascertained that transaction cost theory is limited in its capacity to explain exchange governance in exchange relationships in which the partners are actually able to develop relationship-based governance over time (LAMBE/WITTMANN/SPEKMANN 2001, p. 3). The implications of these findings for the explanatory value of transaction cost theory for the formation and management of logistics relationships is far-reaching. While the theory is obviously suited to explain why customers and logistics service providers engage in cooperative arrangements it can only be utilized to derive conclusions on the design of the actual relationship in a very limited fashion. These conclusions, however, are a central aim of this study as formulated in research questions 1 and 2 (see chapter 2.4.2), because it will be assumed that customer and LSP already have decided upon entering a relationship. Thus, the question remaining is not if a cooperative relationship should be formed, but rather how the parties should act within it. Long-term relationship between firms, based on trust and mutual cooperation, which are replacing the traditional adversarial relationships can improve manufacturing firm performance (HOYT/HUQ 2000, p. 755). The trust between the two parties can play a central role in this context. As

3.2 Introduction to selected theories

61

JARILLO (1988, p. 37) points out, the generation of trust in a network is suited to lower transaction costs arising e.g. from opportunism. This is due to the fact that an atmosphere of trust is beneficial for more efficient problem solving (ZAND 1972; BOSS 1978), since information is exchanged freely and more solutions to a problem are being explored because decision makers do not feel the obligation to protect themselves against the others’ opportunistic behavior (JARILLO 1988, p. 37). As these examples for trust and opportunism show, transaction cost theory allows evaluating whether certain variables are beneficial for the governance of logistics relationships. However, as a single theory it is not sufficient to derive the most important variables necessary for adequate relationship governance and therefore needs the input from further theories which will be presented in the following chapters. After having established these shortcomings which directly effect the answering of research questions 1 and 2 it must be pointed out that the explanatory value of transaction cost theory is very promising with respect to research question 4, which examines the existence of contingency factors that potentially influence the relationship variables and their performance effects. Contingency factors, as will be shown in chapter 3.2.4.1, are those variables that characterize the situation in which a firm’s practices, procedures and processes are established and applied (CLAYCOMB/FRANKWICK 2004, p. 21). Here, the findings of Williamson (1985, p. 2) that transactions can be characterized by the three critical dimensions frequency, uncertainty and asset specificity are of particular importance. They constitute three contingency factors since they directly characterize the situation the customer is affected by. In the case where the relationship between the customer and the LSP is characterized by high asset specificity, because e.g. large investments in equipment and employees have been made, changing the LSP will be accompanied by higher transactions costs for the customer than in a case where the asset specificity is very low. With a similar argumentation it can be shown that with rising levels of uncertainty the transaction costs for the customer rise since contracts and agreements potentially must be modified in the course of a relationship due to unforeseen changes in the behavior of the partner or the environment. Another influencing factor is the frequency of the transaction which in exchange relationships has implications for the amortization of transaction-specific investments. Customers with very frequent transactions will seek longer and deeper relationships with service providers since economies of scale and lower average costs per transaction substantially reduce transaction costs.

62

3 Theoretical framework

3.2.2 Social exchange theory Initially, the transaction cost theory offered an acceptable explanation of governance mechanisms in inter-organizational relationships (HOYT/HUQ 2000, p. 754). However, as institutional markets and exchange practices advance, transaction cost theory seems to be losing some of its explanatory power and concepts such as trust and cooperation are gaining increasing importance for understanding successful buyer-supplier relationships (GHOSHAL/MORAN 1996, pp. 40-42). To overcome these deficiencies and to find new ways to explain relational exchange, researchers of inter-organizational relationships have recently drawn increasingly on social exchange theory (LAMBE/WITTMANN/ SPEKMANN 2001, p. 3). The core explanatory mechanism of this theory is the relational interdependence that develops over time through the interaction of the exchange partners (DWYER/SCHURR/OH 1987; HALLÉN/JOHANSON/SEYED-MOHAMED 1991). Social exchange theory thereby acknowledges the importance of interaction and communication for successful relationships and therefore is a valuable contribution for understanding how exchange relationships should be designed. 3.2.2.1 Basic concepts

Social exchange theory subsumes several approaches which focus on exchange relationships and aim at explaining the behavior of the parties involved (LAMBE/WITTMANN/SPEKMANN 2001; WALLENBURG 2004, p. 77). Its objective is to analyze the formation of relationships as well as to explain why they continue. While the emphasis is on long-term relationships, it is not restricted to a particular context. Private relationships between individuals can be examined as well as buyer-supplier relationships. Social exchange theory may be traced back to “one of the oldest theories of social behavior” which states that any interaction between individuals is an exchange of resources (HOMANS 1958, p. 597). The resources exchanged may be tangible such as goods or money, but also intangible, such as social amenities or friendship (LAMBE/WITTMANN/SPEKMANN 2001, p. 4). The basic assumption of social exchange theory is that “parties enter into and maintain relationships with the expectation that doing so will be rewarding” (LAMBE/WITTMANN/SPEKMANN 2001, p. 4; HOMANS 1958; THIBAUT/KELLEY 1959; BLAU 1964; BLAU 1968; MACNEIL 1980). The origin of the discourse on social exchange can be traced back to the ancient philosopher ARISTOTLE (1999, 1162a341163a24) and his work “Nicomachean Ethics”, in which social exchange is distinguished from economic exchange.

3.2 Introduction to selected theories

63

Of particular influence to the development of the theory were the works of sociologists HOMANS (1958), BLAU (1960), BLAU (1964), and EMERSON (1962) as well as those of social psychologists THIBAUT/KELLEY (1959). According to (BLAU 1968, p. 453) “the first systematic theory that focuses on social behavior as … [exchange]” was developed by HOMANS (1958). BLAU (1964) according to CHADWICK-JONES (1976) may have been the first to use the term “theory of social exchange” to describe his conceptualization of “social interaction as an exchange process”. THIBAUT/KELLEY (1959) are often cited because of their contribution to social exchange theory in the form of their comparison level concept, which is used to explain how parties in the exchange relationship weigh the benefits of the relationship to determine their relationship commitment. The main contribution to the theory by EMERSON (1962) may be found in his article on the effects of power and dependence on exchange relationships in which he theorizes that power imbalances cause relationships to be unstable and thus interdependence is crucial to the continuation of a social exchange relationship. The basic foundational premises of social exchange theory according to LAMBE/WITTMANN/SPEKMANN (2001, p. 6) can be reduced to four points which will be introduced in the following and analyzed in greater detail thereafter. The four premises are: x Exchange interactions result in economic and/or social outcomes. x These outcomes are compared over time to other exchange alternatives to determine dependence on the exchange relationship. x Positive outcomes over time increase firms’ trust on their trading partner(s) and their commitment to the exchange relationship. x Positive exchange interactions over time produce relational exchange norms that govern the exchange relationship. Social exchange theory “views exchange as a social behavior that may result in both economic and social outcomes” (LAMBE/WITTMANN/ SPEKMANN 2001, p. 6). It therefore is different from purely economical theories that only consider the economic results of a relationship because it also includes social components (WALLENBURG 2004, pp. 77-78). As LAMBE/WITTMANN/SPEKMANN (2001, p. 6) state, although economic rewards such as money are important, the social rewards such as emotional satisfaction or the pursuit of the own personal advantage are often valued more. Or as BLAU (1968, p. 455) puts it, the “most important benefits involved in social exchange do not have any material value on which an exact price can be put at all, as exemplified by social approval and respect”.

64

3 Theoretical framework

An exchange relationship over time can be viewed as a series of discrete exchange episodes or interactions, which result in economic, information, product/service, and social exchange (HAKANSSON/WOOTZ 1979; HAKANSSON 1982). The sum of the exchange interactions over time comprise the history of the exchange relationship and are used by the firms to anticipate the future costs and benefits of the relationship (KELLEY/ THIBAUT 1978). Since different costs are associated with the exchange relationship, among them the important opportunity cost of not being in another relationship (LAMBE/WITTMANN/SPEKMANN 2001, p. 8), social exchange theory suggests that parties will remain in the relationship as long as the trade-off between benefits and costs is positive and therefore the satisfactory rewards continue (HOMANS 1958; BLAU 1968). To offer a conceptualization of how firms compare the rewards of one relationship with those of an alternative, THIBAUT/KELLEY (1959) developed the concepts of comparison level (CL) and comparison level of alternatives (CLalt). CL represents the social and economical benefit standard that one feels is deserved in a given relationship and is compared to the outcomes felt to be received through the relationship. If the outcome is below what is perceived to be deserved, dissatisfaction results, while higher outcomes lead to satisfaction. To decide whether a relationship is continued or not, CL alone is not sufficient, since external factors such as dependence can hinder a firm from terminating a relationship. THIBAUT/KELLEY (1959) consequently developed the second standard CLalt. This comparison level of alternatives is used to describe the overall social and economical benefit available from the best possible alternative exchange relationship. If the outcome of a relationship exceeds CLalt, the respective party will have a degree of dependence on the relationship, because it offers greater rewards than could be achieved in an alternative relationship. Therefore, the party will maintain the relationship, even if it is dissatisfied. However, if an alternative relationship can offer greater benefits, the buyer will switch suppliers. Thus, CLalt is the lowest level of rewards that one party is willing to receive from a relationship in order to continue its involvement. As the positive outcomes of a relationship continue, trust between the parties evolves. This is of particular importance, since social exchange is governed to a large degree by social obligations rather than by contracts (BLAU 1968). These obligations result because as one party provides the other with a benefit, it anticipates the other to adequately reciprocate (HOMANS 1958, BLAU 1964). As mutual reciprocations of beneficial actions through multiple interactions occur over time, trust is created

3.2 Introduction to selected theories

65

(HOMANS 1958, BLAU 1964). Therefore, trust creates obligations between exchange partners and leads to closer relationships. In the argumentation of the social exchange theory, trust is also important because it contributes significantly to the level of partner commitment to the relationship (HOMANS 1958, BLAU 1964). The causal relationship between trust and commitment results from the principal of generalized reciprocity, which according to MCDONALD (1981, p. 834) holds that “mistrust breeds mistrust and as such would also serve to decrease commitment in the relationship and shift the transaction to one of more direct shortterm exchanges”. Commitment is important to exchange relationships because it ensures that partners will make the efforts and investments necessary to produce mutually desirable outcomes (DWYER/SCHURR/OH 1987; GANESAN 1994). These mutually desired outcomes with respect to CL and CLalt increase the partners’ tendency to continue the relationship or increase their commitment to the relationship (THIBAUT/KELLEY 1959). Social exchange theory furthermore assumes that over time norms are developed that govern the exchange relationship. These norms, which are explicit or tacit mutually agreed upon rules for behavior, develop over time as the parties in the exchange relationship interact with each other (HOMANS 1958; THIBAUT/KELLEY 1959). Norms increase the efficiency of relationships because they reduce the degree of uncertainty without the difficulties created by using power (THIBAUT/KELLEY 1959). Parties in a relationship adhere to these norms because they believe that doing so will be rewarding (EMERSON 1962; BLAU 1964). In fact, as HOMANS (1958) shows in a group setting, the more one party conforms to the norms of the relationship, the more rewards it will receive. 3.2.2.2 Explanatory value for logistics relationship research

Relationships between customers and logistics service providers, like most other business-to-business relationships, are characterized by the fact that they are not only based on economical but also on social exchange. The concept of social exchange has been suggested by a number of authors to be an important element of successful logistics relationships (BOWERSOX/DAUGHERTY/DROGE/ROGERS/WARDLOW 1989; LALONDE/ COOPER 1989; BOWERSOX 1990; MOORE/CUNNINGHAM III 1999; WALLENBURG 2004; WEBER/WALLENBURG 2004). Social exchange theory allows a deeper insight into how these relationships should be designed in order to enhance their outcomes. As WILSON (1995, p. 335) notes, a substantial body of research “about the variables that make for a successful relationship” exists. These variables, such as trust and commitment, according to LAMBE/WITTMANN/SPEKMANN 2001,

66

3 Theoretical framework

p. 14) have been identified as being critical facilitators of relational exchange. In fact, as WILSON (1995, p. 335) points out, the variables of social exchange theory have been used to create empirical success models. These aim at determining the degree to which relational exchange variables are achieved and examine the degree to which elements of relational exchange lead to enhanced exchange performance (LAMBE/WITTMANN/ SPEKMANN 2001, pp. 14-15). Aside from this strength of social exchange theory, it also has some weaknesses. While transaction cost theory is frequently criticized for its assumption of universal opportunism, according to LAMBE/WITTMANN/ SPEKMANN (2001, p. 26) social exchange theory should be criticized for its implicit assumption that relational exchange is completely devoid of opportunism. This may be one reason why research so far has not been able to show that relational governance can replace formal governance (RINDFLEISCH/HEIDE 1997, p. 50; LAMBE/WITTMANN/SPEKMANN 2001, p. 26), because partners relying solely on norms and trust in relational exchange are more vulnerable to the other party’s opportunism. Nevertheless, the research surrounding social exchange theory is especially interesting for analyzing logistics relationships. As relationship variables such as trust and commitment operationalize the premises of the theory, insights into the design of relationships between customers and LSPs can be gained. Even more so, as shown by LAMBE/WITTMANN/SPEKMANN (2001, p. 15), the research utilizing operationalized facets of social exchange theory is also suited to analyze the benefits and costs that exchange partners receive from the exchange relationship, as it has for instance been done by ANDERSON/NARUS (1984, 1990). Therefore, social exchange theory can indeed be applied to find answers to research question 1 and 2 proposed in chapter 2.4.2. Over the past years, a substantial body of research has made contributions to the question which variables should be included in relationship research. Since the theory as such does not make any perceptions in general aside from trust and commitment, variables must be chosen depending on their suitability for any given context.

3.2 Introduction to selected theories

67

Ɣ

Ɣ

Ɣ

Ɣ Ɣ

Ɣ Ɣ

Ɣ Ɣ Ɣ

Ɣ Ɣ Ɣ

Ɣ Ɣ Ɣ

Ɣ

Ɣ Ɣ Ɣ Ɣ Ɣ

Ɣ Ɣ Ɣ

Ɣ

Ɣ

Ɣ

Ɣ

Ɣ

Ɣ

Ɣ Ɣ Ɣ Ɣ Ɣ Ɣ

Ɣ Ɣ Ɣ Ɣ Ɣ

Ɣ

Ɣ

Ɣ

Ɣ Ɣ Ɣ Ɣ Ɣ

Ɣ Ɣ Ɣ Ɣ

Ɣ Ɣ

Ɣ

Ɣ Ɣ

Ɣ

Ɣ Ɣ Ɣ

Ɣ Ɣ

Ɣ Ɣ

Ɣ Ɣ Ɣ

Ɣ Ɣ

Ɣ

Ɣ Ɣ Ɣ

Ɣ

Ɣ

Ɣ

Others

Ɣ Ɣ

Satisfaction

Communication

Ɣ

Ɣ Ɣ

Norms

Cooperation

Ɣ

Dependence

Commitment

Author(s) ANDERSON/NARUS (1984) ANDERSON/NARUS (1990) ANDERSON/HAKANSSON/JOHANSON (1994) ANDERSON/WEITZ (1989) ANDERSON/WEITZ (1992) DWYER/SCHURR/OH (1987) FRAZIER (1983) GASKI (1984) GASSENHEIMER/HOUSTON/DAVIS (1998) GUNDLACH/ACHROL/MENTZER (1995) GUNDLACH/MURPHY (1993) HEIDE (1994) HEIDE/JOHN (1992) HOUSTON/GASSENHEIMER (1987) KNEMEYER/CORSI/MURPHY (2003) KNEMEYER/MURPHY (2004) LAMBERT/EMMELHAINZ/GARDNER (1999) LUSCH/BROWN (1996) MALONI/BENTON (2000) MOORE (1998) MORGAN/HUNT (1994) RING/VAN DE VEN (1994) SCHURR/OZANNE (1985) SKINNER/GASSENHEIMER/KELLEY (1992) SMITH/BARCLAY (1997)

Trust

Table 3-1. Variables frequently used in relationship research

Ɣ Ɣ

Ɣ Ɣ Ɣ Ɣ Ɣ Ɣ

Ɣ

Ɣ

Ɣ

Ɣ

68

3 Theoretical framework

TATE (1996) WALLENBURG (2004) WHIPPLE/FRANKEL/DAUGHERTY (2002) WILSON (1995) YOUNG/WILKINSON (1989)

Ɣ Ɣ

Ɣ Ɣ

Ɣ Ɣ Ɣ

Ɣ Ɣ

Ɣ

Ɣ

Ɣ Ɣ

Ɣ Ɣ

Ɣ Ɣ

Ɣ

As Table 3-1 shows, the variables most often utilized in relationship research include trust, commitment, dependence, cooperation, communication, norms, and satisfaction. Some of these are also very fitting for logistics relationship research. LALONDE/COOPER (1989) propose that, as trust evolves between customers and LSPs, relationships turn into alliances. According to them, trust is essential to overcome opportunism and achieve mutual advantage, thereby fostering success of the exchange relationship. The same thoughts are displayed when BOWERSOX/DAUGHERTY/DROGE/ROGERS/WARDLOW (1989) find that successful logistics relationships are characterized by mutual trust and commitment where both parties seek to establish jointly rewarding exchange. Other authors suggest that the variables cooperation and communication are important for successful logistics relationships (GARDNER/COOPER 1988; ELLRAM/COOPER 1990). As logistics relationships extend beyond transactional exchanges, parties are motivated to manage risk and uncertainty through cooperation, thereby allowing them to receive benefits normally only attainable through vertical integration. This of course requires the sharing of information, risks, and rewards in order to develop a mutually beneficial relationship. Satisfaction generally also plays an important role in relationships (HOMANS 1958; THIBAUT/KELLEY 1959; BLAU 1964). According to LAMBE/WITTMANN/SPEKMANN (2001, p. 25), it provides an insight into a relationship’s overall performance and thereby may serve as an operationalization of the success of the exchange relationship. Therefore, the satisfaction of the parties with the outcome of a logistics outsourcing relationship can be used as an approximation of its success (ANDERSON/NARUS 1990, p. 46) and thus is suited for answering research question 1 and 2 as proposed in chapter 2.4.2. Dependence, however, will not be analyzed in this study. While a substantial amount of research already exists that examines the importance of dependence for relationships (ANDERSON/NARUS 1990; HALLÉN/JOHANSON/SEYED-MOHAMED 1991; HEIDE 1994; LUSCH/BROWN 1996), this study focuses on behavioral and attitudinal aspects of cooperative relation-

3.2 Introduction to selected theories

69

ships only and does not aim at analyzing the effects of power and dependence in this context. This isolated examination of the effects of selected aspects of exchange relationships will foster a better understanding of the matter as a whole. Once this is accomplished, future research can add further complexity and integrate additional variables such as power and dependence. Only when the causal linkages between the variables introduced above will be further understood also in logistics relationships, the performance drivers in logistics outsourcing arrangements can be identified. This would consequently enable the adequate design of these relationships in the future. 3.2.3 Commitment – trust theory The commitment-trust theory has its roots in the work of MORGAN/HUNT (1994). It is a rather young theory which, based on social exchange theory, views commitment and trust as central elements of exchange relationships while at the same time integrating opportunism into the theory whose implicit exclusion had been a major point of criticism for social exchange theory. Origin of the thoughts of MORGAN/HUNT was the observation that marketing must distinguish between discrete transactions and relational exchange, which in the time of arm’s length relationships commonly was neglected (DWYER/SCHURR/OH 1987, p. 13; MORGAN/HUNT 1994, p. 21; HOYT/HUQ 2000, p. 750). As relational exchange is becoming more important, relationship marketing is required to adequately address partnership issues. MORGAN/HUNT (1994, p. 22) define: “Relationship marketing refers to all marketing activities directed toward establishing, developing, and maintaining successful relational exchanges”. Since the commitment-trust theory therefore enables a deeper insight into the formation of successful exchange relationships even beyond the pure marketing considerations, it presumably is useful for the understanding of logistics relationships and will thus be introduced in greater detail in the following chapter. 3.2.3.1 Basic concepts

Before the commitment-trust theory further emphasized relationship variables as key determinants for successful relationships, the political economy paradigm dominated the discussion. As THORELLI (1986, p. 38) maintained, “power is the central concept in network analysis” since its “mere

70

3 Theoretical framework

existence […] is often sufficient to condition others”. However, since SHERMAN/SOOKDEO (1992, pp. 77-78) point out that roughly one-third of ventures such as strategic alliances are outright failures, MORGAN/HUNT (1994, p. 22) argue that it must be understood which factors truly distinguish productive and effective relational exchanges from those that are unproductive and ineffective. While they state that no doubt exists that many contextual factors contribute to the success and failure of specific relationship marketing efforts, MORGAN/HUNT (1994, p. 22) theorize that the “presence of relationship commitment and trust is central to successful relationship marketing, not power and its ability to ‘condition others’”. The relevance of commitment and trust, which are key to lasting relationship performance, according to MORGAN/HUNT (1994 p. 22) is explained by three facts. The first is that the variables encourage partners to cooperate more closely in order to preserve relation specific investments that ex-ante have been made. Furthermore, commitment and trust help parties to resist attractive short-term alternatives in favor of the expected longterm benefits received when staying with the existing partner. And lastly, commitment and trust in the relationship will enable partners to view high risk actions of one partner as being prudent since they need not fear opportunistic behavior. It can therefore be observed that in relationships where both trust and commitment are present, they produce outcomes promoting efficiency, productivity and effectiveness. This of course has immediate performance implications, since, as MORGAN/HUNT (1994, p. 22) observe, ”commitment and trust lead directly to cooperative behaviors that are conducive to relationship marketing success”. These two thus are key variables for the relationship marketing model which focuses on one party in the relational exchange and this party’s trust and commitment. Because MORGAN/HUNT (1994, pp. 22-27) hypothesize that relationship commitment and trust are key constructs, they position them as mediating variables, between five antecedents and five outcomes (see Figure 3-2). The antecedents relationship termination costs, relationship benefits, shared values, communication and opportunistic behavior determine the level of commitment and trust experienced by the party in the relationship. Depending on these respective levels, the outcome variables acquiescence, propensity to leave, cooperation, functional conflict, and uncertainty are influenced which in turn are important precursors for successful relationships.

3.2 Introduction to selected theories Termination Costs

Acquiescence +

+ Relationship Benefits

_

+

Propensity to Leave

Relationship Commitment

+ Shared Values

71

+

+

+

+

Trust

+ Communication _ Opportunistic Behavior

Cooperation

+

_

Functional Conflict

Uncertainty

Fig. 3-2. Key mediating variables model of relationship marketing (Morgan/Hunt 1994, p. 22)

Lastly, the theory also emphasizes that trust is a major determinant of relationship commitment (ACHROL 1991, p. 89; MORGAN/HUNT 1994, p. 24), because relationships characterized by trust are so highly valued that parties will commit themselves to such relationships (HREBINIAK 1974). 3.2.3.2 Explanatory value for logistics relationship research

Due to its theoretical proximity to social exchange theory, many of the arguments proposed in chapter 3.2.2.2 concerning the explanatory value for logistics relationship research also apply to the commitment-trust theory. Following the argumentation of a large number of authors (DWYER/ SCHURR/OH 1987; ANDERSON/HAKANSSON/JOHANSON 1994; MORGAN/ HUNT 1994; RING/VAN DE VEN 1994; WILSON 1995; LAMBE/WITTMANN/SPEKMANN 2001), little doubt exists that commitment and trust are central variables determining the success of exchange relationships. Since contract logistics relationships as analyzed in this study are especially characterized through a long-term orientation, based on cooperative rather than arm’s length arrangements, the propositions of the commitment-trust theory are particularly relevant (BOWERSOX/DAUGHERTY/DROGE/ ROGERS/WARDLOW 1989; LALONDE/ COOPER 1989; MOORE 1998; LAMBERT/EMMELHAINZ/GARDNER 1999; MALONI/BENTON 2000; KNEMEYER/CORSI/ MURPHY 2003; WALLENBURG 2004). The most remarkable achievement of MORGAN/HUNT (1994) is the integration of the mediating variables commitment and trust with several antecedents and outcomes into a model which, empirically validated, allows deep insights into the designing of successful relationships.

72

3 Theoretical framework

As stated above, commitment and trust certainly are of substantial importance also to logistics relationship research. Among their antecedents, especially shared values, communication and opportunistic behavior are relevant for the analysis of the factors that influence logistics outsourcing relationships and consequently lead to outsourcing performance. MORGAN/HUNT (1994, p. 30) show empirically that shared values increase both commitment and trust while communication efforts and opportunistic behavior influence the trust of the party. For logistics outsourcing relationships it can consequently be concluded that shared values, communication and opportunistic behavior have strong influences on the success of the relationships. The same is true for the variables cooperation and functional conflict. Shown empirically by MORGAN/HUNT (1994, p. 30) to be outcomes of commitment and trust (cooperation) or trust alone (functional conflict) they are also influencing the success of customer-LSP relationships since cooperative behavior in long-term relationships is promising to be more rewarding for both parties than adversarial conduct at arm’s length. MORGAN/HUNT (1994, p. 22) propose five other variables as part of the key mediating variable model either as antecedent or as outcome. Even though they are of importance for understanding the mechanisms of relationship marketing, their explanatory value for logistics relationship research is limited. Uncertainty is, as explained in chapter 3.2.1.2, a contingency factor that is indeed influenced by trust, but at the same time promises a higher explanatory value when utilized as a moderating factor to derive conclusions for the entire relationship. Acquiescence, propensity to leave, and relationship termination costs are variables that have been extensively analyzed in customer loyalty research (WALLENBURG 2004) and will not be focused in this study aiming at understanding the mechanisms of successful logistics outsourcing relationships. Lastly, the relationship benefits modelled by MORGAN/HUNT (1994, p. 22) as an antecedent of relationship commitment, in logistics relationships take on a different role. MORGAN/HUNT understand them as the benefits gained from the relationship as compared to an alternative. In logistics outsourcing relationships, these relationship benefits must be modelled as the outcome or the performance of the outsourcing arrangement and therefore will be part of the performance variable “goal achievement”. 3.2.4 Contingency approach The previous chapters have already shown the importance of exchange relationships. Moreover, it has been argued that since not all relationships

3.2 Introduction to selected theories

73

are the same, an influence of contingency variables may exist that determines the performance of relationships in different contexts. Exemplary are the different transaction characteristics presented in chapter 3.2.1.2. It can therefore be presumed that the performance of logistics relationships is subject to moderating effects of different logistical contingency variables. A possible explanation promising deeper insights into the matter is being offered by the contingency approach (also sometimes: contingency theory) which has been extensively used to describe organizational structures under varying environmental conditions (KLEER 1991, p. 115; PRASAD/TATA/MOTWANI 2001, p. 32). However, in order to understand the explanatory value of the contingency approach for logistics relationship research, it will be discussed – along with its strengths and limitations – in the following chapters. 3.2.4.1 Basic concepts

The origin of the contingency approach was the observation in the 1960’s that organizations show significantly different organizational structures and behaviors. Opposing the dominant belief at the time whereby all organizations could be described by a few universal laws and theories, UDY (1959) and HALL (1963) empirically found that substantial differences between organizations exist. Contingency theory was the attempt of organizational theory to explain these differences (KIESER/KUBICEK 1978, p. 105). The central proposition of contingency theory is, aside from the assumptions that there is no best way to organize and that any way of organizing is not equally effective (GALBRAITH 1973, p. 2), that differences in structure and administrative practices of organizations depend on the nature of the environment to which the organization relates (SCOTT 1992, p. 89). Hence, different forms of organization will have varying degrees of efficiency depending on the environmental conditions. These demands of the environment are also called context and, more specific, contingency or situational variables. Contingency variables encompass all potential factors that can internally or externally affect an organization. Among the most discussed variables in the early days of theory development were the size of the organization and the utilized production technologies as internal factors. Results of the debate were different schools of thought (KIESER/KUBICEK 1978, pp. 106108). Some of the initial findings include that larger organizations with many employees are more bureaucratic than smaller organizations (CAPLOW 1956; RUSHING 1966), that the choice of the production technology has an impact on the organization form (WOODWARD 1958) and

74

3 Theoretical framework

that the standardization of workflows increases the level of bureaucracy within the organization (PERROW 1970). In addition to the internal variables size and technology, especially the influence of the environment on the organization has been in the focus of numerous publications (BURNS/STALKER 1961; LAWRENCE/LORSCH 1967; THOMPSON 1967; KIESER 1974; CHILD 1984). Representatives of this school of thought argue that structural differences of organizations result from environmental differences in general and different environmental dynamics in particular. BURNS/STALKER (1961) for instance could show that organic organizational structures, characterized by large degrees of flexibility and innovation, are superior to other organizational structures in dynamic environments. Only later and considerably influenced by the Aston program (PUGH 1981; PUGH/PHEYSEY 1992), the research focus changed from analyzing one contingency variable at a time to analyzing multiple contingency variables simultaneously. The turning away from mono-causal towards multicausal models resulted after the different schools of thought had all been able to contribute to the explanation of differences between organization structures and administrative practices (KIESER/KUBICEK 1978, p. 108). The research conducted on the basis of contingency theory did not only focus on the influence of one or multiple contingency variables on the structure of organizations. As shown by BURNS/STALKER (1961), the explanation of the influence of different organization structures and administrative practices on the efficiency and effectiveness of organizations is another aim of the contingency approach. An organization is the more efficient and effective, the more appropriate the organizational structure and administrative practices are in the face of relevant situational factors (TOSI/ALDAG/STOREY 1973, p. 27; KIESER/KUBICEK 1978, p. 112). Summing up, the contingency theory according to SCHREYÖGG (1980, p. 308) proposes that “there is an objective necessity to adapt the system (the organization’s structure) congruently to its critical environmental fields in order to secure its survival by achieving the critical level of performance”. The designer of the organization therefore has to comply, at least in the long run, with the constraints which characterize the relationship between environment and organization structure in the view of the contingency approach as deterministic (CHILD 1973, p. 247; PFEFFER/SALANCIK 1978, pp. 225-228). Despite its popularity the contingency approach has been widely criticized (SCHREYÖGG 1980, pp. 309-312). A central point of criticism was the mechanistic view that there is only one best structural answer to a specific contextual situation. In fact, many studies show that a variety of structures exist which enable success in a given environment (SCHREYÖGG

3.2 Introduction to selected theories

75

1980, p. 309). The premises that the organization has full power to change the internal organizational structure at will and at the same time must accept the environment as given, without any possibility of influencing or controlling it, further aggravate the criticism. More disapproval was voiced due to the lack of theory in the contingency approach (HAGE 1974, pp. 18-19). The approach indeed does not offer law-like generalizations. It rather points out that under certain circumstances, differences in organization structure and administrative practices have different implications for a firms’ performance. A last point of criticism of the contingency approach is that it has too often been used in a “quasi-causal” way (KIESER/KUBICEK 1978, p. 135) to ex-post rationalize exploratory results that ex-ante were not proposed by theory-backed hypotheses. STAEHLE (1990, p. 51) therefore concludes that the contingency approach can arbitrarily be used for contents of the most different kinds which is its strength and weakness at the same time, but therefore also constitutes a research approach rather than a theory. After having introduced the contingency approach it must be concluded that its main contribution is its demand to evaluate the structure of organizations and its administrative practices on the grounds of the respective context. The approach therefore must be understood as an idea or a framework rather than a theory with all its implications. 3.2.4.2 Explanatory value for logistics relationship research

As shown above, the contingency approach is suited to explain the influences of different organizational structures and administrative practices on the effectiveness and efficiency of organizations. Since the relationship between a logistics service provider and its customer is an integral part of the customer’s organization structure, the contingency approach is expected to have explanatory power. This will be especially so if the approach it is not used mechanistically but rather as a means to understand the contingencies that potentially influence the relationship. CHOW/HEAVER/HENRIKSSON (1994, p. 26) criticize that the use of contingency models in logistics research has been sparse. They propose that a contingency model – based on the supposition that the fit between logistics organization and strategy and the organization’s environment, product line, production technology, and size will influence performance outcome variables – will overcome the deficits of the widely accepted “one-best-way” paradigm. This demand falls in line with the findings of KNEMEYER/CORSI/ MURPHY (2003, p. 79) who on the basis of the works of LAMBERT/ EMMELHAINZ/GARDNER (1996b) and LAMBERT/EMMELHAINZ/GARDNER

76

3 Theoretical framework

(1999) state that “it should be noted that no particular kind of partnership is better or worse than any other. The key is to obtain the kind of partnership that is most appropriate given the business situation”. Even though not explicitly stated for logistics relationships, KNEMEYER/CORSI/ MURPHY (2003, p. 79) acknowledge the need for an evaluation of the business environment in order to determine the characteristics of the most efficient relationship between two or more parties. After having established the need for the analysis of situational variables also in logistics relationship research, it must be observed that very little research exist on this topic. Notable exceptions are the articles by PERSSON (1997) and PFOHL/ZÖLLNER (1997). PERSSON (1997, p. 286) argues that the writings of LAWRENCE/LORSCH (1967), proposing the two dimensions “differentiation” and “integration” to describe and explain differences in organizational structure, have a substantial meaning for logistics research. Differentiation is defined as the difference in cognitive and emotional orientation among managers in different functional departments while integration refers to the quality of collaboration existing among departments which are cooperating to better address environmental needs the organization is facing. According to PERSSON (1997, p. 286), these concepts ought to appeal to logistics scholars since logistics activities in practice are often scattered among different functional departments for the purpose of functional specialization. It must be noted, however, that this perception is strongly rooted in the belief that logistics services are always provided in-house. In the age of increasing logistics outsourcing, the relationship between the customer and the LSP is gaining ever growing importance and must therefore be the subject of analysis as established in the research questions of this study (see chapter 2.4.2). In their article, PFOHL/ZÖLLNER (1997, pp. 307-310) identify various contingency factors that ought to influence the organization of logistics. While they do not explicitly mention their relevance for the analysis of outsourcing relationships, a large number of them promise to have explanatory value. Environmental relations of the organization are one group of these contingency factors. Since all logistics decisions on the strategic, tactical and operational level depend on environmental information for the design of adequate responses, they can be seen as the situational variables influencing the organizational structure. PFOHL/ZÖLLNER (1997, pp. 307-308) identify the criteria of “complexity and dynamics of environmental relations” as the elements of the environment which are particularly relevant for logistics and provide a large number individual factors (see Figure 33).

3.2 Introduction to selected theories

77

Further contingency variables are the products of the organization and the degree of homogeneity among them. The products determines which organizational measures have to be taken to produce and deliver the products within a certain time limit while the homogeneity among the products in their logistically relevant features is a key factor for designing logistics processes (KIRSCH/BAMBERGER/GABELE/KLEIN 1973, p. 349; COYLE/BARDI/LANGLEY 1992, pp. 383-413). An example for those logistically relevant product features is the degree to which logistical equipment can be used for order processing, transportation, handling, storing, and packaging of all of a firm’s products. Flow of products Complexity of the environmental relations - Number or amount of raw and auxiliary products and operating supplies, and trade goods to be bought - Number or amount of finished products, semifinished products (spare parts) and trade goods - Number of sources of supply - Number of customers to be supplied - Number of deliveries - Variety of transportation, storage and handling procedures for the suppliers and distributed products - Geographical distribution of suppliers and customers

Dynamics of the environmental relations Rate and regularity of change: - The time of delivery and the amount in supply and demand delivered - Channels of procurement and distribution demand - The structure of the suppliers and customers

Flow of information Complexity of the environmental relations Amount of necessary information on actual and potential suppliers and customers -

Location Amount of products in supply and demand Delivery and order cycle Places of supply and support Transportation and packaging requirements

Dynamics of the environmental relations Rate and regularity of change in the necessary information on: - Location - Amounts of products in supply and [sic!] - Delivery and order cycle - Places of supply and support - Transportation and packaging requirements

Fig. 3-3. Relations between the organization and the environment from a logistical viewpoint (PFOHL/ZÖLLNER 1997, p. 308)

PFOHL/ZÖLLNER (1997, p. 308) furthermore propose the contingency factor “degree of homogeneity in the market”. While the same market can be supplied with a variety of finished products, this may lead to diversification and consequently to totally different channels of distribution. The measure for the diversity is the proposed homogeneity, which according to PFOHL/ZÖLLNER (1997) shows to which extent logistical capacities such as order processing, packaging, transportation or handling can be combined on the way to the customer. The more homogeneity in the market of the firm, the easier will it be to design the logistics organization according to the market’s needs.

78

3 Theoretical framework

A further contingency factor is the production technology of the organization (PFOHL/ZÖLLNER 1997, p. 309). They argue that its influence on logistics tasks can be determined by the spatial arrangement of machines which can vary in different states between job-shop and flow shop production. Thus, the more complex the production technology and the associated storage and handling processes, the higher are the demands on the logistics processes. A last contingency factor proposed by PFOHL/ZÖLLNER (1997, p. 310) is the size of the organization. As they argue, the often used criterion of the number of employees does not have any direct link with the integration of logistical tasks into the organization. Rather, they suggest that number of production plants, warehouses and the logistically relevant dependencies between them indicate the size of the organization. The higher the number of plants and warehouses and the more complex the dependencies between them are, the more complex will be the logistics processes and the more demanding will be their design. KLEER (1991, pp. 121-124) identifies similar contingency variables like PFOHL/ZÖLLNER (1997) but utilizes the approach by KIESER/KUBICEK (1983, p. 222) to distinguish between external and internal contingency variables. According to them, external contingency variables are those suited to explain differences between organizational structures and cannot be altered by the organization alone, but depend on other organizations and thereby describe the relationship of the organization to the environment. Internal contingency variables on the other hand are those the organization can influence by itself. The contingency variables proposed by KLEER (1991) are explicitly developed to cover the situational factors important for relationships between LSPs and their customers. Since those relationships are in the focus of this research, these variables are more appropriate in this context than those proposed by PFOHL/ZÖLLNER (1997) a decade earlier which mainly focus on the internal logistics organization. External contingency variables according to KLEER (1991, pp. 121-122) include the complexity and the dynamics of the environmental relations. The complexity is driven by the number of different locations the products are delivered to, the customer structure, the homogeneity of buying patterns with respect to ordinary orders and seasonal differences and the number of LSPs in service. The dynamics are determined by changes in the competition or the customer structure resulting especially from increasing competitive pressure among all parties. However, the dynamics can also be influenced by the changing importance of logistics in the eyes of the customers the LSP’s customer. A last external contingency variable put forward by KLEER (1991, p. 122) is the industry of the customer. Certain

3.3 Theory integration

79

developments of firms may be characteristic for entire industries, e.g. as concentration tendencies lead to increased buying power of firms, entire channels of distribution may change. Therefore, the industry may be suited to give an insight into developments typical for an entire sector of firms. Counterparts to the external variables are the internal contingency factors. According to KLEER (1991, pp. 122-123) these include the products and its features relevant for logistics, the size of the LSP’s customer and the degree of centralization of the logistics activities. A number of the contingency variables introduced in this chapter have explanatory value for logistics relationship research. Since a concentration is necessary in order to answer research question four (see chapter 2.4.2) and to adequately focus on the most important aspects of the issue, chapter 4.5 will see a selection of these contingency variables and their analysis.

3.3 Theory integration The preceding discussion of chapter 3.2 raises the question whether or not it is necessary to resort to four different theories in order to analyze logistics outsourcing relationships. Furthermore, it must be examined which explanatory value the combination of these theories has to offer with respect to the research aims of this study. This chapter will address these questions in the following. Some studies analyzing inter-organizational relationships, both empirically and conceptually, have weaknesses in their theoretical foundation because they do not adequately address the suitability of the utilized theories for the analyzed context. Omitting this argumentation means assuming a generalizability of the results beyond the context the research was conducted in, which should be avoided in order to prevent false interpretations and conclusions. As already mentioned above and discussed further in chapter 5.1, the objects of research in this study are logistics outsourcing relationships between logistics service providers and their customers. Since a number of different relationship designs exist, the focus here will be on long-term relationships and their cooperative aspects as present in contract logistics arrangements (see chapter 2.3.3.2). Especially suited to analyze long-term logistics outsourcing relationships is the commitment-trust theory. It explicitly centers on the analysis of long-term relationships between firms and allows insights into the factors which determine outcome and performance of these arrangements.

80

3 Theoretical framework

Transaction cost theory also contributes to the understanding of the matter. While its main focus is the analysis of the boundaries of the firm and the decision of the organization between market, hierarchy, and cooperation, it also permits the study of different forms of organization across the supply chain. This naturally includes long-term relationships in the logistics outsourcing context. A less obvious theory for the research on logistics outsourcing relationships is the social exchange theory. Superficially, it may seem inadequate since it was originally developed to understand the relationships between individuals and not between organizations. However, since organizations as such cannot make decisions, single human beings or groups of them act as their representatives (WALLENBURG 2004, p. 84). This argument also holds in industrial buying situations. For the people involved, the business relationship is at the same time a social relationship. In the context of this research, those are the logistics responsibles of the customer and their counterparts at the LSP. This of course does not imply that the logistics outsourcing relationship is influenced only by subjective criteria and personal sympathy. It rather displays that additional to purely economic considerations, social aspects of the relationship are of importance as well. Therefore, social exchange theory is adequately suited for the analysis of logistics outsourcing relationships. The contingency approach has been widely criticized for its quasimechanistic approach which proposes that for every context, one single best organizational structure can be found. The criticism concerning this approach (see chapter 3.2.4.1) and its underlying assumptions such as full rationality seem to be by all means justified considering studies where false interpretations of empirical evidence have led to questionable conclusions. The use of the contingency approach nevertheless offers valuable insights into the situational factors of logistics outsourcing relationships and their performance effects. It therefore promises to have value for the answering of research question four. Since in the context of this research question, it will explicitly be refrained from mono-causal interpretations and the accompanying desire to identify the single best organization structure for every context, the use of the contingency approach seems nevertheless justified. Because of the generally very low level of knowledge on logistics outsourcing relationships, it will rather be adopted to identify relevant contingency variables influencing the relationship. These variables can afterwards be utilized to analyze the performance of logistics outsourcing relationship arrangements in different contexts, however, without the demand to have found the only relationship design suitable for the respective context. It thus rather enables a first orientation for the design of logistics outsourcing relationships. Through this procedure, the

3.3 Theory integration

81

strengths of the contingency approach will be used while its main weaknesses have no effect in this context. The general suitability of the theories presented above for analyzing logistics outsourcing relationships does not necessarily constitute the imperative to actually use them all in this study. The necessity to do so rather results from the complexity inherent in the relationships between customers and LSPs. To address this complexity, the theories have been explicitly chosen collectively to develop a model of logistics outsourcing relationships that aims at identifying from different perspectives the relevant factors influencing the performance of outsourcing relationships after the outsourcing decision has been made. An important theory to explain the formation of cooperative relationships is transaction cost theory which offers the economic approach for analyzing logistics outsourcing relationships. Its main explanatory value in the context of this study is the support in identifying factors that enable partner firms to reduce transaction costs in the outsourcing relationship, such as involvement and openness or that increase them, such as opportunism. However, transaction cost theory on its own cannot be sufficient to answer the research questions posed in chapter 2.4.2 for two reasons. First, existing relationships between customers and LSPs are analyzed for which the decision between market, cooperation, and hierarchy has already been taken in favor of cooperation. Second, these relationships are not based on purely economical considerations only, but also involve social factors. To identify these, transaction cost theory can contribute, but if used on its own is not enough. To identify the social factors and the causal linkages between them, theories with an explicit social perspective are necessary. Social exchange theory is particularly suited since its propositions for exchange relationships can readily be adapted to logistics outsourcing relationships. Among its main contribution are the factors commitment and trust, cooperation, and communication as well as the insight that the more social exchange evolves through the behavior of the parties involved, the more stable and successful is the relationship. The character of the relationship is also displayed in the commitmenttrust theory which is of significant importance in this study because it explicitly addresses long-term business relationships. By doing so, it suggests a number of important factors beyond the central variables commitment and trust that also influence outsourcing relationships such as shared values, functional conflict, and opportunism and emphasizes, as does social exchange theory, the importance of communication and cooperation.

82

3 Theoretical framework

In line with the argumentation of WALLENBURG (2004), these three theories therefore are suited to be combined in order to develop a model addressing the performance of logistics outsourcing relationships, comprising both economical and social considerations. It could be criticized that transaction cost theory and social exchange theory are both used in the same model since the former assumes universal opportunism while the latter ignores it. Just for this reason, marketing research generally supports this procedure (LAMBE/WITTMANN/SPEKMANN 2001), because both assumptions conflict with the reality which sees opportunism between the two extremes. It can therefore be summarized that transaction cost-, social exchangeand commitment trust theories allow insights into how logistics outsourcing relationships ought to be designed. However, they alone cannot answer which situational factors affect the relationship variables. Some basic factors can be contributed by transaction cost theory. After all, it proposes that the organizational form of “cooperation” has been chosen instead of market or hierarchy because of the individual values of the transaction characteristics asset specificity, uncertainty, and frequency. These therefore can be understood as contingency variables because their value according to the theory directly influences the level of the transaction costs and therefore determines the level and the characteristics of the cooperation or relationship. Further situational variables are supplied by the contingency approach. It allows the identification of both internal and external factors that influence the performance of logistics outsourcing relationships. Some of these factors include external factors such as environmental uncertainty, composed of environmental complexity and environmental dynamics, the importance of logistics for the customer, and internal factors like the size of the firm, the product range or the centralization of the logistics function.

4 Antecedents and effects of logistics outsourcing performance

After this previous chapter has consolidated the theories and argued their suitability, the following chapters will identify the relevant variables for logistics outsourcing relationships and introduce hypotheses on their causal linkages and performance effects. The previous chapter has pointed out the explanatory value of four different theories for the designing of logistics outsourcing relationships and has indicated that a number of variables affect their performance outcomes. The following chapters will explore these direct performance effects of logistics outsourcing relationships and will identify their antecedents on the basis of the proposed theories and previous research. Aside from the conceptualization, hypotheses will be generated that address the expected direct and indirect effects of these variables on outsourcing performance. In a second step, hypotheses on the anticipated effects of logistics outsourcing performance on logistics performance and firm performance will be presented after the relevant variables will have been established. Finally, it will be analyzed which moderating effects can be expected from both external and internal contingency factors on the models developed on the basis of the different hypotheses.

4.1 Performance of logistics outsourcing relationships The following two chapters will in detail discuss logistics outsourcing performance. At first, chapter 4.1.1 will present the background of the concept of logistics outsourcing performance, its connection to logistics performance and will argue for measuring it by focusing on its outcome. Then, chapter 4.1.2 will conceptualize the construct on the basis of the understanding developed before and introduce its bi-dimensionality.

84

4 Antecedents and effects of logistics outsourcing performance

4.1.1 Background of logistics outsourcing performance As it will be discussed, logistics outsourcing performance is an important antecedent of logistics performance. The latter has been studied by a large number of logistics researchers, who have defined and measured performance in many different ways (CHOW/HEAVER/HENRIKSSON 1994) and thereby provided a valuable starting point for the analysis of logistics outsourcing performance. As CHOW/HEAVER/HENRIKSSON (1995, p. 296) point out, logistics performance is multi-dimensional, reflecting multiple stakeholders and interests. Therefore, the possible desired outcomes are numerous and range from customer satisfaction over environmental responsibility, to overall cost-effectiveness. Important works on the topic of logistics performance include those of MENTZER/KONRAD (1991), CHOW/HEAVER/HENRIKSSON (1994), GASSENHEIMER/STERLING/ROBICHEAUX (1996), STANK/GOLDSBY/VICKERY (1999), DEHLER (2001), STANK/KELLER/DAUGHERTY (2001), STANK/GOLDSBY/VICKERY/SAVITSKIE (2003), KNEMEYER/ MURPHY (2004) and ENGELBRECHT (2004). Logistics performance, which according to CHOW/HEAVER/HENRIKSSON (1995, p. 296) in research is predominantly measured with “soft” perceptual indicators given the difficulty of obtaining “hard” performance measures, is a result of two different variables: on the one hand, it is influenced by the performance of logistics processes performed in-house under the direct responsibility of the LSP’s customer. On the other hand and of particular importance in the context of this research, it is affected by the performance of outsourcing arrangements in which the customer has delegated logistics and other relevant processes and the accompanying responsibility to a logistics service provider. The performance of these outsourced processes, hereafter termed logistics outsourcing performance, is an important strategic issue which has received little attention so far in logistics research. Notable exceptions include STANK/GOLDSBY/VICKERY/SAVITSKIE (2003), KNEMEYER/ MURPHY (2004) and ENGELBRECHT (2004). These authors propose that successful logistics outsourcing can only be realized if the performance of the outsourcing arrangements can be adequately measured. WEBER (2003, pp. 11-12) suggests to utilize the basic economical model comparing input and outcome of logistics processes to measure performance. Building on the work of GUTENBERG (1957), who made this productivity oriented view popular in Germany, WEBER (2003) proposes to structure processes in the four dimensions input, process, output, and outcome. Those were adapted for a general logistics context by WEBER (1986), but

4.1 Performance of logistics outsourcing relationships

85

can also be used to analyze the performance of outsourced logistics services: The input reflects the consumption of materials and services used in logistics processes. According to WEBER (2003, p. 12), it is often implicitly assumed that lower levels of inputs indicate higher efficiency, when for instance the ratio of logistics costs versus total costs is used as a performance measure. The process relates the costs of the input to the production of the logistics service. Measures, such as cost per delivery or cost of an individual warehouse slot per day, are often used to compare the performance of different processes. While the output connects the process costs to the changes in a products’ logistical attributes, concerning for instance its location at different points in time (WEBER 2002, p. 122), the outcome reflects the important perspective which measures if and to which degree the output has satisfied the needs of the customer. WEBER (2003, p. 12) argues that all four dimensions allow the measurement of performance. While the input, process and output perspectives all relate to costs and therefore to efficiency concerns, only the outcome also takes into account the service impact and thus represents effectiveness. The outcome is primarily relevant for the performance evaluation of the logistics outsourcing relationship between the customer and the LSP. While cost efficiency is a major driver of financial performance for the LSP as it reduces costs and thus increases the bottom line, contracts that ex-ante determine the price of the service for the customer are the norm in contract logistics. Consequently, the only relevant performance measures are those indicating whether the logistics service outcome is at least as expected with respect to quality and agreed upon price. Logistics outsourcing performance in this work therefore will be measured in the light of the outcome-orientation as suggested by WEBER (2003, p. 12) and will include measures of both efficiency and effectiveness. The more detailed conceptualization of the construct, which will be argued to be a two-dimensional, will be presented in the following chapter. 4.1.2 Conceptualization of logistics outsourcing performance A number of authors acknowledge that the nature of logistics outsourcing performance is complex and therefore requires very detailed measurement: STANK/GOLDSBY/VICKERY/SAVITSKIE (2003, pp. 32-33) propose a tridimensional construct that incorporates operational performance as well as cost performance and relational performance as antecedents of customer satisfaction with the outsourcing arrangement. KNEMEYER/MURPHY

86

4 Antecedents and effects of logistics outsourcing performance

(2004, p. 40) view outsourcing performance as equally complex, being tridimensional and consisting of the factors operations performance, channel performance and asset reduction performance. Finally, ENGELBRECHT (2004, p. 212) suggests to measure outsourcing performance bidimensionally through the of facets goal achievement and information exchange. As all three works are very relevant for the conceptualization of logistics outsourcing performance, their individual findings will be presented in detail in the following. Realizing that outsourcing performance depends on several antecedents, STANK/GOLDSBY/VICKERY/SAVITSKIE (2003, pp. 32-33) propose a positive relationship between relational performance and logistics outsourcing performance. Implicitly, they also use an outcome-orientation as presented in chapter 4.1.1 by measuring outsourcing performance as a tridimensional construct of operational-, relational-, and cost performance which all directly affect customer satisfaction. They thereby support the hypotheses that outsourcing performance should be measured through both logistics costs (cost performance) and logistics service indicators (operational performance). The finding that relational performance is an important antecedent of outsourcing performance, which is the underlying supposition of this study, has found very strong support in the work of STANK/GOLDSBY/ VICKERY/SAVITSKIE (2003, p. 43) who can statistically show strong positive influence of relational performance on both operational- and cost performance. This is of particular interest since STANK/GOLDSBY/VICKERY (1999, p. 435) found earlier that literature has provided very little guidance on the relationship between operational and relational performance. The proposition on the importance of relational performance is backed by KNEMEYER/MURPHY (2004, p. 38) who conclude that “previous research has established that long-term relationships between buyers and sellers tend to be more successful than transactional agreements, and that dimensions such as trust and communication can enhance relationship success”. They thereby recognize the importance of relationship variables for the performance of outsourcing arrangements that include a LSP. In their model, viewing trust as the central mediating variable between different relationship factors1 and outsourcing performance, the dependent variable of outsourcing performance is a three-dimensional construct, consisting of operations performance, channel performance and asset reduction performance. This acknowledges the complexity of the issue and indicates

1

Relationship variables include specific investments, opportunistic behavior, prior satisfaction, 3PL reputation and communication.

4.1 Performance of logistics outsourcing relationships

87

that a single construct measure of logistics outsourcing performance would be far too limited (KNEMEYER/MURPHY 2004). In his work on logistics outsourcing, ENGELBRECHT (2004, pp. 212218) uses a similar approach. Realizing that for the successful implementation of outsourcing arrangements several variables play a central role, among them the involvement of the parties, the level of conflict, the proactive improvement of the LSP, the customer’s prior outsourcing experience and the length of the project implementation phase (ENGELBRECHT 2004, pp. 202-212), he suggests addressing the complexity of outsourcing performance by utilizing a two-dimensional outsourcing performance construct. ENGELBRECHT (2004, p. 212) states that the performance of outsourcing arrangements must not only be measured by the achievement of the previously set goals alone. A further dimension should include the qualitative aspects, i.e. whether conflicts can be solved openly and in cooperation with an adequate level of information exchange. While these conceptualizations introduced above certainly have very different foci, they all lead to the conclusion that outsourcing performance is a complex issue, requiring complex measurement. All approaches acknowledge the need for operational excellence of the service provided. Together with the understanding that in contract logistics, the expectations for the operational performance are usually codified in a contract prior to the commencement of the outsourcing arrangement, it becomes apparent that the achievement of the operational excellence must be measured, preferably including its two most important aspects: the quality of the service provided and its costs. This will be done in this study through the construct goal achievement which combines facets of both aspects. To excel in this dimension, the LSP must only fulfill the exact specifications of the contract and the corresponding expectations of the customer. However, outsourcing performance is more complex than just doing business as usual. Beyond the specifications of the contract, the LSP can deliver additional value added by exceeding the expectations of the customer. This cannot be measured through the construct goal achievement which due to its nature can only grasp if all deficits that would have prevented the achievement of the previously agreed upon goals have been addressed by the LSP. But to measure whether the LSP has created additional value added by being customer oriented, innovative and proactive, thereby increasing value in the dimensions of service level increases and cost reductions, the construct goal achievement is not sufficient. Therefore, a second dimension, goal exceedance, is included to address the complexity and measure the degree to which the LSP has created additional value added beyond the original expectations of the customers.

88

4 Antecedents and effects of logistics outsourcing performance

Conceptualizing outsourcing performance as a bi-dimension construct is also justified since for the performance in both dimensions, fundamentally different efforts are necessary. While for the achievement of goals fixed in a contract that both parties have previously agreed upon, it is only necessary to fulfill the exact requirements and expectations, exceeding the goals needs very different efforts. Here, the LSP must strive to surpass the expectations by being customer oriented, innovative, and proactive. It may be argued that both aspects could be measured as two opposing extremes on one continuous scale. However, only through dividing the construct into two separate dimensions, it can be measured whether or not different determinants exists that exhibit distinguishably different influences on goal achievement and goal exceedance. Similarly, only with a bi-dimensional construct, different impacts on the consequent performance dimensions of logistics and firm performance can be analyzed. Before the next chapter will identify and conceptualize the various antecedents of logistics outsourcing performance, it must be noted that its two dimensions have of course a common origin. For reaching higher levels of outsourcing goal exceedance in terms of service improvements and cost reductions, simultaneous improvement of goal achievement must be realized. Therefore, the models about to be developed in subsequent chapters will assume a correlation between the two variables.

4.2 Identification of relevant antecedents The performance of logistics outsourcing relationships as presented in chapter 4.1 is influenced by a number of different antecedents. While for some variables a direct causal linkage on the two dimensions of logistics outsourcing performance can be theoretically hypothesized, several factors have indirect effects only. In the following, the effects will be briefly introduced. A detailed argumentation will follow in the chapter 4.3 where the hypotheses on causal linkages between the different variables will be introduced. As suggested by social exchange, commitment-trust and transaction cost theories, the cooperation between the customer and the LSP has a direct effect on goal achievement and goal exceedance. Through the closer working relationship between the parties, they find it easier on the one hand to reach the goals previously agreed upon, while on the other hand, cooperation facilitates the achieving of goals that would not be obtainable without them working together.

4.2 Identification of relevant antecedents

89

According to both social exchange and commitment-trust theories, the factor of communication between the parties supposedly is also an important facilitator of successful logistics outsourcing arrangements as it leads to closer cooperation and fosters the understanding between the two parties. However, it is only an indirect driver of outsourcing performance via its effect on other variables as it will be argued in detail later. Naturally, running the logistics processes as specified in the agreement between customer and LSP is an indispensable prerequisite for successful relationships. However, as the environment and markets change, outstanding performance effects can be achieved only through the constant improvement of the logistics processes through the LSP in order to increase efficiency and effectiveness. This will be measured by the construct proactive improvement. As suggested by social exchange theory and recent research (ENGELBRECHT 2004, p. 276), it has a direct effect on both goal achievement and goal exceedance. Not only will the attention of the LSP to the processes cause better results in terms of meeting the agreed performance levels, the actions taken by the LSP to improve the logistics processes will eventually also lead to exceeding them. Hypothesized effect on outsourcing performance

Variable

Underlying theories

Cooperation

Social Exchange Theory, Commitment-Trust Theory, Transaction Cost Theory

Communication

Social Exchange Theory, Commitment-Trust Theory

Proactive Improvement

Social Exchange Theory

Trust

Social Exchange Theory, Commitment-Trust Theory

Relationship Commitment

Social Exchange Theory, Commitment-Trust Theory

Functional Conflict

Commitment-Trust Theory

Involvement

Transaction Cost Theory

Opportunism

Transaction Cost Theory, Commitment-Trust Theory

Shared Values

Commitment-Trust Theory

Openness

Transaction Cost Theory

Direct

Indirect

9 9 9 9 9 9 9 9 9 9

Fig. 4-1. Antecedents of the performance of logistics outsourcing relationships

Aside from the variables with direct effects, several important constructs promise to have indirect effects that contribute to the overall explanatory value of the resulting model. Commitment and trust, proposed by social exchange and commitmenttrust theories, will have a number of indirect effects as they foster a closer cooperation between the two parties. Direct effects on the two dimensions of outsourcing performance, however, will not be hypothesized because

90

4 Antecedents and effects of logistics outsourcing performance

they constitute behavioral and attitudinal variables without a direct action component. Only through their influence on mediating variables, such as cooperation, they are assumed to have an effect. Commitment-trust theory also suggests that functional conflict contributes to successful relationships. In addition to its indirect effects via other variables in the model, direct effects on outsourcing performance will be assumed as the functional solving of conflicts directly enables the achievement of goals and their exceedance due to less friction caused by open and unsolved conflicts and a closer relationship. Transaction cost theory allows the presumption that the involvement of the parties in the outsourcing relationship is another facilitator of successful relationships with indirect effects on performance. The involvement will e.g. lead to closer cooperation between the parties and foster the proactive improvement of the LSP, as potential problems and optimization potentials are identified earlier. This equally holds true for the factor openness, whose transaction cost reducing character is beneficial for logistics outsourcing arrangements. While openness alone does not have direct effects on outsourcing performance, its positive influence on the communication between the two parties and the functional solving of conflicts will lead to substantial indirect effects. Further indirect effects can be expected from the variable opportunism, whose detrimental effects on relationships are implied by transaction cost and commitment-trust theories and the shared values which according to the commitment-trust theory nurture cooperative and mutually successful relationships. The following chapters will in detail conceptualize the variables that have direct and indirect effects on logistics outsourcing performance. Subsequently, hypotheses on the causal linkages between them will be presented which later will allow the formulation of a model, thus answering research questions one and two (see chapter 2.4.2). 4.2.1 Conceptualization of variables 4.2.1.1 Cooperation

Cooperation has been widely discussed in the literature on customersupplier relationships (ANDERSON/NARUS 1990; METCALF/FREAR/ KRISHNAN 1992; MORGAN/HUNT 1994; WILSON 1995; LAMBE/WITTMANN/SPEKMANN 2001). It refers to situations in which parties work closely together to achieve mutual goals. According to ANDERSON/NARUS (1990, p. 45) cooperation is defined as “similar or complementary coordinated actions taken by firms in interdependent relationships to achieve

4.2 Identification of relevant antecedents

91

mutual outcomes or singular outcomes with expected reciprocation over time”. While coordination, which implies cooperation, has been known to be essential in such areas as channels of distribution for decades, the literature on relationships has focused primarily on power and conflict as focal constructs (MORGAN/HUNT 1994, p. 26). This is exemplified in the work of STERN/EL-ANSARY (1992, p. 312) who point out that “interorganizational coordination is required within a marketing channel”. However, they also state that “power generally must be used in a marketing channel to […] gain cooperation and induce satisfactory role performance”. MORGAN/HUNT (1994, p. 26) conclude that this focus on power results from the absence of a theory that explains cooperation. This deficit, which was addressed already by ALDERSON (1965, p. 239), is partially overcome by the commitment-trust theory which specifically deals with this topic. For the understanding of cooperation in this study the argumentation of MORGAN/HUNT (1994, p. 26) will be followed. They use the definition by ANDERSON/NARUS (1990, p. 45) presented above as a foundation and expand it by emphasizing cooperation’s proactive nature as opposed to the power-related view which involves being coerced into taking interdependent actions (WILSON 1995, p. 338). Cooperation furthermore is characterized as the joint striving toward individual and mutual goals (SCHERMERHORN 1975, p. 847; STERN/REVE 1980, p. 57; BROWN 1981, p. 7) with the implicit or explicit desire to achieve benefits that would not be obtainable without two parties working together. Cooperation thus is not only the inverse of conflict, as it has been conceptualized in earlier research (PEARSON/MONOKY 1976; GATTORNA 1978; ROSS/LUSCH/BROWN 1982), but rather involves actions such as collaborative goal setting, trust, detailed communication, teamwork and unity of purpose (LARSON/ KULCHITSKY 1999, p. 94). For the partners it has according to LAMBE/ WITTMANN/SPEKMANN (2001, p. 23) the advantage that “cooperative behaviors develop and serve to promote greater benefits for the exchange partners”. Cooperation in this study therefore is defined as the joint striving of firms in interdependent relationships towards individual and mutual goals with the desire to achieve benefits which would not be obtainable without the two parties working together. 4.2.1.2 Communication

Information exchange and communication are key constructs in many empirical studies on interorganizational exchange relationships (ANDERSON/ NARUS 1984; FRAZIER/SUMMERS 1984; ANDERSON/NARUS 1990; HEIDE/

92

4 Antecedents and effects of logistics outsourcing performance

JOHN 1992; HEIDE/MINER 1992; MORGAN/HUNT 1994; ELLRAM/ HENDRICK 1995; MOORE 1998; KNEMEYER/CORSI/MURPHY 2003; KNEMEYER/MURPHY 2004; ENGELBRECHT 2004). Following ANDERSON/ NARUS (1990, p. 44) communication can be defined “broadly as the formal as well as informal sharing of meaningful and timely information between firms”. This definition focuses on the efficacy of information exchange rather than the mere quantity or amount, thus going beyond simply exchanging information and aiming at grasping the quality and intensity of the past communication experience between firms. Successful partnerships tend to exhibit effective communication that includes high levels of communication quality, e.g. accurate communication, information sharing about the partners’ changing needs, and joint planning and goal setting (CLAYCOMB/FRANKWICK 2004). The more frequent, intense, and diverse the communication between exchange partners is, the more likely is a buyer-supplier relationship to survive (KENIS/KNOKE 2002). BOWERSOX/DAUGHERTY/DROGE/ROGERS/WARDLOW (1989) believe that a complete and open exchange of operating and strategic information holds logistics alliances together, because joint performance towards shared goals requires open disclosure. Communication, characterized as the complete exchange of information, they argue to be essential to assure that operations of the user and the LSP are synchronized. Communication in this study therefore is understood as the formal as well as informal sharing of meaningful and timely information between firms which is suited to achieve the benefits which are desired outcomes of the relationship. 4.2.1.3 Proactive improvement

Innovation is critical to the success of many firms, including logistics service providers (FLINT/LARSSON/GAMMELGAARD/MENTZER 2005, p. 113). Innovation can generally occur within services, processes, or any social system (SCHUMPETER 1934) and is defined by ROGERS (1995, p. 11) as “an idea, practice, or object that is perceived as new by an individual or other unit of adoption”. In several case studies with logistics service providers, FLINT/LARSSON/ GAMMELGAARD/MENTZER (2005, p. 113) found anecdotal evidence that customers expect service providers to continuously drive innovation in order to increase value creation for their customer as a basis for the own sustained competitiveness. This does not come as a surprise, considering the findings presented earlier that have shown the growing demand in the market for more efficient and more effective logistics solutions. A prereq-

4.2 Identification of relevant antecedents

93

uisite to fulfill these new requirements is a general innovation orientation of the LSP, which in this study will be measured by the construct proactive improvement. Proactive improvement has received very little attention in logistics research. Notable exceptions include the works of ENGELBRECHT (2004, pp. 207-209), WALLENBURG (2004, pp. 102-103), WEBER/WALLENBURG (2004) and FLINT/LARSSON/GAMMELGAARD/MENTZER (2005). While the latter develop a logistics innovation process and conclude that logistics innovation still requires substantial further research (FLINT/LARSSON/ GAMMELGAARD/MENTZER 2005, p. 139), ENGELBRECHT (2004, p. 276) finds empirically that proactive improvement efforts of the LSP are a major driver of logistics outsourcing performance. This finding is supported by WALLENBURG (2004, pp. 237-245), who also shows empirically that LSPs can increase customer loyalty by consequent proactive improvement. Since logistics research aside from ENGELBRECHT (2004) and WALLENBURG (2004) has not dealt with the topic, proactive improvement in this study will in line with them be understood as the extent and intensity of the activities employed by a logistics service provider aiming at improving the logistics systems of the customer. As a reason for the current lack of research on the topic, WALLENBURG (2004, p. 102) suggests that it only plays a decisive role in longer-term logistics relationships, which so far have received very little attention as well. This long-term orientation, however, makes proactive improvement a very valuable construct for the research in the context of this study. 4.2.1.4 Trust

The concept of trust has received widespread attention in social psychology (DEUTSCH 1960; LINDSKOLD 1978; LEWICKI/BUNKER 1995), sociology (STRUB/PRIEST 1976; LEWIS/WEIGERT 1985), economics (WILLIAMSON 1981; DASGUPTA 1988), and consequently also in the marketing related research on buyer-seller relationships (DWYER/SCHURR/OH 1987; ANDERSON/WEITZ 1989; MOORMAN/DESHPANDÉ/ZALTMAN 1993; GANESAN 1994; MORGAN/HUNT 1994; DONEY/CANNON 1997; LANE 2000). Trust has diffused from psychology and sociology into business-tobusiness relationship research and according to WILSON (1995, p. 337) has been established as a “fundamental relationship model building block” and therefore is included in most relationship models. However, while most definitions of trust involve the belief of the one party that the other will act in his best interest (WILSON 1995, p. 337), no common definition on trust exists, leaving a rather fragmented picture.

