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. Bhler/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. Ltke Entrup
Advanced Planning
in Fresh Food Industries
2005. ISBN 978-3-7908-1592-4
U. M. Lwer
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 Hbner
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. Mller
Martina Bihn

Author
Dr. Jan M. Deepen
Ahornallee 14
40468 Dsseldorf
Germany
jan.deepen@whu.edu

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, broad-
casting, 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

Khne-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
firms logistics performance, which in turn positively influences the over-
all 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 cus-
tomer is of particular importance. So far, little to nothing is known on the
detailed influence of this relationship on the outsourcing performance. Ex-
actly here is the starting point of this dissertation. It aims at discovering the
performance effects that arise from the design of the outsourcing relation-
ship. 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 out-
sourcing performance model developed by Deepen through extensive con-
ceptual work in very large parts is supported by the empirical data. It con-
sequently allows a very differentiated insight into the dependence between
the different factors. Most of the hypothesized relationships are highly sig-
nificant. Almost two-thirds of the variance of goal achievement is ex-
plained through the model, and so is more than one-third of the variance of
goal exceedance. This is very encouraging. Viewing the derived hypothe-
ses, it can be observed that the major part finds support. Only three hy-
potheses had to be rejected, nine could not be tested. These results are evi-
dence of a very profound model formulation.
The testing of the relationship between outsourcing performance and lo-
gistics performance also provides some very encouraging results. All four
hypotheses find support. After all, more than ten percent of the level of lo-
gistics services and more than seven percent of the level of logistics costs
VIII Foreword

are explained. Given the multitude of potential influencing factors of logis-

tics 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 con-
cern the effect of the level of logistics costs, which is now shown to posi-
tively influence all three dimensions of firm performance. Deepens disser-
tation 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 as-
pects. It is characterized through excellent methodological skills and a pre-
cise and consistent theoretical development of the empirical research. The
dissertation is built upon a very extensive literature review, which in par-
ticular 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. Jrgen Weber

Preface

It is a common misconception that writing a dissertation largely is a soli-

tary endeavor in my case it certainly has not been. At the Khne-Center
for Logistics Management at the WHU Otto Beisheim School of Man-
agement 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. Jrgen Weber. I would like to thank him for
his guidance, support, and the freedom and responsibility he gave me dur-
ing 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 customers
view on third party logistics relationships, Davids takes the service pro-
viders view. This allowed us not only to share the survey data, but also to
at least daily hold telephone conferences about the progress made, occur-
ring 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 Khne-Center
for Logistics Management, has also substantially contributed to the success
of this dissertation. He has not only always been a very fair critical coun-
terpart 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 al-
ways 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 spon-
sors of my work: The Khne-Foundation, whose support is the basis for all
activities at the Khne-Center for Logistics Management. My appreciation
goes to Klaus-Michael Khne, 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 Khne-Center was
critical for my dissertations success. This is largely due to the people I
met during my time and thanks go to Dr. Christoph Engelbrecht, Dr. An-
dreas Bacher, Dr. Marcus Groll, Dr. Alexander Schmitt, Nils Daecke, Ul-
rich 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 Nout-
sia, and Sonja Schmitt.
The Khne-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. Mas-
cha Sorg, and Dr. Marius Lissautzki.
I am especially grateful to Claudia Warning and Ekin Balik. They al-
ways 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 disserta-
tion 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 sched-
ule 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.

Dsseldorf, 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 re-
cent 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 approxi-
mately 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 global-
ization 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 todays
global supply chains will be largely driven by their continued ability to
provide value to their customers. This value arises from both accommodat-
ing and exceeding customer service expectations in a more cost effective
manner than can be achieved by customers performing the activities them-
selves. While some companies have made the decision to maintain control
of logistics activities, numerous others worldwide have decided to out-
source these activities.
As recently demonstrated by DAUGHERTY/STANK/ELLINGER
(1998), DEHLER (2001), and STANK/GOLDSBY/VICKERY/SAVITS-
KIE (2003), a companys logistics performance has an influence on the
overall firm performance. ENGELBRECHT (2004) contributed to the dis-
cussion by showing that indeed, the outsourcing performance is an impor-
tant driver of the logistics performance. Thus, clearly understanding the
drivers of logistics outsourcing performance is critical knowledge for man-
agers to have in todays 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 logis-

tics 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 re-
search 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. Fur-
thermore, 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 per-
formance 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 num-
ber 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 pre-
cise insights. All of them are multi-dimensional constructs, in each case
focusing on different aspects of outsourcing performance. This makes con-
centration and consolidation necessary, keeping in mind the different as-
pects of logistics outsourcing and everyday business reality.
The logistics outsourcing decision has been identified as a key consid-
eration 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 competen-
cies, 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 reduc-
tions 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 meas-
urement, 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 re-
search conducted in previous years, consolidates it and develops new mod-
els that allow further insights for both academic research as well as the lo-
gistics 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 lo-
gistics 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/ZLLNER 1997). Answering
these questions, however, is of particular relevance as especially in the lo-
gistics 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 contin-
gency models of logistics performance which should include factors such
as the environment, the logistical features of the product range or the pro-
duction technology. While the contingency approach has been criticized in
the past for its mechanistic view that there is only one best structural an-
swer to any specific contextual situation and its lack of theory (HAGE
1974; SCHREYGG 1980), it can contribute to new insights. If ade-
quately used as a measure to understand the environmental contingency of
firms, it can point out that under certain circumstances, differences in or-
ganization structure and administrative practices have different implica-
tions for a firms performance.
4 1 Introduction

1.2 Goal

The intention of this study is to contribute to the logistics outsourcing dis-

cussion 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 re-
search questions which were developed to break down the complex prob-
lem into several smaller parts that can be individually addressed. The re-
search questions are derived from current research deficits and are in detail
discussed in chapter 2.4. In the survey, the relationship between a cus-
tomer 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 logis-
tics outsourcing research (DAUGHERTY/STANK/ELLINGER 1998;
STANK/GOLDSBY/VICKERY/SAVITSKIE 2003; KNEMEYER/MUR-
PHY 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 ap-
proaches exist to choose from. On the other hand, the customer has a par-
ticularly strong interest in the implications of the relationship variables for
the logistics outsourcing performance, which in turn supposedly influences
the firms logistics performance. Consequently, the main influencing fac-
tors of the relationship must be identified and their effects on the outsourc-
ing performance examined.
Additionally, it is of particular interest to see how these different vari-
ables might be integrated into one model that comprises the relationship
variables as well as their performance effects for the outsourcing of the lo-
gistics activities. Here, special consideration must be given to the interde-
pendencies existing between the variables in order to adequately estimate
the causal linkages inside the model and to the design of the logistics out-
sourcing performance construct.
The question of the performance effects of the relationship design be-
tween 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 particu-

lar relevance if it could be shown to positively affect the overall firm per-
formance. 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 out-
sourcing performance construct developed in this study.
Finally, the question of the validity of the answers to the research ques-
tions that deal with the issues presented above must be addressed with re-
spect to the context of the firms. Since the individual context may vary de-
pending on the firms specific and individual situations, some of the causal
linkages in the models are presumably moderated by those contingency va-
riables. It therefore is necessary to first identify contingency variables with
potential moderating effects. In a second step, their influence on the logis-
tics 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 sug-

gested by MENTZER/KAHN (1995) for theoretically rigorous logistics re-
search. In chapter 1, the motivation for the research is outlined. Building
on that, the four research questions that guide this study are presented be-
fore the structure of the study is discussed in detail.
Chapter 2 introduces the basic concepts of this study. Aside from logis-
tics, its origin, development and the status quo of research, logistics out-
sourcing and logistics outsourcing relationships are presented. To approach
the logistics outsourcing discussion, first its origin and definition is intro-
duced. 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 re-
search 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 relation-
ships. 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 mar-
keting. 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 an-
swer 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 per-
formance are discussed. Initially, the measurement of logistics outsourcing
performance is scrutinized and a bi-dimensional construct is conceptual-
ized. Then, the ten most relevant relationship antecedents of logistics out-
sourcing performance are derived from the literature review and the theo-
ries 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 vari-
ables 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 hypothe-
sized before the effects of the latter on firm performance follow, thus pro-
ducing a model of the entire logistics performance chain. Chapter 4.5 fi-
nally 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 repre-
sentativeness 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 de-
sign and modifications are outlined.
The operationalization of the different constructs conceptualized in
chapter 4 is performed in chapter 6. Starting with the antecedents of logis-
tics outsourcing performance in chapter 6.1, all constructs of logistics out-
sourcing-, 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 opera-
tionalized.
 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 perform-
ance 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 conse-
quence 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 signifi-
cant 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 man-
agement 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 sup-
ply 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 flow-
orientation and thereby emphasize conceptual components.
The apparent differences in the understanding of logistics and the result-
ing myriad of definitions are a direct consequence from logistics non-
academic origin. It rather developed as a practical phenomenon which has
gained increasing importance over the past decades for firms in various in-
dustries. 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 pub-
lic 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 bridg-
ing 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 man-
agement concept which, according to PFOHL (1999, pp. V-VI), ought to be
anchored throughout the organization with direct consequences for all cor-
porate functions.

Level of logistics
knowledge Logistics as
supply chain management
4

Logistics as enabler of process

3 orientation within the whole firm

Logistics as
2 coordinative function

Logistics as
1 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 de-
rived from empirical research on the development of logistics in a corpo-
rate context. Most of these concepts indicate that the development of logis-
tics follows three or four distinct phases (BOWERSOX/DAUGHERTY 1987;
WEBER/KUMMER 1998; GPFERT 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 coor-
dination 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 process-
orientation 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 in-
creasing 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 proc-
esses of information exchange are of at least equal importance. In the fol-
lowing 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 ade-
quately addressed.
Historically, the emergence of the first phase of logistical development
was caused by a severe change in the market environment in the 1950s
(DEHLER 2001, p. 13). The traditional suppliers markets turned into buy-
ers 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 respon-
sibilities 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 ob-
tained, 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 ex-
perience curve effects or economies of scale. Furthermore efficiency gains
can be realized on the planning level through the application of mathe-
matical 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 exist-
ing interdependencies are taken into account, e.g. when rising costs in-
curred through higher transportation frequencies are offset by savings
through lower inventory levels.
On the organizational level, a specialization of the logistics function of-
ten leads to the introduction of new departments, combining transporting-,
handling- and warehousing functions. At the same time, a functional divi-
sion 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 exis-
tence 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 dur-
ing 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 fo-
cus towards the entire supply chain, cutting across the boundaries of the
firm and comprising customers as well as suppliers (WEBER 1999, pp. 3-
4).
Starting point for understanding logistics as a coordinative function was
the insufficient consideration of existing interdependencies between differ-
ent functions of the firm. Facilitated by existing structures, especially pro-
curement, production, and distribution functions were optimized independ-
ently. The organizational separation of these functions, however,
historically encouraged the development and cultivation of individual in-
terests, 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-in-
time supply and production, where the required resources are provided ex-
actly when needed. Resulting from the integrated understanding and plan-
ning of the procurement and production functions, cost and performance
benefits emerge.
The focus thus is on influencing the extent and the structure of the de-
mand 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 in-
creased heterogeneity of the function on the one hand and on the other re-
quires an increased interaction with the responsible management of other
functions. The perceived importance of logistics increases during the sec-
ond 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 tar-
geted mainly during the following phases. The second phase is building on
the know-how of the functional specialization, supplemented by substan-
tial inter-organizational and management knowledge needed for the coor-
dination. 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 charac-

terized 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). Histori-
cally, this development was caused by the changing economic environ-
ment. The increasing competitive pressure called for differentiation while
simultaneously reducing costs. For doing so, the purely functionally de-
signed structures and systems proved inapt. Yet, by adopting a stronger
process orientation when supplying logistical services, complexity reduc-
tions 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 employ-
ees rather than the rigid coordination through plans (KIESER/KUBICEK
1992, pp. 106-117).
Because of the transition into a management function, the implementa-
tion 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 sig-
nificance. 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 respon-
sible for supplying production- or maintenance-activities.
In practice, corporate logistics following this understanding are some-
times 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 ade-
quately anchored in the organization. Consequently, the functional spe-
cialization must not necessarily be abandoned when a firm progresses to-
wards the third phase of logistical development. Rather, it is vital to find a
compromise which enables and fosters the coexistence of a functional spe-
cialization for the supply of logistical services and at the same time an-
chors 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 re-
mains a management function, but extends its scope beyond the boundaries
of the firm. Consequently, the concept of process or flow orientation is ex-
tended across the supply chain, encompassing now also suppliers and cus-
tomers, thus ideally spanning from source to sink. Logistics during this
phase, now being called supply chain management (SCM), aims at inte-
grating the entire supply chain.
This understanding of the concept of supply chain management as a
phase of logistical development is not undisputed. As LARSON/ HALL-
DORSSON (2004, pp. 1-7) point out, in the logistics science community ba-
sically four different views of SCM have developed over the years. These
include the traditionalist view which understands SCM as part of logis-
tics and the unionist view which considers logistics as part of SCM. Fur-
thermore, 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, pur-
chasing 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 retrospec-
tive 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 effec-
tiveness. Since most of the internal optimization potentials had already
been exhausted, only those remained that resulted from the inefficient col-
laboration 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 tremen-
dous progress the information and communication technologies made.
Even though supply chains are part of every economy based on the divi-
sion 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 con-
centration 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 func-
tions 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 inter-
organizational 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. Addi-
tionally, endeavors beneficial for one supply chain can be detrimental for an-
other. 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 inter-
organizational flow orientation not only is the answering to the technologi-
cal demands, but also the sufficient willingness and capabilities of the par-
ticipating 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 con-
ducted 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 contin-
gency factor, while they did not address the same sample of firms and in-
dustries. A comparison of the two studies, as WALLENBURG (2004, pp. 44-
45) 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 im-
portance. Furthermore, only 12% of the firms see themselves in the fourth
phase of logistical development and concentrate on supply chain manage-
ment and the implementation of inter-organizational flow orientation.
These results, however, have their own dynamics and are subject to con-
stant 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 12
90
11
80 20
Phase 4
70 57
40 Phase 3
60 82
33 Phase 2
50

40 Phase 1
13
30

20 42 35 22 5
10 8
8 6
1999 2002 1999 2002
Current logistics Aspired logistics
development 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 lo-

gistics can be observed in recent years. It remains an open question, how-
ever, 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 im-
plement logistics as a management function, thereby enabling an inter-
organizational 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 sam-

ple of firms.
This finding is of particular relevance, because DEHLER (2001, pp. 233-
244) 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 fi-
nancial 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 0.47***
Logistics Services
R2: 66.2%
0.47***
Market
Performance

Level of 0.50***
Logistics Costs
R2: 40.6%
0.24***
Financial
Performance
0.65*** Standardized regression weight with significance level
* Significant on 10%-level
** Significant on 5%-level
*** Significant on 1%-level

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 direc-
tion 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 indi-
vidual firm. The options for the firms are to either operate them by them-
selves 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 de-
tail in chapter 2.2.3.2.
The following chapter will first highlight the origin of logistics out-
sourcing and provide a definition, before looking into its benefits and risks.
Then, the current scope of logistics outsourcing in practice will be pre-
sented 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 at-

tention 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; RAZZA-
QUE/SHENG 1998). In the early discussion, different views of the meaning
of logistics outsourcing became apparent. LIEB/MILLEN/VAN WASSEN-
HOVE (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 lim-
ited 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 out-
sourcing is the use of a third-party provider for all or part of an organi-
zations logistics operations and add that its utilization by the firms is in-
creasing. RABINOVICH/WINDLE/DRESNER/CORSI (1999, p. 353) define
logistics outsourcing relationships even more broadly as long and short-
term 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/EMMEL-
HAINZ/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. Ac-
cording 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 global-
ization of business. The continuously growing global markets and the ac-
companying sourcing of parts and materials from other countries has in-
creased 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. 140-
141), 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 outsourc-
ing:
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 char-
acter, 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 ac-
cording to the individual firms perception of the benefits and risks associ-
ated with the particular outsourcing arrangement. Although they are inher-
ently 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 firms 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 logis-
tics services a LSP produces (SCHFER-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 in-
dustries (STLZLE 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. 30-
32), RICHARDSON (1992, p. 22) and RICHARDSON (1995, p. 61) point out
that investments in facilities can be reduced while SHEFFI (1990, pp. 27-
39), 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, logis-
tics outsourcing also allows for a decrease of the workforce and the associ-
ated 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 in-
vestments 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 the-
ses 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 logis-
tics 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, opti-
mized 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 re-
spect 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 facili-
tates 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, un-
expectedly 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 con-
text and the business environment is mentioned by LYNCH (2000b, pp. 9-
11), 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, realiza-
tion came over the years that outsourcing is accompanied by some disad-
vantages 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 qual-
ity and service have been achieved or not (WENTWORTH 2003, pp. 57).
The same holds true for the LSPs 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 re-
duce the complexity of business processes, enabling the firm to concen-
trate on its core business. It must be noted, however, that in the relation-
ship 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 seri-
ous obstacle en route to successful outsourcing. The importance of consid-
ering 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 re-
duction 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 gen-
eral 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 mar-
ket 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 im-

mense 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 abso-
lute 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 Eco-
nomics 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 in-
dustries automotive, consumer goods, and retail ranges between 40% and
50% with the consumer goods industry leading the way. This is an consid-
erable 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 ex-
pand their outsourcing activities. This is consistent to the findings of
LANGLEY/ALLEN/DALE (2004, p. 9) who state that current logistics ex-
penditures directed towards outsourcing in Western Europe are 61% in
2004 and will grow to 68% in the period between 2007 and 2009. The ana-
log numbers for the important logistics market in the U.S. project an in-
crease 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 (2008-
2010). 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 out-
sourcing 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 to-
wards involvement with logistics service providers in general or that the
growth rate for the extent of use of logistics services provided by third par-
ties 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 na-
ture, such as warehousing, transportation, handling, or customs clearance,
while activities more strategic in nature or activities that are directly cus-
tomer-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 in-
creasing outsourcing of more complex activities can be observed, espe-
cially in previously neglected areas close to production and with a distinct
emphasis on the already outsourcing dominated procurement and distribu-
tion activities.
The estimates provided above demonstrate the significance of the mar-
ket for third party logistics services. Since the market has seen strong de-
velopment over the last two decades, the potential for further increase is
mainly limited to areas which so far have not been focused, such as activi-
ties 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) in-
dicates. 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 ex-
panding rapidly. Since complete outsourcing of all logistics tasks does not
make sense in most cases (WALLENBURG 2004, p. 49), the degree of out-
sourcing 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 provid-
ers 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 im-
portant 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 ser-
vices offered by the LSPs are adapted to the needs of the respective cus-
tomers. They range from a narrow spectrum, mainly consisting of ware-
housing 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 ser-
vice 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 lo-
gistics development.
With increasing sophistication of logistics processes, freight forwarders
address the growing needs of the customers by offering coordinating func-
tions and intermediating services. They bundle transportations services, of-
fer 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 transpor-
tation planning and management including providing the associated infor-
mation systems and also sometimes act as carriers by using own asset for
transportation or warehousing. Overall, freight forwarders in their coordi-
nating 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 in-
ter-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 distri-

bution services, the CEP offer integrated services that ensure the distribu-
tion 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, or-
der 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 com-
plex, 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 provid-
ers (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 firms 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 man-
agement and execution of transportation and warehousing []. In addi-
tion, other activities can be included []. Also, [] the contract [is re-
quired] 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 custom-
ers analyzed in the latter must fall under this definition in order to ade-
quately distinguish between advanced logistics outsourcing relationships
and traditional arms lengths sourcing of transportation and/or ware-
housing.
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 mar-
kets 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. There-
fore, 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% in-
volves hypothesis testing. According to MENTZER/KAHN (1995) this is ur-
gently 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 in-
formation 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 disci-
pline 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 be-
cause 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 dis-
cussions 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 func-
tional 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 en-
able the optimization of entire process-chains rather than isolated proc-
esses.
Overall, logistics research has traditionally been confronted with nu-
merous operative and practical problems (WALLENBURG 2004, p. 52), re-
quiring 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 logis-
tics 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 al-
ready been summarized in great detail by several articles (MURPHY/POIST
1998; RAZZAQUE/SHENG 1998; MURPHY/POIST 2000), the following sec-
tion will concentrate on the evolving research published since then (Table
2-1).

Table 2-1. Key logistics outsourcing related research since 1999

Author(s) Key logistics outsourcing related finding
BHATNAGAR/SOHAL/MILLEN Cost savings, customer satisfaction, and flexi-
(1999) bility are the main reasons for outsourcing
BOYSON/CORSI/DRESNER/ Logistics outsourcing is effective in enabling
RABINOVICH (1999) firms to achieve competitive advantage, im-
prove their customer service levels, and reduce
their overall logistics costs
VAN LAARHOVEN/BERG- Some outsourcing relationships are perceived
LUND/PETERS (2000) by the shipper as more successful than others.
Some success factors are: well defined re-
quirements, procedures, systems, and close re-
lationships
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 logis-
tics 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 relation-
ships evolve, a gap exists between what cus-
tomers receive and what they expect to receive
STANK/GOLDSBY/VICKERY/ Relational performance by LSPs is the most
SAVITSKIE (2003) important factor engendering customer satisfac-
tion
LANGLEY/ALLEN/DALE (2004) While 3PL users generally feel that their rela-
tionships with 3PL providers are successful,
they are able to find areas of improvement, such
as implementing capable IT, instituting effec-
tive management and relationship processes and
integrating services and technologies globally.
Customer demands for performance and sophis-
tication 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/GNTHER/ Logistics providers must cope with growing
DANGELMAIER (2005) uncertainty, increasing cost pressure, and more
complex value chains and at the same time ful-
fill increasingly changing customer demands
KNEMEYER/MURPHY (2005) Relationship characteristics such as communi-
cation with the 3PL provider have a more pro-
found impact on logistics outsourcing relation-
ship 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 se-
lecting a 3PL provider

The studies between 1999 and parts of 2003 (BHATNAGAR/SOHAL/

MILLEN 1999; BOYSON/CORSI/DRESNER/RABINOVICH 1999; VAN LAAR-
HOVEN/BERGLUND/PETERS 2000; BAUMGARTEN/WALTER 2000; BAUM-
GARTEN/THOMS 2002; DEEPEN 2003) offer very valuable insights into lo-
gistics outsourcing, but must in large parts be subject to the same criticism
as voiced by MENTZER/KAHN (1995, pp. 240-244) . They are largely de-
scriptive 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. Find-
ings 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 projects 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), KNE-
MEYER/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 lo-
gistics 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. Neverthe-
less, the explanatory value of the degree of outsourcing is quite low. It
must therefore be assumed, that the true drivers behind logistics outsourc-
ing 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 ex-
pression of the achievement of the goals previously set for the outsourcing
project. In a recent study, KNEMEYER/MURPHY (2004, pp. 45-46) further-
more demonstrate by using SEM that various relationship marketing di-
mensions, such as trust, communication and opportunistic behavior, influ-
ence 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; KNE-
MEYER/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. Further-
more, 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 ad-
vance 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/ZLLNER 1997). As already stated in chapter 2.2.3, it is vir-
tually impossible to make a generalized and universally valid recommen-
dation 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 pro-
jects 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 out-
sourcing 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 relation-
ships can, if they turn out successful, give both parties a competitive ad-
vantage 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 re-
lationship. While several models exist that support management during the
decision for or against outsourcing and sometimes even also specify crite-
ria 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 relation-
ships, the next chapters will sequentially present the most important as-
pects, 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 chap-
ter 2.3.2.2 will introduce the current research on the process of partnership
development. After thus having established the need for sustainable rela-
tionships also in the logistics outsourcing context, chapter 2.3.2 will exam-
ine 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 infor-
mation of chapter 2.3 will thus enable the understanding of the status quo
of logistics outsourcing relationship research and the identification of con-
sequential 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 li-
censing, cooperative research or outsourcing agreements. While this, ac-
cording 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, work-
ing partnerships aiming at mutual success are developing. Those they de-
fine 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 fo-
cused on jointly satisfying the requirements of the customer marketplace.
Another term used frequently especially in the supply chain manage-
ment context is collaboration. According to STANK/KELLER/DAUGHERTY
(2001, p. 31) it is a process of decision making among interdependent par-
ties, 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 deci-
sion 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/HEND-
RICK 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 in-
clude 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) there-
fore 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, ad-
dressing both behavioral and monetary aspects, and thus is displaying ex-
actly the understanding of partnerships in this study. This is especially im-
portant with respect to the later development of a logistics outsourcing
relationship model in chapter 4, which will contain a number of character-
istics that originate from this understanding of partnerships. The following
chapters will in detail analyze the development of such partnerships. Spe-
cial attention will first be given to the importance and the effects of part-
nership 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 sim-

ple 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 relation-
ships 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 interorgani-
zational relationships today are commonly referred to as arms-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 re-
lational exchange. According to DWYER/SCHURR/OH (1987, p.12) the idea
of a discrete transaction is the foundation on which the concepts of rela-
tionships 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 ex-
cludes all relational elements and therefore will occur only very rarely in
business relations.
Discrete transactions are an important prerequisite for arms length rela-
tionships. 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 relation-
ship between them will emerge and pure arms 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 arms 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 fac-
tors 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 part-
nership were the benefits expected by the customers. Through their work-
ing 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 lo-
gistics 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 partner-
ships 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 customers
management to make the partnership fail or unprofitability for the seller,
and subsequent service failures. Further factors leading to partnership fail-
ure 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 part-
ners she found that the most prominent factors include poor communica-
tion, 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 cate-
gories suggested by STUART/MCCUTCHEON (1995, pp. 5-6). They identify
perceptual mismatches over the appropriate degree of partnering and fail-
ures in the building process of the relationship as central factors of partner-
ship 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 partner-
ship 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 estab-
lishment 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 produc-
tivity 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 part-
ners.
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 proc-
ess. They criticize, however, that while these models are good tools for as-
sessing 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 arms length contract or a partnership and at the
same time assume that all partnerships are the same. They continue by ob-
serving that while managers mostly decide correctly that outsourcing is
needed, they often make incorrect decisions on the specific type of rela-
tionship 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 mainte-

nance 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) fol-
lows 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 driv-
ers 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 partner-
ship. They include planning, joint operating controls, communications, risk
and reward sharing, trust and commitment, contract style, scope, and fi-
nancial investment. Once it is determined that a particular degree of part-
nership 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 compo-
nents.
The model of LAMBERT/EMMELHAINZ/GARDNER (1999) is suitable to
address the issue of partnership configuration once the outsourcing deci-
sion has been made. It supplies valuable insights on the design of individ-
ual relationships and thereby helps to answer the question how close a re-
lationship 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. There-
fore, 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 perform-
ance effects.
This demand is consistent with that of KNEMEYER/MURPHY (2004, p.
38) who state that the potential causal linkages between relationship di-
mensions such as trust and communication and successful third party lo-
gistics 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 devel-
oped 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 devel-

opment 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 be-
tween the different parties involved ought to be designed.
According to SHETH/PARVATIYAR (1995, p. 399), some social psy-
chologists suggest that competition is inherently destructive and mutual
cooperation is inherently more productive. While few authors such as
PACH (1998, p. 308) doubt this, suggesting that the adversarial approach
to logistics relationships may offer some advantages, because e.g. coopera-
tion may slow down contractors improvements, a broad majority of schol-
ars embraces the idea of the positive effects of cooperation on relation-
ships.
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 (GRNROOS/GUMMESSON 1985; GRNROOS
1989; GRNROOS 1990; GRNROOS 1994; GUMMESSON 1987; GUMMES-
SON 1997). According to MORGAN/HUNT (1994, p. 22), relationship mar-
keting refers to all marketing activities directed toward establishing, de-
veloping, and maintaining successful relational exchanges. As CRAVENS
(1995, pp. 48-57) points out, relational marketing as opposed to transac-
tional 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 re-
search 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 accompa-
nying performance effects on the one hand and the design of these rela-
tionships on the other. As STOCK (1997, p. 528) puts it, a potential applica-
tion 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 market-
ing models that offer insights into the adequate design of relationships. Af-
ter 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 in-
sights 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), HALLN/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 in-
troduced 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 continu-
ity in conventional industrial channel dyads. They formulate a model in
whose center are trust, the degree of communication, and power imbal-
ance. Further constructs include goal congruence, cultural similarity,
stakes, age of the relationship, and the perceived continuity of the relation-
ship. 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 congru-
ence 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 influ-
ences 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 rela-
tionships 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, func-
tionality of conflict and the influence both firms have on one another. Af-
ter 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 manufac-
turing firms. Among the most noteworthy results are that from the distribu-
tors 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 secon-
dary 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 un-
important 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 re-
search.
In their model, MORGAN/HUNT (1994) propose several precursors of re-
lationship commitment and trust in the form of shared values, communica-
tion, opportunistic behavior, relationship benefits, and relationship termi-
nation costs. Commitment and trust in turn impact the five outcome
dimensions acquiescence, propensity to leave, cooperation, functional con-
flict and uncertainty.
Empirically, MORGAN/HUNT (1994, p. 30) can show that with few ex-
ceptions, the proposed causal linkages prevail. The most notable findings
include that trust is influenced by shared values, communication, and op-
portunistic behavior while it in turn influences cooperation, functional con-
44 2 Basic concepts

