Sebastian Raisch (auth.) - Dynamic Strategic Analysis_ Demystifying Simple Success Strategies (2004, Deutscher Universitätsverlag).pdf

Sebastian Raisch Dynamic Strategic Analysis

GABLER EDITION WISSENSCHAFT

Sebastian Raisch

Dynamic Strategic Analysis Demystifying simple success strategies

With a foreword by Prof. Dr. Gilbert Probst

Deutscher UniversiHits-Verlag

Bibliografische Information Der Deutschen Bibliothek Die Deutsche Bibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet uber abrufbar.

Dissertation Universitat Genf, 2004

1. Auflage Dezember 2004 Aile Rechte vorbehalten © Deutscher Universitats-VerlagiGWV Fachverlage GmbH, Wiesbaden 2004 Lektorat: Brigitte Siegel! Nicole Schweitzer Der Deutsche Universitats-Verlag ist ein Unternehmen von Springer Science+Business Media. www.duv.de Das Werk einschlieBlich aller seiner Teile ist urheberrechtlich geschUtzt. Jede Verwertung auBerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verla.gs unzulassig und strafbar. Das gilt insbesondere fUr Vervielfaltigungen, Ubersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in diesem Werk berechtigt auch ohne besondere Kennzeichnung nicht zu der Annahme, dass solche Namen im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei zu betrachten waren und daher von jedermann benutzt werden dUrften. Umschlaggestaltung: Regine Zimmer, Dipl.-Designerin, Frankfurt/Main ISBN-13: 978-3-8244-8249-8 001: 10.1007/978-3-322-81883-6

e-ISBN-13: 978-3-322-81883-6

Foreword Today many business environments are characterized by frequent changes that result in a higher market dynamism and complexity. Global media industries are an up-to-date example of this development: a fast growth period was followed by a rapid downturn. Media companies wrote off a colossal US$ 120 billion from their balance sheets in 2002 - mostly as a result of growth strategies' failure. It seems that despite popular belief growth or market share by itself is less than ever before a guarantor of success. Sebastian Raisch decided to analyze various strategies in order to gain a bett8r understanding of the determinants of successful firm performance. When he decided to go on a research journey into the elements and dynamics of strategic thinking and firm performance he brought along all the ingredients necessary for such a challenging project; he had experience gained through exhaustive consulting work, excellent knowledge of management theories and, as a result of his studies and activities as a research assistant, specific knowledge of strategic management. His research objectives were furthermore clearly defined and, thanks to a meticulously planned and well-organized project as well as his enthusiasm for the topic, he had ready access to an empirical field. In his work he shows that although many insights may be gained from the existing strategy literature, it is only of limited value when analyzing performance in dynamic markets. Analyzing the factors underlying firm success has always been strategic management's foremost activity, but the existing literature fails to capture the increasingly dynamic and complex nature of competition. In a straight forward way Sebastian Raisch demonstrates that strategic management theory is still highly fragmented, often oversimplifies the interrelation between determinants, and remains inherently static in its approach. Researchers as well as reflective practitioners will find the result of his research well worth reading! Corollary to the one-dimensional and static character of most existing approaches, the great majority of empiric studies has been conducted in stable environments. Sebastian Raisch, however, provides a comprehensive strategic analysis that captures the growing

v

complexity and dynamic in the marketplace. In his study he establishes three models, the integrated, the complex and the evolutionary model, to address the current limitations of fragmentation, oversimplification and statism. The integrated model combines the most relevant determinants from competing perspectives into a single model. The complex model captures the most relevant interrelations between these determinants. The evolutionary model encompasses the evolution of these determinants over time. The combined results provide a state-of-the-art view of the factors and effects underlying firm performance. All three models were validated in a field study of the global media industries. Differences in firm performance in dynamic markets such as the global media industry can be explained through a comprehensive research framework that captures the dynamic interaction of a multitude of underlying determinants. This book definitely provides new insights into and a broader understanding of strategy theory and the analysis of firms' success factors. Different strategy perspectives' contributions are not simply critized, but integrated and leveraged in an integrated and dynamic view. For once the media industry does not serve to provide simple illustrations of television markets' consolidation and domination, to provide historic case descriptions of huge mergers, rapid technology changes, regulatory liberalization or fierce competition, but serves as a most interesting empirical field for understanding the dynamics between markets, competition and performance This piece of research allows the success factors in all kinds of industries to be interpreted and understood. It contributed to and also served as a basis for research into company failures (see Probst / Raisch 2004) and leads naturally to the question of what successful companies--balanced companies as Sebastian Raisch would call them--really do as well as providing answers. Prof. Dr. Gilbert Probst

VI

Acknowledgements As Paulo Coelho said in his book The Alchemist, "To realize one's destiny is a person's only obligation. When you want something, all the universe conspires in helping you to achieve it." Little, if anything in life, is accomplished without the assistance, support, and love of others. I am grateful to all who helped with this project and supported me, both intellectually and emotionally, throughout those years. My doctoral years at the University of Geneva, were a source of immense pleasure and rare privilege. First and foremost, I would like to thank my adviser, Professor Gilbert Probst, for the outstanding guidance and support that he provided over the past years. More than anything, however, I deeply appreciate the faith in me that he has continually shown. He is, and will always be, an inspiration to all of us. I am also deeply grateful to Professor Carlos Jarillo for many thought-provoking discussions, for challenging me in ways that made my work stronger, as well as for serving on my committee. A special thank you is due to Professor Yves Fluckiger and Professor Gunther Muller-Stewens whose incisive comments cast new light upon my approaches in this work. I am also indebted to all my friends and colleagues at the University of Geneva. My special thanks go out to Heidi Armbruster, Stefano Borzillo, Claire Di Giovanni, Eva Simeth, and Thomas Straub for a wonderful time in Geneva. I couldn't have asked for better colleagues with whom I shared this experience. lowe more than I can say to my parents, my family, and my long-term friends from Geneva and around the world for their support and care over the years. People come and go, but some remain - thank you for being always there. Finally, I would like to thank Simone for being so patient with me during the last stages of the PhD, for encouraging me in hard times during this project, and for all the passionate years we have lived together. I cannot imagine a more wonderful partner.

VII

To close, I would like to echo the words of Sigmund Freud: "People need not be glued together when they belong together." While the publication of this study concludes my time in Geneva and marks the start of my Post-Doc position at the University of St.Gall, my friends in Geneva will always remain close to me. Sebastian Raisch

VIII

Table of Contents I.

Introduction .............................................................................................................. 1 1.1 1.2 1.3 104 1.5 1.6

II.

Research Interest. ............................................................................................... 1 Research Objectives ........................................................................................... 4 Foundations for a Dynamic ModeL ...................................................................... 6 Model Development & Hypotheses ..................................................................... 9 Research Methodology ..................................................................................... 13 Structure of the Study ....................................................................................... 14

Literature Review ................................................................................................... 17 11.1 11.2 11.3 1104

Industrial Organization ...................................................................................... 18 Corporate Strategy ............................................................................................ 31 Business Strategy ............................................................................................. 48 Resource-based View ....................................................................................... 60

III. Towards a Dynamic Research Model.. ................................................................. 75 111.1 111.2 111.3 lilA

Criticism of Classical Approaches ..................................................................... 76 The Emerging "Dynamic View" ......................................................................... 83 Foundations of a Dynamic Research Model ................................................... 109 Model Development & Hypotheses ................................................................. 123

IV. Research Methodology ....................................................................................... 143 IV.1 Research Design ............................................................................................. 143 IV.2 Questionnaire Development. ........................................................................... 150 IV.3 Survey Implementation ................................................................................... 162 IVA Data Analysis .................................................................................................. 166 V. Research Findings ............................................................................................... 177 V.1 V.2 V.3 VA

The Integrated Model ...................................................................................... 177 The Complex Model ........................................................................................ 199 The Evolutionary Model .................................................................................. 212 Implications ..................................................................................................... 224

VI. Conclusions ......................................................................................................... 235 VI.1 VI.2 VI.3 VI.4

Summary and Implications .............................................................................. 235 Limitations of the Study ................................................................................... 243 Future Research ............................................................................................. 245 Concluding Comments .................................................................................... 246

References .................................................................................................................. 247

IX

List of Figures Figure

Page Frame of Reference...................................................

8

2

Dynamic Research Framework ....... ...................... .......

10

3

The Integrated Model................................................. 11

4

The Structure-Conduct-Performance Paradigm...............

5

The PIMS Competitive Strategy Paradigm...................... 50

6

The Resource-based Model........................................

7

Frame of Reference (2) .............................................. 122

8

Dynamic Research Framework (2) ................................ 124

9

The Integrated Model (2) ............................................. 129

10

Sample Distribution .................................................... 147

11

Response Distribution ................................................. 164

12

Informant Competence ................................................ 165

13

Reliability Analysis ..................................................... 168

14

Convergent Validity Analysis ........................................ 170

15

Discriminant Validity Analysis ....................................... 171

16

Mediation ................................................................. 173

17

Simple Regression Analysis (L1SREL) ............................ 179

18

Simple Regression Analysis (SPSS) .............................. 181

19

Multiple Regression -Industry Perspective ..................... 183

20

Multiple Regression - Corporate Strategy Perspective ...... 184

21

Multiple Regression - Competitive Strategy Perspective .... 185

XI

20

65

Page

Figure 22

Multiple Regression - Firm Capabilities Perspective ......... 186

23

Example for LlSREL Model ......................................... 188

24

Structural Equations Modeling - Industry Submodel ......... 191

25

Structural Equations Modeling - Corporate Strategy Submodel ................................................................ 192

26

Structural Equations Modeling - Competitive Strategy Submodel ................................................................ 193

27

Structural Equations Modeling - Firm Capabilities Submodel ................................................................ 194

28

Structural Equations Modeling - Integrated Model ............ 196

29

I ntegrated Model - Hypothesis Testing .......................... 198

30

Structural Equations Modeling - Complex Model (1/3) ...... 200

31

Structural Equations Modeling - Complex Model (2/3) ...... 202

32

Structural Equations Modeling - Complex Model (3/3) ...... 203

33

Mediation Effects (1/3) - Competitive Strategy & Capabilities ................................................................. 206

34

Mediation Effects (2/3) - Corporate & Competitive Strategy .................................................................. 208

35

Mediation Effects (3/3) - Corporate Strategy & Capabilities .............................................................. 209

36

Complex Model- Hypothesis Testing ............................ 211

37

Cross-Correlations ..................................................... 214

38

Changes in Mean Values ............................................ 216

39

Patterns of Dynamic Change ....................................... 218

40

Regression Effects on Performance .............................. 221

XII

Page

Figure 41

Impact of Changes on Performance .............................. 223

42

Innovativeness in the German Broadcasting Industry ....... 226

43

The ABC Strategy of BSkyB ........................................ 227

44

Diversification Makes the Superstar .............................. 228

45

McGraw-Hill's Balanced Strategy ................................. 229

46

The Integrated Model ................................................ 237

47

The Complex Model .................................................. 238

48

The Evolutionary Model ............................................. 241

XIII

I.

Introduction

During the last few years prominent corporate failures in numerous industries have shown that it is difficult to grasp the magic formula for firm success. For instance, media companies wrote off a colossal US$ 120 billion from their balance sheets during 2002 more than the gross domestic product of the Republic of Ireland. 1 Industry analysts agree that most of the current difficulties are caused by the failure of firm strategies. In their blind rush towards diversification, media companies snapped up competitors at far too high a price. Troubled by these prominent disasters, companies increasingly realize that simplistic and one-sided success strategies can, less than ever, guarantee sustainable success in the complex and dynamic market environments. However, the question regarding the determinants of superior performance remains open.

1.1

Research Interest

The foremost activity as far as strategy theory is concerned, has always been the analysis of factors underlying firm success (Barney 2002). To date literally thousands of studies have analyzed different factors that mayor may not be related to firm success. Four main research perspectives can be distinguished that have subsequently dominated the debate over the past four decades. In the 60s research was dominated by the industrial organization (10) school of thought (Bain 1959). According to this school, superior performance is associated with industries with a highly concentrated structure and high barriers of entry. In the 70s scholars of strategic management related firm performance to corporate strategy, namely to the diversification (Rumelt 1974) and the market share (Buzzell et al 1975) of the firm. In the 80s the focus shifted towards business strategy, promoting product differentiation and cost leadership strategies as key to superior performance (Porter 1980). The 90s were dominated by the resourcebased perspective and the analysis of the impact of firm-specific capabilities on company performance (Barney 1991).

1

The Guardian (19/05/2002)

Despite the strategic management's rich foundations, three significant shortcomings undermine the usability of its findings: Fragmentation: So-called variance decomposition studies have clearly shown that

industry-, corporate- and business-level factors are all important in explaining firm performance (Roquebert et al 1996; Chang / Singh 2000; Bowman / Helfat 2001; McGahan / Porter 1997, 2002). In other words, performance cannot be explained by means of single factors or on one level of analysis alone (Porter 1991, 114). However, the different perspectives tend "to fragment or dichotomize the important parts of the problem rather than integrate them" (item, 95). While building one anothe~, their emphasis strongly differs. Less for theoretical than for academic reasons, the different perspectives remain focused on single determinants, use different labels (often for similar aspects), while ignoring the contributions of other research streams (Pettigrew 2002). An integrated view of the determinants of success thus remains to be developed (Jemison 1981, Rumelt 1991, Stimpert / Duhaime 1997, Teece et a11997, McNamara et aI2003). Determinism: Integrating the different perspectives certainly provides a richer picture,

but still fails to capture the complex reality. The existing literature assumes linear and deterministic relationships between each determinant and performance. Once again the variance decomposition studies mentioned above claim the opposite: There are important interrelations between corporate, industry, and business unit effects (Bowman / Helfat 2001, McGahan / Porter 2002). Success in dynamic industries "requires the interaction of favorable conditions in several of the determinants" (Porter 1991, 114). The interplay between causal variables has thus far been neglected by the literature (Capon et al 1990, Henderson / Mitchell 1997, Stimpert / Duhaime 1997, Black / Farias 2000). An exception is the strategic contingency theory that analyzes the interrelation between single strategic determinants (Le. diversification) and situational organizational or environmental conditions (Ginsberg / Venkatraman 1985). Additional insights can also be gained from the systems theory and complexity theory (Stacey 1995, McKelvey 1999,

2

For instance, Porter's (1980) approaches to competitive strategy rely upon industrial organization's core models. The concept of "imitation barriers" of the resource-based view is closely related to industrial organization's entry barriers theory (Lippman / Rumelt 1982, Rumelt 1984, Reed / DeFillippi 1990).

2

Black / Farias 2000). However, there are very few such studies on strategy, and thus no tested frameworks on the mutual influence between various determinants (Dess et al 1995, Phelps 2001, Farjoun 2002, Pettigrew et al 2002). Static View: Integrating different approaches while accounting for interrelations between determinants may lead to a reasonably realistic picture. However, a picture still fails to show as much as a movie does. Studies have shown that determinants change over time (Le. Porter 1991, Thomas 1996, Teece et al 1997, McGahan / Porter 2002). Despite this evidence, all major theories analyze the influence of determinants at a specific point in time, "without specifying the dynamics through which these outcomes develop" (Barnett / Burgelman 1996, 5). A comprehensive strategy model thus not only integrates competing perspectives, "but a/so adds a dynamic dimension to strategic analysis" (Shay / Rothaermel 1999, 561). More recently, however, two research streams have addressed dynamics: the dynamic capabilities view and the evolutionary theory. Proponents of the dynamic capabilities perspective analyze the development and change of core capabilities over time (Teece et al 1997). The evolutionary theory focuses on the dynamic process of competitive evolution over time (Nelson / Winter 1982, Jacobsen 1992, D'Aveni 1994, Barnett / Burgelman 1996). While both approaches contribute important insights they fail, however, to provide normative theories and empirical validation (Barney 1986, Jacobsen 1992, Nelson 1995, Nelson 1999). In summary, while significant insights into individual determinants of company performance have been gained, the field is still highly fragmented, ignores or oversimplifies the interrelation between determinants, and remains inherently static in its approach. While new research streams such as complexity or evolutionary theory contribute to a more dynamic, and thus more realistic, picture of strategy, they are fragmented and fail to provide testable frameworks and methods (Jacobsen 1992, Nelson 1995, Priem / Butler 2001). The huge potential for cross-fertilization between traditional and more recent dynamic approaches is neglected by both sides (Black / Farias 2000, McNamara et al 2003, Powell / Wakeley 2003). The literature still lacks a more comprehensive and realistic approach to strategic analysis that could capture the complex and dynamic effects that determine company success (Jemison 1981, Porter 3

1991, Dess et al 1995, Sanchez I Heene 1997, Farjoun 2002, Pettigrew et al 2002, McNamara et al 2003).

1.2

Research Objectives

Establishing a truly comprehensive framework that captures all determinants, as well as the full complexity and dynamism of an industry, will be a "Herculean task" (Pettigrew et al 2002, 18). Some scholars, in particular supporters of the complexity theory, argue that success can never be fully explained due to its inherent complexity and the random elements affecting it (Dooley I Van de Ven 1999, McKelvey 1999). Population ecology goes even further and argues that success or failure is mostly a random mutation (Hannan I Freeman 1977, McKelvey I Aldrich 1983). According to this view, companies mostly fail or survive regardless of leaders or strategists' actions. The environment is simply too powerful a force for organizational adaptation of any kind. Even if we do not share this view, we need to be very aware of the limitations imposed by complexity when establishing the research objectives of this study. While we strive for a comprehensive model, we will have to reduce the inherent complexity and dynamics to a practicable level: "Models that exceed a certain level of complexity can never be tested"

(Porter 1991,

116). However, without testing,

science remains

purely

metaphorical and "difficult to distinguish from witchcraft" (McKelvey 1999, 21). While we thus cannot provide an exhaustive model, we can establish a framework that adds more complexity step-by-step by putting together contributions from different perspectives: The overall objective is to develop a model leveraging the contributions of different strategy perspectives for a state-of-the-art view of the determinants of firm performance. In order to concretize and break down the overall objective into manageable tasks, we continue to specify four subordinate research objectives. Above we

mentioned

numerous variance decomposition studies showing that

performance is simultaneously influenced by industry, corporate and business level factors (Le. McGahan I Porter 2002). The first step would thus be to integrate different

4

determinants that have been found important in explaining performance. The literature has suggested that virtually hundreds of variables may have an effect on performance, including such hard-to-grasp variables as luck or chance (Porter 1991, 98; McKelvey 1999, 15; Jarillo 2003, 23). As it is impossible to capture all determinants, the objective should be to include those factors that have been repeatedly found to be important: The first research objective is to integrate the most relevant determinants of firm performance from competing perspectives into a single model. Variance decomposition studies, as well as other research (Le. Henderson / Mitchell 1997, Stimpert / Duhaime 1997, Phelps 2001) have shown important interrelations between individual determinants that underlie performance. The literature has suggested that the full complexity of non-linear ("chaotic") interaction between variables - including feedbacks, side effects, and time delays - may exceed our cognitive capacity to understand that complexity (Dooley / Van de Ven 1999, Sterman 2001). As it is impossible to capture the entire complexity of dynamiC interaction, the objective should be to include those interrelations that have been repeatedly found to be important: The second research objective is to capture the most relevant interrelations between individual determinants in our framework. Studies have shown that determinants of firm performance change and evolve over time (Le. Nelson 1995, Teece et al 1997). While we integrate a dynamic dimension, our analysis has to be limited to a restricted period: The third research objective is to add a dynamic dimension encompassing the development over several years to our framework. The reduction of complexity along several dimensions has the advantage of designing a model that relies on proven theories and remains empirically testable. The model's applicability in an empirical context allows the concrete analysis of selected dynamic industries. We have chosen the global media industries as a particularly interesting example of a dynamic industry to evaluate the model's usefulness for strategic industry analysis:

5

The fourth research objective is to apply the developed framework to global media industries to gain a comprehensive view of the effects underlying performance.

1.3

Foundations for a Dynamic Model

We have seen above that strategy research remains highly fragmented. Furthermore, the different approaches are based on conflicting underlying assumptions. In order to integrate valuable ideas from different perspectives, we first need to establish a common base of fundamental assumptions, or a frame of reference for our model. This frame of reference has to integrate contrary assumptions on three levels of abstraction: Philosophy of Science: First of all, different perspectives are based on differing epistemological assumptions. While traditional approaches are deeply rooted in positivism and realism, more recent dynamic approaches - in particular evolutionary theory - are situated in the constructivist field (Malik I Probst 1982, Nelson 1995, Phelan 2001, Fagerberg 2002). In order to enhance practice and advance theories, it is essential for our study to build frameworks and rely on in-depth empirical research (Porter 1991, 98). At the same time we are aware that "no theory ever attempts to represent or explain the full complexity of some phenomenon" (McKelvey 1999, 15).

Rather than searching for generic truths we should thus focus on establishing "middle range theories" with a high degree of plausibility (Farjoun 2002, 563). Consequently, we

adopt the analytic realism research position as the middle ground between the extremes of positivism and constructivism (Seale 1999). Several models with differing degrees of complexity, based on different assumptions, will be established. While the less complex models allow for formal theorizing, we apply appreciative theorizing - a less formal and more abstract way of theorizing as described by Nelson and Winter (1982, 46) - to the model with higher complexity. Change: At a more concrete level, different perspectives are conflicting in their understanding of change. Traditional approaches to strategy assume that after a rapid tussle, competition, firm behavior and the resource base arrive at an equilibrium. Researchers' interest thus concentrates on systems at rest rather than on the process

6

that led to the current conditions (Foss et al 1995, Barnett / Burgelman 1996). Conversely, evolutionary theory treats competition as an ongoing process that is never at rest. Equilibrium conditions are rarely if ever reached (Schumpeter 1942, Nelson / Winter 1982). For this study we promote an alternative understanding of change: punctuated equilibrium theory (Eldredge / Gould 1972). This view proposes relatively long periods of

relative stability (equilibrium periods) punctuated by relatively short bursts of fundamental change (revolutionary periods). The punctuated equilibrium theory has gained increasing support from empirical studies both in evolutionary biology and management (Miller / Friesen 1980, Tushman et al 1986, Pettigrew 1987, Romanelli / Tushman 1994). For instance, McNamara et al (2003, 275) conclude that market development "reflects a punctuated equilibrium process". From a theoretical perspective, both Schumpeter

himself and more recent strategy scholars treat equilibrium and disequilibrium theories as complementary rather than contradictory (Barney 1986, Porter 1991, Jacobsen 1992, Foss et al 1995, Nelson 1995, Barnett / Burgelman 1996). Punctuated equilibrium theory builds the bridge between the old rivals, equilibrium and disequilibrium theories, thus promoting cross-fertilization between these camps. Complexity and Dynamics: A third line of demarcation not only separates traditional from dynamic approaches, but even runs across different dynamic perspectives: the understanding of complexity and dynamics. Contingency approaches analyze the dynamics between elements of the system at rest, while evolutionary theories define dynamics as a process over time. Drawing on both complexity and evolutionary theory, we present three distinct constructs that describe the different facets of complexity and dynamics. Detail complexity (I) refers to the number of variables that a model takes into account (Senge 1990, Devaney 1993). The more variables (e.g., determinants of performance) we consider, the more realistic, but also the more complex our model. However, even if we were to take all relevant determinants into account, we would still fail to capture the full complexity of the problem. Dynamic complexity (II) refers to a system that captures interactions between causal

variables (Senge 1990). The complexity increases whether we assume simply linear interactions ("periodic") or even non-linear interactions ("chaotic") between variables 7

(Dooley Nan de Ven 1999). While dynamic complexity provides a reasonably accurate picture of how the current system works, it does not explain the development of the system over time. Evolutionary dynamics (III) capture this longitudinal perspective (Nelson 1995, Priem / Butler 2001, Black / Farias 2000). All three distinct and complementary perspectives on complexity and dynamics will be required for a truly comprehensive approach to strategic analysis. Based on the reasoning above, we now continue to establish our model's overall frame of reference (see Figure 1).

Dynamics

Full

~r-----------------------------------~ Disequilibrium

Appreciative Theorizing Evolutionary

Formal Theorizing Equilibrium System

Static

eL-______r -______- .__~_____r~------~ e Simple Detail Dynamic Full ~

Figure 1

Complexity

Frame of Reference

The basic rationale behind the frame of reference is the paradox that increasing levels of complexity and dynamics lead to a better understanding of reality, but at the same time reduce our ability to explain this reality in general and in straightforward theories. A low to medium degree of complexity (simple or detail) can be captured in research models based on formal theorizing (i.e. M1 / M2). Moving towards higher complexity, we quickly reach the limits of formal theorizing. A low degree of dynamic complexity (i.e. M3) can still be captured through formal models. However, higher degrees of dynamic complexity require less formal "appreciative" theorizing, while full complexity eludes any attempt at

8

theorizing due to the involved random elements. The same effect is true for increasing dynamics. A low to medium level of system dynamics (i.e. M1-M3) can be captured through formal theorizing. Evolutionary dynamics, however, require the abolishment of equilibrium assumptions, thus quickly reaching the frontiers of formal theorizing, in particular if combined with some degree of complexity. A simultaneous acceptance of medium to high levels of both dynamics and complexity can only be achieved through appreciative theorizing (i.e. M4). Full dynamics, especially in combination with full complexity, elude any attempt at theorizing. Our recommendation for the establishment of dynamic approaches to strategy is thus to combine multiple models along the axis M1-M4. Lower-level models allow more formal theorizing that deploy the rich body of existing theories and methodologies that rely on equilibrium assumptions. Based on this foundation, additional appreciative theorizing allows "journeys" into higher complexity and dynamism based on more innovative research methods. The combination of two or more such models allows cross-fertilization between a large variety of existing theories and methods. It is also the only practical approach with which to achieve our research objective: the establishment of a more integrative, complex, and dynamic approach to strategic analysis. A single model could never provide the same combination of complexity and dynamism on the one hand, and validity and generalizability on the other.

1.4

Model Development & Hypotheses

We propose a framework for dynamic strategic analysis that unites three sequential models (see Figure 2): The integrated (I), the complex (II), and the evolutionary (III) model. The Integrated Model: The integrated model responds to the first research objective by

addressing the problem of fragmentation through the integration of the most important competing perspectives of strategy theory. The model's main hypothesis suggests that the integrated model improves on any single-perspective approach.

9

(1) The Integrated Model

Figure 2

(2) The Complex Model

(3) The Evolutionary Model

Dynamic Research Framework

Based on a thorough literature review, the most important determinants of company performance are derived from different perspectives. More exactly, we analyze the influence of industry structure, firm capabilities, business strategy, and corporate strategy on company performance within a single model. Based on the 10 literature (Bain 1956, 1959), industry structure is represented by three variables: (I) industry concentration (Bain 1956, Scherer / Ross 1980), (II) entry barriers (Bain 1956), and (III) market growth rate (Aaker / Day 1986, Buzzell / Gale 1987). Corporate strategy (Chandler 1962) is captured by three determinants: (I) product diversification (Gort 1962, Markham 1973, Dess et al 1995), (II) international diversification (Caves 1971, Geringer et al 1989, Hitt et al 1997), and (III) vertical integration (D'Aveni / Ravenscraft 1994, Grant 2002). Competitive strategy is reflected by three variables based on Porter's (1980) "generic strategies": (I) cost leadership, (II) differentiation, and (III) balancer strategy. Finally, firm capabilities (Barney 1986, Rumelt 1987) are explained by means of five constructs: (I) market orientation (Kohli / Jaworski 1990, Slater / Narver 1990), (II) reputation (Fombrun / Shanley 1990, Roberts / Dowling 2002), (III) organizational culture (Denison 1990), (IV) organizational learning (Hurley / Hult 1998, Calantone et al 2002), and (V) innovativeness (Vasquez et al 2001, Calantone 2002). All fourteen variables are

10

presumed to be positively linked to performance. 3 All relevant levels of analysis (industry, strategic business unit, corporate) are covered.

Competitive Strategy

Differentiation Cost Leadership Balancer Strategy

::J

Innovativeness

tJ)

Market Orientation

m Firm Capabilities

Org. Learning Reputation Corporate Culture

~o e-O

(.)

Figure 3

The Integrated Model

The integrated model focuses on the long periods of relative stability. With regard to the frame of reference developed above, the integrated model remains clearly equilibriumbased thus understanding change as a punctuated equilibrium. It attempts to reach a

better understanding of the factors underlying success under the current equilibrium conditions (in Porter's (1991, 105) words, "the cross-sectional problem"). Ignoring system dynamics and longitudinal developments, the model merely strives to grasp detail complexity (Senge 1990). This reduction of real world complexity allows a high degree of formal theorizing (Nelson / Winter 1982) with testable hypotheses on cause-effect

relationships. The Complex Model: The complex model addresses the second research objective by extending the initial model to the most relevant interrelations between individual

3

With the notable exception of international diversification, where a curvilinear relationship was assumed.

11

determinants. The main hypothesis suggests that a model capturing interrelations between determinants improves on any less complex and deterministic models. Drawing heavily on contingency literature, a set of hypotheses on relations between determinants has been established. For business strategy and market structures, we suggested interrelations between a cost leadership strategy and industry concentration (Dutton 1982, Kim / Lim 1988, Miller 1988), as well as between a differentiation strategy and market growth (Hambrick 1983, Miller / Friesen 1984, Kim / Lim 1988). As far as business strategy and capabilities are concerned, we assumed interrelations between a differentiation strategy and market orientation (Slater / Narver 1996, Morgan / Strong 1998, Matsuno / Mentzer 2000, Kumar et al 2001), innovativeness (Snow / Hrebiniak 1980, Hambrick 1983, Vazquez et al 2001), and organizational learning (Slater / Narver 1995, Calantone et al 2002). For business strategy and corporate strategy, we suggested interrelations between product diversification and a cost leadership strategy (Bettis 1981, Hoskisson / Hitt 1988, Hill / Snell 1989), international diversification and a differentiation strategy (Young 1987, Roth / Morrison 1992), as well as vertical integration and a balancer strategy (Armour / Teece 1980, Prabhu / Harrison 1992, D'Aveni / Ravenscraft 1994). Regarding corporate strategy and market structure, we hypothesized a relationship between product diversification and industry concentration (Rumelt 1974, Bass et aI1977, Montgomery 1985). The link between corporate strategy and capabilities is represented by the interrelations between product diversification and organizational learning (Chatterjee / Wernerfelt 1991, Pennings et al 1994), as well as between vertical integration and both innovative ness (Armour / Teece 1980, Buzzell 1983, Heeb 2003) and organizational learning (Armour / Teece 1980, Sorenson 2003). Finally, the relationship between capabilities and market structure is represented by a suggested link between entrepreneurial culture and market growth (Gordon 1991, Kotter / Heskett 1992, Chatman / Jehn 1994, Sorensen 2002). With regards to the frame of reference, the complex model shares equilibrium assumptions with the integrated model. However, the model moves from detail complexity towards dynamic complexity (Senge 1990). Assuming linear relationships between determinants, the inherent complexity of the model is reduced to a level that still allows formal theorizing (Nelson / Winter 1982) that deploys cross-sectional analysis and hypothesis testing. 12

The Evolutionary Model: The evolutionary model reflects the third research objective.

capturing the dynamic development of determinants over time to some extent. More exactly. we search for industry-wide characteristic patterns of dynamic change in key strategic variables underlying firm performance. We then analyze the effects of such changes on the development of firm performance. The main hypothesis suggests that the dynamic model improves on any purely equilibrium-based approaches. Rather than describing the current equilibrium. the focus of the evolutionary model is on the disequilibrium period of revolutionary change and the process leading to the current conditions. Shedding equilibrium assumptions. we refer to appreciative theorizing (Nelson / Winter 1982) for this model. This allows some degree of detail and dynamic complexity to be maintained besides the main focus on evolutionary dynamiCS. In the

tradition of inductive (or exploratory) theorizing. we strive to empirically identify characteristic patterns of change at the industry level. Based on existing research and the criticism thereof. we developed a framework that integrates opposed schools of thought and captures some of the complex and dynamic character of real life strategizing. The models come close to what strategy managers have always had to accomplish in practice: the coherent integration of many complex and dynamic facets of strategy and competition.

1.5

Research Methodology

In accordance with this study's fundamental research position. we applied a multilevel research method combining cross-sectional and longitudinal methods. We had first of all utilized cross-sectional methods to test both the integrated and the complex model in our framework. The established hypotheses on relationships between variables were empirically tested. In order to capture the dynamic development of variables over time. we continued to deploy longitudinal methods to the evolutionary model. The required data for all three models were collected in a field study using survey methodology (Creswell 1994. Churchill 1999).