94

4 Antecedents and effects of logistics outsourcing performance

MORGAN/HUNT (1994, p. 23), for example, define trust “as existing when one party has confidence in an exchange partners’ reliability and integrity”. Similarly, MOORMAN/ZALTMAN/DESHPANDÉ (1992, p. 315) state that “trust is defined as a willingness to rely on an exchange partner in whom one has confidence”. Earlier definitions include the one of SCHURR/OZANNE (1985, p. 940), building on the work of BLAU (1964) and ROTTER (1967), defining trust as “the belief that a party’s word or promise is reliable and that a party will fulfill his/her obligations in an exchange relationship”. While the list of different definitions of trust could be continued virtually indefinitely, LANE (2000, p. 3) observes that most concepts share three common elements: First, a degree of interdependence between the trustor and the trustee generally is assumed. Expectations about a partners’ trustworthiness only become relevant when the outcome of the own activities depend on the others prior action or cooperation (LUHMANN 1979; DASGUPTA 1988) Second, the concepts share the assumption that trust provides a way to cope with risk and uncertainty in exchange relationships. Following economic theory, trusting a partner means potentially being exposed to opportunistic behavior, since the response to any action in a social exchange relationship is usually delayed in time. Therefore, the trustor is confronted with an information problem on how the other party will react. This can be overcome through the somewhat risky pre-commitment of trusting the other (LUHMANN 1979, p. 26). Third, the belief is firmly rooted that the vulnerability resulting from the acceptance of risk will not be taken advantage of by the other party in the relationship. This documents the expectation that with the giving of trust, cooperative behavior and reciprocal action are associated. If one party came to doubt this fundamental exchange, it would have detrimental effects on the trust and will ultimately result in a termination of the relationship. While most concepts on trust can be subsumed under these assumptions, they are diverging on the question of the origin of trust. For some scholars, particularly economists like AXELROD (1984), DASGUPTA (1988) and COLEMAN (1990), trust is a decision made under risk and therefore it is the result of a calculative process. Another group, mainly consisting of sociologists and organizational theorists, insist that common values or moral orientations are the foundations of trust (PARSONS 1951; PARSONS 1969). A third group believes in common cognitions as the basis for trust (BLAU 1964; GARFINKEL 1967; SIMMEL 1978; GIDDENS 1984). Cognitions, defined as “rules that constitute the nature of reality and the frames through which meaning is made” (SCOTT 1995, p. 40) are embodied in the

4.2 Identification of relevant antecedents

95

expectations held about the social order in general and specific interactions in particular (LANE 2000, p. 10). According to GARFINKEL (1967, p. 173), the grounds of trust are “expectations of persistence, regularity, order, and stability in the everyday and routine world”. BLAU (1964, p. 93) furthermore states that social exchange requires trusting each other to discharge specified obligations for services received in the past and substituting them with unspecified obligations on future services, leading to the expectation that in some form or the other they will be rewarded. Consequently, social exchange is not seen as a singular, calculative process, but rather as an open-ended conception of reciprocity (POWELL 1991). In an attempt to consolidate the different assumptions of trust, the views on its origin and the accompanying definitions, the following can be observed: trust requires both the will of at least one party to trust the other and additionally depends on the other party’s trustworthy behavior (WALLENBURG 2004, p. 107). Since trustworthy behavior is a complex issue largely depending on the particular context, trust in this study will be conceptualized along the line of the buying party’s will to trust. This willingness to trust corresponds to the above definition of MOORMAN/ZALTMAN/DESHPANDÉ (1992, p. 315) and directly leads to the definition of DONEY/CANNON (1997, p. 36) that trust is the “perceived credibility and benevolence of a target of trust”. This view, originally put forward by GANESAN (1994, p. 3), will be underlying the understanding of trust in this study as it addresses its most relevant aspects in two dimensions: The first dimension of trust focuses on the objective credibility of an exchange partner. It includes the expectation that the partner’s word or written statement can be relied on (LINDSKOLD 1978). The second dimension, benevolence, marks the extent to which one partner is genuinely interested in the other partner’s welfare and motivated to seek joint benefits. This definition of trust is particularly suited for the context of third party logistics relationships. Here, the customer turns to the LSP with some degree of risk concerning the trust that the LSP will be able to perform effectively and efficiently (credibility) and at the same time is interested in the customer’s best interest (benevolence). The construct trust will in this study consequently measure the degree of trust the customer has in the organization of the LSP. 4.2.1.5 Commitment

Relationship commitment is the most common dependent variable used in buyer-seller relationship studies (WILSON 1995, p. 337) and has been employed among others by DWYER/SCHURR/OH (1987), ANDERSON/WEITZ (1992), MOORMAN/ZALTMAN/DESHPANDÉ (1992), MORGAN/HUNT

96

4 Antecedents and effects of logistics outsourcing performance

(1994), GUNDLACH/ACHROL/MENTZER (1995), WILSON (1995), HOCUTT (1998), MOORE (1998), RODRÍGUEZ/WILSON (2002), WONG/SOHAL (2002), KNEMEYER/CORSI/MURPHY (2003) and WALLENBURG (2004). The reason for its popularity is the widespread understanding that “commitment is an essential ingredient for successful long-term relationships” (GUNDLACH/ACHROL/MENTZER 1995, p. 78). It has originally been defined as “an implicit or explicit pledge of relational continuity between exchange partners” (DWYER/SCHURR/OH 1987, p. 19). Corresponding to this view MOORMAN/ZALTMAN/DESHPANDÉ (1992, p. 316) define ”commitment to the relationship […] as an enduring desire to maintain a valued relationship”. While this definitions grasp the very nature of commitment and can serve as a notional basis, it will be understood in this study more precisely as the relationship commitment proposed by MORGAN/HUNT (1994, p. 23). They draw on the conceptualizations of commitment in social exchange (COOK/EMERSON 1978), marriage (THOMPSON/SPANIER 1983), and organizations (MEYER/ALLEN 1984) and define “relationship commitment as an exchange partner believing that an ongoing relationship with another is so important as to warrant maximum efforts at maintaining it; that is, the committed party believes that the relationship is worth working on to ensure that it endures indefinitely” (MORGAN/HUNT 1994, p. 23). While relationship commitment is fairly new in the discussion of interorganizational relationships (MORGAN/HUNT 1994, p. 23), it has long since been viewed as central in the social exchange literature (see chapter 3.2.2.1 and THIBAUT/KELLEY 1959; BLAU 1964). COOK/EMERSON (1978, p. 728) characterize commitment as “a variable we believe to be central in distinguishing social from economic exchange theory”. Commitment is also perceived as a critical variable in the literatures of organizational and buyer behavior. Organizational commitment, which according to MORGAN/HUNT (1994, p. 23) is one type of relationship commitment that is critical to a firms’ internal relationships, is among the oldest (BECKER 1960) and most widely studied variables (REICHERS 1985) in organizational behavior theory. It is seen as central due to its proposed outcomes such as decreased employee turnover (PORTER/STEERS/ MOWDAY/BOULIAN 1974), higher motivation (FARRELL/RUSBULT 1981) and increased organizational citizenship behavior (WILLIAMS/ANDERSON 1991). Commitment implies a willingness to make short-term sacrifices to realize longer-term benefits (DWYER/SCHURR/OH 1987, p. 19). It is believed to have a positive influence on motivation and involvement (MOWDAY/ PORTER/STEERS 1982), loyalty (KANTER 1972), and performance and obedience to organizational policies (ANGLE/PERRY 1981). Furthermore,

4.2 Identification of relevant antecedents

97

according to MACNEIL (1980), it provides a foundation for the development of social norms of governance, which are considered important mechanisms for regulating long-term relational exchanges and reducing opportunism. Also in the field of service relationship marketing, the importance of commitment is realized. BERRY/PARASURAMAN (1991, p. 139) find that “relationships are built on the foundation of mutual commitment”. Similarly, commitment is viewed as positive in the brand loyalty literature, where ASSAEL (1987, p. 665) defines brand loyalty as “commitment to a certain brand” arising from certain positive attitudes. This is important because manufacturers view brand loyalty as key to superior performance and therefore undertake considerable efforts to foster it among their customers (MORGAN/HUNT 1994, p. 23). Overall, a common theme emerges from the various literatures on relationships (MORGAN/HUNT 1994, p. 23): parties involved in interorganizational exchange relationships identify commitment among exchange partners as key for achieving valuable outcomes for themselves and they work hard to develop and maintain this important facet of their relationships. Therefore, relationship commitment is central to relational exchanges between firms and their partners. While this has been postulated by MORGAN/HUNT (1994, p. 23) in general terms for all relationships, it will be proposed to equally hold true for the more specific relationships between customers and LSPs examined in this study. Before concluding it must be remarked that recently the onedimensional view of commitment has been criticized. Especially in the social and organizational psychology literature, multi-dimensional constructs have been suggested to fully grasp the complexity of commitment. MEYER/ALLEN (1984) and MCGEE/FORD (1987) propose to consider the two dimensions of affective and continuance (or cognitive) commitment. ALLEN/MEYER (1990) add normative commitment as a third dimension. These approaches suggest that individuals or organizations remain in relationships either because they want to or because they feel obliged to do so. While this view, which has seen support by several empirical studies (MEYER/ALLEN/SMITH 1993; HACKETT/BYCIO/HAUSDORF 1994) certainly offers detailed insights, it has as yet not been overly embraced by the research on business-to-business exchange relationships. Exceptions include the studies by WETZELS/DE RUYTER/VAN BIRGELEN (1998) and GRUEN/SUMMERS/ACITO (2000) for general exchange relationships and WALLENBURG (2004) for logistics relationships. WALLENBURG (2004, pp. 225-227) finds that only affective commitment has a significant effect on three dimensions of customer loyalty to a LSP, while cognitive commitment was eliminated from the originally proposed model. Therefore, in

98

4 Antecedents and effects of logistics outsourcing performance

line with the research of MORGAN/HUNT (1994) and in consideration of the current and still quite limited state of logistics relationship research, commitment will be viewed in this study as a one-dimensional construct, reflecting its affective characteristics. This also accommodates the general orientation towards voluntary cooperative behavior that serves as a basis for this study. 4.2.1.6 Functional Conflict

Relationships are characterized through the existence of conflict and disagreements: DWYER/SCHURR/OH (1987, p. 24) state that conflict “is predictable within the relationship just as periods of resource scarcity, misperceptions and changing values […] are inevitable”. Conflict, which can be defined as “divergence of goals and role preferences” (DWYER/ SCHURR/OH 1987, p. 24) can have several destructive consequences, such as hostility, bitterness, violence, polarization of third parties, and isolationism. While some authors suggest that conflict can have benefits for a relationship, such as more frequent and effective communication, a critical review of past actions or a more equitable distribution of resources between the parties (ASSAEL 1969; ROSENBERG 1974), the existence of one or more of the negative characteristics mentioned above according to MORGAN/HUNT (1994, p. 26) can ultimately lead to relationship dissolution. However, when disputes and conflicts are resolved amicably, such disagreement can be referred to as “functional conflict” since they prevent stagnation, stimulate interest and curiosity and provide a “medium through which problems can be aired and solutions arrived at” (DEUTSCH 1969, p. 19). This form of conflict management is beneficial for relationships. While a total suppression of conflicts means a relationship has lost its vitality (DWYER/SCHURR/OH 1987, p. 24), for partner firms that use conflicts as a means of “clearing the air” of potentially harmful tensions and ill-will, conflict can have functional and productive consequences (ANDERSON/NARUS 1990, p. 45). Through making conflict functional, the firms in an exchange relationship can maintain cordial relations and give each other the “benefit of doubt” in conflict situations (HARDY/MAGRATH 1988, p. 103). Functional conflict therefore may increase productivity in relationship marketing (MORGAN/HUNT 1994, p. 26) and may be viewed as “just another part of doing business” (ANDERSON/NARUS 1990, p. 45). Functional conflict will in line with the argumentation above in this study by understood as the amicable and functional resolution of disputes and conflicts, leading to beneficial results for the relationship. As such, it

4.2 Identification of relevant antecedents

99

is perceived as and important feature of logistics outsourcing arrangements. 4.2.1.7 Involvement

According to ENGELBRECHT (2004, pp. 203-204) logistics outsourcing arrangements have strong strategic implications for the customer since they directly effect the boundaries of the firm and therefore significantly influence its core competencies and resources. Due to the complexity of logistics outsourcing and its impact on different business processes of the customer, the implementation should be overseen by inter-disciplinary teams (ENGELBRECHT 2004, p. 203) ideally staffed by the relevant employees of both the customer and the LSP. This demand is supported by the finding that interorganizational collaboration is a success factor in outsourcing arrangements and other business-to-business relationships (BOWERSOX/ CLOSS/STANK 2003; MCHUGH/HUMPHREYS/MCIVOR 2003). While one inter-disciplinary team must lead the implementation process of the logistics outsourcing arrangement, individual tasks can be delegated to subordinate teams (ENGELBRECHT 2004, p. 204). However, he also argues that it is crucial that employees of both sides collaborate efficiently and effectively to ensure optimal implementation performance. This can only be achieved if the logistics service provider is involved intensely at an early stage into the outsourcing process, thereby enabling adequate communication and cooperation. The construct involvement so far has received very little attention in logistics outsourcing research. In this study it will, in-line with the argumentation presented above and the research conducted by ENGELBRECHT (2004), measure to what extent the LSP was involved in the implementation process of the logistics outsourcing and how well the collaboration between the employees of both the customer and the LSP has been functioning. 4.2.1.8 Opportunism

Opportunistic behavior, which is an essential aspect of economic theory, is the overarching concept of transaction cost theory. A large body of literature is detailing its types and forms: key sources include AKERLOF (1970), WILLIAMSON (1975), JENSEN/MECKLING (1976), KLEIN/CRAWFORD/ ALCHIAN (1978) and WILLIAMSON (1985). Opportunistic behavior in the transaction cost literature is defined as “self-interest seeking with guile” (WILLIAMSON 1985, p. 47). Thus, “the essence of opportunistic behavior is the deceit-oriented violation of implicit or explicit promises about one’s

100

4 Antecedents and effects of logistics outsourcing performance

appropriate or required role behavior” (JOHN 1984, p. 279). Since opportunistic behavior in the organizational literature is assumed in the fundamental axioms (see chapter 3.2.1.1), rather than being treated contingently or empirically, DONALDSON (1990, p. 373) states that “this is guilt by axiom”. MORGAN/HUNT (1994, p. 25) suggest that even though guileful and self-interest maximization may be axiomatic in transaction cost analysis, human behavior “may not be so Machiavellian after all”, especially in the long run (BONOMA 1976; JOHN 1984). Therefore, opportunism must not be treated as axiomatic in relationship research, but must rather be understood as an independent variable. As such, opportunism will be understood in this study. The incorporation of trust in exchange relationship models (DWYER/SCHURR/OH 1987, pp. 22-23) provides a suitable vantage point for treating opportunism as an explanatory variable. MORGAN/HUNT (1994, p. 30) and KNEMEYER/MURPHY (2004, pp. 44-46) all could show that opportunistic behavior has direct negative effects on trust and therefore correctly is employed in relationship research. 4.2.1.9 Shared values

The concept of shared values according to MORGAN/HUNT (1994, p. 25) is “the extent to which partners have beliefs in common about what behaviors, goals, and policies are important or unimportant, appropriate or inappropriate, and right or wrong”. They are important for exchange relationships since they have been shown to govern individual exchange relationships between firms (STINCHCOMBE 1986; SHAPIRO 1987) and constitute the basis for fundamental partnership variables such as relationship commitment and trust (DWYER/SCHURR/OH 1987, p. 21; MORGAN/ HUNT 1994, p. 25). Norms for example, as introduced by HEIDE/JOHN (1992, p. 34) are shared values because they refer to the expectations of behavior that are at least partially shared by a group of decision makers. Values are fundamental to definitions of organizational culture (ENZ 1988, p. 287; WIENER 1988, p. 534). SCHEIN (1990, p. 111) observes that it is possible to “distinguish three fundamental levels at which culture manifests itself: (a) observable artifacts, (b) values, and (c) basic underlying assumptions”. Consequently, values reflect culture when they are widely and strongly held (WIENER 1988; SCHEIN 1990). Because it has been suggested that shared values are the best measure to determine the person-organization fit in employment settings (CALDWELL/O'REILLY 1990; CHATMAN 1991), they have become a variable of great interest also in the organizational commitment literature (CHATMAN 1991, p. 451).

4.2 Identification of relevant antecedents

101

This study will use the above stated definition of MORGAN/HUNT (1994, p. 25) and understand shared values as the common believes of partners about what behaviors, goals, and policies are important or unimportant, appropriate or inappropriate, and right or wrong. 4.2.1.10 Openness

The issue of openness between individuals and/or organizations so far has received only little attention of researchers in the area of business exchange relationships, an exception being the work of BENNETT (1996). Until now, the concept has primarily been utilized to analyze openness and informality in interpersonal relationships inside firms, particularly between managers and their subordinates. ASHFORD/ROTHBARD/PIDERIT/DUTTON (1998, p. 46) find that the openness of the top management team, defined as the openness to new ideas and suggestions from subordinates (ASHFORD/ROTHBARD/ PIDERIT/DUTTON 1998, p. 37), is clearly related to perceived organizational support and the quality of the relationship. PREMEAUX/BEDEIAN (2003, p. 1557), using a similar understanding, find that top-management openness is a meaningful predictor of an employee’s ability to speak up and express his views, which is in turn perceived as positive since employees have increasingly become recognized as an invaluable source of ideas for improving a firm’s performance (HARRINGTON 2001, p. 85-88). GUPTA (1987) analyzes the effect of openness between corporate executives and strategic business unit managers on the effectiveness of strategy implementation in different contexts. GUPTA (1987, p. 479) states that openness “refers to the degree to which relations between [strategic business unit’s] managers and their corporate superiors are open and informal and allow for spontaneous and open exchange of information and ideas”. He finds that the upside benefits of a high level of openness in some strategic contexts are much greater than the downside negative effects in the other strategic contexts (GUPTA 1987, p. 493) and concludes that in situations in which the adequate level of openness is not precisely known to corporate executives, “it should be preferable to opt for more, not less, openness across the board”. The beneficial character of openness for exchange relationships is acknowledged in this study as following from the argumentation presented above it may act as a facilitator of improved relationship quality. It will therefore in line with GUPTA (1987, p. 499) be defined as a measure for a relationship that is open and informal and allows for spontaneous and open exchange of information and ideas.

102

4 Antecedents and effects of logistics outsourcing performance

4.3 Formulation of a model of logistics outsourcing performance The previous chapter has conceptualized the variables proposed to directly or indirectly affect logistics goal achievement and goal exceedance. In order to now postulate a consistent model of logistics outsourcing performance, hypotheses on the causal linkages between the different variables must be formulated. Consequently, the following sections 4.3.1.1 through to 4.3.1.10 will introduce these hypotheses. While for all ten variables which are modeled as antecedents of outsourcing performance causal linkages among each other will be hypothesized, direct causal linkages on the outsourcing performance will only be theorized for the constructs of cooperation, proactive improvement, and functional conflict. In the formulation of the hypotheses, two different inputs are utilized: on the one hand in the form of causal linkages that have been proposed previously by researchers from either logistics outsourcing research or from general exchange relationship research and on the other hand by interpreting the theories presented in chapter 3.2. 4.3.1 Hypotheses on causal linkages between the variables 4.3.1.1 Cooperation

In the literature, cooperation has received widespread attention and is viewed as an important variable in exchange relationships (ANDERSON/NARUS 1984; ANDERSON/NARUS 1990; MORGAN/HUNT 1994; WILSON 1995). While the formulated models differ from case to case, consensus has been reached among the majority of scholars that cooperation holds a key role in interorganizational relationships and contributes significantly to their functioning. The role of cooperation in logistics outsourcing relationships, however, so far has received very little attention. Neither its effects on the outsourcing performance, nor its causal linkages with other relationship variables have explicitly been examined. While for some of the causal linkages, works form the general field of exchange relationships can be reverted to, an interpretation of the theories presented in chapter 3.2 promises to have the highest explanatory value. Following transaction cost theory, it can be hypothesized that cooperation has a positive effect on logistics outsourcing performance. As both parties cooperate by working closer together and by pulling in the same direction, transaction costs can be significantly lowered. One the one hand,

4.3 Formulation of a model of logistics outsourcing performance

103

monitoring costs can be reduced as parties understand that cooperation renders monitoring and control less important. On the other hand, cooperation leads to the achievement of previously agreed upon goals and thereby reduces the need to ex-post bargain with or sanction a partner that does not perform according to the agreement, thus reducing enforcement costs. Social exchange theory also suggests that relationships based upon mutual cooperation are more rewarding than those that are not. As relationships extend beyond mere transactional exchanges, parties are motivated to manage risk and uncertainty through cooperation because they believe that doing so will be rewarding. They thereby receive benefits normally only attainable through vertical integration. Therefore, increasing cooperation between customer and LSP should lead to increasing outsourcing performance. These interpretations are backed by the findings of ANDERSON/NARUS (1990, pp. 45-46). They propose that cooperation has a positive effect on customer satisfaction as it facilitates the attaining of desired outcomes, a path which has also received support from other channel researchers such as MALLEN (1963), DWYER (1980), and SIBLEY/MICHIE (1982). At the same time, ANDERSON/NARUS (1990, p. 46) state that satisfaction is a “close proxy for […] perceived effectiveness”. Since goal achievement is measuring just this perceived effectiveness of the LSP when providing the service, the argumentation may be used analogously and it can be supposed that cooperation has a positive effect on logistics goal achievement. While the findings derived from transaction cost theory, social exchange theory, and the literature certainly apply for goal achievement, it may well be assumed that they also hold for goal exceedance, defined as the overachievement of the goals set prior to the outsourcing arrangement. Following transaction cost theory, it can be supposed that the closer the relationship as a consequence of the cooperation is, the lower will be various transaction costs: monitoring costs are reduced as monitoring and control become less important. Furthermore, if closer cooperation leads to exceeding the set goals, ex-post bargaining with or sanctioning the LSP becomes unnecessary, thus reducing enforcement costs. As argued above, social exchange theory suggests that cooperative relationships are more rewarding than adversarial relationships as benefits become attainable that without cooperation would not be achievable. Thus, the closer the cooperation between the two parties is, the more benefits will be available for the partners. In cases of very good cooperation, these benefits may well exceed the expectations the customer had before entering the outsourcing arrangement, thus leading to goal exceedance. On this basis, it can therefore be assumed that the more intense the cooperation between the partners, the more successful will be the outsourcing

104

4 Antecedents and effects of logistics outsourcing performance

relationship and the more likely it is that the previously set expectations are significantly exceeded. Therefore, the following two hypotheses will be formulated: H1a: Cooperation has a positive effect on goal achievement. H1b: Cooperation has a positive effect on goal exceedance. Additional to these direct effects, cooperation also has effects on other antecedents of outsourcing performance. While the causal linkages of other variables on cooperation will be presented in the respective chapters, in the following the effect of cooperation on functional conflict and proactive improvement will be presented. Social exchange theory suggests that as relationships develop and firms shift from discrete to long-term relationships, they will start realizing that cooperative behaviors foster greater benefits for the exchange partners (LAMBE/WITTMANN/SPEKMANN 2001, p. 23) and will downright expect their partners to participate in them for the benefit of the partnership (SPEKMAN/SALMOND/LAMBE 1997, pp. 832-833). As chapter 4.2.1.6 shows, functional conflict is beneficial for exchange relationships and therefore a characteristic desired by the partners. It can thus be hypothesized that as firms engage in cooperative behavior and understand its beneficial character, they will also aim at increasing the functionality of conflict in the relationships. This insight is backed by the research of ANDERSON/NARUS (1990, p. 49), who empirically show that cooperation leads to functional conflict. A further positive aspect of close and long-term oriented relationships is that the LSP can engage in behaviors that would not be rewarding in discrete, short term exchanges. One of these is proactive improvement, which will be argued in chapter 4.3.1.3 to positively influence both the success of the customer and of the LSP. As the level of cooperative behavior rises in the relationship, the LSP can expect to be increasingly rewarded for his efforts as a consequence of the intensified customer satisfaction under the social exchange theory assumption of reciprocity. Therefore, it can be argued that rising levels of cooperation in the relationship lead to stronger proactive improvement on the side of the LSP. These findings lead to the following hypotheses: H2:

Cooperation has a positive effect on functional conflict.

H3:

Cooperation has a positive effect on proactive improvement.

4.3 Formulation of a model of logistics outsourcing performance

105

4.3.1.2 Communication

Communication, just as cooperation, is a variable intensely studied in the exchange relationship literature. Its role has been examined among others in the works of DWYER/SCHURR/OH (1987), ANDERSON/WEITZ (1989), ANDERSON/NARUS (1990), MORGAN/HUNT (1994), MOORE (1998), LAMBERT/EMMELHAINZ/GARDNER (1999), and KNEMEYER/MURPHY (2004) and has almost across the board been evaluated as important for exchange relationships. According to MOORE (1998, p. 28), information exchange and communication are important elements in logistics alliances. This is supported by the findings of SINK/LANGLEY (1997, pp. 176-177) who state that logistics outsourcing involves a high degree of communication and interaction between different organizational levels of the customer and between the customer and the LSP. BOWERSOX/DAUGHERTY/DROGE/ROGERS/WARDLOW (1989, p. 225) even believe that the complete and open exchange of operating and strategic information is the “glue that holds [logistics] alliances together”. The qualitative aspect of communication is often emphasized. Apparently, the way information is exchanged is more important for the success of a relationship than the sheer quantity alone. However, it must also be noted that additionally to the importance of communication quality, some authors suggest that with rising product or service complexity, the amount of required information exchange also rises. CUNNINGHAM/TURNBULL (1982, p. 305) note that “the complexity of the product being purchased [ …has] profound effects on the amount of information exchange which is required and the length of time over which this occurs”. Similarly, METCALF/FREAR/KRISHNAN (1992, p. 29) suggest that while the exchange of information for the purchase of a standard product would not necessarily yield intense ties between buyer and seller, the purchase of a complex product might require close collaboration and communication between the parties for a longer period of time. The same can be assumed in a logistics outsourcing context in which more complex services require more intense communication than standard services do. Social exchange theory as shown in chapter 3.2.2 acknowledges the superior importance of interaction and communication for successful relationships. As parties strive to increase relationship benefits through cooperation, they find the exchanging and sharing of information as a necessary prerequisite. Therefore, social exchange theory permits the argumentation that communication has a positive effect on the outsourcing performance. STANK/GOLDSBY/VICKERY/SAVITSKIE (2003, p. 43) empirically show that relational performance increase both logistics operational and cost per-

106

4 Antecedents and effects of logistics outsourcing performance

formance. They measure relational performance with three indicators, one of them being whether the service provider knows the customers needs well (STANK/GOLDSBY/VICKERY/SAVITSKIE 2003, p. 32). Since this knowledge is obtained primarily through communication, this finding can be seen as an indicator for the importance of communication for goal achievement. Furthermore it can be argued that increasing levels of communication have a positive influence also on goal exceedance. As the LSP better understands the needs of the customer through communicating, he will be able to address them through improving his service, leading to outcomes that exceed the expectations ex-ante set by the customer. The same increased levels of communication, however, may also lead to a continuous adaptation of the goals by the customer, which could potentially cause rising expectations and lead to less customer satisfaction as a result. While obviously some theoretical evidence suggests a link between communication and logistics outsourcing performance, it can be argued that communication alone is not an action that directly affects performance. Certainly, an environment characterized by extensive communication will be more likely to meet performance expectations than one with very limited communication. However, the increased performance will be caused by the effects of communication on factors with direct effects on the performance dimensions, such as cooperation and proactive improvement. On its own, communication has no effect as always an activity is necessary to convert the information exchanged during the communication into an action that leads to measurable performance. Communication therefore is a facilitator of performance, but not a direct antecedent. Following this argumentation, no hypotheses on a direct effect of communication on goal achievement and goal exceedance will be formulated. However, communication has effects on antecedents of outsourcing performance. Again, the causal linkages of other variables on communication will be presented in the respective chapters, while in the following the effect of communication on trust, functional conflict, cooperation, proactive improvement, and opportunism will be presented. According to MORGAN/HUNT (1994, p. 25) and the commitment-trust theory, communication is a “major precursor of trust”. The theory suggests that relevant, timely, and reliable communication fosters trust by resolving disputes and aligning perceptions and expectations (ETGAR 1979, pp. 77-78), a view also supported by the research of ANDERSON/WEITZ (1989) and ANDERSON/NARUS (1990). ANDERSON/NARUS (1990, p. 45) also note that communication is an antecedent of trust, but that in subsequent periods this accumulation of trust leads to better communication. However, since the model in this study will be tested in one particular

4.3 Formulation of a model of logistics outsourcing performance

107

point in time, it will only be posited that communication increases trust and the lagging effects will be neglected (ANDERSON/NARUS 1990; MORGAN/HUNT 1994). As it has been shown above, the commitment-trust theory suggests that communication supports resolving conflicts (ETGAR 1979, pp. 77-78; MORGAN/HUNT 1994, p. 26). As disputes and conflicts are discussed timely in an atmosphere of trust between the relevant representatives of both parties, expectations can be clearly communicated and ways be found to jointly solve conflicts before their detrimental effects affect the relationship. It can thus be presumed that the better and the more intense the communication between customer and LSP is, the more likely is it that conflicts will be handled in a functional way. A further insight into the importance of communication can be gained through social exchange theory. As it has been argued in chapter 3.2.2.2, customers and LSPs in long-term logistics relationships are motivated to manage risks and uncertainty through cooperation in order to receive benefits normally only attainable through vertical integration. This has been shown to especially require the sharing of information and finds the empirical support of METCALF/FREAR/KRISHNAN (1992, p. 39). It can therefore be hypothesized that a positive relationship exists between communication and cooperation. The importance of the LSP’s proactive improvement has been argued in chapter 4.2.1.3. Following the social exchange literature, the LSP will increase this pursuit if he expects an adequate reciprocity. This behavior will be assumed for the long-term relationships between customers and LSPs analyzed in this study. However, the proactive improvement of the LSP also largely depends on its knowledge of the context of the customer’s situation and its issues. Therefore, the LSP’s effort will be greatly supported through relevant, timely, and reliable communication of the customer. Only if the customer communicates its expectations and needs clearly, the LSP is enabled to respond accordingly. Therefore, it can be presumed that communication has a positive effect on the LSPs proactive improvement. As a last hypothesis, the relationship between communication and opportunism shall be examined. While social exchange theory has been criticized for the assumption that relational exchanges are completely devoid of opportunism (LAMBE/WITTMANN/SPEKMANN 2001, p. 26), the same is true for transaction cost theory which assumes universal opportunism. As in reality opportunism most probably will take on a position between those two extremes, it can be derived from the social exchange literature that a relationship profits from any action that is beneficial. Since opportunism according to transaction cost theory is detrimental, it can be argued that

108

4 Antecedents and effects of logistics outsourcing performance

communication between the parties reduces the risk of opportunistic behavior as it increases transparency, exchange, and interaction. Therefore, communication should limit opportunistic behavior, which will in this study be modeled as a negative effect of communication on opportunism. Resulting from the above argumentations, the following hypotheses can be formulated: H4:

Communication has a positive effect on trust.

H5:

Communication has a positive effect on functional conflict.

H6:

Communication has a positive effect on cooperation.

H7:

Communication has a positive effect on proactive improvement.

H8:

Communication has a negative effect on opportunism.

4.3.1.3 Proactive improvement

While it has recently been established that innovation is critical to the success of logistics service providers (FLINT/LARSSON/GAMMELGAARD/ MENTZER 2005, p. 113), LSPs’ proactive improvement as a manifestation of its innovation orientation has received only very little attention, most notably recently by ENGELBRECHT (2004, pp. 207-209) and WALLENBURG (2004, pp. 102-103). The hypothesis of the positive effect of the proactive improvement on goal achievement in logistics outsourcing arrangements is supported by transaction cost theory. As the LSP continues to optimize its services, the customer will consequently experience lower enforcement costs than in a situation without optimization in which the results would not have been as satisfactory. Social exchange theory also suggests positive effects of proactive improvement on goal achievement. As cooperative behaviors between partners increase as a result of the anticipated reciprocity, the LSP can expect to be rewarded for the improved services and therefore will strive to optimize the outcome for the customer in its own interest, leading to a higher goal achievement. Apart from transaction cost and social exchange theories, it can also be argued that the constant efforts of the LSP for proactive improvement will eventually translate into increased goal achievement as ultimately the activities will show measurable effects. This would only not be the case if

4.3 Formulation of a model of logistics outsourcing performance

109

the LSP would continuously fail in his efforts to produce some form of improvement. However, in professional outsourcing arrangements this is very unlikely. These theoretical postulates are supported by recent empirical findings. ENGELBRECHT (2004, p. 276) shows that proactive improvement of the LSP has a strong positive causal effect on the goal achievement of outsourcing projects. This insight is backed by findings of STANK/GOLDSBY/ VICKERY/SAVITSKIE (2003, p. 43) who demonstrate that the relational performance between customer and LSP positively influences both operational and cost performance of the outsourcing arrangements. With respect to proactive improvement this is important, as STANK/GOLDSBY/ VICKERY/SAVITSKIE (2003, p. 32) measure relational performance with three indicators, one of which is that the LSP “makes recommendations for continuous improvement on an ongoing basis”. The importance of proactive improvement for the goal achievement has therefore been established. However, as argued in chapter 4.1, the overall logistics outsourcing performance in this study will be measured also with a second dimension indicating whether the acceded goals have been significantly exceeded in terms of quality of service and costs. Since the proactive improvement through the LSP will lead to continuous improvement of the service quality and/or cost reductions, it will be postulated that proactive improvement positively influences goal exceedance. Not only will the improvements realized by the LSP lower transaction costs below the levels expected by the customer, e.g. through lowering the enforcement costs as sanctioning the LSP becomes less or un-important in an environment where it aims at constantly delivering the best possible service to the customer. Also, goal exceedance will be positively influenced by proactive improvement because the efforts of the LSP going beyond the goals formulated in the contract will eventually lead to surpassing them. In this context, proactive improvement is even more important for goal exceedance than for goal achievement. For the achievement of the latter, it may suffice to only deliver the agreed upon service level. To exceed these goals, however, the LSP must deliver either better service or lower costs, preferably both together. Therefore, the following hypotheses can be derived: H9a: Proactive improvement of the LSP has a positive effect on goal achievement.

110

4 Antecedents and effects of logistics outsourcing performance

H9b: Proactive improvement of the LSP has a positive effect on goal exceedance. 4.3.1.4 Trust

As chapter 4.2.1.4 has argued, trust is essential for virtually all interorganizational relationships (ARROW 1975, p. 24) and is “the cornerstone to the strategic partnership” (SPEKMAN 1988, p. 79). It is therefore included in most relationship models as a fundamental building block (WILSON 1995, p. 337). According to MORGAN/HUNT (1994, pp. 22-23) and the commitment-trust theory, trust is a key mediating variable within effective relational exchanges. Some empirical evidence exists that trust has a direct positive influence on the performance of logistics outsourcing arrangements. KNEMEYER/ MURPHY (2004, p. 45) find support for their hypothesis that trust is positively related to the buyer’s perception of operations performance, while they have to reject their hypotheses regarding the positive effects on both channel and asset reduction performance. This is particularly interesting in the light of the commitment-trust theory which does not postulate a direct effect on relationship performance, but rather suggests indirect effects via the causal linkages on constructs such as commitment, cooperation, and functional conflict. Also, neither social exchange nor transaction cost theories theorize such a direct causal effect of trust on performance. Following these argumentations, no direct effects of trust on logistics outsourcing performance will be hypothesized in this study. Instead, its indirect effects on outsourcing performance via other antecedents will be explored in the following. The commitment-trust theory suggests a strong influence of trust on relationship commitment (MORGAN/HUNT 1994, p. 24). As relationships characterized by trust are valued very highly, the parties will desire to commit themselves to such relationships (HREBINIAK 1974). Also, since relationship commitment means a significant degree of vulnerability, the parties will also seek only trustworthy partners. This hypothesis is supported by social exchange theory, which explains the causal relationship through the principle of generalized reciprocity, stating that mistrust leads to mistrust, thereby reducing relationship commitment, and eventually shifting the relationship to one of more direct short-term exchange (MCDONALD 1981, p. 834). A further outcome of trust suggested by the commitment-trust theory is its positive effect on cooperation (MORGAN/HUNT 1994, p. 26). As already the prisoner’s dilemma experiments of DEUTSCH (1960) show, the initia-

4.3 Formulation of a model of logistics outsourcing performance

111

tion of cooperation requires trust. ANDERSON/NARUS (1990, p. 45) state that „once trust is established, firms learn that coordinated, joint efforts will lead to outcomes that exceed what the firm would have achieved if it acted solely in its own best interest”. This causal linkage was subject to a respecification by ANDERSON/NARUS (1990, p. 48) on the basis of the empirical data. Consequently, they reversed the previously suggested causal linkage. However, since it was later on confirmed in its original form by MORGAN/HUNT (1994, p. 30), this original form will be tested in this study. Consequently, this hypothesis will be of greatest interest in this study. A last causal linkage of trust suggested by the commitment-trust theory is the positive effect on functional conflict. As MORGAN/HUNT (1994, p. 26) state, trust leads partners to believe that future conflicting situations will be functional. They further propose that past cooperation and communication will result in increased functionality of conflict as a result of increasing trust. While the causal linkage between communication and trust is part of their model on the basis of the commitment-trust theory (MORGAN/HUNT 1994, p. 22), the functional chain between trust, cooperation, and functional conflict is not. However, to remedy this shortcoming it will be incorporated in the logistics outsourcing performance model as already argued in chapter 4.3.1.1. On the basis of the argumentations presented above, the formalized hypotheses on the effects of trust in exchange relationships are the following: H10: Trust has a positive effect on relationship commitment. H11: Trust has a positive effect on cooperation. H12: Trust has a positive effect on functional conflict. 4.3.1.5 Commitment

As chapter 4.2.1.5 has shown, relationship commitment is the most common dependent variable used in buyer-seller relationship studies. As MORGAN/HUNT (1994, p. 23) argue, parties involved in exchange relationships identify relationship commitment as key for achieving desired outcomes and therefore work hard to maintain this important aspect of their relationships. In the commitment-trust theory, it thus takes on a central role as a key mediating variable. Alongside with trust, it is positioned between the relationship characteristics that can be influenced by the partners and the out-

112

4 Antecedents and effects of logistics outsourcing performance

come variables acquiescence, propensity to leave, and cooperation (MORGAN/HUNT 1994, pp. 25-26). While acquiescence and propensity to leave are not subject of the research in this study, as they have a customer loyalty orientation, the relationship between relationship commitment and cooperation will be examined. The commitment-trust theory suggests that a party, in this case the customer, committed to a relationship will cooperate with a partner, because of a desire to make the relationship work (MORGAN/HUNT 1994, p. 26). Commitment thus leads directly to cooperative behavior (MORGAN/HUNT 1994, p. 22). On the basis of these argumentations, a positive relationship between commitment and cooperation can be hypothesized. This can also be concluded from social exchange theory. As relationships develop and firms shift from discrete transactions to long-term relationships, the exchange partners will develop commitment to the relationship as they begin anticipating rewards that would not be attainable through discrete transactions. As the customer’s commitment to the relationship increases, so does its understanding for the necessity of cooperative behavior which is an integral part of exchange relationships as a basis for mutual reciprocity (LAMBE/WITTMANN/SPEKMANN 2001, p. 23). Therefore, as the commitment of the customer to the relationship increases, so do its efforts for cooperative behavior, thus leading to increasing cooperation between the partners. Aside from MORGAN/HUNT (1994, p. 30) who find empirical support for the positive effect of relationship commitment on cooperation, it has received very little attention. Particularly its existence in logistics outsourcing relationships has not been studied at all. The hypothesis in this study will therefore be: H13: Relationship commitment has a positive effect on cooperation. 4.3.1.6 Functional conflict

As argued in chapter 4.2.1.6, when occurring conflicts are solved amicably, they can be turned into “functional conflict”, which has beneficial effects on the relationship. The role of functional conflict in logistics outsourcing relationships so far has not received any attention. To predict its effects on logistics goal achievement, social exchange theory offers some valuable insights. As been argued in chapter 3.2.2.2, the theory suggests that parties involved in longer-term relational exchanges are motivated to practice cooperative behavior in order to achieve mutual benefits. This cooperation between partners is decreased as the det-

4.3 Formulation of a model of logistics outsourcing performance

113

rimental effects of conflict take effect on the relationship, thereby also reducing the positive effects high levels of cooperation have on the accomplishment of goals in general (STERN/REVE 1980) and on logistics outsourcing performance in particular. However, if the inevitable conflicts can be solved amicably and functionally rather than becoming openly hostile, the detrimental effects of the conflict on the important cooperative behavior can be reduced. In this case, functional conflict potentially serves as a means of “clearing the air” (ANDERSON/NARUS 1990, p. 45) and has functional and productive consequences. Therefore, it can be hypothesized that functional conflict has a positive effect on goal achievement. Furthermore, functional conflict will also positively influence goal exceedance. When engaging in a logistics outsourcing arrangement with the LSP, the customer will develop expectations about the service provided. Part of these expectations will be the detrimental effects of conflicts, which have been argued to be inevitable in relationships. If the level and weight of these conflicts are now decreased beyond the expectations of the customer, the goals set in the contract will not only be achieved, but even exceeded as now levels of performance are reachable that under the ordinary conflict level of average relationships without functional conflict would never be. While the relationship between functional conflict and outsourcing performance has never been analyzed empirically, the research conducted by SKINNER/GASSENHEIMER/KELLEY (1992, p. 185) supports the argumentation. Empirically, they find that the level of conflict has a negative impact on both the level of cooperation displayed in the relationship and the customer’s satisfaction with the relationship. As now functional conflict decreases the level of conflict in the relationship, the level of cooperation as well as the customer’s satisfaction should rise. Hence, functional conflict will have beneficial consequences for the relationship. Consequently, the hypotheses are as follows: H14a: Functional conflict has a positive effect on goal achievement. H14b: Functional conflict has a positive effect on goal exceedance. 4.3.1.7 Involvement

The degree of involvement of the LSP into the implementation process of logistics outsourcing arrangements as yet has not received extensive attention in the literature as shown in chapter 4.2.1.7. While BOWERSOX/ CLOSS/STANK (2003) and MCHUGH/HUMPHREYS/MCIVOR (2003) argue

114

4 Antecedents and effects of logistics outsourcing performance

on a general level that interorganizational collaboration is a success factor in outsourcing arrangements and other exchange relationships, ENGELBRECHT (2004, pp. 272-277) is the first to find empirical evidence for the hypotheses that the extent of involvement of the LSP in the implementation process leads to improved communication and proactive improvement through the LSP. While these findings are based on exploratory research, they make sense especially in the light of social exchange theory. As partners strive for joint economic and social benefits, they will increase all aspects of cooperative behavior as argued in chapter 3.2.2.2. Therefore, as the LSP and its employees are involved in the outsourcing process, work together, and address conflicts early and functionally, e.g. in a joint project management team, the overall level of cooperation should increase. Furthermore it can be assumed that as the level of involvement of the LSP in the outsourcing arrangement increases, so will the communication between the parties as they understand the potential of early and open discussions for their joint goals. Finally, involvement, leading to more cooperative behavior, will also lead to increased proactive improvement of the LSP as the opportunities of identifying optimization potentials in the customers’ processes increase due to early and thorough involvement. ENGELBRECHT (2004, pp. 272-277), as mentioned above, finds empirical support for the hypotheses that the degree of involvement of the LSP positively influences communication and proactive improvement. The hypothesis that involvement leads to higher outsourcing performance finds no support in his study. Since also none of the theories employed in this study suggest theoretical evidence to that effect, this causal linkage will not be further explored. The hypotheses put forward therefore are: H15: The involvement of the LSP has a positive effect on cooperation. H16: The involvement of the LSP has a positive effect on communication. H17: The involvement of the LSP has a positive effect on proactive improvement. 4.3.1.8 Opportunism

Opportunistic behavior is a critical antecedent in the relationship between customer and LSP and has important practical implications for firms out-

4.3 Formulation of a model of logistics outsourcing performance

115

sourcing logistics services (KNEMEYER/MURPHY 2004, p. 41). If the risk of opportunism in relationships is sufficiently high, considerable resources must be spent that could in other cases be utilized more productively (WATHNE/HEIDE 2000, p. 36). Furthermore, opportunism carries the danger that, if partners cannot guard themselves against exploitation, “valuable deals will not be done” (CALFEE/RUBIN 1993, p. 164). As a result, opportunism generates opportunity costs (WATHNE/HEIDE 2000, p. 36) and therefore “tends to be detrimental to logistics outsourcing relationships” (KNEMEYER/MURPHY 2004, p. 41). Opportunism is one of the two basic behavioral assumptions of transaction cost theory. As chapter 3.2.1.1 has presented, it is treated as a fundamental axiom and therefore, as it is not treated contingently, is “guilt by axiom” (DONALDSON 1990, p. 373) and thus detrimental. Some channel researchers criticize this strong assumption and, as shown in chapter 4.2.1.8, indicate that especially behavior in longer-term relationships may not be characterized through universal and unlimited opportunism (BONOMA 1976; JOHN 1984). The inclusion of opportunism into exchange relationship models therefore promises to have high explanatory value beyond the rather static and one-dimensional axiomatic approach of transaction cost theory. After DWYER/SCHURR/OH (1987, pp. 22-23) proposed that the incorporation of trust into relationship models would provide a suitable vantage point, the commitment-trust theory takes up the argumentation. Consequently, MORGAN/HUNT (1994, p. 25) argue that if a partner believes the other to engage in opportunistic behavior, these perceptions will lead to decreased trust. This negative causal linkage they also show empirically (MORGAN/HUNT 1994, p. 30) and find support in the work of KNEMEYER/MURPHY (2004, p. 46), who specifically analyzed relationships between customers and their LSPs. While opportunism will not be understood as a basic behavioral assumption in the sense of transaction cost theory in this study, but will rather be treated contingently, its effects are still seen as detrimental. As it has a negative influence on the relationship between two parties, social exchange theory also promises insights. While this theory normally does not include opportunism, it argues that cooperation is important for functioning long-term relationships. As opportunistic behavior by a partner is suspected by another, this will have the argued detrimental effects on the cooperation and the process of the two parties working together will be severely impeded. Therefore, it can be presumed that higher levels of opportunism will have negative effects on the cooperation between parties. This negative effect will additionally be aggravated through the fact that

116

4 Antecedents and effects of logistics outsourcing performance

opportunistic behavior will reduce trust, which in turn will have a negative impact on the level of cooperation, as argued in chapter 4.2.1.4. The hypotheses on opportunism consequently are the following: H18: Opportunistic behavior has a negative effect on trust. H19: Opportunistic behavior has a negative effect on cooperation. 4.3.1.9 Shared values

The concept of shared values has been shown in chapter 4.2.1.9 to be important for exchange relationships research, since they have been indicated to govern individual exchange relationships between firms (STINCHCOMBE 1986; SHAPIRO 1987). Also, they are important to exchange relationships between customer and LSP because they foster organizational commitment. MORGAN/HUNT (1994, p. 25) state that the organizational commitment literature distinguishes between two kinds of commitment: one the one hand that brought about by a person “sharing, identifying with, or internalizing the values of the organization” and on the other hand that “brought about by a cognitive evaluation of the instrumental worth of a continued relationship” which is determined through weighing up of the “gains and losses, plusses and minuses, or rewards and punishments”. Thus is becomes clear that a customer sharing the values of the LSP will develop greater relationship commitment, which in turn is beneficial for the relationship. Furthermore, as WILSON (1995, p. 338) points out, mutual goals – as a subset of shared values – encourage both mutuality of interest and stewardship behavior and therefore lead to increased cooperation between the parties. DWYER/SCHURR/OH (1987, p. 21) theorize that shared values contribute to the development of commitment and trust, two constructs of substantial importance for the research conducted in this study. This thought is taken up by MORGAN/HUNT (1994, p. 25) during the development of the commitment-trust theory. Here, shared values are the only antecedent posited to influence both key mediating variables commitment and trust, which subsequently is also supported by empirical findings (MORGAN/HUNT 1994, p. 30). These effects of shared values will consequently be also hypothesized in this study. Beyond these two effects, the commitment-trust theory does not offer further direct insights. However, it does suggest that the existence of shared values among exchange partners leads to relationships that are more

4.3 Formulation of a model of logistics outsourcing performance

117

functional and cooperative than others. It can therefore be presumed that shared values have further positive effects on other constructs. A first hypothesis will be that with increasing levels of shared values, it will be easier for the parties to cooperate. As they understand each others’ motivations and expectations, they will work better together to achieve the mutual goals. Similarly, shared values and the accompanying perceived closeness between the exchange partners will decrease the tendency towards opportunistic behavior. While MORGAN/HUNT (1994, p. 24-30) understand opportunism as a precursor and therefore as an independent variable in their model, for this study it will be argued that the level of acted opportunism can indeed vary. If one party feels closer and more connected to another, it will reduce opportunistic behavior. Thus, increasing levels of shared values should decrease the level of opportunistic behavior. Parties sharing similar values should also find it easier to communicate. As they agree on fundamental issues, such as what is “important or unimportant, appropriate or inappropriate, and right or wrong” (MORGAN/ HUNT 1994, p. 25), they ought to find it easier to share formal and informal meaningful and timely information, because they can better relate to the partners needs than in relationships in which partners do not share the same values. This should also be the case for openness. To the degree that the parties share the same values, they will find it less difficult to openly and informally discuss current issues and problems as the partners will have less fear of being rejected. Consequently, shared values will lead to increased openness in the relationship. The findings on the importance of shared values for exchange relationships lead to the following hypotheses: H20: Shared values have a positive effect on trust. H21: Shared values have a positive effect on relationship commitment. H22: Shared values have a positive effect on cooperation. H23: Shared values have a negative effect on opportunistic behavior. H24: Shared values have a positive effect on communication. H25: Shared values have a positive effect on openness.

118

4 Antecedents and effects of logistics outsourcing performance

4.3.1.10 Openness

As argued in chapter 4.2.1.10, openness is viewed as a major driver for successful strategy implementation. So far, it has not been utilized as a construct in interorganizational exchange relationship research in the logistics domain. However, social exchange theory suggests that openness could be an important antecedent for cooperative and sustainable relationships characterized by commitment and trust which both should be facilitated by high levels of openness in relationships. Corresponding with the demand of several participants in pre-test interviews for this study (see chapter 5.1.3), openness therefore will be viewed as a prerequisite for successful relations between customer and LSP and its role and causal linkages with other constructs will be analyzed. While the concept of openness as a factor in exchange relationships has not received particular attention, even though its importance is increasingly acknowledged in interpersonal relationship research (GUPTA 1987; ASHFORD/ROTHBARD/PIDERIT/DUTTON 1998; PREMEAUX/BEDEIAN 2003). This may be due to the misconception that openness in relationships is included in the construct of communication. However, while certain similarities undoubtedly exist, communication measures the efficient and effective flow of information between the parties (chapter 4.2.1.2), while openness serves as an indicator whether the relationship as such is characterized by informality, spontaneity, and the open exchange of information and ideas. While this connection implies that openness will have a positive effect on communication in a relationship, further causal linkages to other constructs can be hypothesized. As it has been shown in chapter 4.2.1.2, the commitment-trust theory suggests that communication is a major precursor of trust. As openness increases in a relationship, the open exchange of information and ideas should facilitate the reduction of the main reason for mistrust which is to be found in the uncertainty of possible future opportunistic behavior of the partner (WALLENBURG 2004, p. 219). Therefore, openness should not only have an indirect effect on trust via communication, but also a direct influence. The same is true for the positive effect of openness on functional conflict. As much as communication has been shown in chapter 4.2.1.2 to reduce disputes by aligning perceptions and expectations, openness will do the same and facilitate the turning of conflicts into functional exchanges of ideas and perceptions through enabling the informal, spontaneous, and open exchange of ideas and views. Therefore, openness will have a positive effect on functional conflict.

4.3 Formulation of a model of logistics outsourcing performance

119

The hypotheses derived from the above argumentation read in formalized fashion as follows: H26: Openness has a positive effect on communication. H27: Openness has a positive effect on trust. H28: Openness has a positive effect on functional conflict.

4.3.2 Overview of the hypotheses and consequent model After ten variables with supposed direct or indirect effects on the two dimensions of logistics outsourcing performance have been conceptualized in chapter 4.2.1, the preceding chapter 4.3 has introduced hypotheses on the causal linkages between the variables. While for all ten variables modeled as antecedents of outsourcing performance causal linkages with each other are hypothesized, direct causal linkages on outsourcing performance are only theorized for the constructs cooperation, proactive improvement, and functional conflict. Table 4-1 summarizes the hypotheses put forward until this point. Table 4-1. Overview of the hypotheses on the logistics outsourcing performance model Hypotheses H1a

Cooperation has a positive effect on goal achievement

H1b

Cooperation has a positive effect on goal exceedance.

H2

Cooperation has a positive effect on functional conflict.

H3

Cooperation has a positive effect on proactive improvement.

H4

Communication has a positive effect on trust.

H5

Communication has a positive effect on functional conflict.

H6

Communication has a positive effect on cooperation.

H7

Communication has a positive effect on proactive improvement.

H8

Communication has a negative effect on opportunism.

H9a

Proactive improvement of the LSP has a positive effect on goal achievement.

H9b

Proactive improvement of the LSP has a positive effect on goal exceedance.

120

4 Antecedents and effects of logistics outsourcing performance H10

Trust has a positive effect on relationship commitment.

H11

Trust has a positive effect on cooperation.

H12

Trust has a positive effect on functional conflict.

H13

Relationship commitment has a positive effect on cooperation.

H14a

Functional conflict has a positive effect on goal achievement.

H14b

Functional conflict has a positive effect on goal exceedance.

H15

The involvement of the LSP has a positive effect on cooperation.

H16

The involvement of the LSP has a positive effect on communication.

H17

The involvement of the LSP has a positive effect on proactive improvement.

H18

Opportunistic behavior has a negative effect on trust.

H19

Opportunistic behavior has a negative effect on cooperation.

H20

Shared values have a positive effect on trust.

H21

Shared values have a positive effect on relationship commitment.

H22

Shared values have a positive effect on cooperation.

H23

Shared values have a negative effect on opportunistic behavior.

H24

Shared values have a positive effect on communication.

H25

Shared values have a positive effect on openness.

H26

Openness has a positive effect on communication.

H27

Openness has a positive effect on trust.

H28

Openness has a positive effect on functional conflict.

The different hypotheses as a whole serve as the basis for a conceptual model. It is displayed in Figure 4-2 and will be empirically tested in great detail after the operationalization of the constructs in chapter 7.

4.4 Effects of logistics outsourcing performance H22

Shared Values

H21 H20

Commitment

H1a Goal Achievement

H15

H11 H10

H3 H17

Trust

H24

H27 H12

Openness

H23

H28 H18

H2

H4 H16

H9a

H14a

Proactive Improvement

H1b H9b

H19 H6

H7

H26

H14b Functional Conflict

Opportunism

Cooperation

H13

Involvement

H25

H8

121

Goal Exceedance

H5 Communication

Fig. 4-2. Conceptual logistics outsourcing performance model

4.4 Effects of logistics outsourcing performance As it has been argued above, the design of the relationship between the customer and the LSP is hypothesized to have a positive effect on the logistics outsourcing performance. The question of these performance effects, however, is only a strategic one if a connection can be shown between outsourcing performance and both the resulting logistics and firm performance of the customer. If those direct or indirect linkages could not be shown, the aspired finding of this study demanding customers to increase the focus on the relationships with their LSPs would have to be rejected as the resources could be employed more efficiently elsewhere in the firm. Consequently, the relationship model introduced in chapter 4.3.2 must be extended. Even though a study by DEHLER (2001) shows the existence of some causal linkages between logistics performance and firm performance, with a study by ENGELBRECHT (2004) supporting these results, finding also that logistics outsourcing performance positively influences logistics performance, another empirical testing will be conducted in this study to validate the results. This is justified both because of the strategic relevance of the question to the research in this study and the open issues that could not be addressed or resolved by the studies cited above, e.g. concerning causal linkages that were found to be non-significant in either one of the cited works or the connection between the two-dimensional outsourcing performance construct and both logistics and firm performance. Also,

122

4 Antecedents and effects of logistics outsourcing performance

the two-dimensionality of the logistics outsourcing performance construct as such requires empirical testing. The extension of the relationship model will help answering research question three introduced in chapter 2.4.2, which seeks to explore the direct influence of logistics outsourcing performance on logistics performance and its indirect influence on firm performance. In the following, the constructs measuring logistics and firm performance will be introduced and hypotheses will be put forward to develop separate causal models investigating the effects of outsourcing performance of logistics and firm performance. 4.4.1 Logistics performance Logistics performance as argued above is influenced by a combination of two inputs: on the one hand by the performance of the logistics processes outsourced to third parties and on the other hand by the performance of logistics processes still performed in-house by the firm. As CHOW/HEAVER/HENRIKSSON (1994) point out, logistics performance has been studied intensively by a number of logistics scholars, who have defined and measured performance in a variety of different ways. Important works include those of MENTZER/KONRAD (1991), CHOW/HEAVER/ HENRIKSSON (1995), GASSENHEIMER/STERLING/ROBICHEAUX (1996), STANK/GOLDSBY/VICKERY (1999), DEHLER (2001), STANK/KELLER/ DAUGHERTY (2001), STANK/GOLDSBY/VICKERY/SAVITSKIE (2003), KNEMEYER/MURPHY (2004) and ENGELBRECHT (2004). As CHOW/HEAVER/HENRIKSSON (1995, p. 296) argue, logistics performance is not a straightforward construct universally agreed upon. It rather is multi-dimensional, reflecting multiple stakeholders and interests that depend on the current situation and context. Not surprisingly, the desired outcomes from logistics are numerous and range from customer satisfaction over issues such as environmental responsibility to overall costeffectiveness. Depending on the given situation, the measurement of logistics performance therefore must be adapted. As it has been argued above, one goal of this study is to identify the effect the performance of outsourcing arrangements has on overall logistics performance of the customer. Logistics outsourcing performance in this study is measured through the bi-dimensional construct goal achievement and goal exceedance, as introduced in chapter 4.1. However, this construct only enables an insight into the performance of that particular outsourcing arrangement. It does not provide evidence whether or not outsourcing has increased or decreased the overall logistics performance of the firm. It thus

4.4 Effects of logistics outsourcing performance

123

measures the output and not the outcome of logistics outsourcing. To evaluate the outcome, the overall logistics performance of the firm must be measured in order to later on investigate the relationship between outsourcing and logistics performance using structural equation modeling. The conceptualization of the construct of logistics performance will be based on the scale developed and tested by DEHLER (2001, pp. 206-225) and validated as well as refined by ENGELBRECHT (2004, pp. 218-224). As ENGELBRECHT (2004, p. 218) argues, measuring logistics performance this way is most suitable for evaluating performance effects of logistics outsourcing arrangements and therefore fulfills the demand presented above that for the given context and intentions, the proper logistics performance measurement method must be selected. Consequently, the following two chapters will conceptualize the constructs and develop hypotheses on the causal linkages between logistics outsourcing and logistics performance. The operationalization of the constructs will occur in chapter 6.3 in line with the structure of the other chapters of this study. 4.4.1.1 Conceptualization

DEHLER (2001, p. 208) argues that logistics performance consists of the two dimensions of logistics costs and the level of logistics services. While the level of logistics services represents a firm’s capability to supply the customers timely, reliably, and flexibly with qualitatively immaculate products that suit the demand of the market, logistics costs comprise all costs that are incurred in order to provide the chosen level of logistics services of the firm. Since considerable differences in the understanding of logistics services and costs exist, depending on different points of view and intentions, the two constructs will be conceptualized in detail in the following. As WEBER (1995, pp. 90-97) points out, the exact contents and the scope of the term logistics costs are widely discussed and as yet, no common understanding exists. Which costs are subsumed under the term logistics costs largely depends on the internal logistics organization of the firm. Since in the course of this study the absolute amount of the logistics costs is of no importance, because the participants in the survey will be asked to give a subjective estimate of the relative percentage of their firm’s logistics costs with respect to the competition, the exact differentiation of the term logistics costs only needs to be discussed as far as it is required to develop a reliable and valid construct. To conceptualize the construct of logistics costs, DEHLER (2001, p. 120) uses the findings of BAUMGARTEN/BOTT/HAGEN (1997, p. 24), who

124

4 Antecedents and effects of logistics outsourcing performance

pragmatically differentiate three clusters. These are based on their empirical findings of the percentage of firms that perceive the individual costs as part of their logistics costs. The first cluster of logistics costs contains those costs that were indicated by most of the firms, such as warehousing and transportation costs. The second cluster comprises the extended logistics costs that are perceived by at least a third of the firms as part of the logistics costs. Those include costs for picking and packaging as well as costs of capital employed or product returns. The third cluster is formed by the marginal logistics costs, indicated by less than a third of the firms to be perceived as logistics costs. Those include IT-costs for logistics processes, disposal costs or costs for product warrantees. While this differentiation certainly lacks some theoretical and scientific rigor, it does provide a valuable framework for the conceptualization of logistics costs in this study, as it empirically indicates the basis of the understanding of logistics costs among the firms. For reasons of understandability among the participants in the survey and general comparability, the operationalization in chapter 6.3 will mainly focus on the core logistics costs with some general aspects of extended logistics costs, since these constitute the lowest common denominator virtually all firms can agree upon. In the empirical studies of DEHLER (2001, p. 211) and ENGELBRECHT (2004, p. 220) this has proven to be valid, reliable, and effective and will therefore be the understanding of logistics costs in this study. The second dimension of logistics performance as conceptualized by DEHLER (2001, pp. 207-209) is the level of a firm’s logistics services. As it has been argued above, it represents a firm’s capability to supply the customers with the right products of the right quality at the right time in the right place, which is incorporated also in the “4R” concept frequently cited in practice. The level of logistics services according to PFOHL (1996) is experienced by customers through factors such as delivery times, delivery reliability, delivery flexibility, and delivery quality. To conceptualize the construct of the level of logistics services for this study, a framework must be employed that systematizes the concept. ENGELBRECHT (2004, pp. 221-222) points out the difficulty that is connected to its exact definition and measurement. While cost orientation for performance measurement according to WEBER (1987, pp. 107-108) dominates also in a logistics context, measuring the level of logistics services proves more demanding. Consequently, several different measurement approaches exist, centering mostly on a narrow service concept. Since they are most often slightly diffuse and lack adequate contour, ENGELBRECHT (2004, pp. 221-222) suggests to employ a framework de-

4.4 Effects of logistics outsourcing performance

125

veloped by BOWERSOX/CLOSS/HELFERICH (1986, pp. 27-28) to measure the level of logistics services. The framework divides logistics services into three successive levels. On the first level, a firm ensures the availability of logistical objects and resources. This comprises the ability deliver products, material, and information to on-time. Securing this ability is the primary and most elementary service requirement for logistics and constitutes a hygiene factor (HERZBERG 1968, pp. 55-62) that must be fulfilled before other performance increases can be targeted. After the ability to deliver is guaranteed, firms can begin to improve the performance capabilities of their logistics processes on a second level. This includes especially the improvement of delivery times and delivery flexibility. The first factor stands for the ability to convert customer orders into product output suitable for the market, while flexibility measures the logistics processes’ capability to react to present and future changes in customers’ demands. After the first two levels of logistics services are fulfilled, firms must ensure the permanent quality of their logistics processes. This aims at increasing the reliability of the material and information handling processes which is reflected in performance indicators such as delivery reliability or the percentage of damage- and error free logistical handlings. Since the performance indicators as shown above in their entirety grasp the complexity of logistics services and have been shown by ENGELBRECHT (2004, pp. 222-224) to be both valid and reliable, indicators from all three levels will be utilized for the measurement of a firm’s individual levels of logistics services in chapter 6.3. In the following chapter, hypotheses will be developed that address the causal linkages between the logistics outsourcing performance and the newly developed constructs logistics costs and level of logistics services, which together constitute the construct logistics performance. 4.4.1.2 Effects of logistics outsourcing performance on logistics performance

As it has been argued in chapter 4.1, the logistics performance of a firm is basically a combination of two different input variables. One the one hand, it is affected by the performance of in-house logistics processes performed by or under the direct responsibility of the customer. On the other hand, it is influenced by the performance of logistics outsourcing arrangements in which the customer has delegated the process and the accompanying responsibility to a third party. It must therefore be assumed that the quality

126

4 Antecedents and effects of logistics outsourcing performance

of logistics outsourcing has a significant influence on the actual logistics performance of the firm. This causal linkage was first hypothesized by ENGELBRECHT (2004, pp. 277-278). Measuring logistics outsourcing performance both through the goal achievement of the customer and the information exchange between the parties, he supposed positive relationships between the two dimensions and logistics performance, measured through the level of logistics services and the logistics costs. Testing this model empirically, he concluded that while the information exchange positively influences both dimensions of logistics performance on a statistically significant level, goal achievement only positively influenced the reduction of logistics costs. From this, ENGELBRECHT (2004, p. 278) concludes that the quality of the cooperation and the logistics outsourcing process has a direct and lasting positive influence on the firm’s logistics performance. Starting from these findings, the hypotheses can be adapted for the model of logistics outsourcing performance employed in this study, which as shown in chapter 4.1, measures outsourcing performance through a bidimensional construct consisting of goal achievement and goal exceedance and views information exchange or communication as one of its antecedents. It can be supposed that an increase in goal achievement, which reflects both an increase in service levels such as quality and time and a reduction in cost levels, positively influences the overall level of logistics services of the firm and the corresponding overall level of logistics costs. Similarly, goal exceedance grasps whether or not the expectations of the customer regarding the logistical service levels and the cost reductions through the outsourcing arrangement have been significantly exceeded. If that is accomplished through logistics outsourcing, thereby increasing goal exceedance, it can be expected that a positive influence on the overall logistics performance will exist. Therefore, a positive relationship on both the level of logistics services and the logistics costs will be supposed. In a formalized fashion, the four hypotheses developed above read as follows: H29: Goal achievement has a positive effect on the level of logistics services. H30: Goal achievement has a positive effect on the level of logistics costs.

4.4 Effects of logistics outsourcing performance

127

H31: Goal exceedance has a positive effect on the level of logistics services. H32: Goal exceedance has a positive effect on the level of logistics costs. The structural model containing the information from the hypotheses presented above is specified in Figure 4-3.