flict, uncertainty, and relationship commitment. Commitment is further in-

fluenced by shared values, relationship benefits and relationship termina-
tion costs and leads to acquiescence, cooperation and reduces the propen-
sity to leave.
The work of MORGAN/HUNT (1994) was of particular importance to re-
lationship 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 intro-
duced 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, al-
ternatives, adaptation, non-retrievable investments, shared technology,
summative constructs, structural bonds, and social bonds. Even though al-
most a decade has passed since the articles publication, these factors re-
main 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 de-
signed, 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 per-
formance, no statistically sound proof has been put forward. In the years
since the publication of the article by HEIDE/STUMP (1995), some addi-
tional, yet not fundamentally new research, has been produced. Conse-
quently, 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 reluc-
tantly 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 logis-
tics 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 in-
volved are making commitments to establish closer and longer-term rela-
tionships.
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 re-
lationship.
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 moti-
vated 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 de-
mands, 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 market-
ing literature with the individual characteristics proposed by logistics re-
search, 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 partys 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 effec-
tiveness, 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 pre-
vail are that the third partys relationship commitment influences the
buyers relationship commitment positively, while it is influenced nega-
tively by the buyers relationship conflict. Furthermore, MOORE (1998) can
show that the third partys information exchange reduces the relationship
conflict while only the third partys equity enhances the buyers trust in the
third party. Increasing trust, in turn, positively influences the buyers rela-
tionship effectiveness which is also negatively affected by relationship
conflict.
While these findings suggest that both trust and commitment are impor-
tant factors in relationships between customers and LSPs, according to
MOORE (1998, p. 33) they seem to suggest that in terms of logistics alli-
ance 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 lo-
gistics relationship research, a number of questions remain to be answered.
Additional to the fact that several reasonable hypotheses suggested by pre-
vious research could not be supported, the connection between trust, com-
mitment, 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 con-
clusions 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 opera-
tions 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 communica-
tion 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 part-
nership success exists. The model furthermore provides support for the hy-
potheses that communication is positively affecting trust which in turn is
linked to a positive asset reduction performance. Additionally, all antece-
dents 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 per-
formance, the insights they offer for designing successful logistics out-
sourcing 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 implementa-
tion 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 dif-
ferent 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 posi-
tively by the improvement efforts of the LSP. He furthermore shows that
higher levels of improvement efforts of the LSP lead to more communica-
tion, 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 rela-
tional marketing, it holds two important insights. The first is that the im-
plementation of logistics outsourcing projects indeed has a significant in-
fluence 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 con-

structs such as improvement efforts of the LSP seem to have a consider-
able influence as well.
The preceding chapters have in detail analyzed the importance and ef-
fects of partnerships, have introduced models for the process of partnership
development and have given insights into the designing of logistics out-
sourcing relationships. For this, the status quo of relationship marketing
research and its consequences for logistics outsourcing research were pre-
sented. During the discussion, a number of research needs and open ques-
tions were identified. The following chapter presents a summary of those,
identifies the most important research needs and on these grounds devel-
ops 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 un-
derstanding of the underlying mechanisms that drive logistics performance
and enable firms to gain competitive advantages. Following this develop-
ment the topic of logistics outsourcing has been of particular importance as
an ever increasing number of firms decided to utilize logistics service pro-
viders 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/ACKER-
MAN 1995). Unfortunately, the logistics outsourcing debate for the most
part neglected the importance of the relationship between customer and lo-
gistics service provider and thus cannot answer the question which per-
formance implications relationship formation has for the outsourcing ar-
rangement. 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 logis-
tics activities to third parties. While those firms that currently do not out-
source 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 ori-
ented (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 lo-
gistics 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 relation-
ships comes from research in the field of relationship marketing which ex-
plicitly focuses on this issue. The absence of specific logistics oriented re-
search on this topic of course is only relevant if the general results of
relationship marketing are not easily transferable to the relationships be-
tween customers and LSPs. That the outsourcing of logistics activities is a
complex matter has been established above the specificity of the proc-
esses, 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 con-
tribute to this. A general and uncritical transfer of relationship marketing
findings to logistics research therefore is not possible. Instead, the applica-
bility of key relationship marketing concepts for the logistics context must
be examined in detail and therefore be one significant part of future logis-
tics research.
While the works by MOORE (1998), KNEMEYER/MURPHY (2004) and
ENGELBRECHT (2004) as well as numerous articles from relationship mar-
keting (ANDERSON/NARUS 1984; ANDERSON/LODISH/WEITZ 1987;
ANDERSON/WEITZ 1989; HEIDE/JOHN 1990; NOORDEWIER/JOHN/NEVIN
1990; HALLN/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 in-
vestigate 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. Addition-
ally, the question must be addressed to which extent if any perform-
ance effects can be traced back to the way the relationship between cus-
tomer 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 be-
tween the different factors in some of these works is rather thin and does
not meet the requirements concerning adequate scientific rigor as formu-
lated by MENTZER/KAHN (1995). A future model encompassing the rele-
vant 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 ar-
gue 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, as-
suming 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 ef-
fects of contingency variables. This implicitly assumes that within the re-
spective studies, all analyzed relationships are similar and comparable.
However, this disregards the possibility that the various surveyed relation-
ships 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 re-
lationship 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 de-
veloping new approaches which as yet have not been considered in logis-
tics research.

2.4.2 Identification of research questions

The intention of this work is to analyze the relationships between custom-

ers 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 theo-
retical 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 in-
fluence the firms logistics performance. The first research question there-
fore is:

R1 : Which are the main influencing factors of an outsourcing relation-

ship 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 con-
structs as well as their performance effects for the outsourcing of the logis-
tics activities. Here, special consideration must be given to the interde-
pendencies existing between the variables in order to adequately estimate
the causal linkages inside the model and to the design of the logistics out-
sourcing 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 be-
tween 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 influ-
ence on the logistics performance. This in turn may be of particular rele-
vance if it could be shown to positively affect the overall firm perform-
ance. The third research question therefore reads:

R3 : Which influence does the outsourcing performance have on the lo-

gistics 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 individ-
ual 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 sec-
ond 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 logis-
tics 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 un-
structured 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 theo-
retical 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 ex-
tensive literature review and practical observations. This promoted the
substantive justification which led to the development of research ques-
tions 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 relation-
ship 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 re-
search 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 re-
search questions formulated in chapter 2.4.2 will be addressed and conclu-
sions 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 opti-
mal 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 eco-
nomical 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 effec-
tiveness. Central for the choice of the customer would only be the costs
and the corresponding service levels as well as the competence of the ser-
vice 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 ser-
vice provider must be changed in order to reduce costs or increase the ser-
vice level.
Just the introduction of bounded rationality adds a degree of complexity
to relationship management that severely complicates the issue. It there-
fore 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 mi-
cro-economic theories (transaction cost analysis and principal agent the-
ory) and game theory to social exchange theory, the commitment-trust the-
ory, 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 intro-
duce a selection of theories potentially suited if combined to explain how
relationships between customers and LSPs should be designed to maxi-
mize performance outcomes and to adequately address the largely ne-
glected 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 spe-
cifically 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 re-
quested 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 presenta-
tion of the contribution the theories can make to logistics relationship re-
search.

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 inter-
face. Transaction cost theory builds on the work by COASE (1937), who
identified several limitations to the neoclassical paradigm for understand-
ing relationships between firms. Central to neoclassical economics is the
concept of a single product firm, operating in a perfectly competitive in-
dustry 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 ex-
change 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 in-
formation, hence there is no uncertainty regarding prices, product charac-
teristics 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 discover-
ing the appropriate price, the cost for negotiating contracts for each ex-
change 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 an-
other 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 in-
creased. Groundbreaking work in the development of the theory of transac-
tion costs was carried out by WILLIAMSON (1975) and WILLIAMSON
(1979). Gradually, several other theories based on the concept of transac-
tion 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 effi-
ciency 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. Contract-
ing 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 intellec-
tual 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, be-
cause 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 ac-
cepted 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 character-
ized by the three critical dimensions frequency, uncertainty and asset
specificity. The frequency of transactions has implication for the amortiza-
tion 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 econo-
mies of scale (PICOT 1991, p. 347; ANTLITZ 1999, p. 92). Therefore, in-
creasing frequency in transactions should foster hierarchic coordination.
A similar argumentation is used for uncertainty. The higher the uncer-
tainty 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 attrac-

tive.
According to WILLIAMSON (1990, p. 142), the most important factor in-
fluencing the level of transaction costs is the asset specificity. He defines:
asset specificity has reference to the degree to which an asset can be re-
deployed to alternative uses and by alternative users without sacrifice of
productive value. The higher the sacrifice of productive value of an as-
set 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 de-
termined. 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 trans-
actions costs, the costs for using the market as form of coordination in-
crease quicker than those of a cooperation and hierarchy. Therefore, de-
pending on the characteristics of a transaction, the optimal form of
coordination can be determined.
Transaction cost theory furthermore has many different facets and ap-
plications 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 ex-
planatory 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 depend-
ing on the characteristics of a transaction and suggests circumstances un-
der which long-term relationships are preferable over other forms of coor-
dination for a transaction.
However, according to LAMBE/WITTMANN/SPEKMANN (2001, pp. 2-3)
one of transaction cost theorys 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 sev-
eral 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 theorys basic assumption of universal opportun-
ism, especially in relational exchange (HEIDE/JOHN 1992; MORGAN/HUNT
1994). It must therefore be ascertained that transaction cost theory is lim-
ited in its capacity to explain exchange governance in exchange relation-
ships 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 transac-
tion cost theory for the formation and management of logistics relation-
ships is far-reaching. While the theory is obviously suited to explain why
customers and logistics service providers engage in cooperative arrange-
ments it can only be utilized to derive conclusions on the design of the ac-
tual 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 par-
ties should act within it.
Long-term relationship between firms, based on trust and mutual coop-
eration, 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 prob-
lem solving (ZAND 1972; BOSS 1978), since information is exchanged
freely and more solutions to a problem are being explored because deci-
sion 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 the-
ory allows evaluating whether certain variables are beneficial for the gov-
ernance of logistics relationships. However, as a single theory it is not suf-
ficient 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 ex-
planatory 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 vari-
ables that characterize the situation in which a firms 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 cus-
tomer 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 ac-
companied 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 fre-
quent 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 rela-
tional exchange, researchers of inter-organizational relationships have re-
cently 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 interac-
tion of the exchange partners (DWYER/SCHURR/OH 1987; HALLN/JOHAN-
SON/SEYED-MOHAMED 1991). Social exchange theory thereby acknowl-
edges 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 ex-

change relationships and aim at explaining the behavior of the parties in-
volved (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 relation-
ships, it is not restricted to a particular context. Private relationships be-
tween individuals can be examined as well as buyer-supplier relationships.
Social exchange theory may be traced back to one of the oldest theo-
ries of social behavior which states that any interaction between indi-
viduals 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 par-
ties 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, 1162a34-
1163a24) 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 sys-
tematic theory that focuses on social behavior as [exchange] was de-
veloped 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 contri-
bution to social exchange theory in the form of their comparison level con-
cept, which is used to explain how parties in the exchange relationship
weigh the benefits of the relationship to determine their relationship com-
mitment. 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 relation-
ships 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 part-
ner(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 re-
wards 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 in-
volved in social exchange do not have any material value on which an ex-
act 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 com-
prise 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 rela-
tionship (LAMBE/WITTMANN/SPEKMANN 2001, p. 8), social exchange the-
ory suggests that parties will remain in the relationship as long as the
trade-off between benefits and costs is positive and therefore the satisfac-
tory 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) devel-
oped the concepts of comparison level (CL) and comparison level of alter-
natives (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 be-
low 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 re-
lationship exceeds CLalt, the respective party will have a degree of depend-
ence 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 rela-
tionship 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 ac-
tions 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 impor-
tant 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 re-
ciprocity, which according to MCDONALD (1981, p. 834) holds that mis-
trust breeds mistrust and as such would also serve to decrease commitment
in the relationship and shift the transaction to one of more direct short-
term exchanges. Commitment is important to exchange relationships be-
cause it ensures that partners will make the efforts and investments neces-
sary 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 in-
crease 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 relation-
ships 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 ex-

change. In fact, as WILSON (1995, p. 335) points out, the variables of so-
cial exchange theory have been used to create empirical success models.
These aim at determining the degree to which relational exchange vari-
ables 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 op-
portunism. 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 ex-
change are more vulnerable to the other partys opportunism.
Nevertheless, the research surrounding social exchange theory is espe-
cially interesting for analyzing logistics relationships. As relationship vari-
ables such as trust and commitment operationalize the premises of the the-
ory, 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 ex-
change 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 the-
ory 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 contribu-
tions to the question which variables should be included in relationship re-
search. 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

Table 3-1. Variables frequently used in relationship research

Communication
Commitment

Cooperation
Dependence

Satisfaction

Others
Norms
Trust
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)
68 3 Theoretical framework

TATE (1996)
WALLENBURG (2004)
WHIPPLE/FRANKEL/DAUGHERTY

(2002)
WILSON (1995)
YOUNG/WILKINSON (1989)

As Table 3-1 shows, the variables most often utilized in relationship re-
search include trust, commitment, dependence, cooperation, communica-
tion, norms, and satisfaction. Some of these are also very fitting for logis-
tics relationship research.
LALONDE/COOPER (1989) propose that, as trust evolves between cus-
tomers 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 mu-
tual trust and commitment where both parties seek to establish jointly re-
warding 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 uncer-
tainty through cooperation, thereby allowing them to receive benefits nor-
mally only attainable through vertical integration. This of course requires
the sharing of information, risks, and rewards in order to develop a mutu-
ally 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
relationships overall performance and thereby may serve as an operation-
alization of the success of the exchange relationship. Therefore, the satis-
faction of the parties with the outcome of a logistics outsourcing relation-
ship 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 sub-
stantial amount of research already exists that examines the importance of
dependence for relationships (ANDERSON/NARUS 1990; HALLN/JOHAN-
SON/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 depend-
ence 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 fur-
ther complexity and integrate additional variables such as power and de-
pendence.
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 fu-
ture.

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 im-
plicit exclusion had been a major point of criticism for social exchange
theory.
Origin of the thoughts of MORGAN/HUNT was the observation that mar-
keting must distinguish between discrete transactions and relational ex-
change, which in the time of arms length relationships commonly was ne-
glected (DWYER/SCHURR/OH 1987, p. 13; MORGAN/HUNT 1994, p. 21;
HOYT/HUQ 2000, p. 750). As relational exchange is becoming more im-
portant, relationship marketing is required to adequately address partner-
ship 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 understand-
ing 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 vari-

ables as key determinants for successful relationships, the political econ-
omy paradigm dominated the discussion. As THORELLI (1986, p. 38) main-
tained, 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 distin-
guish productive and effective relational exchanges from those that are un-
productive and ineffective.
While they state that no doubt exists that many contextual factors con-
tribute 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 rela-
tionship performance, according to MORGAN/HUNT (1994 p. 22) is ex-
plained 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 par-
ties to resist attractive short-term alternatives in favor of the expected long-
term 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 oppor-
tunistic 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, commit-
ment 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 partys 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 relation-
ships.
 3.2 Introduction to selected theories 71

Termination
Acquiescence
Costs
+ +
Relationship + _ Propensity to
Benefits Relationship Leave
+ Commitment +

Shared Values + Cooperation

+ +

+ Trust +
Functional
Communication
Conflict
_ _

Opportunistic
Uncertainty
Behavior

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 ar-
guments 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/WITT-
MANN/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 arms 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; KNE-
MEYER/CORSI/ MURPHY 2003; WALLENBURG 2004).
The most remarkable achievement of MORGAN/HUNT (1994) is the in-
tegration of the mediating variables commitment and trust with several an-
tecedents 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 im-

portance 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 relation-
ships it can consequently be concluded that shared values, communication
and opportunistic behavior have strong influences on the success of the re-
lationships.
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 arms 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 rela-
tionship marketing, their explanatory value for logistics relationship re-
search is limited. Uncertainty is, as explained in chapter 3.2.1.2, a contin-
gency 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 ex-
tensively 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 re-
lationship commitment, in logistics relationships take on a different role.
MORGAN/HUNT understand them as the benefits gained from the relation-
ship as compared to an alternative. In logistics outsourcing relationships,
these relationship benefits must be modelled as the outcome or the per-
formance 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 re-
lationships. 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 deter-
mines 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 be-
ing offered by the contingency approach (also sometimes: contingency
theory) which has been extensively used to describe organizational struc-
tures under varying environmental conditions (KLEER 1991, p. 115;
PRASAD/TATA/MOTWANI 2001, p. 32).
However, in order to understand the explanatory value of the contin-
gency 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 1960s
that organizations show significantly different organizational structures
and behaviors. Opposing the dominant belief at the time whereby all or-
ganizations could be described by a few universal laws and theories, UDY
(1959) and HALL (1963) empirically found that substantial differences be-
tween organizations exist. Contingency theory was the attempt of organ-
izational theory to explain these differences (KIESER/KUBICEK 1978, p.
105). The central proposition of contingency theory is, aside from the as-
sumptions that there is no best way to organize and that any way of orga-
nizing is not equally effective (GALBRAITH 1973, p. 2), that differences in
structure and administrative practices of organizations depend on the na-
ture 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. 106-
108). 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 tech-
nology 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 re-
sult 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 vari-
ables simultaneously. The turning away from mono-causal towards multi-
causal 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 ex-
planation of the influence of different organization structures and adminis-
trative practices on the efficiency and effectiveness of organizations is an-
other 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 SCHREYGG (1980,
p. 308) proposes that there is an objective necessity to adapt the system
(the organizations structure) congruently to its critical environmental
fields in order to secure its survival by achieving the critical level of per-
formance. The designer of the organization therefore has to comply, at
least in the long run, with the constraints which characterize the relation-
ship 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 criti-
cized (SCHREYGG 1980, pp. 309-312). A central point of criticism was
the mechanistic view that there is only one best structural answer to a spe-
cific contextual situation. In fact, many studies show that a variety of
structures exist which enable success in a given environment (SCHREYGG
 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 ac-
cept 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 contin-
gency approach (HAGE 1974, pp. 18-19). The approach indeed does not of-
fer law-like generalizations. It rather points out that under certain circum-
stances, 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 of-
ten 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 organi-
zations and its administrative practices on the grounds of the respective
context. The approach therefore must be understood as an idea or a frame-
work 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 influ-

ences of different organizational structures and administrative practices on
the effectiveness and efficiency of organizations. Since the relationship be-
tween a logistics service provider and its customer is an integral part of the
customers 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 con-
tingency 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 organizations environment, product line,
production technology, and size will influence performance outcome vari-
ables 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 partner-
ship 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/ZLLNER (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 sub-
stantial meaning for logistics research. Differentiation is defined as the dif-
ference in cognitive and emotional orientation among managers in differ-
ent 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 schol-
ars 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 in-
creasing logistics outsourcing, the relationship between the customer and
the LSP is gaining ever growing importance and must therefore be the sub-
ject of analysis as established in the research questions of this study (see
chapter 2.4.2).
In their article, PFOHL/ZLLNER (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 ex-
planatory value.
Environmental relations of the organization are one group of these con-
tingency 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 influenc-
ing the organizational structure. PFOHL/ZLLNER (1997, pp. 307-308)
identify the criteria of complexity and dynamics of environmental rela-
tions as the elements of the environment which are particularly relevant
for logistics and provide a large number individual factors (see Figure 3-
3).
 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 prod-
ucts 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 logis-
tically 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 firms products.

Flow of products Flow of information

Complexity of the environmental relations Complexity of the environmental relations

- Number or amount of raw and auxiliary products Amount of necessary information on actual
and operating supplies, and trade goods to be and potential suppliers and customers
bought
- Location
- Number or amount of finished products, semi-
- Amount of products in supply and demand
finished products (spare parts) and trade goods
- Delivery and order cycle
- Number of sources of supply
- Places of supply and support
- Number of customers to be supplied
- Transportation and packaging requirements
- 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 Dynamics of the environmental relations

Rate and regularity of change: Rate and regularity of change in the necessary
- The time of delivery and the amount in supply and information on:
demand delivered - Location
- Channels of procurement and distribution - Amounts of products in supply and [sic!]
demand - Delivery and order cycle
- The structure of the suppliers and customers - Places of supply and support
- Transportation and packaging requirements

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

PFOHL/ZLLNER (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 diversifi-
cation and consequently to totally different channels of distribution. The
measure for the diversity is the proposed homogeneity, which according to
PFOHL/ZLLNER (1997) shows to which extent logistical capacities such
as order processing, packaging, transportation or handling can be com-
bined 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 markets needs.
78 3 Theoretical framework

A further contingency factor is the production technology of the organi-

zation (PFOHL/ZLLNER 1997, p. 309). They argue that its influence on lo-
gistics tasks can be determined by the spatial arrangement of machines
which can vary in different states between job-shop and flow shop produc-
tion. 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/ZLLNER (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/ZLLNER (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 de-
veloped 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/ZLLNER (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 pat-
terns 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 increas-
ing 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 LSPs 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 fac-
tors. According to KLEER (1991, pp. 122-123) these include the products
and its features relevant for logistics, the size of the LSPs 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 logis-
tics outsourcing relationships. Furthermore, it must be examined which
explanatory value the combination of these theories has to offer with re-
spect to the research aims of this study. This chapter will address these
questions in the following.
Some studies analyzing inter-organizational relationships, both empiri-
cally and conceptually, have weaknesses in their theoretical foundation be-
cause 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 con-
ducted 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 be-
tween logistics service providers and their customers. Since a number of
different relationship designs exist, the focus here will be on long-term re-
lationships and their cooperative aspects as present in contract logistics ar-
rangements (see chapter 2.3.3.2).
Especially suited to analyze long-term logistics outsourcing relation-
ships is the commitment-trust theory. It explicitly centers on the analysis
of long-term relationships between firms and allows insights into the fac-
tors which determine outcome and performance of these arrangements.
80 3 Theoretical framework

Transaction cost theory also contributes to the understanding of the mat-

ter. While its main focus is the analysis of the boundaries of the firm and
the decision of the organization between market, hierarchy, and coopera-
tion, it also permits the study of different forms of organization across the
supply chain. This naturally includes long-term relationships in the logis-
tics outsourcing context.
A less obvious theory for the research on logistics outsourcing relation-
ships 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 per-
sonal sympathy. It rather displays that additional to purely economic con-
siderations, 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 quasi-
mechanistic 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 conclu-
sions. The use of the contingency approach nevertheless offers valuable in-
sights into the situational factors of logistics outsourcing relationships and
their performance effects. It therefore promises to have value for the an-
swering 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 struc-
ture for every context, the use of the contingency approach seems never-
theless 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 vari-
ables 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 de-
sign of logistics outsourcing relationships. Through this procedure, the
 3.3 Theory integration 81

strengths of the contingency approach will be used while its main weak-
nesses have no effect in this context.
The general suitability of the theories presented above for analyzing lo-
gistics outsourcing relationships does not necessarily constitute the im-
perative to actually use them all in this study. The necessity to do so rather
results from the complexity inherent in the relationships between custom-
ers and LSPs. To address this complexity, the theories have been explicitly
chosen collectively to develop a model of logistics outsourcing relation-
ships that aims at identifying from different perspectives the relevant fac-
tors influencing the performance of outsourcing relationships after the out-
sourcing decision has been made.
An important theory to explain the formation of cooperative relation-
ships 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 opportun-
ism.
However, transaction cost theory on its own cannot be sufficient to an-
swer 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 relation-
ships 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 commitment-
trust theory which is of significant importance in this study because it ex-
plicitly 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 val-
ues, 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, compris-
ing 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 lat-
ter ignores it. Just for this reason, marketing research generally supports
this procedure (LAMBE/WITTMANN/SPEKMANN 2001), because both as-
sumptions conflict with the reality which sees opportunism between the
two extremes.
It can therefore be summarized that transaction cost-, social exchange-
and commitment trust theories allow insights into how logistics outsourc-
ing relationships ought to be designed. However, they alone cannot answer
which situational factors affect the relationship variables. Some basic fac-
tors 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 there-
fore can be understood as contingency variables because their value ac-
cording to the theory directly influences the level of the transaction costs
and therefore determines the level and the characteristics of the coopera-
tion or relationship.
Further situational variables are supplied by the contingency approach.
It allows the identification of both internal and external factors that influ-
ence the performance of logistics outsourcing relationships. Some of these
factors include external factors such as environmental uncertainty, com-
posed of environmental complexity and environmental dynamics, the im-
portance 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 lo-
gistics outsourcing relationships and has indicated that a number of vari-
ables affect their performance outcomes. The following chapters will ex-
plore 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 out-
sourcing performance on logistics performance and firm performance will
be presented after the relevant variables will have been established. Fi-
nally, 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 per-
formance. At first, chapter 4.1.1 will present the background of the concept
of logistics outsourcing performance, its connection to logistics perform-
ance 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 under-
standing 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 perform-
ance in many different ways (CHOW/HEAVER/HENRIKSSON 1994) and
thereby provided a valuable starting point for the analysis of logistics out-
sourcing performance.
As CHOW/HEAVER/HENRIKSSON (1995, p. 296) point out, logistics per-
formance is multi-dimensional, reflecting multiple stakeholders and inter-
ests. 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/GOLDS-
BY/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/HENRIKS-
SON (1995, p. 296) in research is predominantly measured with soft per-
ceptual indicators given the difficulty of obtaining hard performance
measures, is a result of two different variables: on the one hand, it is influ-
enced by the performance of logistics processes performed in-house under
the direct responsibility of the LSPs 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 dele-
gated logistics and other relevant processes and the accompanying respon-
sibility to a logistics service provider.
The performance of these outsourced processes, hereafter termed logis-
tics outsourcing performance, is an important strategic issue which has re-
ceived little attention so far in logistics research. Notable exceptions in-
clude 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 perform-
ance. Building on the work of GUTENBERG (1957), who made this produc-
tivity oriented view popular in Germany, WEBER (2003) proposes to struc-
ture 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 ser-

vices:
The input reflects the consumption of materials and services used in lo-
gistics processes. According to WEBER (2003, p. 12), it is often implicitly
assumed that lower levels of inputs indicate higher efficiency, when for in-
stance 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 individ-
ual 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 lo-
cation at different points in time (WEBER 2002, p. 122), the outcome re-
flects 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 meas-
urement 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 effective-
ness.
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 ex-
pected with respect to quality and agreed upon price. Logistics outsourcing
performance in this work therefore will be measured in the light of the out-
come-orientation as suggested by WEBER (2003, p. 12) and will include
measures of both efficiency and effectiveness. The more detailed concep-
tualization 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 tri-
dimensional 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 tri-

dimensional and consisting of the factors operations performance, channel
performance and asset reduction performance. Finally, ENGELBRECHT
(2004, p. 212) suggests to measure outsourcing performance bi-
dimensionally through the of facets goal achievement and information ex-
change. As all three works are very relevant for the conceptualization of
logistics outsourcing performance, their individual findings will be pre-
sented in detail in the following.
Realizing that outsourcing performance depends on several antecedents,
STANK/GOLDSBY/VICKERY/SAVITSKIE (2003, pp. 32-33) propose a posi-
tive 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 tri-
dimensional 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 (opera-
tional 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 posi-
tive influence of relational performance on both operational- and cost per-
formance. 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 re-
search 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 differ-
ent relationship factors1 and outsourcing performance, the dependent vari-
able of outsourcing performance is a three-dimensional construct, consist-
ing 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. 212-
218) uses a similar approach. Realizing that for the successful implementa-
tion of outsourcing arrangements several variables play a central role,
among them the involvement of the parties, the level of conflict, the proac-
tive improvement of the LSP, the customers prior outsourcing experience
and the length of the project implementation phase (ENGELBRECHT 2004,
pp. 202-212), he suggests addressing the complexity of outsourcing per-
formance by utilizing a two-dimensional outsourcing performance con-
struct. ENGELBRECHT (2004, p. 212) states that the performance of out-
sourcing 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 co-
operation 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 ac-
knowledge 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, pref-
erably 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 con-
tract 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 cus-
tomer. This cannot be measured through the construct goal achievement
which due to its nature can only grasp if all deficits that would have pre-
vented the achievement of the previously agreed upon goals have been ad-
dressed 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 re-
ductions, 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 neces-
sary to fulfill the exact requirements and expectations, exceeding the goals
needs very different efforts. Here, the LSP must strive to surpass the ex-
pectations 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 de-
terminants 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 ante-
cedents 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 real-
ized. 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 in-
troduced. 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 work-
ing relationship between the parties, they find it easier on the one hand to
reach the goals previously agreed upon, while on the other hand, coopera-
tion 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 impor-
tant facilitator of successful logistics outsourcing arrangements as it leads
to closer cooperation and fosters the understanding between the two par-
ties. 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, out-
standing performance effects can be achieved only through the constant
improvement of the logistics processes through the LSP in order to in-
crease 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 per-
formance 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 Direct Indirect
Social Exchange Theory, Commitment-Trust Theory,
Cooperation Transaction Cost Theory 9
Communication Social Exchange Theory, Commitment-Trust Theory 9
Proactive Improvement Social Exchange Theory 9
Trust Social Exchange Theory, Commitment-Trust Theory 9
Relationship Commitment Social Exchange Theory, Commitment-Trust Theory 9
Functional Conflict Commitment-Trust Theory 9
Involvement Transaction Cost Theory 9
Opportunism Transaction Cost Theory, Commitment-Trust Theory 9
Shared Values Commitment-Trust Theory 9
Openness Transaction Cost Theory 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 commitment-
trust 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 relation-
ships. In addition to its indirect effects via other variables in the model, di-
rect effects on outsourcing performance will be assumed as the functional
solving of conflicts directly enables the achievement of goals and their ex-
ceedance 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 success-
ful relationships with indirect effects on performance. The involvement
will e.g. lead to closer cooperation between the parties and foster the pro-
active improvement of the LSP, as potential problems and optimization po-
tentials are identified earlier. This equally holds true for the factor open-
ness, whose transaction cost reducing character is beneficial for logistics
outsourcing arrangements. While openness alone does not have direct ef-
fects on outsourcing performance, its positive influence on the communi-
cation 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 co-
operative and mutually successful relationships.
The following chapters will in detail conceptualize the variables that
have direct and indirect effects on logistics outsourcing performance. Sub-
sequently, hypotheses on the causal linkages between them will be pre-
sented 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 customer-

supplier relationships (ANDERSON/NARUS 1990; METCALF/FREAR/
KRISHNAN 1992; MORGAN/HUNT 1994; WILSON 1995; LAMBE/WITT-
MANN/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 coordi-
nated 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 con-
structs (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 ex-
pand it by emphasizing cooperations proactive nature as opposed to the
power-related view which involves being coerced into taking interdepend-
ent actions (WILSON 1995, p. 338). Cooperation furthermore is character-
ized 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 re-
search (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 be-
haviors 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 em-

pirical 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 for-
mal as well as informal sharing of meaningful and timely information be-
tween firms. This definition focuses on the efficacy of information ex-
change 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 in-
cludes 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, in-
tense, 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) be-
lieve that a complete and open exchange of operating and strategic infor-
mation holds logistics alliances together, because joint performance to-
wards 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 synchro-
nized.
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 ser-

vice 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 or-
der to increase value creation for their customer as a basis for the own sus-
tained 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 re-
search. 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 in-
novation 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 ma-
jor 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 inten-
sity of the activities employed by a logistics service provider aiming at im-
proving 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 psychol-

ogy (DEUTSCH 1960; LINDSKOLD 1978; LEWICKI/BUNKER 1995), sociol-
ogy (STRUB/PRIEST 1976; LEWIS/WEIGERT 1985), economics (WILLIAM-
SON 1981; DASGUPTA 1988), and consequently also in the marketing
related research on buyer-seller relationships (DWYER/SCHURR/OH 1987;
ANDERSON/WEITZ 1989; MOORMAN/DESHPAND/ZALTMAN 1993;
GANESAN 1994; MORGAN/HUNT 1994; DONEY/CANNON 1997; LANE
2000).
Trust has diffused from psychology and sociology into business-to-
business 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 in-
tegrity. Similarly, MOORMAN/ZALTMAN/DESHPAND (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 partys word or
promise is reliable and that a party will fulfill his/her obligations in an ex-
change relationship.
While the list of different definitions of trust could be continued virtu-
ally 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 eco-
nomic theory, trusting a partner means potentially being exposed to oppor-
tunistic behavior, since the response to any action in a social exchange re-
lationship 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 ef-
fects on the trust and will ultimately result in a termination of the relation-
ship.
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 soci-
ologists 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). Cogni-
tions, 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) further-
more 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 ob-
served: trust requires both the will of at least one party to trust the other
and additionally depends on the other partys trustworthy behavior
(WALLENBURG 2004, p. 107). Since trustworthy behavior is a complex is-
sue largely depending on the particular context, trust in this study will be
conceptualized along the line of the buying partys will to trust. This will-
ingness to trust corresponds to the above definition of
MOORMAN/ZALTMAN/DESHPAND (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 dimen-
sions: The first dimension of trust focuses on the objective credibility of an
exchange partner. It includes the expectation that the partners word or
written statement can be relied on (LINDSKOLD 1978). The second dimen-
sion, benevolence, marks the extent to which one partner is genuinely in-
terested in the other partners 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 de-
gree of risk concerning the trust that the LSP will be able to perform effec-
tively and efficiently (credibility) and at the same time is interested in the
customers best interest (benevolence). The construct trust will in this
study consequently measure the degree of trust the customer has in the or-
ganization 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 em-
ployed among others by DWYER/SCHURR/OH (1987), ANDERSON/WEITZ
(1992), MOORMAN/ZALTMAN/DESHPAND (1992), MORGAN/HUNT
96 4 Antecedents and effects of logistics outsourcing performance