13

In line with our fourth research objective, the models were tested in the dynamic global media industries. The global media industries were represented by an area sample of 671 media companies (Churchill 1999). The sample contains all media companies located in three major global markets (United States, Germany, and United Kingdom), operating in four main industry segments (Broadcasting and Cable, Movie Pictures, Print and Publishing, and Music), and having realized revenues of more than US$ 50 million in 2002. Questionnaires were sent to target informants at the Strategic Business Unit (SBU) level of these companies. All informants are part of the senior management team, mainly members of the board and heads of the marketing or strategy department. The questionnaire had been developed deploying measures and scales for all variables that had been used and tested in prior research. Detailed feedback from two pretests had been used to further refine and improve the questionnaire. The mail survey was conducted during September, October, and November 2003. The collected data were then analyzed deploying a multitude of statistical methods, including regression analysis based on SPSS 11 and structural equations modeling based on LlSREL 8.3 (Jbreskog / Sbrbom 2001, Brosius 2002).

1.6

Structure of the Study

In the following Chapter II, we review the extant strategy literature on the determinants of firm performance in detail. This literature review establishes the theoretical foundations of the integrated model of our framework. Chapter III reviews criticisms of traditional strategy literature and introduces the findings

of more recent "dynamic" approaches to strategy. This part prepares the groundwork for the complex and evolutionary models in our framework. Leveraging both the traditional and the dynamic theories reviewed above, we continue to develop the research model and the related hypotheses. Chapter IV outlines the research methodology applied to empirically evaluate the

developed framework in more detail. The chapter describes the research design, the

14

questionnaire development, the survey implementation, and the methods applied for the data analysis. Chapter V is dedicated to the research findings, and presents the multitude of empirical findings on the integrated, complex, and evolutionary models in detail. All findings are discussed in terms of both their theoretical implications and their relevance for (media) industry practitioners. As mentioned above, Chapter VI summarizes the findings and draws conclusions from our study. The limitations of the study are furthermore discussed, as well as directions for future research given and the study closes with concluding comments.

15

II.

Literature Review

Contrary to the propositions of the neoclassical theory of perfect competition, existing empirical evidence suggests that firms earn differential returns (Rumelt 1991). The identification and analysis of factors determining firm success or failure have been at the heart of strategic management since their inception four decades ago. Various conflicting definitions of firm success coexisted in the early years. Michael Porter's seminal work "Competitive Strategy" (1980) is largely merited with focusing the practice on superior

and sustainable (mostly financial) performance as the central goal of firm activity. What are the underlying determinants of superior performance? Early scholars of strategic management, such as Learned et al (1965), defined three essential conditions for firm success that determined the research program of the discipline for the following three decades: 1. A company has to develop a consistent set of goals and an overarching strategy that ensures the implementation of these goals. 2. These goals and strategies need to be aligned with the dynamic environment of the firm and the external opportunities and threats in the industry.

3. The firm must focus on the creation and exploitation of firm-specific distinctive competencies.

Further research into each of these central determinants - strategy, environment, and competences - was conducted largely independent of one another. Research in the 50s and 60s was dominated by the industrial organization school of thought (8ain 1959). According to this school, superior firm performance can only be achieved in industries with a highly concentrated structure facilitating collective agreements between incumbents and in industries with high entry barriers. While clearly based on the industrial organization's theoretical foundation, scholars of strategic management shifted the research focus during the 70s and 80s from the industry towards the individual firm and its strategic actions. They ascribed a central role in firm success to firms' strategic conduct at both the corporate and the business unit level. The late 80s and the 90s were

17

then dominated by the resource-based perspective and the analysis of unique firmspecific resources and capabilities. Our first research objective of establishing a comprehensive model to analyze the dynamics underlying firm performance first and foremost requires an understanding of the causality under equilibrium conditions. According to Porter's postulate, it is impossible to deal analytically with the process "without a rather specific understanding of what underpins a desirable position" (Porter 1991, 105). In order to reach this specific

understanding, we review the sequential development of research on the individual determinants of firm success in the remainder of this chapter. The analysis is segmented into the four essential debates described above: Industrial Organization (11.1), Corporate Strategy (11.2), Business Strategy (11.3), and the Resource-Based View (111.4). This

chapter prepares the ground for both the exploration of more recent dynamic approaches and the establishment of the research model in Chapter 3.

11.1

Industrial Organization

Industrial

organization's

central

approaches

can

be

illustrated

through

recent

developments in the European Pay TV industry. We thus start our review of industrial organization with a short discussion of the differences between the British and the German Pay TV markets. The French newspaper Le Figaro recently labeled Rupert Murdoch, CEO of News Corp, owner of the satellite network BSkyB, as the "king of European Pay TV'.4 Murdoch should be more than pleased by the development of his UK Pay TV operations. However, his stake in the German Pay TV network Premiere remained the black spot on the balance sheet. Premiere had accumulated losses of more than US$ 4 billion and continued to lose US$ 2 million a day. Germany's first and only Pay TV channel may be an extreme example, but nevertheless representative of the majority of the country's broadcasters. Ten out of thirteen channels haven't been able to amortize initial investments and continue to incur losses. The overall market has a negative balance

4

Le Figaro (10106/2002)

18

with accumulated losses of close to US$ 5 billion. At the same time broadcasting has enjoyed a stable profitability in other countries, for instance, the United States. 5 Observers eager to discover the reasons for this market underperformance will certainly embark on an analysis of the specific realities of the German television market. Underperformance of an entire industry when the same segment abroad is profitable points to differences in the underlying industry structure. Scholars of industrial organization emphasize these structural differences between industries as crucial in explaining the respective industry performance (Bain 1959). In the remainder of this segment, we first concentrate on a brief (I) Introduction to Industrial Organization to reach a common understanding of the argumentation of this school of thought. Once the basic parameters have been established, we present the main (/I) Elements of Industry Structure, used to describe different industries, in more detail. Next, the considerable (11/) Empirical Research is reviewed to establish and test the link between individual

elements of market structure and performance. Finally, we conclude this section with a (IV) Criticism and Conclusion part, both summarizing our review of industrial organization

and providing the transition to more recent strategic studies presented in the subsequent section of this chapter.

Introduction to Industrial Organization

The central approach of industrial organization is the structure-conduct-performance (SCP) paradigm. Put simply, the paradigm suggests that firm performance depends significantly on the attributes of the industry environment in which it operates. There is a unidirectional flow of causality from industry structure, through the behavior or conduct of firms, to the collective performance of the industry (see Figure 4). The industry is treated as the unit of study. This paradigm draws largely on the pioneering work of Edward S. Mason (1939) and on major contributions advancing the approach made by Joe S. Bain (1951, 1959).

5

Examples and figures are as of late 2002. During 2002, Premiere was sold and the new management managed to significantly reduce operating losses in 2003. However, the accumulated debt remains a major burden. Improvements have also been realized at other channels, notably at Pr07Sat1.

19

Industry Structure

Figure 4

f----

Conduct (Strategy)

Performance

The Structure-Conduct-Performance Paradigm

Industry structure is defined as the relatively stable economic and technical dimensions

of an industry that provide the context in which competition occurs (Bain 1972). This industry structure depends on basic conditions in both the market demand and supply side. The most important elements of industry structure identified by early scholars of industrial

organization

were

barriers

to

entry,

industry

concentration,

product

differentiation, and scale economies (Bain, 1959). Of these, the notion of industry concentration received the greatest attention by far, particularly in highly concentrated markets including various segments of the media industry. More recent studies have stressed the importance of additional characteristics of market structure, most notably the overall elasticity of demand, market growth, exit barriers, and diversification (Le. Rumelt 1974, Caves / Porter 1977, Porter 1981, Baumol et al 1982, Aaker / Day 1986). The most important elements of market structure are discussed in more detail in the next section. Conduct can be understood as the firm strategy, such as the decision on variables such

as price, product quality and capacity, marketing and advertising, research and development. Firm conduct has been largely assumed as transitory or unimportant unless based on scale economies, which were generally found to be insubstantial. This can be explained by the belief that because structure determines conduct, which in turn determines performance, conduct is a dependent variable that can be ignored (Bain 1959). Most studies thus focus on the direct impact of industry structure on performance. Performance was defined with a micro economist's focus on the extent to which a firm

departs from the Paretian allocative efficiency ideal. The main interest from this perspective is the improvement of welfare from a social viewpoint instead of improving a single firm's performance. This approach provides the basis for analysis of the welfare effects of misallocation (e.g. through monopolistic industry structures).

20

The social interpretation of performance in industrial organization points to the theoretical rationale underlying this school of thought: oligopoly theory. The central hypothesis of oligopoly theories is that there will be an increase in price (and thus firm profit rates) with increased industry concentration (Cournot 1838). The Cournot model predicts a decline in total industry profit in line with a rise in the number of firms present in the marketplace. This effect is further reinforced by the increasing probability of collusive behavior in concentrated markets. Chamberlin (1933) established the concept of the critical concentration ratio: firms that normally act independently begin to act collusively if a certain degree of concentration is reached. This triggers further concentration and thus collusive price increases. This theoretical background partially explains the constriction of industrial organization by industry structure and particularly the focus on market concentration. The key concern for consumer welfare, combined with the necessity to simplify empirical research, explains the neglect of firm conduct (Teece 1984). More recent studies of industrial organization have brought firm conduct back into the scope of research. This work is reviewed later in this study, together with literature from the business policy field that traditionally busies itself with the strategic behavior of the firm.

Elements of Industry Structure

The introduction above reveals the centrality of industry structure for industrial organization. Numerous scholars have analyzed a multitude of individual factors that may (or may not) constitute and shape different industry structures. Porter's well known Five Forces of Competition framework (1980) clusters over twenty structural variables

into five forces

influencing competition

and

profitability.

Other scholars

have

acknowledged Porter's comprehensive approach, but stress the predominance of a smaller number of variables. Empirical studies have strongly supported this argument, as we will see in the following section. A number of elements have been identified as core to explain structural differences between industries (Capon et al 1990). We briefly review the most relevant variables of market structure, notably industry concentration, entry barriers, and market growth rate. Industry Concentration affects the degree of rivalry between the established

competitors in an industry. Concentration refers to the number and size distribution of

21

firms competing in a market. The degree of seller concentration is one determinant of the character and intensity of competition in the industry, establishing perfect competition, monopoly, or anything in between (Bain 1959). A monopolistic position allows the single firm in the market to charge high prices. If the market is dominated by two or a small group of firms, there is still significant potential for collusion in pricing decisions (Scherer / Ross 1980). As the number of firms present in the market increases, price coordination becomes increasingly difficult and there is a growing likelihood that one firm will adopt a price-cutting strategy. The hypothesis is thus a positive impact of market concentration on performance (Mann 1966). The common measurement for industry concentration is the concentration ratio, or the combined market share of the leading producers. In order to measure concentration, it is crucial to define the relevant market along two dimensions: products and regions. In product terms, it is necessary to identify whether the products or services of the firms are substitutes for one another. In regional terms, it is important to identify which group of firms competes for the same buyers. Once the target segment has been identified, concentration can be measured through a number of aggregates like business assets, business income or sales, and labor force employed. Over time, concentration will vary across industries with the height of the barriers to entry: the more difficult entry into an industry, the higher the seller concentration remains. We will now take a closer look at the different barriers to entry. Entry Barriers determine the risk of new competitors entering the market. Generally, if

an industry realizes a return on capital in excess of its cost of capital (e.g. due to a high industry concentration), that industry attracts new firms to the market. Unless the entry of new firms is hampered, the profit rate will fall towards a competitive level. Entry barriers are market conditions allowing concentrated sellers to increasingly raise prices above this competitive level without attracting the entry of new competitors (Bain 1956). An explanation for this lack of entry is the costs that have to be borne by a firm that seeks to enter an industry, but not by firms already active in the industry (Stigler 1963).

22

Bain distinguished three entry barriers categories: (I)

Absolute Cost Advantages: Established firms enjoy a lower average cost than

new entrants due to the acquired know-how, governmental regulation, or privileged access to raw materials. Governmental regulation plays a major role. Some economists claim that governmental regulation including licenses, patents, and copyrights; remain the only effective barriers to entry. One example is the absolute cost advantage that fee-financed public broadcasters enjoy over new entrants (Wirtz 1994). Private companies willing to enter the market have to rely on advertising as the single source of financing allowed by governmental regulation. (II)

Economies of Scale: In various industries the unit cost of a product falls as the

volume of production goes up. An example is the so-called "first copy cost effect' (Owen et aI1974): In the broadcasting industry, the cost of producing a television movie is independent of the number of viewers. Significant economies of scale can be realized by increasing viewer numbers. In the Pay TV business, this effect renders it extremely difficult for new entrants to realize benefits while subscriber numbers are still low. (III)

Product Differentiation: Established firms may possess the advantage of brand

recognition, reputation, and customer loyalty. Customers might be willing to pay a premium for the established brand. New entrants must spend heavily on advertising and promotion to gain the same level of brand awareness. Established TV channels enjoy the advantage of well-known presenters and programs. New entrants have to spend significant sums to poach famous presenters from established channels, or to promote new faces (Neumann 1998).

Additional entry barriers, especially capital requirements, have been identified since Bain's seminal work (Porter 1980): (IV)

Capital Requirements: The initial capital cost to enter an industry can be so large

that only a limited number of investors can bear the required investment. 23

More recent research introduced the concept of exit barriers, closely related to the entry barrier "capital requirements" introduced above (Baumol et al 1982). Exit barriers are basically sunk costs invested in the business that cannot be readily recovered by the firm. It has been argued that exit barriers can also be viewed as entry barriers. The argument is that when exit is costly (due to sunk capital costs), firms are less likely to enter the market in the first place (Caves / Porter 1977; Eaton / Lipsey 1980). In this light, Premiere's accumulated billion-dollar losses mentioned above will certainly deter most companies from investing in the Pay TV market. However, the additional effect is that, due to the high investment that will be lost, competitors remain in the market despite ongoing losses. Premiere is once again a good example of this effect: the shareholders are willing to invest further money to keep operations running despite incurring ongoing high losses over years. Market Growth Rate: High-growth markets are generally regarded as more attractive

than mature or even declining markets. However, growth markets represent both opportunities and substantial risks and challenges (Aaker / Day 1986). It's easier to gain and sustain market shares, and there is less price pressure in growing markets. On the other hand, high-growth markets are more likely to attract competition, require high investments, and are often characterized by fast-paced change. All together these effects nevertheless seem to have a combined positive impact on performance (Buzzell / Gale 1987). This hypothesis has been largely supported by empirical research, as we will see in the following chapter.

Empirical Research

Quantitative empirical research has always been one of the strengths of industrial organization. Bain (1951; 1959) was the first to support the SCP paradigm with empirical evidence. Since then, hundreds of studies have been published on diverse elements of market structure and their influence on performance. Capon et al (1990) provide an overview in their meta-analysis of determinants of financial performance. In this section, we first summarize the overall results of empirical studies on the structure-performance relationship. Next, we assess the relevance of individual elements of industry structure on performance. Finally, we briefly review works pertaining to the media industries.

24

The General

Structure-Performance Linkage:

Empirical

studies have clearly

challenged the generic approach of early scholars of industrial organization in explaining performance as a mere consequence of industry structure. However, a consensus has emerged that industry structure does indeed have some influence on performance. A number of scholars have tempted to measure this influence. Schmalensee (1985) found that differences in market structures account for roughly 20% of the variance in business unit performance and for at least 75% of the variance in industry returns. He concluded that industry effects are far more important than firm-specific factors. Rumelt (1991) contradicted these results and stated that business-unit effects strongly outweigh industry effects on profitability. Long-term industry effects account for 8% in this study (plus another 8% for short-term industry effects). A recent study by Hawawini et al (2003) supports Rumelt's results. Other researchers stressed the importance of both industry and firm effects and their independent effect on performance (Hansen I Wernerfelt 1989; Capon et al 1990). In conclusion, structural variables are insufficient to explain firm performance entirely, but "the overall importance of these factors is beyond dispute" (Hansen I Wernerfelt 1989; 400). Hence, a comprehensive strategic analysis has to comprise a thorough research into market structures. After having established the general importance of market structure in the context of firm performance, we now turn towards the individual elements of market structure. What are the most important elements and what is their impact on economic returns?

The Role of Industry Concentration: As noted above, industry concentration has attracted unparalleled attention: Capon et al (1990) list 100 empirical studies analyzing the link between concentration and performance. Most of these primarily bivariate studies found a significant and positive relationship. Other meta-surveys, such as a review of 45 studies of the US banking industry (Gilbert 1984) and an earlier review of 46 cross-sectional studies (Weiss 1974), came to the same result. Despite a few studies with insignificant or even negative results, it can be reasoned that there is sufficient evidence of a significant and positive bivariate relationship between structure and performance. Whereas the results of bivariate studies are definite, more complex multivariate studies show a slightly different picture. Some researchers found that performance was less

25

influenced by consolidation than by another independent variable: firm efficiencies (Le. Brozen 1971, Smirlock 1985, Evanoff / Fortier 1988). These scholars - later called the Efficiency School - explained concentration as a mere result of the disproportionate growth in market share of the most efficient companies in the market. However, more recent studies comparing both variables clearly support the importance of consolidation above that of efficiency (Montgomery / Wernerfelt 1991, Molyneux / Forbes 1995; Bajtelsmit / Bouzouita 1998). Without entering deeper into this debate, one important conclusion can already be drawn: While exercising significant influence on performance, market consolidation is in turn influenced by other factors such as firm efficiency. Market consolidation is an essential, but insufficient element of strategic analysis. The Importance of Entry Barriers: The analysis of entry barriers has attracted almost

the same attention as the research into market concentration. Capon et al list far more than 100 studies on entry barriers including early work by Bain (1956). A significant and positive effect of entry barriers on performance has been broadly supported through these studies. For instance, Chappell et al (1983) analyzed entry barriers for 63 industries by using a regression model and found sound support for the hypothesized positive correlation between entry barriers and profitability. Furthermore, studies have confirmed Bain's claim that entry barriers influence performance independent of the degree of market concentration (Mann 1966). Most early studies distinguish between a "moderate/low", "substantial", or "very high" overall level of industry barriers. The researcher collects data on a number of industries and classifies these, based on his expert assessment (Bain 1956, Mann 1966, Weiss 1971). More recent studies have focused on individual types of barriers and used empirical proxies to measure entry barriers (Martin 1979). Most attention has been paid to capital requirement (capital/sales ratio), product differentiation (advertising/sales ratio), and economies of scale (measures of minimum efficient scale; cost-disadvantage ratio). The meta-analysis of Capon et al (1990) lists numerous studies that analyzed and confirmed the positive influence of these individual barriers on performance.

26

The Impact of Market Growth: The analysis of the relationship between market growth and performance is more recent than the two classic research fields of concentration and entry barriers described above. Like Bain, early scholars of industrial organization attributed a subordinate role to market growth. The influence of market growth was often seen as ambiguous (Aaker / Day 1986). However, by 1990 more than 80 studies had analyzed this relationship and the vast majority had found a significant and positive correlation (Capon et al 1990). Studies on the Media Industries: Media economics is an emerging area at the intersection of communication and economic studies. Scholars of media economics analyze media companies and markets from an economic perspective (Picard 1989). The existing literature is strongly focused on examining the impact of market structures on company performance leveraging the SCP paradigm of industrial organization (Wirtz 1994, Albarran 1996, Sjurts 1996, Alexander et al 1998). Most studies are limited to individual segments of the media industry, such as that of broadcasting (Owen / Wildman 1992, Wirtz 1994, Neumann 1998), print and publishing (Heinrich 1994a, Lacy / Davenport 1994, Sparks / Sparks 1995, Altmeppen 1996), and movies (Dale 1997, Houcken 1999, Litman 2000). Other studies have focused on the single aspects of the market structure. Researchers have predominantly concentrated on analyzing the growing market concentration in individual media segments such as newspapers and magazines (Picard 1988, Lacy / Davenport 1994, Neiva 1996, McQuivey / McQuivey 1998, Roper 2000, SchOtz 2000), broadcasting (Owen et al 1974, McFayden et al 1980, Litman 1985, Patzold / Roper 1999), and cable television (Chan-Olmsted / Litman 1988, Atkin 1994, Chan-Olmsted 1996, Ford / Jackson 1997, Kelly 2000, Chen 2002). The dominant positions of a small number of media conglomerates and the resulting negative societal effects have been repeatedly criticized (Tunstall / Palmer 1991, Smith 1991, Picard 1996, Albarran / ChanOlmsted 1998). Compaine and Gomery (2000) gained significant attention with the analysis of media concentration in their book Who owns the media? The impressive number of studies on market concentration and other structural elements shows the importance that the media economics literature attributes to market structure in explaining the performance of media companies. While this body of work constitutes a

27

valuable basis for our later research into the media industry, it obscures the limitations of the media economics discipline to a certain extent. As we will see later in this chapter, media economics has maintained an exclusive focus on market structure while largely neglecting variables at the firm level, including strategic actions and firm capabilities.

Criticism and Conclusion

From today's perspective, the structure-conduct-performance paradigm and the classic industrial organization literature can be regarded as the first systematic model for strategic analysis. For the first time analytical techniques were tested in hundreds of studies to obtain empirical evidence on factors underlying performance. The model proves that not all industries are equal in terms of their profitability (Porter 1981 ). Despite the generally acknowledged merits of industrial organization, considerable criticism has been raised, particularly from the business policy field. Four key sets of criticism will be discusses in the remainder of this section: Static View of Industry Structure: The classic industrial organization literature defines

industry structure as a given and definitional stable variable. In reality, frequent changes in industry structure can be observed, as we can see from the current development in the media industry. For instance, the US currently dropped the ban on cross-media ownership in television and print. This will have a dramatic impact on the current industry structure in both segments, lowering barriers to entry while promoting increased competition. In addition, as stated before, industry structure is by no means "given". Firms can fundamentally influence the structure through their actions (Porter 1981). However, firms cannot always change structure. More recent research shows the difficulty in actively erecting entry barriers (McWilliams I Smart 1993). Since all firms in an industry profit from entry barriers, firms that invest in the construction of these barriers bear all the costs while others enjoy the benefits - a classical "free rider" situation. Industry as Unit of Analysis: Industrial organization is clearly limited to the industry as

a unit of analysis. From a business policy perspective, industry development cannot be explained without analyzing firm and business unit behavior. Moreover, the industry

28

focus motivated industrial organization researchers to analyze performance purely from a social viewpoint. Therefore, industrial organization is not very useful in explaining performance differences between companies within the same industry.6 Deterministic Theory and Mixed Empirical Results: Industrial organization literature regards industry structure as the single determinant for performance. This hypothesis of a high correlation between structure and performance, initially supported by bivariate empirical studies, has faced increasing criticism. Researchers from the emerging Efficiency School argued that the effect of other plausible variables (like strategy or firm

competencies) omitted by bivariate studies are misleadingly attributed to market structure (Demsetz 1973). More recent multivariate studies that include other variables (such as firm growth) seem to confirm this view. Different studies have led to controversial results, but the overall picture is clear: while structure has a partial influence on performance, it cannot explain performance entirely (Capon et al 1990). Homogeneity of all Firms in an Industry: The most obvious weakness of the early industrial organization literature is the assumption that all firms within an industry are homogenous (Rumelt 1991). Simple observation of the global media industries reveals that firms competing in the same segment show significant differences in performance. For instance, the six leading competitors in the German print and publishing market realized very different results in 2001: While three companies (WAZ., Bauer, and Burda) reported market share growth and positive operating profits, three others (Springer, Gruner & Jahr, and Holtzbrinck) incurred operating losses. This has been partly explained by the significant losses incurred through high investments in the online activities of the latter group, whereas the former group had followed a more cautious online strategy (Roper 2002). This example illustrates that performance differences between firms cannot be fully explained by industry structure (they all operate in the same environment) or simple differences in size (the second group too consists of larger firms). A similar conclusion can be drawn from most industries. This observation

6

Industrial organization's traditional SCP paradigm includes strategic actions at the firm level (the "Conducf' variable in the paradigm). Conversely, this variable has often been excluded from further analysis and was often treated as a simple intermediate variable. However, later strategy models such as the strategic group concept or the generic strategy types (in particular Porter's work) are widely based on the industrial organization literature.

29

significantly challenges the notion of firm homogeneity within the same industry as postulated by industrial organization economists. More recent work in the tradition of industrial economics thus relativizes the notion of firm homogeneity within the same industry. The most prominent approach is the concept of strategic groups (Porter 1976, 1979, 1980). The central argument of the strategic group concept is that sets of firms with resembling asset configurations within an industry pursue comparable strategies leading to similar performance results. Firms can therefore be clustered into strategic groups according to their key strategy dimensions (e.g. breadth of product line, relative size). Companies within the same strategic group are relatively homogenous in strategic behavior and performance, whereas significant differences can be found between strategic groups. Explanations of performance differences between strategic groups range from "mobility barriers" (Caves / Porter 1977) to firm-specific factors (Porter 1980). Empirical studies failed to establish evidence of performance differences between strategic groups (Porter 1979, Caves / Pugel 1980, Frazier / Howell 1983, Johnson / Thomas 1988). In addition, strong variations have been noted in company performance between individual firms within the same strategic groups (Cool/Schendel 1988, Thomas / Venkatraman 1988, Fliegenbaum / Thomas 1990). Reviews of the empirical literature on strategic groups (McGee / Thomas 1986, Thomas / Venkatraman 1988; Barney / Hoskisson 1990) concluded that empirical evidence to explain variance in performance through group membership is lacking. Barney and Hoskisson (1990) denoted strategic groups as statistical artifacts with little research relevance. Recent studies found that strategic groups had the lowest explanatory power in relation to all other levels of analysis (Rumelt 1991, Roquebert et al 1996, Dixon / Boal 1996). The main contribution of the strategic group concept is certainly the elimination of the questionable assumption of the homogeneity of all firms within an industry. With the removal of this "entry barrier", researchers started to consider the firm as central unit of analysis and strategy was increasingly seen as an important variable in explaining company performance. Whereas the strategic group concept triggered this development, the concept itself remained "stuck in the middle". Firms remain part of a (albeit finer) homogenous structure. While attributing a Significant role in the generation of 30

performance to company strategy, the variable "strategy" itself remained blurred (Cool I Schendel 1987, 1104). Dixon and Boal (1996) found that strategic groups have a lower ability to explain performance than the strategic variables themselves. Their conclusion is: "(. . .) more attention should be directed towards making the individual firm, their resources, capabilities, and strategies the focus of our study in strategy" (1996, 19). We continue following this advice, turning our attention towards the findings of more recent works on corporate and business strategy in the following sections. In summary we can state that the importance of market structure for performance certainly alters from industry to industry and empirical results are somewhat contradictory. However, the influence of basic fundamental structural parameters on industry (and thus company) performance cannot be denied (Porter 1981). Market structure and the different underlying structural elements therefore constitute a first indispensable constituent for an integrative strategic industry analysis.

11.2

Corporate Strategy

We have shown above that industrial organization views the industry (or industry subgroups) as a homogeneous unit. All firms within an industry are assumed to be equal, except for their size. This assumption is at odds with real life observation of firm behavior and performance. Firms within an industry are clearly not all alike, they follow different strategies and there is considerable variance in their rates of return on invested capital (Porter 1979). We have mentioned the example of differences in performance of German print and publishing companies. There are countless other examples both within and outside the media industries. For instance, Cox Communications outperforms the competition in the US Cable and Satellite market, despite being significantly smaller than its competitors. IBM manages to conSistently outperform other computer manufacturers. If market structure alone cannot fully explain performance, what else is at the heart of value creation or destruction? In 2002, Time Warner provided a textbook example of value destruction. The company posted a net loss of nearly US$100biliion - the largest annual loss in US corporate history. The loss in company value occurred after of the blockbuster merger that united

31

the old media goliath Time Warner and new media princess America Online. At the time, the deal was celebrated as the "deal of the millennium" that established a new paradigm in media. Widespread media groups that integrated content and distribution across multiple platforms were proclaimed the stars of the future. Establishing a crosspromotional platform allows content and brands to be deployed across various distribution channels. However, in the current blaze of negative publicity about media conglomerates, the concept of the diversified and integrated media group is suddenly under fire. The pendulum seems to have swung back from diversification to refocusing. Both industry observers and board members have begun to wonder whether Time Warner (and, almost similarly, the French conglomerate VivendiUniversal), expensively pieced together a mere two years ago, would not be better off broken up. In a recent interview, the management thinker and author Jim Collins commented on Time Warner: "If the whole entity went away and left only its constituent parts, would it matter? What

would the world lose if Time Warner ceased to exist as a company?" Collins' comment points to the crucial force behind the described developments at Time Warner: corporate strategy. Porter defined corporate strategy as "what makes the whole add up to more than the sum of its business unit parts" (1987, 43). Time Warner's move towards further diversification was a strategic decision at the corporate level (mainly driven by the two CEOs Gerald Levin and Steve Case). In order to render the merged businesses more valuable than they would be individually, Time Warner now requires a coherent corporate strategy to realize synergies between the constituent units. A decision to split off the merged businesses constitutes a further option for corporate level strategists. The example of Time Warner gives an idea of the important role played by corporate strategy in large, diversified media companies. At the same time it demonstrates the bitter consequences of failed corporate strategies on performance. In order to reach a better understanding of the underlying effects, we review the literature on corporate strategy and performance in the remainder of this section. After some initial comments on (I) Strategy and Strategic Management, we continue to review the literature on (II) Corporate Strategy and the Scope of the Firm. Next, the significant (11/) Empirical Research on the relationship between firm scope and performance is summarized. Finally, we conclude this section with (IV) Conclusions & Criticism of the present research into corporate strategies.

32

Strategy and Strategic Management

Michael Porter is largely merited with defining the research into the conditions of superior firm performance as the core objective of strategic management. By focusing a then incoherent group of researchers on a single common objective, Porter "helped to set the stage for a revolution in the field of strategic managemenf' (Barney 2002, 54). While clearly based on the industrial organization's theoretical foundation, scholars of strategic management shifted the research focus from the industry towards the individual firm and its strategic actions. "Strategy" determines the firm's basic long-term goals and adopts courses of action to achieve these goals (Chandler 1962). Strategic management decides on the allocation of scarce company resources and skills in order to increase performance (Gunnigle / Moore 1994) and to reach long-term objectives (Faulkner / Johnson 1992, 17). Literature on strategic management typically distinguishes between corporate strategy and business strategy (Hofer / Schendel 1978). It is at the level of the individual business units where most competitive interaction occurs (Porter 1980). Business (or competitive) strategy thus focuses on how business units compete within their industry segments by implementing strategies to perform well in the marketplace (Thompson 1993). We will review literature on strategy at the business unit level later in this chapter (11.3). In this section, we turn the attention towards corporate strategy. If strategy is all about understanding performance differences between firms, then corporate strategy has to determine how corporate headquarters add value to the individual component businesses. What decisions made at the corporate level are most important? In order to answer this question, we first need to reach a common understanding of the term corporate strategy. Corporate strategy is concerned with the general positioning of the company and the joint management of a set of business units (Grant 2002). Andrews specified corporate strategy as the decision on "what business the company is in or is to be in" (1971, 25). Hofer and Schendel propound this view and conclude: "scope and resource deployments among businesses are the primary components of corporate strategy" (1978, 27). Porter summarizes that "corporate

33

strategy concerns two different questions: what businesses the corporation should be in and how the corporate office should manage the array of business units" (1987, 43).

Corporate Strategy and the Scope of the Firm

Decisions on the scope of the firm's activities can take three primary directions: (I) product scope, (II) geographical scope, and (III) vertical scope (Jarillo 1987; Grant 2002, 388). First, decisions on product scope, such as the selection of industries in which to operate, have been labeled as the "diversification strategy" (Chandler 1962). Diversification has been defined as "the extent to which firms operate in different businesses simultaneously" (Pitts / Hopkins 1982, 620). Diversification can be understood as an

extension of a corporation's product scope through entry into new industry segments. The relationship between diversification and performance has been one of the most researched linkages in the strategic management literature (Chatterjee / Wernerfelt 1991 ). Second, more recently researchers have also applied the term "diversification" to reflect the increasing geographical scope of a growing number of corporations (Rugman 1979, Ghosal 1987, Geringer et al 1989). In order to distinguish between these two dimensions, we use the term "product diversification" to describe the extension of the corporation's product scope and "international diversification" to represent the extension of the firm's geographical scope. A number of researchers apply the same distinction (Kim et al 1989, Hitt et al 1997). Both forms of diversification are seen as independent (yet potentially interrelated) variables linked to performance (Hitt et a11997, 779). Finally, decisions on the vertical scope refer to the firm's ownership of vertically related activities. The more successive stages of the product value chain are owned or controlled by the firm, the greater the firm's degree of "vertical integration". For instance, Walt Disney Company is a vertically integrated company that controls most steps along the value chain: movie production (Le. Buena Vista), distribution (Le. ABC), and merchandising (Le. Disney Stores).