Goal Achievement

Level of Logistics Services

H29 H31 H30

Goal Exceedance

H32

Level of Logistics Costs

Fig. 4-3. Conceptual logistics performance model

4.4.2 Firm performance Installing effective and cost efficient logistics processes is not an end in itself for firms, but rather an integral part of the effort to reach the overall performance goals of the firm which most commonly are defined as financial objectives (ENGELBRECHT 2004, p. 225). Logistics outsourcing research therefore draws its relevance from the implicit assumption that logistics performance has a measurable and relevant influence on firm performance. As proposed in research question three in chapter 2.4.2, empirical testing of this relationship is a central problem for logistics research and thus will be further analyzed in this study. The measurement of firm performance has received much attention in business administration research, which is reflected in the multitude of different approaches suggested (BHARGAVA/DUBELAAR/RAMASWAMI 1994). Fundamentally, it can be distinguished between “hard” and “soft” performance measures (DALTON/TODOR/SPENDOLINI/FIELDING/PORTER 1989, p. 50), which DEHLER (2001, p. 226) terms objective and subjective. Hard – or objective – performance criteria include sales, gross profit, production or growth while soft – or subjective – measures rely on the indi-

128

4 Antecedents and effects of logistics outsourcing performance

vidual perception of the respondent, be it manager or employee. While DALTON/TODOR/SPENDOLINI/FIELDING/PORTER (1980, p. 50) consider subjective performance criteria to be less indicative of the “bottom line organizational performance”, they will be used for the measurement of firm performance in this study. The reasons for the selection of subjective performance criteria lie in the numerous difficulties that come along with gathering and analyzing objective performance criteria (DEHLER 2001, p. 227). Their suitability for industry spanning empirical research is limited as they strongly rely both on the situation of the firm and the industry. Different levels of transfer prices for instance lower the general comparability of performance indicators just as much as management’s use of its discretion to influence performance measures and balancing procedures prior to supplying the data. Furthermore, asking for hard and objective performance measures in survey research carries the danger that informants will not at all or not truthfully respond to the questions fearing lacking anonymity. These difficulties, coupled with the findings of several studies showing very high correlations between subjective and objective performance measures (DESS/ROBINSON JR. 1984; VENKATRAMAN/RAMANUJAM 1986; NAMAN/SLEVIN 1993; HART/BANBURY 1994), justify the use of subjective performance measures in this study. DEHLER (2001, p. 227) points out the importance of a parallel measurement of different components of firm performance that addresses the plurality of different objectives and enables the differentiated observation of the effects of the two dimensions of logistics performance on different components of firm performance. For this study, the three factors of adaptiveness, market performance, and firm performance have been selected as dimensions of the construct of firm performance. While it was originally developed in the marketing domain (RUEKERT/WALKER JR./ROERING 1985; IRVING 1995), DEHLER (2001) and ENGELBRECHT (2004) demonstrated the suitability of the measures also in a logistics context. The following chapter will conceptualize the three dimensions before in chapter 4.4.2.2 hypotheses on the effects of logistics performance on firm performance will be introduced. 4.4.2.1 Conceptualization

Firm performance, as understood in this study, consists of the three dimensions adaptiveness, market performance, and financial performance. Before they are conceptualized it must be noted that they are not on the same logical level but must rather be understood as interdependent and sequential factors with financial performance as a terminal point.

4.4 Effects of logistics outsourcing performance

129

Since DEHLER (2001) and ENGELBRECHT (2004) have conceptualized the construct in detail and have generally shown its suitability for logistics outsourcing research, the following paragraphs will focus on the most important aspects of the three factors. A further in-depth presentation of the factors will be performed during the operationalization in chapter 6.4. The factor adaptiveness, sometimes also called adaptability or responsiveness, reflects the ability of the firm to adapt to changes in the environment (RUEKERT/WALKER JR./ROERING 1985, p. 15; IRVING 1995). This describes the capability to flexibly react to changes in the environment and varying demands of the customers. Therefore, a firm with high levels of adaptiveness will be able to react quickly to new developments in the market and benefit from the arising opportunities. Market performance is a measure for the effectiveness of a firm in the market (RUEKERT/WALKER JR./ROERING 1985; IRVING 1995) and therefore is a factor of strategic importance. High market performance can lead to increasing levels of customer satisfaction, customer loyalty, and benefits for the end-customer. Further indicators of a high market performance beyond the customer oriented criteria are the achievements of the planned degrees of growth and market share. Finally, the third dimension of firm performance measures the economical success of a firm. Following the argumentation of ENGELBRECHT (2004, p. 229), financial performance is understood as the ultimate indicator of firm performance and will be measured through the level of the firm’s revenue margin compared to that of competing firms as will be argued in detail in chapter 6.4.3. The three dimensions must be considered in their entirety to gain a comprehensive understanding of the effect logistics performance has on the different aspects of firm performance. Hypotheses will be developed in the next chapter that describe the possible causal linkages between the constructs. 4.4.2.2 Effects of logistics performance on firm performance

This chapter will introduce the hypotheses on the causal linkages between the two factors of logistics performance and the three factors of firm performance. It will theorize that a strong indication exists suggesting that logistics performance has a positive effect on the overall firm performance. As mentioned above, this relationship has first been analyzed by DEHLER (2001, pp. 226-244), the results later being confirmed by ENGELBRECHT (2004, pp. 251-255). As Figure 4-4 demonstrates, they found that while the level of logistics costs directly influences the financial performance, the level of logistics services only directly affects the adap-

130

4 Antecedents and effects of logistics outsourcing performance

tiveness and the market performance of the firm. However, since the causal linkages between adaptiveness, market performance and financial performance are very strong, a powerful positive indirect effect from the level of logistics services on the financial performance of the firm exists. In fact, as DEHLER (2001, p. 242) points out, the total effect of the level of logistics services on the financial performance is 0.405, while for the level of logistics costs it is only 0.216, both significant at the 1%-level. Dehler, 2001

R2: 28.7%

Engelbrecht, 2004

R2: 10%

Adaptiveness

Adaptiveness

0.54*** Level of Logistics Services

0.31***

n.s.

0.47***

n.s.

Level of Logistics Services 2

0.47***

R : 66.2%

0.68***

Market Performance

n.s. 0.24***

n.s. 0.50*** R2: 40.6% Financial Performance

***: **: *:

R2: 53% Market Performance

n.s. Level of Logistics Costs

0.13*

Level of Logistics Costs

n.s. 0.16*

0.41*** R2: 23% Financial Performance

1%-level of significance 5%-level of significance 10%-level of significance

Fig. 4-4. Firm Performance Model developed by DEHLER (2001, p. 241) and ENGELBRECHT (2004, p. 254)

Additional to those findings it must be mentioned that several causal linkages originally hypothesized by DEHLER (2001, pp. 233-239) and ENGELBRECHT (2004, pp. 251-255) could not be supported as they were found to be non-significant. As Figure 4-4 indicates, this is the case for the effects of the level of logistics costs on both adaptiveness and market performance as well as for the effect of the level of logistics services on the financial performance. As it has been argued above, the relevance of the causal linkages between logistics performance and firm performance for answering the research questions introduced in chapter 2.4.2 and the resulting research model justifies another empirical testing of the relationship. This is especially relevant in the light of the aspired testing of the moderating effects of situational factors. Therefore, the following paragraphs will briefly introduce the supposed hypotheses including those that did not find support in the studies of DEHLER (2001) and ENGELBRECHT (2004), as long as their theoretical validity can still be presumed.

4.4 Effects of logistics outsourcing performance

131

A positive relationship can be supposed between the level of logistics services and the adaptiveness of the firm. On the one hand, higher levels of logistics service capabilities enhance the ability of firms to react to changes in the product volumes demanded by the customers. On the other hand it can be presumed that the logistics competencies that were necessary and developed for increasing the level of logistics services will be advantageous in other fundamental internal change processes, as for instance the adaptation of distribution channels (ENGELBRECHT 2004, pp. 252-253). Further support for the hypotheses can be found in the argumentation by DEHLER (2001, p. 235), who states that with increasing levels of logistics capabilities, postponement strategies can be implemented that increase the flexibility of the firm to react to demands of the market. The level of logistics services not only has a positive effect on the adaptiveness of the firm, but also on its market performance. Market performance depends very strongly on the satisfaction of customers with the purchased products or services. This customer satisfaction, however, does not only come from the primary capability of the firm to produce a product or service, but also from the excellence of secondary capabilities such as customer service or logistics. Several empirical studies underline the importance of logistics and service elements for the customer (BALLOU 1992; INNIS/LA LONDE 1994; DAUGHERTY/STANK/ELLINGER 1998; EMERSON/ GRIMM 1998). In fact, DAUGHERTY/STANK/ELLINGER (1998, p. 36) state that in many instances, the “distribution service is deemed more important than product quality or price in establishing customer satisfaction”, e.g. when different mail order business all sell the same products. Beyond the positive influence of the level of logistics services on the adaptiveness and the market performance, it can also be argued that it positively affects the financial performance of the firm. This can occur on the one hand indirectly via the positive influence of market performance on financial performance. Higher levels of logistics services lead to increased customer loyalty and serve as a facilitator to win new customers which increases the market performance. This, as it will be argued in hypothesis H34, enables a better utilization of the internal resources and a broader allocation of fixed costs, both directly affecting the profitability of the firm. On the other hand, increased levels of logistics services potentially allow the firm to increase prices. If prices can be achieved that are higher than the increased costs, this will directly lead to a better financial performance of the firm. It can therefore be supposed that with higher levels of logistics performance, also the financial performance will rise. In a formalized fashion, the first three hypotheses read as follows:

132

4 Antecedents and effects of logistics outsourcing performance

H33: The level of logistics services has a positive effect on adaptiveness. H34: The level of logistics services has a positive effect on market performance. H35: The level of logistics services has a positive effect on financial performance. Aside from the effects of the level of logistics services, also the level of logistics costs has an influence on the three dimensions of firm performance. High logistics costs can be argued to restrict the adaptiveness of the firm, since any increase in a given logistical service level is accompanied with a higher level of logistics costs (DEHLER 2001, pp. 235-236), leading to less flexibility and short term maneuverability of the firm. A reduction of the level of logistics costs therefore can be viewed as a relaxation of this restriction and therefore will enable the firm to display a higher degree of adaptiveness. Competitive advantage can be realized not only through differentiation of the service level, but also through cost reductions which enable firms to lower prices. Since the logistics costs often form a substantial part of the overall costs of a firm, a logistics cost reduction will directly enable price reductions which in turn will lead to higher customer satisfaction and therefore, as argued above, to higher market performance. Finally, there is a direct mathematical relationship between the level of logistics costs and the financial performance. The financial performance, being basically a function of the relation between sales and costs, will increase at the same amount at which the logistics costs are lowered. Following the above argumentation, three hypotheses can be formulated: H36: The level of logistics costs has a positive effect on adaptiveness. H37: The level of logistics costs has a positive effect on market performance.

4.4 Effects of logistics outsourcing performance

133

H38: The level of logistics costs has a positive effect on financial performance. After having investigated the causal linkages between logistics performance and firm performance, the following paragraphs will investigate the relationships between the individual factors of firm performance. Causal linkages will be theorized for the relation between adaptiveness, market performance, and financial performance, accounting for the interdependent and sequential nature of the relationships between the factors. The adaptiveness of a firm describes its capability to adjust to changes in market demands. It enables the firm to adapt its products and services to altered customer needs and to react flexibly to new market developments. It can thus be expected that firms with high levels of adaptiveness can better address the needs of their customers, thereby increasing customer satisfaction which in turn is an important prerequisite for market performance. It can therefore be argued that adaptiveness has a positive influence on market performance. Market performance and its revenue effects, which are determined at large by the offered products and services as well as the corresponding prices, in turn directly influences the economical success of the firm and thereby its financial performance. This relationship has been shown in several studies on customer satisfaction and customer loyalty (FORNELL 1992; ANDERSON/SULLIVAN 1993; RUST/ZAHORIK 1993; REICHELD 1996), performance measurement of marketing (AMBLER/KOKKINAKI 1997) and empirical success factor research such as the PIMS study (BUZZEL/GALE 1987). The performance effects as demonstrated in these studies stem from several effects: increasing customer loyalty leads to decreased customer acquisition costs, to higher customer profitability due to longer relationships and to a higher tolerance towards increased prices (REICHELD 1996), while strong growth and high market share also influence the economical performance of a firm positively (CAPON/FARLEY/ HOENIG 1990; SZYMANSKI/BHARADWAJ/VARADARAJAN 1993). After the above presented argumentation, the following two hypotheses can formally be introduced: H39: Adaptiveness has a positive effect on market performance. H40: Market performance has a positive effect on financial performance.

134

4 Antecedents and effects of logistics outsourcing performance

In Figure 4-5, the eight hypotheses developed in this chapter on the causal linkages between logistics performance and firm performance will be graphically displayed. The resulting structural model will be tested in chapter 7.2.2.

Adaptiveness

H33 Level of Logistics Services

H36

H39

H34 Market Performance

H37 Level of Logistics Costs

H40 H38

H35 Financial Performance

Fig. 4-5. Conceptual firm performance model

4.5 Moderating effects In chapters 4.3 and 4.4, three models have been developed that allow general insights into the nature of logistics outsourcing performance on the one hand and into the relationship between outsourcing performance, logistics performance, and firm performance on the other. While the explanatory value of these models can be expected to be very high for the designing of relationships between customers and LSPs, the analysis of the role of moderating factors in line with research question number four is promising to provide additional detailed knowledge. The following chapter will provide some background information on the importance of moderating analyses and the selection of adequate variables before chapter 4.5.2 will introduce the conceptualization of the variables utilized in this study.

4.5 Moderating effects

135

4.5.1 Relevance of adequate contingency variables Before a selection of suitable contingency variables can take place, the relevance of moderating analyses in the logistics context must first be established. This will only be the case if the customers of logistics service providers and the corresponding markets can be shown to be sufficiently heterogeneous to justify the expectation of substantial and individual differences that may influence the models of logistics outsourcing performance as well as logistics- and firm performance. A strong indication for the existence of these differences, which in turn would suggest the possible existence of moderating effects, can be found through the empirical observation of current logistics outsourcing practices. LSPs react to the heterogeneity of their customers by providing a diversified service portfolio whose orientation varies according to the aspired target markets. As will be briefly shown in the following, contract logistics providers as introduced in chapter 2.2.3.2 show distinct organizational differences depending on their customers’ particular situation, e.g. in the dimensions industry, size, geography, or products. As LIEB (2005, p. 23) points out, the earlier attempts by some LSPs to provide every possible service to every industry is “clearly an unworkable strategy”. Instead, he argues, the major logistics service providers have since abandoned this approach and now tend to focus their efforts on a limited number of industry verticals. By doing so, they have developed industry-specific knowledge as a means of differentiating themselves from the competition (LIEB 2005, p. 23). Consequently, the industry of the LSPs’ customers and all its aspects apparently have a sufficiently strong influence on the outsourcing arrangements that LSPs react with organizational change, hence suggesting the existence of moderating effects since if all customers and the corresponding requirements were the same, the logistics service providers would not face the need for organizational adaptation. This argumentation is also supported by the findings of GONZALEZ (2005, pp. 10-12) who sees large LSPs tailoring their services to vertical industries and smaller, niche providers having chosen to accommodate industries only with very specific demands. LIEB (2005, p. 23) also suggests that the differing sizes of their customers have prompted LSPs to adapt their organization structure. As large customers with more standardized logistics processes promise higher margins and yields for logistics service providers, the LSPs have become increasingly customer selective. This is leading to a situation in which the most promising customers resort to the largest LSPs as they offer the most suitable outsourcing arrangements, and many small and medium sized customers are finding themselves with little alternatives for third party logis-

136

4 Antecedents and effects of logistics outsourcing performance

tics services as they have become less attractive customers due to their particular logistics processes (LIEB 2005, p. 23). Further evidence for the diversity of logistics outsourcing customers that suggests the existence of moderating effects is brought forward by QUINN (2005). He states that aspects, such as an increasing geographic complexity caused by the level of the globalization of the customers’ operations, have provoked LSPs to rethinking their organizational structures, developing new capabilities and services, and collaborating more closely with their customers (QUINN 2005, pp. 3-4). The same rethinking as initiated by the geographical complexity can be found when the complexity of the customers’ products force the LSPs to develop new capabilities and services. Summing up these observations, a clear indication is given that logistics service providers have adapted their organizational structures as a reaction to the heterogeneity of their potential customers. Consequently, it must be assumed that this heterogeneity may be responsible for the necessity of different LSP behaviors in different outsourcing situations and contexts, thus suggesting the existence of moderating effects especially in the models of logistics outsourcing performance and logistics performance proposed in chapters 4.3 and 4.4. While the argumentation presented above has concentrated primarily on the importance of moderating analyses in the context of logistics outsourcing arrangements, the study at hand provides the opportunity to also test the potential relevance of the contingency variables on the linkage between logistics performance on firm performance. WEBER (2003, pp. 16-18) argues that depending on the differentiation potential of logistics services for a firm, the impact of logistics performance on the firm performance varies. While in spot market environments with clearly defined logistical requirements the differentiation potential through logistics service quality is very limited while at the same time carrying the risk of producing unjustifiable extra costs, logistics service quality may be the central differentiation factor in long-term relationship with changing and demanding logistical requirements. This clearly indicates that depending on the context of the firm, the effect of logistics performance on firm performance may vary. Further support for this argumentation is provided in the following: Logistics performance is appreciated as a driver of firm performance by many logistics executives (DEEPEN 2003, pp. 139-140). The degree of perceived impact varies, however, from a majority that views logistics as strategically important for their firms to a minority that attributes only a lesser strategic importance. It can be argued that executives that view logistics as strategically important do so because they recognize its potential influence on the firm’s performance and vice versa. Consequently, logistics performance is understood by some firms as relatively unimportant for their

4.5 Moderating effects

137

firm performance, an example for which could be the document logistics function in a retail bank, while others identify it as very relevant, e.g. component logistics in the automotive industry. This clearly indicates the existence of differences among the firms and hence suggests potential moderating effects. While the above argumentation suggests the potential existence of moderating effects also in the firm performance model, it must be acknowledged that the variables exerting influences on this model may substantially go beyond those affecting the logistics outsourcing performance and logistics performance models. However, as will be argued in the following chapter, the variables at hand in this study provide the opportunity to exploratively test the relevance of moderating effects also for the firm performance model. Further analyses going beyond the scope of this study were deliberate postponed and constitute a future research opportunity. Aside from the conceptual insights presented above, moderating effects have also been studied intensively in different areas of business research where they have been found to exert significant influence. Sociological variables that have been intensively researched and have been found to moderate decision- or behavioral models include attributes such as gender (SLAMA/TASHCHIAN 1985; ZEITHAML 1985; JASPER/LAN 1992; GILBERT/ WARREN 1995), age (WALSH 1982; MOSCOVITCH 1982; JOHN/COLE 1986; SMITH/BALTES 1990), or income (FARLEY 1964; SCHANINGER/ SCIGLIMPAGLIA 1981; ZEITHAML 1985; SPENCE/BRUCKS 1997). However, increasingly also models in exchange relationship research that scrutinize effects such as performance, customer satisfaction, and customer loyalty are examined with moderating analyses (GIERING 2000; DEHLER 2001; WALLENBURG 2004). Even though the research in this area must be considered less mature than in the case of the sociological variables that have been in the focus of a far larger amount of studies, significant effects of a large number of variables were found. GIERING (2000, pp. 100-153) argues that the link between customer satisfaction and customer loyalty is moderated by variables from the following five different categories: x Relationship characteristics: trust, information exchange, cooperation, flexibility of the supplier, duration x Customer characteristics: centralization, structural disturbances, risk aversion of the management, behavior under uncertainty, variety seeking, involvement, social persuasibility x Product characteristics: importance for the customer, complexity x Supplier characteristics: reputation, value added for the customer

138

4 Antecedents and effects of logistics outsourcing performance

x Market characteristics: alternatives, technological dynamics, intensity of competition. GIERING (2000, pp. 168-195) empirically finds these variables to have moderating effects on the link between customer satisfaction and customer loyalty, thus documenting the importance of moderating analyses in exchange relationship models. Other studies show mixed results: DEHLER (2001, pp. 244-252) finds that the causal linkages between logistics performance and firm performance are moderated by market dynamics, but not by the intensity of competition. In a study conducted by WALLENBURG (2004, pp. 255-257), neither the level of logistical development nor the strategic importance of logistics for the customer are found to moderate the link between customer loyalty and its determinants. The above presented research findings demonstrate the importance and potential of moderating analyses for exchange relationship research in general. Logistics research so far has only reluctantly taken up this lead. CHOW/HEAVER/HENRIKSSON (1994, p. 26) demand the use of contingency models in logistics performance research. KNEMEYER/CORSI/MURPHY (2003, p. 79) argue that the key to performance is to obtain the kind of relationship between an LSP and its customer that is most appropriate given the business situation, thereby suggesting implicitly the need for analyses of the business context. As chapter 3.2.4.2 has argued, the contingency approach is offering a starting point for the identification of moderating variables in the logistics context. So far, logistics research has not substantially extended beyond the initial conceptual stage. PFOHL/ZÖLLNER (1997, p. 307) suggest a large body of different contingency variables that can be subsumed under the four categories complexity of environmental relations and dynamics of environmental relations for both the flow of products and the flow of information. A detailed description was given in chapter 3.2.4.2. KLEER (1991, pp. 121-124), as outlined in the same chapter, conceptually identifies similar contingency variables but utilizes the segmentation approach by KIESER/KUBICEK (1983, p. 222) to distinguish between internal and external contingency variables. As the following chapter will show, the variables identified by KLEER (1991) will be selected for the moderating analyses in this study as they are most suitable, because they are specifically developed for logistics relationships and are conceptualized broad enough for a later operationalization as a multi-indicator construct. Additionally, selected other variables will also be examined. However, due to the still very immature state of this field of research and the lack of an adequate theory, unlike in the preceding chapters no hypotheses for the moderating effects of the variables will

4.5 Moderating effects

139

be derived. Instead, the variables will first be in detail conceptualized in the following. In chapter 7.3, the moderating effects on the models of outsourcing-, logistics-, and firm performance will then be exploratively tested. 4.5.2 Conceptualization of contingency variables For the identification of the contingency – or situational – variables to be tested in this study it will largely be relied on the work of KLEER (1991), who analyzes the designing of logistics outsourcing relationships via a pure contingency approach. The proposed contingency variables are explicitly developed to cover the situational factors important for relationships between customer and LSP and focus solely on the situation of the customer, not on the context of the relationship. This approach was chosen also for this study since it is the customer who makes the outsourcing decision and can therefore choose a partner that fits its needs sufficiently well. Therefore, contingency variables of the customer are of primary importance. The analyses of the moderating effects of contingency variables influencing the relationship or the LSP would also be desirable. However, analyses that go beyond the scope of the research aims of this study were deliberately postponed for future research. Following the argumentation of KIESER/KUBICEK (1983), KLEER (1991, pp. 120-123) distinguishes between external and internal contingency variables which will be presented in the next two chapters. As already argued in chapter 3.2.4.2, external contingency variables are those suited to explain differences between organizational structures that cannot be altered by the organization alone, but depend on other organizations and thereby describe the relationship of the organization to its environment. Internal contingency variables on the other hand are those the organization can influence by itself. 4.5.2.1 External contingency variables

External contingency variables can generally be divided into two groups based on the dimensions environmental complexity and environmental dynamics. While complexity is primarily referring to relations with customers and comprises the logistics channel structure as well as the number and the diversity of logistics relevant customer relationships, dynamics describe the changes in the relationships with respect to both the customers and the competition. The two groups will be detailed in the following.

140

4 Antecedents and effects of logistics outsourcing performance

The environmental complexity results from the number of partners in the logistics channel on the customers’ side and the nature of the relationships. According to KLEER (1991, p. 121), a first factor increasing the complexity is the number of different locations that products need to be delivered to. Furthermore, the more inhomogeneous the client structure of the LSP’s customer with respect to e.g. size and buying volume, the more demanding is the situation for the logistics system. Further complexity is added through inhomogeneous ordering of the customers that might differ depending on the specific order or the season. Finally, the number of LSPs the customer is working with on a regular basis is adding complexity to the logistics system. The environmental dynamics can be characterized through changes in both the competition and the relationships with the own customers as well as shifting perceptions of the own customers on the importance of logistics. Changes in the competitive landscape through more intense and fierce competition can for instance lead to changing market shares and hence to altered customer structures and corresponding sales volumes. Further dynamics result from changed relationships with customers. As the industries of the customers see increasing competitive pressure and tendencies of concentration, the consequence for the customer can be decreasing numbers of own customers with steadily increasing sales volumes and as a result a shifted balance of power. While KLEER (1991, p. 122) views the customer’s industry as only an indicator of the environmental dynamics, in this study it will be considered to be an individual contingency variable. This is justified because the industry is influenced both by environmental complexity and dynamics as well as several other factors that for reasons of complexity reduction are not subject of this research, like differing political parameters or industry specific regulations. After the preceding discussion of situational variables derived from the contingency approach, the transaction cost argumentation from chapter 3.2.1.2 that indicated that transactions can be characterized by the three critical dimensions asset specificity, uncertainty, and frequency will be taken up. It was argued that they constitute contingency variables since they directly characterize the situation the customer is affected by. While asset specificity and frequency will be classified as internal contingency variables that can be influenced by the customer, uncertainty clearly is an external variable beyond the customer’s control. Close to the variables environmental complexity and dynamics and yet not congruent, it measures the degree of uncertainty the customer has towards the future. This can originate from unpredictable or opportunistic customer behavior, from its changing needs and demands or from a general uncertainty towards the

4.5 Moderating effects

141

development of the customer’s or its customers’ industries. Hence, in this study uncertainty will be understood as related to the concepts of complexity and dynamics, but due to its nature as presented above it is sufficiently distinct and can be analyzed as an independent variable at the same logical level as the two other factors. This intentionally differs from the view often found that complexity and dynamics are determinants of uncertainty. In the context of this study this is justified as the variables originate from differing theoretical backgrounds, namely transaction cost theory and the contingency approach. They therefore constitute variables on the same logical level rather than uncertainty being a consequence of the two others. 4.5.2.2 Internal contingency variables

Internal contingency variables are all those factors that describe the internal situation of the customer and which it can theoretically change independently. They include the six variables derived from the contingency approach: the product range, the firm size, the degree of logistics centralization as well as the asset specificity, the frequency and the processes orientation. The products constitute a central aspect of the customer’s internal situation. Decisive for the designing of outsourcing relationship should be the range of products produced and the corresponding material value as these properties determine the usability of logistics infrastructure for instance for transportation, handling or warehousing processes. A further characteristic of different products is their substitutability which determines the opportunity of the customer to positively differentiate itself through logistics excellence. A second internal contingency variable is the size of the organization. With increasing size, the LSP’s customer can be presumed to have a higher number of production sites, a broader product range, and also a higher sales volume. All three aspects should have implications for the designing of logistics outsourcing relationships and the other models. The final variable derived from the contingency approach is the internal logistics organization. It is primarily the degree of centralization of logistics decisions that affects the relationship with the LSP. A high degree of centralization should not only accelerate the decision making processes as it facilitates communication and cooperation between the employees involved, but should also lead to a higher relationship intensity, while a low degree of centralization should tend to impede or hinder relationship formation. As argued in chapter 3.2.1.2, also the asset specificity and the frequency of the customer’s transactions constitute contingency variables. High asset

142

4 Antecedents and effects of logistics outsourcing performance

specificity implies the need for significant logistics investments on the side of the LSP which might not be fully or entirely retrievable after a termination of the contract. Knowing about high levels of asset specificity should therefore alter the customer’s behavior towards the LSP. Low levels might indicate the possibility to switch LSPs with little cost and effort while high levels point towards the economical necessity to remain in a given outsourcing arrangement as in an outsourcing arrangement with high asset specificity, the LSP will demand a price premium for its additional risk. A similar argumentation holds for the frequency, which has implications for the amortization of transaction-specific investments. Customers with very frequent transactions will seek longer and more intense relationships with LSPs since economies of scale and lower average costs per transaction substantially reduce transaction costs. Another contingency variable to be tested in this study is the process orientation of the customer. First shown by DEHLER (2001, pp. 220-226) to have a positive influence on logistics performance, its moderating effects on outsourcing performance, logistics performance, and firm performance will be analyzed. Process orientation implies the largely failurefree flow of materials and information inside the organization and towards its customers, excellent coordination between all partners required for the production of the goods and generally an orientation towards the achievement of all partners’ joint goals. The higher the process orientation of the firm, the higher is its level of logistics knowledge as presented in chapter 2.1.1. High process orientation can therefore be expected to positively affect the linkage between relationship variables and the outsourcing performance outcomes, while low process orientation must be presumed to have detrimental effects. 4.5.3 Overview of contingency variables The previous two chapters have introduced a variety of external and internal contingency variables that are supposed to have an influence on the effect relationship variables have on the logistics outsourcing performance and furthermore on the relation between outsourcing performance, logistics performance, and firm performance. Figure 4-6 gives an overview of the hypothesized contingency variables, which will in detail be operationalized in chapter 6.6. As argued above, no hypotheses will be developed for the moderating effects these contingency variables potentially exert on the different performance models. Instead, their effects will be subject to explorative testing. This is justified both because of the absence of an adequate theory for

4.5 Moderating effects

143

this specific context and the still very immature state of research in this field. This part of the study is therefore understood as an explorative step which could serve as one of several starting points for future research in this field. External contingency variables

Internal contingency variables



Environmental complexity



Products



Environmental dynamics



Size of the organization



Uncertainty



Degree of logistics centralization



Customer’s industry



Asset specificity



Frequency



Process orientation

Fig. 4-6. External and internal contingency variables

5 Methodology and sample characteristics

To test the hypotheses proposed in the previous chapter, the theoretical and conceptual insights must be challenged with reality through an empirical analysis. For this purpose, a large-scale survey was designed and conducted. This addresses the demand of MENTZER/KAHN (1995, p. 244) discussed in chapter 2.4.3 to apply the adequate scientific rigor also in logistics research which has also been the goal in previous chapters: In chapter 2, the idea for the study has been generated, the literature has been reviewed and reality has been observed, leading to a substantive justification of the research efforts. After in chapter 3 the theories have been introduced that allow an insight into possible research directions, chapter 4 has seen the conceptualization of the research models as well as the development of hypotheses which will form the backbone of the empirical analysis. The following chapter will focus on the methodological basis and the data base of the study. Chapter 5.1 will establish the research object and discuss possible research methods before introducing the questionnaire, its development, and the subsequent data collection. The resulting data base and the characterization will be presented in chapter 5.1.5. In the following, between chapters 5.2 and 5.2.5, the methodological basis for the empirical analysis will be set.

5.1 Survey design The empirical analysis aims at understanding the relationships between the buyers of logistics services from both retailing and manufacturing industries and their logistics service providers. Long-term relationships between those parties, as outlined in chapter 2.2.1 also termed contract logistics, therefore constitute the research object of this study. However, due to the existing degree of diversification and the employment of different LSPs in different parts of many firms, the focus here will be on strategic business units (SBU) and on their particular relationships to their most important LSP. This concentration ensures the focus of the research and facilitates

146

5 Methodology and sample characteristics

the isolated identification of different effects of the proposed relational factors on outsourcing performance that would not be possible if firms were asked to outline their relationships with all LSPs in general. 5.1.1 Methods for data analysis The selection of the method suitable to survey the data needed is directly connected to the research aim as well as the utilized method of analysis. Generally, several different statistical methods are suitable to analyze the relationships between different factors. However, a method is only appropriate if it fulfills a certain set of criteria with respect to the problem (PETER 1997, pp. 128-130; WALLENBURG 2004, p. 124), e.g. the research aims of this study. Relationships between different variables must be tested according to the proposed hypotheses. These variables, which in this study are mainly not directly observable, are called constructs. Examples are cooperation, trust, and commitment. Therefore, a basic requirement for the suitability of a research method is its ability to test causal relationships between constructs (criterion 1). For the analysis of the relationships between latent variables, which cannot be measured directly, these variables must be made measurable. For this, indicators are utilized which are empirically observable and reflect the characteristics of latent variables. The measurement of these indicators generally is subject to measurement errors which can lead to substantially wrong conclusion if they are not appropriately accounted for (HOMBURG 1989, p. 20). The research method therefore must allow for the consideration of measurement errors (criterion 2). Since it is hypothesized that causal linkages not only exist between the relationship variables and outsourcing performance but also among the relationship variables themselves, these interdependencies must be able to be considered by the method (criterion 3). In addition, simultaneous testing of the hypotheses must be possible to estimate the overall quality of the model (criterion 4). Generally, different multivariate methods of analysis based on the principle of regression analysis are potentially suitable to analyze causal relationships. Classical regression analysis, however, proves to be not suitable for the research aims of this study with respect to the criteria introduced above. It only allows the testing of single observed variables which requires an earlier aggregation of the indicators e.g. via factor analysis. The generated factors can then be tested for relationships without being able to model measurement errors. Furthermore, causal linkages between exoge-

5.1 Survey design

147

nous variables cannot be tested as this procedure is based on the assumption of independent regressors. Consequently, a complex set of hypotheses may not be tested simultaneously. Thus, the classical regression analysis does not meet the criteria discussed above and will therefore not be utilized in this study. The present standard for the analysis of different hypotheses between latent constructs is covariance structure analysis or structural equation modeling. Its purpose is to estimate the relationships among a set of observed variables in terms of a generally smaller number of unobserved variables. It allows the explicit consideration of measurement errors, the testing of relationships between exogenous variables and the simultaneous testing of the entire set of hypotheses in a model which will be detailed further in chapters 5.2 through 5.2.5. Therefore, structural equation modeling meets the criteria detailed above and will be utilized in this study even though it also includes some difficulties. As BYRNE (2001, pp. 70-72) points out, the sample size must be very large, the number of the observed variables is limited (BENTLER/CHOU 1987), and the scale of the observed variables must be continuous. This poses a significant challenge for the design of the model and the resulting questionnaire as well as the data collection. 5.1.2 Method of data collection One research aim of this study is the generalizability of its results. This requires high external validity that can only be ensured through a large sample size which is also a prerequisite for using covariance structure analysis. To obtain the large number of responses needed, an on-line survey was used to collect the data. Being almost the modern version of traditional hardcopy mail surveys, they carry similar advantages while at the same time remedying many disadvantages associated with traditional survey research (GRIFFIS/GOLDSBY/COOPER 2003, p. 238). Through a large-scale survey, a large sample size can be collected with comparatively little effort in terms of time and money (KINNEAR/TAYLOR 1996, pp. 331-342). The on-line survey, which has empirically been found to lead to higher response rates than mail surveys (GRIFFIS/GOLDSBY/COOPER 2003, p. 254) is not subject to an interviewer bias which is caused when the informant is biased through the interviewer (CAVUSGIL/ELVEY-KIRK 1998, p. 1165) and which is common in personal interviews. Interviews were, be it personal or over the telephone, considered not suitable for this research also due to their resource needs, as the length of the questionnaire exceeded 200 questions. Given this complexity, the on-line survey offers the respondents the opportunity to flexibly choose the time and location for the an-

148

5 Methodology and sample characteristics

swering which should increase the response rate. The results could furthermore be saved on-line, enabling respondents to stop answering and resuming at a later point in time, e.g. after a friendly reminder. However, there are also drawbacks with any large-scale, questionnaire-based survey, the most important being non-response (HART 1987, p.33). Not only does it reduce the response rate, but more serious it must be established by the researcher that non-respondents are no different than respondents via a non-response bias test. After the method of data collection had been established, the targeted participants of the survey had to be specified. From a theoretical point of view it would be desirable to question all employees of firms that are in any way involved in a relationship with a LSP. However, this would mean an extreme complexity in the data collection and also contains the problem of data aggregation in the end as the combination of discrepant responses of multiple informants is an unresolved issue (KUMAR/STERN/ANDERSON 1993, p. 1636). Furthermore, a dyadic examination, involving both the customer firm and the LSP, could have been desirable but was decided against. While the additional view of the LSP would have provided further insights, the information from the customer will be sufficiently detailed and reliable. Furthermore, a dyadic survey would also mean a considerably higher effort for the data collection, increasing costs, and reducing the amount of responses usable for the later analysis. It was therefore decided to question individuals responsible for logistics decisions of strategic business units as so-called key informants. This is widely accepted and usual in empirical research (BAGOZZI/YI/PHILLIPS 1991, p. 423) as studies suggest that the information from different informants in one firm normally does not differ significantly if the informants are selected carefully (JOHN/REVE 1982, p. 522). However, it is not un-criticized (KUMAR/STERN/ANDERSON 1993, pp. 1635-1637). ERNST (2001, p. 87) states that relying on single informants can lead to substantial measurement problems caused by different motivations or perceptions, restricted cognitive capabilities or diverging levels of information among the informants. These problems occur especially when informants do not possess the required information or are asked to answer for the entire firm. A careful selection of the informants therefore poses a major task for the success of the survey. Among the managers responsible for the outsourcing and the relationship with the LSP it can be presumed that sufficient knowledge for the answering of the questions concerning the outsourcing arrangement exists and that they therefore are suitable key informants. They were therefore targeted as the respondents for the on-line survey.

5.1 Survey design

149

5.1.3 Questionnaire design and pretest The development of the questionnaire was based on the conceptualization of the variables theorized to have an effect on the outsourcing performance as well as the conceptualization of outsourcing-, logistics-, and firm performance. Furthermore, situational factors and general information on the respondent and the respective firm were included. The questionnaire, which will be introduced in detail in the following paragraphs, can be found in its full length in the appendix of this study. To measure the constructs hypothesized to be part of the models introduced in chapter 4, they were operationalized (see chapter 6) and devised to the participants of the survey as statements on a seven point Likertscale, which was anchored with 1 = “strongly disagree” and 7 = “strongly agree” for relationship constructs and 1 = “a lot worse” and 7 = “a lot better” for performance constructs. Multi-item Likert scales are a common and recommended means of collecting data on attitudes, beliefs, values, and other latent constructs (PETERSON 1994). Disagreement exists on the question of how many points the scale should have. To better differentiate between the answers, the wider seven point scale with a neutral middle point was chosen over the narrower five point scale and the six point scale without neutral middle point. To maximize objectivity, it was refrained from using open questions as far as possible. The respondents were furthermore asked to answer all questions with respect to their most important LSP in order to get a consistent picture of one specific relationship. This study is part of an international research project between the Kühne-Center for Logistics Management at the WHU, Otto Beisheim School of Management, in Vallendar, Germany and the Department of Marketing & Logistics, Fisher College of Business, at The Ohio State University, United States of America. The questionnaire for this study was therefore sent out to the sample firms in conjunction with a second study on customer loyalty in logistics from this research project, which appeared to the respondents to be one single survey only. This enabled more information to be gathered from the same respondents with fewer contacts, thereby increasing the response rate of the individual surveys and decreasing the overall amount of time the respondents spent answering as a number of items and constructs could be utilized in both studies.1 1

Both studies shared a total of 68 items, with this study on logistics outsourcing relationships contributing further 94 items and the study on customer loyalty in logistics another 47 items. The number of items therefore totalled 209, as opposed to 277 if both studies had been sent out separately.

150

5 Methodology and sample characteristics

The questionnaire started with a brief introduction of the survey, some basic explanations and an indication of available incentive packages for participants. Then, a general question on the motivation of logistics outsourcing in the firm followed before the main part contained the constructs on logistics outsourcing relationships and customer loyalty in logistics. No distinction was made between the constructs of the two studies and they were arranged in the way that seemed most appropriate with regard to their content. Following the main part, some questions on the logistics and the firm performance followed before the situational factors were surveyed. The end of the questionnaire was formed by statistical questions on the respondent and the firm as well as information on the incentives promised in exchange for the participation. Before mailing out the questionnaire, it was thoroughly tested in a series of pre-test interviews both in Germany and the USA. Eight logistics researchers of the Kühne-Center for Logistics Management reviewed the questionnaire before between July and August 2004 four logistics professors and ten experts from different industries participated in personal interviews to identify potential modification needs in the questionnaire. This procedure has long since been demanded in the literature to avoid logical errors, misunderstandings and misinterpretations (BENNETT 1945, p. 178; CHURCHILL 1991; MALHOTRA 1993). However, since no stringent suggestions exist on the number of pre-test interviews to be conducted, it was decided to complete the phase of questionnaire design after its form had been stable over a number of interviews and it was clear that no further modifications were necessary. On the basis of the pre-test interviews, small modifications were made in the questionnaire, mostly increasing the understandability and slightly changing the layout. The only major modification was the inclusion of the construct openness which primarily has not been discussed intensively in literature, but was viewed by several German pre-test participants as central for logistics outsourcing relationships and different to the construct communication. Overall, the questionnaire was perceived as well structured, understandable and acceptable considering its length with 209 items. 5.1.4 Data collection As argued above, the target respondents for this study are logistics executives that are involved in achieving or responsible for overall logistics performance and for relationships with logistics service providers. Many of these logistics executives in Germany are members of the BVL (“Bundesvereinigung Logistik”), the largest German logistics association. Conse-

5.1 Survey design

151

quently, a cooperation was arranged and selected members of the BVL were mailed the survey. To generalize the results it was important to ensure that the address sample is sufficiently heterogeneous and covers all industries and firm sizes of significant economical importance. On the basis of the BVL database containing approximately 6,800 members, only a selection of the industry was possible. The criterion firm size and the question whether or not the firm is outsourcing logistics services to third parties could not be used for segmentation. Consequently, solely on an industry basis 4,570 addresses were selected that were viewed suitable to represent the basic population of German logistics executives in manufacturing and retailing industries. Additionally, 678 logistics managers from the database of the Kühne-Center for Logistics Management were included in the sample. Duplicate names had previously been removed from the list. All potential participants received a notification E-Mail in August 2004 that briefly introduced the outline of the study and offered the possibility to unsubscribe from the mailing list that later would automatically email the on-line questionnaire. After correcting the sample for mail errors and people that turned out not to be working in logistics, 3,402 contacts remained, 2,789 from the BVL sample and 613 from the sample of the Kühne-Center. The links to the questionnaire were e-mailed to the sample firms in October 2004 together with a cover letter that highlighted the relevance and the importance of the topic and ensured strict confidentiality of the participants’ data. Furthermore, incentives were offered for the participation in the study. Additional to a personalized report of the results of the study, the respondents could chose from a free copy of the book “Erfolg durch Logistik – Erkenntnisse aktueller Forschung” by WEBER/DEEPEN (2003), the free participation at the 2nd annual conference “WHU LogistikSymposium“ or the free participation at a workshop on the topic of logistics outsourcing. No dead-line was set for the participation in the survey. However, after three weeks the contacts that so far had not responded received a friendly reminder. This procedure was repeated again after another three weeks in early December 2004.

152

5 Methodology and sample characteristics

5.1.5 Data base, representativeness and potential biases In total, 579 logistics managers participated in the survey until its end in December 2004. Considering the 3,402 contacts that were initially approached, this results in a response rate of 17.0%. Especially considering the length of the questionnaire with 209 items, this response rate can be regarded as very high (GREER/CHUCHINPRAKARN 1999, p. 76). Some of the returned questionnaires displayed one ore more missing values which must be remedied before a covariance structure analysis can be performed. Different procedures have been suggested to deal with missing values (BYRNE 2001, pp. 289-291): Listwise deletion eliminates all cases from the sample that have at least one missing value, thus reducing the sample significantly. When using pairwise deletion, cases with missing values are not entirely eliminated from the sample, but only excluded from the particular analyses that use the respective variables. Finally, imputation allows to manually insert the missing value on different bases. Mean imputation substitutes the missing value with the arithmetic average of the other cases, regression imputation calculates the missing value on the basis of a regression equation of the complete cases and pattern matching imputation, which is sometimes also called hot-deck imputation (TSIKRIKTSIS 2005, p. 59), replaces the missing value with an observed score from another case in the data for which the response pattern across the variable is very similar. To keep the database as complete and significant as possible, a two step approach was undertaken. First, listwise deletion was utilized to eliminate all cases in which at least on entire construct was missing which could not be obtained from the respondent after sending a friendly personalized request via email. Consequently, 30 cases were removed from the sample, leaving a total of 549 cases of which some still exhibited single or multiple missing values. These missing values were estimated using the pattern matching imputation. In total, 185 or 0.21% of all items were missing values and consequently were replaced with values that were taken from the observed scores of very similar other cases. To be able to generalize the empirical results, they must be representative with respect to the basic population. Therefore, it must be examined if systematic differences exist between firms that have participated in the survey and non-respondents. The potential distortion of the data is called non-response-bias. According to ARMSTRONG/OVERTON (1977, p. 397) it can be assumed that non-respondents are similar to those firms that have participated very late in the survey. Consequently, the sample was split into three parts of equal size on the basis of the answering date. After that, the third of the

5.1 Survey design

153

firms that had answered early was compared to the third that had answered late. A comparison by means of t-tests revealed that only three items out of the 115 indicators which are part of the logistics outsourcing performance, the logistics performance, and the firm performance models, exhibit significant differences of the means on a 10% level. While no recognized threshold value exists in literature, it can be assumed that a rate of less than 3% at a significance level of 10% is very satisfactory (WALLENBURG 2004, p. 134). Therefore, it can be concluded that no non-response bias exists. To also rule out a potential informant bias, the informant competency was assessed on the basis of the respondents’ personal information (KUMAR/STERN/ANDERSON 1993, pp. 1645-1646). 75% of the respondents are either general manager of the firm or logistics manager. Additionally, the average time the informants had already been in their current position was just above 5 years with only 0.4% that had been there for less than one year. Consequently, it can be presumed that the respondents are highly qualified for completing the questionnaire, as they are logistics managers and have sufficient organizational and functional experience. The sample can therefore be assumed not to be informant biased. 5.1.6 Characterization of the participating firms The chosen industries are well represented. As Figure 5-1 indicates, the industries represented strongest are the automotive industry with 16%, retailing with 15% and the electronics, precision mechanics and optics industries with 14%. Other manufacturing industries, such as chemicals and plastics, consumer goods, or manufacturing systems construction are less strong represented. Industries subsumed in the category “other” include pulp and paper, construction, furniture, and several others. Concerning the size of the business units, hereafter called “firms” for reason of complexity reduction, it can be observed that the average revenues are larger than the overall German average. Only 11% of the firms have annual revenues of less than 50 Mio. €, while 27% of the respondents indicated revenues of more than 1 Bill. €. 49% of the firms have revenues between 101 Mio. and 1 Bill. €. Since these firms are very evenly distributed in the sample, the overall results are very satisfactory.

154

5 Methodology and sample characteristics

Food, Beverage and Tobacco

n.a. More than 1 Bn. €

Other 17 Healthcare

6

up to 50 Mio. € 4

Automotive

11 27

16

9

5 9

51-100 Mio. €

Consumer Goods

15

16

Retail

8 10 Chemicals and Plastics

14

101-250 Mio. €

14 Manufacturing Systems Construction

501-1,000 Mio. €

19

Electronics, Precision Mechanics and Optics

251-500 Mio. €

Fig. 5-1. Represented industries and firm sizes in the sample

Since one major goal of this study is the analysis of relationships between customers and logistics service providers, a very interesting point to note are the reasons the firms indicated for the initial outsourcing decision. The reduction of our logistics costs

5.5

To turn fixed costs into variable costs

5.3

To level peaks when order volumes vary

5.0

To reduce our capital employed in logistics processes

4.4

To increase process flexibility and shorten response times

4.0

Our LSP has significantly better logistics skills

3.9

To increase the speed of our logistics

3.8

To increase our capabilities to deliver

3.5

To lower the damage- or error ratio

2.8

Our management capacities are limited

2.5

We consider logistics to be a rather unimportant process

2.3 1

2

“Does not apply”

3

4

5

6

7 “Fully applies”

Fig. 5-2. Motivation for logistics outsourcing

As displayed in Figure 5-2, the top four reasons for logistics outsourcing are all cost related. The reduction of logistics costs is met with the highest approval, followed by the goals to turn fixed costs into variable costs and to level peaks. The objective to increase flexibility through outsourcing ranks only fifth with an average of 4.0 which on the scale between 1 and 7

5.2 Methodological basis for the empirical analysis

155

is precisely the neutral anchor. Points that hint at increased service levels are even less popular, with the increase of delivery capabilities in eighth position and the goal to lower damage- and error ratios in ninth. This clearly indicates that firms currently outsource logistics primarily to reduce costs. Increases in flexibility and service levels are benefits that are noted by the firms. However, their importance is perceived as significantly lower for the outsourcing decision, which obviously is triggered first and foremost through cost reduction considerations. A comparison of these results over time promises to yield further insights. In his study on logistics outsourcing, ENGELBRECHT (2004, pp. 241-243) used the same indicators to asses the reasons for logistics outsourcing. Most notably, the top five reasons are the same in the two studies, with only the increase of process flexibility having moved up to the number 4 reason in the recent survey. This may be interpreted as an increase in cost orientation among the customers as now the top five reasons are cost related. The rest of the indicators have mostly kept their position with respect to the others, which suggests a rather stable motivational situation among outsourcing firms that has not changed dramatically over the past years. A further comparison of the results on the basis of the individual result of the indicators is not possible, as ENGELBRECHT (2004) employed a five point Likert-scale as opposed to the seven point scale used in this study.

5.2 Methodological basis for the empirical analysis In this chapter, the statistical method will be introduced that will be utilized to analyze the gathered data. As it has been argued in chapter 5.1.1, this will be covariance structure analysis, also called structural equation modeling, causal modeling (BAGOZZI 1981a) or causal analysis (HOMBURG 1989). Structural equation modeling is a multivariate statistical method. On the basis of empirically measured covariances of observed, manifest variables, it allows drawing conclusions on the dependence between underlying theoretical, latent variables (HOMBURG 1989, p. 2). Structural equation modeling has its roots in path analysis, a method developed in biometrics to graphically illustrate causal relationship, which was enhanced in the 1960’s by researchers in sociology, the most notable of whom are BLALOCK (1963), DUNCAN (1966), and DUNCAN/HALLER/ PORTES (1968). In the early 1970’s the different approaches were expanded into a general model which can be applied in any research on causal relationships between latent variables through the works of

156

5 Methodology and sample characteristics

KEESLING (1972), JÖRESKOG (1973), WILEY (1973), and JÖRESKOG (1977). Its use in business administration and marketing research can be traced back to BAGOZZI (1980) and BAGOZZI (1981b), where it is now an established standard for the analysis of large-scale surveys. Even though structural equation modeling can simultaneously validate constructs and test hypotheses, a two-step approach as suggested by ANDERSON/GERBING (1988) has become the widely accepted standard. In a first step, reliable and valid measurement models can be identified through confirmatory factor analysis in order to operationalize constructs. The measurement models are then combined in a second step to structural models which allow the testing of hypotheses and causal linkages. 5.2.1 Basics of measurement models Theoretical concepts, or constructs, according to BAGOZZI/PHILLIPS (1982, p. 465) are “abstract, unobservable properties or attributes of a social unit or entity” which cannot be measured directly. However, in order to analyze causal linkages and relationships between constructs, these must be measured. Despite the difficulties for construct validity associated with indirect measurement, this issue has been neglected for a long time and only received increasing attention in the late 1970ies (HOMBURG/GIERING 1996, p. 5). However, the gap was closed through several important works, among them those of JACOBY (1978), BAGOZZI (1979), CHURCHILL (1979), and PETER (1979). Since then, significant advances can be noted in the achievement of valid empirical results (PETER 1981; PETER/ CHURCHILL 1986; GERBING/ANDERSON 1988; BAGOZZI/BAUMGARTNER 1994). For the measurement of a construct, empirically observable indicators are used that reflect the characteristics of a latent variable. Therefore, in a first step the construct must be conceptualized to gain a solid understanding of its true meaning and nature. In a second step, the construct is then operationalized through a number of indicators, also called items, which are suitable to represent the essential aspects and facets of the construct. The indicators are then tested in pre-tests for their relevance and suitability and are modified if necessary. Finally, the tested indicators form the measurement model which on the basis of the empirical data is then tested for validity and reliability in order to later be a part of a structural model. According to HOMBURG/GIERING (1996, p.6), constructs can generally be conceptualized as one-factor or as multi-factor constructs. In the onefactor case, the construct represents exactly one factor on which all measured indicators directly load. Multi-factor constructs contain two or more

5.2 Methodological basis for the empirical analysis

157

factors. They can be either one-dimensional (ANDERSON/GERBING/ HUNTER 1987, p. 435) or multi-dimensional depending on the quantity of different theories underlying the construct. Since in this study only onefactor constructs will be modeled, multi-factor constructs will not be addressed in further detail. The measurement of one-factor constructs can be performed through a single or through several indicators. Using a single indicator only should be restricted to cases of extremely simple variables (JACOBY 1978, p. 93). In the case of more complex constructs, they must be measured employing multiple indicators (JACOBY 1978; CHURCHILL 1979; PETER 1981) in order to guarantee maximum validity of the measurement. Measurement error G1

Indicator x1

O1 Latent, exogenous variable

Measurement error G2

Indicator x2

O2

Fig. 5-3. Generic measurement model of a latent variable with reflective indicators

For construct measurement, formative and reflective indicators can be used (BAGOZZI 1979; BOLLEN/LENNOX 1991; HOMBURG/GIERING 1996). Formative indicators directly affect a factor, which therefore is a function of its indicators as e.g. indices are (WALLENBURG 2004, p. 136). Reflective indicators on the contrary are caused by the factor as displayed in figure 17. Ovals represent the latent, exogenous variable, squares the observable indicators and arrows causal linkages between the latent variable and the indicators. Since reflective indicators are caused by the construct, their measurement is always associated with measurement errors which must be included in the measurement model. In this study, only reflective indicators are used to form the measurement models, since they better capture the variables introduced in chapter 4. Furthermore, the methods to evaluate validity and reliability of formative indicators are still very new and scarcely tested. The measurement of constructs is based on confirmatory factor analysis which constitutes a special case of structural equation modeling. In its course, the measurement model, which can also be formulated as a system of structural equations, is adjusted to the empirical data (JÖRESKOG/SÖRBOM 1982, pp. 404-416; BACKHAUS/ERICHSON/PLINKE/ WEIBER 2003, pp. 334-352). The goal is to estimate the unknown model

158

5 Methodology and sample characteristics

parameters in such a way that the covariance matrix of the measurement model is corresponding as closely as possible to the covariance matrix of the empirical data. The measurement of constructs on the basis of confirmatory factor analysis is only possible if the number of model parameters that need to be estimated is at the most as high as the number of equations underlying the measurement model. Therefore, congeneric measurement models, which are used exclusively in this study and which have no restrictions for factor loading and measurement error variances, require at least three indicators (BAGOZZI 1994, pp. 323-331). Furthermore, consistent with the general procedure all coefficients between the manifest variables and their respective measurement errors will be fixed at 1. The same applies for exactly one random indicator factor loading per measurement model. 5.2.2 Basics of structural models The measurement models as discussed in the previous chapter lead to the formation of structural models. In these structural models, the relationships between the latent variables are examined as displayed in Figure 5-4. Structural model Measurement model of latent, exogenous variable

G1

Measurement model of latent, endogenous variable

] x1

O1 Exogenous variable [

G2

x2

O2

J

O3

y1

H1

O4

y2

H2

Endogenous variable K

Fig. 5-4. Complete causal model consisting of two measurement models and one structural model (BACKHAUS/ERICHSON/PLINKE/WEIBER 2003, p. 350)

The dependent variable is termed endogenous, the independent variable exogenous (BACKHAUS/ERICHSON/PLINKE/WEIBER 2003, p. 336). The causal linkage between the variables is represented by the path coefficient J, which must be standardized for a meaningful interpretation. The sign of the coefficient indicates whether the effect is positive or negative, the absolute value specifies the strength of the relationship. To determine the path coefficients, the measurement and structural models are formulated as linear equation systems and all contained parameters are simultaneously

5.2 Methodological basis for the empirical analysis

159

estimated on the basis of the empirical data. Then, the covariances between the indicator variables are analyzed to draw conclusion on the causal linkages between the latent variables. This reduces the previously many relationships between the indicator variables to a significantly lower number of relationships between latent variables, leading to interpretable models. The parameter estimation is performed in order to see whether the empirically derived covariance matrix fits the matrix resulting from the theoretically hypothesized model. The better the unknown parameters are estimated, the lesser will the differences between the matrices be. For the measurement and minimization of these differences, several discrepancy and estimation functions are available that must be chosen with particular care (BROWNE 1984, pp. 62-83; HOMBURG 1989, pp. 164-185; BACKHAUS/ERICHSON/PLINKE/WEIBER 2003, pp. 362-365). In this study, the maximum-likelihood method (ML) will be utilized which is recommended by HOMBURG/BAUMGARTNER (1995b, pp. 11011102). It represents the international standard in marketing research (BAUMGARTNER/HOMBURG 1996, pp. 149-150) because the ML-method is particularly exact especially in comparison with the alternatives, the unweighted-least-squares (ULS) and the generalized-least-square (GLS) methods. Its application requires a sufficiently large sample size, a multivariate normal distribution of the indicators and the scale of the observed variables to be continuous (BYRNE 2001, p. 70). It must be noted, however, that the maximum-likelihood method is quite robust when the assumption of normally distributed observed variables is violated and still provides valid parameter estimations (BENTLER/CHOU 1987, p. 89; BOOMSMA 1982, pp. 149-173). Especially the assumption of the continuously scaled observed variables has been intensively discussed, as most of the data used in structural equation modeling is typically measured with Likert-scales, which are often implicitly viewed as continuously scaled, but are in fact discontinuous (BYRNE 2001, pp. 70-71). However, BOLLEN/BARB (1981, pp. 232-239) show that Likert-scales with less than 5 categories distort the parameter estimation, while for scales with 5 ore more categories the effect is neglectable. Therefore, the Likert-scales used in this study with 7 categories can be assumed suitable to fulfill the requirement of continuously scaled data. Reliable estimations of the model parameters are generally only possible if the model can be identified. The model can be identified if the covariance matrix of the indicators includes enough information for a clear specification of the model parameters. If there is another model leading to the same estimation of parameters, the model cannot be identified (BAGOZZI/BAUMGARTNER 1994, p. 390). Therefore, the number of parameters to be specified can be at the most as high as the number of equa-

160

5 Methodology and sample characteristics

tions which form the model. This is identical to number of covariances which can be calculated between the observed variables. Therefore, the number of parameters t which have to be specified and the number of indicators q in a model must satisfy the following equation in order to identify the model:

td

q (q  1) 2

(1)

Consequently, the number of parameters t that have to specified cannot exceed the number of empirical covariances q(q+1). The difference between these two variables is considered the degrees of freedom (df). If a model has 0 degrees of freedom, it is clearly determined. However, most of the criteria of the second generation which will be introduced in chapter 5.2.4.2 require a positive number of degrees of freedom, and therefore an over-determined model. 5.2.3 Measurement assessment The explanatory value of any empirical analysis depends strongly on the quality of the underlying measurement. This is especially true in the case of complex latent variables that cannot be directly observed. Therefore, aside from measuring objectively, reliability and validity of the measurement instrument are important. Objectiveness of the data, meaning the independence from the person or group that is conducting the survey, is guaranteed in this study through the on-line survey with a standardized questionnaire. Reliability, which means the measurement is free from random errors (CARMINES/ZELLER 1979, p. 11; CHURCHILL 1979, p. 65; PETER 1979, p. 6; KINNEAR/TAYLOR 1996, p. 232), and validity, which means that it measures what it is intended to measure (CARMINES/ZELLER 1979, pp. 12-13; PETER 1979, pp. 6-7), are two important aspect of the survey that are not as simple to ensure. They will thus be discussed in greater detail in the following chapters. 5.2.3.1 Reliability

Reliability is defined as “the degree to which measures are free of random error” (PETER/CHURCHILL 1986, p. 4). Thus, reliability coefficients estimate the amount of systematic variance in a measure and indicate the extent to which scores on specific measures are repeatable and consistent.

5.2 Methodological basis for the empirical analysis

161

The observed score of a variable consists of the real score, the systematic error, and the random error (PETER 1979; PETER/CHURCHILL 1986). With the increase of the random error component, the measure is becoming more unreliable while the degree of systematic error does not at all reduce reliability as will be discussed in the following chapter. Factors causing the random error are those that are irrelevant for the real score, but nevertheless influence its measurement such as the order of items, respondent fatigue, or the conditions of the measuring situation and is best measured through either the test-retest stability or the internal consistency (HEELER/RAY 1972, p. 361). While test-retest stability cannot be measured due to the nature of this study, the internal consistency is thoroughly evaluated. It measures the correlation between the indicators of a construct through the assessment of split halves (HEELER/RAY 1972, p. 361; HILDEBRANDT 1998, p. 88). It is the higher, the stronger the indicators of a latent variable are correlated. If a significant part of the variance of a construct is explained through the association of the indicators with the latent variable, the influence of measurement errors simultaneously must be smaller (HOMBURG/GIERING 1996, p. 6), leading to a more reliable measurement. The reliability of the measurement is a necessary, yet not a sufficient condition of validity. It is furthermore necessary to test for systematic errors, which is done through the assessment of validity. 5.2.3.2 Validity

Validity is the foundation of the logistics research process (MENTZER/ KAHN 1995, p. 237; GARVER/MENTZER 1999, p. 34). According to MENTZER/FLINT (1997, p. 201), “validity in research is actually a hierarchy of procedures to ensure that what we conclude from a research study can be stated with some confidence (i.e., the conclusion is valid)”. Validity refers to the degree to which a scale measures what it intends to measure. Following CHURCHILL (1979, p. 65), a “measure is valid when the differences in observed scores reflect true differences on the characteristic one is attempting to measure and nothing else”. This is exactly then the case, if a measurement is completely free of systematic errors. A number of different aspects of validity can be distinguished (BAGOZZI/YI/PHILLIPS 1991, pp. 421-422; HOMBURG/GIERING 1996, pp. 7-8). Considering the background of this study and the issue of construct measurement, especially content, convergent, discriminant, and nomological validity are relevant. Content validity refers to the semantic congruence of a construct and its measurement model (HOMBURG/GIERING 1996, p. 7). A measurement

162

5 Methodology and sample characteristics

model has a high content validity if the indicators capture all relevant and theoretically important key facets of the unobservable construct being measured (CHURCHILL 1991, p. 490). Since assessing a scale’s content validity is necessarily qualitative rather than quantitative (PARASURAMAN/ ZEITHAML/BERRY 1988, p. 28), the demand of HEELER/RAY (1972, p. 361) will be followed in this study who suggest to apply “common sense”. In this study, content validity will be ensured during the development of the measurement model, rather than after conducting the survey (NUNNALLY 1978, p. 92; CHURCHILL 1991, pp. 490-491). Convergent validity according BAGOZZI/PHILLIPS (1982, p. 468) is the “degree to which two or more attempts to measure the same concept through maximally dissimilar methods are in agreement”. It exists if the indicators measuring a latent variable exhibit a high correlation (PETER 1981, p. 136) and will be tested in this study through exploratory factor analysis (HOMBURG/GIERING 1996, p. 8). Discriminant validity refers to the degree to which measures of distinct constructs differ (BAGOZZI/PHILLIPS 1982, p. 469), measured in terms of the common variance. Therefore, the correlation of the indicators within individual constructs must be significant and greater than the correlation of the indicators between different constructs (FORNELL/LARCKER 1981, p. 41). Again, this is tested in this study through confirmatory factor analysis. Finally, the nomological validity must be assessed. According to BAGOZZI (1979, p. 14), it “represents the degree to which predictions based on a concept are confirmed within the context of a larger theory”. This refers to the degree that theoretically hypothesized relationships are supported by the analysis of the empirical data, which requires a rigorous theoretical framework for the research models (RUEKERT/CHURCHILL JR. 1984, p. 226; PETER/CHURCHILL 1986, p. 2). Nomological validity in this study is ensured through the solid theoretical framework which was developed in chapter 3 on the basis of which the identification of relationships between the latent variables is possible. 5.2.4 Assessment of measurement and structural models Various different criteria exist to evaluate reliability and validity in structural equation modeling. Generally, criteria of the first and of the second generation can be distinguished (FORNELL 1982). Criteria of the first generation were originally developed in psychometric theory and have been introduced to marketing research through the works of CHURCHILL (1979) and ANDERSON/GERBING/HUNTER (1987).