(1994), GUNDLACH/ACHROL/MENTZER (1995), WILSON (1995), HOCUTT

(1998), MOORE (1998), RODRGUEZ/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 relation-
ships (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/DESHPAND (1992, p. 316) define com-
mitment to the relationship [] as an enduring desire to maintain a val-
ued 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 or-
ganizations (MEYER/ALLEN 1984) and define relationship commitment
as an exchange partner believing that an ongoing relationship with an-
other 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 inter-
organizational 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 ac-
cording to MORGAN/HUNT (1994, p. 23) is one type of relationship com-
mitment that is critical to a firms internal relationships, is among the old-
est (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 real-
ize 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 develop-

ment 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. Simi-
larly, 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 cus-
tomers (MORGAN/HUNT 1994, p. 23).
Overall, a common theme emerges from the various literatures on rela-
tionships (MORGAN/HUNT 1994, p. 23): parties involved in interorganiza-
tional exchange relationships identify commitment among exchange part-
ners 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 be-
tween 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 be-
tween customers and LSPs examined in this study.
Before concluding it must be remarked that recently the one-
dimensional view of commitment has been criticized. Especially in the so-
cial 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 rela-
tionships 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) cer-
tainly 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 commit-
ment 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 dis-

agreements: DWYER/SCHURR/OH (1987, p. 24) state that conflict is pre-
dictable within the relationship just as periods of resource scarcity, mis-
perceptions 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 isolation-
ism. While some authors suggest that conflict can have benefits for a rela-
tionship, such as more frequent and effective communication, a critical re-
view 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 dissolu-
tion.
However, when disputes and conflicts are resolved amicably, such dis-
agreement 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 vi-
tality (DWYER/SCHURR/OH 1987, p. 24), for partner firms that use con-
flicts 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 arrange-

ments.

4.2.1.7 Involvement

According to ENGELBRECHT (2004, pp. 203-204) logistics outsourcing ar-

rangements have strong strategic implications for the customer since they
directly effect the boundaries of the firm and therefore significantly influ-
ence its core competencies and resources. Due to the complexity of logis-
tics outsourcing and its impact on different business processes of the cus-
tomer, 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 ar-
rangements 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 ar-
gues 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 com-
munication and cooperation.
The construct involvement so far has received very little attention in lo-
gistics outsourcing research. In this study it will, in-line with the argumen-
tation presented above and the research conducted by ENGELBRECHT
(2004), measure to what extent the LSP was involved in the implementa-
tion process of the logistics outsourcing and how well the collaboration be-
tween the employees of both the customer and the LSP has been function-
ing.

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 litera-
ture 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 ones
100 4 Antecedents and effects of logistics outsourcing performance

appropriate or required role behavior (JOHN 1984, p. 279). Since oppor-

tunistic behavior in the organizational literature is assumed in the funda-
mental 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 there-
fore 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 behav-
iors, goals, and policies are important or unimportant, appropriate or in-
appropriate, and right or wrong. They are important for exchange rela-
tionships 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 relation-
ship 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 underly-
ing 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 ex-
change 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 organiza-
tional 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 employees ability to speak up
and express his views, which is in turn perceived as positive since employ-
ees have increasingly become recognized as an invaluable source of ideas
for improving a firms performance (HARRINGTON 2001, p. 85-88).
GUPTA (1987) analyzes the effect of openness between corporate execu-
tives 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 busi-
ness units] managers and their corporate superiors are open and infor-
mal 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 ef-
fects 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 ac-
knowledged 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 or-
der to now postulate a consistent model of logistics outsourcing perform-
ance, 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 link-
ages among each other will be hypothesized, direct causal linkages on the
outsourcing performance will only be theorized for the constructs of coop-
eration, 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 in-
terpreting 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 co-
operation holds a key role in interorganizational relationships and contrib-
utes 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 outsourc-
ing 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 coopera-
tion has a positive effect on logistics outsourcing performance. As both
parties cooperate by working closer together and by pulling in the same di-
rection, 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, coopera-
tion 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 mu-
tual cooperation are more rewarding than those that are not. As relation-
ships 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 perform-
ance.
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 sup-
posed 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 over-
achievement 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 vari-
ous transaction costs: monitoring costs are reduced as monitoring and con-
trol become less important. Furthermore, if closer cooperation leads to ex-
ceeding 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 rela-
tionships are more rewarding than adversarial relationships as benefits be-
come 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 enter-
ing the outsourcing arrangement, thus leading to goal exceedance.
On this basis, it can therefore be assumed that the more intense the co-
operation 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 proac-
tive 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 hypothe-
sized that as firms engage in cooperative behavior and understand its bene-
ficial character, they will also aim at increasing the functionality of con-
flict 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 dis-
crete, 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 ef-
forts as a consequence of the intensified customer satisfaction under the
social exchange theory assumption of reciprocity. Therefore, it can be ar-
gued 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 ex-
change relationships.
According to MOORE (1998, p. 28), information exchange and commu-
nication 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 be-
tween different organizational levels of the customer and between the cus-
tomer and the LSP. BOWERSOX/DAUGHERTY/DROGE/ROGERS/WARDLOW
(1989, p. 225) even believe that the complete and open exchange of oper-
ating and strategic information is the glue that holds [logistics] alliances
together.
The qualitative aspect of communication is often emphasized. Appar-
ently, 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 necessar-
ily 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 su-
perior importance of interaction and communication for successful rela-
tionships. As parties strive to increase relationship benefits through coop-
eration, 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 un-
derstands 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 in-
creased levels of communication, however, may also lead to a continuous
adaptation of the goals by the customer, which could potentially cause ris-
ing 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 perform-
ance. 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 improve-
ment. 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. Fol-
lowing this argumentation, no hypotheses on a direct effect of communica-
tion on goal achievement and goal exceedance will be formulated.
However, communication has effects on antecedents of outsourcing per-
formance. Again, the causal linkages of other variables on communication
will be presented in the respective chapters, while in the following the ef-
fect 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 sug-
gests that relevant, timely, and reliable communication fosters trust by re-
solving 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 subse-
quent 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 relation-
ship. 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 bene-
fits normally only attainable through vertical integration. This has been
shown to especially require the sharing of information and finds the em-
pirical support of METCALF/FREAR/KRISHNAN (1992, p. 39). It can there-
fore be hypothesized that a positive relationship exists between communi-
cation and cooperation.
The importance of the LSPs proactive improvement has been argued in
chapter 4.2.1.3. Following the social exchange literature, the LSP will in-
crease 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 customers
situation and its issues. Therefore, the LSPs effort will be greatly sup-
ported through relevant, timely, and reliable communication of the cus-
tomer. 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 op-
portunism shall be examined. While social exchange theory has been criti-
cized 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 be-

havior 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 suc-
cess 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 pro-
active improvement on goal achievement. As cooperative behaviors be-
tween 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 ac-
tivities 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 out-
sourcing projects. This insight is backed by findings of STANK/GOLDSBY/
VICKERY/SAVITSKIE (2003, p. 43) who demonstrate that the relational per-
formance between customer and LSP positively influences both opera-
tional 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 sig-
nificantly exceeded in terms of quality of service and costs. Since the pro-
active improvement through the LSP will lead to continuous improvement
of the service quality and/or cost reductions, it will be postulated that pro-
active 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 influ-
enced by proactive improvement because the efforts of the LSP going be-
yond 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 lat-
ter, 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 interorgani-
zational 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 commit-
ment-trust theory, trust is a key mediating variable within effective rela-
tional 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 posi-
tively related to the buyers 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 theo-
ries 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 re-
lationship 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 sup-
ported 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 prisoners 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 em-
pirical 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 commu-
nication will result in increased functionality of conflict as a result of in-
creasing 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, coopera-
tion, and functional conflict is not. However, to remedy this shortcoming it
will be incorporated in the logistics outsourcing performance model as al-
ready argued in chapter 4.3.1.1.
On the basis of the argumentations presented above, the formalized hy-
potheses 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 com-

mon dependent variable used in buyer-seller relationship studies. As
MORGAN/HUNT (1994, p. 23) argue, parties involved in exchange relation-
ships identify relationship commitment as key for achieving desired out-
comes 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 rela-
tionship 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 cus-
tomer, 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 relation-
ships develop and firms shift from discrete transactions to long-term rela-
tionships, the exchange partners will develop commitment to the relation-
ship as they begin anticipating rewards that would not be attainable
through discrete transactions. As the customers commitment to the rela-
tionship increases, so does its understanding for the necessity of coopera-
tive 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 in-
creases, so do its efforts for cooperative behavior, thus leading to increas-
ing 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 out-
sourcing 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 amica-

bly, they can be turned into functional conflict, which has beneficial ef-
fects on the relationship. The role of functional conflict in logistics out-
sourcing 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 ex-
changes 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 re-
ducing the positive effects high levels of cooperation have on the accom-
plishment of goals in general (STERN/REVE 1980) and on logistics out-
sourcing 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 behav-
ior 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 ex-
ceedance. 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 ordi-
nary conflict level of average relationships without functional conflict
would never be.
While the relationship between functional conflict and outsourcing per-
formance has never been analyzed empirically, the research conducted by
SKINNER/GASSENHEIMER/KELLEY (1992, p. 185) supports the argumenta-
tion. Empirically, they find that the level of conflict has a negative impact
on both the level of cooperation displayed in the relationship and the cus-
tomers satisfaction with the relationship. As now functional conflict de-
creases the level of conflict in the relationship, the level of cooperation as
well as the customers satisfaction should rise. Hence, functional con-
flict 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 atten-
tion 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 imple-
mentation process leads to improved communication and proactive im-
provement 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 em-
ployees are involved in the outsourcing process, work together, and ad-
dress 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 opportuni-
ties of identifying optimization potentials in the customers processes in-
crease due to early and thorough involvement.
ENGELBRECHT (2004, pp. 272-277), as mentioned above, finds empiri-
cal support for the hypotheses that the degree of involvement of the LSP
positively influences communication and proactive improvement. The hy-
pothesis 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 communica-

tion.

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 dan-
ger that, if partners cannot guard themselves against exploitation, valu-
able 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 relation-
ships (KNEMEYER/MURPHY 2004, p. 41).
Opportunism is one of the two basic behavioral assumptions of transac-
tion cost theory. As chapter 3.2.1.1 has presented, it is treated as a funda-
mental 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 be-
yond the rather static and one-dimensional axiomatic approach of transac-
tion cost theory.
After DWYER/SCHURR/OH (1987, pp. 22-23) proposed that the incorpo-
ration of trust into relationship models would provide a suitable vantage
point, the commitment-trust theory takes up the argumentation. Conse-
quently, 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 relation-
ships between customers and their LSPs.
While opportunism will not be understood as a basic behavioral assump-
tion 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 ex-
change theory also promises insights. While this theory normally does not
include opportunism, it argues that cooperation is important for function-
ing long-term relationships. As opportunistic behavior by a partner is sus-
pected by another, this will have the argued detrimental effects on the co-
operation and the process of the two parties working together will be
severely impeded. Therefore, it can be presumed that higher levels of op-
portunism 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 im-
portant 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 commit-
ment. MORGAN/HUNT (1994, p. 25) state that the organizational commit-
ment literature distinguishes between two kinds of commitment: one the
one hand that brought about by a person sharing, identifying with, or in-
ternalizing 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 be-
tween 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 com-
mitment-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 hy-
pothesized 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 to-
wards opportunistic behavior. While MORGAN/HUNT (1994, p. 24-30) un-
derstand opportunism as a precursor and therefore as an independent vari-
able 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 unim-
portant, appropriate or inappropriate, and right or wrong (MORGAN/
HUNT 1994, p. 25), they ought to find it easier to share formal and infor-
mal 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 infor-
mally 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 relation-
ships 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 logis-
tics domain. However, social exchange theory suggests that openness
could be an important antecedent for cooperative and sustainable relation-
ships characterized by commitment and trust which both should be facili-
tated 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 suc-
cessful relations between customer and LSP and its role and causal link-
ages 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 increas-
ingly 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 charac-
terized 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 con-
structs 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 in-
formation and ideas should facilitate the reduction of the main reason for
mistrust which is to be found in the uncertainty of possible future oppor-
tunistic behavior of the partner (WALLENBURG 2004, p. 219). Therefore,
openness should not only have an indirect effect on trust via communica-
tion, but also a direct influence.
The same is true for the positive effect of openness on functional con-
flict. As much as communication has been shown in chapter 4.2.1.2 to re-
duce 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 posi-
tive effect on functional conflict.
 4.3 Formulation of a model of logistics outsourcing performance 119

The hypotheses derived from the above argumentation read in formal-

ized 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 di-
mensions 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 mod-
eled 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 121

Shared Values H22 Cooperation

H21 H13 H1a
Commitment
Goal
H11 H15 Achievement
H20
H10 H3
Involvement
H25
Trust H17 H14a H9a
H24 H2
H27
Proactive
H4
H12 Improvement
Openness
H16 H1b

H19 H9b
H23 H28 H7 H6
H18 H26 Goal
H14b Exceedance
Functional
Conflict
H5
Opportunism H8 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 lo-
gistics outsourcing performance. The question of these performance ef-
fects, however, is only a strategic one if a connection can be shown be-
tween 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 in-
crease the focus on the relationships with their LSPs would have to be re-
jected 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 perform-
ance, with a study by ENGELBRECHT (2004) supporting these results, find-
ing also that logistics outsourcing performance positively influences logis-
tics performance, another empirical testing will be conducted in this study
to validate the results. This is justified both because of the strategic rele-
vance 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. concern-
ing causal linkages that were found to be non-significant in either one of
the cited works or the connection between the two-dimensional outsourc-
ing 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 di-
rect influence of logistics outsourcing performance on logistics perform-
ance 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 in-
vestigating 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 lo-
gistics 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 de-
fined 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 per-
formance 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 de-
sired outcomes from logistics are numerous and range from customer satis-
faction over issues such as environmental responsibility to overall cost-
effectiveness. Depending on the given situation, the measurement of logis-
tics performance therefore must be adapted.
As it has been argued above, one goal of this study is to identify the ef-
fect 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 out-
sourcing 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 out-
sourcing 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 con-
structs and develop hypotheses on the causal linkages between logistics
outsourcing and logistics performance. The operationalization of the con-
structs will occur in chapter 6.3 in line with the structure of the other chap-
ters 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 firms 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 ser-
vices 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 com-
mon understanding exists. Which costs are subsumed under the term logis-
tics 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 firms 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 empiri-

cal 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 logis-
tics costs that are perceived by at least a third of the firms as part of the lo-
gistics 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 un-
derstanding of logistics costs among the firms. For reasons of understand-
ability 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 ef-
fective 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 firms logistics services. As it
has been argued above, it represents a firms capability to supply the cus-
tomers 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 con-
nected 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 ser-
vices proves more demanding. Consequently, several different measure-
ment 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 infor-
mation 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 perform-
ance 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 flexi-
bility. The first factor stands for the ability to convert customer orders into
product output suitable for the market, while flexibility measures the logis-
tics processes capability to react to present and future changes in custom-
ers demands.
After the first two levels of logistics services are fulfilled, firms must
ensure the permanent quality of their logistics processes. This aims at in-
creasing 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, indica-
tors from all three levels will be utilized for the measurement of a firms
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 re-
sponsibility 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 influ-
ence on the firms 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 bi-
dimensional construct consisting of goal achievement and goal exceedance
and views information exchange or communication as one of its antece-
dents.
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 ac-
complished through logistics outsourcing, thereby increasing goal ex-
ceedance, it can be expected that a positive influence on the overall logis-
tics 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 pre-
sented above is specified in Figure 4-3.

Goal H29 Level of

Achievement Logistics Services
H31
H30

Goal H32 Level of

Exceedance 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 it-
self 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 finan-
cial objectives (ENGELBRECHT 2004, p. 225). Logistics outsourcing re-
search therefore draws its relevance from the implicit assumption that lo-
gistics performance has a measurable and relevant influence on firm
performance. As proposed in research question three in chapter 2.4.2, em-
pirical 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 dif-
ferent approaches suggested (BHARGAVA/DUBELAAR/RAMASWAMI 1994).
Fundamentally, it can be distinguished between hard and soft per-
formance 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, pro-
duction 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 or-
ganizational performance, they will be used for the measurement of firm
performance in this study. The reasons for the selection of subjective per-
formance 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. Dif-
ferent levels of transfer prices for instance lower the general comparability
of performance indicators just as much as managements use of its discre-
tion to influence performance measures and balancing procedures prior to
supplying the data. Furthermore, asking for hard and objective perform-
ance measures in survey research carries the danger that informants will
not at all or not truthfully respond to the questions fearing lacking ano-
nymity. These difficulties, coupled with the findings of several studies
showing very high correlations between subjective and objective perform-
ance 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 measure-
ment of different components of firm performance that addresses the plu-
rality 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 adap-
tiveness, 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) demon-
strated the suitability of the measures also in a logistics context. The fol-
lowing chapter will conceptualize the three dimensions before in chapter
4.4.2.2 hypotheses on the effects of logistics performance on firm per-
formance will be introduced.

4.4.2.1 Conceptualization

Firm performance, as understood in this study, consists of the three dimen-

sions adaptiveness, market performance, and financial performance. Be-
fore they are conceptualized it must be noted that they are not on the same
logical level but must rather be understood as interdependent and sequen-
tial 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 im-
portant 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 respon-
siveness, reflects the ability of the firm to adapt to changes in the environ-
ment (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 mar-
ket 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 there-
fore 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 be-
yond 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 economi-
cal success of a firm. Following the argumentation of ENGELBRECHT
(2004, p. 229), financial performance is understood as the ultimate indica-
tor of firm performance and will be measured through the level of the
firms revenue margin compared to that of competing firms as will be ar-
gued 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 per-
formance. It will theorize that a strong indication exists suggesting that lo-
gistics 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*** 0.31***

Level of Logistics n.s. 0.47*** Level of Logistics n.s. 0.13*

Services Services
2
R : 66.2% R2: 53%
0.47*** 0.68***
Market Market
Performance Performance
n.s. n.s.

Level of Logistics n.s. 0.50*** Level of Logistics n.s. 0.41***

Costs Costs

0.24*** R2: 40.6% 0.16* R2: 23%

Financial Financial
Performance 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 per-
formance as well as for the effect of the level of logistics services on the fi-
nancial performance.
As it has been argued above, the relevance of the causal linkages be-
tween logistics performance and firm performance for answering the re-
search questions introduced in chapter 2.4.2 and the resulting research
model justifies another empirical testing of the relationship. This is espe-
cially relevant in the light of the aspired testing of the moderating effects
of situational factors. Therefore, the following paragraphs will briefly in-
troduce 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 neces-
sary and developed for increasing the level of logistics services will be ad-
vantageous 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 adap-
tiveness of the firm, but also on its market performance. Market perform-
ance depends very strongly on the satisfaction of customers with the pur-
chased 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 cus-
tomer service or logistics. Several empirical studies underline the impor-
tance 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 in-
creased customer loyalty and serve as a facilitator to win new customers
which increases the market performance. This, as it will be argued in hy-
pothesis 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 poten-
tially 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 lev-
els 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 adaptive-

ness.

H34: The level of logistics services has a positive effect on market per-
formance.

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 perform-
ance.
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 in-
crease at the same amount at which the logistics costs are lowered.
Following the above argumentation, three hypotheses can be formu-
lated:

H36: The level of logistics costs has a positive effect on adaptiveness.

H37: The level of logistics costs has a positive effect on market per-
formance.
 4.4 Effects of logistics outsourcing performance 133

H38: The level of logistics costs has a positive effect on financial per-
formance.
After having investigated the causal linkages between logistics perform-
ance 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 interdepend-
ent 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 bet-
ter address the needs of their customers, thereby increasing customer satis-
faction 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 de-
creased 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 influ-
ence 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 perform-

ance.
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
H36 H39
Logistics Services

H34
Market
Performance
H37

Level of
H40
Logistics Costs
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 gen-
eral insights into the nature of logistics outsourcing performance on the
one hand and into the relationship between outsourcing performance, lo-
gistics performance, and firm performance on the other. While the ex-
planatory value of these models can be expected to be very high for the de-
signing 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 moderat-
ing 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 es-
tablished. 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 dif-
ferences that may influence the models of logistics outsourcing perform-
ance 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 prac-
tices. LSPs react to the heterogeneity of their customers by providing a di-
versified service portfolio whose orientation varies according to the as-
pired target markets. As will be briefly shown in the following, contract
logistics providers as introduced in chapter 2.2.3.2 show distinct organiza-
tional 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 lim-
ited number of industry verticals. By doing so, they have developed indus-
try-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 influ-
ence 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 indus-
tries and smaller, niche providers having chosen to accommodate indus-
tries only with very specific demands.
LIEB (2005, p. 23) also suggests that the differing sizes of their custom-
ers have prompted LSPs to adapt their organization structure. As large cus-
tomers with more standardized logistics processes promise higher margins
and yields for logistics service providers, the LSPs have become increas-
ingly customer selective. This is leading to a situation in which the most
promising customers resort to the largest LSPs as they offer the most suit-
able outsourcing arrangements, and many small and medium sized cus-
tomers 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 par-
ticular 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 custom-
ers 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 dif-
ferent 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 outsourc-
ing 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) ar-
gues 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 require-
ments 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 fac-
tor in long-term relationship with changing and demanding logistical re-
quirements. 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: Lo-
gistics 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 strategi-
cally 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 firms performance and vice versa. Consequently, logistics per-
formance 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. com-
ponent logistics in the automotive industry. This clearly indicates the exis-
tence of differences among the firms and hence suggests potential moder-
ating effects.
While the above argumentation suggests the potential existence of mod-
erating effects also in the firm performance model, it must be acknowl-
edged that the variables exerting influences on this model may substan-
tially 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 ex-
ploratively test the relevance of moderating effects also for the firm per-
formance 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 cus-
tomer 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 vari-
ables that have been in the focus of a far larger amount of studies, signifi-
cant effects of a large number of variables were found.
GIERING (2000, pp. 100-153) argues that the link between customer sat-
isfaction and customer loyalty is moderated by variables from the follow-
ing 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 see-
king, 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 ex-
change relationship models. Other studies show mixed results: DEHLER
(2001, pp. 244-252) finds that the causal linkages between logistics per-
formance 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 re-
lationship 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/ZLLNER (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 en-
vironmental relations for both the flow of products and the flow of infor-
mation. A detailed description was given in chapter 3.2.4.2. KLEER (1991,
pp. 121-124), as outlined in the same chapter, conceptually identifies simi-
lar contingency variables but utilizes the segmentation approach by
KIESER/KUBICEK (1983, p. 222) to distinguish between internal and exter-
nal 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 rela-
tionships and are conceptualized broad enough for a later operationaliza-
tion 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 preced-
ing 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 out-
sourcing-, 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 ex-
plicitly developed to cover the situational factors important for relation-
ships 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 deci-
sion and can therefore choose a partner that fits its needs sufficiently well.
Therefore, contingency variables of the customer are of primary impor-
tance. The analyses of the moderating effects of contingency variables in-
fluencing 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 ar-
gued in chapter 3.2.4.2, external contingency variables are those suited to
explain differences between organizational structures that cannot be al-
tered by the organization alone, but depend on other organizations and
thereby describe the relationship of the organization to its environment. In-
ternal 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 dy-
namics. 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 LSPs
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 de-
pending 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 logis-
tics. 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 dy-
namics 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 num-
bers of own customers with steadily increasing sales volumes and as a re-
sult a shifted balance of power.
While KLEER (1991, p. 122) views the customers 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 in-
dustry 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 customers control. Close to the variables en-
vironmental 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 customers or its customers industries. Hence, in this

study uncertainty will be understood as related to the concepts of complex-
ity 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 differ-
ing theoretical backgrounds, namely transaction cost theory and the con-
tingency 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 inter-
nal situation of the customer and which it can theoretically change inde-
pendently. They include the six variables derived from the contingency
approach: the product range, the firm size, the degree of logistics centrali-
zation as well as the asset specificity, the frequency and the processes ori-
entation.
The products constitute a central aspect of the customers internal situa-
tion. 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 opportu-
nity of the customer to positively differentiate itself through logistics ex-
cellence.
A second internal contingency variable is the size of the organization.
With increasing size, the LSPs 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 logis-
tics 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 in-
volved, but should also lead to a higher relationship intensity, while a low
degree of centralization should tend to impede or hinder relationship for-
mation.
As argued in chapter 3.2.1.2, also the asset specificity and the frequency
of the customers 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 termina-
tion of the contract. Knowing about high levels of asset specificity should
therefore alter the customers 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 out-
sourcing 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 ef-
fects on outsourcing performance, logistics performance, and firm per-
formance will be analyzed. Process orientation implies the largely failure-
free 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 achieve-
ment 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 af-
fect the linkage between relationship variables and the outsourcing per-
formance 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 inter-
nal contingency variables that are supposed to have an influence on the ef-
fect relationship variables have on the logistics outsourcing performance
and furthermore on the relation between outsourcing performance, logis-
tics performance, and firm performance. Figure 4-6 gives an overview of
the hypothesized contingency variables, which will in detail be operation-
alized in chapter 6.6.
As argued above, no hypotheses will be developed for the moderating
effects these contingency variables potentially exert on the different per-
formance models. Instead, their effects will be subject to explorative test-
ing. 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