34

In the remainder of this section, we review the existing theory on all three variables, starting with product diversification (I), followed by international diversification (II), and vertical integration (III). Product Diversification: In the period between 1950 and 1980, product diversification -

the expansion of firm scope across different product markets - represented a dominant source of corporate growth (Grant 2002, 446). The share of single-business companies among the ranks of the Fortune 500 declined from over 40% in 1949 to less than 15% in 1974 (Rumelt 1982). The same trend applied across Europe and Japan (Whittington et al 1999). During the 1980s, however, large, diversified firms started to reduce their diversification by refocusing on their core businesses (Porter 1987, Markides 1995). The refocusing of the 1980s has been interpreted as attempts by firms to reduce their excessive diversification (Bhagat et al 1990, Markides 1995). Is there an optimal limit to

how much a firm should diversify? How does product diversification impact firms' performance? The product diversification-performance linkage has drawn enormous research attention. Since the initial writings of Ansoff (1957) and Chandler (1962) over 40 years ago, diversification studies have been "a mainstay of strategic management research" (Ramanujam I Varadarajan 1989, 523). Despite this long tradition, there is still "considerable disagreement about precisely how and when diversification can be used to build long-run competitive advantage" (Markides I Williamson 1994, 149). The disputants

in this debate are once more industrial organization economists and scholars from strategic management (Palich et al 2000, 155). While industrial organization economists assumed that diversification and performance are linearly and positively related (Gort 1962, Markham 1973), later inquiries from strategic management postulated that diversification is only beneficial up to a certain degree (Rumelt 1974, Bettis 1981). In the remainder of this section, we review the theoretical assumptions underlying both models. Empirical studies from both sides will be reviewed later in the following section. Since the early beginnings (Gort 1962), researchers in the tradition of industrial organization have assumed a per se positive and linear relationship between product diversification and performance. The linear model rests upon a number of theoretical assumptions, particularly derived from (I) market power theory and (II) internal market

35

efficiency arguments (Grant 2002). Market power theory postulates that diversification leads to increased market power and therefore results in the greater profitability of the diversified firm (Miller 1973, Caves 1981). This is explained by the existence of tools to increase market power, which are only available to diversified firms (Scherer I Ross 1980). A prominent example of such tools is the ability of diversified firms to make use of predatory pricing. The firm can drive rivals from the market through sustained pricecutting. Incurred losses can be funded by cross-subsidization with profits earned from other product lines (Scherer I Ross 1980). Besides the market power theory; justification for diversification activity has more recently been drawn from internal market efficiency arguments (McCutcheon 1991). The diversified firm enjoys flexibility in the capital markets since it can choose between external funding or shifting capital internally between businesses within its portfolio (Stulz 1990, Meyer et al 1992). Internally generated funds are less costly than external capital obtained through debt and equity (Lang et al 1995). In addition, the corporate head office is better positioned to optimize the redistribution of resources among its businesses. Due to superior information access, the head office can efficiently redirect capital from slow-growing, cash-generating operations to rapidly expanding businesses that require investment (Scherer I Ross 1980). Thus, diversification can generate efficiencies unavailable to single-business firms. Besides the dominant arguments along market power and internal market efficiencies, various other theories have been cited as evidence of the positive impact of diversification on performance. For instance, the more efficient exploitation of intangible firm resources such as brand reputation (Markides 1992), tax and financial benefits (Servaes 1996), and risk reduction through a balanced portfolio of businesses (Grant 2002). In contrast to the arguments of industrial organization economists, a number of researchers from strategic management assume a curvilinear relationship between diversification and performance (Palich et al 2000, 158). Based on Wrigley (1970) and Rumelt's (1974, 1982) groundbreaking work, firms are classified according to their diversification strategies as well as to the extent of their diversification. Three main clusters can be distinguished: (I) limited diversification, (II) related diversification, and (III) unrelated diversification. The argumentation related to performance differences between these groups is largely based on the literature of economies of scope and synergies (Kim et al 1989, 46). Limited diversification - the focus on a single industry segment- is seen 36

as unlikely to generate above average returns due to the limited ability to realize economies of scale or between-unit synergies (Lubatkin / Chatterjee 1994). Related diversifiers, however, operate in multiple businesses that make use of the same pool of resources (Nayyar 1992). In other words, they diversify across related segments of the same industry (Palepu 1985). Various synergies can be realized by bundling related products, reusing assets like brand names across multiple operations, deploying learning curve

efficiencies,

and

increasing

purchasing

power (Barney

1997).

Related

diversification is thus seen as superior to limited diversification (Rumelt 1974, Palepu 1985).

However,

while

diversification

is

associated

with

benefits,

increasing

diversification is also associated with higher costs. An increasingly disparate portfolio of businesses leads to high coordination and administrative costs (Markides 1992), control and effort losses (Calvo / Wellisz 1978), and other diseconomies related to organization. Reaching the limits of executives' information-processing ability, inefficiencies occur when managers continue to apply their "dominant logic" to new businesses with completely different strategic conditions (Ravenscraft / Scherer 1987, Hoskisson / Hitt 1988). Consequently, it can be argued that the "marginal costs of diversification increase rapidly as diversification hits high levels" (Palich et al 2000, 159). The benefits from diversification increase up to a certain "optimal level" of diversification and perish with further extension of the degree of diversification. A medium level of diversification (or related diversification) is thus seen as superior to both low (or limited) diversification and high (or unrelated) diversification. Based on these assumptions, Rumelt suggests an inverted-U relationship between diversification and performance (1974, 1982). Before reviewing empirical studies testing the theoretical arguments outlined above, we briefly review literature on the second dimension of diversification -

geographical or

international diversification. International Diversification: Both international trade and foreign direct investment

have soared over the past decade. World trade in commercial services (including media and communications) grew at an annual rate of 14 percent between 1994 and 1999. By 1999, the total stock of foreign direct investment held by all companies worldwide represented US$ 3.6 trillion or 12 percent of the global GDP (Grant 2002, 409). One main driver underlying these dynamics is the increasing international diversification of companies around the globe (Geringer et al 1989, 109). The growing importance of the multinational enterprise has been reflected by a large body of work on international 37

diversification and its competitive and performance implications (Hitt et aI1997). Much in line with the discussion on product diversification, international diversification initially gained economists' attention (Vernon 1966, Rugman 1979, Caves 1982). These authors postulate a linear and positive relationship between the intensity of international diversification and performance. This model is justified through arguments derived from theories on foreign direct investment and market imperfections (Geringer et al 1989, 111; Kim et al 1989,46). A broad geographic scope permits the firm to exploit the benefits of performing more activities internally (Rugman 1981). As a result of the increase in size, internationalization offers possibilities to exploit economies of scale and scope above and beyond the potential of product diversification (Kogut 1985). Inter-country differences in factor costs allow the firm to realize cost advantages (Kogut 1985). Intangible, firm-specific resources can be deployed more efficiently through foreign direct investment than in single markets (Caves 1971). Finally, a global portfolio significantly reduces the level of corporate risk. Similarly to the debate around product diversification, scholars from strategic management increaSingly questioned the assumed

linear relationship

between

international diversification and performance. While acknowledging the advantages of international diversification, these authors stressed the costs entailed by operations on a broad scope (Hitt et al 1997). Realizing synergies across geographically dispersed operations requires significant coordination, leading to increasing transaction costs (Geringer et al 1989, 112). Trade barriers, logistical and distribution costs, and cultural diversity further reduce the

benefits of internationalization

(Hitt et al

1997).

Geographically diverse operations reduce a firm's ability to tailor their products to regional market needs. Firms risk damaging their differentiated market positions (Porter 1985). In general it can be stated that international diversification is highly complex and thus difficult to manage. Hitt et al conclude that moderate levels of international diversification provide an organization with multiple benefits. However, at higher levels of international diversification "the costs of international diversification will eventually exceed the benefits" (1997, 773f). Some scholars, such as Hitt et al (1997) and Geringer

et al (1989), therefore postulate a nonlinear inverted U-shaped relationship between international diversification and performance.

38

Vertical Integration: Besides product and international diversification, a third form of

diversification is widespread in the media industries. Vertical integration is a pattern of diversification where the outputs of one line of business are used as inputs for another line of business (D'Aveni / lIinitch 1992). The vertically integrated firm controls the successive stages of the value chain of its products. Vertical integration can occur in two directions: backward integration and forward integration. Backward integration refers to a firm that acquires suppliers in order to control input production. For instance, Bertelsmann controls a number of printing plants to produce books for their publishing operations. Forward integration signifies taking ownership and control of its own customers. For instance, Bertelsmann controls book clubs and an online bookshop to sell their books directly to end customers. The growing literature in strategy and industrial organization economics relates vertical integration to significant benefits. The transaction cost theory specifically provides some rationale for vertical integration (Williamson 1971, Mahoney 1992). Integrated firms are expected to reduce the costs associated with uncertain market exchanges while providing the benefits of more coordinated activities. Cost reductions may occur in general and administrative expenses, as well as in selling and advertising expenses. Better internal coordination allows for a reduction in unused capacity and inventory carrying costs, as well as for technology and R&D sharing between the different entities (Buzzell 1983, Harrigan 1983). Transactions cost arguments are closely related to market power arguments. Vertical integration increases market power through privileged access to information and the related bargaining advantages (Harrigan 1983). Integrated firms are less dependent on external providers that impose monopolistic market prices (Buzzell 1983, Mahoney 1992). Overall, vertical integration increases the existing barriers to entry (Chatterjee 1991). More recently, however, the usefulness of vertical integration has been increasingly disputed in strategy literature (D'Aveni / Ravenscraft 1994). The disadvantages of vertical integration may be classified into three categories: bureaucratic costs, strategic costs, and production costs (Mahoney 1992, 569). Increasing bureaucratic costs are explained by the additional overheads required to resolve complex problems of control, coordination, and communication caused by vertical integration (D'Aveni / lIinitch 1992). Strategic costs refer to the increased strategic inflexibilities of vertically integrated firms.

39

Integration makes it difficult to adapt quickly to changing market conditions (D'Aveni / lIinitch 1992). Finally, integrated businesses' loss of market pressures leads to increased production costs due to increasing levels of slack and inefficient purchasing (Cyert / March 1963). Besides these reasons, vertical integration has the additional downside of requiring intensive capital investments (Buzzell 1983). The evident question is whether the benefits of vertical integration outweigh its disadvantages. This question is primarily one for empirical analysis. We will thus carry on and review the empirical research on the different forms of diversification addressed above. Our review of the literature on product diversification, international diversification, and vertical integration points to a number of striking similarities. Despite extensive research, there is generally a failure to reach theoretical consensus on the relationship between the different constructs and performance. In order to evaluate the competing theoretical concepts, it is essential to review the existing body of empirical studies. Retracing the present studies will help us draw our own conclusions regarding the existing literature.

Empirical Studies

In this section, five sets of empirical studies on corporate strategy are summarized. Initially, we draw on a number of variance decomposition studies to assess the relative (I) Importance of Corporate Strategy in explaining performance. Based on the overall assessment of the corporate level of analysis, we continue to review empirical studies on the different forms of corporate strategy, namely (/I) Product Diversification, (11/) International Diversification, (IV) Vertical Integration, as well as (V) Studies on the Media Industry. Importance of Corporate Strategy: Earlier in this chapter we mentioned selected studies measuring the relative importance of industry, business, and corporate level effects in explaining business unit performance. As stated above, the initial studies of Schmalensee (1985) and Rumelt (1991) are largely at odds in the importance they assign to industry and business factors. However, the results of both these studies on the role of corporate-level effects are unanimous. Corporate effects explain almost none

40

of the variance in performance. Rumelt concludes, "corporations exhibit little or no ability to affect business-unit returns" (1991, 182). Based on these findings, numerous scholars

suggested that corporate effects on performance do not exist (e.g. Carroll 1993, Ghemawat 1994, Hoskisson et al 1993). Much attention in the strategic management debate subsequently turned towards the business unit level of analysis. More recent studies, however, reflect noticeably larger estimates of corporate effects than do Rumelt (1991) and Schmalensee (1985). Roquebert et al (1996) estimate an average corporate effect of 18 percent. Chang and Singh (2000) found a corporate effect of 11 percent, and Fox et al (1997) one of 8.2% in a simulation run. McGahan and Porter (2002) estimated a corporate effect of 23.7%. Bowman and Helfat review these (and other) variance decomposition studies and argue, "corporate strategic management matters in explaining the variance of profitability" (2001, 20). In their article, Bowman and

Helfat suggest that the poor results of corporate effects in earlier studies can be explained by methodological shortcomings. For instance, many of these studies include single-business firms and define corporate effects as zero for these firms. This produces substantially lower estimates of corporate effects than would occur if single-business firms would be excluded (Bowman I Helfat 2001). McGahan and Porter's (2002) study exemplifies this effect. The estimated corporate effect of 23.7% dropped to 13.7% after the inclusion of single-business firms. Other methodical problems include a too broad definition of industry (increasing the number of single-business firms) and the omission of interaction effects between industry and corporate levels and their impact on performance. Despite the negative effect of such methodical shortcomings, "al/ of the studies (.. .) except Schmalensee (1985) and Rumelt (1991) contain non-negligible estimates of corporate effects" (Bowman I Helfat 2001, 20). Corporate strategy does

indeed matter, especially for the diversified firm. On average, across all present studies, corporate strategy accounts for almost as large a part of variance in performance as industry structure. Having established the overall significance, we turn our attention to the

finer

constructs

underlying

corporate

strategy:

product

and

international

diversification, as well as vertical integration. Product Diversification: A great deal of empirical research has been conducted on the

product diversification-performance linkage over the last three decades. Dess et al (1995) found 32 studies published in the Strategic Management Journal and the 41

Academy of Management Journal since 1980. A more recent review identified 82 relevant quantitative studies on the subject (Palich et al 2000). Despite the size of the research, many authors have stated a lack of consistency and the absence of consensus regarding the diversification-performance linkage (Seth 1990, Hoskisson et al 1993, Dess et al 1995). However, a somehow clearer picture emerges if we separate the studies in the tradition of industrial organization economics from more recent research in strategic management. Early empirical studies looked into the relationship between the degree of diversification and performance. In the tradition of the industrial organization theory, these authors attempted to establish a linear and positive correlation. The evidence for this hypothesis is clearly mixed. Some studies found (mostly modest) support for a linear relationship (Rhoades 1973, Jose et al 1986, Grant et al 1988). Others identified no performance differences between firms with different diversification levels (Gort 1962, Montgomery 1985). A third group found a negative relationship between product diversification and performance (Hill! Snell 1988, Hill! Hansen 1991, Hoskisson et al 1993). Not exactly clear evidence! In addition, empirical work found little evidence for the theoretical arguments underlying the assumption of a linear relationship (Palich et al 2000, 157). There is little evidence for the hypotheSized connection between diversification and market power (Grant 2002) and support for the assumption of increased internal market efficiencies of diversified firms is mixed at best (McCutcheon 1991). In summary, the present empirical work does not provide sufficient evidence for the assumed linear and positive relationship. In the strategy literature, Rumelt (1974) first tested the hypothesis of a curvilinear relationship between product diversification and performance. Rumelt found greater support for the assumed superior performance of firms pursuing related diversification strategies than for firms pursuing unrelated diversification strategies. Many subsequent empirical studies confirmed the view that related diversification could have a positive impact on performance (e.g. Bettis 1981, Christensen! Montgomery 1981, Rumelt 1982, Hoskisson 1987, Hill! Snell 1989). Dess et al (1995) list fifteen studies in support of Rumelt's assumption. There is also growing empirical evidence in support of the curvilinear relationship (Hoskisson ! Hitt 1990, Markides 1992, Lubatkin ! Chatterjee 1994, Tallman ! Li 1996). In their meta-analysis of 55 studies on the product 42

diversification-performance linkage, Palich et al (2000) found some support for the inverted-U model. However, despite much empirical evidence, the support of the curvilinear model is not universal. Some studies found evidence for an equal or even higher performance of unrelated diversification (Bettis / Hall 1982, Michel / Shaked 1984). In summary we conclude that while diversification plays an important role in strategic research, no final agreement has been reached on the overall effect (and its direction) on performance. International Diversification: We have shown above that no clear consensus has been

reached on the relationship between product diversification and performance. The underlying theoretical debate on the international diversification-performance linkage is rather similar to the debate on product diversification. Initially, researchers tested an assumed linear and positive relationship between international diversification and performance. In contrast to the twin debate on product diversification, international diversification has "generally been found to improve operating performance" (Tallman / Li 1996, 179). Vernon (1971) found that multinationals earned a higher return on sales and higher post-tax return on assets than national enterprises among the Fortune 500 corporations. Grant (1987) analyzed 304 British manufacturing companies and found multinationality to be positively associated with superior profitability. A number of other scholars confirmed the findings by providing further empirical evidence on a link between degree of internationalization and various performance measures (Buhner 1987, Daniels / Bracker 1989, Haar 1989, Kim et al 1993). Most of these studies applied measures like the foreign

sales

or the foreign

assets

ration to

represent the degree of

internationalization. Tallman and Li (1996) found evidence for a positive impact of the scope - not the degree or intensity - of internationalization on performance.

While the majority of studies confirm a positive relationship, a few studies failed to establish a significant correlation (e.g. Kumar 1984) or even found a negative relationship between international diversification and performance (Michel / Shaked 1986). Apart from these exceptions, there are two more recent studies with particularly interesting deviating results. Geringer et al (1989) tested the prevalent assumption of a linear and positive relationship, but failed to find confirmation. On the contrary, they found a significant curvilinear relationship: "As the degree of internationalization of 43

multinationals reached higher values, performance also exhibited increased values but then peaked and exhibited diminished levels of performance" (Geringer et al 1989, 117). They conclude that as firms reach higher levels of geographic dispersion, the associated cost escalated rapidly, thus eroding profitability margins. Managers interviewed for their study broadly supported this assumption (Geringer et al 1989, 117). Hitt et al tested these hypotheses with a sample of 295 manufacturing firms and found strong support for

a "curvilinear and inverted U-shaped relationship between international diversification and performance" (1997, 793). These results are interesting from a theoretical point of view and quite close to the current debate on product diversification. However, the empirical support is thus far limited to two single studies. In summary it can be stated that while the number of empirical studies is significantly less, there is stronger consensus on a generally positive influence of international diversification on performance than in the debate on product diversification. However, there is still some contradiction on whether high levels of internationalization lead to increasing or diminishing returns. Further analysis will be required to close this research gap. Vertical Integration: Is it more beneficial for firms to vertically integrate, or, rather, to

"disintegrate" through outsourcing or network organizations? As we have seen above, theoretical arguments are contradictory,

pOinting to both the advantages and

disadvantages of vertical integration. Unfortunately, empirical research into vertical integration is similarly ambiguous. Earlier studies predominantly found a slightly positive relationship between vertical integration and performance. In their meta study, Capon et al (1990) review 14 studies on vertical integration with a significantly higher share of positive than negative relationships reported. Most of these studies identify moderating effects exerting significant impact on the vertical integration-performance relationship. For instance, Buzzell (1983) found a positive relationship; however, the relationship is strongly moderated by factors like market position or business strategy. Moreover, Buzzell and the other earlier studies have been criticized for their choice of the vertical integration measure used (Maddigan I Zaima 1985). Accounting-based measures of the ratio of "value-added" to sales as a proxy for vertical integration are distorted by several factors that have nothing to do with vertical integration (O'Aveni I Ravenscraft 1994). More recent studies use more sophisticated scales, but the results remain mixed. While

44

some studies found an overall positive relationship (D'Aveni / Ravenscraft 1994), others found no significant relationship (Reed / Fronmueller 1990, Prabhu / Harrison 1992). In general, two conclusions can be drawn from the existing literature. First, the empirical work on vertical integration is quite limited and lags significantly behind the elaborate theoretical reasoning (Prabhu / Harrison 1992, 18). Second, the value of vertical integration may depend on the state of other variables (e.g. market conditions, strategy). We will come back to the latter point later in our study. Studies on the Media Industries: Corporate strategy has gained significantly less attention in media industry research than market structure. Nevertheless, diversification is an important and widespread phenomenon in the media industry (Dimmick / Wallschlager 1986, Albarran / Porco 1990). While media economics theory assumes largely positive effects of all three forms of diversification, there are very few empirical studies that verify these claims. The rare exceptions include empirical studies on product diversification (Albarran 1990, Picard / Rimmer 1999) and vertical integration (Ford / Jackson 1997).

Conclusions & Criticism

In the wake of Porter's (1980) influential work on competitive strategy, much research attention turned away from the hitherto predominant realm of corporate strategy. This prevalent trend was fortified when companies shifted their strategic focus from diversification towards restructuring in the late 1980s. A wide range of studies on business-level strategy emerged during the 80s and early 90s "without considering the role played by corporate strategy in creating and sustaining competitive advantage at the business lever (Dess et al 1995, 358). However, ever since the late 90s, corporate

strategy has experienced a modest revival. Variance decomposition studies stressed the importance of corporate-level effects besides business unit and industry level effects (Helfat / Bowman 2001). Other authors claimed the concomitant importance of both corporate and business level strategies (Dess et al 1995, Stimpert / Duhaime 1997). Despite the long tradition and the more recent evidence on the importance of research into corporate strategy, the field remains impeded by considerable methodological problems (Ramanujam / Varadarajan 1989). We briefly review two main methodological

45

criticisms in the remainder of this section: (1) The Omitted Variable Problem, (2) The Unit of Analysis Problem. These methodological issues explain some of the inconsistency in research findings (Stimpert / Duhaime 1997, 565). The Omitted Variable Problem: In his influential study, Rumelt (1974) found that firms

pursuing a strategy of related diversification were the most profitable. In a later study, Rumelt (1982) did check for industry effects and thus failed to reproduce the same results. Bettis and Hall (1982) concluded that the superior performance of a related diversifier was a function of industry membership (particularly due to the high performance of pharmaceutical companies) rather than of diversification strategy. Many later studies on diversification did not monitor industry effects, therefore, "their results suffer from a serious omitted variable problem, and may well be spurious" (Dess et al 1995, 370). Subsequent studies that have checked for industry membership mostly failed to find a significant impact both in the debate on product (Grant et al 1988), international diversification (Grant 1987, Hitt et al 1997), and vertical integration (Prabhu / Harrison 1992). Nevertheless, market structure is generally considered as an important mediator of the relationship between diversification and performance (Ramanujam / Varadarajan 1989). Much along the same line, most studies have been criticized for not monitoring the impact of "unobservable factors" or firm competencies, leading, in turn, to an omitted variable problem (Dess et aI1995). Chandler (1962) stated that the performance impact of diversification strategies depends on the organizational structure in place. It is evident from the following stream of research that administrative mechanisms have to be in place in order to realize diversification benefits, and "structure appears to be the most important of these administrative mechanisms" (Ramanujam / Varadarajan 1989, 536). Chandler (1962) and Williamson (1967) recommend a multidivisional (or "M-form") structure, which involves the deconstruction of the firm into a number of profit centers or divisions, as appropriate for large diversified firms. The assumption that organizational structure (Le. the adoption of the M-form structure) should affect performance has been tested in various empirical studies. Early studies were largely supportive of this hypothesis and associated superior performance with the multidivisional structure (Armour / Teece 1978, Steer / Cable 1978, Ezzamel / Hilton 1980, Teece 1981, Thompson 1981). However, more recent work failed to find evidence of the M-form hypothesis (Cable / Dirrheimer

46

1983, Cable / Yasuki 1985, Ingham 1992, Ezzamel / Watson 1993, Weir 1995). Weir concludes, "(. . .) it is clear that adopting a particular organizational structure does not

automatically result in improved performance" (1995, 32). Beside the lack of empirical evidence, the role of organizational structure is further devaluated by another aspect. Recent studies found that in the meantime all medium-sized and large firms adapted some form of multidivisional structure (Weir 1995). The value of organizational structure in explaining inter-firm performance differences is thus further minimized. In summary, the majority of the diversification-performance studies failed to check for other variables (in particular market structure and capabilities) that have demonstrated significant effects on this relationship (Palich et al 2000, 168). This implies that there is a risk that these studies fail to capture the "complex interactions between industry,

diversification, business strategies, and performance" (Stimpert / Duhaime 1997, 576). The Unit of Analysis Problem: The vast majority of the studies reviewed above focuses

on the linkage between corporate strategy and corporate performance (Summer et al 1990). Researchers have assumed that the corporation is an appropriate unit of analysis for diversification studies. Few empirical analyses focus attention on how diversification strategy influences performance at the business level (D'Aveni / Ravenscraft 1994, Stimpert / Duhaime 1997, 566). However, the problem is that if diversification affects performance, "it is at the business level where its effect will be most apparenf' (Dess et al 1995, 373). Corporate performance is an indirect outcome of corporate strategy, computed by aggregating the performance results of the constituent business units. The linkage between diversification and business unit performance is thus logical beforehand. Consequently, Dess et al (1995, 374) suggest adopting "the business unit as the unit of analysis in diversification studies, as opposed to the corporation as a

whole". Both criticisms imply the interrelation between corporate strategy and other variables, including industry, capabilities, and business strategies. The latter plays a particularly important role. Corporate strategy may impact performance at the business unit level. However, business strategy is equally (if not more) important than corporate level strategy: "Variation in a firm's corporate level strategy and its business level strategy

both help to explain variation in firm profitability" (Beard / Dess 1981,686). It is at the 47

business level where most competitive activities take place (Porter 1980). The following section will thus be dedicated to a review of studies focused on strategic actions at the business unit level and their impact on performance.

11.3

Business Strategy

The last section emphasized the diversified firm and fundamental strategic decisions at the corporate level, such as acquisitions, expansion, and refocusing. However, these decisions are limited to single moves of the company throughout the year. Most competitive activities occur within the single units of the firm. For instance, the German media conglomerate Bertelsmann consists of six principal divisions. The RTL Group operates radio and television stations, BMG develops artists and publishes music, the Direct Group operates book and music clubs and is active in ecommerce, Arvato provides media services, Gruner & Jahr publishes magazines, and Random House publishes books. The competitive challenges vary strongly between the individual divisions. BMG faces a record industry suffering its worst slump in a decade. As consumers continue to download their favorite songs from the Internet for free, album sales slipped by more than 10% in 2002. Bertelsmann announced a cost reduction strategy to turn around the loss-making division. At the same time, the printing business Arvato made a significant EBITA contribution and continued to grow at an average rate of 15% p.a. over the last four years. Arvato has been developed into an international networked provider of media services that exceeds the limits of traditional print production by far. The provided services include print, logistics, IT, and digital storage media. The differentiated product positions Arvato ahead of the competition and the success of this strategy is reflected by the outstanding results achieved despite difficult market conditions. As we can see from the BMG and Arvato examples, the market environments of individual divisions within a diversified firm can differ strongly, as do business strategies. While conditions in one segment may require a differentiated strategy focused on the growth of the business, consolidation and cost reduction may be more appropriate to succeed in other divisions. Strategic moves at the business unit level are more frequent;

48

decisions are made on a regular basis. How can the underlying, more generic and enduring strategic orientations be categorized? What impact does the strategic orientation at the business unit level have on firm performance? In the remainder of this section, we review work from the strategic management literature engaged in finding answers to these questions. Initially we briefly summarize findings from (/) The PIMS Program, the first comprehensive analysis of determinants of performance at the

business unit level. Next we review the most influential (/I) Generic Strategy Typologies. In the following, the considerable body of (11/) Empirical Research is presented, including studies on the media industry. We conclude this section with the (IV) Conclusion & Criticism part.

The PIMS Program

In 1960 General Electric started an internal project under guidance of Sidney Schoeffler to systematically review accounting data in order to identify variables influencing return on investment (ROI) and cash flow at the SBU level (Schoeffler et al 1974). Building on the established database and initial results, the Marketing Science Institute attached to Harvard Business School extended the research to other companies and industries in 1972. The research project, called Profit Impact of Market Strategies (PIMS), became independent in 1974 as the Strategic Planning Institute (SPI). There are now more than 3,000 strategic business units in the database with performance data for at least five years on 150 variables. The PIMS project utilizes ROI as a dependent variable and relates determining variables to ROL The determining variables have been generated inductively through the application of correlation analysis to large amounts of data. The model intended to answer two basic questions (Schoeffler et al 1974): What factors influence SBU performance and how much? How does ROI change in accordance with changing strategies and market conditions? Instead of proceeding from a theoretical framework to the empirical affirmation, the PIMS program tended to move from empirical evidence towards a unifying theory. The ultimate goal of the program though was to identify market place "universal truths" (Gale 1978) or "natural laws" (Schoeffler 1977) that explain the performance of any business in any time period. This research draws largely

49

on prior work from marketing and strategic management. In line with Jemison's (1981) demand for an integrative approach to strategic research, initial PIMS studies comprised 37 distinct factors that had been clustered into three sets of influencing variables (Buzzell / Gale 1987): (1) market structure variables, (2) business strategy variables, and (3) variables determining the competitive pOSition of the strategic business unit. The performance of the SBU is seen as the function of these three sets of variables (see Figure 5). (I) Market Structure variables used by the PIMS program are largely consistent with factors developed by industrial organization scholars (see Chapter 11.1 for description). Eleven factors were used to determine market structure, such as industry growth rate, seller and buyer concentration, investment intensity, and fixed capital intensity. These variables are viewed as largely uncontrollable by management.

Industry Structure Business Strategy

Performance

Competitive Position

Figure 5

The PIMS Competitive Strategy Paradigm

(II) Business Strategy variables comprise the largest number of factors, most notably pricing, advertising / marketing expenses, new product introduction, R&D spending, product quality, and vertical integration. Business strategy, as opposed to corporate strategy, attempts to explain how companies compete on the strategic business unit level (Porter 1987). A business unit is defined as an operating unit selling a distinct set of products to a specific group of customers in competition with a well-defined set of competitors. Based on this understanding, the PIMS research analyzes the firm's

50

strategic activities at the competitive SBU level. The business strategy level is directly controllable by management (Anderson I Paine 1978). (III) Competitive Position (or "Firm Market Position") variables include market share

(absolute or relative to the largest competitors), breadth of product line, relative cost, and relative market entry timing. These factors position the SBU against its main competitors in the market. The competitive position is both influenced by the strategic moves of the company under examination and its competitors. The competitive position level is thus only partially (or indirectly) controllable by management. In order to test the claimed influence of the 37 identified factors on company performance, numerous empirical studies have been conducted. In the first two test phases, 620 businesses were analyzed and the researchers managed to explain more than 70% of the variation in profitability (Schoeffler et al 1974). These and various following studies based on the PIMS database have stressed the performance impact of factors such as market share (Schoeffler et a11974, Buzzell et al 1975), product quality (Schoeffler et al 1974, Craig I Douglas 1982, Phillips et al 1983), vertical integration (Buzzell 1983), market growth (Schoeffler 1977), investment intensity (Schoeffler et al 1974, Gale 1980), and product innovation (Ravenscraft I Scherer 1982, Colliner et al 1984). Despite the breadth of these results, one factor has almost entirely dominated the discussion of findings from the PIMS program: the market share of the SBU. Schoeffler et al (1974, 141) revealed market share as the strongest single influencing factor for business unit profitability. Findings from this study indicated that in company profitability high-market share businesses outperformed low-market share competitors by more than three times. A difference of 10 percent in market share is accompanied by an average difference of almost 5% in pretax ROI (Buzzell et aI1975). These research results led to the general advice to adopt market-share objectives and strategies following the maxim "the bigger the better' (subtitle of Buzzell et al 1975). Once again Mintzberg (1993) proved to be right with his criticism of the policy field's tendency to premature prescriptions: A number of studies in the early 80s started to question the generalized link between market share and profitability (Rumelt I Wensley 51

1981; Woo / Cooper 1981, 1982; Woo 1984). Based on the observation that a number of small firms managed to consistently outperform larger competitors, these authors stressed the superior product quality and lower cost level of these firms as important determinants for performance ry.Joo / Cooper 1982). Small firms are able to achieve superior performance by adopting niche strategies with a strong focus on a target segment such as a regional market (Hamermesh et aI1978). Furthermore, economies of scale do not have the same importance in all industries (Aaker 1984) and the ideal size to achieve maximum economies of scale (or the "minimum efficient scale") can be rather small in several market contexts (Scherer / Ross 1980). In this event, large companies generally suffer from "diseconomies of scale" since they have to bear the increased costs of coordination and administration, while economies of scale are limited. All these effects clearly contradict the general positive effect of market share on profitability. Prescott et al (1986) concluded their empirical study with the statement that the importance of market share is clearly environment and context specific. Additional criticism came from a group of authors led by Robert Jacobson. They argue that the relationship between market share and profitability is not causal but spurious, both being jointly influenced by third factors (Jacobson / Aaker 1985). Omitted variables such as management effectiveness, price, customer satisfaction, or product quality are not modeled into the PIMS model and influence the market share-ROI correlation. Various studies have modeled these factors into their analysis and found that the direct impact of market share on ROI was substantially lower than assumed before (Rumelt / Wensley 1981; Schmalensee 1985, Prescott et a11986; Jacobson 1988, 1990; Hansen / Wernerfelt 1989, Boulding / Staelin 1990, Montgomery / Wernerfelt 1991). The empirical study by Jacobson and Aaker, based on PIMS data (1985), found a relatively minor impact on performance by market structure and criticize strategies targeted at achieving higher market shares. It might come as a surprise, but these studies are perfectly in line with the initial argumentation of the efficiency school (Demsetz 1973). As we have seen before, Demsetz understands market share as a consequence of efficiency, a synonym for these "third factors" such as management quality. Jacobson concludes that "both market share and ROI are joint outcomes of successful strategies that managers have implemented (. ..J" (1988, 78). Market share can thus be understood as a dependent variable influenced by market structures, strategies, and capabilities. Anderson and Paine already concluded that some variables researched within PIMS studies may

52

impact one another and suggested a "causal sequence" (1978, 606). If market share is rather a consequence of business success than a cause (Jacobson / Aaker 1993, 212), it should be seen as an intermediate factor between determinants and measures of performance. Numerous other studies use market share as a dependent variable to measure performance (i.e. Chang / Singh 2000). Despite acknowledging the importance of business strategy, the PIMS program failed to "put a grip" on individual business strategies and thus side-stepped this challenge by analyzing more visible factors, in particular the market share of the firm. Therefore, we will review the concepts of generic business strategies next in order to better capture the essence of strategy.