5.2 Methodological basis for the empirical analysis

163

Reliability and validity criteria of the first generation are solely suited to assess the quality of measurement models. Criteria of the second generation, on the other hand, are based upon confirmatory factor analysis and are suited to assess both measurement and structural models. While they are more powerful than the criteria of the first generation (LONG 1983; GERBING/ANDERSON 1988), their focus is slightly different. Consequently, a two-step process is suggested that uses both sets of criteria complementarily. First, the measurement scales are assessed through criteria of the first generation before in a second step, measurement and structural models are tested with criteria of the second generation. 5.2.4.1 Criteria of the first generation

For the assessment of reliability and validity, exploratory factor analysis, the coefficient alpha and item-to-total correlations as criteria of the first generation are used in this study. The exploratory factor analysis is used to identify the factor structure of a set of indicators (BACKHAUS/ERICHSON/PLINKE/WEIBER 2003, pp. 259232). It serves to analyze whether the measurement model is based on a single factor or if the indicators represent multiple factors. Basis for the analysis are the empirically measured correlations between the indicators without ex-ante formulating hypotheses on the factor structure. Based on the values of the observed variables, an empirical correlation matrix is generated which allows insights into the possible condensation of indicators and builds the foundation for the calculation of the factor loading matrix. Factor loadings represent the association of indicators with a factor and are therefore a measure of their correlation. The relationship between the correlation matrix and the factor loading matrix is described by the fundamental theorem of factor analysis as proposed by THURSTONE (1947) which states that an empirical correlation matrix can be reproduced by the factor loading matrix and its transposed matrix. For a better interpretability of the factor loading matrix, the factor space is rotated around the crossing of the axis. Different rotation procedures are available, the most common of which are varimax and oblimin (BACKHAUS/ERICHSON/PLINKE/WEIBER 2003, pp. 298-332). Varimax assumes orthogonality of the axis and thus independence of the factors, while oblimin is less restrictive and permits any angle between the axes. It is therefore used in this study. The number of extracted factors is determined by the Kaiser criterion (KAISER 1974). According to the criterion, all factors with an eigenvalue higher than one are considered, as they are supposed to have a high ex-

164

5 Methodology and sample characteristics

planatory value. The eigenvalue equates the sum of the squared factor loading over all indicators. Factors with eigenvalues below one explain less variance than one indicator by itself and are eliminated. The scree-test is complementary to the Kaiser criterion and allows the graphical determination of the number of factors to be extracted (BACKHAUS/ERICHSON/ PLINKE/WEIBER 2003, pp. 295-297). The factors are arranged in a coordinate system in the order of decreasing eigenvalues. The graph takes on the shape of a human elbow, whereby the elbow marks the position after which the eigenvalues of the factors are too low to be extracted. Consequently, only factors on the left hand side of the mark are extracted. Based on the exploratory factor analysis, first assumptions can be made on the validity of the measurement models. Convergent and discriminant validity can be assumed when the indicators can be classified as belonging to one factor and the association to other factors is low. This is the case if factor loadings, which mathematically can be interpreted as regression coefficients, for one factor are higher than 0.4 (HOMBURG 1995, p. 93; HOMBURG/GIERING 1996, p. 8) and have a significantly lower factor loading on other factors. Within exploratory factor analysis, the determined variance is also calculated. Validity can be supposed if the extracted factor explains more than 50% of the variance of its indicators (HOMBURG/ GIERING 1996, p. 12) Exploratory factor analysis represents a first step in the assessment of the quality of measurement models which often leads to new insights concerning the structure of the analyzed constructs. While this assesses the validity of the measurement, the reliability is assessed through coefficient alpha and item-to-total correlation. Coefficient alpha, or Cronbach’s alpha (CRONBACH 1951) is considered one of the foundations of measurement theory (PETER 1979, p. 16) and probably is the most widely used criterion of the first generation to assess reliability (CARMINES/ZELLER 1979, p. 44; PETERSON 1994, p. 382; FINN/KAYANDE 1997, pp. 262-263). The widespread use of coefficient alpha can be attributed to its easy calculation and interpretation. It measures the internal consistency of the indicators of a factor and represents the mean of all correlations derived when the items of a factor are divided into two groups in all possible ways and then the sum of the two parts are correlated (CARMINES/ZELLER ,1979, p. 45; MALHOTRA 1993, p. 308; HOMBURG/GIERING 1996, p. 8). The formula of coefficient alpha as proposed by CRONBACH (1951, p. 299) is given below:

5.2 Methodological basis for the empirical analysis

D

§ ¦ Vi n ¨ i ¨1  n 1¨ Vt ©

· ¸ ¸ ¸ ¹

165

(2)

Here, n represents the number of indicators, Vt is the variance of the test scores, and Vi is the variance of item scores after weighting. Alpha values range between zero and one where higher positive values indicate higher levels of reliability. Following the suggestion by NUNNALLY (1978, p. 245), reliabilities of 0.7 will suffice in early stages of research while in many applied settings, even 0.8 might not be enough. Since several scales in this study were either adapted or designed completely new, an early stage of research can be assumed. For the coefficient alpha it must be noted that its value increases with the number of indicators (CHURCHILL/PETER 1984, p. 365), hence scales with more indicators will be more reliable than scales with less indicators.. While in a case of two indicators with a correlation of 0.4 the coefficient alpha takes on a value of 0.57, in a case of 10 indicators with the same correlations, it will be 0.87 (CORTINA 1993, pp. 98-104). Therefore, measurement models with fewer indicators have systematically lower values than those with more indicators. This poses the threat that the utilization of more indicators conceals bad internal consistency. Consequently, the coefficient alpha must always be judged also in terms of the number of indicators used. High reliability can thus be assumed if alpha takes on a value of at least 0.7 with few indicators. Aside from the number of indicators, also their average correlation has an influence on the coefficient alpha. Consequently, it can be improved when indicators with low correlations are deleted from the measurement model (CHURCHILL 1979, p. 68). This is done on the basis of the corrected item-to-total correlation which indicates the correlation between one indicator and all other indicators that represents one factor and consequently is measured individually for all indicators. Since no explicit minimum value is suggested, it serves as a criterion for the elimination of single indicators from the measurement model. The improvement of the coefficient alpha is the higher, the lower the corrected item-to-total correlation of the indicator. It must be observed, however, that the terminology in literature sometimes is quite imprecise and refers to “item-to-total correlation” when actually the “corrected item-to-total correlation” is meant (WALLENBURG 2004, p. 146), even though the former measures the correlation of one indicator with all indicators that represent one factor. In this study, only the

166

5 Methodology and sample characteristics

corrected item-to-total correlation will be used. However, to conform to the common terminology, it will be referred to as item-to-total correlation only. To assess reliability and validity on the basis of the discussed criteria of the first generation poses two distinct disadvantages. On the one hand it is assumed for the calculation of the coefficient alpha that all indicators have the same reliability which means that the measurement errors associated with the individual indicators cannot be analyzed separately. On the other hand, the introduced procedures do not allow inference statistical reliability and validity assessments. Therefore, the criteria are more or less based upon rules of thumb. These disadvantages are overcome by the criteria of the second generation based upon confirmatory factor analysis which will be described in depth in the following chapter. 5.2.4.2 Criteria of the second generation

Confirmatory factor analysis is the basis for the criteria of the second generation to assess the reliability of measurement and structural models. Contrary to the exploratory factor analysis, hypotheses on the relationships between indicators and factors are specified ex-ante. While the exploratory factor analysis serves to empirically identify the presupposed factor structure, the confirmatory factor analysis evaluates it and indicates model fit. For the evaluation of overall measurement and structural model fit, several adaptation measures, fit criteria, and inference statistical measures have been developed which according to HOMBURG/BAUMGARTNER (1995a, p. 165) can be divided into global and local adaptation measures as presented in Figure 5-5. For empirical studies, HOMBURG/BAUMGARTNER (1995a, pp. 171-172) suggest to focus on selected criteria with exceptional explanatory value which will be introduced in the following chapters.

5.2 Methodological basis for the empirical analysis

167

Adaptation measures

Global measures

Local measures

Comparative global measures

Relative global measures CAIC ECVI

Measures for the structural model

Squared multiple correlations Composite reliability Average variance extracted Stand-alone measures

Incremental measures

Measures considering degrees of freedom CFI TLI

Measures for the measurement model

Measures ignoring degrees of freedom

Inference statistical measures

Descriptive measures

F2

NFI

RMSEA Measures considering degrees of freedom

F2/df

Measures ignoring degrees of freedom GFI

AGFI

Fig. 5-5. Adaptation measures

Global adaptation measures Global measures base on the comparison between the empirical covariance matrix and the covariance matrix reproduced by the model and measure the fit between the two. They can by divided into relative global measures and comparative global measures. Relative global measures are not suited to evaluate the fit of a single model but rather allow the comparison of alternative models. They are best used when different models could be selected according to both theory and hypotheses and the decision must be made on the basis of adaptation measures. Both CAIC, the consistent Akaike’s information criterion (AKAIKE 1974; AKAIKE 1987; BOZDOGAN 1987), and the ECVI, the expected cross validation index (BROWNE/CUDECK 1989; BROWNE/CUDECK 1993) will be used in this study for the assessment of structural models. Among different models, the one will be preferred with the lowest values of CAIC and ECVI. For the evaluation of a single model, however, these measures cannot be used, since their value also depends on the complexity of the model. Comparative global measures can be divided into stand-alone and incremental measures. While the former evaluate the fit of the model isolat-

168

5 Methodology and sample characteristics

edly, the latter compare it to a basic model. Stand-alone measures consist of descriptive and inference statistical measures. Inference statistical measures evaluate the model fit on the basis of statistical tests. A first measure is the F2-test with the hypothesis h0 that the empirical covariance matrix S corresponds fully to the covariance matrix of ˆ . The F2-value the model 6

F2

n  1 ˜ F S , 6ˆ

(3)

is calculated with

df

1 ˜ p  q p  q  1  r 2

(4)

degrees of freedom, while n denotes the sample size, p and q the number of endogenous and exogenous indicator variables and r the number of parameters that have to be estimated. Under the condition that h0 is rejected, the p-value denotes the probability of obtaining a F2-value above the observed value. If the p-value is at least 0.05, h0 and therewith the model cannot be rejected at the 5%-level (HOMBURG/GIERING 1996, p. 10). The F2-test has also been criticized. On the one hand, the sample size has a strong influence on the F2-value. As BAGOZZI/YI (1988, p. 77) point out, the probability to reject the model rises with the increase of the sample size. On the other hand, the test measures the absolute correctness of the model while the aim of a model as such is not to be a perfect mirror image of reality, but an abstract approximation. Consequently, the F2-value even increases significantly in complex models if only parts of the empirical covariance matrix S do not correspond to the covariance matrix of the ˆ. model 6 The shortcomings of the F2-test are overcome in the second inference statistical measure RMSEA. The root-mean-squared error of approximation was developed by STEIGER/LIND (1980) and is one of the most common adaptation measures. In contrast to the F2-test, it does not measure the absolute correctness, but rather the approximation of the hypothesized model to the observed data (HOMBURG/BAUMGARTNER 1995a, p. 166). It is calculated by the equation:

RMSEA

F 2  df df (n  1)

(5)

5.2 Methodological basis for the empirical analysis

169

A model is considered as well fit if the RMSEA is below 0.05, while a value up to 0.08 is satisfactory (BROWNE/CUDECK 1993, p. 144). Values up to 0.1 are possibly satisfactory while any value above that is a clear indication for a poor model fit (MACCALLUM/BROWNE/SUGAWARA 1996). However, more recent studies indicate that the RMSEA underestimates the model fit at smaller sample sizes and generally values up to 0.6 indicate a relatively good fit of the model (HU/BENTLER 1999, p. 1). Descriptive measures evaluate the model fit on the basis of minimum values. A descriptive measure ignoring the degrees of freedom is the GFI, the goodness of fit index. It measures the discrepancy between the empiriˆ . It can cal covariance matrix S and the matrix generated by the model 6 range between zero and one, where in the case of ideal adaptation, GFI has a value of 1 (BYRNE 2001, p. 82). According to HOMBURG/BAUMGARTNER (1995a, p. 167), a value above 0.9 can be considered satisfactory. For the calculation, the following formula is used in a ML-setting (JÖRESKOG/SÖRBOM 1982, p. 408):

GFI



tr 6ˆ 1 S  I 1 2 tr 6ˆ 1 S







2

(6)

ˆ represent the covariance matrices as introduced above, I is the S and 6 identity matrix and tr denotes the sum to the diagonal elements of the matrix. The quality of a model is also determined by the number of parameters of the model. With a higher number of parameters, the variances and covariances of a database are easier to explain. In contrast to the GFI, the AGFI (adjusted goodness of fit index) is a descriptive measure that accounts for the number of indicators and degrees of freedom of the model. At the same absolute fit, parsimonious models are evaluated as displaying a higher fit than models with more parameters. Models with higher number of indicators thus are punished by adjusting the GFI according to the number of indicators and degrees of freedom. According to JÖRESKOG/SÖRBOM (1982, p. 408), the AGFI with slightly adapted denomination is calculated as follows, with p being the number of parameters and df the degrees of freedom: AGFI

1

p p  1 1  GFI 2df

(7)

170

5 Methodology and sample characteristics

Values of the AGFI range from zero to one. Some researchers have typically used a value of 0.8 as cut-off value (SHARMA 1996, p. 159), while others demand a more rigorous level of at least 0.9 (HOMBURG/BAUMGARTNER 1995a, p. 168). While the model of logistics outsourcing performance in this study is complex and contains a high number of indicators as several measurement models were newly developed, still the more rigid demand of 0.9 as a minimum value will be favoured in order to ensure the adequate scientific rigor. As the explanatory value of the F2-test is quite limited as argued above, the F2-value will be considered in relation to the number of degrees of freedom df (F2/df). This enables to account for the influence of the degrees of freedom and consequently of the model complexity which leads COTE (2001, p. 87) to demand the use of this measure especially in complex models. HOMBURG (1995, p. 84) proposes a quotient smaller than 3 to accept the model. While only rarely a higher adjusted F2-value is considered acceptable in literature, it is sometimes proposed that only a level under 2.5 indicates a satisfactory model fit (HOMBURG/BAUMGARTNER 1995a, p. 172). The descriptive and inference statistical measures described above belong to the stand-alone measures which asses the model on an isolated basis. Incremental measures which are based on the work of BENTLER/ BONETT (1980) in contrast assess the fit of the model in comparison to a basic model. Here, all indicators are generally assumed to be independent variables, leaving the basic model without explanatory value. A classical incremental measure is the NFI (normed fit index). Since it does not consider the degrees of freedom and is influenced by the sample size (BEARDEN/SHARMA/TEEL 1982; BENTLER/MOOIJAART 1989), it is only rarely used anymore. Therefore, the CFI (comparative fit index) will be used in this study (HU/BENTLER 1995, p. 85). It does consider the degrees of freedom and indicates how the quality of the model is improved by changing from the basic model to the relevant model. It is calculated with the formula:

CFI

max^F r2  df r ;0` max^F b2  df b ; F r2  df r ;0`

(8)

F r2 represents the F2-value of the relevant model and F b2 the F2-value of the basic model. dfr and dfb indicate the degrees of freedom of both models. The CFI has a value between 0 and 1 and indicates a good fit at values above 0.9 (HOMBURG/BAUMGARTNER 1995a, p. 168). While the CFI does not explicitly include the sample size, it does base on the F2-values and

5.2 Methodological basis for the empirical analysis

171

consequently exhibits an undesired rise with an increasing sample size (BOLLEN 1990). Furthermore, the TLI (Tucker-Lewis index) will be used in this study. It is an incremental measure taking into account the degrees of freedom of the model and is recommended by GARVER/MENTZER (1999, p. 41) as a standard criterion for logistics research on the grounds of the general work of MARSH/BALLA/MCDONALD (1988) on ideal fit indices. It is calculated by:

F b2 TLI

df b

F

 2 b

df b

F r2

(9)

df r 1

The denominations used in the formula above are equal to those of the CFI. The TLI, whose values yield a range between 0 and 1, has a satisfactory value when exceeding 0.9 (HOMBURG/BAUMGARTNER 1995a, p. 168; GARVER/MENTZER 1999, p. 41). While it was not explicitly designed to assess the parsimony of models, it does reward simpler models with a higher value and can therefore be used to identify parsimonious models with low complexity and a high model fit. Local adaptation measures Contrary to the global adaptation measures, local adaptation measures are not designed to assess the goodness-of-fit of both measurement and structural models but rather focus on one model type only. Concerning the measurement models the question must be answered how a factor j is measured through the sum of its indicators i. To assess this, the composite reliability (CR) and the average variance extracted (AVE) will be used. Both also allow an evaluation of the convergent validity of the indicators. The composite reliability [ j measures the internal consistency of a group of indicators (BAGOZZI/YI 1988, p. 80). It is measured by the formula:

172

5 Methodology and sample characteristics

2

CR [ j

§ · ¨ ¦ Oij ¸ I jj © i ¹

(10)

2

§ · ¨ ¦ Oij ¸ I jj  ¦ T ii i © i ¹

Here, Oij represents the estimated factor loading, I jj the estimated variance of the latent variable and T ii the estimated variance of the associated measurement error. The value of composite reliability ranges from one to zero, whereat value higher than 0.6 indicate a good model fit (BAGOZZI/YI 1988, p. 82; HOMBURG/BAUMGARTNER 1995a, p. 170). The average variance extracted can be calculated with the following formula from FORNELL/LARCKER (1981, p. 46), again with slightly adapted notation:

AVE [ j

¦O I 2 ij

jj

(11)

i

¦ Oij2 I jj  ¦T ii i

i

The notation of the formula is the same as for the composite reliability. Like many of the indices introduced above, it takes on a value between 0 and 1, while only values above 0.5 can be considered satisfactory (HOMBURG/BAUMGARTNER 1995a, p. 170). Values below 0.5 would indicate that the measurement errors would explain more of the variance of the indicators than the factor itself (WALLENBURG 2004, p. 151). If composite reliability and/or the average variance extracted lie below the proposed threshold value, they can be improved by eliminating indicators with low indicator reliability from the measurement model as long as content validity is guaranteed (see chapter 5.2.3.2). Indicator reliability assesses the reliability of the individual indicators used for the measurement of the construct and indicates the share of the indicator explained through the factor (BAGOZZI/YI 1988, p. 80); the remaining variance is caused by the measurement error. The indicator reliability (IR), sometimes also called individual item reliability, is calculated with the formula whose denomination again is the same as in composite reliability and average variance extracted:

5.2 Methodological basis for the empirical analysis

IR xi

Oij2 I jj Oij2 I jj  T ii

173

(12)

The indicator reliability can take on values between 0 and 1. Values above 0.4 are typically considered satisfactory (HOMBURG/BAUMGARTNER 1995a, p. 170). Apart from the indicator reliability, the t-value of the factor loadings can be used to test the convergent validity of the indicators (BAGOZZI/YI/ PHILLIPS 1991, p. 431). The t-value, also referred to as critical ratio, is calculated as a quotient of the estimated factor loading Oi and the standard error of estimation SE:

t xi

Oi

(13)

SEi

Using a two-sided significance test, t-values of at least |1.96| indicate significant parameter estimates on the 5% level. To test the discriminant validity between factors, the F2-difference-test and the Fornell/Larcker criterion are used. Using the F2-difference-test, discriminant validity can be assessed between two constructs by constraining the estimated correlation parameter I between them to 1 and then performing a F2-difference-test on the values obtained for the constraint and for the unconstrained model (GRANZIN/PAINTER 2001, pp. 82-84). A significantly lower F2-value for the model in which the trait correlations are not constrained to unity would indicate that the traits are not perfectly correlated and that therefore discriminant validity can be presumed. However, the F2-difference-test is not without criticism. Therefore, the Fornell/Larcker criterion will rather be used in this study. FORNELL/ LARCKER (1981) propose that a sufficiently high discriminant validity exists if the average variance extracted of a factor is larger than the squared correlations between the factor and all other factors of the structural model. This means that a factor must explain a higher proportion of the total variance of its own indicators than of any other factor. The assessment of structural models can be performed through squared multiple correlations (R2). They specify the proportion of the variance of an endogenous latent variable that is explained by the variances of all influencing exogenous latent variables (BYRNE 2001, p. 163). The squared multiple correlations take on values between 0 and 1, whereat higher values indicate a higher explanatory value (HOMBURG/BAUMGARTNER

174

5 Methodology and sample characteristics

1995a, pp. 170-171). However, for this criterion, no threshold value will be proposed as its level largely depends on the condition whether the model is aiming at completely explaining a latent variable or if only selected influencing factors are included in the model (HOMBURG/BAUMGARTNER 1995a, p. 170). 5.2.4.3 Overview of criteria for measurement assessment

Figure 5-6 summarizes the different adaptation measures which will be utilized in this study along with the corresponding threshold values. Adaptation Measure

Adaptation of the measurement model

Adaptation of the measurement model and the structural model

Threshold Value

Coefficient Alpha

• 0.6

Explained Factor Variance

• 0.5

Factor loadings

• 0.4

Item-to-total correlation

lowest

t-value of factor loadings

• ~1.96~

Indicator reliability

• 0.4

Composite reliability

• 0.6

Average variance extracted

• 0.5

CAIC

lowest

ECVI

lowest

F2/df

”3

TLI

• 0.9

GFI

• 0.9

AGFI

• 0.8

CFI

• 0.9

Squared multiple correlation

Criteria of the second generation

” 0.08

RMSEA Adaptation of the structural model

Criteria of the first generation

(R2)

Fig. 5-6. Threshold values for adaptation measures

It must be pointed out that single adaptation measures falling below the threshold value must not automatically lead to the rejection of a model (HOMBURG/BAUMGARTNER 1995a, p. 172). For the evaluation of the models, an overall picture of the model, its specific context, and the multiple adaptation measures should also be taken into account. In this sense, a number of authors (FORNELL/LARCKER 1981; BAGOZZI/YI 1988; HOMBURG/BAUMGARTNER 1995a) point out that threshold value must only be understood as recommendations and that the overall judgment of a model has to based on the analysis of all available information, including

5.2 Methodological basis for the empirical analysis

175

both quantitative and qualitative reasoning. Ultimately, researchers and the scientific community must decide whether or not the adaptation measures indicate a model quality which allows drawing substantial conclusions. While this justifies the acceptance of models in which some of the adaptation measures fall below the threshold values and the majority exceeds them, it must not abused to uncritically accept poor models with low fit. This would be detrimental to scientific progress. Instead models, which are accepted despite some measures not meeting the threshold, must be analyzed with particular scrutiny in order to motivate their acceptance. 5.2.5 Basics for model design and modification 5.2.5.1 Measurement models

The development of a measurement model aims at the reliable and valid measurement of a construct. Based on its conceptualization, a group of indicators is selected which reflects the nature of the construct. After these indicators have been tested in pre-tests to determine their relevance and suitability, they are modified if necessary and are finally included in the survey. The operationalization of the measurement models in this study follows the general procedure proposed by HOMBURG/GIERING (1996, pp. 11-14). At first, the group of indicators is tested with the criteria of the first generation before in a second step the more rigorous criteria of the second generation are employed. Not only does this allow assessing the overall quality of the measurement model, but it also provides insights for improvement. Consequently, the criteria are applied in the form of a spiral, wherein single indicators are eliminated from the measurement model if the fit criteria are not met, thus triggering a re-investigation of the model that is terminated only if the model is satisfactory. Following HOMBURG/GIERING (1996, p. 12), in a first step the coefficient alpha is determined for the original group of indicators. If the threshold value of 0.6 is not reached, the item with the lowest item-to-total correlation is eliminated before proceeding with the calculation of the coefficient alpha for the remaining indicators. Once the coefficient alpha meets or exceeds 0.6, an exploratory factor analysis is conducted. Only if solely one factor is extracted, convergent validity can be assumed. Furthermore, the factor must explain at least 50% of the variance of its indicators. If one of these criteria is violated, indictors are again eliminated, this time on the basis of their factor loadings. In a third step, a confirmatory factor analysis is performed that assumes the remaining indicators to be loading on one factor only. Here, both local measures like the composite

176

5 Methodology and sample characteristics

reliability and the average variance extracted are used as well as global measures. For the latter it must be observed, however, that RMSEA, GFI, AGFI, CFI, and TLI require at least one degree of freedom to be estimated. If several of the criteria are clearly not met, further indicators must be eliminated on the basis of their indicator reliability. Resulting from this last step, a final measurement model tested for both reliability and convergent validity is obtained. While the procedure proposed by HOMBURG/GIERING (1996) is especially appealing because of its rigidity, it does not explicitly consider the importance of the content of the indicators during the operationalization of the constructs. As WALLENBURG (2004, p. 155) points out, content validity may be compromised if indicators are eliminated for purely statistical reasons only. He therefore suggest to handle the procedure proposed by HOMBURG/GIERING (1996) more flexibly under certain circumstances. Therefore, in this study it will be paid attention to the fact that after the elimination of indicators from a construct the content validity will still be ensured. WALLENBURG (2004, p. 155) furthermore suggests to parallelize the first two steps of the procedure. While the sequential procedure has the advantage of low complexity in the operationalization process, it does carry the danger that the elimination of indicators on the basis of convergent validity as measured through the item-to-total correlation, leads to the elimination of very reliable indicators – and this even though the valid measurement of a construct depends strongly on the reliable measurement of the indicators. If, on the other hand, both steps are carried out in parallel, the elimination of items cannot only be based on the item-to-total correlation, but also on the results of the exploratory factor analysis and of the indicator reliability. Consequently, the elimination of indicators is always based on both reliability and validity considerations. This procedure will also be applied in this study. 5.2.5.2 Structural models

Covariance structure analysis aims at testing the dependence between different latent variables in a structural model. However, this is only possible if the underlying model shows a sufficient fit. If this is not the case, the model must be modified to allow for valid conclusions. The modification of the model must target the reason for the insufficient fit. Aside from poor data quality, this can either be found in a model that does not reflect reality or in an overly high model complexity (WALLENBURG 2004, p. 155).

5.2 Methodological basis for the empirical analysis

177

If the data contains relationships between variables that have not been included in the model, this will directly lead to a lower model fit. Evidence for this can be found in the modification indices which are stated by the statistics program AMOS. For each fixed parameter specified, they represent the drop in overall F2-value if the parameter were to be specified freely in a subsequent run (BYRNE 2001, p. 90). The model fit can be increased by including formerly not hypothesized relationships into the model. However, this is not always an adequate procedure. The modifications must only be carried out if the observed relationships are plausible and can be explained by theory (BACKHAUS/ERICHSON/PLINKE/WEIBER 2003, p. 380). Otherwise, the model might be over-specified and tailored to the data, leading to a situation in which the results cannot be replicated with a different data set. The results could therefore not be generalized. In this case, the former confirmatory character of covariance structure analysis is altered to a rather exploratory method of analysis (BYRNE 2001, p. 91) which is not desirable. High complexity of a model due to the attempt to maximize explanatory value is another factor that potentially leads to low model fit. The complexity of a model increases disproportionately with the number of variables, since aside from the hypothesized relationships also the interdependencies between the independent variables must be modeled in the structural model. KAUFMANN (2001, pp. 177-178) indicates that beyond 20-30 indicators and 7 variables, sufficient model fit can only rarely be achieved. If a model does not show sufficient fit, it can be improved by eliminating individual variables and the hypothesized causal linkages with other variables from the model. However, as WALLENBURG (2004, p. 156) points out, this may only be done if the variable has no significant direct influence on the explained variables and therefore has no explanatory value for the structural model. In the case where all explaining variables have a significant influence, model complexity can be reduced by splitting the model in two or more separate parts as long as the variables analyzed in the different partial models are not dependent on each other. If the options to improve the model fit introduced above do not suffice, the model must be rejected. Then, an empirical insight into the hypotheses is not possible. Consequently, sufficient fit does only indicate that the model allows statistically valid conclusions on whether the hypotheses must be rejected or not. 5.2.5.3 Moderating analysis

Moderating analyses are conducted using the multi-group analysis function included in the AMOS software package to examine invariance between

178

5 Methodology and sample characteristics

two samples. Since the contingency variables introduced in chapter 4.5.2 were measured in one data set, the moderating analyses require the sample to be split before proceeding to the actual analyses. If the moderator is a latent variable, composed of several indicators rather than a single item only, a single score for the moderator must first be obtained. This is done trough calculating an average score for each case, using the rotated factor loadings of each indicator to weight the individual scores. Subsequently for both single item and latent variable measurements, median splitting is conducted to divide the data set into two parts (DURVASULA/ANDREWS/ LYSONSKI/NETEMEYER 1993; HOMBURG/ GIERING/MENON 2003). The resulting two data sets for the high and the low values of the moderator are then used for the multi-group analyses AMOS. The underlying idea of multi-group analyses is the examination of sample invariance (STEENKAMP/BAUMGARTNER 1998, pp. 80-81). For testing the moderating effects of the variables introduced in chapter 4.5.2, only the structural weights directed at the various performance dimensions of logistics outsourcing, logistics, and the firm are examined. As no sound reasoning is brought forward for other sources of differences such as variances, covariances, or measurement errors, theses are not considered in the following analyses and will always be estimated freely, i.e. separately for the two samples. The procedure used in this study is based on BYRNE (2001, pp. 176197). It tests the hypothesis that both samples’ covariance matrices are different which reads as a formula:

H 0 : ¦ T1 z ¦ T 2

(14)

where 6 T 1 is the model covariance matrix in sub-sample 1 and 6 T 2 is the model covariance matrix for sub-sample 2. This hypothesis is reflected in the so called unconstrained model in which all model parameters are estimated freely. Thus, no equality constraints are imposed between the two samples which therefore are calculated separately. Then, a second estimation is performed, imposing equality constraints on the structural paths leading to all performance dimensions in the respective model. If differences exist in the model in general and in the constrained paths in particular, model adaptation is significantly worse for the constrained model than for the unconstrained model. In multi-group structural equation modeling, adaptation is measured through theF2-statistic (GRANZIN/PAINTER 2001, pp. 82-84) and the F2-difference ('F2) between the two models. Furthermore, the difference in the degrees of freedom ('df) is measured. 'F2 and 'df are calculated via the formulae:

5.2 Methodological basis for the empirical analysis

'F 2

F c2  F u2 , 'df

df c  df u

179

(15)

where F c2 and F u2 are the constrained and unconstrained models’ F2 and

df c and df u are the constrained and unconstrained models’ degrees of freedom. The significance of 'F2 with 'df is assessed through employing the F2-distribution. In the case that the constrained model is significantly worse than the unconstrained model, differences with regard to the constrained structural paths between the two sub-samples exist. In this case, H0 is supported and further analyses must be conducted to precisely locate the differences in the various paths potentially moderated. To do so, the constrained paths are individually examined in an iterative process. For that, the individual paths are restricted to equality one-by-one, and the solution is compared to that of the unconstrained model. If restricting a certain path does not lead to a significant increase in F2, thereby indicating that the path is not different between the two sub-samples, the equality constrained is maintained for the following analyses. If, on the other hand, the F2 is significantly increased by imposing the constrained, a moderating effect is detected and the therefore the path is freely estimated in all subsequent analyses. This procedure is conducted for all paths where a moderating effect can be expected. Eventually, a model with all different paths freely estimated and all invariant paths restricted to equality is estimated. While this already reveals differences in the model, only the comparison of the individual standardized parameter values for the paths in the two sub-samples provide the necessary detail for the in-depth testing of the various moderating effects.

6 Construct operationalization

In chapter 4, logistics outsourcing performance and its antecedents as well as logistics performance and firm performance have been conceptualized in great detail. After chapter 5 has introduced the methodological basis for the research, all constructs relevant for the models will be operationalized in the following.

6.1 Antecedents of logistics outsourcing performance 6.1.1 Cooperation Cooperation has been conceptualized in chapter 4.2.1.1 as the cornerstone of successful relationships. Although it has been intensely studied in customer-supplier relationship research, so far no established scales exist for logistics outsourcing relationships. While the scale proposed by MORGAN/HUNT (1994, p. 35) is formative and not adaptable to the proposed model, the works of FRAZIER (1983) and LARSON/KULCHITSKY (1999) offer a solid basis for the development of a logistics outsourcing related cooperation construct. GUILTINAN/REJAB/RODGERS (1980) empirically find that coordination and cooperation tend to be high when interfirm communications are perceived to be effective in reducing uncertainty and participative decision making is perceived to take place. In addition to this, FRAZIER (1983, p. 73) points out that also four other factors contribute to high levels of cooperation. Aside from high levels of ideological agreement those are goal compatibility, role satisfaction and the use of power in a non-pressurized form. Further insight into the nature of cooperation, this time with a distinct logistics focus, is provided by LARSON/KULCHITSKY (1999, p. 94). They point out that cooperative relationships are characterized by collaborative goal setting, cross-functional coordination, detailed communication, mutual respect, mutual trust, teamwork, and unity of purpose. Starting from these inputs, 8 indicators were developed that incorporate the different aspects of cooperative behavior. The original group of indicators is displayed in Table 6-1.

182

6 Construct operationalization

Table 6-1. Indicators for the measurement of the construct cooperation

Indicator 1 Indicator 2 Indicator 3 Indicator 4

Please indicate your level of agreement with the following statements on your relationship with this LSP The goals of our relationship were jointly set by us and our LSP. Our approach to doing business or organizing activities is very similar to our LSP. In the relationship with our LSP, we always pull together in the same direction. When problems or questions arise during this outsourcing relationship, we make decisions together with our LSP to get to adequate solutions.

Indicator 5 Indicator 6 Indicator 7

If one partner exercises his power in the relationship, he does it in an appropriate way. In our business relationship, both parties fully respect each other. Our employees are working together with the LSP to secure the relationship's success even beyond the previously established responsibilities.

Indicator 8

The LSP is cooperating with us very well.

The conceptual input of FRAZIER (1983, p. 73) is included in four indicators. Indicator 2 reflects high levels of ideological agreement, indicator 4 participative decision making, indicator 5 the use of power in a nonpressurized form and indicator 8 role satisfaction, facilitating general cooperation. The findings of LARSON/KULCHITSKY (1999, p. 94) are embodied in four indicators as well. Indicator 1 addresses the collaborative goal setting, indicator 6 the mutual respect, indicator 7 reflects the need for cross-functional coordination and teamwork and indicator 8 the unity of purpose. As indicated above, FRAZIER (1983) and LARSON/KULCHITSKY (1999) all highlight the importance of communication for cooperation. Since communication is viewed as an independent construct in this research, indicators explicitly targeting information exchange have been excluded from the construct of cooperation. Table 6-2. Adaptation measures for the construct cooperation (8 indicators) Information on the factor cooperation (8 indicators) Coefficient alpha Explained variance

0.87 47.94% 11.130 0.86 0.90

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.82 0.90 0.136 0.87 0.45

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6 Indicator 7 Indicator 8

Item-to-total correlation 0.45 0.74 0.71 0.64 0.49 0.72 0.54 0.73

Indicator reliability 0.22 0.62 0.64 0.43 0.27 0.62 0.34 0.69

t-value of factor loadings 10.76 10.85 9.99 8.85 10.78 9.39 10.98

6.1 Antecedents of logistics outsourcing performance

183

While the exploratory factor analysis reveals that all indicators load on one factor, the explained factor variance fails to meet the threshold value of 50%. Furthermore, the average variance extracted indicates the insufficient reliability of the indicators. Consequently, indicators 1, 5, and 7 having the lowest indicator reliability were eliminated from the measurement model. This results in a measurement model with almost satisfactory fit. However, RMSEA (0.137) and F2/df (11.03) are still insufficient. Further analysis showed that the correlation between the error terms of indicator 8 and those of the indicators 2, 4, and 6 were very high. Since furthermore indicator 8 has the same general notion as indicator 3, indicator 8 was also removed from the measurement model. The resulting measurement model for cooperation with the indicators 2, 3, 4 and 6 displayed in Table 6-3 shows very good adaptation measures and requires no further modification. Table 6-3. Adaptation measures for the construct cooperation (4 items) Information on the factor cooperation (4 indicators) Coefficient alpha Explained variance

0.84 58.07% 0.730 1.00 1.00

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.99 1.00 0.000 0.84 0.57

Information on the indicators Indicator 2 Indicator 3 Indicator 4 Indicator 6

Item-to-total correlation 0.72 0.74 0.61 0.64

Indicator reliability 0.67 0.72 0.44 0.50

t-value of factor loadings 19.83 15.61 16.81

6.1.2 Communication The importance of communication for interorganizational relationships has been emphasized in chapter 4.2.1.2. Aside from the quantitative aspect, it has been pointed out that a key success driver is the quality of the communication between the parties involved. For the measurement of the construct, different scales and concepts have been adapted to fully capture its complexity. ENGELBRECHT (2004, p. 217) develops a scale for the measurement of the information exchange between customers and LSP. From this operationalization, indicators 1 and 2 are used, covering both the frequency of the communication between the parties and its quality. Two further indicators suggested by ENGELBRECHT

184

6 Construct operationalization

(2004) could not be used due to an overlap with other constructs used in this research. In the pre-test interviews, it was also suggested to include a second indicator to measure the quantitative aspects of the information exchange in the relationship. Consequently, indicator 3 was developed which measures the communication intensity between the two parties. Table 6-4. Indicators for the measurement of the construct communication

Indicator 1

Please indicate your level of agreement with the following statements on your relationship with this LSP We frequently discuss possible problems or improvements with the responsible parties of our LSP.

Indicator 2

The exchange of information between the employees of the LSP and our company is working very well.

Indicator 3 Indicator 4

To reach our goals, a lot of meetings and talks with this LSP are necessary. When we exchange information with our LSP, it is always relevant for the progress of the project and our cooperation.

Indicator 5

The exchange of information between us and our LSP takes place as soon as it becomes available.

Indicator 6 Indicator 7

Both sides can always fully rely on the information we exchange. The way we exchange information with our LSP is very suited for solving problems according to both parties' interests.

A conceptualization of the quality of communication beyond the rather general indicator 2 is suggested by MORGAN/HUNT (1994, p. 25) who point out that in order to be perceived as high quality communication, it is required to be relevant, timely, and reliable. These three specifications are incorporated in the indicators 4, 5, and 6. Finally, indicator 7 was included as a general measurement of the quality of the exchange of information and its usability to solve arising problems according to both parties’ interests.

6.1 Antecedents of logistics outsourcing performance

185

Table 6-5. Adaptation measures for the construct communication (6 indicators) Information on the factor communication (6 indicators) Coefficient alpha Explained variance

0.86 51.04% 11.800 0.89 0.94

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.86 0.93 0.140 0.86 0.51

Information on the indicators Indicator 1 Indicator 2 Indicator 4 Indicator 5 Indicator 6 Indicator 7

Item-to-total correlation 0.57 0.68 0.47 0.68 0.71 0.75

Indicator reliability 0.35 0.53 0.23 0.53 0.68 0.73

t-value of factor loadings 13.11 9.67 13.12 14.20 14.46

The resulting seven indicators were tested in an exploratory factor analysis. However, indicator 3 which explicitly covers the quantitative aspects of information exchange only, was found to load on a second factor. It was therefore eliminated from the measurement model. As Table 6-5 indicates, the measurement model with six indicators did not satisfy in all dimensions with insufficient values for RMSEA, TLI, and AGFI. Furthermore, the average variance extracted was only barely above the threshold value of 0.5. To improve the fit of the measurement model, indicators 1 and 4 were eliminated since both showed unsatisfactory indicator reliability well below the threshold of 0.4. Furthermore, their error terms displayed very high correlations with the error terms of the remaining indicators. Table 6-6. Adaptation measures for the construct communication (4 indicators) Information on the factor communication (4 indicators) Coefficient alpha Explained variance

0.86 61.36% 0.597 1.00 1.00

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

1.00 1.00 0.000 0.86 0.61

Information on the indicators Indicator 2 Indicator 5 Indicator 6 Indicator 7

Item-to-total correlation 0.64 0.65 0.75 0.78

Indicator reliability 0.48 0.50 0.71 0.77

t-value of factor loadings 14.89 17.31 17.67

Resulting from the elimination of indicators 1, 3, and 4 is the final measurement model with four indicators as presented in Table 6-6. It

186

6 Construct operationalization

shows very good adaptation measures and can therefore be accepted without further modification as a very satisfactory measurement model for the construct of communication. 6.1.3 Proactive improvement As argued in chapter 4.2.1.3, the importance of proactive improvement of the customers’ logistics processes through the LSP has recently gained increasing attention in logistics research (ENGELBRECHT 2004; WALLENBURG 2004; FLINT/LARSSON/GAMMELGAARD/ MENTZER 2005). For the operationalization of the construct, the scale developed by WALLENBURG (2004, pp. 182-183) on the basis of ENGELBRECHT (2004, pp. 207-209) was selected without modifications. As Table 6-7 shows, a total number of five indicators was selected. The indicators reflect the different aspects important for the improvements in efficiency and effectiveness of the customers’ logistics processes through the LSP. Apart from a generally high level of innovation as measured through indicator 5, the continuous improvement and modification of logistics processes by the LSP is relevant (indicators 1 and 3). Furthermore, it is particularly valuable for the customer if the suggestions for improvements are going beyond the direct responsibility of the LSP (indicator 2) and are proactive (indicator 4). Table 6-7. Indicators for the measurement of the construct proactive improvement

Indicator 1 Indicator 2

Please indicate your level of agreement with the following statements on the improvement of logistics systems. The LSP puts strong effort into continously optimizing logistics processes. The LSP continuously makes suggestions for improvements of activities, even those outside its direct responsibility.

Indicator 3

When the situation changes, the LSP by itself modifies the logistics activities and processes, if this is useful and necessary.

Indicator 4 Indicator 5

The LSP shows initiative by approaching us with suggestions for improvement. The LSP shows a high level of innovation.

As Table 6-8 indicates, the measurement model is showing almost satisfactory adaptation measures. While the criteria of the first generation are satisfactory, the values of F2/df and of the RMSEA as criteria of the second generation are on the verge of being unsatisfactory. In an effort to develop a measurement model with completely sufficient adaptation measures, indicator 1 was eliminated. This was justified on the basis of the very high correlation of its error term with that of indicator 4. Since indicator 4 contains the important aspect of the LSP’s own initiative for improvement and the facet of continuous improvement embodied in indicator 1 is also found

6.1 Antecedents of logistics outsourcing performance

187

in indicators 2 and 3, indicator 1 is eliminated despite its rather high indicator reliability. Table 6-8. Adaptation measures for the construct proactive improvement (5 indicators) Information on the factor proactive improvement (5 indicators) Coefficient alpha Explained variance

0.92 69.59% 2.959 0.99 0.99

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.97 1.00 0.060 0.92 0.70

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5

Item-to-total correlation 0.75 0.74 0.80 0.84 0.82

Indicator reliability 0.61 0.60 0.70 0.81 0.76

t-value of factor loadings 19.53 21.39 23.39 22.64

The resulting measurement model with only 4 indicators shows satisfactory adaptation measures in all dimensions. It can therefore be accepted without any further modifications. Table 6-9. Adaptation measures for the construct proactive improvement (4 indicators) Information on the factor proactive improvement (4 indicators) Coefficient alpha Explained variance

0.91 71.65% 0.561 1.00 1.00

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

1.00 1.00 0.000 0.91 0.72

Information on the indicators Indicator 2 Indicator 3 Indicator 4 Indicator 5

Item-to-total correlation 0.74 0.78 0.85 0.81

Indicator reliability 0.60 0.69 0.83 0.75

t-value of factor loadings 20.78 23.15 21.88

6.1.4 Trust As it has been argued in chapter 4.2.1.4, trust has been established as a fundamental relationship model building block. Its importance for interorganizational exchange relationships is widely recognized. Trust is the perception of the credibility and benevolence of a target of trust by an individual or a group (GANESAN 1994, p. 3; DONEY/CANNON

188

6 Construct operationalization

1997, p. 36). These two aspects will be incorporated in the operationalization of the construct. However, a two-dimensional structure of the construct will not be ex-ante assumed. While some studies find that discriminant validity exists between them, the two dimension are highly correlated (GANESAN 1994; KUMAR/SCHEER/STEENKAMP 1995). Furthermore, DONEY/CANNON (1997, p. 43) point out that while credibility and benevolence could be conceptually distinct, they may be so intertwined in business relationships that in practice they are inseparable. Consequently, the trust of the customer will be treated as a unidimensional construct in this study. Concerning the object of trust, different conceptualizations exist. While most often in buyer-supplier research the organization alone is the object, indications exist that both the organization and the person acting on behalf of the organization are two distinct objects of trust (WALLENBURG 2004, pp. 185-186). While this view certainly has potential for additional explanatory value, this study focuses purely on the relationship between the customer and the organization of the LSP. Personal ties between individual employees and their attitudes are not subject of the research. Consequently, the view of MORGAN/HUNT (1994, p. 35) will be adopted who operationalize trust as the trust in the major supplier which in this case is the most important LSP. Table 6-10. Indicators for the measurement of the construct trust

Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Please indicate your level of agreement with the following statements regarding your LSP This LSP keeps promises it makes to our firm. Regarding problems, this LSP is always honest with us. This LSP is trustworthy. This LSP its genuinely concerned that our business succeeds. The LSP correctly carries out tasks that we cannot directly control. When making important decisions, this LSP considers our welfare as well as its own.

Since the conceptualization of trust is rather diffuse in exchange relationship research, a large number of different scales exist for the operationalization. For this study, the scale introduced by DONEY/CANNON (1997, p. 48) was selected. Having been adapted to the logistics context by WERTZ (2000, pp. 146-147) and WALLENBURG (2004, pp. 186-187), it has shown to be very reliable in the past. Table 6-10 presents the indicators used for the operationalization of the construct in this study. Indicators 1, 3, 4 and 6 with slight modifications come from the scale of DONEY/CANNON (1997, p. 48), indicator 2 from WERNER (1997). Indicator 5 was introduced by WALLENBURG (2004, p. 186). Of these six indicators, 1 to 3 and 5 aim at the credibility of the LSP,

6.1 Antecedents of logistics outsourcing performance

189

while indicators 4 and 6 reflect the customers’ perception of the benevolence of the LSP. Table 6-11. Adaptation measures for the construct trust (6 indicators) Information on the factor trust (6 indicators) 0.92 66.04% 15.392 0.91 0.92

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.81 0.95 0.162 0.92 0.65

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Item-to-total correlation 0.71 0.83 0.86 0.74 0.74 0.75

Indicator reliability 0.58 0.82 0.85 0.57 0.56 0.57

t-value of factor loadings 22.92 23.48 18.44 18.27 18.61

Since a unidimensional structure of the construct was assumed, it was examined first. An exploratory factor analysis revealed that all six indicators load on only one factor. This was also confirmed by a confirmatory factor analysis. The resulting measurement model with six indicators however shows some weaknesses. As Table 6-11 illustrates, both the F2/df with a value of 15.392 and the RMSEA are significantly too high. As a result, the measurement model was modified. Item 6 was eliminated because of the high correlation of its error term with those of the indicators 1, 3, 4, and 5. The remaining measurement model still did not show satisfactory adaptation measures with the F2/df - value at 7.84 and the RMSEA at 0.112. Indicator 4 was eliminated to further increase the adaptation measures. Not only did it display the lowest indicator reliability in the 5-indicator measurement model, its error term was also correlated significantly with the error terms of the other indicators. Finally, its modification index indicated substantially better measures if the indicator was to be dropped from the measurement model. The resulting measurement model with four indicators is presented in Table 6-12. It displays very satisfactory criteria in all dimensions and can therefore be accepted without further modification. It must be noted, however, that both indicators aiming at the measurement of the aspect of benevolence have been eliminated. The resulting measurement model therefore measures primarily the aspect of credibility of the LSP.

190

6 Construct operationalization

Table 6-12. Adaptation measures for the construct trust (4 indicators) Information on the factor trust (4 indicators) 0.90 70.47% 1.384 1.00 1.00

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.99 1.00 0.026 0.90 0.70

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 5

Item-to-total correlation 0.74 0.85 0.84 0.69

Indicator reliability 0.60 0.86 0.84 0.52

t-value of factor loadings 23.98 23.72 17.79

6.1.5 Commitment As it has been argued in chapter 4.2.1.5, relational commitment is central to exchange relationships between buyers and suppliers in general and customers and LSPs in particular. For the context of this study, it has been conceptualized in line with the research of MORGAN/HUNT (1994, p. 23) as organizational commitment which has a strong notion of affective commitment. While relationship commitment is a widely used construct in exchange relationship research, no standardized scale exists. The scale used by MORGAN/HUNT (1994, p. 35) is formative and therefore not suited for covariance structure analysis. Instead, a measurement model was used that originally was developed by ZIMMER (2000, pp. 161-177), extended by WERNER (1997, p. 148) and adapted for the logistics context by WALLENBURG (2004, pp. 189-192). Table 6-13. Indicators for the measurement of the construct commitment

Indicator 1

To what degree do you agree to the following statements on your attitude towards this LSP? We come to our LSP's defense when it is criticized by persons from inside or outside our organization.

Indicator 2 Indicator 3

We would be very sorry personally if we had to terminate the relationship with this LSP. We feel personally offended, when this LSP is criticized by persons from inside or outside of our company.

Indicator 4

We strongly intend to keep up the relationship with this LSP for as long as possible.

As Table 6-13 indicates, the proposed measurement model contains four indicators, three of which have been successfully included in a commitment construct in a logistics context by WALLENBURG (2004, pp. 189192). Indicators 1 and 2 come from ZIMMER (2000, p. 218) and reflect the

6.1 Antecedents of logistics outsourcing performance

191

degree of the customer’s commitment to vouch for the LSP (indicator 1) and the personal commitment of the customer’s employees towards the relationship (indicator 2). The emotional ties of the customer’s employees with the relationship are subject of indicator 3 which was developed by CAHILL (2006, p. 178). Finally, indicator 4 is adapted from WERNER (1997, p. 148) and measures explicitly the desire of the customer to defend the relationship with the LSP also under pressure. As Table 6-14 indicates, the adaptation measures of the measurement model with four indicators were not satisfactory. Especially the values of F2/df at 6.286 and of the RMSEA at 0.098 are not meeting the threshold criteria defined in chapter 5.2.4.3. The insufficient quality of the measurement model is further reflected in the indicator reliability of indicator 3, which is low at 0.30. Since indicator 3 is not necessarily needed to ensure content validity as personal- and emotional commitment are also reflected in indicator 2, it was consequently eliminated from the measurement model. Table 6-14. Adaptation measures for the construct commitment (4 indicators) Information on the factor commitment (4 indicators) 0.81 55.34% 6.286 0.96 0.99

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.94 0.99 0.098 0.82 0.53

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4

Item-to-total correlation 0.66 0.71 0.52 0.69

Indicator reliability 0.52 0.72 0.30 0.67

t-value of factor loadings 17.32 11.88 17.05

After the elimination of indicator 3, the measurement model presented in Table 6-15 with only three remaining indicators resulted. While several criteria of the second generation could not be calculated as a measurement model with only three indicators has no degrees of freedom, the remaining criteria in the form of the coefficient alpha, the explained variance, the composite reliability and the average variance extracted are satisfactory. The measurement model can therefore be accepted without any further modification, leaving the model utilized by WALLENBURG (2004, p. 189192).

192

6 Construct operationalization

Table 6-15. Adaptation measures for the construct commitment (3 indicators) Information on the factor commitment (3 indicators) 0.84 63.47% * * *

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.84 0.65

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated.

Information on the indicators Indicator 1 Indicator 2 Indicator 4

Item-to-total correlation 0.64 0.74 0.72

Indicator reliability 0.49 0.75 0.66

t-value of factor loadings 16.31 16.34

6.1.6 Functional conflict As chapter 4.2.1.6 has pointed out, relationships are characterized through the existence of conflicts. While conflicts between exchange partners can have several destructive consequences, disputes and disagreements that are solved amicably can have beneficial outcomes for relationships. Such conflict resolution is called “functional conflict”. The importance of the construct for exchange relationship research has been recognized by various authors. Consequently, a number of different scales for its measurement exist (ANDERSON/NARUS 1990, p. 49; MORGAN/HUNT 1994, p. 35; CLAYCOMB/FRANKWICK 2004, p. 34). However, no single scale can be identified that fits the broad conceptualization of functional conflict in this study. Furthermore, none of them have as yet been adapted for a logistics context. Therefore, a new scale was developed considering several different inputs from various sources in an effort to develop a new measurement scale, resulting in the seven indicators presented in Table 6-16.

6.1 Antecedents of logistics outsourcing performance

193

Table 6-16. Indicators for the measurement of the construct functional conflict

Indicator 1 Indicator 2

Please indicate your level of agreement with the following statements on your relationship with this LSP When problems occur we always solve them jointly. Especially when solving problems, the exchange of information between the LSP and us is working very well.

Indicator 3 Indicator 4

When problems occur, it is always the same party tackling and solving them. When problems occur, the discussion frequently gets out of hand and ends in an exchange of harsh words.

Indicator 5

Differences of opinion between the LSP and us have significantly decreased the productivity of our working relationship.

Indicator 6

Differences of opinion between the LSP and us are viewed as "just part of doing business" and offer potential benefits for both parties involved.

Indicator 7

Differences when cooperating with this LSP are always settled smoothly.

ENGELBRECHT (2004, p. 217) points out the importance of the exchange of information for solving problems. Only if all parties involved know about the problem and communicate effectively, can conflicts be solved functionally. This aspect is measured through indicator 1, which was slightly adapted to accommodate for the joint problem solving from an indicator ENGELBRECHT (2004, p. 217) uses for the measurement of the construct of information exchange. Indicators 1, 3, and 4 were taken from a scale proposed by CLAYCOMB/FRANKWICK (2004, p. 34) for the measurement of conflict resolution mechanisms and were adapted to functional conflict in logistics outsourcing relationships. CLAYCOMB/FRANKWICK (2004, p. 20) argue that conflict resolution can include a different spectrum of constructive and destructive methods. The more constructive the resolution method is, the more functional will the conflict be. It is of particular importance to solve problems jointly (indicators 1 and 3, reverse coded) and without open and personal confrontation, leading to destructive confrontations (indicator 4, reverse coded). A further aspect for the evaluation of functional conflict is introduced by ANDERSON/NARUS (1990, p. 49). According to them, conflict decreases the productivity of working relationships. This can be overcome by solving conflicts functionally, as measured by indicator 5 (reverse coded) which was directly taken from their scale on the functionality of conflict with only minor modifications for the logistics context. Representing a similar line of thought, indicator 6 was included from the functional conflict construct used by MORGAN/HUNT (1994, p. 35). It measures to what extent the parties involved understand and act upon the potential benefits of functional conflict when dealing with differences in opinion. Again, the item was slightly modified to fit the logistics context.

194

6 Construct operationalization

Finally, indicator 7 was included as a general measure to assess the overall quality of cooperation between the parties involved with a special focus on the functional use of conflicts. Only when differences in opinion and occurring conflicts are settled smoothly, can functional conflict be assumed to exist. Table 6-17. Adaptation measures for the construct functional conflict (5 indicators) Information on the factor functional conflict (5 indicators) Coefficient alpha Explained variance

0.81 50.45% 2.049 0.99 0.99

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.98 1.00 0.044 0.82 0.48

Information on the indicators Indicator 1 Indicator 2 Indicator 5 Indicator 6 Indicator 7

Item-to-total correlation 0.73 0.76 0.47 0.43 0.64

Indicator reliability 0.71 0.81 0.25 0.23 0.51

t-value of factor loadings 23.25 11.96 11.34 18.38

An initial exploratory factor analysis extracted two different factors from the seven indicators presented in Table 6-16. Indicators 3 and 4 therefore had to be eliminated from the measurement model, before the new, five-indicator measurement model was tested again. While the exploratory factor analysis only extracted one factor, several adaptation measures did not meet the desired threshold values. As Table 617 indicates, the average variance extracted is not satisfactory. Furthermore, the indicator reliabilities of indicators 5 and 6 are well below the requested value of 0.4. Consequently, they were also eliminated from the measurement model. The resulting measurement model with three indicators remaining again has a decreased set of adaptation measures only. However, the adaptation measures that can be calculated, such as the coefficient alpha, the explained variance and the average variance extracted, exhibit high and satisfactory adaptation measures. The three indicator measurement model can therefore be accepted without any further modification.

6.1 Antecedents of logistics outsourcing performance

195

Table 6-18. Adaptation measures for the construct functional conflict (3 indicators) Information on the factor functional conflict (3 indicators) 0.85 67.71% * * *

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.86 0.66

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated.

Information on the indicators Indicator 1 Indicator 2 Indicator 7

Item-to-total correlation 0.74 0.79 0.66

Indicator reliability 0.69 0.84 0.50

t-value of factor loadings 20.75 17.90

6.1.7 Involvement As argued in chapter 4.2.1.7, the involvement of the LSP in the implementation process is an essential success factor for the success of logistics outsourcing arrangements. As the construct as yet has not received particular attention in research, the only existing operationalization was developed by ENGELBRECHT (2004, pp. 203-205). Since it also originated from a logistics outsourcing relationship background, it was selected and slightly adapted to suit the research focus of this study. Table 6-19. Indicators for the measurement of the construct involvement Please indicate your level of agreement with the following statements towards your internal relationship and the involvement of your LSP in this outsourcing project Indicator 1 Indicator 2

The LSP was significantly involved in the outsourcing relationship at an early stage. A cross-functional steering committee is responsible for the decision on the outsourcing project and its implementation.

Indicator 3

The employees responsible for the outsourcing project on both sides work together very well.

According to ENGELBRECHT (2004, p. 204), the LSP must be involved in the implementation process at an early stage. This avoids communication problems, enables the exchange of know-how and fosters familiarization between the parties involved. The indicator, which was not modified, is represented as indicator 1 in Table 6-19. Indicator 2 was also taken in a slightly modified form from the scale developed by ENGELBRECHT (2004, p. 203). It reflects the importance of an inter-disciplinary team overseeing the implementation process since it will

196

6 Construct operationalization

have consequences for several functions and business processes in the firm. Indicator 3 was newly developed as an extension of indicator 3. While one interdisciplinary team must lead the implementation process, individual tasks can be delegated to subordinate teams or individuals. For the success of an outsourcing arrangement it is crucial that also these employees collaborate effectively and efficiently with each other and with their counterparts on the side of the LSP. This can only be enabled if the LSP is intensively involved in the outsourcing project from an early stage on. The resulting measurement model with three indicators was then tested with criteria of the first and of the second generation. While a number of adaptation measures could not be calculated due the absence of degrees of freedom, the remaining indicators revealed an unsatisfactory measurement model. While the exploratory factor analysis extracted only one factor, the explained variance with only 37.80% is considerably too low. Furthermore, the average variance extracted with 0.34 does not meet its required threshold value of 0.5. Table 6-20. Adaptation measures for the construct improvement (3 indicators) Information on the factor involvement (3 indicators) 0.61 37.80% * * *

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.61 0.34

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated.

Information on the indicators Indicator 1 Indicator 2 Indicator 3

Item-to-total correlation 0.44 0.39 0.47

Indicator reliability 0.41 0.24 0.48

t-value of factor loadings 7.62 7.29

To improve the measurement model, indicator 2 with an indicator reliability of only 0.24 was eliminated. The resulting measurement model with only two indicators failed to meet the required criteria. While the indicators again loaded on one factor only, the value of the coefficient alpha was 0.613 and the explained variance only 44.278%. Other adaptation measures cannot be calculated with only two indicators and no degrees of freedom. The measurement model with two indicators, like the one with three indicators, cannot be used in the structural model. Since the measurement model has already been reduced to an almost inapt size, it was decided to also remove indicator 3. The remaining indicator 1 encompasses the general nature of the construct and should thus be able to measure on a general

6.1 Antecedents of logistics outsourcing performance

197

scale the degree of involvement of the LSP in the outsourcing project. While statistically this solution is not desirable, it still ensures the content validity of the construct. 6.1.8 Opportunism Opportunistic behavior is detrimental to any exchange relationship as it has been pointed out in chapter 4.2.1.8. It has consequently been studied by many scholars and has been found to be a critical antecedent also of the relationship between the customer and the LSP (KNEMEYER/MURPHY 2004, p. 41). For its measurement, two scales have been selected and adapted to the specific logistics outsourcing relationship context. The scale introduced by MORGAN/HUNT (1994, p. 35) with three indicators was not broad enough, while the scale developed by KNEMEYER/MURPHY (2004, p. 51) with nine indicators was too extensive for this study. Consequently, the essential aspects of both scales were merged into a new scale with six indicators as displayed in Table 6-21. Table 6-21. Indicators for the measurement of the construct opportunism

Indicator 1 Indicator 2 Indicator 3

Please indicate your level of agreement with the following statements on your relationship with this LSP To accomplish its own objectives, sometimes our LSP alters the facts slightly. Our LSP always provides us with a completely truthful picture of its activities. To accomplish its own objectives, sometimes our LSP promises to do things without actually doing them later.

Indicator 4

Our LSP sometimes exaggerates its requirements in order to get what it really needs from us.

Indicator 5 Indicator 6

Our LSP feels that it is alright to do anything within its means to further its own interests. Our LSP feels that honesty does pay when dealing with us.

MORGAN/HUNT (1994) and KNEMEYER/MURPHY (2004) all find that opportunistic behavior is characterized through the alteration of facts by one party to accomplish its own objectives. This was incorporated in form of indicator 1, which was only altered slightly to better fit into the language style of the questionnaire. The same thought as in indicator 1 is also visible in indicator 3, which was also taken from MORGAN/HUNT (1994, p. 35). It reflects opportunistic behavior in the form that promises are only used to pacify in the short run, while in the long no consequences or actions follow. The remaining four indicators were all taken from the scale proposed by KNEMEYER/MURPHY (2004, p. 51). Indicator 2 (reverse coded) measures whether the LSP is always truthful in its portrayal of its activities. If it is

198

6 Construct operationalization

not, the customer will potentially be deceived in the dimensions of its own logistics service capabilities or the true level of logistics costs, both being very detrimental. The same is true if the LSP continuously exaggerated its requirements to get what it really needs. Measured through indicator 4, this behavior may be viewed by some as simply being pragmatic, while the majority will perceive it as dishonest and opportunistic. Indicator 5 is measuring an even stronger version of this dishonesty. If the LSP is doing anything within its means to further its own interest, the relationship will be characterized by this opportunism and suffer in the long run. Finally, indicator 6 (reverse coded) was introduced to assess whether the customer perceives that the LSP feels the importance of honesty in the relationship. Should the relationship be characterized by this understanding, the risk of opportunistic behavior will be substantially reduced. The measurement model with six indicators was first tested with the criteria of the first generation. While the exploratory factor analysis extracted only one factor, the explained variance failed to meet the required value with only 45.19%. The confirmatory factor analysis, as portrayed in Table 6-22, furthermore showed that several other adaptation measures, like the F2/df and the RMSEA were significantly higher than considered acceptable, while others, such as the TLI, the AGFI and the average variance extracted, were slightly too low. Table 6-22. Adaptation measures for the construct opportunism (6 indicators) Information on the factor opportunism (6 indicators) 0.83 45.19% 13.503 0.84 0.93

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.84 0.92 0.151 0.83 0.47

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Item-to-total correlation 0.60 0.49 0.70 0.70 0.54 0.55

Indicator reliability 0.45 0.27 0.64 0.64 0.39 0.33

t-value of factor loadings 10.70 15.39 15.39 12.63 11.67

In an effort to improve the measurement model, indicator 2 with the lowest indicator reliability and an error term that was correlated with several other error term, was eliminated. The resulting measurement with 5 indicators, however, did still not meet all requested criteria either. Especially the explained variance with 48.635% was still not satisfactory. Since

6.1 Antecedents of logistics outsourcing performance

199

also the indicator reliability of the reverse coded indicator 6 had dropped to only 0.27, this indicator was also eliminated from the model. The resulting measurement model with four indicators is presented in Table 6-23. All adaptation measures show satisfactory values. The measurement model can therefore be accepted without any further modifications. Table 6-23. Adaptation measures for the construct opportunism (4 indicators) Information on the factor opportunism (4 indicators) 0.82 54.00% 1.311 1.00 1.00

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.99 1.00 0.024 0.82 0.54

Information on the indicators Indicator 1 Indicator 3 Indicator 4 Indicator 5

Item-to-total correlation 0.59 0.67 0.72 0.58

Indicator reliability 0.44 0.60 0.71 0.42

t-value of factor loadings 14.61 15.09 12.79

6.1.9 Shared values Shared values are important for exchange relationships and, as argued in chapter 4.2.1.9, have been found to govern the relationships between firms. They thereby constitute the basis for fundamental relationship variables such as relationship commitment and trust. While the construct has been used in relationship research, no established scale exists for logistics research. It was therefore decided to adapt the scale proposed by MORGAN/HUNT (1994, p. 35) along the lines of their argumentation what shared values really are which was introduced in chapter 4.2.1.9. Consequently, measurement of the behavior, goals, and policies of the partners must be included in the measurement model. To differentiate between the organizational behavior and goals as well as the personal behavior and goals, separate indicators are developed. Under the assumption that policies are only made on the organization level, it will be measured with only one indicator.

200

6 Construct operationalization

Table 6-24. Indicators for the measurement of the construct shared values

Indicator 1

Please indicate the level of agreement you believe exists between you and your LSP on the following issues If an employee is discovered to have engaged in unethical behavior that results primarily in personal gain, he should be promptly reprimanded.

Indicator 2

In business relationships, companies should not only consider their own short-term advantages, but also keep an eye on the long-term benefits for both parties involved.

Indicator 3

The employees on both sides have fully understood the goals of the relationship and act accordingly.

Indicator 4

The goals of the relationship are defined clearly and are being pursued equally by both parties.

Indicator 5

The basic understanding of our relationship is the same for both sides.

From this argumentation, a measurement model with five indicators results, which is displayed in Table 6-24. Indicator 1 reflects the level of agreement between the customer and the LSP if the personal behavior of employees is found to be detrimental to the firm. High levels of agreement in this issue indicate the existence of shared values. Similarly, indicator 2 measures the agreement on what is considered proper organizational behavior. If the parties have divergent views on the question of cooperation and joint partnership, the effects will most probably be detrimental for the relationship. In such a case, shared values cannot be assumed. Mutual goals are another important aspect. If they are understood and shared on a personal level, the employees will be more likely to work towards a functioning partnership. This is measured by indicator 3. Likewise, goals which are clearly defined and pursued by both organizations require a common set of values and have beneficial consequences for the relationship which is measured by indicator 4. Finally, indicator 5 was included in the measurement model to understand whether the general understanding of both parties regarding their relationship is the same. If it is, it becomes more likely that actions are taken which are favored by both parties. To achieve this, again shared values are a necessary prerequisite. The five-indicator measurement model was tested with an exploratory factor analysis which extracted only one factor. However, several adaptation measures failed to meet the required threshold values as presented in Table 6-25. While the explained variance was just acceptable at 50.88%, the F2/df at 19.577 and the RMSEA at 0.184 were significantly higher than desired. Likewise, the TLI and the AGFI both did not meet the required value of 0.9.

6.1 Antecedents of logistics outsourcing performance

201

Table 6-25. Adaptation measures for the construct shared values (5 indicators) Information on the factor shared values (5 indicators) Coefficient alpha Explained variance

0.82 50.88% 19.577 0.84 0.94

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.81 0.92 0.184 0.83 0.53

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5

Item-to-total correlation 0.42 0.57 0.70 0.73 0.69

Indicator reliability 0.15 0.29 0.66 0.79 0.64

t-value of factor loadings 7.64 8.70 8.83 8.67

In an effort to improve the fit of the measurement model, indicator 1 was eliminated due to its insufficient indicator reliability of only 0.15. The resulting measurement model with only four indicators, displayed in Table 6-26, shows consistently satisfactory adaptation measures. Only the indicator reliability of indicator 2 is below the proposed minimum of 0.4. However, since with indicator one the only other indicator measuring the behavior of the two parties has already been eliminated, it was decided to keep indicator 2 as part of the measurement model in order to ensure content validity. Therefore, no further modifications were made to the fourindicator measurement model. Table 6-26. Adaptation measures for the construct shared values (4 indicators) Information on the factor shared values (4 indicators) 0.84 59.38% 1.521 1.00 1.00

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.99 1.00 0.031 0.86 0.62

Information on the indicators Indicator 2 Indicator 3 Indicator 4 Indicator 5

Item-to-total correlation 0.49 0.73 0.78 0.72

Indicator reliability 0.27 0.66 0.81 0.64

t-value of factor loadings 12.11 12.48 12.02

6.1.10 Openness While openness so far has received little attention in interorganizational exchange relationship research, the argumentation chapter 4.3.1.10 has in-

202

6 Construct operationalization

dicated that it will have the role of an important antecedent for cooperative and sustainable relationships, since important relationship variables such as commitment and trust will be facilitated by high levels of openness. Since so far, openness has not been used as a construct in exchange relationship models, a new scale for its measurement had to be developed. A suitable framework is found in the empirical work of GUPTA (1987, p. 488), which showed that openness between corporate executives and strategic business unit managers has a positive impact on the effectiveness of strategy implementation. Openness in that context according to GUPTA (1987, p. 479) depends on four factors: the relationship must be open, informal, and must allow for spontaneous and open exchange of information and ideas. These four aspects, which were adapted from the scale proposed by GUPTA (1987, p. 499) for the measurement of openness, as well as two additional indicators form the six-indicator measurement model, are presented in Table 6-27. Table 6-27. Indicators for the measurement of the construct openness

Indicator 1 Indicator 2

To what degree do you agree with the following statements on your relationship with this LSP? The relationship with our LSP is informal. When problems or questions arise, we can quickly contact the LSP and jointly find solutions.

Indicator 3 Indicator 4

The relationship with our LSP is very open. We are comfortable exchanging very sensitive information with our LSP if we hope for advantages for our project.

Indicator 5

If one of the parties involved is not fully pleased with something, we immediately and openly talk about it.