Customers 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 con-
ducted. This addresses the demand of MENTZER/KAHN (1995, p. 244) dis-
cussed in chapter 2.4.3 to apply the adequate scientific rigor also in logis-
tics 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 justi-
fication 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 devel-
opment 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 follow-
ing, between chapters 5.2 and 5.2.5, the methodological basis for the em-
pirical 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 indus-
tries 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 appro-
priate 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 con-
structs (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 re-
flect the characteristics of latent variables. The measurement of these indi-
cators generally is subject to measurement errors which can lead to sub-
stantially 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 re-
lationship 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 prin-
ciple of regression analysis are potentially suitable to analyze causal rela-
tionships. 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 re-
quires 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 assump-

tion 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 util-
ized in this study.
The present standard for the analysis of different hypotheses between la-
tent constructs is covariance structure analysis or structural equation mod-
eling. 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 re-
quires high external validity that can only be ensured through a large sam-
ple 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 re-
search (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 re-
sponse 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 per-
sonal 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 respon-
dents 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 fur-
thermore be saved on-line, enabling respondents to stop answering and re-
suming 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 cus-
tomer 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 infor-
mants 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 motiva-
tions 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 per-
formance. 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 intro-
duced in chapter 4, they were operationalized (see chapter 6) and devised
to the participants of the survey as statements on a seven point Likert-
scale, 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 fur-
thermore 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
Khne-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 Uni-
versity, 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 respon-
dents with fewer contacts, thereby increasing the response rate of the indi-
vidual 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 op-
posed 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 out-
sourcing 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 re-
spondent 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 re-
searchers of the Khne-Center for Logistics Management reviewed the
questionnaire before between July and August 2004 four logistics profes-
sors and ten experts from different industries participated in personal in-
terviews 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 sugges-
tions exist on the number of pre-test interviews to be conducted, it was de-
cided 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 modifi-
cations 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 cen-
tral for logistics outsourcing relationships and different to the construct
communication. Overall, the questionnaire was perceived as well struc-
tured, 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 execu-
tives that are involved in achieving or responsible for overall logistics per-
formance and for relationships with logistics service providers. Many of
these logistics executives in Germany are members of the BVL (Bundes-
vereinigung 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 data-
base containing approximately 6,800 members, only a selection of the in-
dustry 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 ad-
dresses 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
Khne-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 re-
mained, 2,789 from the BVL sample and 613 from the sample of the
Khne-Center.
The links to the questionnaire were e-mailed to the sample firms in Oc-
tober 2004 together with a cover letter that highlighted the relevance and
the importance of the topic and ensured strict confidentiality of the partici-
pants 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 Logistik-
Symposium or the free participation at a workshop on the topic of logis-
tics 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 ap-
proached, 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 miss-
ing 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 re-
quest 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 val-
ues 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 representa-
tive 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 sig-
nificant 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 ex-
ists.
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 respon-
dents are either general manager of the firm or logistics manager. Addi-
tionally, 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 in-
dustries represented strongest are the automotive industry with 16%, retail-
ing with 15% and the electronics, precision mechanics and optics indus-
tries 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 reve-
nues 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 distrib-
uted in the sample, the overall results are very satisfactory.
154 5 Methodology and sample characteristics

Food, Beverage n.a.

and Tobacco
More than 1 Bn. up to 50 Mio.
Other
6 Automotive 4
17 11

Health- 16 27 51-100
care 9 Mio.
5

Consumer
9
Goods
15
16
Retail
8 14 101-250
Mio.
10 Manufacturing
14 Systems 501-1,000 19
Construction Mio.
Chemicals
and Plastics Electronics,
251-500 Mio.
Precision Mechanics
and Optics

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

Since one major goal of this study is the analysis of relationships be-
tween 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 3 4 5 6 7
Does not apply 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 fore-
most through cost reduction considerations.
A comparison of these results over time promises to yield further in-
sights. In his study on logistics outsourcing, ENGELBRECHT (2004, pp.
241-243) used the same indicators to asses the reasons for logistics out-
sourcing. Most notably, the top five reasons are the same in the two stud-
ies, 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 in-
crease 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 indi-
vidual 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 util-
ized 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 be-
tween underlying theoretical, latent variables (HOMBURG 1989, p. 2).
Structural equation modeling has its roots in path analysis, a method de-
veloped in biometrics to graphically illustrate causal relationship, which
was enhanced in the 1960s by researchers in sociology, the most notable
of whom are BLALOCK (1963), DUNCAN (1966), and DUNCAN/HALLER/
PORTES (1968). In the early 1970s the different approaches were ex-
panded 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), JRESKOG (1973), WILEY (1973), and JRESKOG

(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 ana-
lyze causal linkages and relationships between constructs, these must be
measured. Despite the difficulties for construct validity associated with in-
direct 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 understand-
ing 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 meas-
urement 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 one-
factor case, the construct represents exactly one factor on which all meas-
ured 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 one-
factor constructs will be modeled, multi-factor constructs will not be ad-
dressed in further detail.
The measurement of one-factor constructs can be performed through a sin-
gle 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 or-
der 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 indica-
tors

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). Reflec-
tive indicators on the contrary are caused by the factor as displayed in fig-
ure 17. Ovals represent the latent, exogenous variable, squares the observ-
able 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 indica-
tors 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
(JRESKOG/SRBOM 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 confir-
matory 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 re-
strictions for factor loading and measurement error variances, require at
least three indicators (BAGOZZI 1994, pp. 323-331). Furthermore, consis-
tent with the general procedure all coefficients between the manifest vari-
ables 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 Measurement model of
latent, exogenous variable latent, endogenous variable

]
G1 x1 O1 O3 y1 H1

J
Exogenous Endogenous
variable [ variable K

G2 x2 O2 O4 y2 H2

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 ab-
solute 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 link-
ages between the latent variables. This reduces the previously many rela-
tionships 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 em-
pirically derived covariance matrix fits the matrix resulting from the theo-
retically hypothesized model. The better the unknown parameters are esti-
mated, 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; BACK-
HAUS/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. 1101-
1102). 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 un-
weighted-least-squares (ULS) and the generalized-least-square (GLS)
methods. Its application requires a sufficiently large sample size, a multi-
variate normal distribution of the indicators and the scale of the observed
variables to be continuous (BYRNE 2001, p. 70). It must be noted, how-
ever, that the maximum-likelihood method is quite robust when the as-
sumption 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 continu-
ously 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 re-
quirement 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 covari-
ance matrix of the indicators includes enough information for a clear speci-
fication 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 pa-
rameters 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 indi-
cators q in a model must satisfy the following equation in order to identify
the model:
q (q  1) (1)
td
2

Consequently, the number of parameters t that have to specified cannot

exceed the number of empirical covariances q(q+1). The difference be-
tween 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 measure-
ment 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 esti-
mate the amount of systematic variance in a measure and indicate the ex-
tent 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 system-
atic error, and the random error (PETER 1979; PETER/CHURCHILL 1986).
With the increase of the random error component, the measure is becom-
ing more unreliable while the degree of systematic error does not at all re-
duce 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 con-
sistency (HEELER/RAY 1972, p. 361). While test-retest stability cannot be
measured due to the nature of this study, the internal consistency is thor-
oughly 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 indica-
tors 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 reli-
able measurement.
The reliability of the measurement is a necessary, yet not a sufficient
condition of validity. It is furthermore necessary to test for systematic er-
rors, 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 hierar-
chy 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 character-
istic 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 nomologi-
cal 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 scales content va-
lidity 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 devel-
oped 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 struc-

tural 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 psychomet-
ric 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 con-
firmatory 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 as-
sessed 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. 259-
232). 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 load-
ing matrix. Factor loadings represent the association of indicators with a
factor and are therefore a measure of their correlation. The relationship be-
tween 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 as-
sumes 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 determi-
nation of the number of factors to be extracted (BACKHAUS/ERICHSON/
PLINKE/WEIBER 2003, pp. 295-297). The factors are arranged in a coordi-
nate 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. Conse-
quently, 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 co-
efficients, for one factor are higher than 0.4 (HOMBURG 1995, p. 93;
HOMBURG/GIERING 1996, p. 8) and have a significantly lower factor load-
ing 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 con-
cerning the structure of the analyzed constructs. While this assesses the va-
lidity of the measurement, the reliability is assessed through coefficient al-
pha and item-to-total correlation.
Coefficient alpha, or Cronbachs 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 al-
pha 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 cor-
related (CARMINES/ZELLER ,1979, p. 45; MALHOTRA 1993, p. 308;
HOMBURG/GIERING 1996, p. 8). The formula of coefficient alpha as pro-
posed by CRONBACH (1951, p. 299) is given below:
 5.2 Methodological basis for the empirical analysis 165

Vi (2)
n i

D 1 
n 1 Vt

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, meas-
urement 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 coef-
ficient alpha must always be judged also in terms of the number of indica-
tors 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 indi-
cator 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 indica-
tor.
It must be observed, however, that the terminology in literature some-
times is quite imprecise and refers to item-to-total correlation when ac-
tually the corrected item-to-total correlation is meant (WALLENBURG
2004, p. 146), even though the former measures the correlation of one in-
dicator 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 reliabil-
ity 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 gen-
eration to assess the reliability of measurement and structural models. Con-
trary to the exploratory factor analysis, hypotheses on the relationships be-
tween indicators and factors are specified ex-ante. While the exploratory
factor analysis serves to empirically identify the presupposed factor struc-
ture, the confirmatory factor analysis evaluates it and indicates model fit.
For the evaluation of overall measurement and structural model fit, sev-
eral 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/BAUM-
GARTNER (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

Relative global Comparative Measures for the Measures for the

measures global measures structural model measurement model
CAIC Squared multiple correlations Composite reliability
ECVI Average variance extracted

Incremental measures Stand-alone

measures

Measures considering Measures ignoring Descriptive measures Inference statistical

degrees of freedom degrees of freedom measures
CFI NFI F2
TLI RMSEA

Measures considering Measures ignoring

degrees of freedom degrees of freedom
F2/df 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 Akaikes information criterion
(AKAIKE 1974; AKAIKE 1987; BOZDOGAN 1987), and the ECVI, the ex-
pected 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 in-
cremental 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 statis-
tical tests. A first measure is the F2-test with the hypothesis h0 that the em-
pirical covariance matrix S corresponds fully to the covariance matrix of
the model 6 . The F2-value

F2 n  1 F S , 6  (3)

is calculated with
1 (4)
df  p  q  p  q  1  r
2

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 pa-
rameters 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 ob-
served 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 co-
variance matrix S do not correspond to the covariance matrix of the
model 6 .
The shortcomings of the F2-test are overcome in the second inference
statistical measure RMSEA. The root-mean-squared error of approxima-
tion was developed by STEIGER/LIND (1980) and is one of the most com-
mon 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:

F 2  df (5)
RMSEA
df (n  1)
 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 in-
dication 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 empiri-
cal covariance matrix S and the matrix generated by the model 6 . It can
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/BAUM-
GARTNER (1995a, p. 167), a value above 0.9 can be considered satisfac-
tory. For the calculation, the following formula is used in a ML-setting
(JRESKOG/SRBOM 1982, p. 408):

GFI 1

tr 6 1 S  I 
2 (6)


tr 6 1 S 
2

S and 6 represent the covariance matrices as introduced above, I is the

identity matrix and tr denotes the sum to the diagonal elements of the ma-
trix.
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 co-
variances of a database are easier to explain. In contrast to the GFI, the
AGFI (adjusted goodness of fit index) is a descriptive measure that ac-
counts 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 num-
ber of indicators and degrees of freedom. According to
JRESKOG/SRBOM (1982, p. 408), the AGFI with slightly adapted de-
nomination is calculated as follows, with p being the number of parameters
and df the degrees of freedom:
p p  1 (7)
AGFI 1 1  GFI 
2df
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 devel-
oped, still the more rigid demand of 0.9 as a minimum value will be fa-
voured 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 ac-
cept 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 be-
long to the stand-alone measures which asses the model on an isolated ba-
sis. 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 con-
sider 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:
max^F r2  df r ;0` (8)
CFI
max^F b2  df b ; F r2  df r ;0`

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 mod-
els. 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 F r2 (9)

df b df r
TLI 2
F b
1
df b

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 satisfac-
tory 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 struc-
tural 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 valid-
ity 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 for-
mula:
172 5 Methodology and sample characteristics

2 (10)

Oij I jj
CR [ j  i
2

Oij I jj  T ii
i i

Here, Oij represents the estimated factor loading, I jj the estimated vari-
ance 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:
(11)
O I 2
ij jj
AVE [ j  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 indi-
cate 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 indica-
tors with low indicator reliability from the measurement model as long as
content validity is guaranteed (see chapter 5.2.3.2). Indicator reliability as-
sesses 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 de-
nomination again is the same as in composite reliability and average vari-
ance extracted:
 5.2 Methodological basis for the empirical analysis 173

Oij2 I jj (12)
IR xi 
Oij2 I jj  T ii

The indicator reliability can take on values between 0 and 1. Values

above 0.4 are typically considered satisfactory (HOMBURG/BAUM-
GARTNER 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:
Oi (13)
t  xi 
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 constrain-
ing the estimated correlation parameter I between them to 1 and then per-
forming a F2-difference-test on the values obtained for the constraint and
for the unconstrained model (GRANZIN/PAINTER 2001, pp. 82-84). A sig-
nificantly lower F2-value for the model in which the trait correlations are
not constrained to unity would indicate that the traits are not perfectly cor-
related 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 ex-
ists 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 to-
tal 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 in-
fluencing exogenous latent variables (BYRNE 2001, p. 163). The squared
multiple correlations take on values between 0 and 1, whereat higher val-
ues 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 se-
lected influencing factors are included in the model (HOMBURG/BAUM-
GARTNER 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 Threshold Value

Coefficient Alpha 0.6

Explained Factor Variance 0.5
Factor loadings 0.4 Criteria of the
Adaptation first generation
Item-to-total correlation lowest
of the
measurement t-value of factor loadings ~1.96~
model
Indicator reliability 0.4
Composite reliability 0.6
Average variance extracted 0.5

CAIC lowest
ECVI lowest
Adaptation F2/df 3 Criteria of the
of the
measurement TLI 0.9 second generation
model and GFI 0.9
the structural
model AGFI 0.8
CFI 0.9
RMSEA 0.08
Adaptation of the
Squared multiple correlation (R2)
structural model

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 multi-
ple 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 adapta-
tion 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 ana-
lyzed 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 in-
dicators 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 gen-
eration 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 im-
provement. 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 coeffi-
cient alpha is determined for the original group of indicators. If the thresh-
old value of 0.6 is not reached, the item with the lowest item-to-total cor-
relation 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. Fur-
thermore, the factor must explain at least 50% of the variance of its indica-
tors. 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 conver-
gent validity is obtained.
While the procedure proposed by HOMBURG/GIERING (1996) is espe-
cially 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 valid-
ity 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 ad-
vantage of low complexity in the operationalization process, it does carry
the danger that the elimination of indicators on the basis of convergent va-
lidity as measured through the item-to-total correlation, leads to the elimi-
nation of very reliable indicators and this even though the valid meas-
urement 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 indica-
tor 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 dif-

ferent 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 (WALLEN-
BURG 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 repre-
sent 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 in-
creased by including formerly not hypothesized relationships into the
model. However, this is not always an adequate procedure. The modifica-
tions 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 analy-
sis 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 com-
plexity of a model increases disproportionately with the number of vari-
ables, since aside from the hypothesized relationships also the interde-
pendencies 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 la-
tent 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 load-
ings of each indicator to weight the individual scores. Subsequently for
both single item and latent variable measurements, median splitting is con-
ducted to divide the data set into two parts (DURVASULA/ANDREWS/
LYSONSKI/NETEMEYER 1993; HOMBURG/ GIERING/MENON 2003). The re-
sulting 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 sam-
ple 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 logis-
tics outsourcing, logistics, and the firm are examined. As no sound reason-
ing is brought forward for other sources of differences such as variances,
covariances, or measurement errors, theses are not considered in the fol-
lowing 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. 176-
197). It tests the hypothesis that both samples covariance matrices are dif-
ferent which reads as a formula:
H 0 : T1  z T 2  (14)

where 6T 1  is the model covariance matrix in sub-sample 1 and 6T 2 

is the model covariance matrix for sub-sample 2. This hypothesis is re-
flected 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 es-
timation is performed, imposing equality constraints on the structural paths
leading to all performance dimensions in the respective model. If differ-
ences exist in the model in general and in the constrained paths in particu-
lar, 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. Further-
more, 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 179

'F 2 F c2  F u2 , 'df df c  df u (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 free-
dom. 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 con-
strained 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 restrict-
ing a certain path does not lead to a significant increase in F2, thereby indi-
cating 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 esti-
mated. While this already reveals differences in the model, only the com-
parison 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 cus-
tomer-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 pro-
posed model, the works of FRAZIER (1983) and LARSON/KULCHITSKY
(1999) offer a solid basis for the development of a logistics outsourcing re-
lated cooperation construct.
GUILTINAN/REJAB/RODGERS (1980) empirically find that coordination
and cooperation tend to be high when interfirm communications are per-
ceived 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 coop-
eration. 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 dis-
tinct logistics focus, is provided by LARSON/KULCHITSKY (1999, p. 94).
They point out that cooperative relationships are characterized by collabo-
rative 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 indica-
tors is displayed in Table 6-1.
182 6 Construct operationalization

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

Please indicate your level of agreement with the following statements on your
relationship with this LSP
Indicator 1 The goals of our relationship were jointly set by us and our LSP.
Indicator 2 Our approach to doing business or organizing activities is very similar to our LSP.
Indicator 3 In the relationship with our LSP, we always pull together in the same direction.
Indicator 4 When problems or questions arise during this outsourcing relationship, we make
decisions together with our LSP to get to adequate solutions.
Indicator 5 If one partner exercises his power in the relationship, he does it in an appropriate way.
Indicator 6 In our business relationship, both parties fully respect each other.
Indicator 7 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 indi-

cators. Indicator 2 reflects high levels of ideological agreement, indicator 4
participative decision making, indicator 5 the use of power in a non-
pressurized form and indicator 8 role satisfaction, facilitating general co-
operation. The findings of LARSON/KULCHITSKY (1999, p. 94) are embod-
ied 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 re-
search, indicators explicitly targeting information exchange have been ex-
cluded 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 0.87 AGFI 0.82
Explained variance 47.94% CFI 0.90
F/df 11.130 RMSEA 0.136
TLI 0.86 Composite reliability 0.87
GFI 0.90 Average variance extracted 0.45
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.45 0.22 -
Indicator 2 0.74 0.62 10.76
Indicator 3 0.71 0.64 10.85
Indicator 4 0.64 0.43 9.99
Indicator 5 0.49 0.27 8.85
Indicator 6 0.72 0.62 10.78
Indicator 7 0.54 0.34 9.39
Indicator 8 0.73 0.69 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 insuffi-
cient reliability of the indicators. Consequently, indicators 1, 5, and 7 hav-
ing 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 0.84 AGFI 0.99
Explained variance 58.07% CFI 1.00
F/df 0.730 RMSEA 0.000
TLI 1.00 Composite reliability 0.84
GFI 1.00 Average variance extracted 0.57
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 2 0.72 0.67 -
Indicator 3 0.74 0.72 19.83
Indicator 4 0.61 0.44 15.61
Indicator 6 0.64 0.50 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 commu-
nication 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 par-
ties 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 indi-
cator 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

Please indicate your level of agreement with the following statements on your
relationship with this LSP
Indicator 1 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 To reach our goals, a lot of meetings and talks with this LSP are necessary.
Indicator 4 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 Both sides can always fully rely on the information we exchange.
Indicator 7 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 qual-
ity of the exchange of information and its usability to solve arising prob-
lems 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 0.86 AGFI 0.86
Explained variance 51.04% CFI 0.93
F/df 11.800 RMSEA 0.140
TLI 0.89 Composite reliability 0.86
GFI 0.94 Average variance extracted 0.51
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.57 0.35 -
Indicator 2 0.68 0.53 13.11
Indicator 4 0.47 0.23 9.67
Indicator 5 0.68 0.53 13.12
Indicator 6 0.71 0.68 14.20
Indicator 7 0.75 0.73 14.46

The resulting seven indicators were tested in an exploratory factor

analysis. However, indicator 3 which explicitly covers the quantitative as-
pects of information exchange only, was found to load on a second factor.
It was therefore eliminated from the measurement model. As Table 6-5 in-
dicates, the measurement model with six indicators did not satisfy in all
dimensions with insufficient values for RMSEA, TLI, and AGFI. Further-
more, 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 reliabil-
ity well below the threshold of 0.4. Furthermore, their error terms dis-
played very high correlations with the error terms of the remaining indica-
tors.

Table 6-6. Adaptation measures for the construct communication (4 indicators)

Information on the factor communication (4 indicators)
Coefficient alpha 0.86 AGFI 1.00
Explained variance 61.36% CFI 1.00
F/df 0.597 RMSEA 0.000
TLI 1.00 Composite reliability 0.86
GFI 1.00 Average variance extracted 0.61
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 2 0.64 0.48 -
Indicator 5 0.65 0.50 14.89
Indicator 6 0.75 0.71 17.31
Indicator 7 0.78 0.77 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 with-
out 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 in-
creasing 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 improve-
ments in efficiency and effectiveness of the customers logistics processes
through the LSP. Apart from a generally high level of innovation as meas-
ured through indicator 5, the continuous improvement and modification of
logistics processes by the LSP is relevant (indicators 1 and 3). Further-
more, it is particularly valuable for the customer if the suggestions for im-
provements are going beyond the direct responsibility of the LSP (indica-
tor 2) and are proactive (indicator 4).

Table 6-7. Indicators for the measurement of the construct proactive improvement
Please indicate your level of agreement with the following statements on the
improvement of logistics systems.
Indicator 1 The LSP puts strong effort into continously optimizing logistics processes.
Indicator 2 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 The LSP shows initiative by approaching us with suggestions for improvement.
Indicator 5 The LSP shows a high level of innovation.

As Table 6-8 indicates, the measurement model is showing almost satis-

factory 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, in-
dicator 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 con-
tains the important aspect of the LSPs 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 indi-

cator reliability.

Table 6-8. Adaptation measures for the construct proactive improvement (5 indi-
cators)
Information on the factor proactive improvement (5 indicators)
Coefficient alpha 0.92 AGFI 0.97
Explained variance 69.59% CFI 1.00
F/df 2.959 RMSEA 0.060
TLI 0.99 Composite reliability 0.92
GFI 0.99 Average variance extracted 0.70
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.75 0.61 -
Indicator 2 0.74 0.60 19.53
Indicator 3 0.80 0.70 21.39
Indicator 4 0.84 0.81 23.39
Indicator 5 0.82 0.76 22.64

The resulting measurement model with only 4 indicators shows satisfac-

tory 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 indi-
cators)
Information on the factor proactive improvement (4 indicators)
Coefficient alpha 0.91 AGFI 1.00
Explained variance 71.65% CFI 1.00
F/df 0.561 RMSEA 0.000
TLI 1.00 Composite reliability 0.91
GFI 1.00 Average variance extracted 0.72
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 2 0.74 0.60 -
Indicator 3 0.78 0.69 20.78
Indicator 4 0.85 0.83 23.15
Indicator 5 0.81 0.75 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 interor-
ganizational 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 operationaliza-

tion of the construct. However, a two-dimensional structure of the con-
struct will not be ex-ante assumed. While some studies find that discrimi-
nant 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 benevo-
lence could be conceptually distinct, they may be so intertwined in busi-
ness 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 ex-
planatory 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. Conse-
quently, 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

Please indicate your level of agreement with the following statements regarding
your LSP
Indicator 1 This LSP keeps promises it makes to our firm.
Indicator 2 Regarding problems, this LSP is always honest with us.
Indicator 3 This LSP is trustworthy.
Indicator 4 This LSP its genuinely concerned that our business succeeds.
Indicator 5 The LSP correctly carries out tasks that we cannot directly control.
Indicator 6 When making important decisions, this LSP considers our welfare as well as its own.

Since the conceptualization of trust is rather diffuse in exchange rela-

tionship research, a large number of different scales exist for the opera-
tionalization. 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 benevo-

lence of the LSP.

Table 6-11. Adaptation measures for the construct trust (6 indicators)

Information on the factor trust (6 indicators)
Coefficient alpha 0.92 AGFI 0.81
Explained variance 66.04% CFI 0.95
F/df 15.392 RMSEA 0.162
TLI 0.91 Composite reliability 0.92
GFI 0.92 Average variance extracted 0.65
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.71 0.58 -
Indicator 2 0.83 0.82 22.92
Indicator 3 0.86 0.85 23.48
Indicator 4 0.74 0.57 18.44
Indicator 5 0.74 0.56 18.27
Indicator 6 0.75 0.57 18.61

Since a unidimensional structure of the construct was assumed, it was

examined first. An exploratory factor analysis revealed that all six indica-
tors load on only one factor. This was also confirmed by a confirmatory
factor analysis. The resulting measurement model with six indicators how-
ever 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 elimi-
nated because of the high correlation of its error term with those of the in-
dicators 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 modi-
fication 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 sat-
isfactory criteria in all dimensions and can therefore be accepted without
further modification. It must be noted, however, that both indicators aim-
ing 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)

Coefficient alpha 0.90 AGFI 0.99
Explained variance 70.47% CFI 1.00
F/df 1.384 RMSEA 0.026
TLI 1.00 Composite reliability 0.90
GFI 1.00 Average variance extracted 0.70
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.74 0.60 -
Indicator 2 0.85 0.86 23.98
Indicator 3 0.84 0.84 23.72
Indicator 5 0.69 0.52 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 cus-
tomers 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 com-
mitment.
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 co-
variance 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

To what degree do you agree to the following statements on your attitude towards
this LSP?
Indicator 1 We come to our LSP's defense when it is criticized by persons from inside or outside our
organization.
Indicator 2 We would be very sorry personally if we had to terminate the relationship with this LSP.
Indicator 3 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 commit-
ment construct in a logistics context by WALLENBURG (2004, pp. 189-
192). Indicators 1 and 2 come from ZIMMER (2000, p. 218) and reflect the
 6.1 Antecedents of logistics outsourcing performance 191

degree of the customers commitment to vouch for the LSP (indicator 1)

and the personal commitment of the customers employees towards the re-
lationship (indicator 2). The emotional ties of the customers 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 measure-
ment 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)
Coefficient alpha 0.81 AGFI 0.94
Explained variance 55.34% CFI 0.99
F/df 6.286 RMSEA 0.098
TLI 0.96 Composite reliability 0.82
GFI 0.99 Average variance extracted 0.53
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.66 0.52 -
Indicator 2 0.71 0.72 17.32
Indicator 3 0.52 0.30 11.88
Indicator 4 0.69 0.67 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. 189-
192).
192 6 Construct operationalization

Table 6-15. Adaptation measures for the construct commitment (3 indicators)

Information on the factor commitment (3 indicators)

Coefficient alpha 0.84 AGFI *
Explained variance 63.47% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.84
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.64 0.49 -
Indicator 2 0.74 0.75 16.31
Indicator 4 0.72 0.66 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 con-
flict 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). How-
ever, 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 pre-
sented in Table 6-16.
 6.1 Antecedents of logistics outsourcing performance 193

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

Please indicate your level of agreement with the following statements on your
relationship with this LSP
Indicator 1 When problems occur we always solve them jointly.
Indicator 2 Especially when solving problems, the exchange of information between the LSP and us
is working very well.
Indicator 3 When problems occur, it is always the same party tackling and solving them.
Indicator 4 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 in-
dicator ENGELBRECHT (2004, p. 217) uses for the measurement of the con-
struct 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 as-
sumed to exist.

Table 6-17. Adaptation measures for the construct functional conflict (5 indica-
tors)
Information on the factor functional conflict (5 indicators)
Coefficient alpha 0.81 AGFI 0.98
Explained variance 50.45% CFI 1.00
F/df 2.049 RMSEA 0.044
TLI 0.99 Composite reliability 0.82
GFI 0.99 Average variance extracted 0.48
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.73 0.71 -
Indicator 2 0.76 0.81 23.25
Indicator 5 0.47 0.25 11.96
Indicator 6 0.43 0.23 11.34
Indicator 7 0.64 0.51 18.38

An initial exploratory factor analysis extracted two different factors

from the seven indicators presented in Table 6-16. Indicators 3 and 4 there-
fore 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 6-
17 indicates, the average variance extracted is not satisfactory. Further-
more, the indicator reliabilities of indicators 5 and 6 are well below the re-
quested 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 ex-
plained variance and the average variance extracted, exhibit high and satis-
factory 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 indica-
tors)

Information on the factor functional conflict (3 indicators)

Coefficient alpha 0.85 AGFI *
Explained variance 67.71% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.86
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.74 0.69 -
Indicator 2 0.79 0.84 20.75
Indicator 7 0.66 0.50 17.90

6.1.7 Involvement

As argued in chapter 4.2.1.7, the involvement of the LSP in the implemen-

tation process is an essential success factor for the success of logistics out-
sourcing 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 lo-
gistics 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 The LSP was significantly involved in the outsourcing relationship at an early stage.
Indicator 2 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 communica-
tion problems, enables the exchange of know-how and fosters familiariza-
tion 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 de-
veloped 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, individ-
ual tasks can be delegated to subordinate teams or individuals. For the suc-
cess of an outsourcing arrangement it is crucial that also these employees
collaborate effectively and efficiently with each other and with their coun-
terparts on the side of the LSP. This can only be enabled if the LSP is in-
tensively 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. Further-
more, 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)
Coefficient alpha 0.61 AGFI *
Explained variance 37.80% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.61
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.44 0.41 -
Indicator 2 0.39 0.24 7.62
Indicator 3 0.47 0.48 7.29

To improve the measurement model, indicator 2 with an indicator reli-

ability of only 0.24 was eliminated. The resulting measurement model with
only two indicators failed to meet the required criteria. While the indica-
tors again loaded on one factor only, the value of the coefficient alpha was
0.613 and the explained variance only 44.278%. Other adaptation meas-
ures cannot be calculated with only two indicators and no degrees of free-
dom.
The measurement model with two indicators, like the one with three in-
dicators, 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 gen-
eral 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 as-
pects 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

Please indicate your level of agreement with the following statements on your
relationship with this LSP
Indicator 1 To accomplish its own objectives, sometimes our LSP alters the facts slightly.
Indicator 2 Our LSP always provides us with a completely truthful picture of its activities.
Indicator 3 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 Our LSP feels that it is alright to do anything within its means to further its own interests.
Indicator 6 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 lan-
guage 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 ac-
tions 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 cri-
teria 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 accept-
able, while others, such as the TLI, the AGFI and the average variance ex-
tracted, were slightly too low.

Table 6-22. Adaptation measures for the construct opportunism (6 indicators)

Information on the factor opportunism (6 indicators)

Coefficient alpha 0.83 AGFI 0.84
Explained variance 45.19% CFI 0.92
F/df 13.503 RMSEA 0.151
TLI 0.84 Composite reliability 0.83
GFI 0.93 Average variance extracted 0.47
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.60 0.45 -
Indicator 2 0.49 0.27 10.70
Indicator 3 0.70 0.64 15.39
Indicator 4 0.70 0.64 15.39
Indicator 5 0.54 0.39 12.63
Indicator 6 0.55 0.33 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 sev-
eral other error term, was eliminated. The resulting measurement with 5
indicators, however, did still not meet all requested criteria either. Espe-
cially 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 meas-
urement model can therefore be accepted without any further modifica-
tions.