Generic Business Strategies

Initially, business strategy was considered as idiosyncratic and unique to each firm. In the early 1980s, research showed a significant shift away from this atomistic view of strategy towards a more "generic" view acknowledging communalities among firms (Dess / Davis 1984). These configurations have been referred to as "generic strategies" (Miles / Snow 1978, Porter 1980), or "gestalts" (Miller 1981) providing integrated models in order to study relationships between strategy and other variables, such as performance, more effectively. In this section we present the two classical approaches of (I) Miles and Snow (1978) and (II) Porter (1980), as well as more recent extensions to these approaches (III). The Miles & Snow Typology: Based on empiric stUdies in four industries, Miles and

Snow identified four types of business-level strategy: prospectors, defenders, analyzers, and reactors. The primary dimension underlying this framework is the rate at which organizations change their products and markets. Prospectors perceive a dynamic environment and are flexible in order to frequently offer new products and realize new market opportunities. Defenders are the opposite of prospectors. They are relatively stable, tend to occupy and defend market niches and focus on constantly improving efficiency. Analyzers are an intermediate type, they are willing and ready to enter new markets or offer new products, but they tend to limit these activities as they affect

53

efficiency. Reactors, as the last group, lack a consistent strategy to confront environmental change and are thus expected to perform poorly. Based on these four strategy types, Miles and Snow examined the interrelation between key strategic attributes within each strategy type. Three key strategic problem sets have been analyzed. First, the entrepreneurial problem focused on product / market entry behavior. Second, the engineering problem concerning the choice of technologies. Third, the administrative problem covers the choice of organization structure and processes. Based on these dimensions, Miles and Snow concluded that prospectors tend to select a flexible and decentralized structure requiring complex coordination and communication mechanisms. Defenders, on the other hand, incorporate an extensive division of labor, high formalization and centralization. Analyzers seem to apply a defender strategy to existing products, while showing prospector attributes in pursuing new opportunities. Miles and Snow suggested that organizations tend to develop internal consistencies that lead to relatively enduring patterns of strategic behavior in aligning the firm to the environment. Strategy types are thus rather stable over time. Prospectors, defenders, and analyzers are expected to show far better performance than reactors with their opportunistic and inconsistent strategic behavior. Following Miles and Snow's concept of generic strategies, other scholars have attempted to capture the complex strategiC activities in a number of clearly defined strategy types (Hofer / Schendel 1978, Abell 1980, Porter 1980, Galbraith / Schendel 1983, Greiner / Schein 1989, Thompson 1993). Most of these approaches have been criticized for vague definitions or too narrow perspectives that make further application difficult. The exception is Porter's model of generic strategies that academics have found particularly interesting and which has received more empirical support than any other typology (Kim / Lim 1988). We will therefore review Porter's model in more detail. Porter's Generic Strategies: Porter (1980) identified the three generic strategies a firm

should pursue, regardless of industry context, to outperform its competitors. The three generic types are cost leadership, differentiation, and focus. The cost leadership strategy focuses on producing a standardized product with the aim to improve efficiency to become the lowest cost producer in the industry. Due to lower costs, prices can be reduced in order to gain market share. This leads to economies of scale that in turn allow further price reductions. Finally, the firm's relative cost advantages lead to a competitive 54

advantage over the competition. Firms that pursue a differentiation strategy focus on offering a clearly differentiated product that is perceived as unique by their customers. Such companies intend to create brand loyalty for their offering and thus, to a certain extent, price inelasticity on the buyer side. Differentiation can be achieved in the breadth of the product (or service) offering, through special features of the product, or a particularly high quality. Finally, focus strategy involves a concentration on one particular market segment or group of buyers. In this smaller market, the company is expected to be able to better fulfill buyer needs than a large, less focused, corporation. Porter emphasized that firms cannot adopt two generic strategies at the same time without being "stuck in the middle", which in turn leads to low performance. The generic strategy types suggested by Porter are quite similar to the typologies of Miles and Snow presented before. Porter's "cost leaders" are similar to Miles and Snow's "defenders". Similarly, Porter's "differentiators" can be compared to Miles and Snow's "prospectors". Extensions to the Classical Typologies: In discussing the Miles and Snow as well as

the Porter models, various researchers contributed to the further development of these concepts. Notably, an interesting discussion emerged around Porter's belief that a firm cannot pursue cost leadership and differentiation strategies simultaneously. Real life examples like McDonalds, Swatch, or Citibank have shown that firms indeed proved successful by employing a "hybrid" strategy combining both cost leadership and a differentiated product. While initial studies confirmed Porter's view (Hambrick 1983, Dess / Davis 1984), later work proposed that adopting the two strategies together can result in superior performance (White 1986; Buzzell / Gale 1987; Wright et al 1990, 1991; Miller 1992; Dess / Miller 1993; Parnell / Wright 1993, Parnell 1997). As an extension to Miles and Snow's typology, Wright et al (1990) proposed an additional strategy type, the balancer. This high performing combination strategy "balances" the need to optimize

costs and differentiate products by combining elements of all three other viable strategy types. The balancer is different from the analyzer as the latter applies alternative strategies to different market spheres instead of compromising between the alternatives.

55

Empirical Research

Considerable empirical research has adopted the concept of generic strategies in various industry settings. The main research focus has been on the relationship between the different strategy types and performance. In order to assess the relevance of generic strategies, we will first review empiric studies on the impact of generic strategies on performance (I). Next, we present the results of research into the significance of hybrid strategies (II). Finally, we summarize studies leveraging the concept of generic strategies for application to the media industry (III). Generic Strategies and Performance: Many notable studies found significant empirical

evidence that companies adopting a viable generic strategy show far better performance than firms that do not (Hambrick 1979, Meyer 1982, McDaniel / Kolari 1987, Conant et al 1990). Applying Porter's model, Dess and Davis (1984) found that firms applying differentiation, cost leadership or focus strategies significantly outperformed companies without a consistent strategy type ("stuck in the middle"). Hambrick (1983) arrived at similar results in an empirical study based on the Miles and Snow typology. Prospectors, defenders, and analyzers showed Significantly higher performance than reactors do. This finding has been confirmed by a large number of studies, including an analysis of the tobacco industry (Miles / Cameron 1982), the semiconductor, automotive, and air transportation industries (Snow / Hrebiniak 1980), the wholesale market (Parnell et al 2000), and the catalogue and mail order industry (Parnell / Wright 1993). The higher performance of viable generic strategies has been proved for profitability measured by return on assets (e.g. Parnell / Wright 1993) and return on investment (e.g. Hambrick 1983, White 1986), sales growth (e.g. Parnell et al 1996, Parnell 1997), and market share increase (Hambrick 1983). In addition to the overall superior performance of the viable strategy types, most researchers found stable performance differences between the different generic strategies. A significantly stronger link between cost leadership (or defender) strategies and profitability has been found compared to differentiation and focus strategies. Dess and Davis (1984) confirmed cost leaders' superior return on assets (RDA) performance compared to that of differentiators. Hambrick (1983) and White (1986) found the same

56

relationship for return on investment (ROI) performance. However, cost leadership firms show lower sales growth than the two other types. With the exception of Dess and Davis' study (showing highest sales growth for analyzers), most studies concluded that differentiators (or prospectors) outperform all other strategy types in sales growth (White 1986, Parnell / Wright 1993, Parnell et al 1996). To draw a first conclusion: there is important empirical evidence that differences in generic strategies have a Significant impact on firm performance. The Significance of Hybrid Strategies: As described above, several authors have extended the Miles & Snow typology with a hybrid or "combinative" strategy that combines the best of both worlds: cost excellence and differentiation. A number of empirical studies attempted to connect the combined strategy type to superior performance. In two early studies, Cappel (1990) and Wright et al (1990) found evidence for a significantly higher performance by firms employing a combinative strategy compared to both cost leadership and differentiation strategies. Subsequently, Parnell (with varying co-authors) has presented a suite of studies in different industry environments analyzing the role of hybrid strategies. Two initial studies on the catalogue and mail order industry and based on the Miles and Snow typology showed a superior profitability performance and competitive growth rates for the analyzer strategy type (Parnell / Wright 1993, Parnell et al 1996). They concluded that combination strategies might be particularly successful in volatile and dynamic industries (like the mail order industry that they examined). Based on this assumption, Parnell (1997) extended the Miles & Snow typology with a "balancer" strategy type (first developed by Wright et al in 1990). He found that both the ROA and sales growth of balancers were Significantly higher than that of all other strategy types. Parnell et al (2000) confirmed this finding in a study of the wholesale grocery market. They found strategy to be an important predictor of ROA, with balancers as the high performers. Yamin et al (1999) detected significant differences in performance between firms with different strategy types, even in a sample of top performers. Combination strategies proved more successful than differentiators, cost leaders, and focused firms.

57

One additional conclusion can be drawn: there is some evidence that hybrid strategies are viable strategic alternatives promoting superior performance, particularly in volatile and dynamic market segments. Application to the Media Industry: Most empirical research in strategic management

has been conducted in stable industries, particularly in the manufacturing industries. In the generic strategy debate, Parnell was the first to select a more dynamic environment, the catalogue and mail order industry (Parnell I Wright 1993, Pamell et al 1996). Both studies confirmed the main assumptions of the Miles and Snow model. Other studies on generic strategies specific to the media (and communication) industries are very limited and show significant shortcomings. Sjurts (1996) analyzes different media industry segments and applies a descriptive approach based on generic strategy types to analyzing strategic firm behavior. She extends Porter's approach with two additional dimensions: strategic behavior and strategic path. Strategic behavior distinguishes between reactive and innovative approaches. Strategic path differentiates internal and external (through mergers and acquisitions) growth strategies. While providing an interesting refinement of the generic strategy concept, this approach fails to provide a framework (or measurements) for empirical research. Chan-Olmsted

and

Jamison

(2001)

analyze

the

strategic

patterns

of

major

telecommunications companies exploiting both Porter's generic strategies and a more recent typology. They argue that Porter's three generic strategies cannot contain all the possible strategies available to communication businesses. Thus, the authors describe five different strategic patterns: (I) focus, (II) best product-differentiation, (III) customersolutions orientation, (IV) lock-in, and (V) strategic alliances. The first two patterns are more or less consistent with Porter's niche and differentiation strategies. "Customersolutions orientation" is a combination of focus (targeted group of customers) and differentiation (unique services) strategies and thus consistent with Miller's (1986) interpretation that focus cannot be pursued without one of the two other strategies. For number four, "lock-in", even the authors themselves fail to provide a single example. Five or "strategic alliances" is hardly an exhaustive strategy for a business unit. Generally it can be concluded that Chan-Olmsted and Jamison's approach remains close to Porter's 58

generic strategies. In addition, this model lacks the required empirical testing to provide a viable alternative to the two proven classical approaches. In summary, despite an emerging discussion on business strategies in the media industry, there are (I) currently no alternatives to the classic generic strategy models, and (II) these models need to be further examined for their validity for the media industry before endless unproven typologies and modifications are created (Mai 2002, 6).

Conclusion & Criticism

Most early criticism, in particular of Porter's model of generic strategies, has been addressed and, for the better part, resolved. Porter's controversial assumption that differentiation and cost leadership strategies are mutually exclusive has been surmounted by extending Miles and Snow's typology to include combination strategies (Wright et al 1990, Parnell et al 2000). Criticism of the assumption that all generic strategies are alternative, viable options across environmental contexts (Miles I Snow 1978, Porter 1980) has led to broad research into the generic strategy-environment relationship (Hambrick 1983, Miller 1988, Marlin et al 1994) to be reviewed later in our study. This research field has recently been extended to capture cultural differences and analyze the validity of generic strategies for countries outside the US (Silva et al 2000, Mai 2002). This effort has led to a more appropriate and specific application of generic strategies. Beside this specific criticism, broad empirical research has demonstrated the usefulness of the generic strategy concept. Significant performance differences have been found between companies adopting different generic strategies. It is clear, however, that generic strategies are broad categorizations and generalizations of idiosyncratic business strategies. Any attempt to categorize complex concepts such as business strategy into a small number of generic strategies will necessarily involve simplification (White 1986, 220). Generic strategies concentrate on certain aspects of business strategies while ignoring others. On the other hand, this Simplification is indispensable in allowing for more effective research into the relationship between strategy and other important variables such as industry structure, firm capabilities, or performance. The

59

concept of generic strategies thus provides an interesting tool for integrated strategic research as attempted in this study. To date both Miles & Snow's as well as Porter's models have been widely adopted by most scholars (Silva et al 2000). Generic strategies capture the most critical

(albeit not all) dimensions,

have strong theoretical

underpinnings, and are tailored to and proven in empirical research. The centrality of corporate and, in particular, business strategy for firm performance has been empirically proven. However, today it is generally acknowledged that the strategyperformance relationship is moderated by a variety of both environmental variables and organizational factors influencing the success or failure of a given strategy (White 1986, Miller 1988, Barney 1991). We have already shown the influence of industry structure and environmental variables above (see Chapter 11.1). More recent research has emphasized the importance of firm-specific factors in explaining company performance. We will thus now turn our attention to a third central construct (besides industry structure and strategy) that has been related to performance: the firm's unique resources and capabilities.

11.4

Resource-based View

With Michael Eisner's inauguration as president of Walt Disney Corporation in 1984, the company experienced its fourth consecutive year of declining earnings and a plummeting share price. In the following four years, Disney's sales more than doubled, net income and market valuation arose to five times the initial level. Despite this unparalleled performance, no fundamental shift in market conditions or the general strategy type had taken place. Disney's turnaround can be largely explained by the mobilization of the company's extensive assets and resources. The company's huge film library was better deployed by licensing content to TV networks and video distributors. Over 28,000 acres of land in Florida owned by Disney were developed into a resort with hotels, theme parks, and residential homes. The Touchstone movie label was leveraged to enter the teenage and adult movie segments. Leading talent was recruited to support this expansion. Above all, the success relied strongly on exploiting Disney's most important asset: the fascination of millions of people across the world with the Disney brand and the Disney characters (Grant 2002).

60

As we can see from the Disney example, strategy can be understood as deploying a firm's resources and capabilities to realize market opportunities. So far, we have largely emphasized the external environment and the role of strategy in line with the work produced by industrial organization and strategic management during the 70s and 80s. In the 90s the focus of strategic management turned towards the analysis of firm-internal factors. Emerging work, accentuating the central role of firm resources and capabilities as basis for strategy and a determinant of performance, coalesced into the so-called resource-based view of the firm (Barney 1991). In the remainder of this section, we provide an Overview of the Resource-based View (I). Next we summarize results from

Empirical Studies (II), testing the relevance of resources for firm performance. Finally, we conclude our analysis of the resource-based view with some Conclusions & Criticism (III) of this field of research.

Overview of the Resource-based View (RBV)

Research into firm-specific sources of performance is not a recent field of strategic research. The roots of the emerging wave of resource-based studies in the early 90s can be traced to the very beginning of management theory. Our review of the resourcebased view will therefore start with a brief description of the Development of the RBV (I), before summarizing the Main Hypothesis of the RBV (II). Next we provide more detail on the differences between individual Types of Resources (III), explore the main

Characteristics of Advantage-Generating Resources (IV), and establish the Relationship between Resources and Superior Performance (V). Development of the RBV: Early roots of the resource-based view can be found in the work of Chamberlin (1933) who identified key capabilities of the firm such as technical know how, reputation, brand awareness, patents, and trademarks, that have also been at the center of the recent debate (Hall 1992, Day 1994). Similarly, Coase commented on the importance of the "allocation of resources in a firm (. ..)" (1937, 389). Selznick (1957) was the first to introduce the term "distinctive competence" to describe unique strengths a firm possesses and which are central to competitive success. The importance of internal firm resources was stressed by most early management theories (e.g. Chandler 1962, Ansoff 1965, Learned et al 1965). Andrews (1971) recommended

61

that an "internal appraisal of strengths and weaknesses, led to identification of distinctive competencies". However, Penrose's (1959) was the first to propose a resource-based

explanation of the firm. The author argued that the heterogeneity of each firm's resource base determines the differences between the individual firms. Firms achieve superior performance due to the underlying resources and distinctive competencies that enable them to leverage these resources. While market-oriented approaches dominated the 70s and 80s, at least two alternative theories persisted in ascribing an important influence to unique firm resources. First of all the efficiency school (Demsetz 1973) stressed the role of firm-specific capabilities, including superior innovativeness or managerial skills as the ultimate source of performance. However, as described above, this school of thought quickly shifted its attention to the intermediate (and better observable) variable of market share. A second research field, the Austrian school of economics, emphasized intangible resources as determinants of business performance (Kirzner 1979). Despite various references to firm-specific resources and capabilities in the strategic literature, it took until the early 1980s before researchers began to consistently develop the resourcebased view. Wernerfelt's article published in 1984 is often seen as the "rebirth" of the resource-based perspective. Main Hypothesis of the RBV: The essence of the resource-based theory is the hypothesis that a firm that effectively establishes and deploys certain firm-specific resources can achieve a sustainable competitive advantage leading to superior economic returns (Barney 1986, Rumelt 1987). This hypothesis is based on three basic assumptions. First, the resource-based view assumes that resources are heterogeneously distributed across competing firms in an industry (Penrose 1959). This assumption differs from the assumption of firm homogeneity within an industry (or strategic group) of the classical microeconomics or industrial organization. Second, the differences in resource distribution between firms are considered stable over time (Barney, 1991). This implies that if competition in factor markets is not perfect, superior rents can be sustained (Barney 1986). The sustainability is explained by the existence of barriers to imitation (Lippman / Rumelt 1982, Rumelt 1984, Reed / Defilippi 1990). Barriers to imitation are similar to the concept of entry barriers in industrial 62

organization, but they are specific to the firm and thus not vulnerable to free riding (Mahoney I Pandian 1992). Reed and Defilippi (1990) introduced the concept of causal ambiguity to explain the emergence of barriers to imitation. The main argument is that

firms are composed of complex bundles of numerous tacit resources specific to the firm which make it difficult for competitors to identify what resources have been ultimately responsible for the firm's superior performance. This causal ambiguity impedes imitation and allows the firm to preserve its resource advantage. Even if the advantageous resources are identifiable, there are often barriers that hamper imitation, like patents, copyrights or the requirement of large-scale investments (Hall 1992, Collis I Montgomery 1995). Finally, not all resources are considered a source of competitive advantage. Resources must fulfill a set of requirements that determine the strategic value of the resource in generating superior returns (Dierickx I Cool 1989, Barney 1991, Schoemaker I Amit 1994). We will review these characteristics after a short introduction into different types of resources. Types of Resources: The first challenge in reviewing the resource-based literature is to

resolve the utter confusion caused by the multitude of terms applied to describe firm resources. Numerous scholars have dedicated ample time to let the interested reader in on the subtle differences between a variety of labels including competencies (sometimes "core" or "distinctive"), capabilities (also whether "core" or "distinctive"), assets ("invisible" or "intangible"), and skills (or "core" skills). To further complicate the issue, other authors stressed the interchangeability of some of these terms. To surmount this ambiguity, we define the term "resources" as embracing all firm-specific assets and capabilities. We further distinguish three subgroups of resources: (I) tangible assets, (II) intangible assets, and (III) capabilities (Fahy 2000). Tangible assets are the fixed and current assets of the firm (Wernerfelt 1989). These assets can be relatively easily observed and the value of these assets can be measured and is reflected in the balance-sheet valuation of the company (Hall 1989). Examples include plants, equipment, land and capital goods. Intangible assets on the other hand, comprise trademarks, patents, brands, knowledge, and company reputation. Deployment of intangibles does not generally lead to asset depreciation (Wernerfelt 1989). The value 63

of a company's intangible assets is reflected by the difference between the balance sheet valuation and the stock market valuation of the firm (Grant 1991). Finally, capabilities comprise the skills of individuals and groups in the firm, as well as the organizational routines and interactions that ensure the coordination between individuals and groups (Grant 1991). Examples include firm culture, teamwork, and trust. These capabilities are difficult to capture and to valuate as they often reside within individuals or groups. Characteristics of Advantage-Generating Resources: One of the principal insights of

the resource-based literature is that there are significant differences in the potential of individual resources to generate sustainable competitive advantage. Various scholars have thus attempted to identify the characteristics of resources generating advantage. Barney (1991) proposes four essential requirements for a resource to be a source of sustainable competitive advantage. The resource has to be (I) valuable, (II) rare, (III) imperfectly imitable, and (IV) non-substitutable. In order to be valuable, the resource has to permit the firm to develop or implement strategies that improve its efficiency and effectiveness by fulfilling customer requirements. In addition, the resource has to be rare; its possession has to be limited to few competitors in the market. Above all, it is central that the resource advantage cannot be imitated or substituted. This condition refers to the concept discussed above of barriers impeding imitation such as causal ambiguity, patents and trademarks. Barney's work can be linked to the three different types of resources distinguished above: tangible assets, intangible assets, and capabilities. Tangible assets are visible and thus relatively easy for competitors to imitate (Grant 1991). Due to their inherent complexity and specificity, intangible assets are far more difficult to imitate or substitute (Itami 1987). However, capabilities are attributed as having the greatest importance for advantage-generation (Collis 1994). Capabilities can be regarded as complex bundles of intangible resources that reside in various individuals, groups, and organizational routines and interactions. Due to this "causal ambiguity", capabilities are extremely difficult to imitate and they are highly valuable as they enable the firm to constantly renew and expand its resource base Barney's work triggered an avalanche of studies identifying various conditions for advantage-generating resources (Grant 1991, Peteraf 1993, Amit / Schoemaker 1993, Collis / Montgomery 1995). There is significant overlap between the individual 64

approaches. One important factor missing in Barney's approach is the question of appropriability (Collis / Montgomery 1995). Intangible resources are often tied to a "host" such as individual employees. These skills, like managers' individual knowledge, can "walk out of the door" when an employee leaves the firm and goes to work for the competition. Therefore, it is important, that the firm has some lever to appropriate the resource. Relationship between Resources and Superior Performance: The resource-based view introduced the concept of sustainable competitive advantage. Advantage is seen as relative to the firm's competitors (Kay 1993). A firm can either possess a differentiation or cost advantage (Porter 1985) or both (Gilbert / Strebel 1989). Competitive advantage is expected to lead to superior economic performance (Bharadwaj 1994). Whether this advantage will be sustainable over time depends on the firm's specific resource base (Barney 1991). The firm's needs to possess resources that fulfill the characteristics of advantage-generating resources described above. However, resources alone do not lead to a sustainable competitive advantage. It is crucial that the firm's resource base is deployed to implement strategies that create value for the customer and acts to defend existing assets and create new ones (Dierickx / Cool 1989, Williams 1992, Amit / Schoemaker 1993, Bharadwaj 1994). In summary, advantage-generating resources have to be deployed to implement strategies in order to gain a sustainable competitive advantage that leads to superior firm performance. Figure 6 provides an overview of the argumentation of the resource-based perspective.

Key Resources

Figure 6

f---

Sustainable Strategy

r--+ Competitive Advantage

The Resource-based Model

65

-

Performance

Empirical Studies

As shown above, the resource-based view claims a direct and important influence by firm-specific resources on a firm's competitive position and thus performance. In order to evaluate this hypothesis, we will first review studies that have empirically assessed the overall relationship between resources and performance and the relative importance of firm-internal resources versus other factors in explaining performance. Once the relevance of resources for company performance has been established, it becomes vital to identify the most important individual resources linked to performance. We will thus focus our attention on reviewing the empirical results of studies that analyzed the influence of individual resources on company performance. Firm versus Industry Effects: Since the advent of the resource-based view, much attention has been paid to the empirical analysis of the relative importance of firm resources and industry effects in explaining performance (Claver et al 2002). Two of the pioneering studies, Schmalensee (1985) and Rumelt (1991), used variance component analysis to study differences in firm performance related to firm and industry effects. Schmalensee found that industry effects had a strong dominance. He concluded that industry effects account for 20 percent of profitability variance, market share for 1%, the remaining 80% were assigned to error. Firm effects on the corporation level were found irrelevant. Rumelt criticized Schmalensee's study on methodological grounds. For instance, he stated that the use of profitability data for a single year makes it impossible to detect firm effects. In order to surmount these shortcomings, Rumelt used multi-year and cross-sectional profitability data and business unit level measures for firm effects. Rumelt found a dominance of firm (business unit) effect over industry effect. Other early studies include Cubbin and Geroski (1987), Wernerfelt and Montgomery (1988), as well as Hansen and Wernerfelt (1989). Cubbin and Geroski found that almost half of the firms they analyzed showed movements in their profitability levels independent of general industry movements. In line with these findings, Hansen and Wernerfelt concluded that firm effects are twice as important as industry factors. However, Wernerfelt and Montgomery found support for a stronger industry effect. Hence, in line with conflicting theoretical guidance, early studies reported contradictory findings on the importance of industry and firm effects (Mauri / Michaels 1998,211).

66

More recent variance decomposition studies have applied more sophisticated methods and thus achieved more consistent results. Roquebert et al (1996) applied a model, similar to that of Rumelt (1991), to manufacturing industries, using ROA data from the COMPUSTAT database over a seven-year period. According to their results, firm effects are three times more important than industry effects. McGahan and Porter (1997) used the same model in an extensive cross-industry analysis of over 12,000 business units in more than 600 different industry segments with performance data for a 14-year period. In respect of manufacturing firms, they confirmed the findings of Roquebert et al: firm effect was more than three times higher than industry effect. However, across all industry segments, they found a more balanced ratio: firm effect was 1.5 times more important than industry effects. In some industry segments, industry even proved to be much more important than firm effects. Mauri and Michaels (1998) confirmed the high results in respect of manufacturing companies and found that firm effect is almost six times more important than industry effect (this ratio drops to four times over a longer period). Chang and Singh (2000) used market share as the dependent variable and found similar evidence. In a cross-industry analysis of 444 different segments, they obtained results that business unit effect is three times more important than industry effect. However, the authors show that the results vary significantly for different segments and dependent on firm size. Claver et al (2002) contributed additional evidence analyzing a set of nondiversified Spanish manufacturing companies. They found that firm effects were eight times more important than industry effects. Finally, Hawawini et al (2003) confirm the predominant importance of business unit level factors, but they also show that these results are highly influenced by a few outperformers and underperformers in the industry. In respect of the large majority of firms, both industry and business unit level factors are almost equally important. To end this review of variance decomposition studies, it can be concluded that there is broad evidence that firm effects are indeed an important source of superior performance (Claver et al 2002, McGahan I Porter 2002). Recent studies managed to explain approximately 50% of the variance in firm performance. Approximately 30-45% in performance differences has been attributed to firm-specific resources and capabilities at the business unit level (compared to 5-20% to industry effects). The remaining percentage has been accredited to differences in corporate & business strategies, and other omitted factors yet to be discovered, or to sheer luck (Claver et al 2002). Despite 67

considerable differences between individual industry segments and firm sizes, the general consensus is that both firm-specific and industry effects are important in explaining the variance of firm profitability (Henderson / Mitchell 1997, McGahan / Porter 2002). Based on this consensus, the next challenge is to identify more specific resources and capabilities that drive performance. Only fine-grained analysis of resources at lower levels of aggregation allows managers and scholars to identify the true sources of superior performance (Mauri / Michaels 1998, 217).