Indicator 6

The LSP is always open and honest with us.

Indicator 1 measures whether the relationship is informal while indicator 2 quantifies the extent to which spontaneous solutions are found between the two parties once problems or questions arise. A further aspect suggested by GUPTA (1987, p. 499) is the general openness of the relationship, reflected through indicator 3. The degree to which both parties openly exchange information and ideas is measured through indicator 4, which takes up the question which is especially important in a logistics context whether both parties exchange sensitive information when necessary. Since the framework put forward by GUPTA (1987) was originally developed for interpersonal relationships, two further indicators were added. Indicator 5 measures the extent to which both parties also act openly as soon as they are dissatisfied with something which is an important prerequisite for beneficial changes. Finally, indicator 6 takes up the general no-

6.1 Antecedents of logistics outsourcing performance

203

tion of openness and combines it with honesty in order to identify if the openness is only superficial or also deeply rooted in both parties’ behavior. The six indicators were tested with an exploratory factor analysis which extracted only one factor and indicated an explained variance of 54.53%. While most of the criteria of the second generation were well above the requested minimum values (compare Table 6-28), the F2/df and the RMSEA failed to meet the threshold values. Table 6-28. Adaptation measures for the construct openness (6 indicators) Information on the factor openness (6 indicators) 0.87 54.53% 7.450 0.94 0.96

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.92 0.96 0.109 0.87 0.52

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Item-to-total correlation 0.59 0.72 0.81 0.53 0.64 0.73

Indicator reliability 0.40 0.65 0.82 0.32 0.46 0.61

t-value of factor loadings 15.53 16.68 11.70 13.64 15.12

To improve the fit of the measurement model, indicator 4 displaying the lowest indicator reliability was eliminated. This posed no problem for the content validity, as also indicator 2 incorporates the understanding of the need to act quickly and thoroughly when necessary, for instance when sensitive data is needed to improve the logistics process outcomes. The resulting five-indicator measurement model again showed good fit, except for the two adaptation measures that also in the six-indicator measurement model were not satisfactory. Especially the value of F2/df with 12.461 was substantially too high. To improve the model fit, indicator 6 was eliminated. While it did not have the lowest indicator reliability, its error term showed very high correlations with the error terms of indicators 1, 2, and 5. Furthermore, its elimination did not jeopardize content validity, as its aspect of general openness is also included in indicator 3, while the aspect of honesty may not be a necessary condition for openness. The resulting measurement model with four indicators presented in Table 6-29 showed very good adaptation measures in all dimensions. It can therefore be accepted without any further modifications.

204

6 Construct operationalization

Table 6-29. Adaptation measures for the construct openness (4 indicators) Information on the factor openness (4 indicators) 0.84 58.58% 0.738 1.00 1.00

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.99 1.00 0.000 0.85 0.59

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 5

Item-to-total correlation 0.58 0.75 0.77 0.60

Indicator reliability 0.40 0.73 0.78 0.43

t-value of factor loadings 15.50 15.65 12.95

6.2 Logistics outsourcing performance As it has been argued in chapter 4.1, the performance of logistics outsourcing arrangements can be and has been measured in many different ways. For this study, an outcome oriented measurement is favored as opposed to input, process or output orientation. On the basis of the conceptualization of logistics outsourcing performance, the following two chapters will in detail operationalize the constructs of goal achievement and goal exceedance. While for goal achievement some existing scales can be used as a basis, the construct of goal exceedance must be developed completely new. 6.2.1 Goal achievement While several scales exist to measure operational outsourcing performance and goal achievement, the scale developed by ENGELBRECHT (2004, pp. 212-218) was selected. On the one hand, it contains several indicators for the measurement of the two important facets improvement of the level of operational logistics performance and cost reductions. On the other hand, it has already been successfully used in a logistics outsourcing study with German manufacturing firms. ENGELBRECHT (2004, pp. 212-214) argues that aside from measuring the operative service and cost performance of logistics outsourcing, qualitative and strategic goals must be part of the goal achievement construct. While their quantification may be substantially more difficult than of service and cost performance, they are equally important. LYNCH (2000a, pp. 186-195) proposes a large number of performance measures in the logis-

6.2 Logistics outsourcing performance

205

tics outsourcing context. However, due to the nature of structural equation modeling, they will be aggregated to a more basic level. According to ENGELBRECHT (2004, p. 212), goal achievement in a logistics outsourcing arrangement depends on the performance in the three dimensions cost, quality, and time. To furthermore also cover qualitative and strategic goals, ENGELBRECHT (2004, p. 212) suggests to also measure the satisfaction of the customer with the LSP and the quality of the relationship. Therefore, the construct of goal achievement has the two aspects of achieving the actual goals agreed upon in the contract between the customer and the LSP and the relationship quality. ENGELBRECHT (2004, pp. 212-214) proposed seven indicators for the measurement of the construct, six of which were selected for this study. All of them were included with only some minor linguistic modifications. They are presented in Table 6-30. Table 6-30. Indicators for the measurement of the construct goal achievement Please indicate your level of agreement with the following statements on how satisfied you are with the relationship between this LSP and your company Indicator 1

Our LSP completely fulfills the goals and expectations we jointly set prior to this logistics outsourcing relationship.

Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

We are very satisfied with our LSP. The relationship with our LSP is very good. Our LSP delivers its services always with the required quality. Our LSP delivers its services always in the required time. Through this cooperation, our logistics costs have been reduced to the level we expected.

Indicator 1 measures the subjective perception of the customer on the outsourcing arrangement as a whole. The better the goals and expectations set prior to the logistics outsourcing are fulfilled, the higher is the goal achievement. Indicators 4, 5, and 6 also directly measure the goal achievement in the three dimensions quality, time, and costs. It must be noted, however, that all three indicators refer to required and expected levels respectively. Therefore, a high score on any of these three indicators reflects the achievement of the previously set goals and expectations only, not their exceedance. To measure the overall satisfaction of the customer with the relationship, indicator 2 was included. This dimension is extended through indicator 3, which measures the subjective perception of the customer of the relationship quality, which is understood as an important prerequisite for successful logistics outsourcing. Only if the relationship between the two parties works out fine, optimal results can be generated.

206

6 Construct operationalization

The indicator “if we had to decide again, we would again choose this LSP” was not included in the scale as several pre-test interviews suggested that the motives for choosing the LSP again are going significantly beyond the mere outsourcing performance. Here, factors such as personal ties, reputation, and others may become important which are not part of the construct goal achievement. The six-indicator construct was tested in an exploratory factor analysis which only extracted one factor. Most of the adaptation measures of first and second generation which are stated in Table 6-31 were satisfactory. However, the values of F2/df at 14.411 and RMSEA at 0.156 were significantly higher than the threshold values in chapter 5.2.4.3. To improve the measurement model, indicator 6 was removed as it had the lowest indicator reliability with 0.15. On the one hand, this does not threaten content validity, as the cost performance is also included in indicator 1, where the expectations can only be fulfilled if the cost performance is adequate. On the other hand, the low indicator reliability suggests that the achievement of the agreed upon cost levels is merely a hygiene factor. Since it is usually fixed in the contract, the customer bears no further risk and therefore does not view it as a determinant of goal achievement. Table 6-31. Adaptation measures for the construct goal achievement (6 indicators) Information on the factor goal achievement (6 indicators) 0.89 64.13% 14.411 0.92 0.93

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.83 0.95 0.156 0.89 0.58

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Item-to-total correlation 0.79 0.85 0.83 0.79 0.70 0.39

Indicator reliability 0.71 0.89 0.85 0.69 0.53 0.15

t-value of factor loadings 30.92 29.44 24.68 19.91 9.40

After the elimination of indicator 6, several of the adaptation measures get worse. F2/df rises to 23.754, while the RMSEA reaches a level of 0.204. As a result indicator 5 was eliminated. While it showed satisfactory indicator reliability, its error term exhibited very high correlations with the error terms of indicators 2 and 4, indicating the view of the customers that delivering services in time is a prerequisite rather than a special achievement,

6.2 Logistics outsourcing performance

207

which is also very much related to the quality of the service (indicator 4) and the overall satisfaction with it (indicator 2). The resulting measurement model presented in Table 6-32 showed excellent adaptation measures. The measurement model can therefore be accepted without any further modifications. Table 6-32. Adaptation measures for the construct goal achievement (4 indicators) Information on the factor goal achievement (4 indicators) 0.93 78.26% 2.729 1.00 1.00

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.98 1.00 0.056 0.94 0.78

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4

Item-to-total correlation 0.81 0.91 0.87 0.79

Indicator reliability 0.71 0.92 0.84 0.66

t-value of factor loadings 31.56 29.16 23.74

6.2.2 Goal exceedance Goal exceedance has been argued in chapter 4.1 to measure whether the logistics outsourcing performance of the LSP has significantly exceeded the goals and expectations set by the customer with regards to the outsourcing arrangement. In order to do so, the LSP must engage in activities going substantially beyond to what is needed to achieve the goals fixed in the contract. Among those activities and behaviors, customer orientation, innovation, and pro-activeness were identified. Since the construct goal exceedance has never been utilized in a logistics outsourcing context before, a new scale was developed. It will be based on the concept of goal achievement introduced in the previous chapter, for which the relevant performance aspects of logistics outsourcing have been identified. For a significant exceedance of the goals and expectations of the customer, the performance in these different aspects, namely quality of service and cost reductions, must go beyond those originally targeted. However, for the goal achievement, also the aspects of overall satisfaction and relationship quality were introduced. Since they constitute universal measures, referring to the state of the relationship in general rather than the level of goal exceedance, they are not included in the exceedance scale.

208

6 Construct operationalization

Resulting from this argumentation, four indicators were selected for the measurement of the construct of goal exceedance which are displayed in Table 6-33. Table 6-33. Indicators for the measurement of the construct goal exceedance

Indicator 1

Please indicate your level of agreement with the following statements on your satisfaction with your cooperation with this LSP The goals and expectations we jointly set prior to entering this relationship have been significantly exceeded.

Indicator 2

We are significantly more satisfied with the quality of the LSP services than we expected.

Indicator 3

Our expectations concerning the reduction of costs through this relationship were significantly exceeded.

Indicator 4

The relationship between actual costs for provided services and the overall service performance is much better than expected.

Indicator 1 measures the subjective perception of the customer with regards to the exceedance of the goals and expectations set prior to the commencement of the outsourcing arrangement. The further these are exceeded by the LSP, the higher is the goal exceedance and consequently also the value for the customer. Indicator 2 and 3 in turn measure the operative exceedance in terms of service quality (indicator 2) and cost reductions (indicator 3). As for indicator 1, it can be argued that surpassing the expectations regarding the quality of logistics services and cost reductions will be a significant value added to the customer and simultaneously lead to goal exceedance. Finally, indicator 4 is added as an overall measure of the relationship between costs and services provided by the LSP. Since arguably costs may even rise when the level of logistics services is increased by the LSP, it must be measured whether this still is of superior value to the customer. As several combinations between increases and decreases in the level of logistics services and logistics costs are thinkable, the indicator measures the perceived value for the customer in terms of exceeding its expectations in aggregate. For all indicators it must be noted, however, that they distinctly refer to exceeding the expected goals and expectations. Therefore, they clearly are nomologically distinct to the measures of simple goal achievement. The four-indicator measurement model was tested in an exploratory factor analysis. While it only extracted one factor and showed a satisfactory explained variance of 58.98%, especially the criteria of the second generation as displayed in Table 6-34 were in large parts insufficient. Especially the TLI at 0.19, the AGFI at -0.05, the RMSEA at 0.585 and the F2/df at 188.642 are by no means acceptable.

6.2 Logistics outsourcing performance

209

Table 6-34. Adaptation measures for the construct goal exceedance (4 indicators) Information on the factor goal exceedance (4 indicators) 0.85 58.98% 188.642 0.19 0.79

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

-0.05 0.73 0.585 0.82 0.55

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4

Item-to-total correlation 0.69 0.70 0.62 0.75

Indicator reliability 0.81 0.84 0.25 0.39

t-value of factor loadings 25.46 12.21 16.23

In an effort to improve the fit of the measurement model, indicator 3 referring to cost reductions was removed due to its low indicator reliability. This was not to be expected since costs reductions have been shown to be a premier motivator for logistics outsourcing in chapter 5.1.6. However, this might be an indication that under the current contract practice of logistics outsourcing arrangements, LSPs either find no room for goal exceedance in terms of cost reductions as they are already operating at the limit, or simply are not motivated to reduce them further than originally agreed. This may lead to the perception among their customers that a substantial exceedance of the originally targeted cost reductions is not possible and therefore is not a relevant aspect of their understanding of goal exceedance. However, its elimination does pose no problem for the content validity, as some cost reduction aspects remain in indicators 1 and 4. After the removal of indicator 3, a three-indicator measurement model results. As Table 6-35 indicates, the reduced set of fit criteria showed satisfactory fit in almost all aspects. While all criteria measuring the quality of the construct as a whole are fully sufficient, only the indicator reliability of indicator 4 at 0.34 is slightly below the requested minimum value of 0.4. However, since all other criteria are satisfactory and indicator 4 due to its comparative nature between logistics service levels and costs is important for the scale, it was decided to keep the indicator and therefore accept the three-indicator measurement model without any further modifications.

210

6 Construct operationalization

Table 6-35. Adaptation measures for the construct goal exceedance (3 indicators) Information on the factor goal exceedance (3 indicators) 0.84 67.48% * * *

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.85 0.66

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated.

Information on the indicators Indicator 1 Indicator 2 Indicator 4

Item-to-total correlation 0.77 0.79 0.56

Indicator reliability 0.80 0.88 0.34

t-value of factor loadings 31.56 29.16

6.3 Logistics performance The measurement of the effect of logistics outsourcing performance on the overall logistics performance of a firm is a main goal of this research and is subject of research question 3 which was introduced in chapter 2.4.2. The conceptualization of the construct of logistics performance was presented in chapter 4.4.1.1. There, it was pointed out that logistics performance is composed of the two dimensions, the level of logistics services and the level of logistics costs. These dimensions were operationalized into measurement models in the research of DEHLER (2001, pp. 206-225) and ENGELBRECHT (2004, pp. 218-224). Due to their suitability for the analysis of the relationships between logistics outsourcing performance, outsourcing performance and firm performance as well as the potential for the examination of contingency variables, they were selected for this study. In the following chapters, the operationalization of the constructs will be presented. 6.3.1 Level of logistics services As pointed out in chapter 4.4.1, logistics performance has been measured in several different ways (CHOW/HEAVER/HENRIKSSON 1994). For this study, the scale developed and tested by ENGELBRECHT (2004, pp. 221224) was selected as it has proven to be both reliable and valid in a largescale German logistics outsourcing survey. The scale developed by ENGELBRECHT (2004, pp. 221-224) is based upon the framework of BOWERSOX/CLOSS/HELFERICH (1986, pp. 27-28) presented in chapter 4.4.1.1. Following it, the measurement model for the

6.3 Logistics performance

211

level of logistics services must include measurement of the firms’ performance on three successive levels: the ability to on-time deliver products, material, and information on a first level, the improvement of the performance capabilities of the firms’ logistics processes on a second level, and the ensuring of the permanent quality of the logistics processes on a third level. ENGELBRECHT (2004, pp. 221-222) suggests five indicators to measure the performance on the three given levels. According to his research, the ability to on-time deliver products, material, and information constitutes a hygiene factor before the other levels may be addressed. For its measurement, the indicator “ability for on-time delivery” is proposed. The second level addresses the performance of the logistics processes in the dimensions of time and flexibility. Time in this context reflects the ability to convert customer orders in product output, e.g. shipped orders. This is measured through the indicators “delivery times” and “order lead times”. Flexibility, on the other hand, marks the ability of a firm to react to changes in the environment. This is measured through the indicators “delivery flexibility”. Finally, the quality of the logistics processes must be permanently ensured, which is measured through an indicator regarding the “degree of damage- and error free logistical activities”. The five indicators developed by ENGELBRECHT (2004, pp. 221-224) were all selected for the measurement model in this study and only slightly adapted linguistically. However, to also include a further qualitative aspect of the abilities to deliver, the indicator “delivery reliability” suggested by DEHLER (2001, p. 209) was included in the measurement model. The final selection of indicators for the measurement model is presented in Table 636. Table 6-36. Indicators for the measurement of the construct level of logistics services

Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

How does your company's performance on logistics service compare to your competitors? Order lead times Delivery times. Ability for on-time delivery. Delivery reliability. Degree of delivery flexibility. Degree of damage- and error-free logistical activities.

To test the measurement model, an exploratory factor analysis was conducted which only extracted one factor. However, as Table 6-37 shows, several adaptation measures indicate insufficient quality of the measurement model. Especially the value of F2/df at 14.448 and the RMSEA at

212

6 Construct operationalization

0.157 are considerably higher than the required threshold values of 3.0 and 0.08 respectively. To improve the measurement model, indicator 3 was eliminated due to the high correlations of its error term with those of the indicators 1, 4, and 6. Apparently, respondents could not adequately differentiate between the ability for on-time delivery and related aspects, such as delivery reliability and order lead times. Table 6-37. Adaptation measures for the construct level of logistics services (6 indicators) Information on the factor level of logistics services (6 indicators) 0.90 59.84% 14.448 0.89 0.93

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.83 0.94 0.157 0.90 0.60

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Item-to-total correlation 0.70 0.76 0.75 0.78 0.74 0.60

Indicator reliability 0.54 0.65 0.68 0.71 0.61 0.39

t-value of factor loadings 18.54 18.98 19.29 17.97 14.27

While the elimination of indicator 3 did improve the values of F2/df and RMSEA somewhat, they still did not meet the required threshold. Consequently, with indicator 1 another item was eliminated from the measurement model. Its error term displayed a very high correlation with the error term of indicator 2 as well as correlations with the error terms of indicators 4 and 5. This suggests that order lead times and delivery times are very hard for the respondents to distinguish or in their understanding resemble related concepts. Therefore, the elimination is justified also on grounds of content validity. The resulting measurement model, displayed in Table 6-38, showed very good adaptation measures. The measurement model with four indicators is therefore accepted without any further modification.

6.3 Logistics performance

213

Table 6-38. Adaptation measures for the construct level of logistics services (4 indicators) Information on the factor level of logistics services (4 indicators) 0.85 59.14% 2.301 0.99 1.00

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.98 1.00 0.049 0.85 0.59

Information on the indicators Indicator 2 Indicator 4 Indicator 5 Indicator 6

Item-to-total correlation 0.67 0.75 0.73 0.60

Indicator reliability 0.57 0.69 0.67 0.44

t-value of factor loadings 18.34 18.13 14.80

6.3.2 Level of logistics costs The level of logistics costs could be measured in a multitude of different ways, depending on the individual understanding of the term logistics costs. In this study, it will be operationalized along the lines of the argumentation in chapter 4.4.1.1. The construct focuses on the core logistics costs with some aspects of extended logistics costs, since they constitute the lowest common denominator that virtually all firms agree upon as will be explained in detail below. In the research of DEHLER (2001, p. 211) and ENGELBRECHT (2004, pp. 220-221), this has proven to be valid, reliable, and effective. Core logistics costs in this context basically are warehousing and transportation costs, while extended logistics costs comprise those for picking and packaging, for the capital employed or for product returns (BAUMGARTEN/BOTT/HAGEN 1997, p. 24). The operationalization developed by ENGELBRECHT (2004, pp. 218-221) proposes five indicators for the measurement model: The core logistics costs are covered through indicators measuring transportation costs and the days of inventory coverage while the extended logistics costs are accounted for in a simplified fashion by quantifying the ratio of logistics costs relative to the total revenue. Two further indicators measure the cost for logistics related IT and personnel as they are the two major cost drivers of logistics functions.

214

6 Construct operationalization

Table 6-39. Indicators for the measurement of the construct level of logistics costs

Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5

How does your company's performance on logistics cost compare to your competitors? Logistics costs relative to total revenue (compensation for the LSP included). Warehousing costs. Transportation costs. IT costs in support of logistics activities. Human resource costs of logistics activities.

Table 6-39 presents the indicators used in this study. Indicators 1, 3, 4, and 5 were directly taken from ENGELBRECHT (2004, pp. 219-221) and only slightly adapted linguistically. Indicator 2 was modified from measuring the days of inventory coverage to the overall warehousing costs. This promises to have a more direct connection to the cost incurred through the warehousing than the mere measurement of the range of coverage through the inventory. As Table 6-40 indicates, the adaptation measures for the measurement model with five indicators were not satisfactory. While the exploratory factor analysis extracted only one factor, the explained variance with 39.12% is well below the required 50%. Also, the average variance extracted is too low at 0.38 and several other fit criteria are not meeting the threshold values. Particularly, the F2/df at 14.242 and the RMSEA at 0.155 are substantially too high. Table 6-40. Adaptation measures for the construct level of logistics costs (5 indicators) Information on the factor level of logistics costs (5 indicators)

Coefficient alpha Explained variance

0.75 39.12% 14.242 0.80 0.95

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.84 0.90 0.155 0.75 0.38

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5

Item-to-total correlation 0.64 0.53 0.51 0.43 0.47

Indicator reliability 0.73 0.36 0.48 0.17 0.22

t-value of factor loadings 12.50 14.04 8.87 9.93

In an effort to improve the quality of the measurement model, indicator 4 was eliminated. It showed the lowest indicator reliability and the lowest item-to-total correlation, indicating the weakest connection to the model of all five constructs. This hints at the fact that logistics IT-costs are comparably independent from other logistics costs or at least are perceived to be.

6.4 Firm performance

215

While the elimination of indicator 4 improved the model fit, the measurement model still was not satisfactory. The same adaptation which were not sufficient before failed to meet the thresholds again. Indicator 5, which measured the cost for logistics personnel, only had an indicator reliability of 0.18. Since these costs are also part of the overall transportation and warehousing costs as well as the ratio of logistics costs to total revenue which are measured in indicators 1 through 3, indicator 5 was also eliminated. The resulting measurement model with three indicators showed very good adaptation measures in almost all dimensions. Explained variance and average variance extracted both exceed the requested minimum values. Only the indicator reliability of indicator 2 is with 0.27 below the proposed 0.4. Since the other criteria are satisfactory and it is desirable for content validity to measure both transportation and warehousing costs, the indicator was maintained in the three-indicator measurement model. It was therefore accepted without any further modifications. Table 6-41. Adaptation measures for the construct level of logistics costs (3 indicators) Information on the factor level of logistics costs (3 indicators)

Coefficient alpha Explained variance

0.74 54.51% * * *

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.76 0.52

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated.

Information on the indicators Indicator 1 Indicator 2 Indicator 3

Item-to-total correlation 0.70 0.46 0.54

Indicator reliability 0.97 0.27 0.40

t-value of factor loadings 8.98 9.88

6.4 Firm performance In chapter 4.4.2 it has been pointed out that the measurement of firm performance has received much attention in business administration research, which is also reflected in the multitude of different scales and measurement approaches available (BHARGAVA/DUBELAAR/RAMASWAMI 1994). For this study, the three factors of adaptiveness, market performance, and financial performance have been selected as dimensions of the construct of firm performance. While they have originally been developed in the marketing domain (RUEKERT/WALKER JR./ROERING 1985; IRVING 1995), they

216

6 Construct operationalization

were adapted for the logistics context by DEHLER (2001) and modified by ENGELBRECHT (2004). In both studies, they have demonstrated their suitability and therefore will be utilized in this research with only minor modifications. It must be noted, however, that as argued in chapter 4.4.2, due to the disadvantages of objective performance criteria in large-scale survey research with individual respondents filling out questionnaires, only subjective performance measures will be used in this study. The constructs are presented in the following chapters. 6.4.1 Adaptiveness As it was argued in chapter 4.4.2.1, a firm’s adaptiveness reflects the ability to adapt to changes in the market (RUEKERT/WALKER JR./ROERING 1985; IRVING 1995). Through the capability of flexible reactions to changes in the environment and varying demands of the market, the firm can benefit from arising opportunities. Both DEHLER (2001, pp. 228-229) and ENGELBRECHT (2004, pp. 227228) use the same three indicators to measure the adaptiveness of a firm, which were also selected for this study. They are presented in Table 6-42. Indicator 1 measures the ability of the firm to adapt its products or services to new customer requirements. Indicator 2 determines how flexibly the firm reacts to new developments in the market, and indicator 3 assesses to what extent it benefits from new opportunities in the market. Table 6-42. Indicators for the measurement of the construct adaptiveness How does your company's overall responsiveness compare to your competitors? Indicator 1 Indicator 2 Indicator 3

Adaptation of products / services to new customer requirements. Reacting to new developments in the market. Benefiting from new opportunities in the market.

Following the standard procedure for the testing of measurement models, an exploratory factor analysis was conducted. It extracted only one factor and showed an explained variance of 76.76%. Even though the set of adaptation measures was reduced as a measurement model with three indicators has no degrees of freedom, the remaining measures showed a very high fit. The measurement model can therefore be accepted without any modifications.

6.4 Firm performance

217

Table 6-43. Adaptation measures for the construct adaptiveness (3 indicators) Information on the factor adaptiveness (3 indicators)

Coefficient alpha Explained variance

0.91 76.76% * * *

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.91 0.77

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated

Information on the indicators Indicator 1 Indicator 2 Indicator 3

Item-to-total correlation 0.78 0.87 0.79

Indicator reliability 0.68 0.92 0.70

t-value of factor loadings 26.30 23.73

6.4.2 Market performance Market performance is a measure for the effectiveness of a firm in the market (RUEKERT/WALKER JR./ROERING 1985; IRVING 1995) and therefore is of a high strategic importance to the firm. Following IRVING (1995), DEHLER (2001, pp. 227-228) developed a measurement model with six indicators for the logistics context from which one had to be eliminated due to insufficient model fit. The scale was also used by ENGELBRECHT (2004, pp. 225-227). The six original indicators will be used in this study and are presented in Table 6-44 and outlined in the following. High market performance can result in customer satisfaction, into additional benefits for the customer (indicator 2) and in customer loyalty, measured thorough customer retention (indicator 3). Furthermore, quantitative aspects allow an insight into the market performance of a firm: new customers (indicator 4), as well as growth (indicator 5) and market share (indicator 6) according to the firm’s own expectations all reflect high market performance. Table 6-44. Indicators for the measurement of the construct market performance How does your company's market performance compare to your competitors? Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Customer satisfaction. Value added provided to the customer. Retention of existing customers. Acquisition of new customers. Achieving our desired growth rate. Achieving our desired market share.

The measurement model consisting of six indicators was tested in an exploratory factor analysis and extracted only one factor. The resulting ad-

218

6 Construct operationalization

aptation measures, however, were by no means satisfactory. Especially the F2/df at 61.403 and the RMSEA at 0.332 were significantly above the requested maximum values. Other criteria of the second generation, such as the TLI, the GFI, the AGFI, and the CFI all failed to reach the required minimum of 0.9. In an effort to improve the quality of the measurement model, indicator 4 was eliminated. While it did display a satisfactory indicator reliability, its error term showed a high correlation with the error terms of all other five indicators. Furthermore, the high correlation of the error terms of indicators 1 and 2 suggested that the respondents did not sufficiently differentiate between customer satisfaction and value added to the customer. Obviously, it is perceived that whenever the value added for the customers is high, so will be its satisfaction. However, this finding did not lead to a modification of the measurement model. Table 6-45. Adaptation measures for the construct market performance (6 indicators) Information on the factor market performance (6 indicators)

Coefficient alpha Explained variance

0.90 58.91% 61.403 0.62 0.73

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.38 0.77 0.332 0.89 0.58

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5 Indicator 6

Item-to-total correlation 0.71 0.69 0.68 0.68 0.77 0.78

Indicator reliability 0.35 0.33 0.36 0.52 0.88 0.88

t-value of factor loadings 11.51 11.92 13.64 16.05 16.06

While these modifications resulted in a measurement model with significantly improved fit criteria, the F2/df-value at 6.442 and the RMSEA at 0.100 were still too high. All other measures exceeded the required threshold values. As the correlation of the error terms between indicators 5 and 6 suggested that also here, respondents could not differentiate between achieving desired levels of market share and growth rate, indicator 5 was eliminated for reasons of content validity, as the market share was perceived as a more important measure for market performance than the growth rate. The resulting measurement model with only four indicators as presented in Table 6-46 showed satisfactory adaptation measures in all dimensions. It could therefore be accepted without any further modifications.

6.4 Firm performance

219

Table 6-46. Adaptation measures for the construct market performance (4 indicators) Information on the factor market performance (4 indicators)

Coefficient alpha Explained variance

0.85 61.73% 0.989 1.00 0.99

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.99 1.00 0.000 0.84 0.57

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 6

Item-to-total correlation 0.77 0.75 0.71 0.57

Indicator reliability 0.65 0.60 0.68 0.41

t-value of factor loadings 23.85 15.34 13.72

6.4.3 Financial performance For the measurement of firm performance, the return on sales with respect to the competition is used as a subjective performance indicator. This reflects a compromise between the explanatory value of the performance indicator and the availability of information to the respondent. More precise performance indicators, such as return on equity, were not used due to their sometimes scarce utilization in practice and the additional information challenge it would have posed for the respondent. The measurement model adapted from ENGELBRECHT (2004, pp. 229231) contains three indicators. Aside from the return on sales in the last year (indicator 1), the average return on sales in the last three years is also measured (indicator 2) to incorporate a view on a longer trend. Finally, the development of the return on sales as a measure for the operative performance of the business is also taken into account (indicator 3). Together, these three indicators reflect the status quo and the development of the overall financial performance of the firm. Table 6-47. Indicators for the measurement of the construct financial performance How does your company's financial performance compare to your competitors? Indicator 1 Indicator 2

Our return on sales last year with respect to our competition was... Our average return on sales over the past three years with respect to our competition was...

Indicator 3

The development of our return on sales during the past three years with respect to our competition was...

The measurement model was tested in an exploratory factor analysis and extracted only one factor. All adaptation measures showed very good re-

220

6 Construct operationalization

sults. The measurement model therefore is fully satisfactory and can be accepted without any modifications. Table 6-48. Adaptation measures for the construct market performance (3 indicators) Information on the factor financial performance (3 indicators)

Coefficient alpha Explained variance

0.95 85.29% * * *

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.95 0.85

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated

Information on the indicators Indicator 1 Indicator 2 Indicator 3

Item-to-total correlation 0.89 0.88 0.89

Indicator reliability 0.86 0.84 0.86

t-value of factor loadings 36.30 37.54

6.5 Discriminant validity of the operationalized constructs After the different measurement models have been operationalized, it must be analyzed whether discriminant validity can be assumed in all dimensions. As introduced in chapter 5.2.3.2, this will be done through the Fornell/Larcker-criterion (FORNELL/LARCKER 1981, p. 41). While the criterion could be applied to relationships between factors only where a lack of discriminant validity could be ex-ante assumed, in this study a more rigorous approach will be followed. For each of the three models proposed in chapter 4, namely logistics outsourcing performance and its antecedents, the effect of logistics outsourcing performance on logistics performance, and the effect of logistics performance on firm performance, discriminant validity between all variables will be tested. While this means a redundant testing of some relationships, the results ensure a necessary and detailed insight into the discriminant validity of all constructs used in this study. The models will be tested in the following three chapters. 6.5.1 Antecedents and dimensions of logistics outsourcing performance As Table 6-49 indicates, the CFA model of logistics outsourcing performance displays some very good adaptation measures.

6.5 Discriminant validity of the operationalized constructs

221

Table 6-49. Adaptation measures for the model logistics outsourcing performance Information on the model logistics outsourcing performance Coefficient alpha Explained variance

* * 1.95 0.95 0.89

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted (AVE)

0.86 0.96 0.042 * *

* These values are only calculated on the facor level.

Information on the indicators Factor

Cooperation

Communication

Proactive Improvement

Trust

Commitment

Functional Conflict Involvement Opportunism

Shared Values

Openness

Goal Achievement

Goal Over-Achievement

Indicator 2 3 4 6 2 5 6 7 2 3 4 5 1 2 3 5 1 2 4 1 2 7 1 1 3 4 5 2 3 4 5 1 2 3 5 1 2 3 4 1 2 4

Indicator reliability 0.61 0.68 0.43 0.60 0.54 0.48 0.71 0.73 0.70 0.83 0.75 0.59 0.63 0.82 0.84 0.55 0.54 0.72 0.65 0.67 0.80 0.57 0.88 0.44 0.64 0.66 0.41 0.26 0.67 0.76 0.69 0.41 0.67 0.82 0.45 0.70 0.90 0.86 0.68 0.82 0.85 0.35

Factor reliability

AVE

0.84

0.57

0.86

0.61

0.91

0.72

0.91

0.71

0.84

0.64

0.86

0.67

0.88

0.88

0.82

0.54

0.86

0.62

0.85

0.59

0.94

0.79

0.85

0.66

These adaptation measures indicate sufficient discriminant validity. Both GFI and AGFI are slightly below the requested minimum value of 0.9. However, since the model is very complex, this is acceptable. The other criteria of the second generation, namely F2/df, TLI, CFI, and

222

6 Construct operationalization

RMSEA are all satisfactory. Furthermore it must be noted that 2 out of 42 indicators fail to meet the required indicator reliability of 0.4. However, for both Shared Values 2 and Goal Exceedance 4, this has already been argued in the respective chapters to be acceptable. Finally, the question of discriminant validity can be answered. The Fornell/Larcker-criterion is tested in Table 6-50, where the average variance extracted (AVE) of each construct is compared to the squared correlations between the different constructs. Altogether, 62 relationships were tested. In only ten cases, the squared correlation between the constructs exceeded one or both terms of the average variance extracted. Thus, a high discriminant validity between the constructs can be assumed. In the case of the pair cooperation/communication, discriminant validity cannot immediately be established. However, when comparing the scales that are used for the measurement of both constructs, it becomes evident that both measurement models measure two distinct constructs. The high correlation between them therefore is not considered as threatening the explanatory value of the model, but is rather seen as an indicator of the textual linkage between the two. Similarly, the squared correlations between cooperation and trust as well as commitment are higher than the average variance extracted for both trust and commitment. A comparison of the scales involved reveals that the constructs measure considerably different constructs. Again, the high correlation between them originates from the strong direct effects that were hypothesized in chapter 4.3.1.1, after which both trust and commitment are direct antecedents of cooperation. The construct of openness shows higher levels of squared correlations than the respective levels of AVE with the constructs cooperation, communication, trust, and functional conflict. These high correlations, however, are not the consequence of missing discriminant validity, but rather the result of the strong direct positive effect that openness is supposed to have on these variables. This effect has been hypothesized in chapter 4.3.1.10 for communication, trust and functional conflict. Since in this chapter, a high correlation has been found between cooperation and communication, it can be argued that the perceived effect of openness on cooperation exists just as much as the positive effect on communication. Furthermore, exploratory factor analyses provide additional support for discriminant validity by showing that if the Kaiser-criterion is relaxed and two factors as a combination of openness and a second one of the above mentioned four factors are extracted, the indicators of the two constructs load on two distinct factors for every case.

6.5 Discriminant validity of the operationalized constructs

223

Table 6-50. Discriminant validity of the antecedents and dimensions of logistics outsourcing performance

Factor (1) Cooperation (2) Communication (3) Proactive Improvement (4) Trust (5) Commitment (6) Functional Conflict (7) Involvement (8) Opportunism (9) Shared Values (10) Openness (11) Goal Achievement (12) Goal Over-Achievement

Factor AVE

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9) (10) (11) (12)

0.57

0.57 0.61 0.72 0.71 0.64 0.67 0.88 0.54 0.62 0.59 0.79 0.66 -

0.61

0.71

0.72

0.38 0.26

squared correlations

-

0.71

0.71 0.61 0.35

0.64

0.64 0.46 0.28 0.56

-

0.67

0.77 0.83 0.34 0.78 0.56

0.88

0.25 0.29 0.05 0.16 0.21 0.27

0.54

0.39 0.29 0.22 0.44 0.29 0.37 0.05

0.62

0.57 0.53 0.29 0.52 0.41 0.63 0.10 0.32

0.59

0.64 0.74 0.27 0.63 0.47 0.82 0.27 0.28 0.51

0.79

0.49 0.42 0.29 0.64 0.47 0.55 0.10 0.31 0.45 0.43

0.66

0.27 0.16 0.29 0.25 0.22 0.22 0.04 0.15 0.21 0.15 0.29

-

Finally, also the construct of functional conflict displays higher squared correlations than the respective AVE-value with the constructs cooperation, communication, and trust. Other than in the case of openness as discussed above, here discriminant validity cannot be assumed. While it could be argued that the strong direct effects these three constructs have on the functional treatment of conflicts as argued in chapter 4.3.1.6 are the reasons for the high correlations, exploratory factor analyses suggest that discriminant validity indeed does not exist. When the Kaiser-criterion is relaxed and two factors as a combination of functional conflict and either cooperation, communication, or trust, are extracted, it shows that none of the combinations are fully discriminant. In all cases, at least one indicator loads not on its own factor, but on the other factor, suggesting a different factor structure. This is a clear sign that functional conflict is not sufficiently discriminant from cooperation, communication, and trust. Therefore, it must be assumed that functional conflict and the other three factors are at least partially measuring the same, leading to lacking discriminant validity. Consequently, the factor functional conflict must be excluded from all further analyses, leaving a remainder of 9 antecedents of logistics outsourcing performance. 6.5.2 Logistics outsourcing performance and logistics performance In chapter 4.4.1.2, several direct effects of the dimensions of logistics outsourcing performance on the dimensions of logistics performance were hypothesized. In an effort to rigorously assess the discriminant validity of

224

6 Construct operationalization

these constructs, a confirmatory factor analysis was performed. Aside from the level of logistics services and level of logistics costs, this includes also the constructs of goal achievement and goal exceedance, which have already been assessed in the previous chapter. As Table 6-51 indicates, the model exhibits very satisfactory adaptation measures. While F2/df, TLI, GFI, AGFI, CFI, and RMSEA all exceed the required threshold values, the indicator reliability of the level of logistics costs 2 at 0.30 fails to meet the recommended value of 0.40. However, this has been argued in chapter 6.3.2 to be acceptable. Table 6-51. Adaptation measures for the model logistics performance Information on the model logistics performance * * 1.54 0.99 0.97

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted (AVE)

0.96 0.99 0.031 * *

* These values are only calculated on the facor level.

Information on the indicators Factor

Goal Achievement

Goal Exceedance

Level of Logistics Services

Level of Logistics Costs

Indicator 1 2 3 4 1 2 4 2 4 5 6 1 2 3

Indicator reliability 0.71 0.92 0.84 0.66 0.83 0.85 0.34 0.57 0.69 0.66 0.45 0.86 0.30 0.45

Factor reliability

AVE

0.94

0.79

0.85

0.66

0.85

0.59

0.76

0.52

For the assessment of the overall discriminant validity, the Fornell/Larcker-criterion was again applied. As Table 6-52 indicates, the average variance extracted of the constructs exceeds the squared correlations between the constructs in every case. Therefore, discriminant validity between all constructs in the model can be assumed.

6.5 Discriminant validity of the operationalized constructs

225

Table 6-52. Discriminant validity of the dimensions of logistics outsourcing performance and logistics performance

Factor (1) Goal Achievement (2) Goal Exceedance (3) Level of Logistics Services (4) Level of Logistics Costs

Factor AVE 0.79 0.66 0.59 0.52

(1) 0.79 0.29 0.08 0.07

(2) 0.66

(3) 0.59 squared correlations

0.08 0.04

(4) 0.52

0.22

-

6.5.3 Logistics performance and firm performance As it has been for logistics outsourcing performance and logistics performance, the discriminant validity for the constructs involved in the model of the effects of logistics performance on firm performance will be assessed. Table 6-53. Adaptation measures for the model firm performance Information on the model firm performance * * 2.60 0.96 0.94

Coefficient alpha Explained variance F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted (AVE)

0.92 0.97 0.054 * *

* These values are only calculated on the facor level.

Information on the indicators Factor Level of Logistics Services

Level of Logistics Costs

Adaptiveness

Market Performance

Financial Performance

Indicator 2 4 5 6 1 2 3 1 2 3 1 2 3 6 1 2 3

Indicator reliability 0.57 0.70 0.65 0.45 0.81 0.32 0.47 0.70 0.90 0.72 0.78 0.74 0.57 0.39 0.85 0.84 0.86

Factor reliability

AVE

0.85

0.59

0.76

0.52

0.91

0.77

0.86

0.60

0.95

0.85

As Table 6-53 indicates, the confirmatory factor analysis reveals satisfactory adaptation measures in virtually all dimensions. While F2/df, TLI, GFI, AGFI, CFI, and RMSEA are all sufficient with regard to the required values, the indicator reliability of the indicators level of logistics costs 2 (0.32) and market performance 6 (0.39) both fail to meet the recommended

226

6 Construct operationalization

value of 0.40. As argued in the respective chapters, these slight deviations are acceptable. To test for discriminant validity, the Fornell/Larcker-criterion was applied again. As Table 6-54, the average variance extracted of the measurement models exceeds in every case the squared correlations between each pair of constructs. Therefore, discriminant validity can be assumed between all constructs. Table 6-54. Discriminant validity of the dimensions of logistics performance and firm performance Factor (1) Level of Logistics Services (2) Level of Logistics Costs (3) Adaptiveness (4) Market Performance (5) Financial Performance

Factor AVE 0.59 0.52 0.77 0.60 0.85

(1) 0.59 0.23 0.15 0.48 0.11

(2) 0.52

(3) 0.77

(4) 0.60

(5) 0.85

0.07 0.23 0.12

squared correlations 0.22 0.14 0.23 -

6.6 Contingency factors The operationalization of the contingency variables is build upon their conceptualization in chapter 4.5.2. Due to the presented divergent nature of the variables, the ways of their operationalization must also differ. While for those deemed suitable, a measurement as a complex, latent construct was attempted, others were operationalized more directly. For the operationalization, existing scales primarily developed by KLEER (1991) were used wherever possible. However, as the following two chapters on the operationalization of both external and internal contingency variables will show, their quality has proven to be very limited. Further use as moderating variables in structural equation modeling can therefore not be recommended before a substantial revision of their underlying concepts is performed. 6.6.1 External contingency variables As argued in chapter 4.5.2.1, the moderating effects of four external contingency variables on the proposed models will be tested in this study. Three of them, namely environmental complexity, environmental dynamics, and uncertainty will be operationalized as constructs with multiple indicators. The fourth variable, the customer’s industry, will be surveyed with a direct question.

6.6 Contingency factors

227

Following the operationalization of KLEER (1991, p. 121) introduced in chapter 4.5.2.1, the four indicators displayed in Table 6-55 are selected for the measurement of environmental complexity. Indicator 1 measures the homogeneity of the customer structure while indicator 2 does the same by weighing the complexity that is added through inhomogeneous ordering of the customers. Indicator 3 is a criterion for geographic complexity and indicator 4 is determining the number of LSPs the customer is regularly working with. Table 6-55. Indicators for the measurement of the construct environmental complexity

Indicator 1

Please indicate your level of agreement with the following statements about the present situation of your company Our customers vary significantly in size, resulting in considerably different sales volumes.

Indicator 2

The orders of our customers are very unevenly distributed and differ depending on the specific order or season.

Indicator 3

Compared to our competitors, we have to deliver our products to a very large number of different locations.

Indicator 4

Compared to our competitors, we are working together with a very large number of LSPs.

An exploratory factor analysis extracted two factors. Indicators 1 and 2 loaded on the first factor, while indicators 3 and 4 loaded on the second. Obviously, a difference exists between the complexity generated through factors the customer can in the broadest sense influence independently, namely the distribution network in the form of delivery locations and LSPs involved, and the complexity induced through inhomogeneity of customer structure and ordering patterns. Generally, it would be worthwhile testing for potential moderating effects of both factors. However, this requires sufficient levels at both the explained variance and the coefficient alpha. Further adaptation measures cannot be calculated as the factors with only two indicators have no degrees of freedom. For factor 1, measuring the complexity induced inhomogeneity of customer structure and ordering patterns (indicators 1 and 2), the explained variance at 49.5% and the coefficient alpha at 0.66 are just satisfactory. The second factor, constituted of indicators 3 and 4, only exhibits an explained variance of 30.1% and a coefficient alpha of 0.45. Therefore, it cannot be treated and utilized for the analysis as one factor. For the further analysis of moderating effects, only the environmental complexity caused by inhomogeneity of customer structure and ordering patterns will therefore be used. The measurement model of environmental dynamics was, according to KLEER (1991, pp. 121-122) operationalized along the lines of expected

228

6 Construct operationalization

changes in both the own competition and the relationships with the customers as well as possible shifting perceptions of the customers on the importance of logistics. Following the argumentation of chapter 4.5.2.1, changes in the competitive landscape are measured through the competitive pressure in the customer’s industry (indicator 1). Changes in the relationships with customers result among other factors from concentration (indicator 2). The more powerful the customers, the more demanding they will be. Finally, the importance of logistics for the customers is assessed through indicators 3 and 4. As the importance of high logistics performance is growing and is increasingly viewed as a source of competitive advantage, the demand on the customer’s logistics processes continuously increases. Table 6-56. Indicators for the measurement of the construct environmental dynamics

Indicator 1 Indicator 2 Indicator 3 Indicator 4

Please indicate your level of agreement with the following statements about the present situation of your company The pressure of competition in our industry is very high. In our customers' industries we see strong tendencies of concentration. The importance of logistics in the eyes of our customers is increasing steadily. Our customers increasingly view a very good logistics performance as important, making it a competitive advantage for us.

While the four indicators presented in Table 6-56 seem to grasp several aspects of environmental dynamics as conceptualized by KLEER (1991, pp. 121-122), an exploratory factor analysis reveals that the indicators load on two distinct factors. One is comprised of indicators 3 and 4 and has an explained variance of 78.6% and a coefficient alpha of 0.88. While these values are very satisfactory and the factor is therefore well suited for a moderating analysis, its character changed somewhat from the original intention. Upon the linguistical analysis of the construct, it is thus renamed to importance of logistics for the customer, representing an importance facet of environmental dynamics on its own. The remaining two indicators 1 and 2 only exhibit an explained variance of 32.0% and a coefficient alpha of 0.49. This is not sufficient for the treatment of the factor as an own construct. For the further analysis of moderating effects on the basis of environmental dynamics, each indicator must therefore be considered independently. In this study, the focus will be on indicator 1 as it reflects one important aspect of the dynamics in the customer’s industry by focusing on the pressure of competition. The focus of the construct thus changes slightly away from the pure environmental dynamics and towards the intensity of competition.

6.6 Contingency factors

229

After the analysis of environmental complexity and dynamics, the related construct uncertainty will be operationalized. As argued in chapter 4.5.2.1, uncertainty is close to these two variables and yet not congruent as it is derived from transaction cost theory and reflects the insecurity the customer has towards the future. The operationalization of the indicators follows the assumptions of transaction cost theory. While a large number of scales exist that have previously operationalized uncertainty, none have done so in an explicit logistics context with the aim to treat it as a contingency variable. Therefore, a specific new operationalization was undertaken. Table 6-57. Indicators for the measurement of the construct uncertainty

Indicator 1 Indicator 2 Indicator 3 Indicator 4

Please indicate your level of agreement with the following statements about the present situation of your company We find it very hard to accurately predict the future behavior of our customers. Our customers sometimes exhibit opportunistic behavior. In our or our customers industries we presently see strong consolidation tendencies and insecurity. We find it very difficult to predict the future needs and requirements of our customers.

Uncertainty in the behavior of the customers is measured through the respective degrees of bounded rationality (indicator 1) and opportunistic behavior (indicator 2), even though the anticipation of opportunism by the own customer may facilitate the adaptation of the organization’s behavior and thus reduce the potentially detrimental effects. The uncertainty about the general environment is reflected in sudden changes in the competitive environment (indicator 3) and changes in the preference structure of the customers (indicator 4). The four indicators were tested in an exploratory factor analysis which only extracted one factor. While most of the adaptation measures showed satisfactory levels as displayed in Table 6-58, the value of F2/df at 5.513 and the RMSEA at 0.091 were not sufficient. In an effort to improve the quality of the measurement model, indicator 2 having the lowest indicator reliability at 0.34 was removed. This does not pose a problem, since the concept of opportunism is already a part of the logistics outsourcing performance model as a separate construct.

230

6 Construct operationalization

Table 6-58. Adaptation measures for the construct uncertainty (4 indicators) Information on the factor uncertainty (4 indicators) Coefficient alpha Explained variance

0.80 51.07% 5.513 0.96 0.99

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.95 0.99 0.091 0.80 0.51

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4

Item-to-total correlation 0.63 0.53 0.63 0.67

Indicator reliability 0.52 0.34 0.53 0.65

t-value of factor loadings 12.10 14.58 15.33

After the elimination of indicator 2 from the measurement model, all adaptation measures as presented in Table 6-59 are fully satisfactory. The measurement model can therefore be accepted without any further modifications. Table 6-59. Adaptation measures for the construct uncertainty (3 indicators) Information on the factor uncertainty (3 indicators) Coefficient alpha Explained variance

0.79 56.72% * * *

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.80 0.57

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated.

Information on the indicators Indicator 1 Indicator 3 Indicator 4

Item-to-total correlation 0.60 0.61 0.69

Indicator reliability 0.48 0.50 0.72

t-value of factor loadings 13.79 13.68

Before finishing this chapter, the final external contingency variable customer’s industry must be introduced. It is composed of the eight industries represented in this survey which were introduced in chapter 5.1.6 as well as “Others”. They are depicted in Table 6-60 and for the moderating analysis will be treated independently.

6.6 Contingency factors

231

Table 6-60. Industries of the participating firms Which industry does your business unit belong to? • Food, Beverage and Tobacco

• Chemicals and Plastics

• Automotive

• Retail

• Consumer Goods

• Healthcare

• Manufacturing Systems Construction

• Others

• Electrotechnology, Precision Engineering and Optics

6.6.2 Internal contingency variables As it has been argued in chapter 4.5.2.2, seven internal contingency variables are hypothesized to have moderating effects on the models analyzed in this study. Three of them, namely products, asset specificity, and process orientation, will be operationalized as constructs with multiple indicators. The remaining three, size of the customer’s organization, degree of centralization of logistics decisions and frequency, can have multiple indicators in some cases. However, these are not measurement models in the sense of chapter 5.2.1, but rather an accumulation of different facets which will be aggregated with different methods. The measurement model of products, which measures the logistical demands caused by the range of a firm’s product, was operationalized along the lines of the conceptualization proposed by KLEER (1991, pp. 122-123). As proposed in chapter 4.5.2.2, the indicators presented in Table 6-61 measure the range of the products produced (indicator 1), their corresponding material value (indicator 2), and the usability of logistics infrastructure for the products (indicator 3). Indicator 4 aims at evaluating whether the products offer the customer the opportunity to positively differentiate itself through logistics excellence. Together, these four indicators should give a broad overview of the customer’s product range from a logistics point of view. Table 6-61. Indicators for the measurement of the construct products

Indicator 1 Indicator 2 Indicator 3 Indicator 4

Please indicate your level of agreement with the following statements rabout the present situation of your company Our product range is very broad with a lot of different products. Our products are so valuable that they require specific logistics measures. Our product range is so broad that we need many different transport-, warehouse- and handling processes. The market offers a range of alternatives for our products so that we can positively differentiate ourselves through logistics excellence.

232

6 Construct operationalization

An exploratory factor analysis for the measurement model showed a disappointing explained variance of only 28.8% with a coefficient alpha of 0.59. While the latter is almost satisfactory, the former clearly indicates that the factor is not adequately operationalized. Obviously, one ore more indicators do not fit into the factor while other important facets might be missing. In order to improve the measurement model, indicator 4 with the lowest indicator reliability was eliminated. However, this only elevates the explained variance to 35.8% which is still not satisfactory. Further elimination of indicators, even up to only 2 remaining indicators, barely improves the explained variance. The highest value attainable with these indicators is 37.0% with indicators 1 and 4 which is still not acceptable. In the light of the complexity of the object product range, it was therefore decided to focus only on indicator 3 in the moderating analysis as it incorporates both the broadness of the product range and the consequences for the logistics processes. The size of the customer’s organization is a further contingency variable. With increasing size, the customer can be expected to have a higher number of productions sites, a broader product ranger and also a higher sales volume, all having implications for the designing of logistics outsourcing arrangements. While a large number of different ways exist to measure the size of an organization, for this study the indicator “What are the current yearly revenues of your business unit (in € Million/year)?” was selected. Not only is it understandable without the need for explanation, it usually is also readily available to the respondents which for other more complex measures might not be the case. The indicator was an open question. Only in the latter analysis, the information was aggregated into the six size segments introduced in chapter 5.1.6 spanning from “up to 50 Mio. €” to “more than 1 Bn. €”. Following KLEER (1991, p. 123), a final variable from the contingency approach introduced in chapter 4.5.2.2 is the degree of centralization of logistics decisions. It was argued that the higher the degree of centralization, the more accelerated is the decision making process and the higher is the relationship intensity. It was measured through the single indicator “The decision-making process on logistics issues is very centralized in our company” which is sufficient to grasp the situation. More complex is the measurement of the construct asset specificity which originates from transaction cost theory. From the large number of scales existing to measure the specificity of assets, the one proposed by JOSHI/STUMP (1999, p. 352) was selected and adapted to the logistics context. They propose to include the dimensions of processes and people which have been operationalized in the form of indicator 2, measuring whether the logistics processes set up by the LSP are particularly specific,

6.6 Contingency factors

233

and indicator 3, measuring if investments into the training of employees have been made that would not be retrievable if the relationship would be terminated. A third indicator was added to account for the particular logistics context with regard to the overall investment of the LSP. Indicator 1 measures whether this investment is usable for the LSP also in later logistics outsourcing arrangements or if it is so specific that it can only be used in this one situation and thereafter is worthless. Table 6-62. Indicators for the measurement of the construct asset specificity

Indicator 1

Please indicate your level of agreement with the following statements on the specific investments your LSP had to make for this relationship To supply us with logistics services, our LSP had to make significant investments (transportation equipment, warehouses, etc.) that it can only use for us.

Indicator 2

The business processes our LSP has set up for our cooperation are so uniquely fitted to our needs that he could not use them for another customer and its products.

Indicator 3

In order to supply us with logistics services, our LSP had to train its employees in such a particular way that they could not use these skills for other customers.

To test the measurement model, an exploratory factor analysis was conducted that extracted only one factor. However, a number of the reduced set of adaptation measures failed to meet the proposed threshold values, most notably the explained variance at 43.41% and the average variance extracted at 0.43. To improve the measurement model, indicator 1 with the lowest indicator reliability of 0.26 was eliminated. This does not threaten the content validity as the original scale of JOSHI/STUMP (1999, p. 352) is now being used. Its explained variance is 51.8% and the coefficient alpha 0.68. Table 6-63. Adaptation measures for the construct asset specificity (3 indicators) Information on the factor asset specificity (3 indicators) Coefficient alpha Explained variance

0.68 43.41% * * *

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

* * * 0.69 0.43

* At only 3 indicators, the measurment model has no degrees of freedom. This value can therefore not be calculated.

Information on the indicators Indicator 1 Indicator 2 Indicator 3

Item-to-total correlation 0.42 0.52 0.55

Indicator reliability 0.26 0.49 0.55

t-value of factor loadings 8.88 8.61

A second internal contingency variable from transaction cost theory as proposed in chapter 4.5.2.2 is frequency. There it was argued that the frequency of transactions has implications for the amortization of transaction-

234

6 Construct operationalization

specific investments, where customers with more frequent transactions will seek longer and deeper relationships with LSPs in order to reduce transaction costs. To capture this context, the indicator “The investments our LSP had to make in employees, processes, or assets will quickly amortize” was developed. The focus on the LSP is justified as in contract logistics, typically the LSP makes investments rather than the customer. The costs for the investments are then part of the price for the overall logistics service. A further construct identified in chapter 4.5.2.2 as an internal contingency variable is process orientation. While this concept was shown to have a positive influence on logistics performance by DEHLER (2001, pp. 220-226) with an extensive, five-dimensional, and indirect construct, this would go beyond the scope of this study. Instead, as also pointed out by DEHLER (2001, pp. 173-175), a direct measurement with only five indicators provides a very close approximation of the results the fivedimensional construct showed. This reduced measurement model includes all relevant aspects and concepts of process orientation. For this study, all five indicators were selected and only slightly linguistically adapted. They are presented in Table 6-64. Table 6-64. Indicators for the measurement of the construct process orientation

Indicator 1

Please indicate your level of agreement with the following statements about your internal logistics organization Our Business Unit has a very smooth, continuous, quick and largely failure-free flow of material and information.

Indicator 2 Indicator 3

Our Business Unit is managed in a very flow- or process-oriented way. Our Business Unit's degree of flow or process orientation is significantly higher than that of our competitors.

Indicator 4 Indicator 5

All business processes are very well coordinated. In our Business Unit a large number of individual interests exists that impede the direct achievement of all our objectives.

While an exploratory factor analysis only extracted one factor, the adaptation measures are slightly worse than those reported by DEHLER (2001, p. 174) and ENGELBRECHT (2004, p. 189) who also used the construct. Especially the F2/df (6.256) and the RMSEA (0.098) are substantially higher than the recommended values proposed in chapter 5.2.4.3.

6.6 Contingency factors

235

Table 6-65. Adaptation measures for the construct process orientation (5 indicators) Information on the factor process orientation (5 indicators) Coefficient alpha Explained variance

0.84 56.40% 6.256 0.96 0.98

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.93 0.98 0.098 0.85 0.53

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 4 Indicator 5

Item-to-total correlation 0.70 0.76 0.57 0.80 0.48

Indicator reliability 0.64 0.76 0.40 0.75 0.27

t-value of factor loadings 22.47 15.33 22.41 12.08

As Table 6-65 indicates, indicator 5 would be a possible starting point to improve the measurement model as it only has an indicator reliability of 0.27. However, this low reliability is not completely unexpected as it was reverse coded. The main driver behind the high F2/df is indicator 4, whose error term shows very high correlations with the error terms of indicators 3 and 5. Since in the measurement models used by both DEHLER (2001, p. 174) and ENGELBRECHT (2004, p. 198), indicator 5 is consistently part of the final measurement models, indicator 4 is eliminated from the measurement model as its character is also contained in several of the other indicators. Table 6-66. Adaptation measures for the construct process orientation (4 indicators) Information on the factor process orientation (4 indicators) Coefficient alpha Explained variance

0.79 51.20% 1.945 0.99 1.00

F²/df TLI GFI

AGFI CFI RMSEA Composite reliability Average variance extracted

0.98 1.00 0.042 0.78 0.48

Information on the indicators Indicator 1 Indicator 2 Indicator 3 Indicator 5

Item-to-total correlation 0.66 0.72 0.52 0.45

Indicator reliability 0.65 0.82 0.35 0.23

t-value of factor loadings 18.20 13.83 10.95

After the elimination, the four indicator measurement model exhibits satisfactory adaptation measures in almost all dimensions. Only the AVE is

236

6 Construct operationalization

slightly lower than recommended at 0.48 and the indicator reliability of indicators 3 and 5 are below the recommended value of 0.4. In the light of the other satisfactory measures, however, the measurement model is accepted without any further modifications.

7 Structural models

This chapter will discuss the empirical findings on the effects of the antecedents developed in chapter 4.2 on the two dimensions of logistics outsourcing performance. Furthermore, it will examine the influence of logistics outsourcing performance on logistics performance and subsequently the effect of logistics performance on firm performance as suggested in chapter 4.4. Finally, the moderating effects of different contingency variables introduced in chapter 4.5 will be analyzed.

7.1 Antecedents and dimensions of logistics outsourcing performance In the following three chapters, first the basic model derived from the empirical analyses of the hypotheses put forward in chapter 4.3.2 on the antecedents and dimensions of logistics outsourcing performance will be presented. Due to insufficient model fit, a simplified model will be developed in chapter 7.1.2 which will be discussed in detail in chapter 7.1.3. 7.1.1 Presentation of the basic model When the relationships between the antecedents and the two dimensions of logistics outsourcing performance are modeled as suggested by the hypotheses in Table 4-1, the model presented in Figure 7-1 results. It consists of the 9 antecedents proposed in chapter 4.2 and the two dimensions of logistics outsourcing performance. As argued in chapter 6.5.1, the only antecedent not to be part of the basic model is functional conflict due to the lack of its discriminant validity.

238

7 Structural models

Commitment

Cooperation

Goal Achievement

Proactive Improvement

Goal Exceedance

Trust Involvement

Communication

Shared Values

Openness

Opportunism

Fig. 7-1. Basic model

This model shows insufficient model fit when using the maximumlikelihood estimation function. While the F2/df, TLI, CFI, and RMSEA all show values well within the desired range or above the respective threshold, both the GFI and the AGFI are well below the desired value of 0.9. However, this must not automatically require the model to be rejected, especially since an AGFI of 0.8 is sometimes proposed as sufficient (see chapter 5.2.4.2). As several authors point out (FORNELL/LARCKER 1981; BAGOZZI/YI 1988; HOMBURG/BAUMGARTNER 1995a), a model can also be accepted if the majority of fit indices show good adaptation measures and only a few are below the required threshold. This is the case with the Basic Model. However, in such cases, special attention should be paid to the CFI as the single most important index as it accounts for sample size (BENTLER 1990; BYRNE 1994). Yet, since GFI and AGFI at 0.87 and 0.85 are substantially below the required 0.9, it must be assumed that the model contains room for improvement. Table 7-1. Adaptation measures of the basic model Model Basic Model

F²/df 2.264

TLI 0.94

Adaptation measures GFI AGFI 0.87 0.86

CFI 0.95

RMSEA 0.048

The insufficient model fit does not result from “forgetting” causal linkages between the factors that compose the model. The statistics program AMOS used to calculate the model does not identify any causal linkages in its “Modification Indices” section which could be included into the model and would at the same time be theoretically justified. Instead, a number of

7.1 Antecedents and dimensions of logistics outsourcing performance

239

correlations between error terms of indicators or cross-loadings between factors and the indicators of other factors, are identified. This is a standard finding especially in complex causal models. The insufficient model fit can therefore be attributed to the complexity of the model in comparison to the empirical database. BENTLER/CHOU (1987, p. 91) and BAGOZZI/YI (1988, p. 82) suggest that the ratio between the sample size and the number of distinct parameters being estimated should be at least 5. For this model, the number of parameters estimated is 104, the sample size 549, giving a ratio of 5.28. While compared to other studies this is a relatively high number, even slightly exceeding the threshold value, is also emphasizes the model’s complexity with respect to the sample size. Therefore, model complexity must be reduced following the argumentation of chapter 5.2.5.2 which introduced the basics of model design and modification by eliminating single factors in order to receive a better model fit. This will be done in the following chapter. 7.1.2 Development of a simplified model The construct of involvement poses a valid starting point for the simplification of the structural model. As it was pointed out in chapter 6.1.7, it is measured by a single indicator only since it showed in the development of the measurement model that the scale originally developed was not satisfactory. In the calculation of the 9-factor basic model, the modification indices as stated by the statistics program AMOS suggest that some reasons for the insufficient model fit can be attributed directly to the construct of involvement. The modification index for including a relationship between involvement and shared values is 57.5, for including involvement and openness it is 68.8. However, since no theoretical foundations exist that would suggest an inclusion of these two causal linkages, they may not be included in the model. When analyzing the modification indices between the error term of involvement and the error terms of the other 8 antecedents of logistics outsourcing performance, several further high correlations are found. Altogether, it is apparent that the construct of involvement is directly responsible for significant amounts of complexity in the 9-factor basic model. However, to reduce model complexity, only those factors may be eliminated that have no or very little influence on the explanatory power of the model, e.g. the squared multiple correlation R2 of both goal achievement and goal exceedance. As Table 7-2 indicates, this is not a problem in this case. The elimination of the factor involvement, leading to a simplified 8factor model, does not only exhibit better adaptation measures in all di-

240

7 Structural models

mensions, it actually improves the explanatory value of the model by increasing the R2 of goal achievement from 60.9% to 63.4% and of goal exceedance from 34.5% to 35.1%. This is a clear sign for improved model fit and quality through a reduction of model complexity. An additional comparison of the two models’ CAIC and ECVI furthermore show substantially lower values for both indices in the simplified, 8-factor model, again suggesting a better model fit for the simplified model. Table 7-2. Fit comparison of the basic and the simplified model Adaptation measures Model Basic Model 8-Factor Model



F²/df

TLI

GFI

AGFI

CFI RMSEA CAIC

ECVI

Goal Achievement

Goal Exceedance

2,264

0.94

0.87

0.86

0.95

0.048

2290.4

3,172

60.9%

34.5%

2,062

0.95

0.89

0.87

0.96

0.044

2052.4

2,777

63.4%

35.1%

The adaptation measures compared to the basic model presented in the previous chapter are almost all very satisfactory. While the model is still complex (the ratio between the sample size and the number of distinct parameters being estimated now is 5.49), GFI and AGFI have reached levels that at 0.89 and 0.87 are sufficiently close to the desired value of 0.9 that in combination with the other indices, especially the CFI at 0.96 and the F²/df at 2.062, a sound model fit is established. Summing up the model comparison, it is evident that the overall model quality both with regard to the adaptation measures and the explanatory value, has increased through the reduction of complexity and has resulted in a very satisfactory model of logistics outsourcing performance. This may be interpreted as an indicator that the construct of involvement is not an elementary part of exchange relationships and their performance outcomes as it does not provide additional explanatory power to the model and at the same time is statistically showing – at least in its current operationalized form – issues with several other constructs. The 8-factor model which results from the above discussion is displayed in Figure 7-2 and will be discussed in detail in the following chapter.

7.1 Antecedents and dimensions of logistics outsourcing performance

241

H22 Commitment

H13

H10

H21 H20 H27

H25

H26 Openness

H23

H3

Communication

H24

H18

H8

H19

Goal Achievement

H9a

H6

H4

Shared Values

H1a Cooperation

H11

Trust

H1b

H7 Proactive Improvement

H9b

Goal Exceedance

Opportunism

Fig. 7-2. Simplified 8-factor model

7.1.3 Discussion of the final simplified model After the reduction of the model to 8 antecedents directly and indirectly influencing the two dimensions of logistics outsourcing performance, the model is taking on a distinct, three level structure that can be seen in Figure 7-2. Five antecedents, namely shared values, trust, commitment, openness, and opportunism, form a block of behavioral and attitudinal variables that only indirectly, namely through their direct effects on the two action variables cooperation and proactive improvement, affect the two dimensions of logistics outsourcing performance, goal achievement and goal exceedance. The factor communication takes on a mediating position between the behavioral/attitudinal variables and the action variables as proposed in chapter 4.3.1.2.

242

7 Structural models 0.15**** Commitment 0.22****

R2: 63.4% 0.77****

0.67****

0.15**

0.42****

Cooperation

Trust

Goal Achievement

n.s.

0.17****

0.24****

0,15****

0.36****

0.80**** 0.23****

Shared Values 0.72****

Communication

0.69**** -0.19*

Openness

-0.28****

n.s.