Table 6-23. Adaptation measures for the construct opportunism (4 indicators)

Information on the factor opportunism (4 indicators)
Coefficient alpha 0.82 AGFI 0.99
Explained variance 54.00% CFI 1.00
F/df 1.311 RMSEA 0.024
TLI 1.00 Composite reliability 0.82
GFI 1.00 Average variance extracted 0.54
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.59 0.44 -
Indicator 3 0.67 0.60 14.61
Indicator 4 0.72 0.71 15.09
Indicator 5 0.58 0.42 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 estab-
lished 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

Please indicate the level of agreement you believe exists between you and your
LSP on the following issues
Indicator 1 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 re-

sults, 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 be-
havior. 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 func-
tioning 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 meas-
urement 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 adapta-
tion 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 0.82 AGFI 0.81
Explained variance 50.88% CFI 0.92
F/df 19.577 RMSEA 0.184
TLI 0.84 Composite reliability 0.83
GFI 0.94 Average variance extracted 0.53
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.42 0.15 -
Indicator 2 0.57 0.29 7.64
Indicator 3 0.70 0.66 8.70
Indicator 4 0.73 0.79 8.83
Indicator 5 0.69 0.64 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 indica-
tor reliability of indicator 2 is below the proposed minimum of 0.4. How-
ever, since with indicator one the only other indicator measuring the be-
havior 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 con-
tent validity. Therefore, no further modifications were made to the four-
indicator measurement model.

Table 6-26. Adaptation measures for the construct shared values (4 indicators)
Information on the factor shared values (4 indicators)
Coefficient alpha 0.84 AGFI 0.99
Explained variance 59.38% CFI 1.00
F/df 1.521 RMSEA 0.031
TLI 1.00 Composite reliability 0.86
GFI 1.00 Average variance extracted 0.62
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 2 0.49 0.27 -
Indicator 3 0.73 0.66 12.11
Indicator 4 0.78 0.81 12.48
Indicator 5 0.72 0.64 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 rela-
tionship 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 stra-
tegic 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, in-
formal, 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 pre-
sented in Table 6-27.

Table 6-27. Indicators for the measurement of the construct openness

To what degree do you agree with the following statements on your relationship
with this LSP?
Indicator 1 The relationship with our LSP is informal.
Indicator 2 When problems or questions arise, we can quickly contact the LSP and jointly find
solutions.
Indicator 3 The relationship with our LSP is very open.
Indicator 4 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 indica-

tor 2 quantifies the extent to which spontaneous solutions are found be-
tween the two parties once problems or questions arise. A further aspect
suggested by GUPTA (1987, p. 499) is the general openness of the relation-
ship, 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 neces-
sary.
Since the framework put forward by GUPTA (1987) was originally de-
veloped 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 prereq-
uisite 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 re-
quested 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)

Coefficient alpha 0.87 AGFI 0.92
Explained variance 54.53% CFI 0.96
F/df 7.450 RMSEA 0.109
TLI 0.94 Composite reliability 0.87
GFI 0.96 Average variance extracted 0.52
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.59 0.40 -
Indicator 2 0.72 0.65 15.53
Indicator 3 0.81 0.82 16.68
Indicator 4 0.53 0.32 11.70
Indicator 5 0.64 0.46 13.64
Indicator 6 0.73 0.61 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 sen-
sitive 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 meas-
urement 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 er-
ror 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 Ta-
ble 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)
Coefficient alpha 0.84 AGFI 0.99
Explained variance 58.58% CFI 1.00
F/df 0.738 RMSEA 0.000
TLI 1.00 Composite reliability 0.85
GFI 1.00 Average variance extracted 0.59
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.58 0.40 -
Indicator 2 0.75 0.73 15.50
Indicator 3 0.77 0.78 15.65
Indicator 5 0.60 0.43 12.95

6.2 Logistics outsourcing performance

As it has been argued in chapter 4.1, the performance of logistics outsourc-

ing 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 ex-
ceedance. 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, quali-
tative and strategic goals must be part of the goal achievement construct.
While their quantification may be substantially more difficult than of ser-
vice 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 lo-
gistics 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 rela-
tionship. Therefore, the construct of goal achievement has the two aspects
of achieving the actual goals agreed upon in the contract between the cus-
tomer 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 We are very satisfied with our LSP.
Indicator 3 The relationship with our LSP is very good.
Indicator 4 Our LSP delivers its services always with the required quality.
Indicator 5 Our LSP delivers its services always in the required time.
Indicator 6 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 lev-
els respectively. Therefore, a high score on any of these three indicators re-
flects the achievement of the previously set goals and expectations only,
not their exceedance.
To measure the overall satisfaction of the customer with the relation-
ship, indicator 2 was included. This dimension is extended through indica-
tor 3, which measures the subjective perception of the customer of the rela-
tionship 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 signifi-
cantly 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 indi-
cator 1, where the expectations can only be fulfilled if the cost perform-
ance 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 fur-
ther risk and therefore does not view it as a determinant of goal achieve-
ment.

Table 6-31. Adaptation measures for the construct goal achievement (6 indicators)
Information on the factor goal achievement (6 indicators)
Coefficient alpha 0.89 AGFI 0.83
Explained variance 64.13% CFI 0.95
F/df 14.411 RMSEA 0.156
TLI 0.92 Composite reliability 0.89
GFI 0.93 Average variance extracted 0.58
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.79 0.71 -
Indicator 2 0.85 0.89 30.92
Indicator 3 0.83 0.85 29.44
Indicator 4 0.79 0.69 24.68
Indicator 5 0.70 0.53 19.91
Indicator 6 0.39 0.15 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 indica-
tor reliability, its error term exhibited very high correlations with the error
terms of indicators 2 and 4, indicating the view of the customers that de-
livering 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 ex-
cellent adaptation measures. The measurement model can therefore be ac-
cepted without any further modifications.

Table 6-32. Adaptation measures for the construct goal achievement (4 indicators)
Information on the factor goal achievement (4 indicators)
Coefficient alpha 0.93 AGFI 0.98
Explained variance 78.26% CFI 1.00
F/df 2.729 RMSEA 0.056
TLI 1.00 Composite reliability 0.94
GFI 1.00 Average variance extracted 0.78
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.81 0.71 -
Indicator 2 0.91 0.92 31.56
Indicator 3 0.87 0.84 29.16
Indicator 4 0.79 0.66 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 out-
sourcing 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 logis-
tics outsourcing context before, a new scale was developed. It will be
based on the concept of goal achievement introduced in the previous chap-
ter, for which the relevant performance aspects of logistics outsourcing
have been identified. For a significant exceedance of the goals and expec-
tations of the customer, the performance in these different aspects, namely
quality of service and cost reductions, must go beyond those originally tar-
geted. However, for the goal achievement, also the aspects of overall satis-
faction and relationship quality were introduced. Since they constitute uni-
versal 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

Please indicate your level of agreement with the following statements on your
satisfaction with your cooperation with this LSP
Indicator 1 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 re-

gards to the exceedance of the goals and expectations set prior to the
commencement of the outsourcing arrangement. The further these are ex-
ceeded 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 op-
erative exceedance in terms of service quality (indicator 2) and cost reduc-
tions (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 ar-
guably 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 sim-
ple goal achievement.
The four-indicator measurement model was tested in an exploratory fac-
tor analysis. While it only extracted one factor and showed a satisfactory
explained variance of 58.98%, especially the criteria of the second genera-
tion 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)
Coefficient alpha 0.85 AGFI -0.05
Explained variance 58.98% CFI 0.73
F/df 188.642 RMSEA 0.585
TLI 0.19 Composite reliability 0.82
GFI 0.79 Average variance extracted 0.55
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.69 0.81 -
Indicator 2 0.70 0.84 25.46
Indicator 3 0.62 0.25 12.21
Indicator 4 0.75 0.39 16.23

In an effort to improve the fit of the measurement model, indicator 3 re-

ferring 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 logis-
tics outsourcing arrangements, LSPs either find no room for goal ex-
ceedance 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 sub-
stantial exceedance of the originally targeted cost reductions is not possi-
ble and therefore is not a relevant aspect of their understanding of goal ex-
ceedance. 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 satis-
factory 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)

Coefficient alpha 0.84 AGFI *
Explained variance 67.48% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.85
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.77 0.80 -
Indicator 2 0.79 0.88 31.56
Indicator 4 0.56 0.34 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 pre-
sented in chapter 4.4.1.1. There, it was pointed out that logistics perform-
ance 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 analy-
sis of the relationships between logistics outsourcing performance, out-
sourcing 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 pre-
sented.

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. 221-
224) was selected as it has proven to be both reliable and valid in a large-
scale 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 perform-

ance on three successive levels: the ability to on-time deliver products, ma-
terial, and information on a first level, the improvement of the performance
capabilities of the firms logistics processes on a second level, and the en-
suring 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 measure-
ment, the indicator ability for on-time delivery is proposed. The second
level addresses the performance of the logistics processes in the dimen-
sions 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 de-
livery 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 6-
36.

Table 6-36. Indicators for the measurement of the construct level of logistics ser-
vices
How does your company's performance on logistics service compare to your
competitors?
Indicator 1 Order lead times
Indicator 2 Delivery times.
Indicator 3 Ability for on-time delivery.
Indicator 4 Delivery reliability.
Indicator 5 Degree of delivery flexibility.
Indicator 6 Degree of damage- and error-free logistical activities.

To test the measurement model, an exploratory factor analysis was con-

ducted which only extracted one factor. However, as Table 6-37 shows,
several adaptation measures indicate insufficient quality of the measure-
ment 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)
Coefficient alpha 0.90 AGFI 0.83
Explained variance 59.84% CFI 0.94
F/df 14.448 RMSEA 0.157
TLI 0.89 Composite reliability 0.90
GFI 0.93 Average variance extracted 0.60
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.70 0.54 -
Indicator 2 0.76 0.65 18.54
Indicator 3 0.75 0.68 18.98
Indicator 4 0.78 0.71 19.29
Indicator 5 0.74 0.61 17.97
Indicator 6 0.60 0.39 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. Conse-
quently, with indicator 1 another item was eliminated from the measure-
ment 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 indica-
tors 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)

Coefficient alpha 0.85 AGFI 0.98
Explained variance 59.14% CFI 1.00
F/df 2.301 RMSEA 0.049
TLI 0.99 Composite reliability 0.85
GFI 1.00 Average variance extracted 0.59
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 2 0.67 0.57 -
Indicator 4 0.75 0.69 18.34
Indicator 5 0.73 0.67 18.13
Indicator 6 0.60 0.44 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 argu-
mentation 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 trans-
portation 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 devel-
oped by ENGELBRECHT (2004, pp. 218-221) proposes five indicators for
the measurement model: The core logistics costs are covered through indi-
cators 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

How does your company's performance on logistics cost compare to your

competitors?
Indicator 1 Logistics costs relative to total revenue (compensation for the LSP included).
Indicator 2 Warehousing costs.
Indicator 3 Transportation costs.
Indicator 4 IT costs in support of logistics activities.
Indicator 5 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 measur-
ing 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 ex-
tracted 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 indi-
cators)
Information on the factor level of logistics costs (5 indicators)
Coefficient alpha 0.75 AGFI 0.84
Explained variance 39.12% CFI 0.90
F/df 14.242 RMSEA 0.155
TLI 0.80 Composite reliability 0.75
GFI 0.95 Average variance extracted 0.38
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.64 0.73 -
Indicator 2 0.53 0.36 12.50
Indicator 3 0.51 0.48 14.04
Indicator 4 0.43 0.17 8.87
Indicator 5 0.47 0.22 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 compa-
rably 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 meas-
urement 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 elimi-
nated.
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 val-
ues. Only the indicator reliability of indicator 2 is with 0.27 below the pro-
posed 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 indi-
cators)
Information on the factor level of logistics costs (3 indicators)
Coefficient alpha 0.74 AGFI *
Explained variance 54.51% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.76
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.70 0.97 -
Indicator 2 0.46 0.27 8.98
Indicator 3 0.54 0.40 9.88

6.4 Firm performance

In chapter 4.4.2 it has been pointed out that the measurement of firm per-
formance has received much attention in business administration research,
which is also reflected in the multitude of different scales and measure-
ment 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 mar-
keting 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 suit-
ability and therefore will be utilized in this research with only minor modi-
fications. 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 sub-
jective 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 firms adaptiveness reflects the abil-

ity 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. 227-
228) 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 Adaptation of products / services to new customer requirements.

Indicator 2 Reacting to new developments in the market.
Indicator 3 Benefiting from new opportunities in the market.

Following the standard procedure for the testing of measurement mod-

els, 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 0.91 AGFI *
Explained variance 76.76% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.91
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.78 0.68 -
Indicator 2 0.87 0.92 26.30
Indicator 3 0.79 0.70 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 there-
fore is of a high strategic importance to the firm. Following IRVING (1995),
DEHLER (2001, pp. 227-228) developed a measurement model with six in-
dicators 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 addi-
tional benefits for the customer (indicator 2) and in customer loyalty,
measured thorough customer retention (indicator 3). Furthermore, quanti-
tative 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 firms own expectations all reflect high mar-
ket 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 Customer satisfaction.

Indicator 2 Value added provided to the customer.
Indicator 3 Retention of existing customers.
Indicator 4 Acquisition of new customers.
Indicator 5 Achieving our desired growth rate.
Indicator 6 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 re-
quested 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 indica-
tors 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 indica-
tors)
Information on the factor market performance (6 indicators)
Coefficient alpha 0.90 AGFI 0.38
Explained variance 58.91% CFI 0.77
F/df 61.403 RMSEA 0.332
TLI 0.62 Composite reliability 0.89
GFI 0.73 Average variance extracted 0.58
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.71 0.35 -
Indicator 2 0.69 0.33 11.51
Indicator 3 0.68 0.36 11.92
Indicator 4 0.68 0.52 13.64
Indicator 5 0.77 0.88 16.05
Indicator 6 0.78 0.88 16.06

While these modifications resulted in a measurement model with sig-

nificantly 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 thresh-
old 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 per-
ceived 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 di-
mensions. 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 indica-
tors)
Information on the factor market performance (4 indicators)
Coefficient alpha 0.85 AGFI 0.99
Explained variance 61.73% CFI 1.00
F/df 0.989 RMSEA 0.000
TLI 1.00 Composite reliability 0.84
GFI 0.99 Average variance extracted 0.57
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.77 0.65 -
Indicator 2 0.75 0.60 23.85
Indicator 3 0.71 0.68 15.34
Indicator 6 0.57 0.41 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 re-
flects a compromise between the explanatory value of the performance in-
dicator 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 informa-
tion challenge it would have posed for the respondent.
The measurement model adapted from ENGELBRECHT (2004, pp. 229-
231) 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 perform-
ance 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 Our return on sales last year with respect to our competition was...
Indicator 2 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 ac-
cepted without any modifications.

Table 6-48. Adaptation measures for the construct market performance (3 indica-
tors)
Information on the factor financial performance (3 indicators)
Coefficient alpha 0.95 AGFI *
Explained variance 85.29% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.95
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.89 0.86 -
Indicator 2 0.88 0.84 36.30
Indicator 3 0.89 0.86 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 dimen-
sions. As introduced in chapter 5.2.3.2, this will be done through the For-
nell/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 lo-
gistics performance, and the effect of logistics performance on firm per-
formance, 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 con-
structs 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 perform-

ance 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 * AGFI 0.86
Explained variance * CFI 0.96
F/df 1.95 RMSEA 0.042
TLI 0.95 Composite reliability *
GFI 0.89 Average variance extracted (AVE) *
* These values are only calculated on the facor level.

Information on the indicators

Factor Indicator Indicator reliability Factor reliability AVE
2 0.61
3 0.68
Cooperation 0.84 0.57
4 0.43
6 0.60
2 0.54
5 0.48
Communication 0.86 0.61
6 0.71
7 0.73
2 0.70
3 0.83
Proactive Improvement 0.91 0.72
4 0.75
5 0.59
1 0.63
2 0.82
Trust 0.91 0.71
3 0.84
5 0.55
1 0.54
Commitment 2 0.72 0.84 0.64
4 0.65
1 0.67
Functional Conflict 2 0.80 0.86 0.67
7 0.57
Involvement 1 0.88 0.88 0.88
1 0.44
3 0.64
Opportunism 0.82 0.54
4 0.66
5 0.41
2 0.26
3 0.67
Shared Values 0.86 0.62
4 0.76
5 0.69
1 0.41
2 0.67
Openness 0.85 0.59
3 0.82
5 0.45
1 0.70
2 0.90
Goal Achievement 0.94 0.79
3 0.86
4 0.68
1 0.82
Goal Over-Achievement 2 0.85 0.85 0.66
4 0.35

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 ar-
gued in the respective chapters to be acceptable.
Finally, the question of discriminant validity can be answered. The For-
nell/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 aver-
age variance extracted. Thus, a high discriminant validity between the con-
structs 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 ex-
planatory value of the model, but is rather seen as an indicator of the tex-
tual 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, com-
munication, trust, and functional conflict. These high correlations, how-
ever, 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 com-
munication, it can be argued that the perceived effect of openness on coop-
eration exists just as much as the positive effect on communication. Fur-
thermore, 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) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Factor AVE 0.57 0.61 0.72 0.71 0.64 0.67 0.88 0.54 0.62 0.59 0.79 0.66
(1) Cooperation 0.57 -
(2) Communication 0.61 0.71 - squared correlations
(3) Proactive Improvement 0.72 0.38 0.26 -
(4) Trust 0.71 0.71 0.61 0.35 -
(5) Commitment 0.64 0.64 0.46 0.28 0.56 -
(6) Functional Conflict 0.67 0.77 0.83 0.34 0.78 0.56 -
(7) Involvement 0.88 0.25 0.29 0.05 0.16 0.21 0.27 -
(8) Opportunism 0.54 0.39 0.29 0.22 0.44 0.29 0.37 0.05 -
(9) Shared Values 0.62 0.57 0.53 0.29 0.52 0.41 0.63 0.10 0.32 -
(10) Openness 0.59 0.64 0.74 0.27 0.63 0.47 0.82 0.27 0.28 0.51 -
(11) Goal Achievement 0.79 0.49 0.42 0.29 0.64 0.47 0.55 0.10 0.31 0.45 0.43 -
(12) Goal Over-Achievement 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 coopera-
tion, communication, and trust. Other than in the case of openness as dis-
cussed 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 rea-
sons for the high correlations, exploratory factor analyses suggest that dis-
criminant validity indeed does not exist. When the Kaiser-criterion is re-
laxed 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 suffi-
ciently discriminant from cooperation, communication, and trust. There-
fore, 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 out-

sourcing 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 al-
ready 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
Coefficient alpha * AGFI 0.96
Explained variance * CFI 0.99
F/df 1.54 RMSEA 0.031
TLI 0.99 Composite reliability *
GFI 0.97 Average variance extracted (AVE) *
* 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 Indicator reliability Factor reliability AVE
1 0.71
2 0.92
0.94 0.79
3 0.84
4 0.66
1 0.83
2 0.85 0.85 0.66
4 0.34
2 0.57
4 0.69
0.85 0.59
5 0.66
6 0.45
1 0.86
2 0.30 0.76 0.52
3 0.45

For the assessment of the overall discriminant validity, the For-

nell/Larcker-criterion was again applied. As Table 6-52 indicates, the av-
erage variance extracted of the constructs exceeds the squared correlations
between the constructs in every case. Therefore, discriminant validity be-
tween 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 per-

formance and logistics performance

Factor (1) (2) (3) (4)

Factor AVE 0.79 0.66 0.59 0.52
(1) Goal Achievement 0.79 - squared correlations
(2) Goal Exceedance 0.66 0.29 -
(3) Level of Logistics Services 0.59 0.08 0.08 -
(4) Level of Logistics Costs 0.52 0.07 0.04 0.22 -

6.5.3 Logistics performance and firm performance

As it has been for logistics outsourcing performance and logistics perform-

ance, 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
Coefficient alpha * AGFI 0.92
Explained variance * CFI 0.97
F/df 2.60 RMSEA 0.054
TLI 0.96 Composite reliability *
GFI 0.94 Average variance extracted (AVE) *
* 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 Indicator reliability Factor reliability AVE
2 0.57
4 0.70
0.85 0.59
5 0.65
6 0.45
1 0.81
2 0.32 0.76 0.52
3 0.47
1 0.70
2 0.90 0.91 0.77
3 0.72
1 0.78
2 0.74
0.86 0.60
3 0.57
6 0.39
1 0.85
2 0.84 0.95 0.85
3 0.86

As Table 6-53 indicates, the confirmatory factor analysis reveals satis-

factory 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 ap-
plied again. As Table 6-54, the average variance extracted of the meas-
urement 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) (2) (3) (4) (5)
Factor AVE 0.59 0.52 0.77 0.60 0.85
(1) Level of Logistics Services 0.59 -
(2) Level of Logistics Costs 0.52 0.23 - squared correlations
(3) Adaptiveness 0.77 0.15 0.07 -
(4) Market Performance 0.60 0.48 0.23 0.22 -
(5) Financial Performance 0.85 0.11 0.12 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 contin-
gency variables will show, their quality has proven to be very limited. Fur-
ther use as moderating variables in structural equation modeling can there-
fore 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 con-

tingency variables on the proposed models will be tested in this study.
Three of them, namely environmental complexity, environmental dynam-
ics, and uncertainty will be operationalized as constructs with multiple in-
dicators. The fourth variable, the customers 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 in-
dicator 4 is determining the number of LSPs the customer is regularly
working with.

Table 6-55. Indicators for the measurement of the construct environmental com-
plexity
Please indicate your level of agreement with the following statements about the
present situation of your company
Indicator 1 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 ef-
fects 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 de-
grees of freedom. For factor 1, measuring the complexity induced inho-
mogeneity 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 ex-
hibits 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 cus-
tomers as well as possible shifting perceptions of the customers on the im-
portance of logistics. Following the argumentation of chapter 4.5.2.1,
changes in the competitive landscape are measured through the competi-
tive pressure in the customers industry (indicator 1). Changes in the rela-
tionships 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 perform-
ance is growing and is increasingly viewed as a source of competitive ad-
vantage, the demand on the customers logistics processes continuously
increases.

Table 6-56. Indicators for the measurement of the construct environmental dy-
namics
Please indicate your level of agreement with the following statements about the
present situation of your company
Indicator 1 The pressure of competition in our industry is very high.
Indicator 2 In our customers' industries we see strong tendencies of concentration.
Indicator 3 The importance of logistics in the eyes of our customers is increasing steadily.
Indicator 4 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 ex-
plained 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 in-
tention. 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 vari-
ance 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
customers 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 re-

lated 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 contin-
gency variable. Therefore, a specific new operationalization was under-
taken.

Table 6-57. Indicators for the measurement of the construct uncertainty

Please indicate your level of agreement with the following statements about the
present situation of your company
Indicator 1 We find it very hard to accurately predict the future behavior of our customers.
Indicator 2 Our customers sometimes exhibit opportunistic behavior.
Indicator 3 In our or our customers industries we presently see strong consolidation tendencies and
insecurity.
Indicator 4 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 re-

spective degrees of bounded rationality (indicator 1) and opportunistic be-
havior (indicator 2), even though the anticipation of opportunism by the
own customer may facilitate the adaptation of the organizations 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 per-
formance 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 0.80 AGFI 0.95
Explained variance 51.07% CFI 0.99
F/df 5.513 RMSEA 0.091
TLI 0.96 Composite reliability 0.80
GFI 0.99 Average variance extracted 0.51
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.63 0.52 -
Indicator 2 0.53 0.34 12.10
Indicator 3 0.63 0.53 14.58
Indicator 4 0.67 0.65 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 modifi-
cations.

Table 6-59. Adaptation measures for the construct uncertainty (3 indicators)

Information on the factor uncertainty (3 indicators)
Coefficient alpha 0.79 AGFI *
Explained variance 56.72% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.80
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.60 0.48 -
Indicator 3 0.61 0.50 13.79
Indicator 4 0.69 0.72 13.68

Before finishing this chapter, the final external contingency variable

customers industry must be introduced. It is composed of the eight indus-
tries 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 vari-

ables are hypothesized to have moderating effects on the models analyzed
in this study. Three of them, namely products, asset specificity, and proc-
ess orientation, will be operationalized as constructs with multiple indica-
tors. The remaining three, size of the customers organization, degree of
centralization of logistics decisions and frequency, can have multiple indi-
cators 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 de-
mands caused by the range of a firms 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 correspond-
ing 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 customers product range from a logistics point of
view.

Table 6-61. Indicators for the measurement of the construct products

Please indicate your level of agreement with the following statements rabout the
present situation of your company
Indicator 1 Our product range is very broad with a lot of different products.
Indicator 2 Our products are so valuable that they require specific logistics measures.
Indicator 3 Our product range is so broad that we need many different transport-, warehouse- and
handling processes.
Indicator 4 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 elimi-
nation of indicators, even up to only 2 remaining indicators, barely im-
proves the explained variance. The highest value attainable with these in-
dicators 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 incor-
porates both the broadness of the product range and the consequences for
the logistics processes.
The size of the customers organization is a further contingency vari-
able. 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 out-
sourcing 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 ques-
tion. 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 centraliza-
tion, 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 con-
text. 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 logis-
tics context with regard to the overall investment of the LSP. Indicator 1
measures whether this investment is usable for the LSP also in later logis-
tics 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
Please indicate your level of agreement with the following statements on the
specific investments your LSP had to make for this relationship
Indicator 1 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 con-

ducted 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 0.68 AGFI *
Explained variance 43.41% CFI *
F/df * RMSEA *
TLI * Composite reliability 0.69
GFI * Average variance extracted 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

Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.42 0.26 -
Indicator 2 0.52 0.49 8.88
Indicator 3 0.55 0.55 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 fre-
quency 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 transac-
tion 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, typi-
cally 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 contin-
gency 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 indica-
tors provides a very close approximation of the results the five-
dimensional 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

Please indicate your level of agreement with the following statements about your
internal logistics organization
Indicator 1 Our Business Unit has a very smooth, continuous, quick and largely failure-free flow of
material and information.
Indicator 2 Our Business Unit is managed in a very flow- or process-oriented way.
Indicator 3 Our Business Unit's degree of flow or process orientation is significantly higher than that
of our competitors.
Indicator 4 All business processes are very well coordinated.
Indicator 5 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 ad-
aptation measures are slightly worse than those reported by DEHLER
(2001, p. 174) and ENGELBRECHT (2004, p. 189) who also used the con-
struct. Especially the F2/df (6.256) and the RMSEA (0.098) are substan-
tially 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 indica-
tors)

Information on the factor process orientation (5 indicators)

Coefficient alpha 0.84 AGFI 0.93
Explained variance 56.40% CFI 0.98
F/df 6.256 RMSEA 0.098
TLI 0.96 Composite reliability 0.85
GFI 0.98 Average variance extracted 0.53
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.70 0.64 -
Indicator 2 0.76 0.76 22.47
Indicator 3 0.57 0.40 15.33
Indicator 4 0.80 0.75 22.41
Indicator 5 0.48 0.27 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 meas-
urement model as its character is also contained in several of the other in-
dicators.

Table 6-66. Adaptation measures for the construct process orientation (4 indica-
tors)
Information on the factor process orientation (4 indicators)
Coefficient alpha 0.79 AGFI 0.98
Explained variance 51.20% CFI 1.00
F/df 1.945 RMSEA 0.042
TLI 0.99 Composite reliability 0.78
GFI 1.00 Average variance extracted 0.48
Information on the indicators
Item-to-total correlation Indicator reliability t-value of factor loadings
Indicator 1 0.66 0.65 -
Indicator 2 0.72 0.82 18.20
Indicator 3 0.52 0.35 13.83
Indicator 5 0.45 0.23 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 in-
dicators 3 and 5 are below the recommended value of 0.4. In the light of
the other satisfactory measures, however, the measurement model is ac-
cepted without any further modifications.
7 Structural models

This chapter will discuss the empirical findings on the effects of the ante-
cedents developed in chapter 4.2 on the two dimensions of logistics out-
sourcing performance. Furthermore, it will examine the influence of logis-
tics 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 vari-
ables 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 em-
pirical analyses of the hypotheses put forward in chapter 4.3.2 on the ante-
cedents and dimensions of logistics outsourcing performance will be pre-
sented. 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 hy-
potheses 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 lo-
gistics outsourcing performance. As argued in chapter 6.5.1, the only ante-
cedent not to be part of the basic model is functional conflict due to the
lack of its discriminant validity.
238 7 Structural models

Commitment

Goal
Cooperation
Achievement
Trust

Involvement

Shared Values Communication

Openness
Proactive Goal
Improvement Exceedance

Opportunism

Fig. 7-1. Basic model

This model shows insufficient model fit when using the maximum-
likelihood estimation function. While the F2/df, TLI, CFI, and RMSEA all
show values well within the desired range or above the respective thresh-
old, 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, es-
pecially 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

Adaptation measures
Model F/df TLI GFI AGFI CFI RMSEA
Basic Model 2.264 0.94 0.87 0.86 0.95 0.048

The insufficient model fit does not result from forgetting causal link-
ages 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 thresh-
old value, is also emphasizes the models 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 de-
sign 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 simplifica-
tion 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 satis-
factory. In the calculation of the 9-factor basic model, the modification in-
dices 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 di-
rectly responsible for significant amounts of complexity in the 9-factor ba-
sic model.
However, to reduce model complexity, only those factors may be elimi-
nated 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 8-
factor 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 in-

creasing the R2 of goal achievement from 60.9% to 63.4% and of goal ex-
ceedance 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 com-
parison of the two models CAIC and ECVI furthermore show substantially
lower values for both indices in the simplified, 8-factor model, again sug-
gesting a better model fit for the simplified model.