Key Resources and Capabilities: Variance decomposition studies have Significantly supported the general acceptance of the resource-based perspective (Levinthal 1995, 20). However, the analysis in these studies remains at the macro level. None of the authors made claims about the individual factors and processes underlying their results (McGahan / Porter 1997, 15). In order to better understand the role of resources within an industry and to provide managers with the means to take action, it is vital to concentrate on concrete internal factors underlying performance. This research has been complicated

by the

fact that

intangible

resources

and

capabilities

constitute

"unobservable" factors (Godfrey / Hill 1995) due to their high degree of tacitness and causal ambiguity. The less a resource can be identified and understood, the greater the likelihood that it is an important source of superior performance. Following this logic, strategy researchers must often fail to discover influential and concrete resources (Collis 1994). Early empiric research has thus been criticized as conceptually vague and tautological, providing little support for a linkage to superior performance (Williamson 1999, Priem and Butler 2001). Initial work examining concrete capabilities in distinct functional areas (e.g. Snow / Hrebiniak 1980, Hitt / Ireland 1986) was therefore increasingly displaced by efforts to identify more conceptual "meta capabilities" and their impact on performance. Grant (1991) argued that the most important capabilities are likely to arise from an integration of individual intangible assets and capabilities. Hamel and Prahalad (1990) used the term "core competencies" to describe these central capabilities. Examples in the literature of such central capabilities include market orientation (Slater / Narver 1994), brand recognition (Aaker 1991), firm reputation (Fombrun / Shanley 1990), corporate culture (Kotter / Heskett 1992), knowledge (Winter 1987), and product quality (Anderson et al 1994). As mentioned above, the resource-based perspective failed to agree on a

68

common framework to classify and evaluate these resources and capabilities. The first theoretical studies on the media industries acknowledged the importance of core capabilities for success in the media industries, but remain limited to generic reflections (Miller / Shamsie 1996, Campling / Michelson 1998, Habann 1999 / 2000 / 2001). Consequently, we do not attempt to provide an exhaustive overview of potential valuegenerating factors at this time. For the moment we concentrate on the three factors that most attracted other researchers' attention and that have been repeatedly proven in empirical studies as central in explaining performance: (I) market orientation, (II) reputation, and (III) corporate culture. Market Orientation: The marketing concept to relate customer orientation to firm

success was first articulated by Peter Drucker in 1954. However, it took until the 1980s until the concept of market orientation emerged in the wake of the resource-based literature (Shapiro 1988, Day / Wensley 1988). As defined by Slater and Narver (1999, 1165), firms with a strong market orientation "seek to understand customers' expressed and latent needs, and develop superior solutions to those needs". Approaches to the market orientation concept can be classified into two perspectives: (I) market-oriented culture and (II) market-oriented activities. The cultural viewpoint highlights the need for an organization's culture to be oriented towards markets, customers, and competitors (Deshpande / Webster 1989, Slater / Narver 1995, Turner / Spencer 1997). From an activity perspective, it is crucial to implement organization-wide processes to collect market intelligence about customers and competitors, to disseminate this intelligence across departments, and integrate it with strategic decision-making processes (Kohli / Jaworski 1990, Deng / Dart 1994, Jaworski et al 2000). Hult and Ketchen conclude that the main characteristic of market orientation is "system-wide attention to markets (customers,

competitors and other entities in the environment) throughout the

organization" (2001, 901). Day (1994) positioned the market orientation approach in the resource-based tradition. He understands market orientation as a combination of two distinctive capabilities: market sensing and customer linkage. A firm that possesses these capabilities has a higher market focus. The increased market focus permits tracking and satisfying customer needs better by creating superior value for them (Kohli / Jaworski 1990). This in turn leads to a competitive advantage and superior performance (Day 1994). The 69

basic argumentation is thus that market orientation will lead to superior company performance. Narver and Slater (1990) developed scales to empirically test this assumption. In an empirical survey of 140 SBUs of a single corporation, Narver and Slater found that market orientation had a substantial positive effect on profitability. Ruekert (1992) confirmed these findings in a questionnaire survey of 400 managers of a high tech firm. Both studies were criticized for their restriction to a single company (leading to a common respondent bias problem). However, the growing body of research that followed has provided unambiguous support for a strong relationship between market orientation and various measures of performance (e.g. Greenley 1995, Pelham 1997, Van Egeren I O'Connor 1998, Kumar et al 1998). Slater and Narver themselves confirmed the importance and generality of their initial work in a later study drawing on a more diverse population (Slater I Narver 2000). Most of these studies draw a direct and linear link between market orientation and performance. Recently, a growing number of studies emerged that analyze the moderating influence of other variables on the relationship between market orientation and performance. We will review some of this work later in this study. Reputation: In line with Weiss et al (1999, 75) reputation can be defined as "a global perception of the extent to which an organization is held in esteem or regard". Reputation

is determined by the firm's past actions. Reputation assets include brand equity, corporate reputation, and corporate image. Examples of companies with a particular reputation in the media industry include Disney and AOL, which are repeatedly cited amongst Fortune's Most Admired Corporations. Reputation has been emphasized as an important strategic asset by scholars of the resource-based perspective. The intangible and complex character of reputation makes it extremely difficult to imitate this asset in the short term (Rumelt 1987, Dierickx I Cool 1989, Hall 1993). The firm's ability to drive reputation is embedded inside the firm and associated with a high degree of causal ambiguity (Dowling 2001). Consequently, the differences in reputation between firms have remained relatively stable over time (Barney 2001). Reputation is a valuable asset due to its positive effect on both the product differentiation and cost advantage of the firm. Companies with high reputation can charge a certain premium due to customers' preference for products of these firms (Klein I Leffler 1981, 70

Milgrom , Roberts 1986). This effect is explained by the perception of reputation as a signal for underlying product quality and is expected to be particularly strong in markets with high levels of uncertainty. The cost advantage of firms with a high reputation stems from the best resources' increased motivation to work for such a firm and to work harder for the same or a lower remuneration (Stigler 1962). In addition, suppliers are less anxious when contracting high reputation firms, leading to lower contracting and monitoring costs (Roberts' Dowling 2002). Podolny (1993) states that the positive correlation between high reputation and these beneficial outcomes creates a "virtuous circle" that motivates firms with a good reputation to further enhance their reputation. This provides additional support for the relative persistence of reputation over time as stated above. The manifold benefits of a good reputation and their persistence over time provide the ground for the assumption of a strong positive relationship between reputation and performance (Fombrun 1996). A growing body of empirical studies confirmed the link between reputation and the associated benefits (McGuire et al 1990, Fombrun , Shanley 1990, Herremans et al 1993, Podolny 1993, Landon' Smith 1997, Roberts' Dowling 2002). For instance, Roberts and Downing (2002) found evidence of a sustainable superior performance by firms with a strong reputation. Organizational Culture:

The organizational culture phenomenon

has attracted

considerable research attention in such diverse fields as organizational behavior, sociology, anthropology, and management studies. Deshpande and Webster reviewed more than 100 studies and concluded that organizational culture can be defined as "the pattern of shared values and beliefs that help individuals understand organizational functioning and thus provide them with the norms for behavior in the organization" (1989, 4). Early studies have already drawn the link between culture and performance. In the late 1920s, the well-known Hawthorne studies emphasized the central importance of culture for work group productivity (van der Post, de Coning 1998). Drucker concluded that management and culture are closely linked. Managers should thus actively influence and shape company culture (Drucker 1973). Despite these early roots, it was not until the late 1970s that broad research on organizational culture emerged (Le. Ouchi 1981, Pascale' Athos 1981, Deal' Kennedy 71

1982, Schein 1992, Peters / Waterman 1982, Denison 1990). The central conclusion of this body of work is that while all firms have organizational cultures, some have stronger cultures than others leading to superior performance (Kotter / Heskett 1992). A culture is perceived as strong if the values of the culture are widely shared and strongly held within the organization. Successful organizations are thus distinguished by their ability to create strong cultures that are consistent with their chosen strategies (Peters / Waterman 1982). Two empirical studies confirmed the general relationship between strong cultures and performance (Denison 1990, Gordon / Di Tomaso 1992). However, both authors argue that sustainable competitive advantage can only be achieved if the culture is sufficiently flexible to adapt to changes in environmental conditions and has unique qualities that cannot be imitated. This is in line with Barney's argument that the value of culture as a strategic resource depends upon the rarity, imitability and sustainability of the culture concerned (Barney 1986/ 1991). While additional studies confirmed the importance of culture for firm performance (Marcoulides / Heck 1993; van der Post / de Coning 1998), the research interest turned towards the forms or characteristics of well-performing cultures. Gordon and DiTomaso (1992) found that adaptable and flexible cultures led to better performance than stabile cultures. Deshpande et al (1993) extended this research by analyzing the relationship between four culture types and performance. Along the two dimensions organic (high flexibility, spontaneity and individuality) to mechanistic processes (high control, stability and order) and internal (smoothing activities, integration) to external pOSitioning (competition, market focus), they distinguish four culture types labeled clan, hierarchy, adhocracy, and market. While most firms do have elements of all four types, one dominant culture type is expected to emerge over time. The assumption is that the four culture types imply varying degrees of business performance. In an empirical study Deshpande et al confirmed that all four types are well represented and that market cultures are associated with the best performance, followed by adhocracy. Both clan and hierarchical culture are linked to poor performance. In a later empirical study Deshpande and Farley (1999), this time using a classification schema for culture types developed by Quinn (1988), confirmed the primacy of competitive and entrepreneurial cultures over bureaucratic and consensual cultures in generating superior performance. In a study of UK firms, Ogbonna and Harris (2000) found a strong and direct link between competitive and innovative forms of culture and organizational performance, while bureaucratic and 72

community cultures were not directly related to performance. Homburg and Pflesser (2000) analyzed 1,100 German SBUs and found support for a strong relationship between market-oriented and innovative cultures, on the one hand, and performance, particularly in dynamic environments, on the other. Some recent studies emphasized the importance of organizational culture for media companies (Daymon 2000, Elashmawi 2000, Kung 2000, Li-Chuan 2001).

Conclusions & Criticism

The important contribution of the resource-based view to the field of strategic management was acknowledged by the Strategic Management Journal's 1994 best paper award to Birger Wernerfelt. The resource-based literature has greatly enhanced the understanding of firm-internal sources of competitive advantage. Where scholars of industrial organization failed to explain performance differences within industries, or within the narrower confines of strategic groups, resource-based theory managed to relate these differences to resource asymmetries. The research into the conditions for advantage-generating resources helped to explain why these asymmetries persist over time, even in conditions of open competition. We have shown before that numerous variance decomposition studies showed the significant influence of organizational factors. However, despite this overall empirical support, the RBV's contributions remained conceptual rather than empirical. This can be explained by the often tacit and "unobservable" character of resources (Godfrey I Hill 1995). Faced with the complexity of a real firm, it is often difficult, if not impossible to detect those resources that account for superior performance. Recently an increasing number of empirical studies emerged that focus on more observable "meta capabilities". The problem is that firm management finds these relatively abstract constructs difficult to capture and actively manage. The main criticism voiced against the resource-based perspective is that of circular reasoning and tautology (Williamson 1999, Priem I Butler 2001). The identification of core resources often has an "ex post quality": successful companies are analyzed for underlying resources and these then labeled as valuable. This tradition has some

73

inherent shortcomings. First, the identified resources are only valuable within the examined context and situation. Second, the process through which the resource provides competitive advantage remains in a black box (Priem I Butler 2001). Third, resources can be turned into disadvantages (or "core rigidities") swiftly when environmental conditions shift (Leonard-Barton

1992). Finally, the purely static

description of "valuable" resources does not reveal how these strategic assets have been created in the first place. This criticism can be surmounted by adopting a more "dynamic" approach that focuses on the firm's ability to develop new resources. In the following chapter, we review the emerging literature of dynamic approaches to strategic analysis. Our review comprises a discussion of recent studies concerned with the dynamic evolution of firm capabilities in the tradition of the resource-based view. This body of work addresses most of the criticisms raised above.

74

III.

Towards a Dynamic Research Model

Chapter two provided a detailed review of four different schools of thought that dominated the strategy literature over the past four decades. Extensive research helped to establish sound theoretical foundations in each research field. Based on these foundations, each approach claims to provide a satisfactory explanation for performance differences between firms. However, with four competing paradigms, one question is particularly interesting: who is right? Can performance be fully explained by industry characteristics as claimed by industrial organization economists? Or do the firm's unique capabilities make the difference? Does corporate strategy play the decisive role or is it perhaps strategic moves on the business unit level? Simple observation of the media industries indicates that the reality might be far less clear-cut than the textbook approaches described above. Let us consider a concrete example from the media industry as an illustration: the recent performance of the US media company Liberty Media, for instance. The company reported significant operating losses in 2001 and 2002. Liberty Media had invested US$ billions in the quest to gain control of the European cable market. Both UPC and Telewest, Liberty's major investments in Europe, are currently in a financial restructuring process after accumulating debts of more than US$ 20 billion. The Dutch cable company UPC filed for bankruptcy protection as part of the rescue refinancing. Analysts repeatedly related the drop in Liberty's performance to the failure of the company's international diversification strategy. Consequently, one could assume a linkage between international diversification and performance in the tradition of the corporate strategy literature. However, if we take a closer look, the picture is rather more complex than this. First of all, the failure of the European expansion can subsequently be explained by the complete collapse of Liberty's business strategy. The magic formula applied in the US a business model based on the three pillars market consolidation, triple-play, and programming control -

failed overseas. Market consolidation is complicated by

fragmented markets and a thicket of regulatory and cultural barriers. Triple-play, a

75

differentiated product bundle of digital TV, high-speed Internet, and telephony services, is challenged by hefty competition from entrenched telecom monopolies and wireless operators. Finally, the strategy to use cable networks as an outlet for Liberty's own programming assets put Liberty at loggerheads with regulators and a coalition of both public and private broadcasters. Without going into greater detail, the example shows that explaining Liberty's performance purely from a Single perspective - corporate strategy - fails to provide the full picture. On the contrary, both corporate and business strategies, as well as differences in market structures between Europe and the US, playa significant role in explaining Liberty's operating performance. Moreover, all these factors are related. Triple-play cannot be that bad a strategy if Liberty chairman John Malone made a fortune employing it in the US. However, the success of this strategy depends on the present market structures, the available capabilities, as well as the timing. What had been right in the US in the 90s is not necessarily the right strategy for Europe ten years later. This example provides a rather more complex and dynamic picture of factors underlying performance than the classic strategic management theories described above. In our quest to establish a more complex and dynamic picture of company performance, we thus continue with a review of the emerging (111.1) Criticism of Classical Approaches to uncover the shortcomings of the mainstream approaches in capturing the new market realities. Thereafter we review the disjointed body of recent stUdies addressing these shortcomings under the heading of (111.2) The Emerging "Dynamic School". Based on both classical and dynamic approaches, we continue to establish the (111.3) Foundations of the Dynamic Research Model. Finally, the last section, (111.4) Model Establishment and Hypotheses, describes our model for dynamic strategic analysis that will be applied to

the global media industries at a later stage.

111.1

Criticism of Classical Approaches

Recently, numerous scholars have described growing complexity and dynamism as an important new context in strategic management. Prahalad and Hamel (1994) provide an overview of the forces impacting the nature of competition, including deregulation, 76

excess

capacity,

increasing

consolidation

and

diversification,

technological

discontinuities, and (like in our Liberty example) globalization. These changes lead to the firm experiencing an increasing external and internal complexity (Lowendahl I Revang 1998, Pettigrew et al 2002). Companies operate in a growing number of local and vertical markets that subsequently require complex internal procedures for organization and decision-making. Alavi et al conclude, "institutions seek to adapt to changes in an increasingly

complex

and

dynamic

world'

(1997,

1329).

Furthermore,

these

developments imply intensifying competition that forces companies to permanently review and adjust their strategy. D'Aveni (1994) coined the term "hyper-competition" to describe these competitive dynamics (see also Thomas 1996, Brown / Eisenhardt 1998, Hamel 2000). In the light of these changes, criticism of the classical linear and one-dimensional approaches in strategic management has emerged. Various authors postulate a more complex and dynamic - and thus more realistic - understanding of strategy. Despite the rich foundations of strategic management, three significant shortcomings have been increasingly criticized: The (I) Fragmentation, (2) Determinism, and (3) Static View of traditional approaches. Fragmentation: The example of Liberty Media outlined above shows that company

performance should be traced to a multitude of influencing factors rather than the onedimensional explanations of traditional approaches in strategic management. This finding is very much in line with the results of recent studies in the strategy literature. As mentioned before, so-called variance decomposition studies measure the relative importance of factors at the industry, corporate, and business unit levels in explaining differences in firm performance. Recent studies provide clear evidence that factors at all three levels significantly influence profitability (Roquebert et al 1996, McGahan / Porter 1997, Chang / Singh 2000, Bowman I Helfat 2001, McGahan / Porter 2002). McGahan and Porter (2002, 850) reconcile different studies and conclude: "The robust findings suggest that the research has successfully shown that industry, corporate-parent, and business-specific influences are all importanf'. McNamara et al (2003, 274) confirm this view: "We find evidence that stable factors at industry, corporate, and business-unit levels (. ..) all significantly affect business performance". In other words, performance cannot be sufficiently explained by factors at one level of analYSis alone. Much along the 77

same lines, Porter states: "Competitive advantage in more sophisticated industries and industry segments (. . .) rarely results from strength in a single determinanf' (1991, 114). While there is renewed support for this argument, researchers have always argued that the success of an enterprise rarely depends on a single factor alone (Thompson 1967, Schendel/Hatten 1972, Lenz 1980). However, traditional approaches in strategic management are often focused on single aspects or perspectives. Industrial organization economists focus their analysis on industry effects, often neglecting firm-specific differences. Studies into diversification are mostly limited to corporate-level effects. The resource-based perspective remains widely focused on internal factors, thus disregarding industry effects. In general, most strategy scholars have chosen to limit their observations by specializing in the level of analysis and in the disciplinary frame of reference (Pettigrew et al 2002, 3). Due to paradigmatic constraints "each successive model seemingly ignored the contributions of its predecessor" (Shay / Rothaermel 1999, 561). Past results are "pushed off-stage by the hard edge of exclusivity of a new paradigm or frame of reference" (Pettigrew et al 2002,

9). Consequently, "research has tended to fragment or dichotomize the important parts of the problem rather than integrate them" (Porter 1991, 95). Many other authors share Porter's criticism (Le. Schendel 1994, Farjoun 2002). Sanchez and Heene (1997, 304) state that the traditional theory base for strategic management has become "progressively more fragmented into multiple unconnected streams of research and practice". More precisely, many authors criticize the "split personality" of strategiC management with one stream of theory ignoring internal aspects of organizations, while a second major stream largely ignores the firm's competitive interactions (De Wit / Meyer 1994, Grant 2002). Both streams, industrial organization economics and business policy, have contributed important insights, but remain unconnected conceptually and in practice (Hansen / Wernerfelt 1989, Sanchez / Heene 1997). Derived from industrial organization and the SCP paradigm are the five forces model and the industry life cycle model that became the dominant models with which to analyze the firm's external environment (Porter 1980). The value chain model, on the other hand, has become the standard model for analyzing the internal side of the firm (Porter 1985, Barney 1997). Henderson and Mitchell (1997) conclude that both sides have typically been content with thoroughly 78

analyzing one side while applying very simple models to capture the other. Hardly any research has simultaneously considered strategy, organizational, and environmental constructs within an industry (Parnell et al 1996, 41), having rather focused on a different element of the strategy picture (Farjoun 2002, 566). The fragmented and polarized theoretical concepts forming the foundation of traditional strategic management theory "bear little resemblance to new, more dynamic modes of competition evident in many industries" (Sanchez / Heene 1997, 303). The authors thus call for a framework that integrates the "many useful conceptually unconnected insights developed in prior strategy theory" (Sanchez / Heene 1997, 304). This claim is in line with Shay and Rothaermel (1999, 559), promoting an integrative and multi-perspective approach that "allows us to bridge the gaps between the competing strategic analysis models". They continue to stress the interrelatedness of the different streams of theory and the potential that exists for synergies to be deployed through integration. Various other

authors

have

emphasized

this

potential

for

cross-fertilization

between

complementary approaches (Hansen / Wernerfelt 1989, Barney 1992, Zajac 1992, Seth / Thomas 1994, Pettigrew et al 2002). McNamara et al (2003) conclude that industrial organization, corporate, and resource-based perspectives all have the potential to contribute to the explanation of business performance. Researchers and managers should thus "address industry-, corporate-, business-, and other-level factors that together drive business success or failure" (McNamara et al 2003, 274). Stimpert and Duhaime (1997) recognize a critical gap in the strategy literature that could be filled by a comprehensive framework that integrates the industry characteristics, as well as corporate and business strategy. Hoskisson et al (1999) conclude their review of four decades of strategy research with the criticism that multiple theoretical paradigms lack integration. The authors particularly pOint out the need for integration of research on business- and corporate-level strategy. Dess et al (1995, 358) make the same claim and stress that "research in strategic management cannot and should not be confined to any one level, given their interdependent and interconnected nature". They conclude that integrative research studies across two or more levels of strategy represent "some of the most fruitful avenues for future investigations in strategic managemenf' (Dess et al 1995, 388).

79

Ironically, the request by Hoskisson et al (1999) and by Dess et al (1995) is almost identical to the appeal for integrated research as voiced 15 to 20 years earlier by Beard and Dess (1981, 686). In the same year, Jemison (1981, 601) stated "strategic

management has reached the point where integrative research approaches are necessary for continued progress in the field'. Jemison argues that strategy research has generated a variety of research streams, each based on different paradigms, units of analysis, and causal assumptions. While each stream provides valuable insights, there remains "enormous potential for cross-fertilization ofthese ideas" (Jemison 1981, 602). Jemison's judgment of the situation and the need for integration is identical to the various appeals emerging over the last decade, as discussed above. Despite two decades of criticism, not much has been done to overcome the fragmentation. One of the few frameworks, and certainly the most common one, that bridges the division between internal and external analysis is the SWOT model (Andrews 1971). Despite its age and certain theoretical limitations, the framework remains a primary consulting and teaching tool at business schools (Hill / Westbrook 1997, Barney 1997) - it could be argued through a lack of alternatives. However, the increasingly complex and dynamic conditions faced by companies in various industries today, increase the need for a more integrated approach to strategic analysis. Determinism: Integrating the different approaches certainly provides a richer picture, but

still fails to capture reality entirely. In our Liberty Media example, we have seen that the different factors influencing performance also exhibit significant interaction. For instance, the success of Liberty's business strategy is influenced by the respective market structures. This finding is again supported by the variance composition studies mentioned above: "The results not only confirm the importance of each of the effects,

they also point to important relationships between the effects" (McGahan / Porter 2002, 848). These studies find covariance effects between corporate, industry, and business unit effects (Bowman / Helfat 2001, McGahan / Porter 2002). Hansen and Wernerfelt (1989, 409) recommend moving "beyond value decomposition studies and consider

various interactions between economic and organizational variables". A recent special issue of the Journal of Business Research (Vol. 51, No.3) published a number of articles providing evidence that factors "interact in complex ways in affecting

performance outcomes" (Phelps 2001, 168). In an earlier special issue of the Strategic 80

Management Joumal (Vol. 18) dedicated to the same subject, Henderson and Mitchell (1997) pointed to the reciprocal interaction between market environment, firm capabilities, strategy and performance. Game theoretical models focus on this strategic interaction (Brandenburger / Nalebuff 1996). The reciprocal causality between these factors is also implicit in the notion of dynamic fit (Itami 1987). Porter (1991, 113f) recognizes that the effect of one determinant depends on the state of others and models these "mutual reinforcement of the determinants" into his framework. Much as with the lack of integration discussed above, numerous scholars remarked on the

research

gap

regarding

interaction

between

individual

factors

influencing

performance. However, much less has been done to actually fill this gap. Stimpert and Duhaime (1997) maintain that the strategy literature has generally failed to examine how factors like industry characteristics, corporate, and business strategy influence one another. They continue to remark, "few studies have considered or empirically tested the possibility that these factors might also have important indirect influences on performance" (Stimpert / Duhaime 1997, 561). Henderson and Mitchell share this judgment and conclude that studies of the mutual influence between capabilities and competition "could be one of the next great opportunities for the field of strategy research" (1997, 6). The theoretical differentiation between research streams has certainly hampered the recognition of the multiple and reciprocal causality between the individual elements (Stimpert / Duhaime 1997, Farjoun 2002, Pettigrew et al 2002). Phelps (2001, 168) states the "need for complex models involving interacting factors" in order to understand dynamic realities. Dess et al (1995, 383) opt for an "interdependency approach" that captures the effect of strategic variables at one level on those at another level, and their combined interactive effect on performance. Others establish a need to go beyond stating the existence of interactions between individual factors by analyzing the exact mechanisms through which interaction occurs (Rindova / Fombrun 1999, Farjoun 2002). Static View: Integrating different approaches and taking the various interactions between determinants into account can provide a quite realistic picture. However, a picture still doesn't show as much as a movie since determinants change and evolve over time (Le. Porter 1991, Thomas 1996, Teece et a11997, McGahan / Porter 2002). As previously stated, what had been a successful strategy for Liberty Media ten years ago 81

might no longer be appropriate. The competitive landscape in cable television has changed

due

to

technological

advances,

regulatory

changes,

and

growing

internationalization. Despite the obvious presence of change processes, strategic analysis's classical approaches share the common belief that at any given time theoretical assumptions correspond to empirical conditions (Jacobsen 1992). These approaches focus on a single point in time "without specifying the dynamics through which these outcomes develop" (Barnett I Burgelman 1996, 5). This implies "historical efficiency": Cause-effect relationships quickly result in steady state equilibrium, regardless of the historical path (March I Olsen 1989, Carroll I Harrison 1994). Strategic approaches are therefore "mostly associated with analysis of situations presumed to be at resf' (Nelson 1995, 48). Cockburn and his colleagues conclude that most frameworks allow "ex post accounts of the way in which a winning gambler chose to put her money on red rather than on black at the roulette table" (2000, 1124). These models explain the relative explanatory power of different determinants of performance in detail, but not how these entities have been formed (Jacobsen 1992, Cockburn et al 2000, Farjoun 2002). Porter (1991,105) states that these approaches "are agnostic as to the process by which the superior positions were attained". Shay and Rothaermel (1999) extend this criticism with their statement that traditional approaches are appropriate for present-time analysis, but not sufficient to project the future. However, the real challenge for strategy lies in creating tomorrow's competitive advantage today (Hamel I Prahalad 1989). Classical approaches fail to capture this increasingly dynamic nature of gaining competitive advantage: "Most strategic models are static and therefore void of any dynamic description" (Shay I Rothaermel1999, 561). Due to these shortcomings in classical approaches, a growing group of researchers postulates a more dynamic analysis (Le. Nelson I Winter 1982, Porter 1991, Nelson 1995, Barnett I Burgelman 1996, Sanchez I Heene 1997, Black I Farias 2000, Eisenhardt

I Martin 2000). This proposal can be traced to Marshall (1948), who at that time already criticized the static and mechanistic view based on the equilibrium assumptions of most economic approaches. He believed that science should aim to understand change and not only the forces underlying the current situation. Nelson (1995, 51) states more precisely that economic analysis has to be evolutionary and needs to show how a

82

particular equilibrium, and not another one came to be. Hill and Deeds (1996) add that a focus on competition as a process is more important than the simplistic analysis of competition under equilibrium conditions (Hill / Deeds 1996). Farjoun (2002) notes a growing interest in viewing strategy in dynamic and process terms. A comprehensive strategy model not only integrates different perspectives, "but also adds a dynamic dimension to strategic analysis" (Shay / Rothaermel1999, 561). This can be achieved by

an extension of the time period under examination (Hill / Deeds 1996, McGahan / Porter 2002, Pettigrew et al 2002). The present section has combined criticisms of strategic management's classical approaches from various sources. While significant insights into the different determinants of company performance have been gained, the field is still highly fragmented, ignores or oversimplifies the interrelation between determinants, and remains inherently static in its approach. The simple, mechanistic perspectives come into conflict with the increasingly sophisticated market conditions faced by more and more companies today. Today's business problems "require frame-breaking approaches" that can bridge the abyss that has grown between rival paradigms (liinitch et al 1996,218). A more complex and richer picture of strategy and its interactive relationship with environmental and organizational factors has to be established. While to date there's no comprehensive approach that simultaneously addresses the three research gaps outlined above, a number of studies have emerged over the last decade that provide inSights into individual aspects of the problem. In order to come closer to our research objective of establishing a dynamic model for strategic analysis, we continue with a review of the emerging body of work aiming at a more dynamic and complex understanding of strategy.

111.2

The Emerging "Dynamic View"

At first, we need to clarify that to date there is no such thing as a "dynamic view". As discussed above, there are various scholars who advocate a more complex and dynamic approach to strategy. Moreover, a growing number of publications respond to this appeal by addressing a variety of dynamic topics. However, despite their common focus on dynamism, these activities remain largely isolated from one another. This is due, first of 83

all, to the strongly differing definitions of "dynamic" inherent in the individual studies. Unfortunately, the cleavage goes considerably deeper than the simple question of controversial definitions. This dispute merely reflects the conflicting underlying epistemological assumptions on the nature of competition. The purpose of this section is to review and structure this fragmented field in order to contribute to the emergence of a "dynamic view" of strategy. We therefore classify the present literature into three research streams in accordance with their shared underlying beliefs. Initially, we review studies that are close to the classical approaches to strategy as presented in chapter two. These studies have two main orientations in common. First, in line with traditional approaches, they analyze competition under equilibrium conditions. Second, they are "dynamic", as they assume an inherent dynamic interaction between individual determinants that jointly impact firm performance. Consequently, they overcome the simple deterministic linkages between individual factors and performance assumed by traditional approaches. Due to the focus on the interdependence between determinants, we refer to this work as (I) Contingency Approaches. Next, we present studies that depart from traditional equilibrium approaches. The socalled (II) Dynamic Capabilities stream of research assumes that competition encounters frequent changes, moving from one equilibrium to another. The ability of the firm to dynamically adapt its capabilities to new challenges is thus critical to ensure a sustainable competitive advantage. Finally, the (III) Evolutionary Perspective completely contradicts the reasoning of classical approaches. Proponents of this view assume competition never attains equilibrium. The dynamic process of competitive evolution over time is thus the main focus of this research stream. In the remainder of this section, we review the shared underlying beliefs, as well as the results of research into each of the three fields in more detail. This review provides the basis for the development of our model for dynamic strategic analysis later in this chapter. We conclude this section with a (IV) Conclusions and Criticism section in order to evaluate the present state of the literature.

84

Contingency Approaches

Strategic management's classical approaches aim to establish universal truths that are valid across different businesses, regardless of the individual situation (Hambrick I Lei 1985). For instance, the generic strategy approach implies equal usefulness across a variety of environments (Miles I Snow 1978, Porter 1980). However, as we learn from the realities of business life, much in strategy is situational. Sometimes diversification may work and sometimes it may fail. Contingency theory thus assumes that there is no one best way that is equally effective under all conditions (Ginsberg I Venkatraman 1985, 421). However, the contingency approach to strategy suggests, "that for a certain set of

organizational and environmental conditions, an optimal strategy exists" (Ginsberg I Venkatraman 1985, 422). The contingency theory thus holds a middle ground position between universal views and situation-specific views that contradict all generalization beyond single firms (Hambrick I Lei 1985, 764). The term contingency is derived from the Latin expression "contingere" and means situational, or dependant on other factors. Consequently, the contingency approach can be

defined

as

"identifying

and

developing

functional

relationships

between

environmental, management and performance variables" (Luthans I Stewart 1977, 183). While contingency theories are widely recognized in organizational (Burns I Stalker 1961, Lawrence I Lorsch 1967) and leadership theory (Fiedler 1967), strategic contingency theory is still in its infancy (Lee I Miller 1996, 729). Nevertheless, various scholars have argued that the contingency theory is particularly appropriate for strategic management, particularly in dynamic market environments (Hambrick I Lei 1985, Miller 1988, Lee I Miller 1996, Aragon-Correa I Sharma 2003). Below we outline contingency approaches' (I) Shared Beliefs in strategic management with the focus on their particular understanding of dynamics and of the nature of competition. We then review studies that analyze the interrelation between the determinants of firm performance as discussed in chapter two: (II) Business Strategy - Market Structure, (III) Business Strategy -

Capabilities, (IV) Corporate Strategy - Business Strategy, (V) Corporate Strategy Market Structure, (VI) Corporate Strategy - Capabilities, and (VII) Capabilities - Market Structure.

85

Shared Beliefs: Contingency approaches take on the same fundamental research orientations as the different classical schools of strategic thought presented in chapter two (Zajac et al 2000). In the tradition of epistemological positivism (and the underlying ontological belief in realism), these scholars believe in a social reality that exists objectively, i.e. independently of the individual researcher. Consequently, the researcher attempts to establish objective, inter-subjectively stable theories by applying particular research methods (Seale 1999). In our case, strategists focus on establishing law-like cause-effect relationships between individual determinants and performance that may be generalized to a large number of companies. Neoclassical theorists additionally concentrate on competition under equilibrium conditions. This implies "historical efficiency" expecting competition to swiftly arrive at a steady-state equilibrium (March / Olsen 1989). Since equilibrium is the condition of a system at rest (Hill / Deeds 1996), this allows theories based on the premise that stable, underlying regularities (or strategic laws) govern business behavior and determine performance to be established (Jacobsen 1992). As the system is assumed at rest, these regularities or theories are expected to correspond to empirical patterns that are observable at any given time (Barnett / Burgelman 1996, 5). For example, a positive relationship between differentiation strategies and performance, proven for a sufficient number of firms at a single point in time, would thus be assumed as generally valid for all firms, regardless of their individual situation. As described above, classical strategic approaches (and the underlying equilibrium assumptions) have been criticized as inherently static (Jacobsen 1992, Nelson 1995). In contrast, contingency approaches possess a certain dynamic component absent in classical strategic theories. They analyze dynamic interactions among different factors influencing performance. Thereby they address the second criticism of classical strategic approaches as voiced above. The cause-effect linkage between e.g. business strategy and performance is no longer assumed as independent of the context. Third factors, like market structure, may impact this relationship. The effect may depend on the favorable interplay between these factors. To date this "dynamic" aspect of contingency approaches has gained little recognition. Due to the underlying equilibrium assumptions and the failure to capture the evolutionary aspects, these approaches have often been regarded as purely static. However, a system that captures the contingent relationship between individual determinants is inherently dynamic (Porter 1991, 113). As Porter

86

points out, such a system "begins to address a dynamic theory of strategy early in the

chain of causality" (Porter 1991, 115). While blind to the longitudinal problem, the contingency theory grasps the dynamics under equilibrium conditions well. Competitive Strategy - Market Structure: The linkage between competitive strategy

and market structure has gained more research attention than any other interrelation between determinants of firm performance. At the same time it is the only linkage that has been widely analyzed by classical approaches to strategic management. As reviewed in chapter 2.1, the structure-conduct-performance paradigm of industrial organizations suggests a direct impact of the market environment on business strategy (Bain 1959). However, the linkage between structure and strategy has often been downplayed, with market conditions' direct impact on performance being preferred (Scherer / Ross 1980). Researchers from the "efficiency school" argued to the contrary that firms with a superior strategy attain higher market shares and subsequently influence the structure of the market (Demsetz 1973). Contrary to industrial organization economists, the efficiency school thus assumes that business strategy has an impact on market structure. Later research confirms both views and presumes a reciprocal interrelation between business strategy and market structure (Porter 1991). However, the influence of market structure on strategic firm conduct has generally been seen as more influential (Farjoun 2002). Numerous frameworks have been established to help firms in evaluating the influence of market structure on their strategic decisions, including Porter's Five Forces (1980), the Boston Consulting Group's Growth Matrix, and D'Aveni's Hypercompetition Model (1994). In the theoretical debate, scholars of the industrial organization tradition speCifically encouraged research into the market structure-business strategy relationship. The strategic group concept, for example, combines market and strategy dimensions to position firms (Porter 1979). Moreover, a number of researchers started to analyze the impact of different market environments on the relationship between generic strategies and performance. Hambrick (1983) was the first to empirically contradict Miles and Snow's (1978) claim that strategy types are viable alternatives regardless of the market environment. He confirmed earlier suggestions that the environment can indeed influence both strategy selection and viability (Le. Burns / Stalker 1961). These findings contributed to the emerging contingency view of generic strategies (Hill 1988, Murray 87

1988, Lee / Miller 1996}. In general, cost leadership strategies are better suited to stable environments (Kim / Lim 1988, Miller 1988, McCabe 1990). However, they are less appropriate in environments characterized by fast-paced change, since economies of scale may swiftly turn into diseconomies of scale (Miller 1988). Differentiation strategies, in contrast, are more effective in dynamic environments, since the organization is more focused on quickly changing products, pursuing new technologies, and entering new markets (Bourgeois et al 1978, Hambrick 1983, Miller / Friesen 1984, Kim / Lim 1988). This leads to the conclusion that firms "with appropriate business level strategyenvironment fit will exhibit higher performance" (Marlin et al 1994, 171). Competitive Strategy - Capabilities: The main strategy paradigm is rooted in the idea

of matching firm resources with the corresponding business strategy (Chandler 1962, Andrews 1971, Schendel/Hofer 1979). It is generally accepted that there is no universal set of strategic choices that is optimal for all businesses (Ginsberg / Venkatraman 1985). Significant research thus focused on the interrelation between generic strategy types and the firm's internal features. Both Porter (1980) and Snow and Hrebiniak (1980) stressed the necessity of appropriate sets of distinctive competences in order to successfully implement generic strategies. Empirical studies confirmed that firms whose internal factors were well aligned with the chosen competitive strategy generally performed better than firms that lacked this alignment (White 1986, Miller 1988, Kumar et al 1998). Both Snow and Hrebiniak (1980) and Hambrick (1983) found that prospectors outperform defenders in terms of entrepreneurship and innovative ness (measured according to relative higher R&D and marketing expenses). Miller (1988) established the relationship between strategy types and organizational structure variables. Conant et al (1990) found that the marketing competencies of prospector organizations are Significantly superior to other strategy types, confirming the findings of earlier studies. In short, research into the relationship between strategiC types and capabilities suggests that prospectors are generally more entrepreneurial, innovative and market oriented than analyzers, and in particular, defenders. A second stream of literature regarding capabilities and business strategy emerged within the resource-based perspective. Grant (1991, 114) critically mentioned that "the link between strategy and the firm's resources has suffered comparative neglecr'. To surmount this shortcoming, Grant suggested a framework that positions firm resources

88

and capabilities as the foundation for strategy. A successful strategy needs to be designed around the firm's most critical resources and capabilities. In addition, the strategy must also be focused on identifying resource gaps and upgrading the firm's pool of resources in order to remain competitive. Grant concluded, "the resources and capabilities of a firm are the central considerations in formulating its strategy' (1991,

133). More recent research in the Resource-based View (RBV) has advanced Grant's initial thinking. Particular attention has been paid to the linkage between market orientation and business strategy. Business strategy has been found to moderate the strength of the relationship between market orientation and business performance (Slater! Narver 1994, Morgan! Strong 1998, Matsuno ! Mentzer 2000, Vazquez et al 2001, Kumar et al 2002). The positive impact of market orientation on performance is thus greater for prospectors and analyzers than for other strategy types. Vazquez et al (2001) furthermore established a link between market orientation and innovativeness and related both capabilities to prospector or differentiation strategies. Hurley and Hult (1998) added the link

between

market

orientation,

innovativeness,

and

organizational

learning.