-0.27**** -0.36****

0.33**** Proactive Improvement

Goal Exceedance 0.32****

R2: 35.1% Opportunism

0.65**** n.s. * ** *** ****

Standardized regression weight with significance level Not significant Significant on 10%-level Significant on 5%-level Significant on 1%-level Significant on 0.1%-level

Fig. 7-3. Simplified 8-factor model with standardized parameter coefficients

In chapter 4.3.2, 31 hypotheses on causal linkages between different factors were originally proposed. Due to the lack of discriminant validity, functional conflict was removed from the model. Additionally, the factor involvement was eliminated to reduce model complexity. Consequently, all hypotheses involving these two factors could not be tested. This involved 9 hypotheses, namely H2, H5, H12, H14a-b, H15-17, and H28. The remaining hypotheses were tested and will be discussed in the following. The first two hypotheses H1a and H1b examine the direct influence that cooperation exerts on goal achievement and goal exceedance, respectively. Model results indicate strong support for both hypotheses, demonstrating that cooperation directly influences both of the performance dimensions in logistics outsourcing arrangements. For the effect of cooperation on goal achievement the path coefficient is 0.77, for cooperation on goal exceedance it is 0.33. Both paths are positive and significant at the 0.1%-level, suggesting very strong relationships. Cooperation therefore can be viewed as a cornerstone of successful logistics outsourcing arrangements. Two further hypotheses, H9a and H9b, imply that proactive improvement through the LSP exerts similar, positive influences on both goal achievement and goal exceedance. The model results provide mixed support for the two hypotheses. While the path coefficient for the relationship between

7.1 Antecedents and dimensions of logistics outsourcing performance

243

proactive improvement and goal exceedance is 0.32 and significant at the 0.1%-level, thereby supporting H9b, the path from proactive improvement to goal achievement is not significant. H9a must therefore be rejected. Apparently, proactive improvement of the LSP has a particularly strong impact on exceeding the goals ex-ante agreed upon, while it has no significant influence on the dimension of goal achievement. Therefore, proactive improvement is a facilitator of exceptional performance while it seems unnecessary to merely meet the expectations of the customer. After having discussed the direct effects of action variables on the two dimensions of logistics outsourcing performance, now the effects of the remaining six antecedents will be analyzed that have indirect effects only. In hypothesis H3 it was supposed that cooperation has a direct positive effect on the proactive improvement of the LSP. This finds support in the high path coefficient of 0.80 which is significant at the 0.1%-level. This means that cooperation not only is an important facilitator of an increased proactive improvement through the LSP, but that it influences the outsourcing performance outcomes both directly and indirectly. Therefore, total effects can be calculated for cooperation which at 0.80 for goal achievement and 0.59 for goal exceedance are remarkably stronger than the isolated direct effects of 0.77 and 0.33. Cooperation and proactive improvement are hypothesized to have the common antecedent communication. The relationship between communication and cooperation (H6) finds support given a standardized parameter value of 0.24, which is significant again at the 0.1%-level of significance. Recall that this is the relationship that illustrated very strong correlation in the measurement model analysis performed in chapter 6.5.1. Hypothesis H7 however, does not find support through the empirical analysis. While the path for the effect of communication on proactive improvement is significant at the 10%-level, it is negative at a value of -0.19. This indicates that communicating and exchanging information alone do not lead to the increase of the proactive improvement of the LSP, therefore requiring H7 to be rejected. However, communication can in the long run have a positive effect on proactive improvement through the mediation of cooperation, trust, opportunism, and commitment. Its total effect is positive at 0.16, suggesting that communication leads to higher levels of cooperation which then in turn drives proactive improvement. Communication was further hypothesized to have effects on trust (H4) and opportunism (H8). Both hypotheses find support in the model results. The path coefficient from communication to trust is 0.23 and significant at the 0.1%-level, indicating that increased communication works trustbuilding as originally hypothesized. Also, communication serves as a means to decrease opportunistic behavior of the LSP as argued in chapter

244

7 Structural models

4.3.1.2. This is supported by finding the predicted negative standardized parameter value of -0.28, again significant at the 0.1%-level. After having discussed the effects of the mediating variable communication, the effects of the antecedents that form the block of behavioral/attitudinal variables on communication, cooperation, and proactive improvement as well as direct effects on each other can be examined. A valuable starting point is provided by the factor trust, which is an important building block of relationship models (WILSON 1995, p. 337). As hypothesized in H10 and H11, trust is supposed to have a direct positive effect on both relationship commitment and cooperation. The path coefficient from trust to relationship commitment is at 0.67 and 0.1%-level of significance very strong and supports H10. Evidently, as proposed by the commitment-trust theory, parties in relationships that are characterized by high levels of trust will seek to commit themselves to such relationships. Hypothesis H11, proposing a positive effect of trust on cooperation also finds support in the model. The standardized parameter value is 0.24 and again significant at the 0.1%-level. This further confirms the proposition that once trust is established, the parties involved learn that cooperative efforts lead to results that could not be achieved if both parties worked solely in their own interest. Analyzing the relationship between relationship commitment and cooperation, support is found for hypothesis H13, which proposed a direct positive effect induced by relationship commitment. The path coefficient is 0.22 at the 0.1%-level of significance. Apparently, parties that are committed to the relationship will increasingly cooperate in an effort to make the relationship work. Opportunism is another important factor in the relationship between customers and LSPs. It was hypothesized that opportunism would have negative effects on both trust (H18) and cooperation (H19), thereby making it a detrimental factor for logistics outsourcing arrangements. These hypotheses both find support in the data. The standardized path coefficient for the relationship between opportunism and trust is as expected negative at -0.27 at a level of significance of 0.1%. This indicates that if the customer believes the LSP to engage in opportunistic behavior, this perception will lead to decreased levels of trust with all following negative consequences, such as lower levels of relationship commitment and cooperation. The hypothesis on the direct negative effect of opportunism on cooperation as proposed in H19, however, did not find support. The path coefficient is not significant. This suggests that opportunism alone is not the main driver of reduced cooperation. This is rather evoked through the indirect detrimental effects of opportunism which have been discussed above.

7.1 Antecedents and dimensions of logistics outsourcing performance

245

Looking at the total effect of opportunism on cooperation, the effect of 0.20 therefore does not come as a surprise. A very important building block of the logistics outsourcing performance model is the factor shared values. Altogether, six hypotheses were formulated on its direct effects. All of them find support in the model. Hypothesis H20 proposed a positive effect of shared values on trust. The standardized parameter value for this linkage is 0.17 at the 0.1%-level of significance, suggesting that the existence of shared values will contribute to building trust in the LSP. A similar argumentation holds for the relationship between shared values and relationship commitment (H21). The path coefficient of 0.15, significant at the 5%-level, indicates that with higher levels of shared values, the customer will be increasingly more committed to the outsourcing arrangement, leading in turn to increased cooperative behavior. This relationship was subject to an own hypothesis. H22 proposed a direct effect on cooperation which is supported by a standardized parameter value of 0.15 at a 0.1%-level of significance. Apparently, higher levels of shared values and the accompanying understanding of each others motivations, needs, and expectations, leads customers and LSPs to work more closely together to achieve mutual goals. Similarly, higher levels of shared values seem to decrease the tendency on the part of the LSP for opportunistic behavior. Assumed in H23, this hypothesis finds support in a strong negative path coefficient of -0.36 at again a 0.1%-level of significance. This shows that increasing levels of shared values in a logistics outsourcing arrangement lead to lower levels of opportunistic behavior and thereby to better results. In hypotheses H24 and H25, it was furthermore proposed that shared values also facilitate higher levels of communication and openness. H24 is supported by a path coefficient of 0.23 at the 0.1%level of significance. This provides the insight that LSP and customer will find it easier to formally and informally exchange meaningful and timely information because they can better relate to their partners in the relationship. Similarly, openness is facilitated by higher levels of shared values. The corresponding hypothesis H25 boasts a path coefficient of 0.72 and is significant at the 0.1%-level. This strongly supports the hypothesis that if customer and LSP share a similar set of values, they will find it less difficult to openly and informally discuss current issues and problems since the fear of being rejected by the other party is substantially decreased. Finally, two hypotheses concerning the effect of the factor openness will be discussed. Hypothesis H26 proposed a positive direct influence of openness on communication. The standardized parameter value of 0.69 significant at the 0.1%-level provides strong support for this hypothesis. Openness serves as a facilitator through its facets informality, spontaneity, and the open exchange of information and ideas for the efficient and effective

246

7 Structural models

flow of information as measured by the construct of communication. Hypothesis H27, proposing a direct effect of openness on trust, also finds strong support. The standardized parameter value is 0.36, significant at the 0.1%-level. This indicates that openness serves as a facilitator to create trust between customer and LSP as both parties find it easier to trust each other through the open exchange of information and ideas, thus creating an atmosphere of honesty. Summing up the model discussion, 19 out of the original 31 hypotheses found support. 3 were rejected while another 9 could not be tested due to a preceding elimination of the construct functional conflict and involvement from the model. An overview of the hypotheses is presented in Table 7-3. Table 7-3. Hypotheses for the logistics outsourcing performance model Hypotheses

Supported

H 1a

Cooperation has a positive effect on goal achievement.

9

H 1b

Cooperation has a positive effect on goal exceedance.

9

H2

Cooperation has a positive effect on functional conflict.

H3

Cooperation has a positive effect on proactive improvement.

9

H4

Communication has a positive effect on trust.

9

H5

Communication has a positive effect on functional conflict.

H6

Communication has a positive effect on cooperation.

H7

Communication has a positive effect on proactive improvement.

H8

Communication has a negative effect on opportunism.

Rejected

Not Tested

8

8 9 8 9 8

H 9a

Proactive improvement of the LSP has a positive effect on goal achievement.

H 9b

Proactive improvement of the LSP has a positive effect on goal exceedance.

9

H 10

Trust has a positive effect on relationship commitment.

9

H 11

Trust has a positive effect on cooperation.

9

H 12

Trust has a positive effect on functional conflict.

H 13

Relationship commitment has a positive effect on cooperation.

8 9

H 14a

Functional conflict has a positive effect on goal achievement.

8

H 14b

Functional conflict has a positive effect on goal exceedance.

8

H 15

The involvement of the LSP has a positive effect on cooperation .

8

H 16

The involvement of the LSP has a positive effect on communication.

8

H 17

The involvement of the LSP has a positive effect on proactive improvement .

H 18

Opportunistic behavior has a negative effect on trust .

8 9 8

H 19

Opportunistic behavior has a negative effect on cooperation.

H 20

Shared values have a positive effect on trust.

9

H 21

Shared values have a positive effect on relationship commitment .

9

H 22

Shared values have a positive effect on cooperation.

9

H 23

Shared values have a negative effect on opportunistic behavior .

9

H 24

Shared values have a positive effect on communication.

9

H 25

Shared values have a positive effect on openness.

9

H 26

Openness has a positive effect on communication.

9

H 27

Openness has a positive effect on trust.

9

H 28

Openness has a positive effect on functional conflict.

8

7.2 Effects of logistics outsourcing performance

247

After the model of logistics outsourcing performance and its antecedents have been discussed in great detail, it must be mentioned that beyond the direct effects presented in Figure 7-3, a number of variables also have indirect effects, being mediated by other variables. Combined, the two effects compose the total effect. These will not be further discussed at length. However, Table 7-4 provides an insight into these effects. The most interesting finding certainly is that except for the factor opportunism, which has a negative total effect on every other variable in the model, all other antecedents have positive total effects on both goal achievement and goal exceedance. This again is strong support for the importance of all 8 antecedents for logistics outsourcing performance. Table 7-4. Total effects in the logistics outsourcing performance model Factor (1) Shared Values (2) Openness (3) Communication (4) Opportunism (5) Trust (6) Commitment (7) Cooperation (8) Proactive Improvement (9) Goal Exceedance (10) Goal Achievement

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

0.72 0.73 -0.57 0.75 0.65 0.81 0.52 0.44 0.65

0.69 -0.19 0.57 0.39 0.50 0.28 0.26 0.40

-0.28 0.31 0.21 0.42 0.16 0.19 0.34

-0.27 -0.18 -0.20 -0.16 -0.12 -0.16

0.67 0.57 0.46 0.34 0.46

0.22 0.18 0.13 0.18

0.80 0.59 0.80

0.32 0.04

0.00

-

7.2 Effects of logistics outsourcing performance The previous chapter has established the model of logistics outsourcing performance and its antecedents. The results suggest the strategic importance for the customers to not only agree to an outsourcing decision, but to also actively select the adequate LSP and to manage the relationship. As is was pointed out in chapter 4.4, however, logistics outsourcing can only be considered a truly strategic issue of utmost importance to customers if logistics outsourcing performance has measurable effects on both logistics and firm performance. To analyze this question, chapter 7.2.1 will investigate the effects of logistics outsourcing performance on logistics performance while chapter 7.2.2 will analyze the effect of logistics performance on firm performance.

248

7 Structural models

7.2.1 Logistics outsourcing performance and logistics performance 7.2.1.1 Presentation of the model

Following the hypotheses developed in chapter 4.4.1.2, both goal achievement and goal exceedance are supposed to have a direct effect on the two dimensions of logistics performance, the level of logistics services and the level of logistics costs. The resulting four hypotheses are displayed in Figure 7-4. Goal Achievement

Level of Logistics Services

H29 H30

H31

Goal Exceedance

Level of Logistics Costs

H32

Fig. 7-4. Hypotheses on the logistics performance model

Upon testing, the model showed very satisfactory adaptation measures. The adjusted F², at 2.540 is well within the range generally deemed acceptable. The CFI as the single most important indicator is very strong at 0.98, while also the other indicators are well above the required threshold value of 0.9 as can be seen in Table 7-5. The model can therefore be accepted without any further modification. Table 7-5. Adaptation measures of the logistics performance model Adaptation measures Model Logistics Performance



F²/df

TLI

GFI

AGFI

CFI

RMSEA

Logistics Services

Logistics Costs

2.540

0.97

0.96

0.94

0.98

0.053

10.8%

7.3%

Before the individual results of the model will be discussed in the next chapter, it must be mentioned, however, that within the model the object of analysis changes. While for the two dimensions of logistics outsourcing performance only the relationship of the customer to its single most important LSP is analyzed, the two dimensions of logistics performance investigate the entire logistics performance of the customer. Aside from the most important relationship to one LSP, this also includes all other relationships with third parties as well as all logistics processes still produced in-house.

7.2 Effects of logistics outsourcing performance

249

The model must therefore be viewed as a partial model only which can be expected to deliver significantly lower squared multiple correlations (R2) than the logistics outsourcing performance model. The results will be discussed in the following chapter. 7.2.1.2 Discussion of the model

The empirical analysis of the effect of logistics outsourcing performance on logistics performance provides some very interesting insights at a very high statistical level. All four hypotheses that were proposed in chapter 4.4.1.2 find support as presented in Figure 7-5. R2: 10.8% Goal Achievement

Level of Logistics Services

0.18**** 0.21****

0.20****

R2: 7.3% Goal Exceedance

0.10*

0.65**** n.s. * ** *** ****

Level of Logistics Costs

Standardized regression weight with significance level Not significant Significant on 10%-level Significant on 5%-level Significant on 1%-level Significant on 0.1%-level

Fig. 7-5. Logistics performance model

Hypothesis H29 examines the direct influence that goal achievement exerts on the level of logistics services. Model results indicate support for this hypothesis with a standardized parameter value of 0.18 at the 0.1%-level of significance. This finding demonstrates that the achievement of the goals that were specified between the customer and the LSP has a direct and lasting positive influence on the firms’ logistics performance. This confirms the rather intuitive insight that the specifications of the contract between the parties will set the limit for the service performance of the LSP at a level where it will be beneficial for the customer. However, recalling the change in the object of analysis, the strength of the path at 0.18 is remarkable as it displays the high relevance of the most important LSP for the customer’s overall level of logistics services.

250

7 Structural models

The next hypothesis H30 implies that goal achievement has a direct positive effect also on the level of logistics costs. The model provides support also for this hypothesis with a path coefficient of 0.21 with a level of significance of 0.1%. This relationship, which is even slightly stronger than the effect of goal achievement on the level of logistics services, indicates that when the goals and expectations of the customer concerning the outsourcing arrangement are met, the level of logistics costs can be decreased significantly. Summing up the above discussion on hypotheses H29 and H30, it is clear that goal achievement is a significant driver of both dimensions of logistics performance, making this outsourcing relationship an important strategic issue. After having pointed out the importance of goal achievement, goal exceedance as the second dimension is analyzed. Consistent with theory, hypothesis H32 finds support, proposing that goal exceedance has a positive effect on the level of logistics costs. The path coefficient of 0.10 is significant at the 10% level. This finding suggests that an over-performance on the side of the LSP does indeed lead to a reduction of the logistics costs. However, compared to the other effects, the influence is not particularly strong. This may be attributable to the fact that in current logistics outsourcing arrangements the price for the service, and therewith the costs for the customer, are exactly specified in the contract. Improvements reduce the price for the customer and simultaneously the revenue of the LSP. Consequently, the LSP has no incentive to proactively further reduce the costs, unless it is directly motivated, e.g. by being allowed to keep a share of the reduced costs. While this practice is becoming increasingly popular, it by no means has become an established industry standard. The importance of goal exceedance, which as pointed out above is primarily driven by the proactive improvement of the LSP, is further demonstrated by the findings on hypothesis H31. The standardized parameter value for the effect of goal exceedance on the level of logistics services is 0.20 at the 0.1%-level of significance. This indicates that by exceeding the goals in the single most important logistics outsourcing arrangement, the logistics service level of the customer is significantly increased. Apparently, the surpassing of the expectations of the customer though the LSP provides higher levels of logistics quality and process flexibility which in turn significantly influence the customer’s overall logistics service quality.

7.2 Effects of logistics outsourcing performance

251

Table 7-6. Hypotheses for the logistics performance model Hypotheses

Supported

H 29

Goal achievement has a positive effect on the level of logistics services.

9

H 30

Goal achievement has a positive effect on the level of logistics costs.

9

H 31

Goal exceedance has a positive effect on the level of logistics services.

9

H 32

Goal exceedance has a positive effect on the level of logistics costs.

9

Rejected

Not Tested

All four hypotheses as displayed in Table 7-6 find empirical support. It can therefore be observed that goal exceedance is an important facet of logistics outsourcing performance that together with the goal achievement is a major driver of logistics performance. This should be even more visible in further empirical research, where beyond the single most important outsourcing relationship a larger number of relationships could be analyzed. For the moment, just 10.8% of the variance of the level of logistics services and 7.3% of the level of logistics costs can be explained through the two dimensions of logistics outsourcing performance. Since this is only a partial model, the effect of logistics outsourcing performance over all relationships can legitimately be assumed to be significantly higher, indicating an even higher strategic importance of the issue than the above presented results originally suggest. 7.2.2 Logistics performance and firm performance After the previous chapter has established the importance of logistics outsourcing for the overall logistics performance, this chapter will analyze the connection between the latter and the individual firm performance. Following the argumentation from chapter 4.4.2.2, both dimensions of logistics performance are proposed to have direct effects on all three dimensions of firm performance. 7.2.2.1 Presentation of the model

Following the hypotheses developed earlier, both the level of logistics services and the level of logistics costs are supposed to exert positive effects on the three dimensions of firm performance, adaptiveness, market performance and financial performance. Additional, positive direct effects were proposed from adaptiveness on market performance as well as from the latter to financial performance. The resulting eight hypotheses are displayed in Figure 7-6.

252

7 Structural models

Adaptiveness

H33 Level of Logistics Services

H36

H39

H34 Market Performance

H37 Level of Logistics Costs

H40 H35

H38

Financial Performance

Fig. 7-6. Hypotheses on the firm performance model

As Table 7-7 indicates, the model shows satisfactory adaptation measures in almost all dimensions. The CFI as the single most important index is particularly strong at 0.95, alongside with TLI, GFI and RMSEA. The AGFI of 0.89 just barely fails to meet the recommended threshold value of 0.9. However, since this index is extremely demanding on the one hand and the difference to the threshold extremely little, it is of no further concern. The adjusted F², at 3.586 is well below the highest minimum value proposed in literature of 5.0 (STANK/GOLDSBY/VICKERY/SAVITSKIE 2003) and only slightly above the threshold value suggested in chapter 5.2.4.3. Since the remaining indicators suggest sound model fit and the adjusted F² is known for its tendency to reject models on grounds of complexity, the model is nevertheless accepted without any further modifications. Table 7-7. Adaptation measures of the firm performance model Adaptation measures Model Firm Performance



F²/df

TLI

GFI

AGFI

CFI

RMSEA

Adaptiveness

Market Performance

Financial Performance

3.586

0.94

0.92

0.89

0.95

0.069

12.9%

50.4%

22.3%

Comparing the model fit to that found by DEHLER (2001) and ENGELBRECHT (2004), the difference in some of the fit criteria, most notably the higher F²/df in this study, might surprise. However, this is largely due to a modification in the model discussed already above. Other than assumed by both DEHLER (2001) and ENGELBRECHT (2004), no correlation was introduced between the two dimensions of logistics performance. This creates a very high modification index between the two factors which is at

7.2 Effects of logistics outsourcing performance

253

large responsible for the high values of F² in this model. However, since no theoretical evidence would suggest a correlation between the two factors, it will be abstained from including a correlation and from modifying the model in this or any other direction. 7.2.2.2 Discussion of the model

The results of the full model analysis provide some very interesting insights into the effect of logistics performance on firm performance. In total, seven out of the eight hypotheses developed in chapter 4.4.2.2 found support while only one had to be rejected. The findings presented in Figure 7-7 substantially advance the models proposed by DEHLER (2001, p. 241) and ENGELBRECHT (2004, p. 254) by providing empirical evidence for several causal linkages their empirical data failed to support. Hypothesis H33 examines the direct influence the level of logistics services exerts on the adaptiveness of the customer. This hypothesis is supported by the empirical findings with a standardized parameter value of 0.33 at the 0.1%-level of significance. This suggests that through improving their logistics service capabilities, firms indeed enhance flexibility to react to changes in the market and in particular sharpen their ability to answer to changes in product volumes demanded by the customers. The positive effect of the level of logistics services is also visible in the findings on hypothesis H34 which analyzed its effects on market performance. This relationship was found to be particularly strong with a path coefficient of 0.55 again at the 0.1%-level of significance. Apparently, market performance, which depends very strongly on the satisfaction of the customer with a purchased product or service, is driven not only by the primary capability of the firm to produce the product or service, but to a large degree also by its secondary capability to properly distribute it. Beyond these two hypotheses, hypothesis H35 also proposed a positive effect of the level of logistics services on the financial performance of the firm. However, this relationship was found not to be significant. Evidently, higher logistics performance levels today do not enable firms to increase prices to an extent that would be reflected in a directly increased financial performance. This, however, is an indicator for the supposition that logistics so far is viewed also by the customers of the customer as a hygiene factor which has to be fulfilled in an acceptable manner but gives little room for differentiation that would translate into substantially increased prices. Therefore, apart from the direct effects the level of logistics services has on the market performance, logistics do not offer direct effects on the financial performance.

254

7 Structural models R2: 12.9% Adaptiveness

0.33**** Level of Logistics Services

0.14***

0.24**** R2: 50.4%

0.55**** 0.21**** Level of Logistics Costs

Market Performance

0.43**** 0.15****

n.s.

R2: 22.3% Financial Performance

0.65**** n.s. * ** *** ****

Standardized regression weight with significance level Not significant Significant on 10%-level Significant on 5%-level Significant on 1%-level Significant on 0.1%-level

Fig. 7-7. Firm performance model

Aside from the effect of the level of logistics services, hypotheses were also developed for the level of logistics costs. Hypothesis H36 proposed a direct positive effect from the level of logistics costs on the adaptiveness of the firm. This is supported by the empirical finding of a path coefficient of 0.14 which is significant at the 1%-level. As argued in chapter 4.4.2.2, higher levels of logistics costs lead to less flexibility and short term maneuverability of the firm. These restrictions can obviously be reduced by lowering the logistics costs, leading to higher flexibility and therewith to more adaptiveness on the part of the customer. The level of logistics costs also increases a firm’s market performance. Support for the corresponding hypothesis H37 was found in a standardized parameter coefficient of 0.21 between the two constructs, significant at the 0.1%-level. This suggests that through lowering logistics costs, firms are indeed enabled to substantially lower their overall product costs, highlighting the high impact of logistics costs on total costs. Lower product costs directly enable prices reductions, leading to higher customer satisfaction, which is the main prerequisite for higher market performance. Lower levels of logistics costs finally also lead to a higher financial performance of the firm. This was proposed in hypothesis H38 and supported by the finding of a path coefficient of 0.15 at the 0.1%-level of significance. Evidently, the supposed direct mathematical relationship between lower logistics costs and lower total costs exists. Since financial performance is a function of sales and costs, it increases by the same amount as the

7.2 Effects of logistics outsourcing performance

255

logistics costs are overall reduced. The fact that the effect is not particularly strong was to be expected as for the majority of products the share of the logistics costs of the total product cost is substantial and yet other costs like raw material and production costs are often significantly higher. Beyond the hypotheses discussed above, two further hypotheses were developed to grasp the relationship between adaptiveness, market performance and financial performance. Hypothesis H39 examined the effect of adaptiveness on market performance. Support for the positive effect was found as reflected in a path coefficient of 0.24 at the 0.1%-level of significance. Apparently, higher levels of adaptiveness enable firms to adapt their products or services to altered customers needs and thereby to react flexibly to market developments. Therefore, adaptiveness enables a better addressing of customer needs and thereby the increase of customer satisfaction, which in turn is an important prerequisite for market performance. Finally, hypothesis H40 proposed a positive effect of market performance on financial performance. This hypothesis is supported by the empirical findings. The standardized parameter value is very strong with 0.43 at the 0.1%-level of significance. Evidently, firms with higher market performance indeed enjoy higher customer loyalty, leading to decreased customer acquisition costs, to higher customer profitability due to longer and stronger relationships, and to a higher tolerance towards price increases. Furthermore, strong growth and higher market shares of course also positively influence the economical performance of the firm. After having tested the eight hypotheses, it can be assessed that both dimensions of logistics performance have measurable, direct, and strong effects on the three dimensions of firm performance. An overview of the hypotheses can be found in Table 7-8. Table 7-8. Hypotheses for the firm performance model Hypotheses

Supported

H 33

The level of logistics services has a positive effect on adaptiveness.

9

H 34

The level of logistics services has a positive effect on market performance.

9

H 35

The level of logistics services has a positive effect on financial performance.

H 36

The level of logistics costs has a positive effect on adaptiveness.

9

H 37

The level of logistics costs has a positive effect on market performance.

9

H 38

The level of logistics costs has a positive effect on financial performance.

9

H 39

Adaptiveness has a positive effect on market performance.

9

H 40

Market performance has a positive effect on financial performance.

9

Rejected

Not Tested

8

Only analyzing the direct effects of logistics performance on firm performance as presented in Figure 7-7, however, does not allow drawing conclusions on the individual importance of the two dimensions of logis-

256

7 Structural models

tics performance. For this matter, the total effects of both the level of logistics services and the level of logistics costs must be explored which can be found in Table 7-9. Since the most important figure for most firms is the “bottom line”, the financial performance, the individual importance of the two dimensions is decided just there. The analysis shows that both dimensions have a distinct positive total effect on the financial performance of the firm. While the effect of the level of logistics costs at 0.26 is slightly larger than the one of the level of logistics services at 0.23, it must be noted that they are almost equally strong. This must lead to the conclusion that when determining a firm’s logistics strategy, not only cost reductions through logistics should be aimed at. Rather, similar attention should be given to increasing the level of logistics services, since via their effect on both adaptiveness and market performance they have a substantial indirect effect on financial performance. Therefore, firms that only view the cost reduction potential of logistics gamble away the potential positive effects of increased logistics service levels. This particularly has implications for logistics outsourcing arrangement, which as argued above, today are most often engaged with the clear motivation of the customer to reduce costs while the potential to increase service levels only plays a minor role – if at all. Table 7-9. Total effects in the firm performance model Factor Adaptiveness Market Performance Financial Performance

Level of Logistics Services

Level of Logistics Costs

Adaptiveness

Market Performance

0.33 0.63 0.23

0.14 0.24 0.26

0.24 0.11

0.43

The effect described above was not found by DEHLER (2001, p. 242), who reported a total effect of the level of logistics services on the financial performance of 0.41 and of the level of logistics costs of 0.22. This may be a statistical effect since in the model proposed in this study, seven out of the eight hypotheses were found to be significant while DEHLER (2001) only found five to be significant. It may also be, however, that this finding means that in the four years between the two studies, the field of logistics has substantially advanced. As logistics as such and the offerings of LSPs have become increasingly standardized and commoditized, the focus of the firms may have shifted towards the price of the services. In this light, it is not surprising to find the increased importance of the logistics costs as reflected in the total effects, where the costs now appear to have an even slightly bigger impact than the services. However, further analysis is needed to investigate this potential trend which cannot be performed in this thesis.

7.3 Contingency variables

257

7.3 Contingency variables After the previous chapters have discussed the dimensions and antecedents of logistics outsourcing performance as well as the effect of outsourcing performance on logistics performance and the effect of logistics performance on firm performance in detail, the question remains whether the corresponding three models are moderated by the contingency variables which were introduced in chapter 4.5.2. The next chapters will present explorative empirical findings and in the following discuss external and internal contingency variables for each of the three models separately. For the analyses, the procedure and method for moderating analyses described in chapter 5.2.5.3 will be followed. 7.3.1 Moderating effects on the model of logistics outsourcing performance As the focus of this study is to examine the performance effects of exchange relationships in the logistics context, the analyses of moderating effects in the model of logistics outsourcing performance will focus on the performance relevant causal linkages – links that only affect the antecedents of outsourcing performance among themselves will not be tested. Since the construct of functional conflict as one of three performance relevant factors had to be eliminated from the model due to missing discriminant validity, only the paths from cooperation and proactive improvement on goal achievement and goal exceedance respectively remain to be analyzed. Chapter 7.3.1.1 will discuss the moderating effects of external contingency variables on the four remaining paths while chapter 7.3.1.2 will analyze the effect of internal contingency variables. 7.3.1.1 External contingency variables

Performing a structural equation modeling multi-group analysis on the moderating effects of external contingency variables on the performance relevant paths in the logistics outsourcing performance model shows in the first step that only for three of the proposed five moderating variables significant differences in the F²-value can be observed. As Table 7-10 indicates, the differences in the F²-values of environmental complexity, environmental dynamics, and uncertainty are all significant at levels well below the 10%-level of significance. The importance of logistics to the customer does not appear to have a significant moderating effect. However, since its significance level of 14.4% is just slightly above the de-

258

7 Structural models

manded threshold, it will be included in the more detailed analyses, because this level still leaves the possibility of finding a few individual moderating effects on some paths. Table 7-10. Moderating effects of external contingency variables on the logistics outsourcing performance model Logistics Outsourcing Performance Moderating Variable

Environmental Complexity

'F² P

Importance of Logistics

Industry of the Shipper

Environmental Dynamics

Uncertainty

13.64

10.20

16.55

6.85

See Text

0.009

0.037

0.002

0.144

See Text

After having completed the necessary first step of the analyses on the moderating effects, a number of potential moderators have been identified. In the further analyses, the effects of these moderating variables on the four individual paths between cooperation and proactive improvement and the two dimensions of logistics outsourcing performance are examined in detail. It shows that even though the F2-difference in the first step of the analyses suggested the existence of moderated paths, the individual analyses presented in Table 7-11 find that only very few moderations exist. The relationship between cooperation and goal achievement is moderated by environmental complexity, environmental dynamics, and uncertainty. In all cases the path, which in the original model has a standardized parameter value of 0.77 significant at the 0.1%-level, is slightly stronger when the external contingency variables’ specification is low. This suggests that an increasingly difficult situation for the customer of the LSP as caused by a more complex, more dynamic, or more uncertain environment weakens the effect of cooperation on goal achievement. In such environments, which can be e.g. the consequence of a demanding customer structure, logistically complex manufacturing processes, or increased competitive pressure, external factors disturb the relationship between customer and LSP and thus impede cooperative behavior. Consequently, the effect of cooperation on goal achievement is lower than in the following cases: When environmental complexity, environmental dynamics, and uncertainty are lower, both parties can better concentrate on the logistics outsourcing arrangement as less interferences influence the relationship, which can be characterized in this case as more stable. Hence, the cooperation can unfold its full potential in creating solutions and fostering the achievement of goals that with adversarial or non-cooperative behavior would not be attainable. For the relationship between the cooperation and goal exceedance, a strong moderating effect of the importance of logistics to the customer was

7.3 Contingency variables

259

found. It seems that at levels of high importance, cooperation becomes more important for the goal exceedance as necessary process stability can better be ensured if both parties work closely together. Furthermore, even a slight decrease of cooperation can have large consequences for the goal exceedance. These findings are especially interesting in the light of the findings above that the link between cooperation and goal achievement is not moderated by the importance of logistics. Table 7-11. Causal linkages moderated by external contingency variables in the logistics outsourcing performance model Logistics Outsourcing Performance Moderating Variable Moderated Path

Cooperation on Goal Achievment Cooperation on Goal Exceedance Proactive Improvement on Goal Achievement Proactive Improvement on Goal Exceedance

Environmental Complexity

Environmental Dynamics

Uncertainty

Importance of Logistics

Industry of the Shipper

See Text

Low

0.79****

0.79****

0.82****

-

High

0.73****

0.71****

0.71****

-

See Text

Low

-

-

-

0.21***

See Text

High

-

-

-

0.40****

See Text

Low

-

-

-

-

See Text

High

-

-

-

-

See Text

Low

0.41****

-

-

-

-

High

0.25****

-

-

-

-

(n.s.), *, **, ***, ****: Not significant, Path coefficient significant at the 10%-, 5%-, 1%-, or 0.1%-level

Finally, the empirical analysis suggests that in environments with higher complexity, a moderating effect for the path between proactive improvement and goal exceedance exists. For low complexity, the standardized parameter value is 0.41 at the 0.1%-level of significance, while for high complexity it is only 0.25 at the 0.1%-level. This may point to the fact that under low environmental complexity, both the customer and the LSP find the logistics processes better controllable and consequently, sufficient energy and room for continuous and proactive improvements remain for the LSP. The higher the environmental complexity, the more difficult is it for the LSP to handle the logistics processes at the same level of quality as in more predictable environments. Keeping this level, however, consumes both energy and room that in other situations would have been available for its improvement efforts. Therefore, the full potential of proactive improvement cannot be realized and its effect on the goal exceedance is weaker than under lower complexity. All other paths between cooperation and proactive improvement on both goal achievement and goal exceedance are not moderated by the external contingency variables. While this means on the one hand that the model proves to be quite robust against moderating effects in general, this was not expected following the argumentation presented in chapter 4.5. As a

260

7 Structural models

preliminary finding, it must therefore be established that the empirical support for the existence of moderating effects in the logistics outsourcing context is very weak. The reasons for this are not subject of this study. However, they represent a research need for the future. The above discussion has shown that in some instances, the effect of cooperation and proactive improvement on the logistics outsourcing performance is stronger in environments characterized by lower levels of complexity and dynamics. This could mean that the more complex and dynamic the industry environment is, the closer will the cooperation be and the more important will the LSP’s proactive improvement be. The empirical findings on the moderating effect of the customer’s industry show that for the majority of the industries, no moderating effect can be detected. While for the Food, Beverage and Tobacco as well as the Healthcare sector limited sample size did not allow calculations, no significant effects could be found in Consumer Goods, Manufacturing Systems Construction, Chemicals and Plastics, and Retail industries. Only the Automotive as well as the Electronics, Precision Mechanics and Optics industries display significant moderating effects. In the Automotive industry, the paths from cooperation on goal achievement and goal exceedance are moderated. While for Automotive companies the standardized path coefficient from cooperation on goal achievement is 0.72 at the 0.1%-level of significance, it is 0.78 at the same significance level for the rest. For the relationship between cooperation and goal exceedance it is not significant in the Automotive industry and 0.37 at the 0.1%-level of significance for the others. Since the Automotive industry generally is viewed as highly competitive and therefore complex and dynamic, the finding especially on the path on goal achievement suggests the following: in a very complex and dynamic environment like the Automotive industry, the effect of cooperation on outsourcing performance is weaker than in other industries, as different issues and obstructions prevent or at least restrict the full development of cooperative behavior and its beneficial effects. The findings from the Electronics, Precision Mechanics and Optics industries further support these findings. The same paths as in the Automotive industry are moderated plus the relationship between proactive improvement and goal achievement. The path coefficient between cooperation and goal achievement is 0.55 for the industry and 0.78 for the rest, both being significant at the 0.1%-level. Recalling the above findings, this does not surprise as the specific industries are very complex and dynamic indeed. The path from cooperation on goal exceedance is also moderated. It is 0.50 for the specific industries and only 0.32 for the rest, again at the 0.1%-level of significance. This marks the only finding where higher

7.3 Contingency variables

261

levels of complexity and dynamics lead to a stronger effect of cooperation on one of the dimensions of logistics outsourcing performance. This may be due to the fact that in this particular environment, cooperative behavior is especially helpful in achieving and exceeding goals, as it promotes better solutions of problems, reduces frictions between the partners, and generally enhances the collaboration. As this conflicts with the findings presented above, further research must establish the reasons for these diverging effects which will not be examined in this study due to its exploratory focus. A last finding in the individual analysis of the Electronics, Precision Mechanics and Optics industries was made for the relationship between proactive improvement and goal achievement. While for all remaining industries the effect is not significant, just as in the non-moderated model, it is positive at 0.30 at the 1%-level of significance for the specific industries. This implies that here, proactive improvement is so important that it has a substantial positive effect on the goal achievement. This finding, together with the one presented above on cooperation, might suggest that the Electronics, Precision Mechanics and Optics industries are more advanced in their logistics processes and relationships with the respective LSPs than the firms of other industries. As the discussion has shown, only in 8 out of the 20 possible cases, empirical support could be found for moderating effects. While it is not the aim of this study to explain the exact reasons for not finding moderating effects on several individual causal linkages, this poses a substantial research need and opportunity for the future. 7.3.1.2 Internal contingency variables

The multi-group analysis of the moderating effects of internal contingency variables on the performance relevant paths of the logistics outsourcing performance model showed that only for two of the proposed six moderating variables significant differences in the F²-value can be observed. As Table 7-12 indicates, only for the variables asset specificity and the degree of centralization, the difference in the F²-value is significant at levels below the 10% significance level. For the other four variables, no significant moderating effects were found.

262

7 Structural models

Table 7-12. Moderating effects of internal contingency variables on the logistics outsourcing performance model Logistics Outsourcing Performance Moderating Variable

Products

Asset Specificity

Process Orientation

Size of the Shipper

Degree of Centralization

Frequency

'F²

4.71

11.40

3.47

1.96

9.97

2.67

P

0.318

0.022

0.483

0.743

0.041

0.614

Table 7-13 gives an overview of the moderating effects found in the detailed analyses. With higher levels of asset specificity, the causal linkages between proactive improvement and the two dimensions of outsourcing performance become stronger. The path coefficient from proactive improvement on goal achievement is negative at -0.04 for low levels of asset specificity, while for high levels it is 0.13, both being significant at the 0.1%-level of significance. Similarly, under low asset specificity the path coefficient from proactive improvement on goal exceedance is 0.23, while under high specificity it is 0.40, again significant at the 0.1%-level. This indicates that with higher levels of asset specificity, the proactive improvement of the LSP has a substantially stronger effect on the outsourcing performance, e.g. because its impact under the influence of more complex products, processes, or customer demands is significantly higher than for goods with low asset specificity. Processes and products with high asset specificity require the LSP to develop individualized and specialized solutions. This may cause a deeper engagement by the service provider compared to standard solutions, which then fosters an increased effect of the proactive improvement on the performance dimensions. The only other contingency variable that exhibits measurable moderating effects is the degree of logistics centralization of the LSP’s customer. The path from cooperation to goal exceedance is moderated: While the standardized parameter value is 0.20 at the 1%-level of significance for low levels of centralization, it is 0.44 significant at the 0.1%-level of significance for high levels of centralization. This suggests that the more centralized the customer and LSP can work together, the more important is the cooperation between the parties to reach outsourcing performance. Both parties may find it easier there to cooperate, to coordinate their activities, and to collaborate.

7.3 Contingency variables

263

Table 7-13. Causal linkages moderated by internal contingency variables in the logistics outsourcing performance model Logistics Outsourcing Performance Moderating Variable Products

Asset Specificity

Process Orientation

Size of the Shipper

Degree of Centralization

Frequency

-

Moderated Path

Cooperation on Goal Achievment Cooperation on Goal Exceedance Proactive Improvement on Goal Achievement Proactive Improvement on Goal Exceedance

Low

-

-

-

-

-

High

-

-

-

-

-

-

Low

-

-

-

-

0.20***

-

High

-

-

-

-

0.44****

-

Low

-

-0.04****

-

-

0.11**

-

High

-

0.13****

-

-

(n.s.)

-

Low

-

0.23****

-

-

0.43****

-

High

-

0.40****

-

-

0.22***

-

(n.s.), *, **, ***, ****: Not significant, Path coefficient significant at the 10%-, 5%-, 1%-, or 0.1%-level

For the effect of proactive improvement on the goal achievement, no significant effect is found for high levels of centralization, while for low levels it is 0.11 at the 5%-level of significance. Since this effect is not significant in the main model presented in chapter 7.1.3, it is interesting to find that in decentralized environments, proactive improvement and the according behavior of the LSP has indeed a significant influence on goal achievement. A similar effect can be observed for the effect of proactive improvement on goal exceedance. While for low levels of centralization the effect has a strength of 0.43 at the 0.1%-level of significance, it is only 0.22, significant at the 1%-level, for high levels. Evidently, the effect of proactive improvement of the LSP is much stronger in decentralized environments. This may be due to the fact that with direct contact to operative units of the customer at multiple levels, the LSP is closer to a variety of customer opinions, suggestions, and expressions of general satisfaction. Consequently, this additional knowledge will facilitate the identification of improvement potentials. While it must be recalled that from the many proposed moderating effects only very few have shown to be significant, from the analyses it can be learned that the effect of proactive improvement on outsourcing performance is stronger in decentralized environments and at higher levels of asset specificity. In total, the empirical evidence for moderating effects is very limited. This means that while evidently some moderations exist, the model again has shown to be very robust against moderation of contingency variables. As shown in the previous chapter, a large number of moderations, here a total of 19, were found not to be significant. While

264

7 Structural models

again, it is not the aim of this study to explain why the moderating effects frequently are found to be non-significant, this seems to be a valuable topic for further analysis. 7.3.2 Moderating effects on the model of logistics performance In chapter 4.5.1 it was established that it is highly relevant to test for potential moderating effects of both external and internal contingency variables also in the logistics performance model presented in chapter 7.2.1.2. Consequently, chapter 7.3.2.1 will discuss the effects of external contingency variables, while chapter 7.3.2.2 will analyze the effects of internal contingency variables. 7.3.2.1 External contingency variables

The analysis showed that neither environmental complexity, nor environmental dynamics or the importance of logistics to the customer have significant moderating effects. As Table 7-14 further shows, only uncertainty and the industry of the customer in one case have a measurable moderating effect. For both cases, the difference in the F²-value is well below the 10% significance level proposed in chapter 5.2.5.3. Table 7-14. Moderating effects of external contingency variables on the logistics performance model Logistics Performance Moderating Variable

Environmental Complexity

Environmental Dynamics

Uncertainty

Importance of Logistics

Industry of the Shipper

'F²

4.48

3.01

8.80

4.08

See Text

P

0.345

0.556

0.066

0.396

See Text

When analyzing the moderating effect of the variable uncertainty in detail, a significant effect is found on the relationship between goal achievement and the level of logistics costs. While it is not significant for high levels of uncertainty, it is 0.34 at the 0.1%-level of significance for low uncertainty. This effect for low levels of uncertainty is stronger than the effect in the original model where the path coefficient was only 0.21. This may be due to the fact that as higher levels of the contingency variables lead to more complexity, uncertainty, and insecurity for the customer, deficits in its outsourced logistics processes have a stronger negative effect for the overall logistics performance. This is the case because in very complex and demanding environments, a slightly smaller logistics outsourcing

7.3 Contingency variables

265

performance will lead to a significantly lower logistics performance. Vice versa – as in the case of low uncertainty shown above – in less complex and undemanding environments the customer can afford slight deficits in the outsourcing performance, as its consequences will not have effects that are as strong on the overall logistics performance. Consequently, in less uncertain environments, the effect of goal achievement on the level of logistics costs is stronger than in uncertain environments as demonstrated by the path coefficient of 0.34 for low uncertainty compared to the 0.21 of the original model. Table 7-15. Causal linkages moderated by external contingency variables in the logistics performance model Logistics Performance Moderating Variable Moderated Path

Environmental Complexity

Environmental Dynamics

Uncertainty

Importance of Logistics

Industry of the Shipper

Low

-

-

-

-

See Text

High

-

-

-

-

See Text

Goal Achievement on Level of Low

-

-

0.34****

-

-

High

-

-

(n.s.)

-

-

Low

-

-

-

-

-

High

-

-

-

-

-

Low

-

-

(n.s.)

-

-

High

-

-

0.20***

-

-

Goal Achievement on Level of Logistics Services

Logistics Costs

Goal Exceedance on Level of Logistics Services

Goal Exceedance on Level of Logistics Costs

(n.s.), *, **, ***, ****: Not significant, Path coefficient significant at the 10%-, 5%-, 1%-, or 0.1%-level

A further moderating effect is detected for the link of goal exceedance on the level of logistics costs. While it is not significant for low uncertainty, for high uncertainty it is 0.20 significant at the 1%-level. This is higher than the effect in the original model which in chapter 7.2.1.2 was shown to be 0.10, significant on the 10%-level. This suggests that in uncertain environments the effect of goal exceedance on the level of logistics costs is higher than in more stable situations. This conflicts with the findings presented above on the link of goal achievement on the level of logistics costs and again represents a research need for the future. As Table 7-15 further shows, the effect of goal achievement on the level of logistics services is also moderated by the industry of the customer. However, from the six industries that can be analyzed due to sufficient sample size, only the Retail sector shows a moderation. Here, the effect has a strength of 0.38, significant at the 0.1%-level, while the remaining industries only have an effect of 0.15, significant at the 5%-level. This suggests that in the Retail industry, which can be considered quite advanced in terms of logistics processes compared to other industries, the ef-

266

7 Structural models

fect of goal achievement for the level of logistics services is substantially stronger than for the other industries. However, summing up the discussion on the moderating effects of external contingency variables on the logistics performance model, it must be concluded that again, very limited evidence is found for moderating effects as a large number of potential moderations have found to be not significant. It will be a task for future research to determine the reasons why 17 out of a total of 20 possible effects did not find support, which as argued above is not subject of the research of this study. For now, it can be concluded that the model, even more so than the logistics outsourcing performance model, is evidently very robust against moderations by external contingency variables and therefore generally applicable. 7.3.2.2 Internal contingency variables

The analyses of the moderating effects of internal contingency variables showed that for none of the six variables, a significant difference in the F²value exists. Thus it can be determined that no moderations exist and the internal situation of the firm apparently has no role in determining the size of the standardized parameter coefficients between the dimensions of outsourcing performance and logistics performance. The model thus has again demonstrated its robustness against the moderation of contingency variables. The results are presented in Table 7-16. Still, further research is required to determine why exactly none of the potential moderating effects did find empirical support. Table 7-16. Moderating effects of internal contingency variables on the logistics performance model Logistics Performance Moderating Variable

Products

Asset Specificity

Process Orientation

Size of the Shipper

Degree of Centralization

Frequency

'F²

4.68

0.43

1.21

1.06

2.04

2.17

P

0.322

0.980

0.877

0.901

0.728

0.704

7.3.3 Moderating effects on the model of firm performance In chapter 4.5.1, it was proposed that the causal linkages between the dimensions of logistics performance and those of firm performance are moderated by both external and internal contingency variables. In the following two chapters, first the effects of the external variables will be discussed in

7.3 Contingency variables

267

chapter 7.3.3.1, before in chapter 7.3.3.2 the moderating effects of the internal variables will be presented. 7.3.3.1 External contingency variables

As Table 7-17 shows, four out of the five external contingency variables display significant differences in F²-values and therewith indicate the existence of moderating effects. Solely the importance of logistics to the customer does not seem to have a moderating influence. Table 7-17. Moderating effects of external contingency variables on the firm performance model Firm Performance Moderating Variable

Environmental Complexity

'F² P

Importance of Logistics

Industry of the Shipper

Environmental Dynamics

Uncertainty

19.30

29.34

16.73

7.61

See Text

0.013

0.000

0.033

0.472

See Text

As Table 7-18 indicates, only five out of a total of 40 paths are moderated. These effects will be discussed in the following. For the variable environmental complexity, a moderating effect is found on the relationship between adaptiveness and market performance. While it is not significant for high levels of complexity, for low levels the path coefficient is 0.36 significant at the 0.1%-level. Since the strength of the relationship in the original model is 0.24, this suggests that adaptiveness and the associated flexibility have a stronger impact on the firm performance dimensions in situations characterized by low complexity. This may be due to the fact that as environmental complexity decreases, firms may find it easier to adapt to their environment and to address the needs and demands of their own customers in a more appropriate way, thus leading to an increased market performance. Similar observations can be made for the moderating effects of environmental dynamics. Here, the relationship between adaptiveness and market performance is moderated in the same fashion. For high levels of dynamics, the strength of the path coefficient is only 0.12 at the 5%-level of significance, while it is 0.37, significant at the 0.1%-level, for low environmental dynamics. As for complexity, is seems as though the impact of flexibility on market performance is distinctly stronger in more stable and predictable environments as firms find it easier to adapt to their environment. The effect of the level of logistics costs on adaptiveness is also moderated by the environmental dynamics. While the effect under high dynamics

268

7 Structural models

is not significant, for low levels it is significant at the 0.1% level with a strength of 0.30 and thus substantially stronger than in the main model where the path coefficient is only 0.14. This suggests that the impact of logistics costs reductions on the adaptiveness of a firm is particularly strong in stable and predictable environments. This may be due to the fact that especially in stable and thus predictable environments, the price of its own products and services becomes increasingly important for the customer of the LSP as its own customers’ choices increase due to better planning opportunities. The price, determined to a significant portion by the level of logistics costs, therefore takes on an important part also in the flexibility to react to the changing demands of the market – the lower the costs, the more flexible can the customer react. Table 7-18. Causal linkages moderated by external contingency variables in the firm performance model Firm Performance Moderating Variable Moderated Path

Environmental Complexity

Environmental Dynamics

Uncertainty

Importance of Logistics

Industry of the Shipper

-

Low

-

-

-

-

High

-

-

-

-

-

Level of Logistics Services on Low

-

-

0.65****

-

-

High

-

-

0.50****

-

-

Level of Logistics Services on Low

-

-

-

-

-

Level of Logistics Services on Adaptiveness

Market Performance

Financial Performance

Level of Logistics Costs on Adaptiveness

Level of Logistics Costs on Market Performance

Level of Logistics Costs on Financial Performance

Adaptiveness on Market Performance

Market Performance on Financial Performance

High

-

-

-

-

-

Low

-

0.30****

-

-

-

High

-

(n.s.)

-

-

-

Low

-

-

-

-

-

High

-

-

-

-

-

Low

-

-

-

-

-

High

-

-

-

-

-

Low

0.36****

0.37****

-

-

See Text

High

(n.s.)

0.12**

-

-

See Text

Low

-

-

-

-

-

High

-

-

-

-

-

(n.s.), *, **, ***, ****: Not significant, Path coefficient significant at the 10%-, 5%-, 1%-, or 0.1%-level

In the same direction goes the finding on the moderating effect of uncertainty. Here, the standardized parameter value for the effect of the level of logistics services on the market performance is 0.65 for low uncertainty and 0.50 for high levels, both significant at the 0.1%-level. Seemingly, under conditions of lower uncertainty the effect of service level changes on market performance is stronger. The more uncertain and unstable the envi-

7.3 Contingency variables

269

ronment, the less important apparently is the level of logistics services for market performance and the more important become other factors that were not subject of research in this study. Apparently, the customers of the LSP’s customer find these other factors more important with increasing levels of uncertainty. Therefore, the level of logistics services seems to be a substantially stronger driver of the purchasing decision in stable and predictable environments, while under stronger uncertainty, other motivations prevail. A final moderating effect comes from the industry of the customer. Here, only a single moderation can be found in the Manufacturing Systems Construction industry. All other industries do not show any significant moderating effects. However, in the Manufacturing Systems Construction industry the relationship between adaptiveness and market performance is moderated. For this industry it has a standardized parameter value of 0.52, for the remaining industries it is 0.21, both significant at the 0.1%-level. Building on the above findings of this chapter that showed that especially in situations with lower environmental complexity and dynamics this effect can be observed, it can be argued that the Manufacturing Systems Construction industry compared to other industries is in such a rather stable and predictable situation. Consequently, the impact of adaptiveness on market performance there is stronger than on average. For the moderating effect of external contingency variables on the firm performance model it must be concluded that while some moderations were detected, the vast majority of paths are not moderated. Therefore, only very limited evidence for the relevance of contingency variables in this context was found. Hence, the model is very robust against moderating effects and therefore is rather generally applicable. 7.3.3.2 Internal contingency variables

After the firm performance model has proven to be quite robust in the analyses of the external contingency variables, the test of internal variables revealed that only products and asset specificity exert a significant influence, while no moderating effects could be detected for the remaining four variables. The differences in the F²-values and the corresponding significance levels are presented in Table 7-19. Especially the variable products with a total of four shows a large number of moderating effects (see Table 7-20). The effect of the level of logistics costs on adaptiveness, 0.14 in the original model, is 0.29 at the 0.1%level of significance for a low logistical complexity of the product range and not significant for a high complexity. This suggests that the less com-

270

7 Structural models

plex the product range, the higher is the impact of logistics costs reductions for the adaptiveness of the firm. Table 7-19. Moderating effects of internal contingency variables on the firm performance model Firm Performance Moderating Variable

Products

Asset Specificity

Process Orientation

Size of the Shipper

Degree of Centralization

'F²

27.90

15.77

11.22

12.12

6.97

11.93

P

0.000

0.046

0.190

0.146

0.540

0.154

Frequency

Similar is the situation for the moderation of the relationship between the level of logistics costs and financial performance. While the effect is not significant for a high complexity of the product range, it is 0.25, significant at the 0.1%-level, for a low complexity. In comparison to the strength of the path coefficient in the original model of 0.15, this is significantly stronger. It can therefore be argued that for firms with a logistically very standardized product range, the effect of logistics cost reductions for the financial performance is substantially stronger than for firms with complex and customized products. This may be due to the often very low profit margin of firms operating with standardized products and commodities. Here, a slight absolute cost reduction may have comparably large effects on the relative financial performance. As it was found for several external contingency variables, the effect of adaptiveness on market performance is moderated. Under high product range complexity it is only 0.17 significant at the 1%-level, while under low complexity it is 0.34 at the 0.1%-level of significance. It can therefore be argued that especially firms with a logistically standardized and commoditized product range can reap the benefits of adaptiveness for an increased market performance. This may be due to fact that the flexibility gained is disproportionately appreciated by their customers which have the choice between several products with very few differentiating characteristics. Finally, the relationship between market performance and financial performance is moderated by the product range. Under low product range complexity, the effect is 0.30, under high complexity it is 0.51, both significant at the 0.1%-level. Evidently, firms with a broad and individualized product range find it easier to get a price premium for their goods, translating into higher financial performance. This, however, suggests that while firms with commoditized and standardized products have advantages in terms of flexibility, the effect of market performance on the bottom line is stronger with individualized and non-standardized products. Alternatively,

7.3 Contingency variables

271

it can also be argued that firms displaying market performance which produce commodities and other logistically simple products have smaller profit margins than other firms. Table 7-20. Causal linkages moderated by internal contingency variables in the firm performance model Firm Performance Moderating Variable Products

Asset Specificity

Process Orientation

Size of the Shipper

Moderated Path

Degree of Centralization

Frequency

Low

-

0.47****

-

-

-

-

High

-

0.20***

-

-

-

-

Level of Logistics Services on Low

-

-

-

-

-

-

High

-

-

-

-

-

-

Low

-

-

-

-

-

-

Level of Logistics Services on Adaptiveness

Market Performance

Level of Logistics Services on Financial Performance

Level of Logistics Costs on Adaptiveness

Level of Logistics Costs on Market Performance

Level of Logistics Costs on Financial Performance

Adaptiveness on Market Performance

Market Performance on Financial Performance

High

-

-

-

-

-

-

Low

0.29****

-

-

-

-

-

High

(n.s.)

-

-

-

-

-

Low

-

-

-

-

-

-

High

-

-

-

-

-

-

Low

0.25****

-

-

-

-

High

(n.s.)

-

-

-

-

-

Low

0.34****

-

-

-

-

-

High

0.17***

-

-

-

-

-

Low

0.30****

-

-

-

-

-

High

0.51****

-

-

-

-

-

(n.s.), *, **, ***, ****: Not significant, Path coefficient significant at the 10%-, 5%-, 1%-, or 0.1%-level

A last moderating effect is found for the variable asset specificity on the causal linkage between the level of logistics services and adaptiveness. While its path coefficient is 0.20 and significant at the 1%-level for high asset specificity, it is 0.47 at the 0.1%-level of significance for low asset specificity. Apparently, the impact of the level of logistics services on adaptiveness is strongest when the asset specificity is low. This suggests that the less specific the logistics processes of the customer are, the higher is the impact of the outsourced services on the flexibility of the customer. Vice versa, very specific assets hinder the customer in developing additional adaptiveness as logistics processes tend to be more complex and difficult to handle. Summing up the discussion of this chapter, it must again be remarked that the vast majority of causal linkages are not moderated by internal contingency variables. This is substantially less evidence for the relevance of

272

7 Structural models

contingency variables than originally expected. It must therefore be concluded that the firm performance model, like the other two models, is very robust against moderating effects and therefore rather generally applicable. What remains to be shown are the reasons for the absence of so many potential moderating effects; 43 out of a total of 48 moderations were found to be non-significant. As argued above, it is not the aim of this study to in detail investigate the reasons for the absence. However, this will be a worthwhile research subject for future logistics research with implication for both theory and practice.

8 Summary and results

The following chapters will summarize the findings of this study, present managerial implications and will develop recommendations for further research on the basis of questions that either could not be addressed in this study or that only emerged during its course.

8.1 Main results In this study, a number of deficits in current logistics research were identified. While it had been established in previous research that logistics outsourcing performance is a driver of logistics performance, which in turn is a facilitator of firm performance, the main antecedents of logistics outsourcing performance were not known. In fact, in the measurement of logistics outsourcing performance, substantial deficits existed which also cast doubts on the earlier findings on its influence on both logistics- and firm performance. In addition to that, the role of contingency variables as moderators of the relationships between the antecedents of logistics outsourcing performance and its effects on logistics- and subsequently of that on firm performance was unclear. The intention of this study was to contribute to the logistics outsourcing discussion by analyzing the relationships between logistics service providers and their customers as the main drivers of outsourcing performance. Technical aspects of the outsourcing implementation were not addressed. On the basis of a literature review, the deficits in current research as discussed above were identified and four research questions developed that guided this study. Consequently, ten relationship variables as antecedents of logistics outsourcing performance were identified, using transaction cost-, social exchange-, and commitment-trust theory as a theoretical framework. Contingency variables were derived using the contingency approach. To answer the four research questions, a large-scale empirical study was conducted. A total of 3.402 German logistics executives in manufacturing and retailing industries received the e-mail request to participate in an on-

274

8 Summary and results

line survey with 209 items. 579 managers responded, giving the study a response rate of 17.0%. After correcting the database for missing values, a total of 549 usable cases remained. The data was analyzed using covariance structure analysis. Recommendations and standards from literature concerning the conduct of empirical studies, as well the analyses and fit criteria were rigorously obeyed. In the following, the four research questions will be individually addressed and the key results presented. R1 :

Which are the main influencing factors of an outsourcing relationship between a customer and a LSP and what influence do they have on the outsourcing performance?

Conducting a literature review as well as employing a theoretical framework that included transaction cost-, social exchange-, and commitment-trust theories, ten relationship variables were identified that supposedly influence logistics outsourcing performance. Those are cooperation, proactive improvement, and functional conflict as well as communication, trust, relationship commitment, involvement, shared values, openness, and opportunism. On the basis of the existing literature and new argumentations, a new outsourcing performance construct was developed. It contains the two dimensions goal achievement and goal exceedance. Goal achievement measures whether the expectations of the customer with respect to the service quality and the costs of the outsourcing arrangement have been met. Goal exceedance, on the other hand, grasps whether the LSP has created additional value added by being customer oriented, innovative and proactive, thereby increasing value in the dimensions of service level increases and cost reductions. The separate measurement of the two dimensions was necessary as on the one hand, fundamentally different efforts and actions are necessary to excel in any of the two dimensions, and on the other hand only separate measurement can establish whether they truly are distinct. Through the literature review and the utilization of the theoretical framework, direct effects on goal achievement and goal exceedance were hypothesized for cooperation, proactive improvement, and functional conflict. The remaining seven variables were hypothesized to have indirect effects only, all mediated through one or more of the three variables with direct effects. These findings directly lead to the second research question which will be addressed in the following:

8.1 Main results

R2 :

275

How are the influencing factors of outsourcing relationships and their performance effects to be integrated into one model and which interdependencies must be considered between them?

After the operationalization of the constructs on the basis of the empirical analysis, it was found that the construct of logistics outsourcing performance indeed is bi-dimensional, consisting of both goal achievement and goal exceedance. For the ten relationship variables, discriminant validity was also assessed, finding that all variables are showing discriminant validity with the exception of functional conflict which was found not to be sufficiently discriminant from cooperation, communication, and trust. It was therefore excluded from further analysis. A basic model with nine relationship variables and the two logistics outsourcing performance dimensions was formulated which exhibited slightly unsatisfactory fit criteria. Consistent with literature recommendations and standards, the model was modified to improve its fit. As a consequence, the variable involvement was eliminated, which lead to the acceptance of the model. After the reduction of the model to eight antecedents directly and indirectly influencing the two dimensions of logistics outsourcing performance, the final model displays a distinct three level structure. Five antecedents, namely shared values, trust, commitment, openness, and opportunism, form a block of behavioral and attitudinal variables that only indirectly, namely through their direct effects on the two action variables cooperation and proactive improvement, affect the two dimensions of logistics outsourcing performance. The factor communication takes on a mediating position between the behavioral/attitudinal variables and the action variables. A large number of the hypotheses developed for the model found support in the empirical analysis. Cooperation exerts a direct influence on both goal achievement and on goal exceedance, while the effect on the former is substantially stronger than on the latter. It furthermore has a direct positive effect on proactive improvement. Other than for cooperation, the effect of proactive improvement of the LSP on the goal achievement is not significant. However, the influence of proactive improvement on goal exceedance is approximately as strong as the effect of cooperation. It can therefore be concluded that while cooperation is the main driver of goal achievement and does facilitate goal exceedance, the proactive improvement of the LSP must not be neglected if an exceedance of the goals is aspired by the customer.

276

8 Summary and results

The indirect effects of the behavioral/attitudinal variables on outsourcing performance were also found to be important. The shared values are a central variable, exerting direct positive effects on cooperation, communication, relationship commitment, trust, and openness and a negative effect on opportunism. It therefore constitutes the backbone of the relationship between customer and LSP. Antecedents of cooperation are the relationship commitment and trust, which also has a direct effect on commitment. The factor openness directly influences both trust and communication and therefore only has indirect effects on the action variables cooperation and proactive improvement. Opportunism, to the contrary exerts a negative effect on trust, while the hypothesized negative effect on cooperation is not significant. Finally, the variable communication takes on a mediating role between behavioral/attitudinal variables and action variables. It positively influences the cooperation and contrary to the hypothesis negatively influences the proactive improvement. This indicates that communicating and exchanging information alone do not lead to increase the proactive improvement of the LSP. However, communication can in the long run have a total positive effect on proactive improvement through the mediation of cooperation, trust, opportunism, and commitment. Overall, the explanatory value of the model of logistics outsourcing performance is very high. The R2 of goal achievement is 63.4%, the one of goal exceedance is 35.1%. This suggests that the relationship variables are the main driver of logistics outsourcing performance and therefore require special attention by both the customer and the LSP. All relationship variables that remained in the main model have a positive total effect on logistics outsourcing performance and on all other variables, even though some of them may only have indirect effects. The exception in the model is the variable opportunism, which has a negative effect on all antecedents and the two dimensions of logistics outsourcing performance. The next question addressed in the study was the effect the newly developed construct of logistics outsourcing performance has on the logistics performance and if the hypothesized link between logistics- and firm performance could be confirmed. Research question three was the following: R3 :

Which influence does the outsourcing performance have on the logistics performance and on the firm performance?

In the study, a basic model of logistics performance was developed that hypothesized direct effects of both goal achievement and goal exceedance on the two dimensions of logistics performance, the level of logistics services and the level of logistics costs.

8.1 Main results

277

The empirical analysis found support for all four hypotheses. Goal achievement influences almost equally strong the two dimensions of logistics performance. The effect of goal exceedance on the level of logistics services is even slightly stronger than the respective effect of goal achievement, while its effect on the level of logistics costs is measurable, but notably weaker. These findings suggest that goal achievement is an important facilitator of logistics performance by increasing both the service level and reducing the overall logistics costs. Goal exceedance is of utmost importance to firms that want to increase the level of logistics services through outsourcing, as it has an even stronger effect than the goal achievement. With the model, 10.8% of the variance of the level of logistics services and 7.3% of the level of logistics costs can be explained through the two dimensions of logistics outsourcing performance. Since this is only a partial model focusing only on the most important LSP of the customer, the effect of logistics outsourcing performance over all relationships can legitimately be assumed to be substantially higher, indicating an even higher strategic importance of the issue than the above presented results originally suggest. The second part of research question three contains the testing of the model that hypothesized the effect of logistics performance on firm performance. Logistics performance was measured with the two dimensions presented above, while firm performance is a tri-dimensional construct consisting of adaptiveness, market performance, and financial performance. The findings of existing literature on the model could be confirmed and substantially extended. The level of logistics services strongly influences the adaptiveness and the market performance of the firm, while the effect on the financial performance is not significant. The level of logistics costs on the other hand has a moderate effect on the financial performance and on the adaptiveness of the firm. Furthermore, it has an effect of medium strength on the market performance. As expected, the adaptiveness has a medium effect on the market performance, which in turn has a strong effect on the financial performance. These results allow new insights into the importance of logistics performance. While only the level of logistics costs have a direct effect on the important aspect of financial performance, the total effect on it exerted by the level of logistics services is almost as strong as those of the level of logistics costs due to the indirect effects via the firm’s adaptiveness and market performance. Therefore it must be concluded that while the level of logistics costs, which as has been shown above are significantly affected by logistics outsourcing, are an important aspect, customers must also at-

278

8 Summary and results

tend to the level of logistics services. The lever this aspect of logistics provides is virtually as strong as the one of cost reductions. Consequently, neglecting it would mean giving away a vast potential. As discussed already in the introduction to this study, the role of the influence of contingency variables on the models presented above remains unclear. Therefore, research question four was developed: R4 :

Which contingency variables for the logistics outsourcing context can be identified and which moderating effects can be observed in the outsourcing performance model and its causal linkages to logistics and firm performance?