Table 7-2. Fit comparison of the basic and the simplified model
Adaptation measures R
Goal Goal
Model F/df TLI GFI AGFI CFI RMSEA CAIC ECVI
Achievement Exceedance
Basic Model 2,264 0.94 0.87 0.86 0.95 0.048 2290.4 3,172 60.9% 34.5%
8-Factor Model 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 pa-
rameters 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
F/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 out-
comes as it does not provide additional explanatory power to the model
and at the same time is statistically showing at least in its current opera-
tionalized 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 H1a
H21 Goal
H11 Cooperation
Achievement
Trust

H20 H9a
H6
H4
H27
H3
Shared Values H24 Communication

H25 H26
H7 H1b
Openness H19
H8
H18 Proactive Goal
H23 Improvement Exceedance
H9b

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 in-
fluencing the two dimensions of logistics outsourcing performance, the
model is taking on a distinct, three level structure that can be seen in Fig-
ure 7-2. Five antecedents, namely shared values, trust, commitment, open-
ness, 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 dimen-
sions of logistics outsourcing performance, goal achievement and goal ex-
ceedance. The factor communication takes on a mediating position be-
tween 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%
0.67**** 0.77****
0.15** Goal
0.42**** Cooperation
Achievement
Trust
n.s.
0.17****
0,15**** 0.24****
0.36****
0.80****
Shared Values 0.23**** Communication

0.72**** 0.69****
-0.19* 0.33****
Openness
-0.28**** n.s.
Proactive Goal
-0.27**** Improvement Exceedance
-0.36****
0.32****
R2: 35.1%

Opportunism 0.65**** Standardized regression weight with significance level

n.s. 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 hy-
potheses 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, respec-
tively. Model results indicate strong support for both hypotheses, demon-
strating that cooperation directly influences both of the performance di-
mensions in logistics outsourcing arrangements. For the effect of
cooperation on goal achievement the path coefficient is 0.77, for coopera-
tion on goal exceedance it is 0.33. Both paths are positive and significant
at the 0.1%-level, suggesting very strong relationships. Cooperation there-
fore can be viewed as a cornerstone of successful logistics outsourcing ar-
rangements.
Two further hypotheses, H9a and H9b, imply that proactive improvement
through the LSP exerts similar, positive influences on both goal achieve-
ment 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. Ap-
parently, proactive improvement of the LSP has a particularly strong im-
pact on exceeding the goals ex-ante agreed upon, while it has no signifi-
cant influence on the dimension of goal achievement. Therefore, proactive
improvement is a facilitator of exceptional performance while it seems un-
necessary 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 out-
sourcing performance outcomes both directly and indirectly. Therefore, to-
tal 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 communi-
cation 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 sig-
nificant 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 trust-
building 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 communica-
tion, the effects of the antecedents that form the block of behav-
ioral/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 im-
portant building block of relationship models (WILSON 1995, p. 337). As
hypothesized in H10 and H11, trust is supposed to have a direct positive ef-
fect on both relationship commitment and cooperation. The path coeffi-
cient 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 ef-
forts lead to results that could not be achieved if both parties worked solely
in their own interest.
Analyzing the relationship between relationship commitment and coop-
eration, support is found for hypothesis H13, which proposed a direct posi-
tive effect induced by relationship commitment. The path coefficient is
0.22 at the 0.1%-level of significance. Apparently, parties that are commit-
ted 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 hy-
potheses 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 cus-
tomer believes the LSP to engage in opportunistic behavior, this perception
will lead to decreased levels of trust with all following negative conse-
quences, such as lower levels of relationship commitment and cooperation.
The hypothesis on the direct negative effect of opportunism on coopera-
tion 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 perform-
ance model is the factor shared values. Altogether, six hypotheses were
formulated on its direct effects. All of them find support in the model. Hy-
pothesis H20 proposed a positive effect of shared values on trust. The stan-
dardized parameter value for this linkage is 0.17 at the 0.1%-level of sig-
nificance, suggesting that the existence of shared values will contribute to
building trust in the LSP. A similar argumentation holds for the relation-
ship 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 pa-
rameter value of 0.15 at a 0.1%-level of significance. Apparently, higher
levels of shared values and the accompanying understanding of each oth-
ers motivations, needs, and expectations, leads customers and LSPs to
work more closely together to achieve mutual goals. Similarly, higher lev-
els 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 sig-
nificance. 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 relation-
ship. 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 diffi-
cult 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 open-
ness on communication. The standardized parameter value of 0.69 signifi-
cant at the 0.1%-level provides strong support for this hypothesis. Open-
ness 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. Hy-

pothesis 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 Rejected Not Tested

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. 8
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. 8
H6 Communication has a positive effect on cooperation. 9
H7 Communication has a positive effect on proactive improvement. 8
H8 Communication has a negative effect on opportunism. 9
H 9a Proactive improvement of the LSP has a positive effect on goal achievement. 8
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. 8
H 13 Relationship commitment has a positive effect on cooperation. 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 . 8
H 18 Opportunistic behavior has a negative effect on trust . 9
H 19 Opportunistic behavior has a negative effect on cooperation. 8
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 indi-
rect 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 inter-
esting finding certainly is that except for the factor opportunism, which has
a negative total effect on every other variable in the model, all other ante-
cedents have positive total effects on both goal achievement and goal ex-
ceedance. This again is strong support for the importance of all 8 antece-
dents for logistics outsourcing performance.

Table 7-4. Total effects in the logistics outsourcing performance model

Factor (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
(1) Shared Values -
(2) Openness 0.72 -
(3) Communication 0.73 0.69 -
(4) Opportunism -0.57 -0.19 -0.28 -
(5) Trust 0.75 0.57 0.31 -0.27 -
(6) Commitment 0.65 0.39 0.21 -0.18 0.67 -
(7) Cooperation 0.81 0.50 0.42 -0.20 0.57 0.22 -
(8) Proactive Improvement 0.52 0.28 0.16 -0.16 0.46 0.18 0.80 -
(9) Goal Exceedance 0.44 0.26 0.19 -0.12 0.34 0.13 0.59 0.32 -
(10) Goal Achievement 0.65 0.40 0.34 -0.16 0.46 0.18 0.80 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 impor-
tance 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 lo-
gistics outsourcing performance has measurable effects on both logistics
and firm performance. To analyze this question, chapter 7.2.1 will investi-
gate the effects of logistics outsourcing performance on logistics perform-
ance 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 achieve-

ment 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 Fig-
ure 7-4.

Goal H29 Level of

Achievement Logistics Services

H30 H31

Goal H32 Level of

Exceedance Logistics Costs

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

Upon testing, the model showed very satisfactory adaptation measures.

The adjusted F, at 2.540 is well within the range generally deemed ac-
ceptable. 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 ac-
cepted without any further modification.

Table 7-5. Adaptation measures of the logistics performance model

Adaptation measures R
Logistics Logistics
Model F/df TLI GFI AGFI CFI RMSEA
Services Costs
Logistics Performance 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 impor-
tant LSP is analyzed, the two dimensions of logistics performance investi-
gate 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 0.18**** Level of

Achievement Logistics Services

0.21**** 0.20****

R2: 7.3%

Goal 0.10* Level of

Exceedance Logistics Costs

0.65**** Standardized regression weight with significance level

n.s. 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 ex-
erts 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, re-
calling 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 customers overall level of logistics services.
250 7 Structural models

The next hypothesis H30 implies that goal achievement has a direct posi-
tive 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 sig-
nificance 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 out-
sourcing 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 dimen-
sions of logistics performance, making this outsourcing relationship an
important strategic issue.
After having pointed out the importance of goal achievement, goal ex-
ceedance as the second dimension is analyzed. Consistent with theory, hy-
pothesis H32 finds support, proposing that goal exceedance has a positive
effect on the level of logistics costs. The path coefficient of 0.10 is signifi-
cant 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 out-
sourcing 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 pri-
marily driven by the proactive improvement of the LSP, is further demon-
strated 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. Appar-
ently, 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 customers overall logistics service quality.
 7.2 Effects of logistics outsourcing performance 251

Table 7-6. Hypotheses for the logistics performance model

Hypotheses Supported Rejected Not Tested

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

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 lo-
gistics 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 out-
sourcing 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 par-
tial model, the effect of logistics outsourcing performance over all rela-
tionships 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 out-
sourcing for the overall logistics performance, this chapter will analyze the
connection between the latter and the individual firm performance. Fol-
lowing the argumentation from chapter 4.4.2.2, both dimensions of logis-
tics performance are proposed to have direct effects on all three dimen-
sions of firm performance.

7.2.2.1 Presentation of the model

Following the hypotheses developed earlier, both the level of logistics ser-
vices and the level of logistics costs are supposed to exert positive effects
on the three dimensions of firm performance, adaptiveness, market per-
formance 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 dis-
played in Figure 7-6.
252 7 Structural models

Adaptiveness

H33
Level of
H36 H39
Logistics Services

H34
Market
Performance
H37

Level of
H40
Logistics Costs
H38 H35

Financial
Performance

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

As Table 7-7 indicates, the model shows satisfactory adaptation meas-

ures 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 con-
cern. The adjusted F, 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 F
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 R
Market Financial
Model F/df TLI GFI AGFI CFI RMSEA Adaptiveness
Performance Performance

Firm 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 nota-
bly the higher F/df in this study, might surprise. However, this is largely
due to a modification in the model discussed already above. Other than as-
sumed 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 F in this model. However, since
no theoretical evidence would suggest a correlation between the two fac-
tors, 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 in-
sights into the effect of logistics performance on firm performance. In to-
tal, 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 ser-
vices exerts on the adaptiveness of the customer. This hypothesis is sup-
ported by the empirical findings with a standardized parameter value of
0.33 at the 0.1%-level of significance. This suggests that through improv-
ing their logistics service capabilities, firms indeed enhance flexibility to
react to changes in the market and in particular sharpen their ability to an-
swer 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 perform-
ance. This relationship was found to be particularly strong with a path co-
efficient of 0.55 again at the 0.1%-level of significance. Apparently, mar-
ket 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 logis-
tics 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
0.14*** 0.24****
Logistics Services

R2: 50.4%
0.55****
Market
Performance
0.21****

Level of
Logistics Costs 0.43****
n.s.
0.15**** R2: 22.3%

Financial
Performance
0.65**** Standardized regression weight with significance level
n.s. 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 ma-
neuverability 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 firms 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, highlight-
ing 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 per-
formance 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 signifi-
cance. Evidently, the supposed direct mathematical relationship between
lower logistics costs and lower total costs exists. Since financial perform-
ance 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 particu-
larly 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 per-
formance 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 perform-
ance.
Finally, hypothesis H40 proposed a positive effect of market perform-
ance on financial performance. This hypothesis is supported by the em-
pirical findings. The standardized parameter value is very strong with 0.43
at the 0.1%-level of significance. Evidently, firms with higher market per-
formance indeed enjoy higher customer loyalty, leading to decreased cus-
tomer 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 posi-
tively 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 Rejected Not Tested

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. 8
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

Only analyzing the direct effects of logistics performance on firm per-

formance 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 logis-
tics 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 firms lo-
gistics strategy, not only cost reductions through logistics should be aimed
at. Rather, similar attention should be given to increasing the level of logis-
tics services, since via their effect on both adaptiveness and market per-
formance 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 ar-
rangement, which as argued above, today are most often engaged with the
clear motivation of the customer to reduce costs while the potential to in-
crease service levels only plays a minor role if at all.

Table 7-9. Total effects in the firm performance model

Level of Logistics Level of Logistics
Factor Services Costs
Adaptiveness Market Performance

Adaptiveness 0.33 0.14 - -

Market Performance 0.63 0.24 0.24 -
Financial Performance 0.23 0.26 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 re-
flected 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 perform-
ance on firm performance in detail, the question remains whether the cor-
responding three models are moderated by the contingency variables
which were introduced in chapter 4.5.2. The next chapters will present ex-
plorative empirical findings and in the following discuss external and in-
ternal 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 ex-

change relationships in the logistics context, the analyses of moderating ef-
fects in the model of logistics outsourcing performance will focus on the
performance relevant causal linkages links that only affect the antece-
dents of outsourcing performance among themselves will not be tested.
Since the construct of functional conflict as one of three performance rele-
vant factors had to be eliminated from the model due to missing discrimi-
nant validity, only the paths from cooperation and proactive improvement
on goal achievement and goal exceedance respectively remain to be ana-
lyzed. Chapter 7.3.1.1 will discuss the moderating effects of external con-
tingency 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 sig-
nificant differences in the F-value can be observed. As Table 7-10 indi-
cates, the differences in the F-values of environmental complexity, envi-
ronmental 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. How-
ever, 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, be-

cause this level still leaves the possibility of finding a few individual mod-
erating effects on some paths.

Table 7-10. Moderating effects of external contingency variables on the logistics

outsourcing performance model
Logistics Outsourcing Performance

Moderating Environmental Environmental Importance of Industry of the

Uncertainty
Variable Complexity Dynamics Logistics Shipper

'F 13.64 10.20 16.55 6.85 See Text

P 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 analy-
ses presented in Table 7-11 find that only very few moderations exist.
The relationship between cooperation and goal achievement is moder-
ated by environmental complexity, environmental dynamics, and uncer-
tainty. 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 sug-
gests 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 environ-
ments, which can be e.g. the consequence of a demanding customer struc-
ture, logistically complex manufacturing processes, or increased competi-
tive 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 un-
fold its full potential in creating solutions and fostering the achievement of
goals that with adversarial or non-cooperative behavior would not be at-
tainable.
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 Environmental Environmental Importance of Industry of the
Uncertainty
Moderated Path Complexity Dynamics Logistics Shipper

Cooperation on Low 0.79**** 0.79**** 0.82**** - See Text

Goal Achievment High 0.73**** 0.71**** 0.71**** - See Text
Cooperation on Low - - - 0.21*** See Text
Goal Exceedance High - - - 0.40**** See Text
Proactive Improvement Low - - - - See Text
on Goal Achievement High - - - - See Text
Proactive Improvement Low 0.41**** - - - -
on Goal Exceedance 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 improve-
ment and goal exceedance exists. For low complexity, the standardized pa-
rameter 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 en-
ergy 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 im-
provement 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 per-
formance is stronger in environments characterized by lower levels of
complexity and dynamics. This could mean that the more complex and dy-
namic the industry environment is, the closer will the cooperation be and
the more important will the LSPs proactive improvement be.
The empirical findings on the moderating effect of the customers indus-
try 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 Health-
care sector limited sample size did not allow calculations, no significant
effects could be found in Consumer Goods, Manufacturing Systems Con-
struction, Chemicals and Plastics, and Retail industries. Only the Automo-
tive 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 sug-
gests the following: in a very complex and dynamic environment like the
Automotive industry, the effect of cooperation on outsourcing perform-
ance 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 in-
dustries further support these findings. The same paths as in the Automo-
tive industry are moderated plus the relationship between proactive im-
provement 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 dy-
namic indeed. The path from cooperation on goal exceedance is also mod-
erated. 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 bet-
ter solutions of problems, reduces frictions between the partners, and gen-
erally enhances the collaboration. As this conflicts with the findings pre-
sented above, further research must establish the reasons for these
diverging effects which will not be examined in this study due to its ex-
ploratory 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 in-
dustries 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 indus-
tries. This implies that here, proactive improvement is so important that it
has a substantial positive effect on the goal achievement. This finding, to-
gether 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, em-
pirical 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 re-
search 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 moderat-
ing variables significant differences in the F-value can be observed. As
Table 7-12 indicates, only for the variables asset specificity and the degree
of centralization, the difference in the F-value is significant at levels be-
low 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 Process Size of the Degree of

Products Asset Specificity Frequency
Variable Orientation Shipper Centralization

'F 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 de-
tailed 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 im-
provement 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 im-
provement 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 solu-
tions. This may cause a deeper engagement by the service provider com-
pared 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 moderat-
ing effects is the degree of logistics centralization of the LSPs 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 sig-
nificance for high levels of centralization. This suggests that the more cen-
tralized 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 Asset Process Size of the Degree of
Products Frequency
Moderated Path Specificity Orientation Shipper Centralization

Cooperation on Low - - - - - -
Goal Achievment High - - - - - -

Cooperation on Low - - - - 0.20*** -

Goal Exceedance High - - - - 0.44**** -
Proactive Improvement Low - -0.04**** - - 0.11** -
on Goal Achievement High - 0.13**** - - (n.s.) -

Proactive Improvement Low - 0.23**** - - 0.43**** -

on Goal Exceedance 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 sig-
nificant 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, signifi-
cant at the 1%-level, for high levels. Evidently, the effect of proactive im-
provement 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. Conse-
quently, this additional knowledge will facilitate the identification of im-
provement potentials.
While it must be recalled that from the many proposed moderating ef-
fects only very few have shown to be significant, from the analyses it can
be learned that the effect of proactive improvement on outsourcing per-
formance 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 contin-
gency 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 po-
tential moderating effects of both external and internal contingency vari-
ables 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 contin-
gency 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 environ-

mental dynamics or the importance of logistics to the customer have sig-
nificant 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 F-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 Environmental Environmental Importance of Industry of the

Uncertainty
Variable Complexity Dynamics Logistics Shipper

'F 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 de-

tail, a significant effect is found on the relationship between goal achieve-
ment and the level of logistics costs. While it is not significant for high lev-
els 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 com-
plex 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 lo-
gistics 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 Environmental Environmental Importance of Industry of the
Uncertainty
Moderated Path Complexity Dynamics Logistics Shipper

Goal Achievement on Level of Low - - - - See Text

Logistics Services
High - - - - See Text

Goal Achievement on Level of Low - - 0.34**** - -

Logistics Costs
High - - (n.s.) - -

Goal Exceedance on Level of Low - - - - -

Logistics Services
High - - - - -

Goal Exceedance on Level of Low - - (n.s.) - -

Logistics Costs
High - - 0.20*** - -
(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 uncer-
tainty, 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 uncer-
tain environments the effect of goal exceedance on the level of logistics
costs is higher than in more stable situations. This conflicts with the find-
ings presented above on the link of goal achievement on the level of logis-
tics 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 ad-
vanced 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 ex-
ternal 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 signifi-
cant. 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 con-
cluded that the model, even more so than the logistics outsourcing per-
formance 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 F-
value 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 out-
sourcing performance and logistics performance. The model thus has again
demonstrated its robustness against the moderation of contingency vari-
ables. The results are presented in Table 7-16. Still, further research is re-
quired 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 Process Size of the Degree of

Products Asset Specificity Frequency
Variable Orientation Shipper Centralization

'F 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 di-
mensions of logistics performance and those of firm performance are mod-
erated 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 in-
ternal 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 F-values and therewith indicate the exis-
tence of moderating effects. Solely the importance of logistics to the cus-
tomer does not seem to have a moderating influence.

Table 7-17. Moderating effects of external contingency variables on the firm per-
formance model
Firm Performance

Moderating Environmental Environmental Importance of Industry of the

Uncertainty
Variable Complexity Dynamics Logistics Shipper

'F 19.30 29.34 16.73 7.61 See Text

P 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 moder-
ated. These effects will be discussed in the following. For the variable en-
vironmental 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 sig-
nificant 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 envi-
ronmental 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 envi-
ronmental 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 environ-
ment.
The effect of the level of logistics costs on adaptiveness is also moder-
ated 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 lo-
gistics 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 op-
portunities. 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 Environmental Environmental Importance of Industry of the
Uncertainty
Moderated Path Complexity Dynamics Logistics Shipper

Level of Logistics Services on Low - - - - -

Adaptiveness
High - - - - -

Level of Logistics Services on Low - - 0.65**** - -

Market Performance
High - - 0.50**** - -

Level of Logistics Services on Low - - - - -

Financial Performance
High - - - - -

Level of Logistics Costs on Low - 0.30**** - - -

Adaptiveness
High - (n.s.) - - -

Level of Logistics Costs on Low - - - - -

Market Performance
High - - - - -

Level of Logistics Costs on Low - - - - -

Financial Performance
High - - - - -

Adaptiveness on Market Low 0.36**** 0.37**** - - See Text

Performance
High (n.s.) 0.12** - - See Text

Market Performance on Low - - - - -

Financial Performance
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 uncer-
tainty. 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, un-
der 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
LSPs 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 pre-
dictable 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 Con-
struction 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 influ-
ence, while no moderating effects could be detected for the remaining four
variables. The differences in the F-values and the corresponding signifi-
cance levels are presented in Table 7-19.
Especially the variable products with a total of four shows a large num-
ber of moderating effects (see Table 7-20). The effect of the level of logis-
tics 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 reduc-
tions for the adaptiveness of the firm.

Table 7-19. Moderating effects of internal contingency variables on the firm per-
formance model
Firm Performance

Moderating Process Size of the Degree of

Products Asset Specificity Frequency
Variable Orientation Shipper Centralization

'F 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

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, sig-
nificant 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 signifi-
cantly 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 commodi-
ties. Here, a slight absolute cost reduction may have comparably large ef-
fects 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 com-
moditized product range can reap the benefits of adaptiveness for an in-
creased 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 characteris-
tics.
Finally, the relationship between market performance and financial per-
formance is moderated by the product range. Under low product range
complexity, the effect is 0.30, under high complexity it is 0.51, both sig-
nificant 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, translat-
ing 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 pro-
duce 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 Asset Process Size of the Degree of
Products Frequency
Moderated Path Specificity Orientation Shipper Centralization

Level of Logistics Services on Low - 0.47**** - - - -

Adaptiveness
High - 0.20*** - - - -

Level of Logistics Services on Low - - - - - -

Market Performance
High - - - - - -

Level of Logistics Services on Low - - - - - -

Financial Performance
High - - - - - -

Level of Logistics Costs on Low 0.29**** - - - - -

Adaptiveness
High (n.s.) - - - - -

Level of Logistics Costs on Low - - - - - -

Market Performance
High - - - - - -

Level of Logistics Costs on Low 0.25**** - - - - -

Financial Performance
High (n.s.) - - - - -

Adaptiveness on Market Low 0.34**** - - - - -

Performance
High 0.17*** - - - - -

Market Performance on Low 0.30**** - - - - -

Financial Performance
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 addi-
tional adaptiveness as logistics processes tend to be more complex and dif-
ficult 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 con-
tingency variables. This is substantially less evidence for the relevance of
272 7 Structural models

contingency variables than originally expected. It must therefore be con-

cluded 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 po-
tential 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 re-
search 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 identi-

fied. While it had been established in previous research that logistics out-
sourcing performance is a driver of logistics performance, which in turn is
a facilitator of firm performance, the main antecedents of logistics out-
sourcing performance were not known. In fact, in the measurement of lo-
gistics 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 out-
sourcing 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 provid-
ers 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 dis-
cussed 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 ap-
proach.
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 re-
sponse rate of 17.0%. After correcting the database for missing values, a
total of 549 usable cases remained. The data was analyzed using covari-
ance 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 ad-
dressed and the key results presented.

R1 : Which are the main influencing factors of an outsourcing rela-

tionship 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 commit-
ment-trust theories, ten relationship variables were identified that suppos-
edly 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 di-
mensions goal achievement and goal exceedance. Goal achievement
measures whether the expectations of the customer with respect to the ser-
vice 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 proac-
tive, 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 con-
flict. The remaining seven variables were hypothesized to have indirect ef-
fects only, all mediated through one or more of the three variables with di-
rect effects. These findings directly lead to the second research question
which will be addressed in the following:
 8.1 Main results 275

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?
After the operationalization of the constructs on the basis of the empiri-
cal analysis, it was found that the construct of logistics outsourcing per-
formance indeed is bi-dimensional, consisting of both goal achievement
and goal exceedance. For the ten relationship variables, discriminant valid-
ity 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 out-
sourcing 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 indi-
rectly influencing the two dimensions of logistics outsourcing perform-
ance, the final model displays a distinct three level structure. Five antece-
dents, 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 lo-
gistics outsourcing performance. The factor communication takes on a me-
diating position between the behavioral/attitudinal variables and the action
variables.
A large number of the hypotheses developed for the model found sup-
port 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 di-
rect 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 improve-
ment of the LSP must not be neglected if an exceedance of the goals is as-
pired by the customer.
276 8 Summary and results

The indirect effects of the behavioral/attitudinal variables on outsourc-

ing performance were also found to be important. The shared values are a
central variable, exerting direct positive effects on cooperation, communi-
cation, 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 relation-
ship 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 ef-
fect 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 influ-
ences the proactive improvement. This indicates that communicating and
exchanging information alone do not lead to increase the proactive im-
provement 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 per-
formance 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 vari-
ables that remained in the main model have a positive total effect on logis-
tics 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 de-
veloped construct of logistics outsourcing performance has on the logistics
performance and if the hypothesized link between logistics- and firm per-
formance 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 ser-
vices 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 logis-
tics 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 ser-
vice level and reducing the overall logistics costs. Goal exceedance is of
utmost importance to firms that want to increase the level of logistics ser-
vices 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 par-
tial model focusing only on the most important LSP of the customer, the
effect of logistics outsourcing performance over all relationships can le-
gitimately 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 per-
formance. 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 perform-
ance.
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 ef-
fect on the financial performance.
These results allow new insights into the importance of logistics per-
formance. 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 lo-
gistics costs due to the indirect effects via the firms 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 pro-
vides is virtually as strong as the one of cost reductions. Consequently, ne-
glecting it would mean giving away a vast potential.
As discussed already in the introduction to this study, the role of the in-
fluence of contingency variables on the models presented above remains
unclear. Therefore, research question four was developed:

R4 : Which contingency variables for the logistics outsourcing con-

text can be identified and which moderating effects can be ob-
served 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 environ-
mental complexity, environmental dynamics, uncertainty, and the cus-
tomers industry. Internal variables on the other hand can be influenced di-
rectly by the customer. They include the customers products, the size of
the organization, the degree of centralization of logistics decisions, the as-
set 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 meas-
uring the intensity of competition in the customers industry and the other
one the importance of logistics to the customer of the LSPs customer.
All three performance models, for logistics outsourcing-, logistics-, and
firm performance, were tested for moderating effects of the eleven contin-
gency 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 moder-
ating effects in the different models were tested. Uncertainty and the prod-
ucts show 4 moderations each, while environmental complexity, asset
specificity, and the centralization of logistics decisions show 3. The re-
maining 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 to-
tal 150 or 85.2%, was not subject of this study. However, it will be an in-
teresting 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 recommenda-
tions for the practice, the key findings have several managerial implica-
tions 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 as-
pects are important differentiators in a world that is characterized by in-
creasing competition. Logistics outsourcing should therefore be a relevant
option for the firms as its implementation and extension promises signifi-
cant 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 ex-
ceed 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, communica-
tion, 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 ar-
rangement 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 logis-
tics service levels, the customer must choose an LSP that continuously im-
proves 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 out-
sourcing. To a certain degree, this motivation and its according actions are
justified, as it directly leads to a higher financial performance of the cus-
tomer. At the same time, it also increases the firms 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 rea-
son for outsourcing.
The potential of logistics service level increases, however, is substan-
tially underestimated as the current, strongly cost-oriented logistics out-
sourcing motivation among firms has shown. While it is more difficult to
measure progress in this field, its potential benefits are vast. It does not di-
rectly influence the bottom line of the firm, but it increases the firms
flexibility and the market performance substantially stronger than cost re-
ductions. 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 LSPs
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 in-
deed, some contingency variables moderate individual aspects of the per-
formance models that means, however, that the remaining parts of the re-
spective 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 con-
texts and individual situations.

8.3 Recommendations for further research

The findings of this study do not only offer insights into managerial impli-
cations, 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 cor-
relation 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. How-
ever, 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 impor-
tant LSP was examined. It remains to be shown how much of the remain-
ing 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 firms logistics performance.
In this study, a total of seven significant effects of the dimensions of lo-
gistics 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 ef-
fects. The reasons for these deviations represent very fertile ground for fur-
ther 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 de-
mands 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 op-
erationalization or conceptualization. Furthermore, it may be that better re-
sults concerning moderating effects could be obtained if not only the con-
text 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 ex-
tended.
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 re-
spondents only. In a world that is characterized by an ever increasing glob-
alization of the markets, driven by better communication methods, less re-
strictions for foreign trade, and a general internationalization of firms both
with respect to purchasing and selling of goods and services, it must be ex-
amined if these results also hold in other regional and cultural environ-
ments. Studies in different countries would enable an insight into regional
differences that could be a valuable insight for firms dealing with the real-
ity of internationalization.
Appendix: Questionnaire

WHU-BVL-Studie:
"Erfolg durch Logistik-Outsourcing"

Prof. Dr. Jrgen Weber

Dipl.-Kfm. David L. Cahill
Dipl.-Kfm. Jan Deepen

Khne-Zentrum fr Logistikmanagement
WHU - Otto-Beisheim-Hochschule
Burgplatz 2, 56179 Vallendar

Als Dankeschn fr Ihre Teilnahme erhalten Sie:

x einen individuellen Benchmarking-Bericht (Vergleich Ihres Unter-

nehmens mit den Werten Ihrer Branche, der Ihnen nach Abschluss
der Studie zugesandt wird.)
sowie zustzlich
x eine Ausgabe des Buches "E-Commerce in der Logistik: Quanten-
sprung oder business as usual?" (151 S.) von Prof. Dr. Jrgen 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 Frage-
bogen ausfllen:

x Alle Daten werden anonym ausgewertet und streng vertraulich be-

handelt

x Bitte fllen Sie alle Fragen so gut wie mglich aus, auch wenn
manche Fragen hnlich erscheinen. Dies ist aus methodischen Grn-
den ntig. Wenn Sie die genaue Antwort nicht kennen, bitten wir Sie
bewut um Ihre subjektive Einschtzung.

x Bitte beziehen Sie alle Fragen zu Ihrem Unternehmen immer auf

Ihre Geschftseinheit bzw. auf den Teilbereich des Unternehmens,
fr dessen Logistik Sie (mit)verantwortlich sind.

x Falls Sie verschiedene Logistikdienstleister (LDL) benutzen, bezie-

hen Sie alle Fragen auf den aus Ihrer Sicht fr Ihr Unternehmen
bzw. Ihre Geschftseinheit wichtigsten LDL.

x Der in dieser Studie verwendete Begriff Logistik-Outsourcing be-

zieht sich auf die dauerhafte Fremdvergabe logistischer Dienst-
leistungen an einen Logistikdienstleister.

x Sie knnen die Beantwortung des Fragebogens nach jeder ab-

geschickten Seite unterbrechen und dann durch den in der Email
enthaltenen Link fortsetzen. Die eingegebenen Daten werden auto-
matisch nach jeder Seite gespeichert. Das Dankeschn gibt es fr
alle vollstndig ausgefllten Fragebgen.

x Bitte verwenden Sie zur Navigation innerhalb des Fragebogens aus-

schlielich die "zurck" und "weiter" Felder, da es sonst zu tech-
nischen Problemen kommen kann.