Organizational learning has been found to be an important driver of product innovation and the ability to realize a differentiation strategy (Slater! Narver 1995, Calantone et al 2002). Corporate Strategy - Competitive Strategy: Despite the obvious interrelation of strategic actions at the corporate and business levels, very little research attempted to investigate this relationship (Dess et aI1995). However, some studies suggested that the link between corporate strategy and performance is moderated by strategic decision making at the business level. Empirical studies found that higher levels of product diversification were associated with lower levels of R&D expense and capital investments (Bettis 1981, Hoskisson! Hitt 1988, Hill! Snell 1989). The financial controls employed by diversified firms may therefore instill an orientation that discourages risk taking, research and development activities, and investment in new growth areas (Hoskisson ! Hitt 1988). Stimpert and Duhaime confirm these findings and conclude that "diversification influences performance indirectly by influencing strategic decision making at the business lever (1997, 577). They stress the importance of a better understanding

89

of the complex interrelation between corporate strategy, business strategy, and performance. Similar to the limited research into the product diversification-business strategy linkage, the interrelation between international diversification and business strategy has scarcely been investigated (Dess et al 1995). One exception is the study by Roth and Morrison (1992) that found the level of international involvement to be accompanied by differences in business-level strategy. In an earlier study, Young (1987) had already found a contingent relationship between business strategy and the degree of internationalization. International companies have been found to place greater strategiC emphasis on innovation, new product development, and marketing orientation - characteristics of analyzers and prospectors - than their domestic counterparts (Roth / Morrison 1992, 483). Vertically integrated businesses have been found to generate more new products and innovate more quickly, because of their participation in many of the production and distribution activities in which change can occur (Buzzell 1983, Heeb 2003). They also show higher research expenditures than less integrated firms (Armour / Teece 1980). Not surprisingly, firms in high growth markets have been found to be more vertically integrated (Prabhu / Harrison 1992). At the same time, vertical integration has been found to significantly reduce transaction and overhead costs (D'Aveni / Ravenscraft 1994). D'Aveni and Ravenscraft (1994, 1194) suggest "that vertical integration can be used as part of a low-cost strategy". If we consider all the findings summarized above, a

balancer strategy might come closest to resembling the findings, combing the cost and product innovation focus of vertical integration. However, the relationship has not been empirically tested to date. In summary it can be stated that a number of contingent relationships between the different types of corporate strategy and business strategy have been detected by the extant literature, but further research will be required to better analyze these interrelations. Corporate Strategy - Market Structure: Akin to the corporate strategy-business strategy linkage, the contingency of corporate strategy on the market context has gained 90

little research attention and the present findings are highly controversial (Datta et al 1991). Early studies on industrial organization economics were concerned with the effects of diversification on market structure. These studies viewed diversification merely as an element of market structure (Rhoades 1973, 1974). Broadly speaking, diversification was associated with market concentration, thus increasing the diversified firm's market power and performance (Berry 1974). Later research into strategic management proposed the converse idea that market structure affects diversity (Christensen / Montgomery 1981, Bettis / Hall 1982, Wernerfelt / Montgomery 1986, Grant et aI1988). These studies state that performance differences, spuriously attributed to variations in diversification strategy, are actually due to differences in the underlying market structures. For instance, Christensen and Montgomery (1981) propose that firms in highly profitable and concentrated markets mostly pursue related diversification, while firms exposed to less favorable market conditions follow unrelated diversification strategies. The higher performance of firms adopting related diversification is thus attributed to differences in market structure rather than to differences in diversification strategy. However, studies monitoring industry effects nonetheless found a significant relationship between diversification strategy and performance, thus rejecting Christensen and Montgomery's argument (Rumelt 1982). Datta et al (1991) conclude that while numerous studies seek to reassert the importance of market structure, they are inconclusive with respect to its impact on the diversification-performance relationship. However, industry structure variables are expected to play an important role in moderating the diversification-performance relationship. Datta et al postulate that "any theory of diversification must, by definition, be contingency based - a diversification strategy ( ...) needs to be matched with the environmental context or industry conditions" (1991, 550). This implies the need for additional studies to examine the interrelationship between degree and type of diversification and industry structure in determining performance. A first study has taken up this challenge: Stimpert and Duhaime (1997) empirically analyzed the impact of industry profitability on a firm's diversification decisions. They found that industry profitability exerts a strong negative influence on the extent of diversification. Firms operating in profitable industries have sufficient opportunities for profitable growth within these market segments, while companies in less profitable industries are forced to diversify into other markets. This finding is particularly interesting 91

as it provides empirical confirmation of earlier hypotheses. Rumelt (1974, 82) suggested that "for a great many firms, diversification is the means employed to escape from declining prospects in their original business area". Later studies echoed Rumelt's "escape hypothesis" and concluded that firms located in markets that constrain their growth or profitability are likely to pursue more diversification activity (Christensen I Montgomery 1981, Montgomery 1985). This argument implies that adopting a diversification strategy is more appropriate for companies operating in less favorable market conditions than for firms in more attractive segments. There is some anterior support for this suggestion: Two empirical studies found diversification to be positively related to profitability in slow growth and low concentration industries, but no relationship or a negative relationship in high growth and high concentration industries (Miller 1973, Bass et al 1977). The influence of different market conditions has also been discussed in the literature on the vertical integration-performance relationship. Some positive effects of vertical integration, such as the ability to secure access to supplies, have been stressed as particularly important in fast growing and volatile industries (Prabhu I Harrison 1992). However vertical integration has generally been regarded as rather negative for fast growing and volatile environments. Bureaucratic costs and coordination (or information processing) are already challenging in dynamic environments, these challenges become even more acute with vertical integration (Jones I Hill 1988, Prabhu I Harrison 1992, D'Aveni I Ravenscraft 1994). Moreover, vertical integration makes it particularly difficult to adapt to changing market conditions (D'Aveni I lIinitch 1992). Thus, on the whole, it has been hypothesized that vertical integration is more appropriate in concentrated industries than in fast growing or volatile industries. Empirical evidence shows that concentrated industries indeed exhibit more overall integration than other industries (D'Aveni I Ravenscraft 1994). However, the link to performance is somewhat less clear. Some studies confirmed the negative effect of market growth (Prabhu I Harrison 1992) or, vice versa, the positive effect of emphasis on the vertical integration-performance relationship (MacDonald 1985). Studies on the vertical integration-risk relationship found similar evidence: vertical integration reduces the risk in the highly concentrated and stable oil industry (Mitchell 1976), but increases the risk of bankruptcy in more turbulent environments (D'Aveni I llinitch 1992). Despite

92

this evidence, others failed to find any significant market condition effects on the vertical integration-performance relationship (Buzzell 1983, D'Aveni / Ravenscraft 1994). In summary, we can state that despite sound theoretical support and some empirical evidence, further research into the complex relationships between corporate strategy and environment will be required. Corporate Strategy - Capabilities: The linkage between internal organization and

strategy is central to strategic management. Research on the fit between corporate strategy and organizational structure, and their combined affect on performance, has a long history (Chandler 1962, Rumelt 1974). Empirical research has generally confirmed the suggestion that a multidivisional (or M-form) structure positively affects the diversified firm's performance (Le. Teece 1981). In their review of the diversification literature, Ramanujam and Varadarajan (1989) conclude that the benefits of diversification are not automatically realized. On the contrary, they depend on the presence of managerial attributes such as organizational structure or distinctive skills. More recent research has shifted the focus from organizational structure to firm resources and capabilities. Hitt and Ireland were the first to empirically confirm that "the group of corporate level distinctive competencies related to performance varies by the type of diversification strategy employed by the firm" (1986, 411f). Chatterjee and Wernerfelt suggest that "to understand the link between diversification and performance we need to consider the resource profile of the firm" (1991, 41). They found that firms with significant intangible assets (measured according to research intensity or innovativeness) were strongly associated with related diversification. This may indicate that intangible assets are less flexible than, e.g., financial assets and thus more applicable in related diversification. Performance is consequently not a function of diversification strategy per se, but "the appropriateness of the diversification strategy given the resource profile of the firm" (Chatterjee / Wernerfelt 1991, 45). Lei et al (1996) claim that the ability of the firm to realize benefits from diversification (such as economies of scale and scope) relies heavily on the leveraging of core capabilities generated through organizational learning. Pennings et al (1994) confirm in their study that diversification success is dependent on organizational knOW-how and learning.

93

In prior studies a strong linkage was established between vertical integration and innovation (Armour / Teece 1980. Buzzell 1983. Heeb 2003). These studies found that vertically integrated firms show a high degree of innovativeness. they generate more new products and stimulate R&D expenditures. In addition. Armour and Teece (1980) claim that vertical integration increases the knowledge exchange between hitherto disconnected entities and creates a common "language". thus facilitating new product development. Sorenson (2003) developed the link between organizational learning and vertical integration further. In his empirical study. he found strong evidence that vertically integrated firms learn more efficiently than their less interdependent rivals. particularly in volatile environments. In summary. while it has often been claimed that corporate strategy is closely dependent on the underlying capabilities of the firm. there is only minor evidence of the importance of relationships between concrete capabilities and different types of corporate strategy. Capabilities - Market Structure: Exploring the linkage between organizational and

managerial characteristics and the environment has traditionally been at the heart of the organizational contingency theory (White 1986. 218). The contingency theory posits that firm performance is the result of the proper alignment of organizational structures with exogenous context variables (i.e. Burns / Stalker 1961. Lawrence / Lorsch 1967). Scholars of the resource-based view later pursued this tradition. While early studies in the resource-based tradition were mainly focused on internal factors. more recent approaches bridge the divide between internal and environmental influences. The notion of "core competences" introduced by Prahalad and Hamel (1990) provided a first framework with which to integrate organizational and competitive dynamics. More recently. empirical evidence has shown that the effectiveness of capabilities varies with the market dynamism and competitive environment (Brush / Artz 1999. Eisenhardt / Martin 2000). Miller and Shamsie (1999). for example. showed that knowledge-based resources in motion picture studios only had a positive effect on performance in uncertain and risky market environments. The business environment has moreover been found to affect the process that allows the development of dynamic capabilities (Helfat 1997. Teece et al 1997). Henderson and Mitchell (1997) argue that capabilities and competitive environment have reciprocal relationships. Firms develop capabilities that 94

shape the environment that, in turn, further shapes capabilities. This argument is in line with Levinthal and Myatt's (1994) earlier study, empirically confirming a "co-evolution of capabilities and industry" through multiple feedback relationships between firm

capabilities and industry structures. Based on these findings, the first authors postulated a "contingent resource-based view" (Brush I Artz 1999, Zajac et al 2000, Barney 2001, Priem I Butler 2001, Arag6n-Correa I Sharma 2003). Brush and Artz concluded, "the value of resources is contingent on the context in which itis used" (1999, 246).

While a cohesive "contingent RBV" is just about to emerge, an impressive body of studies has already been conducted on the contingency of single capabilities. Researchers have, for instance, drawn attention to the major influence of industry characteristics on organizational culture (Gordon 1991, Kotter I Heskett 1992, Burt et al 1994, Chatman I Jehn 1994, Sorensen 2002). While it is now widely accepted that organizational culture is contingent on market environment, it remains less clear how organizational culture is contingent with market environment. Chatman and Jehn (1994) fail to provide solid empirical support for their hypothesis that innovative cultures are more appropriate in markets characterized by intensive technologies and high growth. Sorensen (2002) finds some support for the assumption that strong cultures lead to increased performance in stable environments, while incurring great difficulties in more volatile environments. Research into the contingent relationships of other capabilities struggles with much the same problems. For instance, it is generally accepted that the influence of firm innovation capability is strongly moderated by environmental context variables (Tidd 2001). However, far less agreement has been reached on the nature and direction of the influence of these context variables. While some scholars postulate a positive correlation between high industry concentration and innovation (in the tradition of Schumpeter's concept of "creative destruction"), others found none or a negative effect (Acs I Audretsch 1988, Lunn 1989, Koeller 1995). Yet another group of researchers suggests that the degree of environmental uncertainty and complexity has a strong influence on the innovation-performance relationship (Damanpour 1996, Tidd 2001). This hypothesis, however, remains to be empirically tested. In this section, we attempted to review the broad and diverse literature on contingent relationships between constructs that underlie firm performance. As we have seen, the importance of capturing these interactions has been well established, while Significant 95

work remains to be done in order to clarify the concrete linkages between individual factors. We do not attempt to provide a full critique of contingency theories at this point. However, it is important to state that contingency theories (as with all positivist theorizing) risk taking a somewhat deterministic approach. Various real life examples show that managers enjoy considerable strategic freedom. For example, in any industry successful companies can be found following strategies contrary to generic strategy types. Despite a certain level of strategic discretion, however, it is important to build models that discover the mechanisms underlying strategic decisions and performance in most companies. In our understanding, contingencies do influence the impact of determinants on performance, but they constrain rather than fully determine "best practice".

Dynamic Capabilities

While contingency approaches contribute a certain dynamic aspect to traditional approaches, they share the limitations of equilibrium approaches. In order to capture more dimensions of a dynamic system, these restrictions have to be overcome. A first stream of research that goes beyond equilibrium concepts is the discussion around socalled "dynamic capabilities". In the tradition of the resource-based perspective, these authors analyze the role of strategic firm capabilities in generating competitive advantage for the firm. However, in contrast to earlier work, they overcome the static approach of identifying existing capabilities through an analysis of the capability creation process over time (Helfat 2000). In this section, we review the (I) Shared Beliefs of these approaches, and present two core dynamic capabilities central to the current discussion: (II) Organizational Learning and (III) Innovation. Shared Beliefs: Although the resource-based view initially emphasized dynamic change

over time (Le. Penrose 1959, Wernerfelt 1984), much of the following literature remained static in concept. Typically, a certain type of resource - e.g. market orientation or organizational culture as described above - is theoretically linked to competitive advantage. Next the value of the resource is ascertained through the presence of advantage-creating characteristics. This approach has "a distinctly ex post quality; once a firm is recognized as successful, the resources behind the success are labeled as

96

valuable" (Foss et al 1995,8). In addition, the processes through which these resources provide advantage "remain in a black box' (Priem / Butler 2001, 33). Based on the assumed equilibrium conditions, resources once identified as valuable are assumed to preserve this quality over time. More recently, some researchers in the resource-based view have contradicted this assumption: "RBV's emphasis on long-term competitive advantage is often unrealistic in high-velocity markets" (Eisenhardt / Martin 2000, 1118). Proponents of the "dynamic capabilities" perspective argue that in turbulent and fast changing market environments, firms need to continually evolve and change their resources (Teece et al 1997). Capabilities cannot remain static without the risk of losing their value through changing market conditions (Lei et al 1996). What are valuable resources today may no longer be appropriate tomorrow. The focus thus shifts to the development and change of core capabilities. This implies a process perspective: "The simple observation that competencies may lead to advantage is only half the battle: the other half is the understanding where competencies come from" (Cockburn et al 2000, 1128). The specific processes or routines firms use to alter their resource base have been described as "dynamic capabilities". Teece and his colleagues define dynamic capabilities as the "firm's ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments" (1997, 516). Dynamic capabilities, or the ability to adapt firm resources to changes in the competitive environment more quickly than competitors, are seen as the most important source for sustained competitive advantage in rapidly changing markets (Eisenhardt / Martin 2000). Zott (2003) linked dynamiC capabilities to differential firm performance within an industry. The dynamic capabilities perspective is inherently dynamic in its emphasis on the development process of firm capabilities in turbulent environments. Teece et al (1997, 515) state that the term 'dynamic' "refers to the capacity to renew competences so as to achieve congruence with the changing business environmenf'. In contrast to contingency approaches, dynamism refers less to the inherent interplay of determinants underlying 97

performance than to the dynamic evolution of a group of determinants - capabilities over time. Teece et al (1997) point to the works of Schumpeter (1942) and Nelson and Winter (1982) as the intellectual roots of the dynamic capabilities perspective. Competition is thus seen as the Schumpeterian world of frequent changes that rarely reaches equilibrium conditions. This position is directly opposed to traditional equilibrium assumptions. However, most subsequent studies on dynamic capabilities take a less radical position. For instance, Lei and his colleagues state: "Actually, over time organizations altemate between equilibrium and disequilibrium" (1996, 561). Competition is thus viewed as "punctuated equilibrium" (Gersick 1991); periods of relative stability are interrupted by

phases of revolutionary change. This position allows for a synthesis of the RBV and dynamic capabilities perspective. While the former is appropriate for the relatively stable (or "equilibrium") phase, the latter is required to capture more volatile phases of change. Eisenhardt and Martin (2000) take a similar position. They state that sometimes it is effective to strengthen the current resource configurations using the RBV's pathdependent leverage logic. However, in dynamic markets, it is just as important to deploy dynamic capabilities to build new resource configurations using a path-breaking change logic. Both logics are expected to alternate in their importance in line with the punctuated equilibrium theory. Temporary stable equilibrium conditions are interrupted by phases of change: "The goal is a series of temporary competitive advantages" (Eisenhardt / Martin 2000, 1118). This claim is in accordance with the generally accepted view that the RBV and dynamic capabilities perspective are complementary rather than mutually exclusive approaches (Makadok 2001). While the general importance of the dynamic capabilities construct has been widely recognized, the present studies have often been criticized as vague or tautological. This is largely due to lack of empirically grounded research into individual dynamic capabilities and their contribution to competitive advantage (Williamson 1999, Priem / Butler 2000). Dynamic capabilities have often been declared as unique and idiosyncratic processes that cannot be generalized across firms (Teece et al 1997). More recent studies reaffirm the idiosyncratic character, but assert that "specific dynamic capabilities also exhibit common features that are associated with effective processes across firms"

(Eisenhardt / Martin 2000, 1108). Two dynamic capabilities have gained particular

98

research attention in the past years: leaming and innovation. Both discussions profit from the vast literature on both subjects from adjoining research fields like organizational science. The integration of this work into the debate on strategic management provides a sound foundation for analyzing these important dynamic capabilities and their contribution to firm performance. In the remainder of this section, we continue to review the current discussion on these two central dynamic capabilities. Organizational Learning: Zollo and Winter (2002) observe that much attention has

been spent on explaining why dynamic capabilities are important and how they work, but the question of where they come from remains open. The authors continue to investigate the mechanisms through which organizations develop dynamic capabilities. The argument is made that dynamic capabilities are shaped through deliberate leaming processes such as knowledge articulation and codification. This argument has a long tradition in the resource-based perspective. The accumulation and development of competencies is understood as a process of continuous leaming (Bamey 1986, Hamel 1991, Nelson 1991, Petroni 1998). Prahalad and Hamel (1990) suggest that core competences are based on collective leaming in the organization. At the most basic level, organizationalleaming is the development of new knowledge (Fiol / Lyles 1985). Knowledge is thus the basic foundation of capabilities (Winter 1987). Leonard-Barton (1992) discussed core capabilities as a function of the firm's ability to create knowledge. Similarly, Lei et al (1996) note that core competences are developed from organizational leaming. In order to be effective, core competencies "must be continually evolving and changing via continuous organizational learning" (Lei et al 1996,

550). Pisano concludes, "without learning, it is difficult to imagine from where a firm's unique skills and competencies would come" (1994, 86).

In other words, leaming is the driving force behind the creation and evolution of competence. However, Teece et al (1997) labeled processes that create and develop competencies as "dynamic capabilities". Leaming itself can thus be understood as the firm's preeminent dynamic capability and source of competitive advantage (Lei et al 1996, Bierly / Chakrabarti 1996, Eisenhardt / Martin 2000).

99

The vast literature on organizational leaming postulates the need for businesses competing in dynamic environments to pursue leaming in order to maintain and improve performance (Fiol / Lyles 1985). Organizational adaptation to environmental changes through leaming is often seen as the essence of strategic management (Cyert / March 1963, Miles / Snow 1978, Chakravarthy 1982). Comments such as "the ability to leam faster than your competitor may be the only sustainable competitive advantage"

(DeGeus 1988, 71) have been frequently made. However, the emerging "knowledgebased view" (Grant 1996, Spender 1996) is the first consistent effort to tap findings from

the organizational leaming literature for strategic management. Knowledge is seen as the major source of competitive advantage (Leonard-Barton 1992, Conner / Prahalad 1996, Grant 1996). Teece summarizes: 'The essence of the firm is its ability to create, transfer, assemble, integrate and exploit knowledge assets. Knowledge assets underpin competences, and competences in tum underpin the firm's product and service offerings to the market" (1998, 75).

Various scholars thus postulate a better integration of the resource-based theory, organizational leaming and

knowledge management (Le.

Probst et al

1998).

Organizational leaming, by definition a dynamic process, is likely to be incorporated into many theoretical models of strategic management (Hoskisson et al 1999, 19). Despite the broad theoretical appraisal of the important role of organizational leaming and knowledge, few empirical studies attempted to link these concepts to firm performance. Bierly and Chakrabarti (1996) found significant performance differences between firms applying different generic knowledge strategies. A number of marketing scholars acknowledged the importance of leaming orientation to overall firm performance (Slater / Narver 1995, Hurley / Hult 1998). Leaming orientation refers to the organization-wide activity of creating and using knowledge to enhance competitive advantage. In their empirical analysis, Calantone et al (2002) found that leaming orientation had a profound effect on firm performance. Their findings confirm earlier empirical research by Baker and Sinkula (1999). Calantone and his colleagues conclude, "Ieaming orientation facilitates the generation of resources and skills essential for firm performance" (2002, 522).

100

Innovation: Innovation is closely related to organizational learning. Defined as the

generation, acceptance, and implementation of new ideas, processes, products, and services (Thompson 1965), innovation involves the acquisition, distribution, and use of new knowledge (Daman pour 1991). A learning climate clearly promotes firm innovation (Hurley / Hult 1998). Despite the close relationship, learning orientation and innovation are distinct constructs: "Learning orientation emphasizes the organizational value of obtaining knowledge, whereas innovativeness focuses on the organization's willingness to change" (Calantone et al 2002, 522). Some authors have pointed to innovation as an important dynamic capability. Teece and Pisano (1994, 541) defined dynamic capabilities as a subset of competencies "which allow the firm to create new products and processes and respond to changing market circumstances". This definition is almost identical to temporary definitions of innovation (i.e. D'Aveni 1994). Hurley and Hult (1998, 44) conceptualize innovativeness as "the central mechanism by which organizations develop capabilities and adapt to their environments". Eisenhardt and Martin (2000) describe the product development process as an important dynamic capability. Nevertheless, Lawson and Samson (2001) were the first to explicitly draw together knowledge from various fields to propose that innovation management can be viewed as dynamic capability. Building on the dynamic capabilities literature, they propose the "innovation capability" construct as the primary engine for performance improvement. The positive impact of innovation on firm performance has been empirically confirmed in a number of studies (i.e. Geroski 1994, Crepon et a11998, Klomp / van Leeuwen 2001). This link has been explained by organizations with a high innovation speed's ability to gain market share and defend the latter against competition by erecting of "isolating mechanisms" (Rumelt 1987). Firms must be innovative to survive in volatile market environments and innovativeness is reflected by the firm's rate of adoption of innovations and its willingness to change. Numerous studies by marketing scholars have empirically tested the relationship between firm innovativeness and firm performance (Deshpande et al 1993, Vazquez et al 2001, Calantone et al 2002). They confirm prior findings that ascribe a central role to innovativeness in explaining firm success.

101

In summary, the present literature suggests that both learning orientation and innovativeness, while closely interrelated, exert an independent and significant influence on firm performance. Both constructs can be seen as dynamic capabilities, central to the ongoing creation and development of the firm's base of resources and competences.

Evolutionary Perspective

As previously described, the dynamic capabilities perspective is particularly appropriate in describing volatile market environments. Many business environments today have been described as chaotic and complex, characterized by fast-paced change (Le. O'Aveni 1994). Strategy researchers thus began to develop theories that incorporate conditions of uncertainty, dynamism, and aperiodicity (Black / Farias 2000). While the research subjects and perspectives are divergent, many of these studies at a deeper level share a well-defined common core that ties these different viewpoints together. This core consists of different arguments derived from evolutionary economics (Schumpeter 1942, Nelson / Winter 1982). We describe these (I) Shared Beliefs next and continue to provide a brief summary of the most relevant (/I) Evolutionary Approaches to Strategy. Shared Beliefs: Evolutionary economics itself is a highly disparate research field with

obvious differences between traditional evolutionary economics (mainly Schumpeter's work) and the new wave of evolutionary theorists (Nelson and Winter onwards). However, the fundamentals are the same and can be summarized in three interrelated arguments that together define evolutionary theories (Nelson 1995, Fagerberg 2002). First, the purpose of evolutionary arguments is to "explain the movement of something over time, or to explain why that something is what it is at a moment in time in terms of how it got there; that is, the analysis is expressly dynamic" (Oosi / Nelson 1994, 154).

This argument goes back to Schumpeter (1942) who analyzed capitalist development as an evolutionary process characterized by continuous economic change. The term "evolutionary" is a metaphor that has been adopted from biology, where evolutionary ideas have been dominant for more than a century (going back to Charles Darwin's publications in the 19th century). The main factor driving the economic evolution is innovation or the technological competition between firms (Schumpeter 1942, Nelson / 102

Winter 1982), an idea borrowed from Marx (1887). Evolution in its economic sense thus describes "the changes in the economic process brought about by innovation" (Schumpeter 1939, 86). Entrepreneurs or firms' innovation activities constantly disrupt equilibrium forces. Consequently, evolutionary theory aims to understand aspects or sources of economic change: "The focus of attention is on a variable or set of them that is changing over time and the theoretical quest is for an understanding of the dynamic process behind the observed change" (Nelson 1995, 54). Second, while involving random elements, evolutionary processes are characterized by strong regularities (Dosi 1988, Nelson 1995). There is, for instance, the sequence of innovation and imitation (Schumpeter 1939). The economic reward associated with a successful innovation is transitory; it vanishes when a sufficient number of imitators enter the market. The successful introduction of large innovations causes temporary market growth. These effects contribute to fairly regular "business cycles" of varying lengths. Third, the unpredictability of the future (due to the open-ended character of evolution), combined with the complexity of economic decision-making, prevents the economic actor from making fully rational decisions (Nelson / Winter 1982). Neoclassical economists' assumptions of the rational actor are thus replaced by the notion of "bounded rationality" (Cyert / March 1963). Firms are assumed to follow decision rules (or "routines") that determine behavior (Nelson I Winter 1982). Routines are not necessarily rational, but they remain until more efficient routines are found. New routines are found through "search" defined as R&D or learning process (Nelson 1995,69). While all three assumptions can be found in evolutionary writings within strategic management, particular attention has been paid to the aspect of dynamic economic change. Evolutionary theories in strategy take a process-oriented approach (Foss et al 1995) and "explicitly question how strategic outcomes develop" (Barnett / Burgelman 1996, 6). Taking an evolutionary perspective of strategy means taking the pace and path of strategic change into account through application of dynamic models.

103

Evolutionary Approaches to Strategy: While numerous strategic studies share the evolutionary beliefs outlined above, it is difficult to classify these approaches into distinct research fields. A multitude of concepts and the associated specific terminology complicate the necessary consolidation of the field. For example, some authors translate findings from complexity theory into strategic thinking (e.g. Black / Farias 2000), whereas others apply the chaos theory to strategy (e.g. Levy 1994). While these efforts are promising, they remain limited to a few studies to date. However, two streams of research stand out: since the early 1990s much strategic writing on the evolutionary tradition has emerged under the two labels "hypercompetition" (D'Aveni 1994) and "Austrian school of Strategy" (Jacobson 1992).

D'Aveni's concept of "hypercompetition" is based on Schumpeter's model of competition as a process of permanent change (or "creative destruction"). Hypercompetition is an industry environment characterized by inherent instability and change. Firms in these industries need to continuously build new competitive advantages through innovation, thus destroying the opponent's competitive advantage and creating disequilibrium. This quest for profit by establishing competitive advantage constitutes the driving force behind competition. The permanent rivalry for competitive advantage means that competitive advantage is temporary. Corporate competencies are thus eroded through frequent discontinuities (Hamel 2000). Superior performance over time can therefore only be achieved through continuous recreation and renewal of competitive advantage. Proponents of the hypercompetitive perspective assume that more and more industries change from static to dynamic (or "Schumpeterian") competition. This "hypercompetitive shift" is explained as forces of change, including globalization, advances in information

technology, deregulation, and demographic shifts (Thomas 1996). Based on the hypercompetitive assumption, numerous studies emerged that moved away from the theoretical paradigm of static competition towards Schumpeterian competition (liinitch et al 1996). Because hypercompetition is a dynamic process, these studies often take a longitudinal approach. Webb and Pettigrew (1999), for instance, explore the dynamics of industry and firm strategy development in the UK insurance market. Craig (1996) describes the dynamic movement of firms in the Japanese beverage industry - evolving from one hypercompetitive arena to the next. Melin (1992), on the other hand, analyzes internationalization strategies in dynamic and process terms. 104

Schumpeter's view of competition as a dynamic process of rivalry is also the major theme of the more recent writings in the tradition of the Austrian school of economics, or the "Austrian school of strategy" (Jacobson 1992, Young et aI1996). In opposition to traditional industrial organization economists, Austrians tend to view market structures as dynamic and in a state of disequilibrium (Hayek 1945). The market structure is the by-product of entrepreneurs' actions. Entrepreneurs constantly apply new capabilities and strategies in competition to realize new profit opportunities. Competitive advantage therefore arises from the superior creativity and alertness of some entrepreneurs (Kirzner 1973). The focus of the Austrian perspective is thus on the dynamic interaction between firm actions and industry structure. Young et al (1996), for example, use a dynamic model of competitive activity to analyze the linkages between firm actions, environment, and performance by applying a longitudinal empirical research approach. Abernathy and Utterback (1975) and following empirical research (Le. Utterback 1994) suggest that patterns of technological evolution cause a particular pattern of industry structure evolution. The Austrian school can furthermore be seen as a predecessor of the dynamic capability perspective. The dynamic capability perspective shares the Austrian interest in unobservable factors, including innovativeness and other entrepreneurial skills. Much of this work is in the Austrian tradition, for example Levinthal and Myatt's (1994) analysis of the co-evolution of capabilities and industry, or Zajac and colleagues' (2000) longitudinal research into the dynamics of strategic fit between environmental forces and firm resources. In summary it can be stated that evolutionary thinking in strategic management has been mainly applied to longitudinal studies of competitive dynamics, as well as the evolution of industry structures and capabilities over time. Evolutionary theory thus contributes a process perspective to strategic analysis that helps to overcome the limitations of traditional static (equilibrium-based) approaches.