A total of ten contingency variables were developed on the basis of a literature review and the contingency approach. They can be divided into external and internal contingency variables: External variables that cannot be influenced directly by the organization of the customer are environmental complexity, environmental dynamics, uncertainty, and the customer’s industry. Internal variables on the other hand can be influenced directly by the customer. They include the customer’s products, the size of the organization, the degree of centralization of logistics decisions, the asset specificity of the outsourced processes, the frequency, and the process orientation of the customer. Like the constructs of the models introduced above, the contingency variables were operationalized on the basis of the empirical data. Several variables had to be substantially modified on these grounds. The variable environmental dynamics was split into two different constructs, one measuring the intensity of competition in the customer’s industry and the other one the importance of logistics to the customer of the LSP’s customer. All three performance models, for logistics outsourcing-, logistics-, and firm performance, were tested for moderating effects of the eleven contingency variables. However, only very limited evidence is found for these moderating effects. In total, out of a potentially possible 176 moderations, only 26 find empirical support. For every contingency variable, 16 moderating effects in the different models were tested. Uncertainty and the products show 4 moderations each, while environmental complexity, asset specificity, and the centralization of logistics decisions show 3. The remaining variables all show only one or none at all. In total, the external variables moderate 16 out of 80 causal linkages (20.0%), while the internal variables moderate 10 out of 96 (10.4%). These findings suggest that the models developed are very robust against moderations and therefore are generally applicable. To investigate

8.2 Managerial implications

279

the reasons why certain moderating effects did not find any support, in total 150 or 85.2%, was not subject of this study. However, it will be an interesting field for future research as the results found in the context of this research question are not fully satisfactory.

8.2 Managerial implications While it was not the explicit goal of this study to develop recommendations for the practice, the key findings have several managerial implications that will be discussed in the following: Firms must realize the potential of logistics outsourcing and should extent their efforts. Logistics outsourcing serves to reduce the overall logistics costs of the firm, thereby conserving resources that can be utilized elsewhere, and on the other hand strengthens the logistics service levels of the firm. Both aspects are important differentiators in a world that is characterized by increasing competition. Logistics outsourcing should therefore be a relevant option for the firms as its implementation and extension promises significant potential. Logistics outsourcing performance has different facets – depending on the focus, the results of the outsourcing arrangement will differ. In an outsourcing arrangement, the customer can achieve its goals or exceed them. Both is possible simultaneously, but it can also occur detached from each other. Through goal achievement, the logistical service levels will increase, while the level of logistics costs decreases. However, the goal exceedance has an even higher effect on the logistical service levels – firms that focus on increasing their service levels should therefore always expect their LSP to exceed their expectations. This will also further reduce the costs, but not as strong as the initial goal achievement: most of the cost reduction potential is raised by the LSP already on the basis of the original agreement. The true driver of the outsourcing performance is the relationship with the LSP, not the extent of outsourced logistics processes. Factors such as cooperation between customer and LSP, communication, openness, trust, shared values between the partners, but also the commitment of the customer to the relationship, and the constant proactive improvement of the processes through the LSP make the outsourcing arrangement more successful. Opportunism of the LSP is always detrimen-

280

8 Summary and results

tal. The customer must focus on the relationship when outsourcing – the most critical part of outsourcing starts after the decision to outsource has been made. The customer needs a holistic and detailed understanding of the relationship, because all factors are important. Depending on the goals of the customer, the focus may change: if cost reductions are the ultimate aim, it may be sufficient to select an LSP that achieves the goals that are expected by the customer. If the goals are significant increases of the logistics service levels, the customer must choose an LSP that continuously improves the logistics processes – proactively. Logistics performance is a main driver of firm performance. However, most firms do not raise its full potential. The reduction of logistics costs is the most common motivation for outsourcing. To a certain degree, this motivation and its according actions are justified, as it directly leads to a higher financial performance of the customer. At the same time, it also increases the firm’s flexibility to react to changing market needs and increases the market performance. Since it is also easy to measure, it does not surprise that it is the most prominent reason for outsourcing. The potential of logistics service level increases, however, is substantially underestimated as the current, strongly cost-oriented logistics outsourcing motivation among firms has shown. While it is more difficult to measure progress in this field, its potential benefits are vast. It does not directly influence the bottom line of the firm, but it increases the firm’s flexibility and the market performance substantially stronger than cost reductions. Through those performance increases, it has a total effect on the financial performance that is almost as strong as those of cost reductions. That does not mean that firms should not aim at reducing their logistics costs. But it does mean that a lot of additional potential can be raised if the firms also try to increase their logistics service levels. While this may take more intense efforts in the short run, it does promise financial rewards in the long run. One of the main levers is the proactive improvement of the LSP. If the customer selects an LSP that displays high levels of continuous improvement in the first place or succeeds in improving the current LSP’s behavior – which according to WALLENBURG (2003) usually is not very far developed in terms of proactive improvement – then the higher goal exceedance that also leads to higher levels of logistics services may be reachable with comparably little efforts. The models proposed in this study apply generally and for all firms in manufacturing and retail industries – the specific context of the firm only has a very limited impact.

8.3 Recommendations for further research

281

The argument that certain firms are just different than others has lost a lot of its weight in the logistics context. The research has shown that indeed, some contingency variables moderate individual aspects of the performance models – that means, however, that the remaining parts of the respective model is not moderated. Furthermore, in almost 90% of the cases, the models are not moderated at all. This means that firms should carefully keep in mind what their particular situation is – a framework is offered by the external and internal contingency variables developed in this study. But when deciding what decision is appropriate in their particular context, the firms should remember that the performance models and the consequent implications that were introduced in this chapter, apply for almost all contexts and individual situations.

8.3 Recommendations for further research The findings of this study do not only offer insights into managerial implications, but also allow recommendations for further research. They will be discussed in the following: The relationship variables identified in this study have high explanatory power for the outsourcing performance with a R2 of goal achievement of 63.4% and of goal exceedance of 35.1%. The question remains which variables account for the remaining variance. Further research is needed to establish which other aspects, such as the technical implementation of the outsourcing, may provide additional explanatory value. This is especially relevant with respect to the goal exceedance, whose squared multiple correlation is substantially lower than that of goal achievement. The dimensions of logistics outsourcing performance have been shown to positively influence the two dimensions of logistics performance. However, the squared multiple correlations were not found to be particularly high with 10.8% for the level of logistics services and 7.3% for the level of logistics costs. The model must therefore be considered a partial model. In this study, only the relationship between the customer and its most important LSP was examined. It remains to be shown how much of the remaining variance can be explained if more, preferably all, relationships to other LSPs are included in the model. Furthermore, the role of the logistics processes that have remained in-house could be integrated into the model, thus allowing for a total model of a firm’s logistics performance. In this study, a total of seven significant effects of the dimensions of logistics performance on the three dimensions of firm performance have been found. This marks a substantial increase compared to the findings of

282

8 Summary and results

DEHLER (2001) and ENGELBRECHT (2004) who used very similar scales but only found five significant effects and considerably different total effects. The reasons for these deviations represent very fertile ground for further research. Whether the concept of logistics as understood by the firms has simply farther evolved or the commoditization and standardization of logistics processes and services has altered the importance of logistics costs and service levels for the firm performance, only a detailed analysis of the changes of the concept of logistics over time and the changing demands of the market, both with respect to buyers and suppliers, enables a fundamental understanding of this development. The number of moderating effects of contingency variables in this study has been extremely small with respect to the potentially possible number. Only 14.8% of the possible effects were found to moderate the different causal linkages significantly. This must be attributed on the one hand to the fact that the variables conceptualized by KLEER (1991), which were operationalized for this study, have failed to work out as expected. It must now be examined whether this is attributable to either an insufficient operationalization or conceptualization. Furthermore, it may be that better results concerning moderating effects could be obtained if not only the context of the customer were targeted. Also, contingency variables focusing on the relationship between the LSP and its customer as well as the context of the LSP alone could supply further insights. Through this research, the initial findings of this study that certain aspects of the relationship between customer and LSP as well as its performance implications could be extended. Finally, it must be noted that all results put forward in this study have been reached on the basis of a survey that was conducted with German respondents only. In a world that is characterized by an ever increasing globalization of the markets, driven by better communication methods, less restrictions for foreign trade, and a general internationalization of firms both with respect to purchasing and selling of goods and services, it must be examined if these results also hold in other regional and cultural environments. Studies in different countries would enable an insight into regional differences that could be a valuable insight for firms dealing with the reality of internationalization.

Appendix: Questionnaire

WHU-BVL-Studie: "Erfolg durch Logistik-Outsourcing" Prof. Dr. Jürgen Weber Dipl.-Kfm. David L. Cahill Dipl.-Kfm. Jan Deepen Kühne-Zentrum für Logistikmanagement WHU - Otto-Beisheim-Hochschule Burgplatz 2, 56179 Vallendar

Als Dankeschön für Ihre Teilnahme erhalten Sie: x einen individuellen Benchmarking-Bericht (Vergleich Ihres Unternehmens mit den Werten Ihrer Branche, der Ihnen nach Abschluss der Studie zugesandt wird.) sowie zusätzlich x eine Ausgabe des Buches "E-Commerce in der Logistik: Quantensprung oder business as usual?" (151 S.) von Prof. Dr. Jürgen Weber et al. oder x die kostenfreie Teilnahme am 3. WHU Logistiksymposium (2005) an der WHU

284

Appendix: Questionnaire

Bitte lesen Sie die folgenden Hinweise, bevor Sie den Fragebogen ausfüllen: x Alle Daten werden anonym ausgewertet und streng vertraulich behandelt x Bitte füllen Sie alle Fragen so gut wie möglich aus, auch wenn manche Fragen ähnlich erscheinen. Dies ist aus methodischen Gründen nötig. Wenn Sie die genaue Antwort nicht kennen, bitten wir Sie bewußt um Ihre subjektive Einschätzung. x Bitte beziehen Sie alle Fragen zu Ihrem Unternehmen immer auf Ihre Geschäftseinheit bzw. auf den Teilbereich des Unternehmens, für dessen Logistik Sie (mit)verantwortlich sind. x Falls Sie verschiedene Logistikdienstleister (LDL) benutzen, beziehen Sie alle Fragen auf den aus Ihrer Sicht für Ihr Unternehmen bzw. Ihre Geschäftseinheit wichtigsten LDL. x Der in dieser Studie verwendete Begriff Logistik-Outsourcing bezieht sich auf die dauerhafte Fremdvergabe logistischer Dienstleistungen an einen Logistikdienstleister. x Sie können die Beantwortung des Fragebogens nach jeder abgeschickten Seite unterbrechen und dann durch den in der Email enthaltenen Link fortsetzen. Die eingegebenen Daten werden automatisch nach jeder Seite gespeichert. Das Dankeschön gibt es für alle vollständig ausgefüllten Fragebögen. x Bitte verwenden Sie zur Navigation innerhalb des Fragebogens ausschließlich die "zurück" und "weiter" Felder, da es sonst zu technischen Problemen kommen kann.

Für Fragen stehen Ihnen Dipl.-Kfm. David Cahill oder Dipl.-Kfm. Jan Deepen gerne unter 0261-6509-471 oder [email protected] zur Verfügung!

Appendix: Questionnaire

285

A. Allgemeine Fragen zum Logistik-Outsourcing Ein wesentlicher Grund für uns, einen Teil unserer Logistik outzusourcen, war/waren: Die Senkung der Logistikkosten. Die Variabilisierung unserer Fixkosten. Der Spitzenausgleich bei Kapazitätsschwankungen. Die niedrigere Kapitalbindung. Der Geschwindigkeitsgewinn für unsere Logistik. Die flexibleren Abläufe und kürzeren Reaktionszeiten. Die geringere Schadens- und Fehlerquote. Die verbesserte Lieferfähigkeit. Die Tatsache, dass dieser LDL ein deutlich höheres Logistik--Know-how besitzt. Die Tatsache, dass Logistik für uns zu den eher unwichtigen Prozessen gehört. Der Engpass bei unseren Managementkapazitäten.

Trifft gar nicht zu

{ { { { { { { { {

Trifft voll zu

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { { { { { { {

Wichtig: Bitte beziehen Sie sich ab hier immer auf den für Sie wichtigsten Logistikdienstleister!

Welche Logistikleistungen führt der für Sie wichtigste Logistikdienstleister (LDL) mit welchem Schwerpunkt aus? Planung und Steuerung der Transporte Transportdurchführung Internationales Freight Forwarding (Frachtversand) Zollabfertigung Cross-Docking Lagerhaltung Lager-Management Verpacken, Kommissionieren Montagetätigkeiten Retourenabwicklung IT-Systeme der Logistik Logistik-Koordination (Lead Logistics Management) Beratungsleistungen

Grad der Durchführung Gar nicht Vollständig

Leistungs-schwerpunkt Be-schaffung

Produktion

Distribution

{ { { { { { { { { { { { { { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { { { { { {

{ { { { { { { { {

{ { { { { { { { {

{ { {

286

Appendix: Questionnaire

B. Zusammenarbeit mit Ihrem Logistikdienstleister Inwieweit treffen die folgenden Aussagen bzgl. Ihrer Zufriedenheit über die Zusammenarbeit mit Ihrem Logistikdienstleister (LDL) zu? Die Ziele und Erwartungen, die wir im Vorfeld der Zusammenarbeit festgelegt hatten, werden durch diesen LDL vollständig erfüllt. Wir sind sehr zufrieden mit diesem LDL. Die Zusammenarbeit mit diesem LDL kann man als sehr gut bezeichnen. Dieser LDL erbringt seine Leistungen in der geforderten Qualität. Dieser LDL erbringt seine Leistungen in der geforderten Zeit. Unsere Logistikkosten sind durch das Outsourcing in dem von uns vorhergesagten Ausmaß gesenkt worden. Mit der Art und Weise des Umgangs sind wir sehr zufrieden. Konflikte in der Zusammenarbeit mit diesem LDL werden immer reibungslos beigelegt. Die Beziehung zu diesem LDL kann man als sehr gut bezeichnen. Dieser LDL gibt eigene Einsparungen aus verbesserten Prozessabläufen etc. in angemessenem Umfang an uns weiter. Dieser LDL bereichert sich auf unsere Kosten. Wir fühlen uns durch diesen LDL fair behandelt. Wir profitieren in gleichem Maße vom Outsourcing wie dieser LDL.

Inwieweit stimmen Sie den folgenden Aussagen hinsichtlich Ihres Verhältnisses zu Ihrem Logistikdienstleister zu? Dieser LDL hält immer die Zusagen, die er uns macht. Dieser LDL ist bei auftretenden Problemen immer offen und ehrlich zu uns. Dieser LDL ist sehr vertrauenswürdig. Dieser LDL ist stark daran interessiert, dass wir erfolgreich sind. Auch Dinge, die wir nicht oder nur unter großem Aufwand kontrollieren können, erledigt dieser LDL Bei wichtigen Entscheidungen berücksichtigt der LDL auch unsere Interessen. Wir verteidigen diesen LDL, wenn er durch Mitglieder unseres Unternehmens oder durch Personen von Wir würden es persönlich sehr bedauern, wenn wir die Geschäftsbeziehung mit diesem LDL aufgeben Wir fühlen uns persönlich angegriffen, wenn dieser LDL durch Mitglieder unseres Unternehmens oder durch Wir haben den festen Willen, die Beziehung so lange wie möglich aufrechtzuerhalten.

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Appendix: Questionnaire Inwieweit stimmen Sie den folgenden Aussagen zur Kooperation mit diesem Logistikdienstleister zu? Die Ziele unserer Zusammenarbeit wurden von uns und diesem LDL gemeinsam festgelegt. Hinsichtlich der Art und Weise, zusammen zu arbeiten, Geschäfte zu machen und Projekte durchzuführen, Dieser LDL und wir ziehen in allen Belangen an einem Strang. Wenn auf übergeordneter Ebene noch Probleme oder Unklarheiten hinsichtlich der Outsourcing-Kooperation auftauchen, fällen wir notwendigen Entscheidungen gemeinsam mit dem LDL. Wenn im Rahmen unserer Geschäftsbeziehung einer der Partner seine Macht ausspielt, geschieht das in einer angemessenen Art und Weise. In unserer Geschäftsbeziehung respektieren sich beide Seiten vollkommen. Auch abseits der vorher festgelegten Zuständigkeiten arbeiten unsere Mitarbeiter mit dem LDL zusammen, um den Erfolg der Kooperation sicherzustellen. Die Kooperation mit diesem LDL funktioniert sehr gut. Inwieweit beschreiben die folgenden Aussagen die relevanten Teilschritte auf dem Weg zum Outsourcing? Einige Führungskräfte in unserem Unternehmen sind gegen die Outsourcing-Entscheidung. Auf Seiten der operativen Mitarbeiter in unserem Unternehmen gibt es viele Widerstände gegen das Outsourcing-Projekt. Zwischen dem LDL und uns treten immer wieder Konflikte auf der Managementebene auf. Auf der operativen Ebene kommt es häufig zu Problem oder Konflikten zwischen unseren Mitarbeitern und denen des LDL. Insgesamt ist die Beziehung zu diesem LDL sehr konfliktreich.

Trifft gar nicht zu

287

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

{ { { { { { { { { { { { { { { { { { { { {

{ { { { { { { Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

288

Appendix: Questionnaire

Inwieweit beschreiben die folgenden Aussagen die laufende Zusammenarbeit mit diesem Logistikdienstleister? Wir besprechen mögliche Probleme und Verbesserungen regelmäßig mit den verantwortlichen Mitarbeitern des LDL. Der Austausch von Informationen zwischen unseren Mitarbeitern und denen dieses LDL verläuft sehr gut. Um unsere gesteckten Ziele zu erreichen, sind viele Treffen und Gespräche mit diesem LDL notwendig. Wenn wir Informationen mit diesem LDL austauschen, sind diese immer relevant für das Vorankommen des Projektes bzw. unserer Geschäftsbeziehung. Der Austausch von Informationen zwischen diesem LDL und uns findet immer sofort statt, sobald sie zur Verfügung stehen. Auf die Informationen, die zwischen uns und diesem LDL ausgetauscht werden, können sich beide Seiten vollkommen verlassen. Die Art und Weise, wie wir mit diesem LDL Informationen austauschen, ist besonders gut dazu geeignet, Probleme in beiderseitigem Interesse zu lösen. Inwieweit stimmen Sie den folgenden Aussagen zu dem Verhältnis bzw. der Zusammenarbeit mit diesem Logistikdienstleister zu? Der LDL war frühzeitig und umfassend eingebunden. Für das mit diesem LDL durchgeführte Outsourcing ist ein funktionsübergreifendes Projektteam verantwortlich, in dem alle relevanten Funktionen unseres Unternehmens vertreten sind. Die auf beiden Seiten für das Outsourcing verantwortlichen Mitarbeiter arbeiten sehr gut zusammen. Den alltäglichen Austausch zwischen uns und diesem LDL kann man als informell bezeichnen. Wenn Fragen oder Probleme auftreten, können wir diesen LDL spontan kontaktieren und sofort gemeinsam beginnen, an Lösungen zu arbeiten. Die Beziehung zwischen uns und diesem LDL ist sehr offen. Wir tauschen mit diesem LDL auch sehr sensitive Daten aus, wenn wir uns dadurch einen Vorteil erhoffen. Wenn eine der beiden Parteien mit etwas nicht zufrieden ist, sagen wir uns das klar und deutlich. Der LDL ist immer vollkommen offen und ehrlich zu uns.

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

{ { { { { { { { { { { { { { { { { { { { {

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { {

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Appendix: Questionnaire Inwieweit stimmen Sie den folgenden Aussagen zu Ihrem Verhältnis bzw. der Zusammenarbeit mit diesem Logistikdienstleister zu? Dieser LDL und wir bewältigen auftretende Problem immer gemeinsam. Der Informationsaustausch bei der Behandlung von Problemen funktioniert sehr gut. Wenn Probleme in der Zusammenarbeit mit diesem LDL auftauchen, werden sie immer von derselben der beiden Parteien angefasst und beseitigt. Wenn es zwischen dem LDL und uns zu Problemen kommt, wird die Diskussion oft unsachlich und auch durch den Austausch von "Unfreundlichkeiten" geprägt.

Trifft gar nicht zu

289

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

{ { { { { { {

Probleme zwischen dem LDL und uns haben in der Vergangenheit die Produktivität unserer Beziehung stark beeinträchtigt. Probleme zwischen dem LDL und uns werden als Chance verstanden, die beiden Partnern die Möglichkeit zu Verbesserungen eröffnen.

{ { { { { { {

Inwieweit treffen die folgenden Aussagen auf die persönlichen Beziehungen zwischen Mitarbeitern Ihres Unternehmens und denen des Logistikdienstleisters zu? Wir pflegen viele persönliche Kontakte mit diesem LDL. Die Zusammenarbeit mit diesem LDL klappt auch auf der persönlichen Ebene sehr gut. Wenn wir die Beziehung zu diesem LDL beenden würden, würden ich oder einige meiner Mitarbeiter und Kollegen einen guten Geschäftsfreund verlieren. Wir haben enge persönliche Beziehungen zu Mitarbeitern dieses LDL.

Trifft gar nicht zu

Inwieweit treffen folgende Aussagen auf Ihren Hauptansprechpartner beim Logistikdienstleister zu? Diese Person ist sehr vertrauenswürdig. Diese Person macht keine falschen Versprechen. Diese Person besitzt großes Know-how und ist ein guter Manager. Diese Person spricht die gleiche Sprache wie wir. Diese Person berücksichtigt auch unsere Interessen und Bedürfnisse.

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Welche Position hat Ihr Hauptansprechpartner beim LDL? {

Geschäftsführer/Vorstand

{

Niederlassungsleiter

{

Bereichsleiter

{

Vertriebsleiter

{

Key Account Manager

{

Andere: __________________

290

Appendix: Questionnaire

Inwieweit treffen folgende Aussagen zur Verbesserung Ihrer Logistiksysteme durch diesen Logistikdienstleister zu? Der LDL arbeitet intensiv daran, die Logistikprozesse fortlaufend zu optimieren. Der LDL gibt uns laufend Anstöße zu Verbesserungen auch außerhalb seines direkten Zuständigkeitsbereiches. Bei veränderten Rahmenbedingungen modifiziert der LDL aus eigenem Antrieb die Logistiksysteme bzw. abläufe soweit sinnvoll und notwendig. Der LDL spricht uns aus Eigeninitiative mit Verbesserungsvorschlägen an. Dieser LDL ist generell sehr innovativ.

Trifft gar nicht zu

Inwieweit treffen folgende generelle Aussagen zu diesem Logistikdienstleister zu? Der LDL bietet insgesamt einen exzellenten Service. Die Leistungen des LDL sind hervorragend. Der LDL bietet eine sehr hohe Qualität.

Trifft gar nicht zu

Inwieweit treffen die folgenden Fragen bzgl. Ihrer Zufriedenheit über die Zusammenarbeit mit diesem Logistikdienstleister zu? Die Ziele und Erwartungen, die wir im Vorfeld der Zusammenarbeit festgelegt hatten, wurden in deutlichem Maße übertroffen. Unsere Anforderungen an die Qualität der Leistungen dieses LDL sind deutlich positiv übertroffen worden. Unsere Logistikkosten sind auf Grund der Zusammenarbeit mit diesem LDL deutlich niedriger als erwartet. Die tatsächlichen Kosten sind im Verhältnis zur erbrachten Gesamtleistung wesentlich besser, als wir es vorher erwartet hätten.

Trifft gar nicht zu

Inwieweit treffen folgende Aussagen zur bisherigen und zukünftigen Zusammenarbeit mit diesem Logistikdienstleister zu? Wir werden diesen LDL auch zukünftig weiter nutzen. Aus heutiger Sicht gehen wir davon aus, vorhandene Verträge mit dem LDL bei deren Auslaufen zu verlängern. Wenn wir mit unserem heutigen Wissen nochmals vor der ursprünglichen Entscheidung über die Zusammenarbeit mit diesem LDL stünden, würden wir die Geschäftsbeziehung erneut eingehen. Wir werden die Leistungen, die wir von diesem LDL in Anspruch nehmen, bei Auslaufen des Vertrages höchst wahrscheinlich nicht neu ausschreiben, sondern direkt mit diesem LDL verhandeln.

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { {

{ { { { { { {

Appendix: Questionnaire Inwieweit treffen folgende Aussagen zu Ihrer Einstellung gegenüber diesem Logistikdienstleister zu? Wir haben in der Organisation angeregt, diesen LDL für zukünftige Projekte bevorzugt zu berücksichtigen. Ich erwähne diesen LDL gegenüber Kollegen häufig sehr positiv. Ich empfehle diesen LDL auch nach außen hin häufig weiter. Wir empfehlen diesen LDL häufig weiter.

Trifft gar nicht zu

Inwieweit treffen folgende Aussagen bzgl. der Investitionen in die Geschäftsbeziehung zu? Wir haben viel Arbeitszeit (Mannjahre) investiert, die umsonst wäre, wenn wir den LDL wechseln würden. Wir haben hohe Sachinvestitionen getätigt, die erheblich an Wert verlieren würden, wenn wir den LDL wechselten. Wir haben hohe Investitionen in IT-Systeme getätigt, die erheblich an Wert verlieren würden, wenn wir den LDL wechselten. Wir haben sehr viel in die Weiterentwicklung des LDL investiert. Wir haben viel Arbeitszeit (Mannjahre) in die Integration unserer Prozesse mit denen des LDL investiert, die umsonst wäre, wenn wir den LDL wechseln würden.

Trifft gar nicht zu

Bitte geben Sie den Grad der Übereinsstimmung an, den Sie in den folgenden Fragen zwischen Ihrem Unternehmen und diesem Logistikdienstleister vermuten: Mitarbeiter, die nur auf den eigenen Vorteil bedacht sind anstatt den Vorteil der Firma zu verfolgen, sollten zurechtgewiesen werden. In Beziehungen sollten Unternehmen nicht nur auf den eigenen kurzfristigen Vorteil bedacht sein, sondern vielmehr den langfristigen Nutzen für die beteiligten Unternehmen im Auge haben. Die Mitarbeiter beider Seiten haben die Ziele der Zusammenarbeit vollständig verstanden und handeln auch danach. Die Ziele der Zusammenarbeit sind vollkommen klar definiert und werden von beiden Unternehmen in gleicher Art und Weise verfolgt und angestrebt. Das grundlegende Verständnis über die Art der Zusammenarbeit ist bei beiden Seiten sehr ähnlich und kompatibel.

Sehr niedrige Übereinstimmung

Inwieweit treffen folgende generelle Aussagen zu diesem Logistikdienstleister zu? Die Preise dieses LDL sind im Vergleich zu anderen LDL sehr günstig. Das Preis-Leistungs-Verhältnis bei diesem LDL ist sehr gut. Die Preise dieses LDL sind im Vergleich zur Eigenerstellung sehr günstig.

Trifft gar nicht zu

291

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Sehr hohe Übereinstimmung

{ { { { { { { { { { { { { {

{ { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { {

292

Appendix: Questionnaire

Inwieweit treffen folgende Aussagen zur bisherigen und zukünftigen Zusammenarbeit mit diesem Logistikdienstleister zu? In Zukunft wird dieser LDL einen größeren Anteil an unserem Auftragsvolumen erhalten. Beim Outsourcing anderer als der bisherigen Logistikleistungen werden wir diesen LDL bevorzugt berücksichtigen. Neue Leistungen werden wir zunächst diesem LDL anbieten, bevor wir sie ausschreiben. In den nächsten Jahren werden wir stärker auf diesen LDL zurückgreifen als bisher.

Trifft gar nicht zu

Inwieweit beschreiben die folgenden Aussagen die laufende Zusammenarbeit mit diesem Logistikdienstleister? Dieser LDL verändert manchmal Fakten so, dass sich seine Interessen besser begründen und vertreten lassen. Was der LDL uns gegenüber als seine momentanen Aktivitäten und Leistungen für uns angibt, entspricht immer der Realität. Um seine eigenen Ziele besser zu erreichen, verspricht uns der LDL manchmal Dinge, die er später überhaupt nicht einhält. Wenn der LDL mit uns über seine Bedürfnisse spricht, übertreibt er manchmal, um seine Ziele besser zu erreichen. Dieser LDL würde alles in seiner Macht stehende tun, um seine eigenen Ziele zu erreichen. Dieser LDL hat das Gefühl, dass sich im Umgang mit uns Ehrlichkeit lohnt.

Trifft gar nicht zu

Inwieweit treffen folgende Aussagen zu Ihrer Einstellung gegenüber diesem Logistikdienstleister zu? Meiner Meinung nach haben wir zuviel investiert, um einen Abbruch der Geschäftsbeziehung in Betracht zu ziehen. Auch wenn wir es wollten, wäre es sehr schwierig, die Geschäftsbeziehung zu diesem LDL aufzugeben. Wir fühlen uns verpflichtet, die Geschäftsbeziehung zu diesem LDL aufrechtzuerhalten. Dieser LDL hat zuviel in die Beziehung investiert, als dass wir ihm einen Abbruch der Geschäftsbeziehung zumuten könnten.

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

Appendix: Questionnaire Inwieweit treffen folgende Aussagen zu Alternativen und Ihrer Abhängigkeit von diesem Logistikdienstleister zu? Es gibt viele alternative LDL, die gleichwertige Leistungen bei gleichen Kosten anbieten können. Es gibt viele alternative LDL, die gleichwertige Leistungen bei höheren Kosten anbieten können. Es gibt viele alternative LDL, die gleichwertige Leistungen bei niedrigeren Kosten anbieten können. Dieser LDL beeinflusst die Qualität und Zuverlässigkeit unserer Logistikprozesse so sehr, dass sein Wegfall uns sehr stark treffen würde. Bei einem Wechsel des LDL würden insbesondere die neuen Verhandlungen sehr viel Aufwand verursachen. Die Suche und Auswahl eines gleichwertigen LDL ist sehr aufwendig. Wir können diesen LDL leicht ersetzen. Ein kurzfristiger Wechsel dieses LDL wäre mit sehr viel Aufwand verbunden.

Trifft gar nicht zu

Inwieweit beschreiben die folgenden Aussagen die spezifischen Investitionen, die für den Logistikdienstleister in Ihrem Fall notwendig waren? Um Logistikdienstleistungen für uns zu erbringen, musste der LDL hohe Investitionen (Transportmittel, Lagerhäuser o.ä.) tätigen, die er ausschließlich für uns verwenden kann. Der LDL kann die für uns entwickelten Geschäftsabläufe nicht ohne große Veränderungen für einen anderen Kunden verwenden. Damit der LDL Leistungen für uns erbringen kann, musste er seine Mitarbeiter so schulen, dass sie dieses Wissen bei einem anderen Kunden nicht verwenden könnten. Die Investitionen, die dieser LDL für unsere Beziehung in Menschen, Prozesse oder Vermögenswerte tätigen musste, werden sich sehr schnell amortisieren.

Trifft gar nicht zu

293

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { {

{ { { { { { { { { { { { { {

{ { { { { { {

$FKWXQJ%LWWHEHDQWZRUWHQ6LHGLHIROJHQGHQ)UDJHQLQ%H]XJDXI ,KU8QWHUQHKPHQXQG,KUH.XQGHQ

294

Appendix: Questionnaire

C. Fragen zu Ihrer Unternehmenssituation Inwieweit beschreiben die folgenden Aussagen die zur Zeit in Ihren Unternehmen vorherrschende Unsicherheit? Es fällt uns schwer, das künftige Verhalten unserer Kunden präzise vorherzusehen. Es kommt oft vor, dass sich Kunden uns gegenüber opportunistisch oder unfair verhalten. In unserer Branche oder der unserer Kunden gibt es aufgrund der Wettbewerbssituation zur Zeit eine spürbare Unsicherheit über die Zukunft. Es fällt uns schwer einzuschätzen, wie sich die Wünsche und Bedürfnisse unserer Kunden in Zukunft entwickeln werden.

Trifft gar nicht zu

Inwieweit beschreiben die folgenden Aussagen die momentane Situation Ihres Unternehmens? Unser Absatzvolumen ist von Kunde zu Kunde sehr unterschiedlich. Das Bestellverhalten unser Kunden ist sehr uneinheitlich und schwankt je nach Auftrag oder Saison beträchtlich. Verglichen mit unseren Wettbewerbern müssen unsere Produkte an eine hohe Anzahl von Lieferpunkten geliefert werden. Verglichen mit unseren Wettbewerbern greifen wir auf eine sehr große Anzahl von Logistikdienstleistern zurück. Der Verdrängungswettbewerb innerhalb unserer Branche ist sehr stark. In den Branchen unserer Kunden sind starke Konzentrationstendenzen zu beobachten. Die Bedeutung der Logistik nimmt in den Augen unserer Kunden immer weiter zu. Eine sehr gute Logistik wird von unseren Kunden zunehmend als wichtig empfunden und ist deswegen für uns ein großer Wettbewerbsvorteil.

Trifft gar nicht zu

Inwieweit beschreiben die folgenden Aussagen das Produktprogramm Ihres Unternehmens? Wir fertigen ein sehr breites Produktprogramm mit einer Reihe von unterschiedlichen Produkten. Unsere Produkte sind so wertvoll, dass sie besondere logistische Maßnahmen erfordern. Unsere Produkte sind so unterschiedlich, dass für sie sehr viele unterschiedliche Transport-, Umschlagund/oder Lagerprozesse konzipiert werden müssen. Für unsere Produkte gibt es echte Alternativen am Markt, so dass wir uns über eine erstklassige Logistik positiv herausstellen können.

Trifft gar nicht zu

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { {

Appendix: Questionnaire Inwieweit treffen die folgenden eher grundsätzlichen Aussagen zur Ausgestaltung der Logistik Ihrer Geschäftseinheit zu? Entscheidungen zu logistischen Fragen werden in unserem Unternehmen sehr zentral gefällt. Unsere Geschäftseinheit beherrscht einen reibungslosen, durchgängigen, schnellen und störungsarmen Material- und Informationsfluss. Unsere Geschäftseinheit wird insgesamt sehr flussbzw. prozessorientiert geführt. Unsere Geschäftseinheit besitzt einen deutlich höheren Grad an Fluss- bzw. Prozessorientierung als unsere Wettbewerber. Sämtliche Prozesse der Leistungserstellung sind in unserer Geschäftseinheit gut aufeinander abgestimmt. Es existieren in unserer Geschäftseinheit zahlreiche Einzelinteressen, die der Erreichung der Ziele unserer Geschäftseinheit im Wege stehen.

Trifft gar nicht zu

295

Trifft voll zu

{ { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { { {

D. Fragen zu Logistik- und Unternehmenserfolg Wie schätzen Sie die Logistikkosten Ihrer Geschäftseinheit Sehr viel schlechter im Vergleich zum Wettbewerb ein? Logistikkosten relativ zum Umsatz (inkl. Vergütung des { { { { { { LDL) Lagerkosten { { { { { { Transportkosten { { { { { { Durch Logistikprozesse hervorgerufene IT-Kosten { { { { { { Personalkosten der Logistik { { { { { {

Sehr viel besser

{ { { { {

Sehr viel Sehr viel Wie schätzen Sie die Logistikleistung Ihrer schlechter besser Geschäftseinheit im Vergleich zum Wettbewerb ein? Durchlaufzeiten { { { { { { { Lieferzeit { { { { { { { Lieferfähigkeit { { { { { { { Liefertreue { { { { { { { Lieferflexibilität und Reaktionszeiten (Zeit, Menge) { { { { { { { Schadens- und Fehlerfreiheit der logistischen Prozesse { { { { { { {

Wie schätzen Sie den Markterfolg Ihrer Geschäftseinheit im Vergleich zum Wettbewerb ein? Kundenzufriedenheit Kundennutzen Bindung bestehender Kunden Gewinnung/Akquisition von Neukunden Erreichung des angestrebten Wachstums Erreichung des angestrebten Marktanteils

Sehr viel schlechter

{ { { { { {

Sehr viel besser

{ { { { { {

{ { { { { {

{ { { { { {

{ { { { { {

{ { { { { {

{ { { { { {

296

Appendix: Questionnaire

Wie schätzen Sie die Flexibilität Ihrer Geschäftseinheit im Vergleich zum Wettbewerb ein? Anpassung der Produkte/Dienstleistungen an neue Kundenbedürfnisse Reaktion auf neue Entwicklungen am Markt Nutzung neuer Marktchancen

Sehr viel schlechter

Wie verhält sich Ihrer Einschätzung nach die Umsatzrendite Ihrer Geschäftseinheit im Vergleich zum Wettbewerb? Unsere Umsatzrendite war im letzten Geschäftsjahr im Vergleich zu unseren Wettbewerbern... Unsere Umsatzrendite war im Durchschnitt der letzten drei Geschäftsjahre im Vergleich zu der unserer Wettbewerber... Die Entwicklung unserer Umsatzrendite war in den letzten drei Jahren im Vergleich zu der unserer Wettbewerber...

Sehr viel schlechter

Sehr viel besser

{ { { { { { { { { { { { { { { { { { { { { Sehr viel besser

{ { { { { { { { { { { { { { { { { { { { {

Vielen Dank, dass Sie den Fragebogen bis hierhin ausgefüllt haben! Wir bitten Sie jetzt noch, einige statistische Fragen zu beantworten. Wie der gesamte Fragebogen unterliegen auch diese Fragen strengster Vertraulichkeit! Bitte beantworten Sie folgende Fragen zu Ihrer Person : In welcher Funktion sind Sie tätig? { Leiter Teilbereich der Logistik

{

Geschäftsführer

{

Leiter Logistik

{

Mitarbeiter Logistikbereich

{

Andere: ________________

Seit wie vielen Jahren sind Sie in dieser Position tätig?

____ Jahre

Seit wie vielen Jahren sind Sie in Ihrem Unternehmen tätig?

____ Jahre

Welcher Branche gehört Ihre Geschäftseinheit an? (Bitte nur eine Branche ankreuzen.)

{ { {

Nahrungs- und Genussmittel

{

Maschinen- und Apparatebau

{

Chemie/ Kunststoff

Fahrzeugbau

{

Elektrotechnik/ Feinmechanik/ Optik

{

Handel

Konsumgüter

{

andere: __________________

{

Health Care

Wie groß ist das Umsatzvolumen Ihrer Geschäftseinheit derzeit? (in Mio. € / Jahr) bis 10 { 11-24 { 25-50 { 50-100 100-250 { 250-500 { 500-1.000 { Über 1.000

{ {

Appendix: Questionnaire Wie viele Mitarbeiter arbeiten ungefähr für Ihre Geschäftseinheit? Wie viele Logistikdienstleister nutzt Ihre Geschäftseinheit insgesamt? Welchen Anteil an Ihren gesamten Logistikkosten haben externe Logistikdienstleister ungefähr (in %)?

297

____ Mitarbeiter ____ Logistikdienstleister ____ Prozent

Füllen Sie die folgenden Fragen bitte wieder für den gleichen Logistikdienstleister aus, für den Sie zuvor auch schon die Fragen beantwortet haben. Welchen Anteil am Gesamtvolumen der fremdvergebenen Logistikdienstleistungen Ihrer Geschäftseinheit hat dieser Logistikdienstleister? Seit wie vielen Jahren arbeiten Sie mit diesem Logistikdienstleister zusammen? Seit wie vielen Jahren arbeiten Sie mit diesem Logistikdienstleister so zusammen, dass Sie von einer engeren Geschäftsbeziehung sprechen würden? Wie lang ist die Gesamtlaufzeit des aktuellen Vertrages mit diesem Logistikdienstleister? Wie lang ist die Restlaufzeit dieses Vertrages mit diesem Logistikdienstleister?

____ Prozent

____ Jahre (z.B. 2,5) ____ Jahre (z.B. 2,5)

____ Jahre (z.B. 2,5) ____ Jahre (z.B. 2,5)

298

Appendix: Questionnaire

Vielen Dank, dass Sie an unserer Studie teilgenommen haben!

Sie können jetzt das Dankeschön für Ihre Teilnahme auswählen. Bitte wählen Sie hier das Dankeschön für Ihre Teilnahme aus. Alle Teilnehmer erhalten zusätzlich einen individuellen Benchmarking-Bericht. Ich habe als Dankeschön bereits am 2. WHU Logistiksymposium in Vallendar teilgenommen.

{

Ich hätte gerne das Buch: "E-Commerce in der Logistik: Quantensprung oder business as usual?" von Prof. Dr. Jürgen Weber et al..

{

Ich möchte am 3. WHU Logistiksymposium (2005) an der WHU teilnehmen.

{

Bitte füllen Sie die folgenden Felder aus, damit wir Sie bzgl. Ihres Dankeschöns kontaktieren können oder heften Sie eine Vistenkarte an. Bitte geben Sie auf jeden Fall Name und E-Mail-Adresse an. Falls Sie sich für das Buch entscheiden, benötigen wir auch Ihre Anschrift. Nachname:

___________________________________________

Vorname:

___________________________________________

E-Mail-Adresse:

___________________________________________

Unternehmen:

___________________________________________

Straße:

___________________________________________

Postleitzahl:

___________________________________________

Ort:

___________________________________________

Telefon:

___________________________________________

Falls Sie Anmerkungen oder Kritik haben, können Sie uns das im Folgenden gerne mitteilen: ___________________________________________________________ ___________________________________________________________ ___________________________________________________________ ___________________________________________________________

Vielen Dank für Ihre Mitarbeit!

List of figures

Fig. 2-1. Fig. 2-2. Fig. 2-3. Fig. 2-4. Fig. 3-1. Fig. 3-2. Fig. 3-3. Fig. 4-1. Fig. 4-2. Fig. 4-3. Fig. 4-4. Fig. 4-5. Fig. 4-6. Fig. 5-1. Fig. 5-2. Fig. 5-3. Fig. 5-4. Fig. 5-5. Fig. 5-6. Fig. 7-1. Fig. 7-2. Fig. 7-3. Fig. 7-4. Fig. 7-5. Fig. 7-6. Fig. 7-7.

The four phases of logistics development..................................10 Logistics development in Germany between 1999 and 2002 ....17 Performance effects of logistics.................................................18 Framework of logistics research ................................................53 The optimal form of organization depending on transaction dimensions .................................................................................59 Key mediating variables model of relationship marketing ........71 Relations between the organization and the environment from a logistical viewpoint ........................................................77 Antecedents of the performance of logistics outsourcing relationships ...............................................................................89 Conceptual logistics outsourcing performance model .............121 Conceptual logistics performance model.................................127 Firm Performance Model developed by Dehler and Engelbrecht ..............................................................................130 Conceptual firm performance model .......................................134 External and internal contingency variables ............................143 Represented industries and firm sizes in the sample................154 Motivation for logistics outsourcing........................................154 Generic measurement model of a latent variable with reflective indicators..................................................................157 Complete causal model consisting of two measurement models and one structural model .............................................158 Adaptation measures................................................................167 Threshold values for adaptation measures ...............................174 Basic model..............................................................................238 Simplified 8-factor model ........................................................241 Simplified 8-factor model with standardized parameter coefficients...............................................................................242 Hypotheses on the logistics performance model......................248 Logistics performance model...................................................249 Hypotheses on the firm performance model ............................252 Firm performance model .........................................................254

List of tables

Table 2-1. Key logistics outsourcing related research since 1999 ..........30 Table 3-1. Variables frequently used in relationship research ................67 Table 4-1. Overview of the hypotheses on the logistics outsourcing performance model ..............................................................119 Table 6-1. Indicators for the measurement of the construct cooperation...........................................................................182 Table 6-2. Adaptation measures for the construct cooperation (8 indicators) ........................................................................182 Table 6-3. Adaptation measures for the construct cooperation (4 items) ...............................................................................183 Table 6-4. Indicators for the measurement of the construct communication.....................................................................184 Table 6-5. Adaptation measures for the construct communication (6 indicators) ........................................................................185 Table 6-6. Adaptation measures for the construct communication (4 indicators) ........................................................................185 Table 6-7. Indicators for the measurement of the construct proactive improvement ........................................................186 Table 6-8. Adaptation measures for the construct proactive improvement (5 indicators)..................................................187 Table 6-9. Adaptation measures for the construct proactive improvement (4 indicators)..................................................187 Table 6-10. Indicators for the measurement of the construct trust..........188 Table 6-11. Adaptation measures for the construct trust (6 indicators) ........................................................................189 Table 6-12. Adaptation measures for the construct trust (4 indicators) ........................................................................190 Table 6-13. Indicators for the measurement of the construct commitment .........................................................................190 Table 6-14. Adaptation measures for the construct commitment (4 indicators) ........................................................................191 Table 6-15. Adaptation measures for the construct commitment (3 indicators) ........................................................................192

302

List of tables

Table 6-16. Indicators for the measurement of the construct functional conflict ................................................................193 Table 6-17. Adaptation measures for the construct functional conflict (5 indicators) ........................................................................194 Table 6-18. Adaptation measures for the construct functional conflict (3 indicators) ........................................................................195 Table 6-19. Indicators for the measurement of the construct involvement .........................................................................195 Table 6-20. Adaptation measures for the construct improvement (3 indicators) ........................................................................196 Table 6-21. Indicators for the measurement of the construct opportunism .........................................................................197 Table 6-22. Adaptation measures for the construct opportunism (6 indicators) ........................................................................198 Table 6-23. Adaptation measures for the construct opportunism (4 indicators) ........................................................................199 Table 6-24. Indicators for the measurement of the construct shared values ...................................................................................200 Table 6-25. Adaptation measures for the construct shared values (5 indicators) ........................................................................201 Table 6-26. Adaptation measures for the construct shared values (4 indicators) ........................................................................201 Table 6-27. Indicators for the measurement of the construct openness...............................................................................202 Table 6-28. Adaptation measures for the construct openness (6 indicators) ........................................................................203 Table 6-29. Adaptation measures for the construct openness (4 indicators) ........................................................................204 Table 6-30. Indicators for the measurement of the construct goal achievement .................................................................205 Table 6-31. Adaptation measures for the construct goal achievement (6 indicators) ........................................................................206 Table 6-32. Adaptation measures for the construct goal achievement (4 indicators) ........................................................................207 Table 6-33. Indicators for the measurement of the construct goal exceedance ...................................................................208 Table 6-34. Adaptation measures for the construct goal exceedance (4 indicators) ........................................................................209 Table 6-35. Adaptation measures for the construct goal exceedance (3 indicators) ........................................................................210 Table 6-36. Indicators for the measurement of the construct level of logistics services .....................................................211

List of tables

303

Table 6-37. Adaptation measures for the construct level of logistics services (6 indicators) ..........................................................212 Table 6-38. Adaptation measures for the construct level of logistics services (4 indicators) ..........................................................213 Table 6-39. Indicators for the measurement of the construct level of logistics costs .......................................................................214 Table 6-40. Adaptation measures for the construct level of logistics costs (5 indicators) ...............................................................214 Table 6-41. Adaptation measures for the construct level of logistics costs (3 indicators) ...............................................................215 Table 6-42. Indicators for the measurement of the construct adaptiveness .........................................................................216 Table 6-43. Adaptation measures for the construct adaptiveness (3 indicators) ........................................................................217 Table 6-44. Indicators for the measurement of the construct market performance .........................................................................217 Table 6-45. Adaptation measures for the construct market performance (6 indicators)...................................................218 Table 6-46. Adaptation measures for the construct market performance (4 indicators)...................................................219 Table 6-47. Indicators for the measurement of the construct financial performance .........................................................................219 Table 6-48. Adaptation measures for the construct market performance (3 indicators)...................................................220 Table 6-49. Adaptation measures for the model logistics outsourcing performance .........................................................................221 Table 6-50. Discriminant validity of the antecedents and dimensions of logistics outsourcing performance...................................223 Table 6-51. Adaptation measures for the model logistics performance .........................................................................224 Table 6-52. Discriminant validity of the dimensions of logistics outsourcing performance and logistics performance ...........225 Table 6-53. Adaptation measures for the model firm performance ........225 Table 6-54. Discriminant validity of the dimensions of logistics performance and firm performance .....................................226 Table 6-55. Indicators for the measurement of the construct environmental complexity ...................................................227 Table 6-56. Indicators for the measurement of the construct environmental dynamics ......................................................228 Table 6-57. Indicators for the measurement of the construct uncertainty ...........................................................................229

304

List of tables

Table 6-58. Adaptation measures for the construct uncertainty (4 indicators) ........................................................................230 Table 6-59. Adaptation measures for the construct uncertainty (3 indicators) ........................................................................230 Table 6-60. Industries of the participating firms.....................................231 Table 6-61. Indicators for the measurement of the construct products ...231 Table 6-62. Indicators for the measurement of the construct asset specificity.............................................................................233 Table 6-63. Adaptation measures for the construct asset specificity (3 indicators) ........................................................................233 Table 6-64. Indicators for the measurement of the construct process orientation ...............................................................234 Table 6-65. Adaptation measures for the construct process orientation (5 indicators) ........................................................................235 Table 6-66. Adaptation measures for the construct process orientation (4 indicators) ........................................................................235 Table 7-1. Adaptation measures of the basic model .............................238 Table 7-2. Fit comparison of the basic and the simplified model .........240 Table 7-3. Hypotheses for the logistics outsourcing performance model ...................................................................................246 Table 7-4. Total effects in the logistics outsourcing performance model ...................................................................................247 Table 7-5. Adaptation measures of the logistics performance model ...248 Table 7-6. Hypotheses for the logistics performance model.................251 Table 7-7. Adaptation measures of the firm performance model..........252 Table 7-8. Hypotheses for the firm performance model .......................255 Table 7-9. Total effects in the firm performance model .......................256 Table 7-10. Moderating effects of external contingency variables on the logistics outsourcing performance model ......................258 Table 7-11. Causal linkages moderated by external contingency variables in the logistics outsourcing performance model...259 Table 7-12. Moderating effects of internal contingency variables on the logistics outsourcing performance model ......................262 Table 7-13. Causal linkages moderated by internal contingency variables in the logistics outsourcing performance model...263 Table 7-14. Moderating effects of external contingency variables on the logistics performance model ..........................................264 Table 7-15. Causal linkages moderated by external contingency variables in the logistics performance model.......................265 Table 7-16. Moderating effects of internal contingency variables on the logistics performance model ..........................................266

List of tables

305

Table 7-17. Moderating effects of external contingency variables on the firm performance model.................................................267 Table 7-18. Causal linkages moderated by external contingency variables in the firm performance model .............................268 Table 7-19. Moderating effects of internal contingency variables on the firm performance model.................................................270 Table 7-20. Causal linkages moderated by internal contingency variables in the firm performance model .............................271

References

Achrol, R. S. (1991): Evolution of the marketing organization: New forms for turbulent environments, in: Journal of Marketing, Vol. 55 (4): pp. 77-93. Ackerman, K. B. (1996): Pitfalls in logistics partnerships, in: International Journal of Physical Distribution & Logistics Management, Vol. 26 (3): pp. 35-37. Akaike, H. (1974): A New Look at the Statistical Model Identification, in: IEEE Transactions on Automatic Control, Vol. 19 (6): pp. 716-723. Akaike, H. (1987): Factor analysis and AIC, in: Psychometrika, Vol. 52 (3): pp. 317-332. Akerlof, G. A. (1970): The Market for "Lemons": Quality Uncertainty and the Market Mechanism, in: Quarterly Journal of Economics, Vol. 84 (3): pp. 488-500. Alchian, A. A. (1965): Some Economics of Property Rights, in: Il Politico, Vol. 30 (4): pp. 816-829. Alchian, A. A. and H. Demsetz (1972): Production, Information Costs, and Economic Organization, in: American Economic Review, Vol. 62 (5): pp. 777-795. Alderson, W. (1965): Dynamic Marketing Behavior. Homewood, IL. Allen, N. J. and J. P. Meyer (1990): The Measurement and Antecedents of Affective, Continuance and Normative Commitment to the Organization, in: Journal of Occupational Psychology, Vol. 63 (1): pp. 1-18. Alt, R. and B. Schmid (2000): Logistik und E-Commerce – Perspektiven durch zwei sich wechselseitig ergänzende Konzepte, in: Zeitschrift für Betriebswirtschaft, Vol. 70 (1): pp. 75-99.

308

References

Ambler, T. and F. Kokkinaki (1997): Measures of Marketing Success, in: Journal of Marketing Management, Vol. 13 (7): pp. 665-678. Amit, R. and P. J. H. Schoemaker (1993): Strategic Assets and Organizational Rent, in: Strategic Management Journal, Vol. 14 (1): pp. 33-46. Anderson, E., L. M. Lodish and B. A. Weitz (1987): Resource Allocation Behavior in Conventional Channels, in: Journal of Marketing Research, Vol. 24 (February 1987): pp. 85-97. Anderson, E. and B. Weitz (1989): Determinants of Continuity in Conventional Industrial Channel Dyads, in: Marketing Science, Vol. 8 (4): pp. 310-323. Anderson, E. and B. Weitz (1992): The Use of Pledges to Build and Sustain Commitment in Distribution Channels, in: Journal of Marketing Research, Vol. 29 (February 1992): pp. 18-34. Anderson, E. W. and M. W. Sullivan (1993): The Antecedents and Consequences of Customer Satisfaction for Firms, in: Marketing Science, Vol. 12 (2): pp. 125-143. Anderson, J. C. and D. W. Gerbing (1988): Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach, in: Psychological Bulletin, Vol. 103 (3): pp. 411-423. Anderson, J. C., D. W. Gerbing and J. E. Hunter (1987): On the Assessment of Unidimensional Measurement: Internal and External Consistency, and Overall Consistency Criteria, in: Journal of Marketing Research, Vol. 24 (4): pp. 432-437. Anderson, J. C., H. Hakansson and J. Johanson (1994): Dyadic business relationships within a business network context, in: Journal of Marketing, Vol. 58 (4): pp. 1-15. Anderson, J. C. and J. A. Narus (1984): A Model of the Distibutor's Perspective of Distributor-Manufacturer Working Relationsships, in: Journal of Marketing, Vol. 48 (Fall 1984): pp. 62-74.

References

309

Anderson, J. C. and J. A. Narus (1990): A Model of Distributor Firm and Manufacturer Firm Working Partnerships, in: Journal of Marketing, Vol. 54 (1): pp. 42-58. Angle, H. L. and J. L. Perry (1981): An Empirical Assessment of Organizational Commitment and Organizational Effectiveness, in: Administrative Science Quarterly, Vol. 26 (1): pp. 1-14. Antlitz, A. (1999): Unternehmensgrenzen und Kooperationen: Makecooperate-or-buy im Zusammenspiel von Kompetenz- und Strategieentwicklung. Wiesbaden. Aristotle (1999): Nicomachean Ethics. Indianapolis. Armstrong, J. S. and T. S. Overton (1977): Estimating Nonresponse Bias in Mail Surveys, in: Journal of Marketing Research, Vol. 14 (3): pp. 396-402. Arrow, K. J. (1970): The Organization of Economic Activity: Issues Pertinent to the Choice of Market Versus Nonmarket Allocation, in: R. H. Haveman and J. Margolis: Public Expenditure and Policy Analysis, 2. Edition. Chicago: pp. 67-81. Arrow, K. J. (1975): Gifts and Exchanges, in: E. Phelps: Altruism, Morality and Economic Theory. New York: pp. 13-28. Ashford, S. J., N. P. Rothbard, S. K. Piderit and J. E. Dutton (1998): Out On a Limb: The Role of Context and Impression Management in Selling Gender-Equity Issues, in: Administrative Science Quarterly, Vol. 43 (1): pp. 23-57. Assael, H. (1969): Constructive Role of Interorganizational Conflict, in: Administrative Science Quarterly, Vol. 14 (4): pp. 573-583. Assael, H. (1987): Consumer Behavior and Marketing Action, 3. Edition. Boston, MA. Axelrod, R. (1984): The evolution of cooperation. New York. Backhaus, K., B. Erichson, W. Plinke and R. Weiber (2003): Multivariate Analysemethoden - Eine anwendungsorientierte Einführung, 10. Edition. Berlin, Heidelberg, New York.

310

References

Bagchi, P. K. and H. Virum (1996): European Logistics Alliances: A Management Model, in: International Journal of Logistics Management, Vol. 7 (1): pp. 93-108. Bagozzi, R. P. (1979): The Role of Measurement in Theory Construction and Hypothesis Testing: Toward a Holistic Model, in: O. C. Ferrel, S. Brown and C. Lamb: Conceptual and Theoretical Developments in Marketing. Chicago: pp. 15-32. Bagozzi, R. P. (1980): Causal Models in Marketing. New York et al. Bagozzi, R. P. (1981a): Causal Modeling: A General Method for Developing and Testing Theories in Consumer Research, in: K. B. Monroe: Advances in Consumer Research, Vol. VIII. Ann Arbor, MI: pp. 195202. Bagozzi, R. P. (1981b): An Examination of the Validity of Two Models of Attitude, in: Multivariate Behavioral Research, Vol. 16 (3): pp. 323-359. Bagozzi, R. P. (1994): Structural Equation Models in Marketing Research: Basic Principles, in: R. P. Bagozzi: Principles of Marketing Research. Cambridge, MA: pp. 317-385. Bagozzi, R. P. and H. Baumgartner (1994): The Evaluation of Structural Equation Models and Hypothesis Testing, in: R. P. Bagozzi: Principles of Marketing Research. Cambridge, MA: pp. 386-422. Bagozzi, R. P. and L. W. Phillips (1982): Representing and Testing Organizational Theories: A Holistic Construal, in: Administrative Science Quarterly, Vol. 27 (3): pp. 459-489. Bagozzi, R. P. and Y. Yi (1988): On the Evaluation of Structural Equation Models, in: Journal of the Academy of Marketing Science, Vol. 16 (1): pp. 74-94. Bagozzi, R. P., Y. Yi and L. W. Phillips (1991): Assessing Construct Validity in Organizational Research, in: Administrative Science Quarterly, Vol. 36 (3): pp. 421-458. Balling, R. (1997): Kooperationen - Strategische Allianzen, Netzwerke, Joint-Ventures und andere Organisationsformen zwischenbetrieblicher

References

311

Zusammenarbeit in Theorie und Praxis. Frankfurt am Main, Berlin, Bern, New York, Paris, Wien. Ballou, R. H. (1992): Business Logistics Management, 3. Edition. Englewood Cliffs. Bardi, E. J. and M. Tracey (1991): Transportation Outsourcing: A Survey of US Practices, in: International Journal of Physical Distribution & Logistics Management, Vol. 21 (3): pp. 15-21. Baumgarten, H., A. Bott and T. Hagen (1997): Logistik und Kosten Ergebnisse der Untersuchung Trends und Strategien in der Logistik 2000. Berlin. Baumgarten, H. and J. Thoms (2002): Trends und Strategien in der Logistik - Supply Chains im Wandel. Berlin. Baumgarten, H. and S. Walter (2000): Trends und Strategien in der Logistik 2000+. Berlin. Baumgartner, H. and C. Homburg (1996): Applications of structural equation modeling in marketing and consumer research: A review, in: International Journal of Research in Marketing, Vol. 13 (2): pp. 139161. Bearden, W. O., S. Sharma and J. E. Teel (1982): Sample Size Effects on Chi Square and Other Statistics Used in Evaluating Causal Models, in: Journal of Marketing Research, Vol. 19 (4): pp. 425-430. Bechtel, C. and J. Jayaram (1997): Supply Chain Management: A Strategic Perspective, in: International Journal of Logistics Management, Vol. 8 (1): pp. 15-34. Becker, H. (1960): Notes on the Concept of Commitment, in: American Journal of Sociology, Vol. 66: pp. 32-42. Bennett, A. S. (1945): Some Aspects of Preparing Questionnaires, in: Journal of Marketing, Vol. 10 (2): pp. 175-179. Bennett, R. (1996): Relationship Formation and Governance in Consumer Markets: Transactional Analysis Versus the Behaviourist Approach, in: Journal of Marketing Management, Vol. 12 (5): pp. 417-436.

312

References

Bentler, P. M. (1990): Fit Indexes, Lagrange Multipliers, Constraint Changes, and Incomplete Data in Structural Models, in: Multivariate Behavioral Research, Vol. 25 (2): pp. 163-172. Bentler, P. M. and D. G. Bonett (1980): Significance Tests and Goodness of Fit in the Analysis of Covariance Structures, in: Psychological Bulletin, Vol. 88 (3): pp. 588-606. Bentler, P. M. and C.-P. Chou (1987): Practical Issues in Structural Modeling, in: Sociological Methods & Research, Vol. 16 (1): pp. 78117. Bentler, P. M. and A. Mooijaart (1989): Choice of Structural Model via Parsimony: A Rationale based on Precision, in: Psychological Bulletin, Vol. 106: pp. 315-317. Berglund, M., P. van Laarhoven, G. Sharman and S. Wandel (1999): Third-Party Logistics: Is There a Future?, in: International Journal of Logistics Management, Vol. 10 (1): pp. 59-70. Berry, L. L. and A. Parasuraman (1991): Marketing Services: Competing through Quality. New York. Bhargava, M., C. Dubelaar and S. Ramaswami (1994): Reconciling Diverse Measures of Performance: A Conceptual Framework and Test of a Methodology, in: Journal of Business Research, Vol. 31 (2/3): pp. 235246. Bhatnagar, R., A. S. Sohal and R. Millen (1999): Third party logistics services: a Singapore perspective, in: International Journal of Physical Distribution & Logistics Management, Vol. 29 (9): pp. 569-587. Blalock, H. M. (1963): Making Causal Inferences for Unmeasured Variables from Correlations among Indicators, in: American Journal of Sociology, Vol. 69: pp. 53-62. Blau, P. M. (1960): A Theory of Social Integration, in: The American Journal of Sociology, Vol. 65 (6): pp. 545-556. Blau, P. M. (1964): Exchange and Power in Social Life. New York.

References

313

Blau, P. M. (1968): Interaction: Social Exchange, in: D. L. Sills: International Encyclopedia of the Social Sciences. New York: pp. 452-458. Bollen, K. and R. Lennox (1991): Conventional wisdom on measurement: A structural equation perspective, in: Psychological Bulletin, Vol. 110 (2): pp. 305-314. Bollen, K. A. (1990): Overall Fit in Covariance Structure Models: Two Types of Sample Size Effects, in: Psychological Bulletin, Vol. 107 (2): pp. 256-259. Bollen, K. A. and K. H. Barb (1981): Pearsons's r and coarsely categorized measures, in: American Sociological Review, Vol. 46 (2): pp. 232239. Bonoma, T. V. (1976): Conflict, Cooperation and Trust in Three Power Systems, in: Behavioral Science, Vol. 21: pp. 499-514. Boomsma, A. (1982): The Robustness of LISREL against Small Sample Sizes in Factor Analysis Models, in: K. G. Jöreskog and H. Wold: Systems Under Direct Observation. Amsterdam, New York, Oxford: pp. 149-173. Boss, R. W. (1978): Trust and Managerial Problem Solving Revisited, in: Group & Organization Studies, Vol. 3 (3): pp. 331-342. Bovet, D. (1991): Logistics Strategies for Europe in the Nineties, in: Planning Review, Vol. 19 (4): pp. 12-15 and 46-48. Bowersox, D. J. (1990): The Strategic Benfits of Logistics Alliances, in: Harvard Business Review, Vol. 86 (4): pp. 36-45. Bowersox, D. J., D. J. Closs and O. K. Helferich (1986): Logistical Management: A Systems Integration of Physical Distribution, Manufacturing Support, and Materials Procurement. New York. Bowersox, D. J., D. J. Closs and T. P. Stank (2003): How to Master Cross-Enterprise Collaboration, in: Supply Chain Management Review, Vol. 7 (4): pp. 18-27.

314

References

Bowersox, D. J. and P. J. Daugherty (1987): Emerging Patterns of Logistical Organization, in: Journal of Business Logistics, Vol. 8 (1): pp. 46-60. Bowersox, D. J., P. J. Daugherty, C. R. Droge, D. S. Rogers and D. L. Wardlow (1989): Leading Edge Logistics: Competitive Logistics for the 1990's. Oak Brook, IL. Boyle, B., F. R. Dwyer, R. A. Robicheaux and J. T. Simpson (1992): Influence Strategies in Marketing Channels: Measures and Use in Differen Relationship Structures, in: Journal of Marketing, Vol. 29: pp. 462-473. Boyson, S., T. M. Corsi, M. Dresner and E. Rabinovich (1999): Managing effective third party logistics relationships: What does it take?, in: Journal of Business Logistics, Vol. 20 (1): pp. 73-98. Bozdogan, H. (1987): Model Selection and Akaike's Information Criterion (AIC): The General Theory and its Analytical Extensions, in: Psychometrika, Vol. 52 (3): pp. 345-370. Bradley, P. (1994a): Contract Logistics: It's All About Costs, in: Purchasing (October 1994): pp. 56A3-A14. Bradley, P. (1994b): Cozy up, but stay tough, in: Purchasing (March 17th, 1994): pp. 47-51. Bretzke, W.-R. (1993a): Outcontracting logistischer Leistungen - Voraussetzungen, Potentiale und Grenzen, in: Zeitschrift für Logistik, Vol. 15 (1): pp. 8-12. Bretzke, W.-R. (1993b): Pro und Contra Outsourcing von Logistikdienstleistungen, in: Beschaffung aktuell (6/93): pp. 37-39. Brown, J. R. (1981): A Cross-Channel Comparison Of Supplier-Retailer Relations, in: Journal of Retailing, Vol. 57 (4): pp. 3-18. Browne, M. (1984): Asymptotically Distribution-Free Methods for the Analysis of Covariance Structures, in: British Journal of Mathematical and Statistical Psychology, Vol. 37: pp. 62-83.

References

315

Browne, M. and J. Allen (2001): Logistics Out-Sourcing, in: A. M. Brewer, K. J. Button and D. A. Hensher: Handbook of Logistics and Supply Chain Management. Oxford. Browne, M. W. and R. Cudeck (1989): Single Sample Cross-Validation Indices for Covariance Structures, in: Multivariate Behavioral Research, Vol. 24 (4): pp. 445-455. Browne, M. W. and R. Cudeck (1993): Alternative Ways of Assessing Model Fit, in: K. A. Bollen and J. S. Long: Testing Structural Equation Models. Newbury Park, London, New Delhi: pp. 136-162. Burns, T. and G. Stalker (1961): The Management of Innovation. Oxford, New York. Buzzel, R. D. and B. T. Gale (1987): The PIMS Principles: Linking Strategy to Performance. New York. Byrne, B. M. (1994): Structural Equation Modeling with EQS and EQS/Windows. Thousand Oaks. Byrne, B. M. (2001): Structural Equation Modeling With AMOS - Basic Concepts, Applications, and Programming. Mahwah, London. Byrne, P. (1993): A new road map for contract logistics, in: Transportation & Distribution, Vol. 34 (4): pp. 58-62. Cahill, D. L. (2006): Customer Loyalty in Third Party Logistics Relationships: Findings from Studies in the USA and Germany - Forthcoming Dissertation. Vallendar. Caldwell, D. F. and C. A. O'Reilly (1990): Measuring Person-Job Fit With a Profile-Comparison Process, in: Journal of Applied Psychology, Vol. 75 (6): pp. 648-657. Calfee, J. E. and P. H. Rubin (1993): Nontransactional Data in Managerial Economics and Marketing, in: Managerial & Decision Economics, Vol. 14 (2): pp. 163-173. Caplow, T. (1956): The Effect of Increasing Size on Organizational Structure in Industry, in: Transactions of the Third World Congress of Sociology, Vol. 2: pp. 157-164.