Fr Fragen stehen Ihnen Dipl.-Kfm. David Cahill oder Dipl.-Kfm.

Jan Deepen gerne unter 0261-6509-471 oder cahill@whu.edu zur
Verfgung!
 Appendix: Questionnaire 285

A. Allgemeine Fragen zum Logistik-Outsourcing

Ein wesentlicher Grund fr uns, einen Teil unserer Logistik Trifft gar Trifft voll
outzusourcen, war/waren: nicht zu zu
Die Senkung der Logistikkosten. { { { { { { {
Die Variabilisierung unserer Fixkosten. { { { { { { {
Der Spitzenausgleich bei Kapazittsschwankungen. { { { { { { {
Die niedrigere Kapitalbindung. { { { { { { {
Der Geschwindigkeitsgewinn fr unsere Logistik. { { { { { { {
Die flexibleren Ablufe und krzeren Reaktionszeiten. { { { { { { {
Die geringere Schadens- und Fehlerquote. { { { { { { {
Die verbesserte Lieferfhigkeit. { { { { { { {
Die Tatsache, dass dieser LDL ein deutlich hheres { { { { { { {
Logistik--Know-how besitzt.
Die Tatsache, dass Logistik fr uns zu den eher { { { { { { {
unwichtigen Prozessen gehrt.
Der Engpass bei unseren Managementkapazitten. { { { { { { {

Wichtig: Bitte beziehen Sie sich ab hier immer auf den fr Sie wich-
tigsten Logistikdienstleister!

Grad der Durchfhrung Leistungs-schwerpunkt

Welche Logistikleistungen fhrt der fr Sie Gar nicht Voll- Be-schaf- Pro- Dis-
wichtigste Logistikdienstleister (LDL) mit stndig fung duk- tribu-
tion tion
welchem Schwerpunkt aus?
Planung und Steuerung der Transporte { { { { { { { { { {
Transportdurchfhrung { { { { { { { { { {
Internationales Freight Forwarding { { { { { { { { { {
(Frachtversand)
Zollabfertigung { { { { { { { { { {
Cross-Docking { { { { { { { { { {
Lagerhaltung { { { { { { { { { {
Lager-Management { { { { { { { { { {
Verpacken, Kommissionieren { { { { { { { { { {
Montagettigkeiten { { { { { { { { { {
Retourenabwicklung { { { { { { { { { {
IT-Systeme der Logistik { { { { { { { { { {
Logistik-Koordination (Lead Logistics { { { { { { { { { {
Management)
Beratungsleistungen { { { { { { { { { {
286 Appendix: Questionnaire

B. Zusammenarbeit mit Ihrem Logistikdienstleister

Inwieweit treffen die folgenden Aussagen bzgl. Ihrer Trifft gar Trifft voll
Zufriedenheit ber die Zusammenarbeit mit Ihrem nicht zu zu
Logistikdienstleister (LDL) zu?
Die Ziele und Erwartungen, die wir im Vorfeld der { { { { { { {
Zusammenarbeit festgelegt hatten, werden durch
diesen LDL vollstndig erfllt.
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 { { { { { { {
Qualitt.
Dieser LDL erbringt seine Leistungen in der geforderten { { { { { { {
Zeit.
Unsere Logistikkosten sind durch das Outsourcing in { { { { { { {
dem von uns vorhergesagten Ausma 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 { { { { { { {
Prozessablufen etc. in angemessenem Umfang an
uns weiter.
Dieser LDL bereichert sich auf unsere Kosten. { { { { { { {
Wir fhlen uns durch diesen LDL fair behandelt. { { { { { { {
Wir profitieren in gleichem Mae vom Outsourcing wie { { { { { { {
dieser LDL.

Inwieweit stimmen Sie den folgenden Aussagen hinsichtlich Trifft gar Trifft voll
Ihres Verhltnisses zu Ihrem Logistikdienstleister zu? nicht zu zu
Dieser LDL hlt immer die Zusagen, die er uns macht. { { { { { { {
Dieser LDL ist bei auftretenden Problemen immer offen { { { { { { {
und ehrlich zu uns.
Dieser LDL ist sehr vertrauenswrdig. { { { { { { {
Dieser LDL ist stark daran interessiert, dass wir { { { { { { {
erfolgreich sind.
Auch Dinge, die wir nicht oder nur unter groem { { { { { { {
Aufwand kontrollieren knnen, erledigt dieser LDL
Bei wichtigen Entscheidungen bercksichtigt der LDL { { { { { { {
auch unsere Interessen.
Wir verteidigen diesen LDL, wenn er durch Mitglieder { { { { { { {
unseres Unternehmens oder durch Personen von
Wir wrden es persnlich sehr bedauern, wenn wir die { { { { { { {
Geschftsbeziehung mit diesem LDL aufgeben
Wir fhlen uns persnlich angegriffen, wenn dieser LDL { { { { { { {
durch Mitglieder unseres Unternehmens oder durch
Wir haben den festen Willen, die Beziehung so lange { { { { { { {
wie mglich aufrechtzuerhalten.
 Appendix: Questionnaire 287

Inwieweit stimmen Sie den folgenden Aussagen zur Trifft gar Trifft voll
Kooperation mit diesem Logistikdienstleister zu? nicht zu zu
Die Ziele unserer Zusammenarbeit wurden von uns und { { { { { { {
diesem LDL gemeinsam festgelegt.
Hinsichtlich der Art und Weise, zusammen zu arbeiten, { { { { { { {
Geschfte zu machen und Projekte durchzufhren,
Dieser LDL und wir ziehen in allen Belangen an einem { { { { { { {
Strang.
Wenn auf bergeordneter Ebene noch Probleme oder { { { { { { {
Unklarheiten hinsichtlich der Outsourcing-Kooperation
auftauchen, fllen wir notwendigen Entscheidungen
gemeinsam mit dem LDL.
Wenn im Rahmen unserer Geschftsbeziehung einer { { { { { { {
der Partner seine Macht ausspielt, geschieht das in
einer angemessenen Art und Weise.
In unserer Geschftsbeziehung respektieren sich beide { { { { { { {
Seiten vollkommen.
Auch abseits der vorher festgelegten Zustndigkeiten { { { { { { {
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 Trifft gar Trifft voll
relevanten Teilschritte auf dem Weg zum Outsourcing? nicht zu zu
Einige Fhrungskrfte in unserem Unternehmen sind { { { { { { {
gegen die Outsourcing-Entscheidung.
Auf Seiten der operativen Mitarbeiter in unserem { { { { { { {
Unternehmen gibt es viele Widerstnde gegen das
Outsourcing-Projekt.
Zwischen dem LDL und uns treten immer wieder { { { { { { {
Konflikte auf der Managementebene auf.
Auf der operativen Ebene kommt es hufig zu Problem { { { { { { {
oder Konflikten zwischen unseren Mitarbeitern und
denen des LDL.
Insgesamt ist die Beziehung zu diesem LDL sehr { { { { { { {
konfliktreich.
288 Appendix: Questionnaire

Inwieweit beschreiben die folgenden Aussagen die Trifft gar Trifft voll
laufende Zusammenarbeit mit diesem Logistikdienstleister? nicht zu zu
Wir besprechen mgliche Probleme und { { { { { { {
Verbesserungen regelmig mit den verantwortlichen
Mitarbeitern des LDL.
Der Austausch von Informationen zwischen unseren { { { { { { {
Mitarbeitern und denen dieses LDL verluft sehr gut.
Um unsere gesteckten Ziele zu erreichen, sind viele { { { { { { {
Treffen und Gesprche mit diesem LDL notwendig.
Wenn wir Informationen mit diesem LDL austauschen, { { { { { { {
sind diese immer relevant fr das Vorankommen des
Projektes bzw. unserer Geschftsbeziehung.
Der Austausch von Informationen zwischen diesem { { { { { { {
LDL und uns findet immer sofort statt, sobald sie zur
Verfgung stehen.
Auf die Informationen, die zwischen uns und diesem { { { { { { {
LDL ausgetauscht werden, knnen 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
lsen.

Inwieweit stimmen Sie den folgenden Aussagen zu dem Trifft gar Trifft voll
Verhltnis bzw. der Zusammenarbeit mit diesem nicht zu zu
Logistikdienstleister zu?
Der LDL war frhzeitig und umfassend eingebunden. { { { { { { {
Fr das mit diesem LDL durchgefhrte Outsourcing ist { { { { { { {
ein funktionsbergreifendes Projektteam verantwortlich,
in dem alle relevanten Funktionen unseres
Unternehmens vertreten sind.
Die auf beiden Seiten fr das Outsourcing { { { { { { {
verantwortlichen Mitarbeiter arbeiten sehr gut
zusammen.
Den alltglichen Austausch zwischen uns und diesem { { { { { { {
LDL kann man als informell bezeichnen.
Wenn Fragen oder Probleme auftreten, knnen wir { { { { { { {
diesen LDL spontan kontaktieren und sofort
gemeinsam beginnen, an Lsungen 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.
 Appendix: Questionnaire 289

Inwieweit stimmen Sie den folgenden Aussagen zu Ihrem Trifft gar Trifft voll
Verhltnis bzw. der Zusammenarbeit mit diesem nicht zu zu
Logistikdienstleister zu?
Dieser LDL und wir bewltigen 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" geprgt.
Probleme zwischen dem LDL und uns haben in der { { { { { { {
Vergangenheit die Produktivitt unserer Beziehung
stark beeintrchtigt.
Probleme zwischen dem LDL und uns werden als { { { { { { {
Chance verstanden, die beiden Partnern die
Mglichkeit zu Verbesserungen erffnen.

Inwieweit treffen die folgenden Aussagen auf die Trifft gar Trifft voll
persnlichen Beziehungen zwischen Mitarbeitern Ihres nicht zu zu
Unternehmens und denen des Logistikdienstleisters zu?
Wir pflegen viele persnliche Kontakte mit diesem { { { { { { {
LDL.
Die Zusammenarbeit mit diesem LDL klappt auch auf { { { { { { {
der persnlichen Ebene sehr gut.
Wenn wir die Beziehung zu diesem LDL beenden { { { { { { {
wrden, wrden ich oder einige meiner Mitarbeiter und
Kollegen einen guten Geschftsfreund verlieren.
Wir haben enge persnliche Beziehungen zu { { { { { { {
Mitarbeitern dieses LDL.

Inwieweit treffen folgende Aussagen auf Ihren Trifft gar Trifft voll
Hauptansprechpartner beim Logistikdienstleister zu? nicht zu zu
Diese Person ist sehr vertrauenswrdig. { { { { { { {
Diese Person macht keine falschen Versprechen. { { { { { { {
Diese Person besitzt groes Know-how und ist ein { { { { { { {
guter Manager.
Diese Person spricht die gleiche Sprache wie wir. { { { { { { {
Diese Person bercksichtigt auch unsere Interessen { { { { { { {
und Bedrfnisse.

Welche Position hat Ihr Hauptansprechpartner beim LDL?

{ Geschftsfhr- { Niederlassungs- { Bereichsleiter
er/Vorstand leiter
{ Vertriebsleiter { Key Account { Andere: __________________
Manager
290 Appendix: Questionnaire

Inwieweit treffen folgende Aussagen zur Verbesserung Trifft gar Trifft voll
Ihrer Logistiksysteme durch diesen Logistikdienstleister zu? nicht zu zu
Der LDL arbeitet intensiv daran, die Logistikprozesse { { { { { { {
fortlaufend zu optimieren.
Der LDL gibt uns laufend Anste zu Verbesserungen { { { { { { {
auch auerhalb seines direkten
Zustndigkeitsbereiches.
Bei vernderten Rahmenbedingungen modifiziert der { { { { { { {
LDL aus eigenem Antrieb die Logistiksysteme bzw. -
ablufe soweit sinnvoll und notwendig.
Der LDL spricht uns aus Eigeninitiative mit { { { { { { {
Verbesserungsvorschlgen an.
Dieser LDL ist generell sehr innovativ. { { { { { { {

Inwieweit treffen folgende generelle Aussagen zu diesem Trifft gar Trifft voll
Logistikdienstleister zu? nicht zu zu
Der LDL bietet insgesamt einen exzellenten Service. { { { { { { {
Die Leistungen des LDL sind hervorragend. { { { { { { {
Der LDL bietet eine sehr hohe Qualitt. { { { { { { {

Inwieweit treffen die folgenden Fragen bzgl. Ihrer Trifft gar Trifft voll
Zufriedenheit ber die Zusammenarbeit mit diesem nicht zu zu
Logistikdienstleister zu?
Die Ziele und Erwartungen, die wir im Vorfeld der { { { { { { {
Zusammenarbeit festgelegt hatten, wurden in
deutlichem Mae bertroffen.
Unsere Anforderungen an die Qualitt der Leistungen { { { { { { {
dieses LDL sind deutlich positiv bertroffen worden.
Unsere Logistikkosten sind auf Grund der { { { { { { {
Zusammenarbeit mit diesem LDL deutlich niedriger als
erwartet.
Die tatschlichen Kosten sind im Verhltnis zur { { { { { { {
erbrachten Gesamtleistung wesentlich besser, als wir
es vorher erwartet htten.

Inwieweit treffen folgende Aussagen zur bisherigen und Trifft gar Trifft voll
zuknftigen Zusammenarbeit mit diesem nicht zu zu
Logistikdienstleister zu?
Wir werden diesen LDL auch zuknftig weiter nutzen. { { { { { { {
Aus heutiger Sicht gehen wir davon aus, vorhandene { { { { { { {
Vertrge mit dem LDL bei deren Auslaufen zu
verlngern.
Wenn wir mit unserem heutigen Wissen nochmals vor { { { { { { {
der ursprnglichen Entscheidung ber die
Zusammenarbeit mit diesem LDL stnden, wrden wir
die Geschftsbeziehung erneut eingehen.
Wir werden die Leistungen, die wir von diesem LDL in { { { { { { {
Anspruch nehmen, bei Auslaufen des Vertrages hchst
wahrscheinlich nicht neu ausschreiben, sondern direkt
mit diesem LDL verhandeln.
 Appendix: Questionnaire 291

Inwieweit treffen folgende Aussagen zu Ihrer Einstellung Trifft gar Trifft voll
gegenber diesem Logistikdienstleister zu? nicht zu zu
Wir haben in der Organisation angeregt, diesen LDL fr { { { { { { {
zuknftige Projekte bevorzugt zu bercksichtigen.
Ich erwhne diesen LDL gegenber Kollegen hufig { { { { { { {
sehr positiv.
Ich empfehle diesen LDL auch nach auen hin hufig { { { { { { {
weiter.
Wir empfehlen diesen LDL hufig weiter. { { { { { { {

Inwieweit treffen folgende Aussagen bzgl. der Investitionen Trifft gar Trifft voll
in die Geschftsbeziehung zu? nicht zu zu
Wir haben viel Arbeitszeit (Mannjahre) investiert, die { { { { { { {
umsonst wre, wenn wir den LDL wechseln wrden.
Wir haben hohe Sachinvestitionen gettigt, die { { { { { { {
erheblich an Wert verlieren wrden, wenn wir den LDL
wechselten.
Wir haben hohe Investitionen in IT-Systeme gettigt, { { { { { { {
die erheblich an Wert verlieren wrden, 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 wre, wenn wir den LDL
wechseln wrden.

Bitte geben Sie den Grad der bereinsstimmung an, den Sehr niedrige Sehr hohe
berein- berein-
Sie in den folgenden Fragen zwischen Ihrem Unternehmen
stimmung stimmung
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 fr die beteiligten
Unternehmen im Auge haben.
Die Mitarbeiter beider Seiten haben die Ziele der { { { { { { {
Zusammenarbeit vollstndig 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 Verstndnis ber die Art der { { { { { { {
Zusammenarbeit ist bei beiden Seiten sehr hnlich und
kompatibel.

Inwieweit treffen folgende generelle Aussagen zu diesem Trifft gar Trifft voll
Logistikdienstleister zu? nicht zu zu
Die Preise dieses LDL sind im Vergleich zu anderen { { { { { { {
LDL sehr gnstig.
Das Preis-Leistungs-Verhltnis bei diesem LDL ist sehr { { { { { { {
gut.
Die Preise dieses LDL sind im Vergleich zur { { { { { { {
Eigenerstellung sehr gnstig.
292 Appendix: Questionnaire

Inwieweit treffen folgende Aussagen zur bisherigen und Trifft gar Trifft voll
zuknftigen Zusammenarbeit mit diesem nicht zu zu
Logistikdienstleister zu?
In Zukunft wird dieser LDL einen greren Anteil an { { { { { { {
unserem Auftragsvolumen erhalten.
Beim Outsourcing anderer als der bisherigen { { { { { { {
Logistikleistungen werden wir diesen LDL bevorzugt
bercksichtigen.
Neue Leistungen werden wir zunchst diesem LDL { { { { { { {
anbieten, bevor wir sie ausschreiben.
In den nchsten Jahren werden wir strker auf diesen { { { { { { {
LDL zurckgreifen als bisher.

Inwieweit beschreiben die folgenden Aussagen die Trifft gar Trifft voll
laufende Zusammenarbeit mit diesem Logistikdienstleister? nicht zu zu
Dieser LDL verndert manchmal Fakten so, dass sich { { { { { { {
seine Interessen besser begrnden und vertreten
lassen.
Was der LDL uns gegenber als seine momentanen { { { { { { {
Aktivitten und Leistungen fr uns angibt, entspricht
immer der Realitt.
Um seine eigenen Ziele besser zu erreichen, verspricht { { { { { { {
uns der LDL manchmal Dinge, die er spter berhaupt
nicht einhlt.
Wenn der LDL mit uns ber seine Bedrfnisse spricht, { { { { { { {
bertreibt er manchmal, um seine Ziele besser zu
erreichen.
Dieser LDL wrde alles in seiner Macht stehende tun, { { { { { { {
um seine eigenen Ziele zu erreichen.
Dieser LDL hat das Gefhl, dass sich im Umgang mit { { { { { { {
uns Ehrlichkeit lohnt.

Inwieweit treffen folgende Aussagen zu Ihrer Einstellung Trifft gar Trifft voll
gegenber diesem Logistikdienstleister zu? nicht zu zu
Meiner Meinung nach haben wir zuviel investiert, um { { { { { { {
einen Abbruch der Geschftsbeziehung in Betracht zu
ziehen.
Auch wenn wir es wollten, wre es sehr schwierig, die { { { { { { {
Geschftsbeziehung zu diesem LDL aufzugeben.
Wir fhlen uns verpflichtet, die Geschftsbeziehung zu { { { { { { {
diesem LDL aufrechtzuerhalten.
Dieser LDL hat zuviel in die Beziehung investiert, als { { { { { { {
dass wir ihm einen Abbruch der Geschftsbeziehung
zumuten knnten.
 Appendix: Questionnaire 293

Inwieweit treffen folgende Aussagen zu Alternativen und Trifft gar Trifft voll
Ihrer Abhngigkeit von diesem Logistikdienstleister zu? nicht zu zu
Es gibt viele alternative LDL, die gleichwertige { { { { { { {
Leistungen bei gleichen Kosten anbieten knnen.
Es gibt viele alternative LDL, die gleichwertige { { { { { { {
Leistungen bei hheren Kosten anbieten knnen.
Es gibt viele alternative LDL, die gleichwertige { { { { { { {
Leistungen bei niedrigeren Kosten anbieten knnen.
Dieser LDL beeinflusst die Qualitt und Zuverlssigkeit { { { { { { {
unserer Logistikprozesse so sehr, dass sein Wegfall
uns sehr stark treffen wrde.
Bei einem Wechsel des LDL wrden insbesondere die { { { { { { {
neuen Verhandlungen sehr viel Aufwand verursachen.
Die Suche und Auswahl eines gleichwertigen LDL ist { { { { { { {
sehr aufwendig.
Wir knnen diesen LDL leicht ersetzen. { { { { { { {
Ein kurzfristiger Wechsel dieses LDL wre mit sehr viel { { { { { { {
Aufwand verbunden.

Inwieweit beschreiben die folgenden Aussagen die Trifft gar Trifft voll
spezifischen Investitionen, die fr den Logistikdienstleister nicht zu zu
in Ihrem Fall notwendig waren?
Um Logistikdienstleistungen fr uns zu erbringen, { { { { { { {
musste der LDL hohe Investitionen (Transportmittel,
Lagerhuser o..) ttigen, die er ausschlielich fr uns
verwenden kann.
Der LDL kann die fr uns entwickelten { { { { { { {
Geschftsablufe nicht ohne groe Vernderungen fr
einen anderen Kunden verwenden.
Damit der LDL Leistungen fr uns erbringen kann, { { { { { { {
musste er seine Mitarbeiter so schulen, dass sie dieses
Wissen bei einem anderen Kunden nicht verwenden
knnten.
Die Investitionen, die dieser LDL fr unsere Beziehung { { { { { { {
in Menschen, Prozesse oder Vermgenswerte ttigen
musste, werden sich sehr schnell amortisieren.

$FKWXQJ%LWWHEHDQWZRUWHQ6LHGLHIROJHQGHQ)UDJHQLQ%H]XJDXI
,KU8QWHUQHKPHQXQG,KUH.XQGHQ
294 Appendix: Questionnaire

C. Fragen zu Ihrer Unternehmenssituation

Inwieweit beschreiben die folgenden Aussagen die zur Zeit Trifft gar Trifft voll
in Ihren Unternehmen vorherrschende Unsicherheit? nicht zu zu
Es fllt uns schwer, das knftige Verhalten unserer { { { { { { {
Kunden przise vorherzusehen.
Es kommt oft vor, dass sich Kunden uns gegenber { { { { { { {
opportunistisch oder unfair verhalten.
In unserer Branche oder der unserer Kunden gibt es { { { { { { {
aufgrund der Wettbewerbssituation zur Zeit eine
sprbare Unsicherheit ber die Zukunft.
Es fllt uns schwer einzuschtzen, wie sich die { { { { { { {
Wnsche und Bedrfnisse unserer Kunden in Zukunft
entwickeln werden.

Inwieweit beschreiben die folgenden Aussagen die Trifft gar Trifft voll
momentane Situation Ihres Unternehmens? nicht zu zu
Unser Absatzvolumen ist von Kunde zu Kunde sehr { { { { { { {
unterschiedlich.
Das Bestellverhalten unser Kunden ist sehr { { { { { { {
uneinheitlich und schwankt je nach Auftrag oder Saison
betrchtlich.
Verglichen mit unseren Wettbewerbern mssen unsere { { { { { { {
Produkte an eine hohe Anzahl von Lieferpunkten
geliefert werden.
Verglichen mit unseren Wettbewerbern greifen wir auf { { { { { { {
eine sehr groe Anzahl von Logistikdienstleistern
zurck.
Der Verdrngungswettbewerb 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
fr uns ein groer Wettbewerbsvorteil.

Inwieweit beschreiben die folgenden Aussagen das Trifft gar Trifft voll
Produktprogramm Ihres Unternehmens? nicht zu zu
Wir fertigen ein sehr breites Produktprogramm mit { { { { { { {
einer Reihe von unterschiedlichen Produkten.
Unsere Produkte sind so wertvoll, dass sie besondere { { { { { { {
logistische Manahmen erfordern.
Unsere Produkte sind so unterschiedlich, dass fr sie { { { { { { {
sehr viele unterschiedliche Transport-, Umschlag-
und/oder Lagerprozesse konzipiert werden mssen.
Fr unsere Produkte gibt es echte Alternativen am { { { { { { {
Markt, so dass wir uns ber eine erstklassige Logistik
positiv herausstellen knnen.
 Appendix: Questionnaire 295

Inwieweit treffen die folgenden eher grundstzlichen Trifft gar Trifft voll
Aussagen zur Ausgestaltung der Logistik Ihrer nicht zu zu
Geschftseinheit zu?
Entscheidungen zu logistischen Fragen werden in { { { { { { {
unserem Unternehmen sehr zentral gefllt.
Unsere Geschftseinheit beherrscht einen { { { { { { {
reibungslosen, durchgngigen, schnellen und
strungsarmen Material- und Informationsfluss.
Unsere Geschftseinheit wird insgesamt sehr fluss- { { { { { { {
bzw. prozessorientiert gefhrt.
Unsere Geschftseinheit besitzt einen deutlich hheren { { { { { { {
Grad an Fluss- bzw. Prozessorientierung als unsere
Wettbewerber.
Smtliche Prozesse der Leistungserstellung sind in { { { { { { {
unserer Geschftseinheit gut aufeinander abgestimmt.
Es existieren in unserer Geschftseinheit zahlreiche { { { { { { {
Einzelinteressen, die der Erreichung der Ziele unserer
Geschftseinheit im Wege stehen.

D. Fragen zu Logistik- und Unternehmenserfolg

Wie schtzen Sie die Logistikkosten Ihrer Geschftseinheit Sehr viel Sehr viel
im Vergleich zum Wettbewerb ein? schlechter besser
Logistikkosten relativ zum Umsatz (inkl. Vergtung des { { { { { { {
LDL)
Lagerkosten { { { { { { {
Transportkosten { { { { { { {
Durch Logistikprozesse hervorgerufene IT-Kosten { { { { { { {
Personalkosten der Logistik { { { { { { {

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

Wie schtzen Sie den Markterfolg Ihrer Geschftseinheit Sehr viel Sehr viel
im Vergleich zum Wettbewerb ein? schlechter besser
Kundenzufriedenheit { { { { { { {
Kundennutzen { { { { { { {
Bindung bestehender Kunden { { { { { { {
Gewinnung/Akquisition von Neukunden { { { { { { {
Erreichung des angestrebten Wachstums { { { { { { {
Erreichung des angestrebten Marktanteils { { { { { { {
296 Appendix: Questionnaire

Wie schtzen Sie die Flexibilitt Ihrer Geschftseinheit im Sehr viel Sehr viel
Vergleich zum Wettbewerb ein? schlechter besser
Anpassung der Produkte/Dienstleistungen an neue { { { { { { {
Kundenbedrfnisse
Reaktion auf neue Entwicklungen am Markt { { { { { { {
Nutzung neuer Marktchancen { { { { { { {

Wie verhlt sich Ihrer Einschtzung nach die Sehr viel Sehr viel
Umsatzrendite Ihrer Geschftseinheit im Vergleich zum schlechter besser
Wettbewerb?
Unsere Umsatzrendite war im letzten Geschftsjahr im { { { { { { {
Vergleich zu unseren Wettbewerbern...
Unsere Umsatzrendite war im Durchschnitt der letzten { { { { { { {
drei Geschftsjahre im Vergleich zu der unserer
Wettbewerber...
Die Entwicklung unserer Umsatzrendite war in den { { { { { { {
letzten drei Jahren im Vergleich zu der unserer
Wettbewerber...

Vielen Dank, dass Sie den Fragebogen bis hierhin ausgefllt 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 { Geschftsfhrer { Leiter Logistik
ttig?
{ Leiter Teilbereich der { Mitarbeiter { Andere:
Logistik Logistikbereich ________________
Seit wie vielen Jahren sind Sie in dieser Position ttig? ____ Jahre
Seit wie vielen Jahren sind Sie in Ihrem Unternehmen ttig? ____ Jahre

Welcher Branche gehrt Ihre Geschftseinheit an? (Bitte nur eine Branche ankreuzen.)

{ Nahrungs- und Maschinen- und

{ { Chemie/ Kunststoff
Genussmittel Apparatebau
{ Elektrotechnik/
Fahrzeugbau { { Handel
Feinmechanik/ Optik
{ andere:
Konsumgter { { Health Care
__________________

Wie gro ist das Umsatzvolumen Ihrer Geschftseinheit derzeit? (in Mio. / Jahr)
bis 10 { 11-24 { 25-50 { 50-100 {
100-250 { 250-500 { 500-1.000 { ber 1.000 {
 Appendix: Questionnaire 297

Wie viele Mitarbeiter arbeiten ungefhr fr Ihre ____ Mitarbeiter

Geschftseinheit?
Wie viele Logistikdienstleister nutzt Ihre ____ Logistikdienstleister
Geschftseinheit insgesamt?
Welchen Anteil an Ihren gesamten Logistikkosten ____ Prozent
haben externe Logistikdienstleister ungefhr (in %)?

Fllen Sie die folgenden Fragen bitte wieder fr den gleichen Lo-
gistikdienstleister aus, fr den Sie zuvor auch schon die Fragen be-
antwortet haben.

Welchen Anteil am Gesamtvolumen der ____ Prozent

fremdvergebenen Logistikdienstleistungen Ihrer
Geschftseinheit hat dieser Logistikdienstleister?
Seit wie vielen Jahren arbeiten Sie mit diesem ____ Jahre (z.B. 2,5)
Logistikdienstleister zusammen?
Seit wie vielen Jahren arbeiten Sie mit diesem ____ Jahre (z.B. 2,5)
Logistikdienstleister so zusammen, dass Sie von einer
engeren Geschftsbeziehung sprechen wrden?
Wie lang ist die Gesamtlaufzeit des aktuellen Vertrages ____ Jahre (z.B. 2,5)
mit diesem Logistikdienstleister?
Wie lang ist die Restlaufzeit dieses Vertrages mit ____ Jahre (z.B. 2,5)
diesem Logistikdienstleister?
298 Appendix: Questionnaire

Vielen Dank, dass Sie an unserer Studie teilgenommen haben!