105

Conclusions & Criticism

The principal merit of the different approaches presented above is their contribution to a more dynamic and thus more realistic picture of the factors and processes underlying firm performance. The contingency theory is undoubtedly right in assuming that there is no one best way or universal set of strategic choices that is equally effective under all conditions (Galbraith 1973). It is largely due to the dynamic capabilities perspective that important economic constructs like learning and innovation found their way back to the heart of strategic theorizing (Hoskisson et al 1999, 19). In addition, the evolutionary perspective represents an important step forward in breaking through the limitations of the equilibrium assumptions in order to understand the process that led to the current competitive situation (Nelson 1995). While these approaches are particularly strong in their criticism of traditional approaches to strategy, and to a lesser extent in theorizing, there are a number of factors that severely limit their usefulness and explanatory power. We will review three of the most important impediments obstructing the further development of the "dynamic perspective" of strategy and point out the work that remains to be done. Two hostile camps: As observed above, traditional approaches to strategy (including

the contingency perspective) analyze competition under equilibrium conditions, while proponents of the evolutionary and dynamic capabilities perspectives understand competition as a process that never or rarely reaches equilibrium. While the two orientations are often seen as complementary rather than mutually exclusive (Barney 1986, Nelson 1995, McNamara et al 2003), the two camps are busy building walls between each other. For instance, it is obvious that resource-based view and dynamic capabilities perspective are nothing more than two sides of the same medal. The former analyzes the role of existing resource configurations, while the latter contributes insights into the process that led to their establishment. With the rare exception of Makadok (2001) the research streams remain widely separate. The cleavage between the evolutionary and traditional approaches to strategy is even deeper. Much of the evolutionary literature positions their approaches as a replacement of rather than complementary to traditional approaches (i.e. Hill / Deeds 1996). Consequently, these authors reject the results of more than four decades of strategic thinking based on

106

equilibrium assumptions as wrong and irrelevant. The huge potential for crossfertilization between the camps is neglected by both sides (Black I Farias 2000, 101). Some authors, however, including Barney (1986) and Porter (1991), called for more interaction between the different camps. These claims certainly point to the right direction. Integration between the equilibrium and evolutionary perspectives is necessary to further advance both discussions. Above all this requires a common theoretical foundation that will accommodate elements of both perspectives. As Powell and Wakeley state: "Evolutionary economics alone will not be sufficient to provide all the normative framework for a strategic manager" (2003, 156). A fragmented field: Unfortunately, the fragmentation within the dynamic camp is even stronger. The dynamic capability perspective, for instance, limits their longitudinal analysis of capabilities. Evolutionary theory, in contrast, deals predominantly with understanding the evolution of industries and offers "only a relatively narrow description of the firm and its resources" (Foss et al 1995, 7). There are no obvious guidelines as to which level of analysis to adopt, leaving much scope for confusion (Powell I Wakeley 2003). There is no clear definition of "dynamics": while the contingency school understands dynamics as inherent to a system at rest, evolutionary theorists analyze dynamics as a process over time. The dynamic field thus remains fragmented, characterized by a multitude of disjointed research efforts (Porter 1991, 95). Each effort is supported by a limited number of researchers reluctant to integrate findings from other fields. Progress in dynamic research has therefore been rather slow. Lee and Miller, for instance, state that the "strategic contingency theory is still in its infancy" (1996, 729). Much the same is true for the dynamic capabilities perspective (Priem I Butler 2001, 35). Farjoun concludes that the dynamic perspective's "transition away from fragmentation, static, and linearity has remained incomplete and uneven" (2002, 570). In order to provide a true alternative to mechanistic approaches, the dynamic perspective needs more integrative efforts to develop a consistent set of concepts and models. Lack of research frameworks & methods: Barney (1986) points to the difficulties associated with formal theorizing based on the concept of Schumpeterian competition. He finds that the evolutionary concept has resisted translation into normative theories of strategy (Barney 1986, 795). Nelson agrees and concludes that economists who eschew 107

equilibrium conditions "must pay an analytical price" (1995, 50). Due to the resulting complexity, "the proponents of evolutionary theory are struggling with both techniques and standards" (Nelson 1995, 85). Evolutionary theories are also less traceable and

convenient in terms of their cause-effect logic. Jacobsen adds that Austrian economics "fails to provide a framework for making a positive contribution" (1992, 795). This might

be due to the fact that "no two Austrians have ever completely agreed on methodology" (Littlechild 1978, 22). A comparable need for new methodologies and testable frameworks has been stated for the RBV and the dynamic capabilities perspective (Barney 2001, Priem / Butler 2001). The failure to provide testable frameworks and methods is therefore a major drawback for the dynamic perspective. Empirical validation is rather difficult with most dynamic approaches remaining purely theoretical. However, without such applicable analytical models and empirical evidence, the dynamic perspective is merely a form of art (Farjoun 2002, 584). Since strategic management finds its rationale in the quest for normative outputs, practitioners require concrete, practical assistance. In order to develop a dynamic or evolutionary support to strategic decision-making, "substantial development work is now required" (Powell / Wakeley 2003, 160).

At a general level we can state that the dynamic perspective is a promising research field that has the potential to contribute Significantly to further advances in the field of strategic management. However, in order to gain practical relevance for "mainstream" researchers and practitioners, the field has to move from generating attractive ideas to more formal theorizing. More integrative dynamic approaches that provide applicable frameworks testable in empirical studies are required. The strenuous road towards this goal could be Significantly facilitated by tapping into the tremendous body of work of more conventional strategy scholars. Cross-fertilization will help to advance dynamic research and to establish the linkages between these and other - equally valuable - approaches to strategy. In the following two sections, we attempt to advance the present literature into this direction, by establishing an integrative dynamic framework for strategic analysis.

108

111.3

Foundations of a Dynamic Research Model

We draw nearer to our target of establishing an integrative dynamic model for strategic analysis. In the previous chapter, the present strategy literature on determinants of firm success was reviewed and the most relevant "ingredients" for a dynamic model were presented. In chapter two, we discussed traditional theories on firm performance while we have reviewed approaches attempting to provide a more dynamic understanding of processes underlying firm success in the present chapter. However, while the "ingredients" for a dynamic model are there, they do not seem to fit. We have seen above that both classical and dynamic approaches are highly fragmented. Furthermore, the different approaches rest on conflicting assumptions. This significantly complicates the task of establishing a truly integrative dynamic model. Metaphorically speaking, we need a recipe to transform the incongruous ingredients into an elaborate dinner. The "recipe" denotes a set of common underlying assumptions that provide a frame of reference for our model. In order to establish this frame of reference, we begin with some thoughts on philosophy of science. As discussed above, classical and dynamic approaches are based on differing epistemological assumptions. Our first effort is thus to achieve a common (I) Reference to Philosophy of Science for our model. At a more concrete level, classical and dynamic

models are conflicting in their understanding of change. While traditional approaches are based on equilibrium assumptions, dynamic models assume an evolutionary process. In order to approximate these disparate assumptions, we provide our model with a consistent (II) Reference to Change. A third line of demarcation not only separates the traditional from the dynamic approaches, but runs across different dynamic perspectives: The understanding of complexity and dynamics. For instance, contingency approaches analyze dynamics between different elements of a system at rest, whereas the evolutionary perspective defines dynamics as a process over time. We thus continue to strive for a common (III) Reference to Complexity and Dynamics. Finally, the three different references will be integrated into an (IV) Overall Frame of Reference providing the required "recipe" for the later preparation of our dynamic model.

109

Reference to Philosophy of Science

What do we mean with "Philosophy of Science"? Science can be understood as the attempt to understand. explain or predict the world we live in. Scientists use particular methods to investigate the world and to construct theories with which to explain it. The philosophy of science analyzes these methods of enquiry and questions the underlying assumptions the scientist takes for granted (Rosenberg 2000). In the past century. the philosophy of science pointed out a number of problems that question the claimed absolute validity of scientific theories. We briefly review three central problems related to (I) Scientific Reasoning. (II) Scientific Explanation. and (III) Scientific ObseNation. Dealing with these issues will sensitize the reader to the inherent limitations of scientific knowledge. Most noteably. the review of the philosophy of science allows us to draw conclusions for the (IV) Research Position we adopt in this study. The Problem of Scientific Reasoning: Scientists of strategic management tell us that

firms that are highly innovative show higher performance. How do they reach this conclusion? Scientists arrive at certain beliefs through a process of reasoning (Okasha 2002). Economists most often use inductive patterns of reasoning (Phelan 2001). For instance. they found a statistically significant correlation between firm innovativeness and performance when analyzing a sample of 100 firms. Based on this evidence. they inferred that firms that haven·t been examined will behave Similarly to firms of the same kind that have been examined. Hume (1966) claims that this "uniformity of nature" cannot be proven unless the argument is really tested in all firms at all places and times. As this is rationally impossible (in particular for the future). Hume concludes that it is impossible to verify general laws as conclusively true through empirical testing. Later philosophers argued that while the infinite truth of a theory cannot be guaranteed. it can be gradually confirmed. The "concept of probability" implies that scientific knowledge that has been confirmed in multiple empirical tests might not be certain. but is nonetheless highly probable. In contrast. Popper (1959) argued that the problem of induction makes it impossible to prove a theory. Instead. theories need to be falsifiable in

110

order to be scientific. This means that the theory has to make some definite predictions that can be tested. A single disconfirming instance could then disapprove a theory. Poppers theory was later criticized as overly simplistic. First of all, scientist can always blame failure on methodological errors or other factors. Furthermore, in reality, scientists do not automatically reject falsified theories (Putnam 1991). Usually, they adapt their theories to encompass new evidence, which is also a good thing to do in order to achieve progress in theorizing. Consequently, neither confirmation nor falsification can ultimately prove or disapprove scientific knowledge (Curd / Cover 1998, Rosenberg 2000, Phelan 2001). The Problem of Scientific Explanation: Some philosophers believe that the key to scientific explanation lies in the concept of causality (Salmon 1989). In order to explain a phenomenon, it is simply necessary to uncover what caused it. For instance, strategy researchers attempt to reveal the causes of performance differences between firms. However, Hume (1966) argues that while we may observe that, for example, innovative firms often show higher performance, this does not prove that the former brings about the latter. He concludes that there is no causal connection. While other philosophers soften this conclusion, they point to other limitations of scientific explanation. In short, scientists are always specialized in explaining certain aspects of a phenomenon. Management scholars, for example, study the behavior of firms. But firms are in turn composed of human beings, and human beings are living organisms, hence physical entities. However, explanation on the latter levels is left to scientists in biology or physics. Explanation in management is clearly reduced to some elements of the phenomenon. Furthermore, science fails to explain everything. Every explanation implies fundamental laws and principles that themselves remain unexplained (Curd / Cover 1998, Rosenberg 2000, Okasha 2002). The Problem of Scientific Observation: Traditionally, researchers position themselves between the two extremes of positivist and constructivist assumptions (Seale 1999). A positivist epistemology (and the underlying ontological belief in realism) assumes that the outside world can be observed through the researcher in an objective, inter-subjectively stable way. Consequently, researchers in the positivist camp can establish universal, 111

law-like relationships based on their assumptions and empirical testing. For instance, the Bain / Mason industrial organization model and Porter's later work on competitive strategy are positioned in this tradition. They apply empirical testing to verify hypotheses and establish general cause-effect relationships between individual determinants and firm

success.

Opposing

this

stance,

a researcher

positioning

himself within

epistemological constructivism (and thus in the ontological realm of idealism) states there is a "real world" that can be objectively accessed. Observation is seen as nothing more than a subjective construction in the mind of the observer with, evolutionary theorists, for example, having a stronger constructivist orientation (Hill / Deeds 1996). Rather than building complex theories, they are aware of the construction involved in the research process and only assume the "bounded rationality" of economic actors (Foss et al 1995, Fagerberg 2002). Many theories, such as the linkage between diversification and performance, have been successfully tested in different settings. How can we explain these findings, if everything is construction? Realists' no miracles argument states that it is obviously better not to believe in miracles if other plausible arguments exist. Positivists conclude that empirically tested theories might not be correct down to the last detail, but they are approximately true. Constructivists counter that historically, there have been many cases of empirically successful theories that are now believed to be false, such as the PIMS theorizing on the role of market share. From the constructivist perspective, perception is always conditioned by underlying beliefs and truth thus relative to the present, generally accepted, paradigm (Kuhn 1962, Okasha 2002). Research Position: What implications for our study do we draw from the works emanating from the philosophy of science? First and foremost, we need to be very aware of the limitations of our work. For instance, our analysis will be limited to a single industry, and furthermore to a number of firms in this industry, and to certain representatives of each firm. Rather than searching for "generic truths", we should thus focus on establishing "middle range theories" (Farjoun 2002, 563) with a high degree of plausibility. We need to be conscious to the fact that science has always been about reducing the complexity of the world to simple regularities (Phelan 2001, 130). We will not succeed (and not even attempt) to capture every aspect of the dynamic and complex process underlying performance (Porter 1991, 98). As McKelvey (1999,15) states: "No 112

theory ever attempts to represent or explain the full complexity of some phenomenon". Some idealization is always involved, for instance, in terms of a limited number of determinants, a reduced number of interactions, or an exclusion of earlier developments (Suppe 1977, 223f). To some extent our model thus does not explain "real world" behavior; it only attempts to explain "model" behavior. This concept allows us to have multiple models at the same time. However, this does not rule out formal theorizing and quantitative empirical testing (Suppe 1977,228). Without theorizing and testing, science remains purely metaphorical and "difficult to distinguish from witchcraft (McKelvey 1999, 21). In fact, in order to inform practice and to advance theories, it is indispensable to build frameworks and rely on in-depth empirical research (Porter 1991,98). The limitations of theorizing and model building are less bothersome "if models are limited to description and interpretation without attempting to prescribe and predicf' (Thierart I Forgues 1995, 28). Prescription would imply perfect knowledge of the interrelationships between variables and their dynamics over time that we do not have. Global reality is impossible to grasp. Consequently, we take a research position of analytic (or transcendental) realism as the middle ground between the positivist and the constructivist traditions (Seale 1999). This "bridging across boundaries" (Gioia I Pitre 1990) has the advantage of being able to capture the best of both worlds. We are thus consistently aware of the limitations and the constructed nature of our research, but nevertheless strive for some degree of plausibility, reliability, and trustworthiness (Seale 1999). As stated above, this position allows us to establish multiple models. For the different models, we can thus apply theorizing at different levels of abstraction (Nelson I Winter 1982, 46). We use more "formal theorizing" when basing our research on well-established and tested theories, while applying rather "appreciative theorizing" to less explored fields of analysiS. As Nelson and Winter state, both forms perfectly complement each other (item, 46ff).

Reference to Change

Having established our general epistemological research position, we now turn our attention to the characteristics of change processes. Change specifically refers to the

113

firm's market environment, but also to alterations in the firm's strategic behavior or resource base. How do firms and industries evolve over time? How do they adapt (or fail to adapt) to changing environments? Research on these questions is inevitably influenced by basic assumptions of the nature of change. As mentioned above, traditional approaches to strategic management assume that competition, firm behavior, and the resource base converge towards a stable equilibrium position. Evolutionary approaches contradict this assumption and model competition as a dynamic process that never reaches equilibrium. In this section, we briefly reexamine the main assumptions of both the (I) Equilibrium Theory and (II) Disequilibrium Theory. We thereafter present a third perspective of change that is well-established in biology and is increasingly gaining recognition in economics: (III) Punctuated Equilibrium Theory. Finally, we present our (IV) Research Position and outline the advantages of our choice over alternative approaches. Equilibrium Theory: Most traditional theories in strategic management do not pay much

attention to organizational or environmental changes. Cause-effect relationships between individual determinants and performance are expected to rapidly play themselves out to achieve a steady-state equilibrium. Underlying this particular view is the idea of an efficient historical process - an evolutionary process that moves rapidly to equilibrium (Farjoun 2002). Change is thus a quickly optimizing force that brings about stable regularities. Researchers' interest thus concentrates on systems under equilibrium conditions, rather than on the process that led to the current conditions (Barnett / Burgelman 1996). A well-known application of the equilibrium theory is industrial economics' concept of competition initially developed by Mason (1939) and Bain (1956). As presented above, industry structures are seen as largely stable over time. Later research in strategic management applied the same concept to strategy (Le. Porter 1980) and firm capabilities (e.g. Barney 1991). To date the equilibrium theory remains the most used and widely accepted concept of change in strategic management (Foss et aI1995). Disequilibrium Theory: In contrast to the equilibrium theory, other researchers show

more interest in change processes. Based on Darwin's model of evolution as a slow stream of small mutations gradually shaped by environmental selection into new forms, 114

the disequilibrium theory assumes a process of incremental, cumulative change (Gersick 1991). Schumpeter (1942) used similar assumptions to describe competition in capitalist societies. Competition, according to Schumpeter, is an evolutionary process that is never at rest. Equilibrium conditions are rarely if ever reached, revolution is ongoing. We have seen above that researchers from the Austrian school of economics (Jacobson 1992) and evolutionary theory (Nelson / Winter 1982) have drawn heavily on Schumpeter's work. Both perspectives treat competition as an ongoing process driven by entrepreneurial discovery rather than a static notion. Environments are continually shifting,

corporate

competences

are

eroded

by

frequent

discontinuities,

and

"hypercompetition" requires continuous adaptation of firm strategies (D'Aveni 1994, Thomas 1996). Punctuated Equilibrium Theory: Some evolutionary biologists have challenged

Darwin's model that underlies the disequilibrium theory. Eldredge and Gould (1972) postulate a different view of evolution as a punctuated equilibrium. They propose that relatively long periods of stability (equilibrium periods) are punctuated by relatively short bursts

of

fundamental

change

(revolutionary

periods).

Revolutionary

periods

substantively disrupt established patterns and set the foundation for new equilibrium periods (Romanelli / Tushman 1994). During equilibrium periods, persistent underlying structures permit small incremental changes that may be perceived as turbulence on the surface, but the system's basic organization and activity patterns remain the same. The existence of relatively long stable phases is explained by the concept of "deep structure" (Gersick 1991). Deep structure is the set of fundamental "choices" a system has made concerning the internal organization of its parts and the basic activity patterns of these entities. These deep structures are highly stable due to path dependencies and self-reinforcements through multiple feedback loops. In addition, other effects like human resistance to change, sunken costs or existing power structures support the stability of the system. Even the objective presence of a need for change (due to internal or external changes) does not necessarily lead to revolution. The presence of a severe crisis or "burning platform" is essential in setting the stage for revolutionary change (Gersick 1991 ).

115

Research Position: As we strive for a dynamic research model, we can clearly rule out

the equilibrium theory as a basis for our work. Equilibrium theory fails to capture the process component of change. It is simply contrary to all real-life observation for markets to never change or firms to never alter their strategies. Among the remaining alternatives, the disequilibrium theory and punctuated equilibrium theory, we strongly favor the latter. This choice can be explained by the (I) empirical, (II) theoretical, and (III) methodological advantages of punctuated equilibrium approaches.

Empirically, the punctuated equilibrium theory has gained increasing support both outside and within economics. Punctuated equilibrium models have, for instance, been applied to evolutionary biology (Eldredge / Gould 1972), philosophy of science (Kuhn 1962), and psychology (Levinson 1986). More recently, the concept has been applied to management science. Empirical studies found strong support for the punctuated equilibrium model as the most common mode of organizational transformation. Both Miller & Friesen (1980) and Romanelli & Tushman (1994), in analyzing large numbers of firms, found that organizational transformations most frequently follow the patterns described in the punctuated equilibrium model. This was found true for both changes in strategy and structure. Smaller incremental adjustment during equilibrium phases did not amount to fundamental changes (Romanelli / Tushman 1994, 1157). Similar results can be found in case studies of corporate transformations, including that of Tushman et al (1986), who examined the life histories of four organizations, as well as that of Pettigrew (1987). Besides being applied to organizational transformation, the punctuated equilibrium theory has also been applied to analyze environmental changes. In an early study, Tushman and Anderson (1986) demonstrated that technological progress within industries evolves through periods of incremental change punctuated by technological breakthroughs. Furthermore, such discontinuities are rare - in fact, they occurred only eight times in a total of 190 years observed across three industries. In a recent study, McNamara and colleagues (2003) contradict claims that markets have become increasingly hypercompetitive. According to these findings, managers face markets that are no more dynamic today than they were in the 1970s. Perceptions of increasing environmental instability are explained by cognitive biases - past events seem more logical and stable than current or future (Mintzberg 1991). Instead of hypercompetition or 116

disequilibrium, McNamara et al (2003, 275) conclude that market development may rather "reflect a punctuated equilibrium process". While further research may be required, the present empirical studies clearly support the punctuated equilibrium theory as the predominant pattern of change for both organizational and environmental variables. Besides the empirical evidence, there are also a number of arguments for the selection of the punctuated equilibrium theory that can be derived from the theoretical debate. First of all it is interesting to note that contrary to more recent "neo-Schumpeterian" writings (Nelson / Winter 1982, Jacobson 1992, D'Aveni 1994), Schumpeter himself did not rule out equilibrium phases. As Jacobson (1992, 787) notes: "Schumpeter's notion of the market being, at times, in equilibrium separates him from the 'mainstream' Austrian viewpoint. His aim was rather to develop "a theory about economic evolution as a complement (not substitute) to the static equilibrium theory" (Fagerberg 2002, 7). Akin to Schum peter, prominent scholars from strategiC management understand the equilibrium and disequilibrium theories to be more complementary than contradictory (Le. Porter 1991, 105; Jacobson 1992, 784; Nelson 1995, 50; Barnett / Burgelman 1996, 17). For instance, Barney (1986, 792) stated that equilibrium and disequilibrium approaches are "not logically contradictory, but rather, taken together, form a more complete picture of competitive forces facing firms". Barney stresses equilibrium competition as strategically more relevant for most of the time, while "Schumpeterian shocks" occasionally gain importance when interrupting equilibrium phases (1986, 798). Much along the same lines, Foss et al suggest a compromise position by preserving some elements of both equilibrium and disequilibrium positions (1995, 11). The punctuated equilibrium theory provides the theoretical foundation for these claims. Bridging the old rivals - the equilibrium and disequilibrium theories - the punctuated equilibrium perspective welcomes theorizing from both camps, promoting crossfertilization and the emergence of a richer understanding of competition and strategy. Finally, the cross-fertilization may also affect the methodological realm. As stated above, proponents of the dynamic perspective often criticized cross-sectional studies as inappropriate to capture dynamic processes (Le. Jacobson 1992, 791f). In the same manner, traditional strategists criticize the reluctance of disequilibrium studies to use 117

analytical tools, or to develop testable propositions (Le. Barney 1986, 795). Under punctuated equilibrium assumptions, methods applied by both sides are justified. During long periods of relative stability, results of cross-sectional studies do have certain relevance. However, they have to be complemented by longitudinal analysis in order to capture revolutionary changes and provide a richer picture of the dynamic realities that elude cross-sectional studies. Surprisingly or not, both Porter (1991) and Nelson (1995) admit that methods used by both sides might perfectly complement each other. Cross-sectional statistics allow for formal testing of concrete assumptions with clear cause-effect relationships, while longitudinal qualitative studies provide more "realistic" explanations. Gersick (1991, 33) thus suggests "multilevel research" as most appropriate from a punctuated equilibrium perspective.

Reference to Complexity and Dynamics

Taking a punctuated equilibrium perspective acknowledges the importance of change and dynamic evolution over time. In contrast, contingency approaches deal with the dynamics between individual determinants of firm performance. Proponents of a more integrative approach to strategy claim a better understanding of the complex processes underlying performance. What different forms of complexity and dynamics need to be distinguished? How are complexity and dynamics interrelated? What can be done to capture complexity and dynamics in our research model? Drawing on both complexity theory and evolutionary theory, we present three distinct constructs describing various facets of complexity and dynamics. More precisely, we distinguish (I) Detail Complexity, (II) Dynamic Complexity, and (III) Evolutionary Dynamics. Based on the review of related concepts, we draw conclusions for our (IV) Research Position?

7

Much of the complexity literature reviewed in this chapter is based on earlier writings of systems dynamics (i.e. Wiener 1961, Gomez / Probst 1987, Vester 1988). This literature has been reviewed, but is not presented in detail in this study.

118

Detail Complexity: Let's assume for the moment that we would like to explain the success of Real Madrid, the world's leading soccer team. If we ask five soccer fans to explain Real's dominant position, we might indeed get five different answers. Some might assert that one famous player makes all the difference; others may point to the financial strength of the team, to the coach, to the management, or (and these are probably fans of their archival Barcelona) to the partiality of referees. The more of these factors we take into account to explain Real's performance, the more complicated our explanation gets. In mathematical terms, we can talk of the dimension of the causal system. Dimensionality relates to the number of different independent variables affecting the output of a system (Dooley / Van de Ven 1999,364). A simple explanation analyzes one or a few variables, much as the majority of the present theories in strategic management do. Taking a higher number of variables into account would create a "complicated system" (Devaney 1993). Senge (1990) speaks of "detail complexity" when referring to

the number of variables captured by a model. Integrating more variables provides a more realistic picture. In our model, we thus strive to integrate variables from various perspectives that could explain firm performance. However, while complexity theory describes systems that capture extremely high numbers of variables, so-called "white noise" systems (Dooley / Van de Ven 1999), Senge points to the limitations of such approaches for management science (1990, 71f). Human beings have cognitive limits when mastering larger number of variables simultaneously. It is impossible to model hundreds or thousands of determinants into a model. Due to this enormous detail complexity, all rational explanations remain inherently incomplete (Senge 1990, 365). However, theorists are free to obtain relatively higher (or lower) detail complexity. Dynamic Complexity: Even if theoretically, we could take into account all variables related to performance, we would still fail to fully capture the inherent complexity of the problem. Let's go back to our soccer example for a moment. Excellent players and a high budget certainly explain some part of Real Madrid's superior performance. However, other clubs like FC Barcelona or AC Milan have comparable talent and budgets - why do they fail to reach the same level of success? For instance, while all three clubs have heavily invested in world-class players, the interplay between these players did not work out in Barcelona and Milan. Expensive players frequently remain on 119

the bench while less talented players who fit into the tactical system better are allowed to play. From this simple example alone, we can see that interrelations between individual factors (such as players, money, tactics etc) contribute a great deal to performance. Senge (1990, 71f) speaks of "dynamic complexity" when referring to systems that capture interactions between causal factors. Contrary to the static concept of "detail complexity", such systems are dynamic: "Change is the natural state of being of the complex system" (Black 2000, 522). In keeping with the nature of this dynamic interaction, Dooley and Van de Ven (1999) further distinguish between systems with linear interaction ("periodic") and nonlinear interaction ("chaotic") between variables. Complexity theory (Stacey 1995, McKelvey 1999) has mainly focused on the latter type, complex systems with nonlinear relationships between system elements that interact (e.g. through feedback loops). However, far from assuming random interaction, "complexity scientists are seeking simple rules that underpin complexity" (Phelan 2001, 130). Much as Senge above, Sterman (2001, 10) discusses "the mismatch between the dynamic complexity of the systems we have created and our cognitive capacity to understand that complexity". We can neither understand the full range of feedbacks operating in the system, nor anticipate all the side effects and time delays our actions may trigger (Sterman 2001). However, a lower degree of "dynamic complexity" can be captured by models and tested (Anderson 1999). This can be achieved by concentrating on the most important interactions and/or assuming linear relationships between determinants (Melcher / Melcher 1980, 236). Evolutionary Dynamics: If we assume for the moment that Real Madrid's success during the turn of the millennium is due to its tremendous forwards, this does not say much about its future success. Crucial players may end their career or move on. A new coach may change the tactical system. The world soccer association FIFA might reduce the number of foreign players allowed in the European soccer league, and so on. While "dynamic complexity" provides a pretty accurate picture of how the current system works, it does not necessarily explain the development of the system over time. Changes in key determinants may change the whole system. Evolutionary economics thus define 120

dynamics as a process over time (Nelson 1995, 54). "Evolutionary dynamics" explain the current status of the system in terms of how it got there. Several strategy researchers claimed that such a "temporal componenf' (Priem / Butler 2001, 35) or "longitudinal perspective" (Porter 1991, 105) should be added to existing static theories. Some

applications of complexity theory include this dynamism in the system over time (Black / Farias 2000, 104). Research Position: First of all, we conclude that the three perspectives of complexity and dynamics as described above are independent of one another as well as being complementary. Furthermore, all three perspectives will be required for a truly integrative and dynamic approach to strategic analysis. This conclusion is in line with Porter's (1991) prior work on dynamic strategy. He stresses that every fully dynamic model first of all needs a theory that links firm and environmental characteristics to outcomes. Next, the dynamic interactions between the selected model parameters need to be captured. Finally, assuming "an understanding of the cross-sectional problem, the longitudinal problem takes on prime importance" (Porter 1991, 105). Supporting Porter's view, other

authors state that the evolutionary component is "less a distinct theory than a longitudinal perspective on some other underlying cross-sectional or decontextualized theory"

(Spender 1996, 45; similar: Malik / Probst 1982, 157; Barnett / Burgelman 1996, 17). The three perspectives can thus be analyzed independent of one another, however, all three need to be integrated to achieve a fully dynamic model. While we strive for a dynamic model, we will later have to reduce the inherent complexity and dynamics to a practicable level: "Models that exceed a certain level of complexity can never be tested" (Porter 1991, 116). Complexity is context dependent and in many instances very subjective in definition: "Things are only as complex as the actor / observer makes them" (Black 2000, 522). Complexity science has a rather positivist research agenda to discover generative rules and equations that are capable of explaining some of the observed complexity (Phelan 2001, 133). We should follow this example and reduce the complexity of our general model where required, in order to allow profound empirical testing.

121

Overall Frame of Reference

From the philosophy of science we derived the insight that our model cannot fully capture complex reality. However, theorizing can strive for high plausibility by applying different levels of abstraction, including both formal and appreciative theorizing. Furthermore, we made a strong case for understanding change as punctuated equilibrium. This perspective enjoys growing empirical support and combines the strong

points of both the equilibrium and disequilibrium models. Finally, we have shown that a fully dynamic model needs to account for three different perspectives: detail complexity, dynamic complexity, and evolutionary dynamics. How can these different insights be

integrated? Figure 7 provides an overview of the overall frame of reference combining the different elements.

Dynamics ffi~-------------------------------------'

Disequilibrium

Full

Appreciative Theorizing Evolutionary

Formal Theorizing Equilibrium System

Static

eL-______~--------~--~----,L--------4 ffi Dynamic Full e Simple Detail Figure 7

Complexity

Frame of Reference

The basic rationale behind the frame of reference is the paradox that increasing levels of complexity and dynamics lead to a better understanding of reality, but at the same time reduce our ability to explain this reality in general and straightforward theories. A low to medium degree of complexity (simple or detail) can be captured in research models based on formal theorizing (Le. M1 I M2). Moving towards higher complexity, we quickly 122

reach the limits of formal theorizing. A low degree of dynamic complexity (Le. M3) can still be captured by formal models; higher degrees of dynamic complexity require less formal "appreciative" theorizing, while full complexity escapes any attempt at theorizing due to the involved random elements. The same effect is true for increasing dynamics. A low to medium level of system dynamics (Le. M3) can be captured by formal theorizing. Evolutionary dynamics, however, require abolishing equilibrium assumptions, thus quickly reaching the frontiers of formal theorizing, particularly if combined with some degree of complexity. A simultaneous acceptance of medium to high levels of both dynamics and complexity can only be recognized through appreciative theorizing (Le. M4). Full dynamism, especially in combination with full complexity, is outside any attempt at theorizing. Our recommendation for dynamic approaches to strategy is thus to combine multiple models along the axis M1-M4. Lower level models allow more formal theorizing that deploys the rich existing body of theories and methodologies based on equilibrium assumptions. Based on this foundation, additional appreciative theorizing allows for "journeys" into higher complexity and dynamism based on more innovative research methods. The combination of two or more such models allows for cross-fertilization between a large variety of existing theories and methods. It is also the only practical approach with which to achieve our research objective: The establishment of a more integrative, complex, and dynamic approach to strategic analysis. A single model can never provide the same combination of complexity and dynamism on the one side, and validity and generalizability on the other side.

111.4

Model Development & Hypotheses

In this section, we respond to our first research objective in establishing a comprehensive research framework for the strategic analysis of dynamic and complex industries. First, we present the (I) Overall Research Framework which comprises three distinct models: the integrated, the complex, and the evolutionary model. Each model is briefly described in terms of its particular contribution to our research objective. The models are then positioned within the overall frame of reference developed in the

123

previous section. Next, we develop distinct sets of research hypotheses for each model, which will guide our further analysis. We exhaustively develop (II) Research Propositions 1: The Integrated Model, (III) Research Propositions 2: The Complex Model, and (IV) Research Propositions 3: The Evolutionary Model. Finally, we draw first (V) Conclusions

regarding our established framework and then introduce the empirical part of this study.