316

References

Capon, N., J. U. Farley and S. Hoenig (1990): Determinants of Financial Performance: A Meta-Analysis, in: Management Science, Vol. 36 (10): pp. 1143-1159. Carmines, E. G. and R. A. Zeller (1979): Reliability and Validity Assessment. Newbury Park, London, New Dehli. Cavinato, J. L. (1989): The Logistics of Contract Manufacturing, in: International Journal of Physical Distribution & Materials Management, Vol. 19 (1): pp. 13-21. Cavusgil, S. T. and L. A. Elvey-Kirk (1998): Mail survey response behavior, in: European Journal of Marketing, Vol. 32 (11/12): pp. 9391192. Chadwick-Jones, J. K. (1976): Social Exchange Theory: Its Structure and Influence in Social Psychology. London, New York, San Francisco. Chatman, J. A. (1991): Matching People and Organizations: Selection and Socialization in Public Accounting Firms, in: Administrative Science Quarterly, Vol. 36 (3): pp. 459-484. Child, J. (1973): Organizations: A Choice for Man, in: J. Child: Man and Organization. London: pp. 234-257. Child, J. (1984): Organization - A Guide to Problems and Practice, 2. Edition. London. Chow, G., T. D. Heaver and L. E. Henriksson (1994): Logistics Performance: Definition and Measurement, in: International Journal of Physical Distribution & Logistics Management, Vol. 24 (1): pp. 17-28. Chow, G., T. D. Heaver and L. E. Henriksson (1995): Strategy, Structure and Performance: A Framework for Logistics Research, in: Logistics and Transportation Review, Vol. 31 (4): pp. 285-308. Churchill, G. A. (1979): A Paradigm for Developing Better measures of Marketing Constructs, in: Journal of Marketing Research, Vol. 16 (1): pp. 64-73.

References

317

Churchill, G. A. (1991): Marketing Research - Methodological Foundations, 5. Edition. Orlando. Churchill, G. A. and J. P. Peter (1984): Research Design Effects on the Reliability of Rating Scales: A Meta-Analysis, in: Journal of Marketing Research, Vol. 21 (4): pp. 360-375. Claycomb, C. and G. L. Frankwick (2004): A Contingency Perspective of Communication Conflict Resolution and Buyer Search Effort in Buyer Supplier Relationships, in: Journal of Supply Chain Management, Vol. 40 (1): pp. 18-34. Coase, R. H. (1937): The Nature of the Firm, in: Economica, Vol. 4: pp. 386-405. Coleman, J. S. (1990): Foundations of Social Theory. Cambridge. Cook, K. S. and R. M. Emerson (1978): Power, Equity and Commitment in Exchange Networks, in: American Sociological Review, Vol. 43 (5): pp. 721-739. Cooper, J. C. (1993): Logistics Strategies for Global Businesses, in: International Journal of Physical Distribution & Logistics Management, Vol. 23 (4): pp. 12-23. Corsten, D. (2003): Zur empirischen Methode in der Logistikforschung Bestandsaufnahme und Leitfaden zur Konzeptualisierung und Operationalisierung von Logistikkonstrukten, in: Logistik Management, Vol. 5 (3): pp. 49-60. Cortina, J. M. (1993): What Is Coefficient Alpha? An Examination of Theory and Applications, in: Journal of Applied Psychology, Vol. 78 (1): pp. 98-104. Cote, J. (2001): Structural Equations Modeling - VI.C. Improving Model Fit by Correlating Errors, in: Journal of Consumer Psychology, Vol. 10 (1/2): pp. 83-100. Coyle, J. J., E. J. Bardi and C. J. Langley (1992): The Management of Business Logistics, 5. Edition. St. Paul.

318

References

Cravens, D. W. (1995): The changing role of the sales force, in: Marketing Management, Vol. 4 (2): pp. 48-57. Cronbach, L. J. (1951): Coefficient Alpha and the Internal Structure of Tests, in: Psychometrika, Vol. 16 (3): pp. 297-334. Crutchfield, T. N. (2001): The Effect of Trust and Commitment on Retention of High-Risk Professional Service Customers, in: Services Marketing Quarterly, Vol. 22 (2): pp. 17-27. CSCMP (2005): Supply Chain and Logistics Terms and Glossary. Bellevue. Cunningham, M. T. and P. W. Turnbull (1982): Inter-Organizational Personal Contact Patterns, in: H. Hakansson: International Marketing and Purchasing of Industrial Goods. New York: pp. 304-316. Dalton, D. R., W. D. Todor, M. J. Spendolini, G. J. Fielding and L. W. Porter (1980): Organization Structure and Performance: A Critical Review, in: Academy of Management Review, Vol. 5 (1): pp. 49-64. Dasgupta, P. (1988): Trust as a Commodity, in: D. Gambetta: Trust: Making and Breaking Cooperative Relations. New York: pp. 49-72. Datamonitor (2000): German Logistics Client Survey. Frankfurt. Daugherty, P. J., T. P. Stank and A. E. Ellinger (1998): Leveraging Logistics/Distribution Capabilities: The Effect of Logistics Service on Market Share, in: Journal of Business Logistics, Vol. 19 (2): pp. 35-51. Deepen, J. M. (2003): Die Rolle von Logistikdienstleistern im Supply Chain Management - Eine konzeptionell-empirische Betrachtung, in: J. Weber and J. M. Deepen: Erfolg durch Logistik - Erkenntnisse aktueller Forschung. Bern, Stuttgart, Wien. Dehler, M. (2001): Entwicklungsstand der Logistik - Messung - Determinanten - Erfolgswirkungen. Wiesbaden. Demsetz, H. (1967): Toward a Theory of Property Rights, in: American Economic Review, Vol. 57 (2): pp. 347-359.

References

319

Dess, G. G. and R. B. Robinson Jr. (1984): Measuring Organizational Performance in the Absence of Objective Measures: The Case of the Privately-held Firm and Conglomerate Business Unit, in: Strategic Management Journal, Vol. 5 (3): pp. 265-273. Deutsch, M. (1960): The Effect of Motivational Orientation on Trust and Suspicion, in: Human Relations, Vol. 13: pp. 123-139. Deutsch, M. (1969): Conflicts: Productive and Destructive, in: Journal of Social Issues, Vol. 25 (1): pp. 7-41. Donaldson, L. (1990): The Ethereal Hand: Organizational Economics and Management Theory, in: Academy of Management Review, Vol. 15 (3): pp. 369-381. Doney, P. M. and J. P. Cannon (1997): An examination of the nature of trust in buyer-seller relationships., in: Journal of Marketing, Vol. 61 (2): pp. 35-51. Duffy, R. and A. Fearne (2004): The Impact of Supply Chain Partnerships on Supplier Performance, in: International Journal of Logistics Management, Vol. 15 (1): pp. 57-71. Duncan, O. D. (1966): Path Analysis: Sociological Examples, in: American Journal of Sociology, Vol. 72 (1): pp. 1-16. Duncan, O. D., A. O. Haller and A. Portes (1968): Peer Influences on Aspirations: A Reinterpretation, in: American Journal of Sociology, Vol. 74 (2): pp. 119-137. Durvasula, S., J. C. Andrews, S. Lysonski and R. G. Netemeyer (1993): Assessing the Cross-national Applicability of Consumer Behavior Models: A Model of Attitude toward Advertising in General, in: Journal of Consumer Research, Vol. 19 (4): pp. 626-636. Dwyer, F. R. (1980): Channel-Member Satisfaction: Laboratory Insights, in: Journal of Retailing, Vol. 56 (2): pp. 45-65. Dwyer, F. R., P. H. Schurr and S. Oh (1987): Developing Buyer-Seller Relationships, in: Journal of Marketing, Vol. 51 (April 1987): pp. 11-27.

320

References

Dyer, J. H. (1997): Effective Interfirm Collaboration: How Firms Minimize Transaction Costs and Maximize Transaction Value, in: Strategic Management Journal, Vol. 18 (7): pp. 535-556. Dyer, J. H., D. S. Cho and W. Chu (1998): Strategic supplier segmentation: The next 'best practice' in supply chain management, in: California Management Review, Vol. 40 (2): pp. 57-77. Ellram, L. M. (1992): International Purchasing Alliances: An Empirical Study, in: The International Journal of Logistics Management, Vol. 3 (1): pp. 23-36. Ellram, L. M. (1995a): A Managerial Guideline for the Development and Implementation of Purchasing Partnerships, in: International Journal of Purchasing and Materials Management, Vol. 31 (2): pp. 9-16. Ellram, L. M. (1995b): Partnering Pitfalls and Success Factors, in: International Journal of Purchasing and Materials Management, Vol. 31 (2): pp. 35-44. Ellram, L. M. and M. C. Cooper (1990): Supply Chain Management, Partnerships, and the Shipper-Third Party Relationship, in: The International Journal of Logistics Management, Vol. 1 (2): pp. 1-10. Ellram, L. M. and T. E. Hendrick (1995): Partnering characteristics: A dyadic perspective, in: Journal of Business Logistics, Vol. 16 (1): pp. 41-64. Emerson, C. J. and C. M. Grimm (1998): The relative importance of logistics and marketing customer service: A strategic perspective, in: Journal of Business Logistics, Vol. 19 (1): pp. 17-32. Emerson, R. M. (1962): Power-Dependence Relations, in: American Sociological Review, Vol. 27 (February 1962): pp. 31-41. Engelbrecht, C. (2003): Logistik-Outsourcing: Erfolgsfaktoren und Erfolgswirkung – Erkenntnisse aus der Praxis, in: J. Weber and J. M. Deepen: Erfolg durch Logistik - Erkenntnisse aktueller Forschung. Bern, Stuttgart, Wien. Engelbrecht, C. (2004): Logistikoptimierung durch Outsourcing: Erfolgswirkungen und Erfolgsfaktoren. Wiesbaden.

References

321

Enz, C. A. (1988): The Role of Value Congruity in Intraorganizational Power, in: Administrative Science Quarterly, Vol. 33 (2): pp. 284-304. Ernst, H. (2001): Erfolgsfaktoren neuer Produkt - Grundlagen für eine valide empirische Forschung. Wiesbaden. Etgar, M. (1979): Sources and Types of Intrachannel Conflict, in: Journal of Retailing, Vol. 55 (1): pp. 61-78. Farley, J. U. (1964): Why Does "Brand Loyalty" Vary Over Products, in: Journal of Marketing Research (JMR), Vol. 1 (4): pp. 9-14. Farrell, D. and C. E. Rusbult (1981): Exchange Variables as Predictors of Job Satisfaction, Job Commitment, and Turnover: The Impact of Rewards, Costs, Alternatives, and Investments, in: Organizational Behavior and Human Performance, Vol. 27: pp. 78-95. Fawcett, S. E., L. Birou and B. Cofield Taylor (1993): Supporting Global Operations through Logistics and Purchasing, in: International Journal of Physical Distribution & Logistics Management, Vol. 23 (4): pp. 3-11. Finn, A. and U. Kayande (1997): Reliability assessment and optimization of marketing measurement, in: Journal of Marketing Research, Vol. 34 (2): pp. 262-275. Flint, D. J., E. Larsson, B. Gammelgaard and J. T. Mentzer (2005): Logistics Innovation: A Customer Value-Oriented Social Process, in: Journal of Business Logistics, Vol. 26 (1): pp. 113-147. Fornell, C. (1982): A Second Generation of Multivariate Analysis. New York. Fornell, C. (1992): A National Customer Satisfaction Barometer: The Swedish Experience, in: Journal of Marketing, Vol. 56 (1): pp. 6-21. Fornell, C. and D. F. Larcker (1981): Evaluating Structural Equation Models with Unobservable Variables and Measurement Error, in: Journal of Marketing Research, Vol. 18 (1): pp. 39-50.

322

References

Foster, T. A. and E. J. Muller (1990): Third Parties: Your Passport to Profits, in: Distribution, Vol. 89 (10): pp. 30-32. Frazier, G. L. (1983): Interorganizational Exchange Behavior in Marketing Channels: A Broadenend Perspective, in: Journal of Marketing, Vol. 47 (Fall 1983): pp. 68-78. Frazier, G. L. and J. O. Summers (1984): Interfirm Influence Strategies and Their Application within Distribution Channels, in: Journal of Marketing, Vol. 48 (Summer 1984): pp. 43-55. Galbraith, J. (1973): Designing Complex Organizations. Reading. Ganesan, S. (1994): Determinants of Long-Term Orientation in BuyerSeller Relationships, in: Journal of Marketing, Vol. 58 (April 1994): pp. 1-19. Gardner, J. T. and M. C. Cooper (1988): Elements of Strategic Partnership, in: J. E. McKeon: Partnerships: A Natural Evolution in Logistics. Cleveland, OH: pp. 15-32. Gardner, J. T., M. C. Cooper and T. Noordewier (1994): Understanding shipper-carrier and shipper-warehouser relationships: Partnerships revisited., in: Journal of Business Logistics, Vol. 15 (2): pp. 121-143. Garfinkel, H. (1967): Studies in Ethnomethodology. Englewood Cliffs. Garver, M. S. and J. T. Mentzer (1999): Logistics Research Methods: Employing Structural Equation Modeling to Test for Construct Validity, in: Journal of Business Logistics, Vol. 20 (1): pp. 33-57. Gaski, J. F. (1984): The Theory of Power and Conflict in Channels of Distribution, in: Journal of Marketing, Vol. 48 (3): pp. 9-29. Gassenheimer, J. B., F. S. Houston and J. C. Davis (1998): The Role of Economic Value, Social Value, and Perceptions of Fairness in Interorganizational Relationship Retention Decisions, in: Journal of the Academy of Marketing Science, Vol. 26 (4): pp. 322-337. Gassenheimer, J. B., J. U. Sterling and R. A. Robicheaux (1996): Longterm channel member relationships, in: International Journal of Physical Distribution & Logistics Management, Vol. 26 (5): pp. 94-114.

References

323

Gattorna, J. (1978): Channels of Distribution Conceptualisations: A State-of-the-Art Review, in: European Journal of Marketing, Vol. 12 (7): pp. 471-491. Gattorna, J. (1991): Building Relationships in Distribution Channels, in: International Journal of Physical Distribution & Logistics Management, Vol. 21 (8): pp. 36-39. Gentry, J. J. (1996): Carrier involvement in buyer-supplier strategic partnerships, in: International Journal of Physical Distribution & Logistics Management, Vol. 26 (3): pp. 14-25. Gerbing, D. W. and J. C. Anderson (1988): An Updated Paradigm for Scale Development Incorporating Unidimensionality and Its Assessment, in: Journal of Marketing Research, Vol. 25 (2): pp. 186-192. Ghoshal, S. and P. Moran (1996): Bad for practice: A critique of the transaction cost theory, in: Academy of Management Review, Vol. 21 (1): pp. 13-47. Giddens, A. (1984): The Constitution of Society: Outline of the Theory of Structuration. Cambridge. Giering, A. (2000): Der Zusammenhang zwischen Kundenzufriedenheit und Loyalität - Eine Untersuchung moderierender Effekte. Wiesbaden. Gilbert, F. W. and W. E. Warren (1995): Psychographic Constructs and Demographic Segments, in: Psychology & Marketing, Vol. 12 (3): pp. 223-237. Gonzalez, A. (2005): Study finds 3PLs searching for differentiation, in: Modern Materials Handling, Vol. 60 (5): pp. 10-12. Göpfert, I. (1999): Stand und Entwicklungen der Logistik, in: Logistik Management, Vol. 1 (1): pp. 19-34. Granzin, K. L. and J. J. Painter (2001): Motivational Influences on 'Buy Domestic' Purchasing: Marketing Management Implications from a Study of Two Nations, in: Journal of International Marketing, Vol. 9 (2): pp. 73-96.

324

References

Greer, T. V. and N. Chuchinprakarn (1999): Business Respondents' Behavior: Main and Interaction Effects of Delivery Method, Questionnaire, Length, and Time of the Week, in: Journal of Business-toBusiness Marketing, Vol. 6 (1): pp. 59-88. Griffis, S. E., T. J. Goldsby and M. Cooper (2003): Web-bases and Mail Surveys: A Comparison of Response, Data and Cost, in: Journal of Business Logistics, Vol. 24 (2): pp. 237-258. Grönroos, C. (1989): Defining Marketing: A Market-Oriented Approach, in: European Journal of Marketing, Vol. 23 (1): pp. 52-60. Grönroos, C. (1990): Relationship Approach to Marketing in Service Contexts: The Marketing and Organizational Behavior Interface, in: Journal of Business Research, Vol. 20: pp. 3-11. Grönroos, C. (1994): Quo Vadis, Marketing? Toward a Relationship Marketing Paradigm, in: Journal of Marketing Management, Vol. 10: pp. 347-360. Grönroos, C. and E. Gummesson (1985): Service Orientation in Industrial Marketing - Working Paper. Helsinki, Stockholm. Gruen, T. W., J. O. Summers and F. Acito (2000): Relationship Marketing Activities, Commitment, and Membership Behaviors in Professional Associations, in: Journal of Marketing, Vol. 64 (3): pp. 34-49. Guiltinan, J. P., I. B. Rejab and W. C. Rodgers (1980): Factors Influencing Coordination in a Franchise Channel, in: Journal of Retailing, Vol. 56 (3): pp. 41-58. Gummesson, E. (1987): The New Marketing: Developing Long term Interactive Relationships, in: Long Range Planning, Vol. 20 (4): pp. 1020. Gummesson, E. (1997): Relationship marketing as a paradigm shift: some conclusions from the 30R approach, in: Management Decision, Vol. 35 (4): pp. 267-272. Gundlach, G. T., R. S. Achrol and J. T. Mentzer (1995): The structure of commitment in exchange, in: Journal of Marketing, Vol. 59 (1): pp. 78-92.

References

325

Gundlach, G. T. and P. E. Murphy (1993): Ethical and legal foundations of relational marketing exchanges, in: Journal of Marketing, Vol. 57 (4): pp. 35-46. Gupta, A. K. (1987): SBU Strategies, Corporate-SBU Relations, and SBU Effectiveness in Strategy Implementation, in: Academy of Management Journal, Vol. 30 (3): pp. 477-500. Gutenberg, E. (1957): Betriebswirtschaftslehre als Wissenschaft. Acadmic commemorative speech on the occasion of the foundation of the University on May, 22nd, 1957, Krefeld: pp. 5-38. GVB (2000): Die GVB (Gesellschaft für Verkehrsbetriebswirtschaftslehre) Top 100 in Europa: Der Club der Umsatzmilliardäre, in: Logistik Heute (1-2): pp. 32-35. Hackett, R. D., P. Bycio and P. A. Hausdorf (1994): Further Assessments of Meyer and Allen's (1991) Three-Component Model of Organizational Commitment, in: Journal of Applied Psychology, Vol. 79 (1): pp. 15-23. Hage, J. (1974): Communication and Organizational Control: Cybernetics in Health and Welfare Settings. New York. Hakansson, H. (1982): International Marketing und Purchasing of Industrial Goods: An Interaction Approach. Chichester, New York, Brisbane, Toronto, Singapore. Hakansson, H. and B. Wootz (1979): A Framework of Industrial Buying and Selling, in: Industrial Marketing Management, Vol. 8 (1): pp. 28-39. Hall, R. H. (1963): The Concept of Bureaucracy: An Empirical Assessment, in: American Journal of Sociology, Vol. 69: pp. 32-40. Hallén, L., J. Johanson and N. Seyed-Mohamed (1991): Interfirm Adaptation in Business Relationships, in: Journal of Marketing, Vol. 55 (2): pp. 29-37. Hamel, G., Y. L. Doz and C. K. Prahalad (1989): Collaborate with your competitors - and Win, in: Harvard Business Review, Vol. 76 (1): pp. 133-139.

326

References

Han, S.-L., D. T. Wilson and S. P. Dant (1993): Buyer-Supplier Relationships Today, in: Industrial Marketing Management, Vol. 22: pp. 331-338. Hardy, K. G. and A. J. Magrath (1988): Marketing Channel Management: Strategic Planning and Tactics. Glenview, IL. Harrington, B. E. (2001): Organizational Performance and Corporate Social Capital: A Contingency Model, in: Research in the Sociology of Organizations, Vol. 18: pp. 83-106. Hart, S. (1987): The Use of the Survey in Industrial Market Research, in: Journal of Marketing Management, Vol. 3 (1): pp. 25-38. Hart, S. and C. Banbury (1994): How Strategy-Making Processes Can Make A Difference, in: Strategic Management Journal, Vol. 15 (4): pp. 251-269. Heeler, R. M. and M. L. Ray (1972): Measure Validation in Marketing, in: Journal of Marketing Research, Vol. 9 (4): pp. 361-370. Heide, J. B. (1994): Interorganizational governance in marketing channels, in: Journal of Marketing, Vol. 58 (1): pp. 71-85. Heide, J. B. and G. John (1990): Alliances in Industrial Purchasing: The Determinants of Joint Action in Buyer-Supplier Relationships, in: Journal of Marketing Research, Vol. 27 (February 1990): pp. 24-36. Heide, J. B. and G. John (1992): Do Norms Matter in Marketing Relationships?, in: Journal of Marketing, Vol. 56 (April 1992): pp. 32-44. Heide, J. B. and A. S. Miner (1992): The Shadow of the Future: Effects of Anticipated Interaction and Frequency of Contact on Buyer-Seller Cooperation, in: Academy of Management Journal, Vol. 35 (2): pp. 265291. Heide, J. B. and R. L. Stump (1995): Performance Implications of Buyer-Supplier Relationships in Industrial Markets - A Transaction Cost Explanation, in: Journal of Business Research, Vol. 32: pp. 57-66.

References

327

Hennart, J.-F. (1993): Explaining the Swollen Middle: Why Most Transactions are a Mix of 'Market' and 'Hierarchy', in: Organization Science, Vol. 4 (4): pp. 529-547. Herzberg, F. (1968): One more time: How do you motivate employees?, in: Harvard Business Review, Vol. 46 (1): pp. 52-62. Heskett, J. L. (1977): Logistics - essential to strategy, in: Harvard Business Review, Vol. 55 (6): pp. 85-96. Hildebrandt, L. (1998): Kausalanalytische Validierung in der Marketingforschung, in: L. Hildebrandt and C. Homburg: Die Kausalanalyse - Instrument der empirischen betriebswirtschaftlichen Forschung. Stuttgart: pp. 86-110. Hobbs, J. E. (1996): A transaction cost approach to supply chain management, in: Supply Chain Management, Vol. 1 (2): pp. 15-27. Hocutt, M. A. (1998): Relationship dissolution model: Antecedents of relationship commitment and the likelihood of dissolving a relationship, in: International Journal of Service Industry Management, Vol. 9 (2): pp. 189-200. Homans, G. C. (1958): Social Behavior as Exchange, in: American Journal of Sociology, Vol. 63: pp. 597-606. Homburg, C. (1989): Exploratorische Ansätze der Kausalanalyse als Instrument der Marketingplanung. Frankfurt am Main, Bern, New York. Homburg, C. (1995): Kundennähe von Industriegüterunternehmen - Konzeption, Erfolgsauswirkungen, Determinanten. Wiesbaden. Homburg, C. and H. Baumgartner (1995a): Beurteilung von Kausalmodellen - Bestandsaufnahme und Anwendungsempfehlungen, in: Marketing - ZFP, Vol. 17 (3): pp. 162-176. Homburg, C. and H. Baumgartner (1995b): Die Kausalanalyse als Instrument der Marketingforschung, in: Zeitschrift für Betriebswirtschaft, Vol. 65 (10): pp. 1091-1108.

328

References

Homburg, C. and A. Giering (1996): Konzeptualisierung und Operationalisierung komplexer Konstrukte - Ein Leitfaden für die Marketingforschung, in: Marketing - ZFP, Vol. 18 (1): pp. 5-24. Homburg, C., A. Giering and A. Menon (2003): Relationship Characteristics as Moderators of the Satisfaction-Loyalty Link: Findings in a Business-to-Business Context, in: Journal of Business-to-Business Marketing, Vol. 10 (3): pp. 35-62. Houston, F. S. and J. B. Gassenheimer (1987): Marketing and Exchange, in: Journal of Marketing, Vol. 51 (4): pp. 3-18. Hoyt, J. and F. Huq (2000): From arms-length to collaborative relationships in the supply chain - An evolutionary process, in: International Journal of Physical Distribution & Logistics Management, Vol. 30 (9): pp. 750-764. Hrebiniak, L. G. (1974a): Effects of Job Level and Participation on Employee Attitudes and Perceptions of Influence, in: Academy of Management Journal, Vol. 17 (4): pp. 649-662. Hrebiniak, L. G. (1974b): Job Technology, Supervision, and Work-Group Structure, in: Administrative Science Quarterly, Vol. 19 (3): pp. 649662. Hu, L.-t. and P. M. Bentler (1995): Evaluating Model Fit, in: R. H. Hoyle: Structural Equation Modeling: Concepts, Issues, and Applications. London, New Dehli: pp. 76-99. Hu, L.-t. and P. M. Bentler (1999): Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria Versus New Alternatives, in: Structural Equation Modeling, Vol. 6 (1): pp. 1-55. Innis, D. E. and B. J. La Londe (1994): Customer service: The key to customer satisfaction, customer loyalty, and market share, in: Journal of Business Logistics, Vol. 15 (1): pp. 1-27. Irving, E. L. (1995): Marketing quality practices, Dissertation. Chapel Hill. Jacoby, J. (1978): Consumer Research: A State of the Art Review, in: Journal of Marketing, Vol. 42 (2): pp. 87-96.

References

329

Jarillo, J. C. (1988): On strategic networks, in: Strategic Management Journal, Vol. 9 (1): pp. 31-41. Jasper, C. R. and P.-N. R. Lan (1992): Apparel Catalog Patronage: Demographic, Lifestyle and Motivational Factors, in: Psychology & Marketing, Vol. 9 (4): pp. 275-296. Jensen, M. C. and W. H. Meckling (1976): Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, in: Journal of Financial Economics, Vol. 3 (4): pp. 305-360. John, D. R. and C. A. Cole (1986): Age Differences in Information Processing: Understanding Deficits in Young and Elderly Consumers, in: Journal of Consumer Research, Vol. 13 (3): pp. 297-315. John, G. (1984): An Empirical Investigation of Some Antecedents of Opportunism in a Marketing Channel, in: Journal of Marketing Research, Vol. 21 (3): pp. 278-289. John, G. and T. Reve (1982): The Reliability and Validity of Key Informant Data from Dyadic Relationships in Marketing Channels, in: Journal of Marketing Research, Vol. 19 (4): pp. 517-524. Jöreskog, K. G. (1973): A General Method for Estimating a Linear Structural Equation System, in: A. S. Goldberger and O. D. Duncan: Structural Equation Models in the Social Sciences. New York: pp. 85-112. Jöreskog, K. G. (1977): Structrual Equation Models in the Social Sciences: Specification, Estimation, and Testing, in: P. R. Krishnaiah: Applications of Statistics. Amsterdam: pp. 265-287. Jöreskog, K. G. and D. Sörbom (1982): Recent Developments in Structural Equation Modeling, in: Journal of Marketing Research, Vol. 19 (4): pp. 404-416. Joshi, A. W. and R. L. Stump (1999): Determinants of Commitment and Opportunism: Integrating and Extending Insights from Transaction Cost Theory and Relational Exchange Theory, in: Canadian Journal of Administrative Sciences, Vol. 16 (4): pp. 334-352.

330

References

Kaiser, H. F. (1974): An Index of Factorial Simplicity, in: Psychometrika, Vol. 39 (1): pp. 31-36. Kalwani, M. U. and N. Narayandas (1995): Long-Term ManufacturerSupplier Relationships: Do they pay off for Supplier Firms?, in: Journal of Marketing, Vol. 59 (1): pp. 1-16. Kanter, R. M. (1972): Commitment and Community: Communes and Utopias in Sociological Perspective. Cambridge. Kanter, R. M. (1994): Collaborative Advantage: The Art of Alliances, in: Harvard Business Review, Vol. 72 (4): pp. 96-108. Kaufmann, L. (2001): Internationales Beschaffungsmanagement: Gestaltung strategischer Gesamtsysteme und Management einzelner Transaktionen. Wiesbaden. Keesling, J. W. (1972): Maximum Likelihood Approaches to Causal Flow Analysis, Dissertation. Chicago. Kelley, H. H. and J. W. Thibaut (1978): Interpersonal Relations: A Theory of Interdependece. New York, Chichester, Brisbane, Toronto. Kenis, P. and D. Knoke (2002): How Organizational Field Networks Shape Organizational Tie-Formation Rates, in: Academy of Management Review, Vol. 27 (2): pp. 275-293. Kieser, A. (1974): Der Einfluss der Umwelt auf die Organisationstruktur der Unternehmung, in: Zeitschrift für Organisation, Vol. 43 (6): pp. 302314. Kieser, A. and H. Kubicek (1978): Organisationstheorien II: Wissenschaftstheoretische Anforderungen und kritische Analyse klassischer Ansätze. Stuttgart, Berlin, Köln, Mainz. Kieser, A. and H. Kubicek (1983): Organisation, 2. Edition. Berlin, New York. Kieser, A. and H. Kubicek (1992): Organisation, 3. Edition. Berlin, New York.

References

331

Kinnear, T. C. and J. R. Taylor (1996): Marketing Research - An Applied Approach, 5. Edition. New York. Kirsch, W., I. Bamberger, E. Gabele and K. Klein (1973): Betriebswirtschaftliche Logistik. Wiesbaden. Klaus, P. (1993): Die dritte Bedeutung der Logistik, Nuremberg Logistics Working Paper No. 3, University of Erlangen-Nürnberg. Klaus, P. (1999): Logistik als „Weltsicht“, in: J. Weber and H. Baumgarten: Handbuch Logistik – Management von Material- und Warenflußprozessen. Stuttgart: pp. 15-32. Klaus, P. (2003): Die "Top 100" der Logistik - Marktgrössen, Marktsegmente und Marktführer in der Logistikdienstleistungswirtschaft. Hamburg. Kleer, M. (1991): Gestaltung von Kooperationen zwischen Industrie- und Logistikunternehmen - Ergebnisse theoretischer und empirischer Untersuchungen. Berlin. Klein, B., R. G. Crawford and A. A. Alchian (1978): Vertical Integration, Appropriable Rents, and the Competitive Contracting Process, in: Journal of Law & Economics, Vol. 21 (2): pp. 297-326. Knemeyer, A. M., T. M. Corsi and P. R. Murphy (2003): Logistics outsourcing relationships: Customer perspectives, in: Journal of Business Logistics, Vol. 24 (1): pp. 77-109. Knemeyer, A. M. and P. R. Murphy (2004): Evaluating the Performance of Third-Party Logistics Arrangements: A Relationship Marketing Perspective, in: The Journal of Supply Chain Management, Vol. 40 (1): pp. 35-51. Knemeyer, A. M. and P. R. Murphy (2005): Exploring the Potential Impact of Relationship Characteristics and Customer Attributes on the Outcomes of Third-party Logistics Arrangements, in: Transportation Journal, Vol. 44 (1): pp. 5-19. Kumar, N., L. K. Scheer and J.-B. E. M. Steenkamp (1995): The effects of perceived interdependence on dealer attitudes, in: Journal of Marketing Research, Vol. 32 (3): pp. 348-356.

332

References

Kumar, N., L. W. Stern and J. C. Anderson (1993): Conducting interorganizational research using key informants, in: Academy of Management Journal, Vol. 36 (6): pp. 1633-1651. Kummer, S. (1993): Outsourcing von Logistikleistungen: Controlling wirkt qualitätssichernd, in: Beschaffung aktuell (11/93): pp. 29-31. Lacity, M. C., L. P. Willcocks and D. F. Feeny (1995): IT Outsourcing: Maximize Flexibility and Control, in: Harvard Business Review, Vol. 73 (3): pp. 84-93. LaLonde, B. J. and M. C. Cooper (1989): Partnerships in Providing Customer Service: A Third-Party Perspective. Oak Brook, IL. LaLonde, B. J. and A. B. Maltz (1992): Some Propositions about Outsourcing the Logistics Function, in: The International Journal of Logistics Management, Vol. 3 (1): pp. 1-11. Lambe, C. J., C. M. Wittmann and R. E. Spekmann (2001): Social Exchange Theory in Research on Business-to-Business Relational Exchange, in: Journal of Business-to-Business Marketing, Vol. 8 (3): pp. 1-36. Lambert, D. M., M. A. Emmelhainz and J. T. Gardner (1996a): Developing and Implementing Supply Chain Partnerships, in: International Journal of Logistics Management, Vol. 7 (2): pp. 1-17. Lambert, D. M., M. A. Emmelhainz and J. T. Gardner (1996b): So you think you want a partner?, in: Marketing Management, Vol. 5 (2): pp. 24-41. Lambert, D. M., M. A. Emmelhainz and J. T. Gardner (1999): Building Successful Logistics Partnerships, in: Journal of Business Logistics, Vol. 20 (1): pp. 165-181. Lane, C. (2000): Theories and Issues in the Study of Trust, in: C. Lane and R. Bachmann: Trust within and between Organizations. Oxford: pp. 1-30.

References

333

Langley, C. J., G. R. Allen and M. J. Colombo (2003): Third-Party Logistics Study - Results and Findings of the 2003 Eighth Annual Study. Atlanta. Langley, C. J., G. R. Allen and T. A. Dale (2004): Third-Party Logistics Study - Results and Findings of the 2004 Ninth Annual Study. Langley, C. J., E. v. Dort, A. Ang and S. R. Sykes (2005): Third Party Logistics - Results and Findings of the 10th Annual Study. Larson, P. D. and A. Halldorsson (2004): Logistics versus supply chain management: An international survey, in: International Journal of Logistics: Research & Applications, Vol. 7 (1): pp. 17-31. Larson, P. D. and J. D. Kulchitsky (1999): Logistics improvement programs - the dynamics between people and performance, in: International Journal of Physical Distribution & Logistics Management, Vol. 29 (2): pp. 88-102. Lawrence, P. R. and J. W. Lorsch (1967): Organization and Environment. Boston. Lewicki, R. J. and B. B. Bunker (1995): Trust in Relationsships: A Model of Development and Decline, in: B. B. Bunker and J. Z. Rubin: Conflict, Cooperation, and Justice: Essays Inspired by the Work of Morton Deutsch. San Francisco. Lewis, J. D. and A. Weigert (1985): Trust as a Social Reality, in: Social Forces, Vol. 63 (4): pp. 967-985. Lieb, R. C. (1992): The Use of Third-Party Logistics Services by Large American Manufacturers, in: Journal of Business Logistics, Vol. 13 (2): pp. 29-42. Lieb, R. C. (2005): The 3PL Industry: Where It's Been, Where It's Going, in: Supply Chain Management Review, Vol. 9 (6): pp. 20-27. Lieb, R. C. and B. A. Bentz (2004): The Use of Third-Party Logistics Services by Large American Manufacturers: The 2003 Survey, in: Transportation Journal, Vol. 43 (3): pp. 24-33.

334

References

Lieb, R. C., R. A. Millen and L. N. van Wassenhove (1993): Third-party Logistics Services: A Comparison of Experienced American and European Manufacturers, in: International Journal of Physical Distribution & Logistics Management, Vol. 23 (6): pp. 35-44. Lieb, R. C. and H. L. Randall (1996): A Comparison of the Use of ThirdParty Logistics Services by Large American Manufacturers 1991,1994 and 1995, in: Journal of Business Logistics, Vol. 17 (1): pp. 305-320. Lindskold, S. (1978): Trust Development, the GRIT Proposal, and the Effects of Conciliatory Acts on Conflict and Cooperation, in: Psychological Bulletin, Vol. 85 (4): pp. 772-793. Long, J. S. (1983): Confirmatory Factor Analysis: A preface to LISREL. Beverly Hills. Luhmann, N. (1979): Trust and Power. Chichester. Lusch, R. F. and J. R. Brown (1996): Interdependency, contracting, and relational behavior in marketing channels, in: Journal of Marketing, Vol. 60 (4): pp. 19-38. Lynch, C. F. (2000a): Logistics Outsourcing: A Management Guide. Oak Brook, IL. Lynch, C. F. (2000b): Managing the Outsourcing Relationship, in: Supply Chain Management Review, Vol. 4 (4): pp. 90-96. Lynch, M. E., S. J. Imada and J. H. Bookbinder (1994): The Future of Logistics in Canada: A Delphi-Based Forecast, in: Logistics & Transportation Review, Vol. 30 (1): pp. 95-112. MacCallum, R. C., M. W. Browne and H. M. Sugawara (1996): Power Analysis and Determination of Sample Size for Covariance Structure Modeling, in: Psychological Methods, Vol. 1 (2): pp. 130-149. Macneil, I. R. (1978): Contracts: Adjustment of Long-Term Economic Relations under Classical, Neoclassical, and Relational Contract Law, in: Northwestern University Law Review, Vol. 72 (6): pp. 854-904. Macneil, I. R. (1980): The New Social Contract: An Inquiry Into Modern Contractual Relations. New Haven, London.

References

335

Malhotra, N. (1993): Marketing Research - An Applied Orientation. Englewood Cliffs. Mallen, B. (1963): A Theory of Retailer-Supplier Conflict, Control, and Cooperation, in: Journal of Retailing, Vol. 39: pp. 24-32 and pp. 51-52. Maloni, M. and W. C. Benton (2000): Power Influences on the Supply Chain, in: Journal of Business Logistics, Vol. 21 (1): pp. 49-73. Marsh, H. W., J. R. Balla and R. P. McDonald (1988): Goodness-of-Fit in Confirmatory Factor Analysis: The Effect of Sample Size, in: Psychological Bulletin, Vol. 103 (3): pp. 391-410. McCabe, J. V. (1990): Outside Managers Offer Packaged Export Expertise, in: Journal of Business Strategy, Vol. 11 (2): pp. 20-23. McDonald, G. W. (1981): Structural Exchange and Marital Interaction, in: Journal of Marriage and the Family, Vol. 43 (4): pp. 825-839. McGee, G. W. and R. C. Ford (1987): Two (or More?) Dimensions of Organizational Commitment: Reexamination of the Affective and Continuance Commitment Scales, in: Journal of Applied Psychology, Vol. 72 (4): pp. 638-642. McGinnis, M. A., C. M. Kochunny and K. B. Ackerman (1995): Third Party Logistics Choice, in: The International Journal of Logistics Management, Vol. 6 (2): pp. 93-102. McHugh, M., P. Humphreys and R. McIvor (2003): Buyer-supplier relationships and organizational health, in: Journal of Supply Chain Management, Vol. 39 (2): pp. 15-25. McIvor, R. (2000): A practical framework for understanding the outsourcing process, in: Supply Chain Management: An International Journal, Vol. 5 (1): pp. 22-36. Mentzer, J. T. and D. J. Flint (1997): Validity in Logistics Research, in: Journal of Business Logistics, Vol. 18 (1): pp. 199-216. Mentzer, J. T. and K. B. Kahn (1995): A framework of logistics research, in: Journal of Business Logistics, Vol. 16 (1): pp. 231-250.

336

References

Mentzer, J. T. and B. P. Konrad (1991): An efficiency/effectiveness approach to logistics performance analysis, in: Journal of Business Logistics, Vol. 12 (1): pp. 33-62. Metcalf, L. E., C. R. Frear and R. Krishnan (1992): Buyer-Seller Relationships: An Application of the IMP Interaction Model, in: European Journal of Marketing, Vol. 26 (2): pp. 27-46. Meyer, J. P. and N. J. Allen (1984): Testing the 'Side-Bet Theory' of Organizational Commitment: Some Methodological Considerations, in: Journal of Applied Psychology, Vol. 69 (3): pp. 372-378. Meyer, J. P., N. J. Allen and C. A. Smith (1993): Commitment to Organizations and Occupations: Extension and Test of a Three-Component Conceptualization, in: Journal of Applied Psychology, Vol. 78 (4): pp. 538-551. Moore, K. R. (1998): Trust and Relationship Commitment in Logistics Alliances: A Buyer Perspective, in: International Journal of Purchasing and Materials Management, Vol. 34 (1): pp. 24-37. Moore, K. R. and W. A. Cunningham III (1999): Social exchange behaviour in logistics relationships: A shipper perspective, in: International Journal of Physical Distribution & Logistics Management, Vol. 29 (2): pp. 103-121. Moorman, C., R. Deshpandé and G. Zaltman (1993): Factors Affecting Trust in Market Research Relationships, in: Journal of Marketing, Vol. 57 (1): pp. 81-101. Moorman, C., G. Zaltman and R. Deshpandé (1992): Relationships between Providers and Users of Market Research: The Dynamics of Trust within and between Organizations, in: Journal of Marketing Research, Vol. 29 (3): pp. 314-328. Morgan, R. M. and S. D. Hunt (1994): The Commitment-Trust Theory of Relationship Marketing, in: Journal of Marketing, Vol. 58 (3): pp. 2038.

References

337

Moscovitch, M. (1982): A neuropsychological approach to perception and memory in normal and pathological aging, in: F. I. M. Craik and S. Trehub: Aging and cognitive processes. New York: pp. 55-78. Mowday, R. T., L. W. Porter and R. M. Steers (1982): Organizational Linkages: The Psychology of Commitment, Absenteeism, and Turnover. New York. Murphy, P. R. and J. M. Daley (1994): Logistics Issues in International Sourcing: An Exploratory Study, in: International Journal of Purchasing & Materials Management, Vol. 30 (3): pp. 22-27. Murphy, P. R. and R. F. Poist (1998): Third-Party Logistics Usage: An Assessment of Propositions Based on Previous Research, in: Transportation Journal, Vol. 37 (4): pp. 26-35. Murphy, P. R. and R. F. Poist (2000): Third-Party Logistics: Some Use versus Provider Perspectives, in: Journal of Business Logistics, Vol. 21 (1): pp. 121-133. Naman, J. L. and D. P. Slevin (1993): Entrepreneurship and the Concept of Fit: A Model and Empirical Tests, in: Strategic Management Journal, Vol. 14 (2): pp. 137-153. Noordewier, T. G., G. John and J. R. Nevin (1990): Performance Outcomes of Purchasing Arrangements in Industrial Buyer-Vendor Relationships, in: Journal of Marketing, Vol. 54 (4): pp. 80-93. North, D. C. (1990): Institutions, Institutional Change and Economics Performance. Cambridge. Nunnally, J. C. (1978): Psychometric Theory. New York, St. Louis, San Francisco. Paché, G. (1998): Logistics outsourcing in grocery distribution: A European perspective, in: Logistics Information Management, Vol. 11 (5): pp. 301-308. Parasuraman, A., V. A. Zeithaml and L. L. Berry (1988): SERVQUAL: A Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality, in: Journal of Retailing, Vol. 64 (1): pp. 12-40.

338

References

Parsons, T. (1951): The Social System. Glencoe, New York. Parsons, T. (1969): Politics and Social Structure. New York. Pearson, M. and J. F. Monoky (1976): The Role of Conflict and Cooperation in Channel Performance, in: K. L. Bernhardt: Marketing 17761976 and Beyond: pp. 240-244. Penrose, E. T. (1959): The Theory of the Growth of the Firm. New York. Perrow, C. (1970): Organizational Analysis: A Sociological View. Belmont. Persson, G. (1997): Organization design strategies for business logistics, in: International Journal of Physical Distribution & Logistics Management, Vol. 27 (5/6): pp. 282-291. Peter, J. P. (1979): Reliability: A Review of Psychometric Basics and Recent Marketing Practices, in: Journal of Marketing Research, Vol. 16 (1): pp. 6-17. Peter, J. P. (1981): Construct Validity: A Review of Basic Issues and Marketing Practices, in: Journal of Marketing Research, Vol. 18 (2): pp. 133-145. Peter, J. P. and G. A. Churchill (1986): Relationships Among Research Design Choices and Psychometric Properties of Rating Scales: A MetaAnalysis, in: Journal of Marketing Research, Vol. 23 (1): pp. 1-10. Peter, S. I. (1997): Kundenbindung als Marketingziel - Identifikation und Analyse zentraler Determinanten. Wiesbaden. Peterson, R. A. (1994): A Meta-analysis of Cronbach's coefficient alpha, in: Journal of Consumer Research, Vol. 21 (2): pp. 381-391. Pfeffer, J. and G. Salancik (1978): The external control of organizations. New York. Pfohl, H.-C. (1996): Logistiksysteme, 5. Edition. Berlin, Heidelberg, New York.

References

339

Pfohl, H.-C. (1999): Logistik 2000plus: Visionen - Märkte - Ressourcen. Berlin. Pfohl, H.-C. and W. Zöllner (1997): Organization for logistics: The contingency approach, in: International Journal of Physical Distribution & Logistics Management, Vol. 27 (5/6): pp. 306-320. Picot, A. (1991): Ein neuer Ansatz zur Gestaltung der Leistungstiefe, in: Zeitschrift für betriebswirtschaftliche Forschung, Vol. 43 (4): pp. 336357. Porter, L. W., R. M. Steers, R. T. Mowday and P. V. Boulian (1974): Organizational Commitment, Job Satisfaction, and Turnover Among Psychiatric Technicians, in: Journal of Applied Psychology, Vol. 59 (5): pp. 603-609. Porter, M. E. and M. B. Fuller (1986): Coalitions and Global Strategy, in: M. E. Porter: Competition in Global Industries. Boston: pp. 315-343. Powell, W. W. (1991): Neither market nor hierarchy: network forms of organization, in: G. Thompson, J. Frances, R. Levacic and J. Mitchell: Markets, Hierarchies and Networks. London: pp. 265-276. Prahalad, C. K. and G. Hamel (1990): The Core Competence of the Corporation, in: Harvard Business Review, Vol. 68 (3): pp. 79-91. Prasad, S., J. Tata and J. Motwani (2001): International Supply Chain Management: Learning and Evolving Networks, in: Global Journal of Flexible Systems Management, Vol. 2 (2): pp. 31-36. Premeaux, S. F. and A. G. Bedeian (2003): Breaking the Silence: The Moderating Effects of Self-Monitoring in Predicting Speaking Up in the Workplace, in: Journal of Management Studies, Vol. 40 (6): pp. 15371562. Pugh, D. S. (1981): The Aston Program Perspective - The Aston Program of Research: Retrospect and Prospect, in: A. H. Van De Ven and W. F. Joyce: Perspectives on Organization Design and Behaviour. New York, Chichester, Brisbane: pp. 135-166.

340

References

Pugh, D. S. and D. C. Pheysey (1992): Some Developments in the Study of Organizations, in: Management International Review, Vol. 32 (Special Issue 1/1992): pp. 63-73. Quinn, J. P. (2005): Greater Expectations for Global 3PLs, in: Supply Chain Management Review, Vol. 9 (5): pp. P3-P16. Rabinovich, E., R. Windle, M. Dresner and T. M. Corsi (1999): Outsourcing of integrated logistics functions - An examination of industry practices, in: International Journal of Physical Distribution & Logistics Management, Vol. 29 (6): pp. 353-373. Ranaweera, C. and A. Neely (2003): Some moderating effects on the service quality-customer retention link, in: International Journal of Operations & Production Management, Vol. 23 (2): pp. 230-248. Rao, K. and R. R. Young (1994): Global Supply Chains, in: International Journal of Physical Distribution & Logistics Management, Vol. 24 (6): pp. 11-19. Rao, K., R. R. Young and J. A. Novick (1993): Third Party Services in the Logistics of Global Firms, in: Logistics & Transportation Review, Vol. 29 (4): pp. 363-370. Razzaque, M. A. and C. C. Sheng (1998): Outsourcing of Logistics Functions: A Literature Survey, in: International Journal of Physical Distribution & Logistics Management, Vol. 28 (2): pp. 89-107. Reicheld, F. F. (1996): The Loyalty Effect: The Hidden Force behind Growth, Profits, and Lasting Value. Boston. Reichers, A. E. (1985): A Review and Reconceptualization of Organizational Commitment, in: Academy of Management Review, Vol. 10 (3): pp. 465-476. Richardson, H. L. (1992): Outsourcing: The Power Worksource, in: Transportation & Distribution (July 1992): pp. 22-24. Richardson, H. L. (1993): Why Use Third Parties?, in: Transportation & Distribution (January 1993): pp. 29-31.

References

341

Richardson, H. L. (1995): Logistics help for the challenged, in: Transportation & Distribution (January 1995): pp. 60-64. Rindfleisch, A. and J. B. Heide (1997): Transaction cost analysis: Past, present, and future applications, in: Journal of Marketing, Vol. 61 (4): pp. 30-54. Ring, P. S. and A. H. Van De Ven (1994): Developmental Processes of Cooperative Interorganziational Relationships, in: Academy of Management Review, Vol. 19 (1): pp. 90-118. Rodríguez, C. M. and D. T. Wilson (2002): Relationship Bonding and Trust as a Foundation for Commitment in U.S.-Mexican Strategic Alliances: A Structural Equation Modeling Approach, in: Journal of International Marketing, Vol. 10 (4): pp. 53-76. Rogers, E. M. (1995): Diffusion of Innovations, 4. Edition. New York. Rosenberg, L. J. (1974): A New Approach to Distribution Conflict Management, in: Business Horizons, Vol. 17 (5): pp. 67-74. Ross, R. H., R. F. Lusch and J. R. Brown (1982): Power Dependency in the Marketing Channel: A Methodological Note, in: B. J. Walker: An Assessment of Marketing Thought and Practice, 1982 Educator's Conference Proceeds, Series No. 48. Chicago, IL: pp. 194-198. Rotter, J. B. (1967): A New Scale for the Mearsurement of Interpersonal Trust, in: Journal of Personality, Vol. 35 (4): pp. 651-665. Ruekert, R. W. and G. A. Churchill Jr. (1984): Reliability and Validity of Alternative Measures of Channel Member Satisfaction, in: Journal of Marketing Research, Vol. 21 (2): pp. 226-233. Ruekert, R. W., O. C. Walker Jr. and K. J. Roering (1985): The Organization of Marketing Activities: A Contingency Theory of Structure and Performance, in: Journal of Marketing, Vol. 49 (1): pp. 13-25. Rushing, W. A. (1966): Organizational Size and Administration: The Problems of Causal Homogenity and Heterogenous Category, in: Pacific Sociological Review, Vol. 9: pp. 100-108.

342

References

Rust, R. T. and A. J. Zahorik (1993): Customer Satisfaction, Customer Retention, and Market Share, in: Journal of Retailing, Vol. 69 (2): pp. 193-215. Schäfer-Kunz, J. and C. Tewald (1998): Make-or-buy Entscheidungen in der Logistik. Wiesbaden. Schaninger, C. M. and D. Sciglimpaglia (1981): The Influence of Cognitive Personality Traits and Demographics on Consumer Information Acquisition, in: Journal of Consumer Research, Vol. 8 (2): pp. 208-216. Schein, E. H. (1990): Organizational Culture, in: American Psychologist, Vol. 45 (2): pp. 109-119. Scherer, S. (2004): Stepping-stones or traps? The consequences of labour market entry positions on future careers in West Germany, Great Britain and Italy, in: Work, employment and society, Vol. 18 (2): pp. 369-394. Schermerhorn, J. R. (1975): Determinants of Interorganizational Cooperation, in: Academy of Management Journal, Vol. 18 (4): pp. 846-856. Schmitt, A. (2006): 4PL-Providing als strategische Option für Kontraktlogistikdienstleister: Eine konzeptionell-empirische Betrachtung Forthcoming Dissertation. Vallendar. Schrage, M. (1990): Shared minds: The New Technologies of Collaboration. New York. Schreyögg, G. (1980): Contingency and Choice in Organization Theory, in: Organization Studies, Vol. 1 (4): pp. 305-326. Schumpeter, J. A. (1934): The Theory of Economic Development: An Inquiry into Profits, Capital, Credit, Interest and the Business Cycle. Cambridge. Schurr, P. H. and J. L. Ozanne (1985): Influences on Exchange Processes: Buyers' Preconceptions of a Seller's Trustworthiness and Bargaining Toughness, in: Journal of Consumer Research, Vol. 11 (4): pp. 939953. Scott, W. R. (1992): Organizations: Rational, Natural, and Open Systems, 3. Edition. Englewood Cliffs.

References

343

Scott, W. R. (1995): Institutions and Organizations. Thousand Oaks, CA. Shapiro, S. P. (1987): The Social Control of Impersonal Trust, in: American Journal of Sociology, Vol. 93 (3): pp. 623-658. Sharma, S. C. (1996): Applied Multivariate Techniques. New York. Sheffi, Y. (1990): Third Party Logistics: Present and Future Prospects, in: Journal of Business Logistics, Vol. 11 (2): pp. 27-39. Shemwell, D. J., J. J. Cronin and W. R. Bullard (1994): Relational Exchange in Services: An Empirical Investigation of Ongoing Customer Service-provider Relationships, in: International Journal of Service Industry Management, Vol. 5 (3): pp. 57-68. Sherman, S. and R. Sookdeo (1992): Are strategic alliances working, in: Fortune, Vol. 126 (6): pp. 77-78. Sheth, J. N. and A. Parvatiyar (1995): The Evolution of Relationship Marketing, in: International Business Review, Vol. 4 (7): pp. 397-418. Sibley, S. D. and D. A. Michie (1982): An Exploratory Investigation of Cooperation in a Franchise Channel, in: Journal of Retailing, Vol. 58 (4): pp. 23-45. Simmel, G. (1978): The Philosophy of Money. London. Sink, H. L. and C. J. Langley (1997): A Managerial Framwork for the Acquisition of Third-Party Logistics Services, in: Journal of Business Logistics, Vol. 18 (2): pp. 163-189. Skinner, S. J., J. B. Gassenheimer and S. W. Kelley (1992): Cooperation in Supplier-Dealer Relations, in: Journal of Retailing, Vol. 68 (2): pp. 174-193. Skjoett-Larsen, T. (2000): Third party logistics - from an interorganizational point of view, in: International Journal of Physical Distribution & Logistics Management, Vol. 30 (2): pp. 112-127. Skjoett-Larsen, T., C. Thernoe and C. Andersen (2003): Supply Chain Collaboration - Theoretical Perspectives and Empirical Evidence, in: In-

344

References

ternational Journal of Physical Distribution & Logistics Management, Vol. 33 (6): pp. 531-549. Slama, M. E. and A. Tashchian (1985): Selected Socioeconomic and Demographic Characteristics Associated with Purchasing Involvement, in: Journal of Marketing, Vol. 49 (1): pp. 72-82. Smith, B. J. and D. W. Barclay (1997): The effects of organizational differences and trust on the effectiveness of selling partner relationships, in: Journal of Marketing, Vol. 61 (1): pp. 3-21. Smith, J. and B. B. Baltes (1990): Wisdom-related knowledge: Age/cohort differences in response to life planning problems, in: Development Psychology, Vol. 26: pp. 494-505. Spekman, R. E. (1988): Strategic Supplier Selection: Understanding Long-Term Buyer Relationships, in: Business Horizons, Vol. 31 (4): pp. 75-81. Spekman, R. E., D. J. Salmond and C. J. Lambe (1997): Consensus and Collaboration: Norm-regulated Behaviour in Industrial Marketing Relationships, in: European Journal of Marketing, Vol. 31 (11/12): pp. 832856. Spence, M. T. and M. Brucks (1997): The Moderating Effects of Problem Characteristics on Experts' and Novices' Judgements, in: Journal of Marketing Research, Vol. 34 (2): pp. 233-247. Staehle, W. (1990): Management - Eine verhaltenswissenschaftliche Perspektive, 6. Edition. München. Stank, T. P., T. J. Goldsby and S. K. Vickery (1999): Effect of service supplier performance on satisfaction and loyalty of store managers in the fast food industry, in: Journal of Operations Management, Vol. 17 (4): pp. 429-447. Stank, T. P., T. J. Goldsby, S. K. Vickery and K. Savitskie (2003): Logistics service performance: Estimating its influence on market share, in: Journal of Business Logistics, Vol. 24 (1): pp. 27-55.

References

345

Stank, T. P., S. B. Keller and P. J. Daugherty (2001): Supply Chain Collaboration and Logistical Service Performance, in: Journal of Business Logistics, Vol. 22 (1): pp. 29-48. Steenkamp, J.-B. E. M. and H. Baumgartner (1998): Assessing Measurement Invariance in Cross-National Consumer Research, in: Journal of Consumer Research, Vol. 25 (1): pp. 78-90. Steiger, J. H. and J. C. Lind (1980): Statistically based tests for the number of common factors, Paper presented at the Annual Spring Meeting of the Psychonometric Society. Iowa City. Stern, L. W. and A. I. El-Ansary (1992): Marketing Channels, 4. Edition. Englewood Cliffs, NJ. Stern, L. W. and T. Reve (1980): Distribution Channels as Political Economies: A Framework for Comparative Analysis, in: Journal of Marketing, Vol. 44 (3): pp. 52-64. Stinchcombe, A. L. (1986): Norms of Exchange: A Basic Typology of Exchange Relations, in: A. L. Stinchcombe: Stratification and Organization: Selected Papers. Cambridge: pp. 231-267. Stock, J. R. (1997): Applying theories from other disciplines to logistics, in: International Journal of Physical Distribution & Logistics Management, Vol. 27 (9/10): pp. 515-539. Stölzle, W. (1996): Grundzüge des Outsourcing von Entsorgungsleistungen, in: Zeitschrift für Betriebswirtschaft, ZfB-Ergänzungsheft 2/96: pp. 121-145. Straube, F., H.-C. Pfohl, W. A. Günther and W. Dangelmaier (2005): Ein Blick auf die Agenda des Logistik-Managements 2010: Trends und Strategien in der Logistik. Berlin. Strub, P. J. and T. B. Priest (1976): Two Patterns of Establishing Trust: The Marijuana User, in: Sociological Focus, Vol. 9 (4): pp. 399-411. Stuart, F. I. (1993): Supplier Partnerships: Influencing Factors and Strategic Benfits, in: International Journal of Purchasing and Materials Management, Vol. 29 (4): pp. 22-28.

346

References

Stuart, F. I. and D. McCutcheon (1995): Problem Sources in Establishing Strategic Alliances, in: International Journal of Purchasing and Materials Management, Vol. 31 (1): pp. 3-9. Szymanski, D. M., S. G. Bharadwaj and P. R. Varadarajan (1993): An analysis of the market share-profitability relationship, in: Journal of Marketing, Vol. 57 (3): pp. 1-18. Tate, K. (1996): The elements of a successful logistics partnership, in: International Journal of Physical Distribution & Logistics Management, Vol. 26 (3): pp. 7-13. Thibaut, J. W. and H. H. Kelley (1959): The Social Psychology of Groups. New York. Thompson, J. (1967): Organizations in Action. New York. Thompson, L. and G. B. Spanier (1983): The End of Marriage and Acceptance of Marital Termination, in: Journal of Marriage and the Family, Vol. 45 (February 1983): pp. 103-113. Thorelli, H. B. (1986): Networks: Between Markets and Hierarchies, in: Strategic Management Journal, Vol. 7 (1): pp. 37-51. Thurstone, L. L. (1947): Multiple Factor Analysis: A Development and Expansion of the Vectors of Mind. Chicago. Tosi, H., R. Aldag and R. Storey (1973): On the Measurement of the Environment: An Assessment of the Lawrence and Lorsch Environmental Uncertainty Subscale, in: Administrative Science Quarterly, Vol. 18 (1): pp. 27-36. Trunick, P. A. (1992): Carving a Niche in Global Logistics, in: Transportation & Distribution, Vol. 33 (2): pp. 57-58. Tsikriktsis, N. (2005): A review of techniques for treating missing data in OM survey research, in: Journal of Operations Management, Vol. 24 (2005): pp. 53-62. Udy, S. H. (1959): "Bureaucracy" and "Rationality" in Weber's Organizational Theory. An Empirical Study, in: American Sociological Review, Vol. 24: pp. 791-795.

References

347

van Laarhoven, P., M. Berglund and M. J. Peters (2000): Third-party logistics in Europe - five years later, in: International Journal of Physical Distribution & Logistics Management, Vol. 30 (5): pp. 425-442. Venkatraman, N. and V. Ramanujam (1986): Measurement of Business Performance in Strategy Research: A Comparison of Approaches, in: Academy of Management Review, Vol. 11 (4): pp. 801-814. Virum, H. (1993): Third Party Logistics Development in Europe, in: Logistics and Transportation Review, Vol. 29 (4): pp. 355-361. Wallenburg, C. M. (2003): Kundenbindung in der Logistik: Eine empirische Untersuchung zu Geschäftsbeziehungen zwischen Logistikdienstleistern und ihren Kunden, in: J. Weber and J. M. Deepen: Erfolg durch Logistik: Erkenntnisse aktueller Forschung. Bern, Stuttgart, Wien. Wallenburg, C. M. (2004): Kundenbindung in der Logistik - Eine empirische Untersuchung zu ihren Einflussfaktoren. Bern, Stuttgart, Wien. Walsh, D. A. (1982): The development of visual information processes in adulthood and old age, in: F. I. M. Craik and S. Trehub: Aging and cognitive processes. New York: pp. 99-125. Wathne, K. H. and J. B. Heide (2000): Opportunism in Interfirm Relationships: Forms, Outcomes, and Solutions, in: Journal of Marketing, Vol. 64 (4): pp. 36-51. Weber, J. (1986): Zum Begriff Logistikleistung, in: Zeitschrift für Betriebswirtschaft, Vol. 56 (12): pp. 1197-1212. Weber, J. (1987): Logistikkostenrechnung. Berlin, Heidelberg, New York, London, Paris, Tokyo. Weber, J. (1995): Logistik-Controlling: Leistungen - Prozeßkosten Kennzahlen, 4. Edition. Stuttgart. Weber, J. (1999): Stand und Entwicklungsperspektiven des LogistikControlling, WHU Research Paper No. 61. Vallendar. Weber, J. (2002): Logistikkostenrechnung, 2. Edition. Berlin.

348

References

Weber, J. (2003): Macht Logistik erfolgreich? - Konzeptionelle Überlegungen und empirische Ergebnisse, in: Logistik Management, Vol. 5 (3): pp. 11-22. Weber, J. and J. M. Deepen (2003): Erfolg durch Logistik - Erkenntnisse aktueller Forschung. Bern, Stuttgart, Wien. Weber, J. and S. Kummer (1994): Logistikmanagement: Führungsaufgaben zur Umsetzung des Flußprinzips im Unternehmen. Stuttgart. Weber, J. and S. Kummer (1998): Logistikmanagement, 2. Edition. Stuttgart. Weber, J. and C. M. Wallenburg (2004): Zusatzbeauftragung von Logistikdienstleistern - Empirische Ergebnisse und konzeptionelle Überlegungen zu entsprechenden Defiziten, in: Logistik Management, Vol. 6 (4): pp. 34-46. Weitz, B. A. and S. D. Jap (1995): Relationship Marketing and Distribution Channels, in: Journal of the Academy of Marketing Science, Vol. 23 (4): pp. 305-320. Wentworth, F. (2003): Outsourcing Services: The Case Against, in: Journal of the Institute of Logistics & Transport, Vol. 5 (2): pp. 57-59. Werner, H. (1997): Relationales Beschaffungsverhalten: Ausprägungen und Determinanten. Wiesbaden. Wernerfelt, B. (1984): A Resource-based View of the Firm, in: Strategic Management Journal, Vol. 5 (2): pp. 171-180. Wertz, B. (2000): Management von Lieferanten-ProduzentenBeziehungen - Eine Analyse von Unternehmensnetzwerken in der deutschen Automobilindustrie. Wiesbaden. Wetzels, M., K. de Ruyter and M. van Birgelen (1998): Marketing service relationships: The role of commitment., in: Journal of Business & Industrial Marketing, Vol. 13 (4/5): pp. 406-423. Whipple, J. M., R. Frankel and P. J. Daugherty (2002): Information Support for Alliances: Performance Implications, in: Journal of Business Logistics, Vol. 23 (2): pp. 67-81.

References

349

Whybark, D. C. (1990): Education and Global Logistics, in: Logistics & Transportation Review, Vol. 26 (3): pp. 261-270. Wiener, Y. (1988): Forms of Value Systems: A Focus on Organizational Effectiveness and Cultural Change and Maintenance, in: Academy of Management Review, Vol. 13 (4): pp. 534-545. Wiley, D. E. (1973): The Identification Problem for Structural Equation Models with Unmeasured Variables, in: A. S. Goldberger: Structural Equation Models in the Social Sciences. New York: pp. 69-83. Williams, L. J. and S. E. Anderson (1991): Job Satisfaction and Organizational Commitment as Predictors of Organizational Citizenship and In-Role Behaviors, in: Journal of Management, Vol. 17 (3): pp. 601617. Williamson, O. E. (1975): Markets and Hierachies: Analysis and Antitrust Implications. New York. Williamson, O. E. (1979): Transaction-Cost Economics: The Governance of Contractual Relations, in: Journal of Law & Economics, Vol. 22 (2): pp. 233-261. Williamson, O. E. (1981): The Economics of Organization: The Transaction Cost Approach, in: American Journal of Sociology, Vol. 87 (3): pp. 548-577. Williamson, O. E. (1985): The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. New York. Williamson, O. E. (1990): Transaction Cost Economics, in: R. Schmalensee and R. D. Willig: Handbook of Industrial Organization. Amsterdam, New York, Oxford, Tokyo: pp. 135-182. Wilson, D. T. (1995): An Integrated Model of Buyer-Seller Relationships, in: Journal of the Academy of Marketing Science, Vol. 23 (4): pp. 335345. Wong, A. and A. S. Sohal (2002): An examination of the relationship between trust, commitment and relationship quality, in: International Journal of Retail & Distribution Management, Vol. 30 (1): pp. 34-50.

350

References

Wong, Y.-Y., T. E. Maher, J. D. Nicholson and N. P. Gurney (2000): Strategic Alliances in Logistics Outsourcing, in: Asia Pacific Journal of Marketing and Logistics, Vol. 12 (4): pp. 3-21. Woodward, J. (1958): Management and Technology. London. Young, L. C. and I. F. Wilkinson (1989): The Role of Trust and Cooperation in Marketing Channels: A Preliminary Study, in: European Journal of Marketing, Vol. 23 (2): pp. 109-122. Zand, D. E. (1972): Trust and Managerial Problem Solving, in: Administrative Science Quarterly, Vol. 17 (2): pp. 229-239. Zeithaml, V. A. (1985): The New Demographics and Market Fragmentation, in: Journal of Marketing, Vol. 49 (3): pp. 64-75. Zhu, J., H. S. Lean and S. K. Ying (2002): The Third-Party Logistics Services and Globalization of Manufacturing, in: International Planning Studies, Vol. 7 (1): pp. 89-104. Zimmer, P. (2000): Commitment in Geschäftsbeziehungen. Wiesbaden.