Sie knnen jetzt das Dankeschn fr Ihre Teilnahme auswhlen.

Bitte whlen Sie hier das Dankeschn fr Ihre Teilnahme aus. Alle Teilnehmer
erhalten zustzlich einen individuellen Benchmarking-Bericht.
Ich habe als Dankeschn bereits am 2. WHU Logistiksymposium in Vallendar {
teilgenommen.
Ich htte gerne das Buch: "E-Commerce in der Logistik: Quantensprung oder {
business as usual?" von Prof. Dr. Jrgen Weber et al..
Ich mchte am 3. WHU Logistiksymposium (2005) an der WHU teilnehmen. {

Bitte fllen Sie die folgenden Felder aus, damit wir Sie bzgl. Ihres Danke-
schns kontaktieren knnen oder heften Sie eine Vistenkarte an. Bitte ge-
ben Sie auf jeden Fall Name und E-Mail-Adresse an. Falls Sie sich fr das
Buch entscheiden, bentigen wir auch Ihre Anschrift.

Nachname: ___________________________________________

Vorname: ___________________________________________

E-Mail-Adresse: ___________________________________________

Unternehmen: ___________________________________________

Strae: ___________________________________________

Postleitzahl: ___________________________________________

Ort: ___________________________________________

Telefon: ___________________________________________

Falls Sie Anmerkungen oder Kritik haben, knnen Sie uns das im Folgen-
den gerne mitteilen:

___________________________________________________________
___________________________________________________________
___________________________________________________________
___________________________________________________________

Vielen Dank fr Ihre Mitarbeit!

List of figures

Fig. 2-1. The four phases of logistics development..................................10

Fig. 2-2. Logistics development in Germany between 1999 and 2002 ....17
Fig. 2-3. Performance effects of logistics.................................................18
Fig. 2-4. Framework of logistics research ................................................53
Fig. 3-1. The optimal form of organization depending on transaction
dimensions .................................................................................59
Fig. 3-2. Key mediating variables model of relationship marketing ........71
Fig. 3-3. Relations between the organization and the environment
from a logistical viewpoint ........................................................77
Fig. 4-1. Antecedents of the performance of logistics outsourcing
relationships ...............................................................................89
Fig. 4-2. Conceptual logistics outsourcing performance model .............121
Fig. 4-3. Conceptual logistics performance model.................................127
Fig. 4-4. Firm Performance Model developed by Dehler and
Engelbrecht ..............................................................................130
Fig. 4-5. Conceptual firm performance model .......................................134
Fig. 4-6. External and internal contingency variables ............................143
Fig. 5-1. Represented industries and firm sizes in the sample................154
Fig. 5-2. Motivation for logistics outsourcing........................................154
Fig. 5-3. Generic measurement model of a latent variable with
reflective indicators..................................................................157
Fig. 5-4. Complete causal model consisting of two measurement
models and one structural model .............................................158
Fig. 5-5. Adaptation measures................................................................167
Fig. 5-6. Threshold values for adaptation measures ...............................174
Fig. 7-1. Basic model..............................................................................238
Fig. 7-2. Simplified 8-factor model ........................................................241
Fig. 7-3. Simplified 8-factor model with standardized parameter
coefficients...............................................................................242
Fig. 7-4. Hypotheses on the logistics performance model......................248
Fig. 7-5. Logistics performance model...................................................249
Fig. 7-6. Hypotheses on the firm performance model ............................252
Fig. 7-7. 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 Organiza-
tion, 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 ergnzende Konzepte, in: Zeitschrift fr
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 Organ-

izational Rent, in: Strategic Management Journal, Vol. 14 (1): pp. 33-46.

Anderson, E., L. M. Lodish and B. A. Weitz (1987): Resource Alloca-

tion Behavior in Conventional Channels, in: Journal of Marketing Re-
search, Vol. 24 (February 1987): pp. 85-97.

Anderson, E. and B. Weitz (1989): Determinants of Continuity in Con-

ventional 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 Sus-
tain Commitment in Distribution Channels, in: Journal of Marketing Re-
search, Vol. 29 (February 1992): pp. 18-34.

Anderson, E. W. and M. W. Sullivan (1993): The Antecedents and Con-

sequences of Customer Satisfaction for Firms, in: Marketing Science,
Vol. 12 (2): pp. 125-143.

Anderson, J. C. and D. W. Gerbing (1988): Structural Equation Model-

ing 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 As-

sessment of Unidimensional Measurement: Internal and External Con-
sistency, and Overall Consistency Criteria, in: Journal of Marketing Re-
search, Vol. 24 (4): pp. 432-437.

Anderson, J. C., H. Hakansson and J. Johanson (1994): Dyadic busi-

ness relationships within a business network context, in: Journal of Mar-
keting, Vol. 58 (4): pp. 1-15.

Anderson, J. C. and J. A. Narus (1984): A Model of the Distibutor's Per-

spective of Distributor-Manufacturer Working Relationsships, in: Jour-
nal 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 Organ-

izational Commitment and Organizational Effectiveness, in: Administra-
tive Science Quarterly, Vol. 26 (1): pp. 1-14.

Antlitz, A. (1999): Unternehmensgrenzen und Kooperationen: Make-

cooperate-or-buy im Zusammenspiel von Kompetenz- und Strategie-
entwicklung. 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 Per-

tinent 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, Moral-

ity 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): Multivari-

ate Analysemethoden - Eine anwendungsorientierte Einfhrung, 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 Mar-
keting. 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 Devel-

oping and Testing Theories in Consumer Research, in: K. B. Monroe:
Advances in Consumer Research, Vol. VIII. Ann Arbor, MI: pp. 195-
202.

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 Or-

ganizational 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 Va-

lidity 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. Engle-

wood Cliffs.

Bardi, E. J. and M. Tracey (1991): Transportation Outsourcing: A Sur-

vey 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 Lo-

gistik - Supply Chains im Wandel. Berlin.

Baumgarten, H. and S. Walter (2000): Trends und Strategien in der Lo-

gistik 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. 139-
161.

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 Stra-

tegic 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 Bulle-
tin, 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. 78-
117.

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: Compet-

ing through Quality. New York.

Bhargava, M., C. Dubelaar and S. Ramaswami (1994): Reconciling Di-

verse Measures of Performance: A Conceptual Framework and Test of a
Methodology, in: Journal of Business Research, Vol. 31 (2/3): pp. 235-
246.

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 Vari-

ables from Correlations among Indicators, in: American Journal of So-
ciology, 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: Interna-

tional Encyclopedia of the Social Sciences. New York: pp. 452-458.

Bollen, K. and R. Lennox (1991): Conventional wisdom on measure-

ment: 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 catego-

rized measures, in: American Sociological Review, Vol. 46 (2): pp. 232-
239.

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. Jreskog and H. Wold: Sys-
tems 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: Plan-
ning 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, Manufac-
turing 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 Lo-

gistical 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): In-

fluence 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): Manag-

ing 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: Psycho-
metrika, Vol. 52 (3): pp. 345-370.

Bradley, P. (1994a): Contract Logistics: It's All About Costs, in: Purchas-
ing (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 - Voraus-

setzungen, Potentiale und Grenzen, in: Zeitschrift fr Logistik, Vol. 15
(1): pp. 8-12.

Bretzke, W.-R. (1993b): Pro und Contra Outsourcing von Logistikdienst-

leistungen, 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. Ox-

ford, New York.

Buzzel, R. D. and B. T. Gale (1987): The PIMS Principles: Linking Strat-

egy 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: Transporta-
tion & Distribution, Vol. 34 (4): pp. 58-62.

Cahill, D. L. (2006): Customer Loyalty in Third Party Logistics Relation-

ships: 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 Manage-

rial Economics and Marketing, in: Managerial & Decision Economics,
Vol. 14 (2): pp. 163-173.

Caplow, T. (1956): The Effect of Increasing Size on Organizational Struc-

ture in Industry, in: Transactions of the Third World Congress of Soci-
ology, 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 As-

sessment. Newbury Park, London, New Dehli.

Cavinato, J. L. (1989): The Logistics of Contract Manufacturing, in: In-

ternational Journal of Physical Distribution & Materials Management,
Vol. 19 (1): pp. 13-21.

Cavusgil, S. T. and L. A. Elvey-Kirk (1998): Mail survey response be-

havior, in: European Journal of Marketing, Vol. 32 (11/12): pp. 939-
1192.

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 Sci-
ence 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. Edi-

tion. London.

Chow, G., T. D. Heaver and L. E. Henriksson (1994): Logistics Per-

formance: 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, Struc-

ture and Performance: A Framework for Logistics Research, in: Logis-
tics 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 Founda-

tions, 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: In-

ternational 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 Operatio-
nalisierung 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: Market-
ing 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 Reten-

tion of High-Risk Professional Service Customers, in: Services Market-
ing Quarterly, Vol. 22 (2): pp. 17-27.

CSCMP (2005): Supply Chain and Logistics Terms and Glossary. Belle-
vue.

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 Re-
view, in: Academy of Management Review, Vol. 5 (1): pp. 49-64.

Dasgupta, P. (1988): Trust as a Commodity, in: D. Gambetta: Trust: Mak-

ing 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 Lo-

gistics/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 - Determi-

nanten - 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 Partner-

ships on Supplier Performance, in: International Journal of Logistics
Management, Vol. 15 (1): pp. 57-71.

Duncan, O. D. (1966): Path Analysis: Sociological Examples, in: Ameri-

can 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 Mod-
els: 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 Mini-

mize 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 segmenta-

tion: 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: Inter-

national 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 Interna-
tional 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 lo-

gistics and marketing customer service: A strategic perspective, in: Jour-
nal of Business Logistics, Vol. 19 (1): pp. 17-32.

Emerson, R. M. (1962): Power-Dependence Relations, in: American So-

ciological Review, Vol. 27 (February 1962): pp. 31-41.

Engelbrecht, C. (2003): Logistik-Outsourcing: Erfolgsfaktoren und Er-

folgswirkung Erkenntnisse aus der Praxis, in: J. Weber and J. M. Dee-
pen: Erfolg durch Logistik - Erkenntnisse aktueller Forschung. Bern,
Stuttgart, Wien.

Engelbrecht, C. (2004): Logistikoptimierung durch Outsourcing: Er-

folgswirkungen 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 fr eine va-

lide 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 Re-
wards, Costs, Alternatives, and Investments, in: Organizational Behav-
ior 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: Jour-
nal 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 Market-

ing 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 Mar-
keting, Vol. 48 (Summer 1984): pp. 43-55.

Galbraith, J. (1973): Designing Complex Organizations. Reading.

Ganesan, S. (1994): Determinants of Long-Term Orientation in Buyer-

Seller Relationships, in: Journal of Marketing, Vol. 58 (April 1994): pp.
1-19.

Gardner, J. T. and M. C. Cooper (1988): Elements of Strategic Partner-

ship, 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 revis-
ited., 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 Interor-
ganizational Relationship Retention Decisions, in: Journal of the Acad-
emy of Marketing Science, Vol. 26 (4): pp. 322-337.

Gassenheimer, J. B., J. U. Sterling and R. A. Robicheaux (1996): Long-

term 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 part-

nerships, 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 Assess-
ment, 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 Loyalitt - 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.

Gpfert, 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, Question-
naire, Length, and Time of the Week, in: Journal of Business-to-
Business 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.

Grnroos, C. (1989): Defining Marketing: A Market-Oriented Approach,

in: European Journal of Marketing, Vol. 23 (1): pp. 52-60.

Grnroos, 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.

Grnroos, C. (1994): Quo Vadis, Marketing? Toward a Relationship

Marketing Paradigm, in: Journal of Marketing Management, Vol. 10:
pp. 347-360.

Grnroos, C. and E. Gummesson (1985): Service Orientation in Indus-

trial Marketing - Working Paper. Helsinki, Stockholm.

Gruen, T. W., J. O. Summers and F. Acito (2000): Relationship Market-

ing 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 Influ-

encing Coordination in a Franchise Channel, in: Journal of Retailing,
Vol. 56 (3): pp. 41-58.

Gummesson, E. (1987): The New Marketing: Developing Long term In-

teractive Relationships, in: Long Range Planning, Vol. 20 (4): pp. 10-
20.

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. Acad-

mic commemorative speech on the occasion of the foundation of the
University on May, 22nd, 1957, Krefeld: pp. 5-38.

GVB (2000): Die GVB (Gesellschaft fr Verkehrsbetriebswirtschaftsleh-

re) Top 100 in Europa: Der Club der Umsatzmilliardre, in: Logistik
Heute (1-2): pp. 32-35.

Hackett, R. D., P. Bycio and P. A. Hausdorf (1994): Further Assess-

ments of Meyer and Allen's (1991) Three-Component Model of Organ-
izational 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 Indus-

trial 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 Assess-

ment, in: American Journal of Sociology, Vol. 69: pp. 32-40.

Halln, L., J. Johanson and N. Seyed-Mohamed (1991): Interfirm Adap-

tation 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 Rela-

tionships Today, in: Industrial Marketing Management, Vol. 22: pp.
331-338.

Hardy, K. G. and A. J. Magrath (1988): Marketing Channel Manage-

ment: Strategic Planning and Tactics. Glenview, IL.

Harrington, B. E. (2001): Organizational Performance and Corporate So-

cial 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 chan-

nels, 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: Jour-
nal of Marketing Research, Vol. 27 (February 1990): pp. 24-36.

Heide, J. B. and G. John (1992): Do Norms Matter in Marketing Rela-

tionships?, 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. 265-
291.

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 Trans-
actions 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 Busi-

ness Review, Vol. 55 (6): pp. 85-96.

Hildebrandt, L. (1998): Kausalanalytische Validierung in der Marketing-

forschung, in: L. Hildebrandt and C. Homburg: Die Kausalanalyse - In-
strument der empirischen betriebswirtschaftlichen Forschung. Stuttgart:
pp. 86-110.

Hobbs, J. E. (1996): A transaction cost approach to supply chain man-

agement, in: Supply Chain Management, Vol. 1 (2): pp. 15-27.

Hocutt, M. A. (1998): Relationship dissolution model: Antecedents of re-

lationship 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 Jour-

nal of Sociology, Vol. 63: pp. 597-606.

Homburg, C. (1989): Exploratorische Anstze der Kausalanalyse als In-

strument der Marketingplanung. Frankfurt am Main, Bern, New York.

Homburg, C. (1995): Kundennhe von Industriegterunternehmen - Kon-

zeption, Erfolgsauswirkungen, Determinanten. Wiesbaden.

Homburg, C. and H. Baumgartner (1995a): Beurteilung von Kausalmo-

dellen - Bestandsaufnahme und Anwendungsempfehlungen, in: Marke-
ting - ZFP, Vol. 17 (3): pp. 162-176.

Homburg, C. and H. Baumgartner (1995b): Die Kausalanalyse als In-

strument der Marketingforschung, in: Zeitschrift fr Betriebswirtschaft,
Vol. 65 (10): pp. 1091-1108.
328 References

Homburg, C. and A. Giering (1996): Konzeptualisierung und Operatio-

nalisierung komplexer Konstrukte - Ein Leitfaden fr die Marketingfor-
schung, in: Marketing - ZFP, Vol. 18 (1): pp. 5-24.

Homburg, C., A. Giering and A. Menon (2003): Relationship Character-

istics as Moderators of the Satisfaction-Loyalty Link: Findings in a
Business-to-Business Context, in: Journal of Business-to-Business Mar-
keting, 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 relation-

ships 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 Em-

ployee Attitudes and Perceptions of Influence, in: Academy of Man-
agement 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. 649-
662.

Hu, L.-t. and P. M. Bentler (1995): Evaluating Model Fit, in: R. H.

Hoyle: Structural Equation Modeling: Concepts, Issues, and Applica-
tions. London, New Dehli: pp. 76-99.

Hu, L.-t. and P. M. Bentler (1999): Cutoff Criteria for Fit Indexes in Co-
variance Structure Analysis: Conventional Criteria Versus New Alterna-
tives, 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: Manage-

rial Behavior, Agency Costs and Ownership Structure, in: Journal of Fi-
nancial Economics, Vol. 3 (4): pp. 305-360.

John, D. R. and C. A. Cole (1986): Age Differences in Information Proc-

essing: 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 Op-

portunism 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 Infor-
mant Data from Dyadic Relationships in Marketing Channels, in: Jour-
nal of Marketing Research, Vol. 19 (4): pp. 517-524.

Jreskog, K. G. (1973): A General Method for Estimating a Linear Struc-

tural Equation System, in: A. S. Goldberger and O. D. Duncan: Struc-
tural Equation Models in the Social Sciences. New York: pp. 85-112.

Jreskog, K. G. (1977): Structrual Equation Models in the Social Sci-

ences: Specification, Estimation, and Testing, in: P. R. Krishnaiah: Ap-
plications of Statistics. Amsterdam: pp. 265-287.

Jreskog, K. G. and D. Srbom (1982): Recent Developments in Struc-

tural 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 Ad-
ministrative 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 Manufacturer-

Supplier 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: Gestal-

tung strategischer Gesamtsysteme und Management einzelner Transak-
tionen. 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 The-

ory 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 Manage-
ment Review, Vol. 27 (2): pp. 275-293.

Kieser, A. (1974): Der Einfluss der Umwelt auf die Organisationstruktur

der Unternehmung, in: Zeitschrift fr Organisation, Vol. 43 (6): pp. 302-
314.

Kieser, A. and H. Kubicek (1978): Organisationstheorien II: Wissen-

schaftstheoretische Anforderungen und kritische Analyse klassischer
Anstze. Stuttgart, Berlin, Kln, 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 Ap-

plied Approach, 5. Edition. New York.

Kirsch, W., I. Bamberger, E. Gabele and K. Klein (1973): Betriebswirt-

schaftliche Logistik. Wiesbaden.

Klaus, P. (1993): Die dritte Bedeutung der Logistik, Nuremberg Logistics

Working Paper No. 3, University of Erlangen-Nrnberg.

Klaus, P. (1999): Logistik als Weltsicht, in: J. Weber and H. Baumgar-

ten: Handbuch Logistik Management von Material- und Wa-
renfluprozessen. Stuttgart: pp. 15-32.

Klaus, P. (2003): Die "Top 100" der Logistik - Marktgrssen, Marktseg-

mente und Marktfhrer in der Logistikdienstleistungswirtschaft. Ham-
burg.

Kleer, M. (1991): Gestaltung von Kooperationen zwischen Industrie- und

Logistikunternehmen - Ergebnisse theoretischer und empirischer Unter-
suchungen. Berlin.

Klein, B., R. G. Crawford and A. A. Alchian (1978): Vertical Integra-

tion, 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 out-

sourcing 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 Per-
spective, in: The Journal of Supply Chain Management, Vol. 40 (1): pp.
35-51.

Knemeyer, A. M. and P. R. Murphy (2005): Exploring the Potential Im-

pact of Relationship Characteristics and Customer Attributes on the Out-
comes of Third-party Logistics Arrangements, in: Transportation Jour-
nal, 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 Market-
ing Research, Vol. 32 (3): pp. 348-356.
332 References

Kumar, N., L. W. Stern and J. C. Anderson (1993): Conducting interor-

ganizational research using key informants, in: Academy of Manage-
ment Journal, Vol. 36 (6): pp. 1633-1651.

Kummer, S. (1993): Outsourcing von Logistikleistungen: Controlling

wirkt qualittssichernd, 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 Cus-

tomer Service: A Third-Party Perspective. Oak Brook, IL.

LaLonde, B. J. and A. B. Maltz (1992): Some Propositions about Out-

sourcing the Logistics Function, in: The International Journal of Logis-
tics Management, Vol. 3 (1): pp. 1-11.

Lambe, C. J., C. M. Wittmann and R. E. Spekmann (2001): Social Ex-

change Theory in Research on Business-to-Business Relational Ex-
change, in: Journal of Business-to-Business Marketing, Vol. 8 (3): pp.
1-36.

Lambert, D. M., M. A. Emmelhainz and J. T. Gardner (1996a): Devel-

oping 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): Build-

ing 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 Lo-

gistics 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 Logis-
tics: Research & Applications, Vol. 7 (1): pp. 17-31.

Larson, P. D. and J. D. Kulchitsky (1999): Logistics improvement pro-

grams - 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 Environ-

ment. 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 Mor-
ton 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 Euro-
pean 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 Third-

Party 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 Ef-
fects of Conciliatory Acts on Conflict and Cooperation, in: Psychologi-
cal 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 & Trans-
portation 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: Psycho-
logical Bulletin, Vol. 103 (3): pp. 391-410.

McCabe, J. V. (1990): Outside Managers Offer Packaged Export Exper-

tise, 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 Con-
tinuance 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 Man-
agement, Vol. 6 (2): pp. 93-102.

McHugh, M., P. Humphreys and R. McIvor (2003): Buyer-supplier re-

lationships and organizational health, in: Journal of Supply Chain Man-
agement, Vol. 39 (2): pp. 15-25.

McIvor, R. (2000): A practical framework for understanding the outsourc-

ing 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 re-

search, in: Journal of Business Logistics, Vol. 16 (1): pp. 231-250.
336 References

Mentzer, J. T. and B. P. Konrad (1991): An efficiency/effectiveness ap-

proach to logistics performance analysis, in: Journal of Business Logis-
tics, Vol. 12 (1): pp. 33-62.

Metcalf, L. E., C. R. Frear and R. Krishnan (1992): Buyer-Seller Rela-

tionships: 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 Or-

ganizational 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 Or-

ganizations 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 be-

haviour in logistics relationships: A shipper perspective, in: Interna-
tional Journal of Physical Distribution & Logistics Management, Vol.
29 (2): pp. 103-121.

Moorman, C., R. Deshpand 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. Deshpand (1992): Relationships be-

tween 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. 20-
38.
 References 337

Moscovitch, M. (1982): A neuropsychological approach to perception and

memory in normal and pathological aging, in: F. I. M. Craik and S. Tre-
hub: 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: Transporta-
tion 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 Out-

comes of Purchasing Arrangements in Industrial Buyer-Vendor Rela-
tionships, in: Journal of Marketing, Vol. 54 (4): pp. 80-93.

North, D. C. (1990): Institutions, Institutional Change and Economics Per-

formance. Cambridge.

Nunnally, J. C. (1978): Psychometric Theory. New York, St. Louis, San

Francisco.

Pach, G. (1998): Logistics outsourcing in grocery distribution: A Euro-

pean 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 Coop-

eration in Channel Performance, in: K. L. Bernhardt: Marketing 1776-
1976 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. Bel-

mont.

Persson, G. (1997): Organization design strategies for business logistics,

in: International Journal of Physical Distribution & Logistics Manage-
ment, Vol. 27 (5/6): pp. 282-291.

Peter, J. P. (1979): Reliability: A Review of Psychometric Basics and Re-

cent 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 Meta-
Analysis, 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 - Mrkte - Ressourcen.

Berlin.

Pfohl, H.-C. and W. Zllner (1997): Organization for logistics: The con-
tingency 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 fr betriebswirtschaftliche Forschung, Vol. 43 (4): pp. 336-
357.

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. 1537-
1562.

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 (Spe-
cial 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): Out-

sourcing 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 ser-

vice quality-customer retention link, in: International Journal of Opera-
tions & 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 Organiza-

tional 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: Transpor-

tation & 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 Man-
agement Review, Vol. 19 (1): pp. 90-118.

Rodrguez, C. M. and D. T. Wilson (2002): Relationship Bonding and

Trust as a Foundation for Commitment in U.S.-Mexican Strategic Alli-
ances: A Structural Equation Modeling Approach, in: Journal of Interna-
tional 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 Man-

agement, 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 Con-
ference 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 Or-

ganization 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.

Schfer-Kunz, J. and C. Tewald (1998): Make-or-buy Entscheidungen in

der Logistik. Wiesbaden.

Schaninger, C. M. and D. Sciglimpaglia (1981): The Influence of Cogni-

tive 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 Coop-

eration, in: Academy of Management Journal, Vol. 18 (4): pp. 846-856.

Schmitt, A. (2006): 4PL-Providing als strategische Option fr Kontraktlo-

gistikdienstleister: Eine konzeptionell-empirische Betrachtung -
Forthcoming Dissertation. Vallendar.

Schrage, M. (1990): Shared minds: The New Technologies of Collabora-

tion. New York.

Schreygg, 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 In-

quiry into Profits, Capital, Credit, Interest and the Business Cycle. Cam-
bridge.

Schurr, P. H. and J. L. Ozanne (1985): Influences on Exchange Proc-

esses: Buyers' Preconceptions of a Seller's Trustworthiness and Bargain-
ing Toughness, in: Journal of Consumer Research, Vol. 11 (4): pp. 939-
953.

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: Ameri-

can 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 Ex-

change in Services: An Empirical Investigation of Ongoing Customer
Service-provider Relationships, in: International Journal of Service In-
dustry 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): Coopera-

tion in Supplier-Dealer Relations, in: Journal of Retailing, Vol. 68 (2):
pp. 174-193.

Skjoett-Larsen, T. (2000): Third party logistics - from an interorganiza-

tional 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 dif-

ferences 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: Devel-
opment 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 Rela-
tionships, in: European Journal of Marketing, Vol. 31 (11/12): pp. 832-
856.

Spence, M. T. and M. Brucks (1997): The Moderating Effects of Prob-

lem Characteristics on Experts' and Novices' Judgements, in: Journal of
Marketing Research, Vol. 34 (2): pp. 233-247.

Staehle, W. (1990): Management - Eine verhaltenswissenschaftliche Per-

spektive, 6. Edition. Mnchen.

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): Lo-

gistics 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 Col-

laboration 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 Meas-

urement 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 num-
ber 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 Organiza-
tion: Selected Papers. Cambridge: pp. 231-267.

Stock, J. R. (1997): Applying theories from other disciplines to logistics,

in: International Journal of Physical Distribution & Logistics Manage-
ment, Vol. 27 (9/10): pp. 515-539.

Stlzle, W. (1996): Grundzge des Outsourcing von Entsorgungsleistun-

gen, in: Zeitschrift fr Betriebswirtschaft, ZfB-Ergnzungsheft 2/96: pp.
121-145.

Straube, F., H.-C. Pfohl, W. A. Gnther 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 Strate-

gic Benfits, in: International Journal of Purchasing and Materials Man-
agement, Vol. 29 (4): pp. 22-28.
346 References

Stuart, F. I. and D. McCutcheon (1995): Problem Sources in Establish-

ing Strategic Alliances, in: International Journal of Purchasing and Ma-
terials 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: In-

ternational 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 Ac-

ceptance of Marital Termination, in: Journal of Marriage and the Fam-
ily, 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 En-
vironment: 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: Transpor-

tation & 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 Organiza-

tional 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: Lo-

gistics and Transportation Review, Vol. 29 (4): pp. 355-361.

Wallenburg, C. M. (2003): Kundenbindung in der Logistik: Eine empiri-

sche Untersuchung zu Geschftsbeziehungen zwischen Logistik-
dienstleistern 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 empiri-

sche 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 cog-
nitive processes. New York: pp. 99-125.

Wathne, K. H. and J. B. Heide (2000): Opportunism in Interfirm Rela-

tionships: Forms, Outcomes, and Solutions, in: Journal of Marketing,
Vol. 64 (4): pp. 36-51.

Weber, J. (1986): Zum Begriff Logistikleistung, in: Zeitschrift fr Be-

triebswirtschaft, Vol. 56 (12): pp. 1197-1212.

Weber, J. (1987): Logistikkostenrechnung. Berlin, Heidelberg, New York,

London, Paris, Tokyo.

Weber, J. (1995): Logistik-Controlling: Leistungen - Prozekosten -

Kennzahlen, 4. Edition. Stuttgart.

Weber, J. (1999): Stand und Entwicklungsperspektiven des Logistik-

Controlling, WHU Research Paper No. 61. Vallendar.

Weber, J. (2002): Logistikkostenrechnung, 2. Edition. Berlin.

348 References

Weber, J. (2003): Macht Logistik erfolgreich? - Konzeptionelle berle-

gungen 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: Fhrungsaufga-

ben zur Umsetzung des Fluprinzips im Unternehmen. Stuttgart.

Weber, J. and S. Kummer (1998): Logistikmanagement, 2. Edition.

Stuttgart.

Weber, J. and C. M. Wallenburg (2004): Zusatzbeauftragung von Lo-

gistikdienstleistern - Empirische Ergebnisse und konzeptionelle berle-
gungen zu entsprechenden Defiziten, in: Logistik Management, Vol. 6
(4): pp. 34-46.

Weitz, B. A. and S. D. Jap (1995): Relationship Marketing and Distribu-

tion Channels, in: Journal of the Academy of Marketing Science, Vol.
23 (4): pp. 305-320.

Wentworth, F. (2003): Outsourcing Services: The Case Against, in: Jour-

nal of the Institute of Logistics & Transport, Vol. 5 (2): pp. 57-59.

Werner, H. (1997): Relationales Beschaffungsverhalten: Ausprgungen

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-Produzenten-

Beziehungen - Eine Analyse von Unternehmensnetzwerken in der deut-
schen Automobilindustrie. Wiesbaden.

Wetzels, M., K. de Ruyter and M. van Birgelen (1998): Marketing ser-

vice 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 Organ-

izational Commitment as Predictors of Organizational Citizenship and
In-Role Behaviors, in: Journal of Management, Vol. 17 (3): pp. 601-
617.

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 Transac-

tion 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. Schmalen-

see 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. 335-
345.

Wong, A. and A. S. Sohal (2002): An examination of the relationship be-

tween trust, commitment and relationship quality, in: International Jour-
nal 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 Co-

operation 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: Adminis-

trative Science Quarterly, Vol. 17 (2): pp. 229-239.

Zeithaml, V. A. (1985): The New Demographics and Market Fragmenta-

tion, 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 Geschftsbeziehungen. Wiesbaden.