Overall Research Framework

In the first section of this chapter, we went to great lengths to review the widespread criticism of classical approaches as fragmented (Le. Porter 1991), deterministic (Le. Henderson I Mitchell 1997), and static (Le. Barnett I Burgelman 1996). The dynamic research framework presented in this section responds to this criticism by providing a more integrative, complex, and evolutionary approach. We follow the advice of Pettigrew and colleagues (2002, 10), recommending that the ''jigsaw should be put together piece by piece rather than attempting a Herculean synthesis". Therefore, we propose a

framework that unites three sequential models (see figure below): the integrated (1), the complex (2), and the evolutionary (3) model.

(1) The Integrated Model

Figure 8

(2) The Complex Model

Dynamic Research Framework

124

(3) The Evolutionary Model

The Integrated Model: The review of the writings from four dominant perspectives of

strategic management in chapter two (industrial organization, business strategy, corporate strategy, and resource-based view) has shown that each of these schools provides valuable and complementary insights into the determinants underlying corporate success. Variance decomposition studies prove that all four perspectives are required to explain performance (i.e. McGahan / Porter 2002). Despite these insights, the present approaches remain often fragmented and polarized (Porter 1991, Shay / Rothaermel 1999). The integrated model addresses the problem of fragmentation by capturing the most important determinants derived from the different perspectives on strategic management. More exactly, we analyze the influence of market structure, firm capabilities, business strategy, and corporate strategy on company performance within a single model. Our approach spans three levels of analysis: industry, corporate, and business unit. Based on a "comparative" approach to multi-level analysis, the model looks at the relative (but not yet interactive) effect of factors at multiple levels on performance (Dess et al 1995, 382f). The integrated model thus provides an answer to the multitude of claims for more integrated research that bridges two or more levels of strategic research (Beard / Dess 1980, Jemison 1981, Dess et al 1995, Hoskisson et al 1999, Shay / Rothaermel 1999, McNamara et al 2003). How does the integrated model fit into the frame of reference developed above? First of all, the integrated model remains clearly equilibrium-based. We analyze determinants underlying firm performance at a given point in time. Understanding change as a punctuated equilibrium process, the integrated model focuses on the long phases of relative stability. During these phases, the system may not be exactly at rest, but the fundamental mechanisms remain stable (Romanelli / Tushman 1994; Nelson 1995, 50). Due to the relative temporary stability, it makes sense to reach a better understanding of the factors underlying success under current equilibrium conditions. Porter (1991, 105) stresses that a thorough understanding of the currently desirable position (in his terms the "cross-sectional problem") has to come first, before dealing analytically with the process of getting there.

125

By taking equilibrium assumptions, the model consequentially reduces reality by excluding longitudinal (or evolutionary) developments. In addition, reality is further simplified by excluding dynamics between individual parts of the model (or determinants) from analysis. Instead of showing dynamic complexity, the model merely strives to grasp detail complexity (Senge, 1990).

This reduction of real world complexity allows a high degree of formal theorizing to be applied (Nelson / Winter 1982). We can thus tap into significant established work from various classical perspectives on strategy. Based on the literature review in chapter two, we will be able to derive testable hypotheses on cause-effect relationships. However, while we attempt to grasp detail complexity through formal theorizing, it has to be pointed out that we will not be able to capture every determinant underlying performance (Porter 1991, 98; McKelvey 1999, 15). There are hundreds of variables that may influence performance, and it is impossible to incorporate all of them into a practicable model (Jarillo 2003, 23). Nevertheless, we are confident that our detailed review of the existing strategic literature allows us to select and incorporate the most relevant determinants of firm performance into our model. Depending on the results of the empirical testing, the model is flexible to changes in the individual determinants it incorporates. New evidence may suggest dropping or replacing old determinants, or to add new determinants to the model. The flexibility of the model to such changes makes it valuable for future research and allows the model to be further improved of to incorporate increasing detail complexity.

The Complex Model: While the integrative model includes various determinants, it disregards the inherent interdependence of the determinants. However, empirical evidence points to the important relationships between individual determinants (Bowman / Helfat 2001, Phelps 2001, McGahan / Porter 2002). But present studies fail to recognize the multiple and reciprocal causality between individual elements (Henderson / Mitchell 1997, Stimpert / Duhaime 1997, Farjoun 2002). Various authors have thus remarked on the need for more complex models that capture these interactions and their indirect influence on performance (i.e. Phelps 2001). The complex model reflects these claims in providing an "interdependency approach" (Dess et al 1995, 383) that captures the interactive effect of variables at multiple levels on performance. 126

The complex model shares equilibrium assumptions with the integrated model. "Model behavior" differs from reality through the exclusion of longitudinal developments (Suppe 1977, 223f). However, the model moves from detail complexity towards dynamic complexity (Senge 1990). As we have seen in the last section, low levels of dynamic

complexity allow for a certain degree of formal theorizing, while analysis into higher levels of complexity are restricted to appreciative theorizing (Nelson I Winter 1982). More exactly, the line of demarcation runs between systems with linear and non-linear interactions among variables. Formal theorizing largely fails to acknowledge more complex reciprocal and non-linear interactions, including various feedback effects. This failure can be explained by both the significant complexity of such systems, and the limited availability of prior theorizing on these systems. Conversely, assuming linear relationships between determinants allows us to draw on findings from the contingency theory and to apply traditional cross-sectional analysis for formal testing. The contingency theory pOints to a number of predominant linear relationships between determinants. We will thus apply formal theorizing to describe the most influential linear relationships between determinants. While this approach allows for a considerable understanding of dynamic complexity, it fails, as stated before, to account for the full complexity of dynamic interactions. Akin to the integrated model, we seek a thorough identification and analysis of the qualitatively most important effects, rather than a high quantity of interactions. This approach furthermore contributes to the advancement of scientific knowledge on interaction between determinants. As the body of knowledge on these interactions grows, the level of dynamic complexity captured by future models can also increase incrementally. The Evolutionary Model: The complex model explains a fair deal more of the complex

interaction underlying performance than the simple integrated model. However, both models share an inherently static character. Taking a punctuated equilibrium perspective, we acknowledge that the present equilibrium conditions may be interrupted by revolutionary phases of discontinuous change. The dynamic evolution of determinants underlying performance is an important (though most often neglected) element of strategic analysis (Nelson I Winter 1982, Porter 1991, Barnett I Burgelman 1996, Black I Farias 2000). 127

The evolutionary model adds a longitudinal perspective to our dynamic framework. Rather than describing the current equilibrium, the focus of the evolutionary model is on the disequilibrium phases of change and the process that leads to the current conditions. Shedding equilibrium assumptions, formal theorizing is rather complicated. Due to the extremely limited prior evolutionary research into strategy, we apply appreciative theorizing to the evolutionary model. This moreover allows us to maintain some degree of detail and dynamic complexity, besides the main focus on evolutionary dynamics. The evolutionary model thus completes our dynamic framework through less formal, but significantly more dynamic and complex modeling. From an overall perspective, the three models perfectly complement one another. Sequentially, they provide a more integrative, complex, and dynamic picture of strategic interactions underlying the success of firms. The many useful, but conceptually unconnected, insights from different schools of thought are pragmatically integrated. The strengths and weaknesses of different approaches to theorizing (and the related methodology) are carefully balanced to allow for insights that are simultaneously realistic, valid, applicable, and generalizable. The model comes close to what strategy managers have always had to accomplish in practice - the coherent integration of many complex and dynamic facets of strategy and competition.

Research Propositions 1: The Integrated Model

As described above, the integrated model is based on equilibrium assumptions and allows for formal theorizing. In this section, we formulate a first series of research hypotheses drawing on findings from our literature research in chapter two. Cause-effect relationships

between

individual

determinants

and

company

performance

are

established for four traditional fields of strategy theory: (I) Industry Structure, (II) Corporate Strategy, (III) Competitive Strategy, and (IV) Firm Capabilities. Figure 9 provides an overview of the cause-effect relationships to be established in this section. The hypotheses form the basis of the later empirical research in this study. Based on our prior argumentation in this chapter, the main hypothesis of the integrated model is as follows:

128

Explaining variance in firm performance requires an integrated model embracing the most important determinants from competing perspectives of strategic management.

H1

I Competitive Strategy

Industry Structure

Differentiation Cost Leadership Balancer Strategy Innovativeness

Firm Capabilities

MarketOrientation arg. Learning Reputation Corporate Culture

~...

I

o

...o0..

I Concentration Industry II

II

I

Entry Barriers

~

7

I I

i

i I I I I

Corporate Strategy

Market Growth Rate

r--

..

~

:::::::'J

~

:. ~

I Diversification Product II Diversification International I I II

Vertical Integration

I

(.)

Figure g

The Integrated Model

In order to test and further detail the lead hypothesis, we continue to formulate hypotheses for the

underlying

relationships

between

single

determinants and

performance. These hypotheses are based on the four strategy perspectives discussed above: Industry Structure: 8 The central construct of theorizing in industrial economics, industry structure, is represented by three variables: (I) industry concentration, (II) entry barriers, and (III) market growth rate. While various other structural variables have been discussed in the literature, these three have been identified as core to explain structural differences between industries (Capon et aI1990).

B

Please refer to chapter 11.1 for a detailed review of the relevant industrial organization literature.

129

Industry Concentration. Industrial organization posits that industry structure variables determine firm strategy that, in turn, determines business performance (Bain 1959). The degree of seller concentration is seen as one main influencer of industry structure. Firms in highly concentrated markets profit from their monopoly power and the possibility of collusion with other firms (Scherer I Ross 1980). This position allows firms to charge higher prices, thus increasing their performance (Mann 1966). Consequently, we expect to find that: H1.1

Increasing levels of industry concentration will positively impact business unit performance.

Entry Barriers. Markets with high entry barriers reduce the risk of new competitors entering the market. The presence of strong entry barriers allows concentrated sellers to persistently raise prices without attracting new competitors (Bain 1956). Entry barriers include high capital requirements, as well as absolute cost advantages, economies of scale, and the product differentiation of established players. Bain's (1956) claim that entry barriers influence performance regardless of the degree of market concentration has been supported by numerous empirical studies. Hence, H1.2

The increasing strength of entry barriers will be positively linked to business unit performance.

Market Growth Rate. Attractive markets are often characterized by high growth. While growth markets bear a higher risk, often require higher investments, and are more likely to attract new competitors, they facilitate market share growth and allow for higher prices (Aaker I Day 1986). All in all, these effects should have a combined positive effect on performance (BuzzellI Gale 1987): H1.3

High market growth rates will positively impact business unit performance.

130

Corporate Strategy:9 Corporate strategy's concern with the firm's general positioning

and the joint management of its business units dominates its decisions on diversification (Chandler 1962). Diversification can take three different forms: (I) product diversification, (II) international diversification, and (III) vertical integration (Grant 2002). All three forms are seen as independent constructs that collectively represent corporate strategy and are linked to firm performance (Hitt et al 1997). Product Diversification. Extending a firm's product scope through entry into new market

segments can lead to superior performance (Gort 1962, Markham 1973). A moderate level of product diversification, or "related diversification", allows for synergies and efficiencies between units, thus increasing profitability (Dess et al 1995). However, higher levels of product diversification, or "unrelated diversification", are also associated with increasing administrative costs and diseconomies related to organization (Markides 1992). Studies on the media industries assume a generally positive relationship (Albarran 1990). Empirical studies' findings are rather mixed. While the majority confirms the Significant effect of diversification on performance, the direction of this effect varies between positive, negative or curvilinear (Dess et al 1995). Based on the specific findings for the media industries, we formulate the following hypothesis: 1o

I H1.4

Product diversification is positively related to business unit performance.

Intemational Diversification. A broad geographical scope permits the firm to better exploit

economies of scale and scope, to realize cost advantages, and to make more efficient use

of its

resource

base

(Caves

1971,

Kogut

1985).

However,

increasing

internationalization also entails growing transaction and coordination costs (Geringer et al 1989). More recent studies suggest a curvilinear relationship between international diversification and performance (Geringer et al 1989, Hitt et al 1997). Based on these studies, we suggest the following hypothesis:

9 10

Please refer to chapter 11.2 for a detailed review of the relevant literature on corporate strategy. The alternative hypotheses (negative, curvilinear) will also be tested. An additional argument for a positive relationship is that the investigated media companies' diversification is limited to related fields.

131

H1.5

International diversification has a curvilinear relationship to business unit performance.

Vertical Integration. Vertically integrated businesses enjoy reduced general and administrative, other selling, advertising, and R&D expenditures (D'Aveni / Ravenscraft 1994). Conversely, vertical integration also leads to increased production costs, strategic inflexibilities, and coordination overheads (Mahoney 1992). Empirical studies disagree on the combined overall impact of these positive and negative effects on performance. On the whole, results from empirical studies point to an overall positive relationship between vertical integration and performance (Capon et aI1990). Thus,

I H1.6

Vertical integration is positively related to business unit performance.

Competitive Strategy: 11 Most competitive interaction occurs at the business unit level

(Porter 1980). These competitive actions have been clustered into "generic strategies" (Miles / Snow 1978, Porter 1980). Firms applying a consistent generic strategy deliver a performance superior to those firms that do not (Conant et al 1990). Porter's strategy types of "cost leadership" and "differentiation" have frequently been associated with increased performance: H1.7

A cost leadership strategy will be positively linked to business unit performance.

H1.8

A differentiation strategy will be positively linked to business unit performance.

Furthermore, the "balancer" strategy, by focusing on both cost excellence and differentiation, has been connected to high profitability (Wright et al 1990, Parnell 1997, Parnell et al 2000). H1.9

11

A balancer strategy will be positively linked to business unit performance.

Please refer to chapter 11.3 for a detailed review of the relevant literature on business strategy.

132

Firm Capabilities: 12 The essence of the resource-based perspective on strategic management is the theory that a firm that effectively establishes and deploys certain firm-specific resources can achieve superior performance (Barney 1986, Rumelt 1987). Numerous resources have been identified and related to performance. However, we restrict our research to five capabilities that have been shown as particularly important to explain performance: (I) market orientation, (II) reputation, (III) organizational culture, (IV) organizational learning, and (V) innovation. The latter two capabilities are derived from the "dynamic capabilities" perspective and represent the ability of the firm to permanently enhance and renew the existing resource base of the firm. The recognition of these factors provides the integrated model with a "touch of dynamism". While inherently static, the model already incorporates capabilities that are crucial for a company's ability to adapt to changing market conditions. Market Orientation. The market orientation concept refers to an organization's market

focus and customer orientation. Firms with high market orientation create superior value for their customers leading to improved company performance (Kohli I Jarworski 1990, Narver I Slater 1990). H1.10

Increasing market performance.

orientation

is

positively

related

to

business

unit

Reputation. A firm's reputation is a relatively stable strategic asset with a positive effect

on both product differentiation and cost advantage of the firm (Rumelt 1987). A positive reputation allows the charging of higher prices (Klein I Leffler 1981, Milgrom I Roberts 1986) and reduces contracting, monitoring, and factor costs (Stigler 1962, Roberts I Dowling 2002). Reputation is thus linked to superior company performance (Fombrun I Shanley 1990, Landon I Smith 1997, Roberts I Dowling 2002). H1.11

Reputation is positively related to business unit performance.

Organizational Culture. Successful organizations are distinguished by their particular

cultures that support chosen strategies (Peters I Waterman 1992). While several studies 12

Please refer to chapter 11.4 for a detailed review of the resource-based literature. and chapter 111.2 for a discussion of the dynamic capabilities perspective.

133

confirmed the importance of culture on firm performance (Denison 1990, Gordon I Di Tomaso 1992, Marcoulides I Heck 1983), more recent studies have shed light on the particular characteristics of well-performing cultures. In particular, entrepreneurial cultures have been found to consistently outperform other types of organizational cultures (Deshpande et al 1993, Deshpande and Farley 1999, Ogbonna I Harris 2000, Homburg I Pflesser 2000). H 1.12

"Entrepreneurial" cultures are positively linked to business unit performance

Organizational Learning. Continuous learning is essential to create and develop new

capabilities and is thus a main source of competitive advantage (Pisano 1994, Lei et al 1996, Eisenhardt I Martin 2000). Learning orientation, the organization-wide activity of creating and using knowledge to enhance competitive advantage, is central to firm performance (Slater I Narver 1995, Hurley I Hult 1998, Baker I Sinkula 1999, Calantone et al 2002). Based on the theoretical and empirical evidence, it is proposed that: H 1.13

Learning orientation is positively linked to business unit performance.

Innovation. Similar to learning, innovation is crucial to allow the firm to create new

products and processes that respond to changing market circumstances (Teece I Pisano 1994, Hurley I Hult 1998). Firm innovativeness has thus been described as an important dynamic capability of the firm (Lawson I Samson 2001), exerting a positive impact on firm performance (Geroski 1994, Crepon et al 1998, Vasquez et al 2001, Calantone 2002). Hence, H1.14

Firm innovativeness is positively related to business unit performance.

Research Propositions 2: The Complex Model

The integrated model attempts to capture the most important determinants of firm performance. Based on the integrated model, we now move towards a better understanding of the various interactions between individual determinants. The complex 134

model draws heavily on our discussion of the contingency theory earlier in this chapter. While still assuming equilibrium conditions, we establish assumptions on relationships between determinants and their combined influence on performance. The focus is on the most relevant interactions and assumes linear cause-effect relationships. Based on the findings from the contingency literature discussed above, the main hypothesis for the complex model is as follows: H2

Explaining variance in firm performance requires a complex model capturing the most relevant interrelations between individual determinants.

In order to evaluate and further concretize this main hypothesis, we continue to

establish a number of additional hypotheses: Competitive Strategy - Market Structure: Research by both industrial organization

economists and proponents of the efficiency school has confirmed a reciprocal relationship between competitive strategy and market structure (Bain 1959, Demsetz 1973, Porter 1991). More precisely, the emergent contingency view of generic strategies suggests that different types of competitive strategy are more or less appropriate in different market environments (Hill 1988, Murray 1988, Lee / Miller 1996). Cost leadership strategies have been found to be best suited to stable environments (Dutton 1982, Kim / Lim 1988, Miller 1988, McCabe 1990). Stable environments are often characterized by high industry concentration. Consequently, we assume: H2.1

The relationship between cost leadership strategies and business unit performance is positively impacted by the degree of industry concentration.

Differentiation strategies, in contrast, are more efficient in dynamic environments (Hambrick 1983, Miller I Friesen 1984, Kim / Lim 1988). Dynamic environments are often characterized by high market growth. Hence, H2.2

The relationship between differentiation strategies and business unit performance is positively impacted by increasing market growth.

135

Competitive Strategy - Capabilities: A perfect match between firm capabilities and

selected business strategies is crucial for firm success (Porter 1980, Snow I Hrebiniak 1980, White 1986, Miller 1988, Grant 1991). For instance, the successful implementation of a differentiation strategy requires a high degree of market orientation (Slater I Narver 1996, Morgan I Strong 1998, Matsuno I Mentzer 2000, Kumar et al 2001, Vazquez et al 2001), innovativeness (Snow I Hrebiniak 1980, Hambrick 1983, Vazquez et al 2001), and organizational learning (Slater I Narver 1995, Calantone et al 2002). All three capabilities are seen as fundamental elements of a differentiation strategy. Thus, H2.3

The effect of a differentiation strategy on business unit performance is partly explained by market orientation.

H2.4

The effect of a differentiation strategy on business unit performance is partly explained by innovativeness.

H2.5

The effect of a differentiation strategy on business unit performance is partly explained by organizational learning.

Corporate Strategy - Competitive Strategy: Numerous authors stress the importance

of the interrelations of strategic actions at the corporate and the business unit levels (Dess et al 1995, Stimpert I Duhaime 1997). However, little research has been undertaken to shed light on these interactions. The exceptions are a number of studies that associate attributes of cost leadership strategies (such as low levels of R&D expenses and capital investments) with high levels of product diversification (Bettis 1981, Hoskisson I Hitt 1988, Hill I Snell 1989). Based on this limited evidence, we formulate the following hypothesis: H2.6

The effect of product diversification on business unit performance is partly explained by a cost leadership strategy.

Conversely, characteristics of differentiation strategies, such as innovation, new product development, and customer orientation, have been associated with high levels of international diversification (Young 1987, Roth I Morrison 1992).

136

H2.7

The effect of international diversification on business unit performance is partly explained by a differentiation strategy.

Vertical integration has been linked to the attributes of differentiation strategies (such as innovation or product development) as well as cost leadership strategies (Armour I Teece 1980, Buzzell 1989, Prabhu I Harrison 1992, D'Aveni I Ravenscraft 1994). We therefore conclude: H2.8

The effect of vertical integration on business unit performance is partly explained by a balancer strategy.

Corporate Strategy - Market Structure: Scholars from both industrial organization and

strategic management have assumed a strong interrelation between degree and type of diversification and industry structure (Datta et al 1991). However, to date not much progress has been made to investigate these interrelations. Hence, there is some support for the supposition that successful diversification is more likely in companies operating in less attractive markets than firms in more profitable markets (Rumelt 1974, Christensen I Montgomery 1981, Montgomery 1985). Based on empirical findings of Miller (1973) and Bass and colleagues (1977), we argue: H2.9

The relationship between product diversification and business performance is negatively impacted by industry concentration.

unit

Much along the same lines, market conditions are expected to exert a strong influence on the vertical integration-performance relationship. Vertical integration is expected to be more appropriate in stable, concentrated industries than in dynamic industries. While empirical results are far from being unequivocal, the majority supports this assumption (MacDonald 1985, Jones I Hill 1988, D'Aveni I llinitch 1992, Prabhu I Harrison 1992, D'Aveni I Ravenscraft 1994). Therefore: H2.10

The relationship between vertical integration and business unit performance is positively impacted by industry concentration.

137

Corporate Strategy - Capabilities: The success of corporate strategies depends on

the firm's available resource profile (Hitt / Ireland 1986, Chatterjee / Wernerfelt 1991). Pennings et al (1994) found that the success of product diversification was strongly associated with organizational learning and know-how. Therefore, H2.11

The effect of product diversification on business unit performance is partly explained by organizational learning.

Vertical integration has been related to the firm's innovative ness. Returns from innovation have been found to be particularly high in vertically integrated firms (Armour / Teece 1980, Buzzell 1983, Heeb 2003). Therefore, H2.12

The effect of vertical integration on business unit performance is partly explained by innovativeness.

The success of vertical integration moves has also been linked to organizational learning. Vertical integration stimulates and, in turn, benefits from organizational learning (Armour / Teece 1980, Sorenson 2003). Therefore, H2.13

The effect of vertical integration on business unit performance is partly explained by organizational learning.

Capabilities - Market Structure: The effectiveness of capabilities greatly varies with

the market environment's different characteristics (Brush / Artz 1999, Eisenhardt / Martin 2000, Zajac et al 2000). Capabilities and competitive environments have reCiprocal relationships (Levinthal / Myatt 1994, Henderson / Mitchell 1997). Corporate culture has, for example, been linked to market environment (Gordon 1991, Kotter / Heskett 1992, Burt et al 1994, Chatman / Jehn 1994, Sorensen 2002). Although empirical evidence is rather mixed to date, the following formulated hypothesis is based on present studies: H2.14

The relationship between entrepreneurial culture and business performance will be positively impacted by market growth rate.

138

unit

Research Propositions 3: The Evolutionary Model

Having established the integrated and complex models, we continue with the evolutionary model. This model overcomes the static character of the first two models. The evolutionary model concentrates on phases of discontinuous change, rather than equilibrium conditions, thus adding a longitudinal perspective to our overall research framework. Besides the main focus on evolutionary dynamics, we intend to discern some of the inherent detail and dynamic complexity. We apply appreciative theorizing to acknowledge the limited prior research and the high degree of complexity and dynamics that we intend to capture. Appreciative theorizing belongs to the tradition of theory building rather than theory testing. In contrast with the deductive (theory driven hypothesis testing) approach of the first two models, we now apply an inductive (datadriven generalization) approach. Based on our review of evolutionary theory above, the main hypothesis for the evolutionary model is as follows: H3

Explaining variance in firm performance requires an evolutionary model capturing the most relevant developments and changes over time.

Instead of allowing concrete hypotheses to restrict our research focus right from the beginning, we initially content ourselves with two general research propositions derived from the existing literature. The first research proposition is based on the assumption of consistent patterns of change across industries (i.e. Tushman / Anderson 1986). Webb and Pettigrew (1999, 601), for example, found "characteristic patterns of strategy development that occur at both firm and industry level over time". This research relies on earlier theorizing on the diffusion of common strategies within industries (Spender 1989). Firms within an industry face similar challenges. If the firms approach these problems in a rational manner, it is highly likely that they will arrive at the same strategic decisions (Porter 1980). The first research question is thus: H3.1

Are there industry-wide characteristic patterns of dynamic change in key strategic variables underlying firm performance?

139

Let's assume for the moment that such characteristic patterns of change do exist. The next question is then, how the different changes affect business unit performance. For instance, if we find that at some point in time, firms turn their focus towards cost leadership strategies, it is particularly interesting to detect the consequences of this move. Did the increased focus on cost leadership lead to an increase in performance? A number of studies have analyzed the dynamic interaction between different strategic variables and performance over time (Le. Young et al 1996). This leads us to the second research question of the evolutionary model: H3.2

Are there interrelations between changes in different strategic variables and performance over time?

Conclusions

The present chapter has been dedicated to the development of a research framework for dynamic strategic analysis. Based on existing research and the criticism thereof, we developed a framework that integrates different schools of thought and captures some of the complex and dynamic character of real life strategizing. Our intention was to find the happy medium between static and dynamic, simplicity and complexity, equilibrium and process. Only the middle way permits the benefits of both the extensive traditional strategy research and the promising dynamic school to be enjoyed, while minimizing the inherent shortcomings of both sides. The cross-fertilization will be beneficial on both theoretical and empirical grounds. In addition, we separated the theoretical backbone of our overall framework from the concrete research models and hypotheses. This approach leads to the significant flexibility and modularity of the framework. While the overall framework may be too broad for some purposes, researchers can select a single model (e.g. the integrated model) from the overall framework to analyze particular fields of interest. Even more, individual components of each model can be removed or exchanged for new elements. For instance, the integrated model comprises a number of strategic variables that have been appraised as most important in the explanation of firm performance. However, new factors might occur over time, while the relative importance of others will be further 140

questioned. The modular character of the model allows these changes to be comprehended. Finally, the end of this chapter does not terminate the further development of our research framework. By applying the framework in empirical research, new insights will be gained that will contribute to further improvement of the framework. The demarcation line between formal and appreciative theorizing will, for example, be pushed further outwards, as additional evidence and theorizing become available. The ultimate value of our research framework depends significantly on its applicability and usefulness for empirical research. Having established the theoretical ground in the past two chapters, we thus continue to present the research methodology in the following chapter.

141

IV.

Research Methodology

In this chapter, we describe the empirical study conducted in order to validate the proposed theoretical models. The chapter is broken down into four main sections: (1) Research Design, (2) Questionnaire Development, (3) Survey Implementation and (4) Data Analysis. The first section presents the overall plan for the empirical study,

establishing guidelines for data collection and analysis. The second section summarizes the process of questionnaire development, including pretesting procedures and the operationalization and measurement of the main constructs. Section three describes the survey implementation process. The final section provides an overview of the statistical methods applied in hypothesis testing, as well as an evaluation of the reliability and validity of the constructs operationalized in this study.

IV.1

Research Design

The initial section presents the fundamental elements of our empirical study: What research methods have been selected? How did we obtain the required data? What sample of firms did we select? What is the unit of analysis? Who were the main informants?

Research Setting

The methodology to be employed in this study is designed to be predominantly descriptive rather than exploratory in nature. Based on a partial test of existing theory, it

is expected that the relationships hypothesized in our theoretical models will be confirmed. Descriptive Studies can be classified into two basic types: (I) cross-sectional and (II) longitudinal (Churchill 1999). The cross-sectional study typically involves a number of variables measured once for a sample from the population of interest. In other words, it provides a snapshot of the variables of interest at a single point in time. Longitudinal studies, on the other hand, measure variables for a fixed sample repeated

143

over time. They provide a series of pictures that, when pieced together, provide a movie of the changes accruing over time. In accordance with the fundamental research position of this study, we apply a "multilevel research method" combining both cross-sectional and longitudinal methods. 13 We first utilize cross-sectional methods to test the Integrated Model and the Complex Model in our overall research framework. Cross-sectional methods are the predominate

mode of analysis in empirical strategy research (Bowen / Wiersema 1999) as they provide stable results in terms of relationships among dependent and independent variables (Langley 1999). While strong in realism, cross-sectional studies fail to capture the dynamic development of variables over time (Miller / Friesen 1982a). In order to capture this dynamic component, we then apply longitudinal methods to the Evolutionary Model in our research framework. As far as the latter model, we depart somewhat from

our descriptive approach and allow for some exploratory analysis. 14

Data-Collection Method

Both cross-sectional and longitudinal analyses rely on cross-sectional data. 15 The required data can be collected from primary or secondary sources (Churchill 1999). The researcher collects primary data specifically for the study at hand. Secondary data, in contrast, has been collected for some other purpose and is reused for the immediate analysis. Due to time and cost economies, researchers should refer to currently available secondary data as far as available (Churchill 1999). For instance, most studies applying longitudinal methods use secondary data sources, such as Compustat databases, to obtain firm-level data over a number of years. Unfortunately, the available secondary sources fail to provide the breadth of data we consider necessary to capture the inherent complexity of our model. For example, reliable information on firm capabilities cannot be found in databases or other publicly available sources. We thus revert to primary data acquired in a field study.

13 14

15

Please refer to chapter 111.3 for a detailed explanation and justification of the fundamental research position of this study. Further details on the statistical methods applied can be found in Chapter IV.4. Our longitudinal analysis relies on panel data that combine cross~sectional and time series information.

144

Primary data can be acquired through communication or observation (Churchill 1999). The recommended method in the case of descriptive research is communication through the direct questioning of respondents. In order to test specific hypotheses, a structured and standardized questionnaire is considered the most productive. The survey methodology facilitates the generalization from a representative sample of companies within an industry to the entire industry (Creswell 1994}. It further allows all required data for the numerous variables of the Integrated Model and the Complex Model to be collected. However, the field study does have some limitations in respect of the Evolutionary Model. Key informants in the participating companies can only provide reliable information for a limited period of time. We considered it safe to ask for information on a period of four years: The two previous years, the current year and the following year. The recollection of developments further in the past may have faded, or may have become unreliable over time. Taking a significantly longer period into account would thus have required a separate, additional method of data collection. Due to the already significantly restricted time available for our empirical study, we refrained from the additional effort. Despite this limitation, we are confident that the reduced observation period will allow for valuable findings. The analyzed period had been particularly dynamic, and included the peak of a long boom period, the following downturn, a phase of stagnation and first signs of a moderate rebound.

Sampling Frame

Another objective of the empirical part of this study is to test the proposed model in the context of the dynamic global media industry. The media industry was selected for a number of reasons. First, numerous recent media company failures around the world have shown that the simple "success formulas" of the past (such as size or diversification) are insufficient to explain performance. More sophisticated analysis is required. Second, the media industry is inherently complex and dynamiC (Albarran I Moellinger 2002). The converging nature of different market segments, the fast technological changes and the growing internationalization will further enhance this complexity. Third, very few quantitative empirical studies have been conducted on the 145

media industry. The few existing studies are mostly restricted to the analysis of industry concentration (Wirth / Bloch 1995). Finally, the writer has spent several years in the media industry which provides a good foundation from which to conduct this study. In our survey, the global media industry is represented by a sample of media companies. It is common in scientific research to take a sample (or selection) of the total population and to infer from the results of the subset to the larger group (Churchill 1999). Statistical inference requires a probability sample where the sample elements are objectively selected through a specific process and not according to the researcher's whims. Probability sampling generally starts from a list of all population elements to draw the sample. However, there is currently no database containing all media companies on a global basis. If, as in our case, detailed lists are unavailable (or their construction is too costly), the area sampling method offers the researcher the best alternative (Churchill 1999). An area sample is a probability sample in which the population is divided into mutually exclusive and exhaustive subsets and from these a subset is selected. The advantage is that the researcher simply needs the list of population elements for the subset. We thus continue to cluster an area sample from the overall global media industry population along the following steps: 1.

To be included in the area sample, a company has to gain there is largest share of revenues in one of the following four industry segments, as represented by the respective SIC codes: (1) Broadcasting & Cable (SIC-Codes 4832, 4833, & 4841), (2) Movie Production & Distribution (7812 to 7841), (3) Print & Publishing (2711, 2721,2731,2741, and 5942), and (4) Music Production & Distribution (3652 and 5735).

2.

To be included, a company's headquarters has to be located in one of the following three global media markets: (1) United States, (2) Germany, and (3) United Kingdom. These three countries represent a large share of the global media market. All other markets have been excluded, reducing the overall scope to a manageable sample.

146

3.

In addition, a company's media businesses had to realize revenues of at least US$ 50 million (as of FY 2002) to be included. Smaller businesses have been excluded for a number of reasons (including the little data that is available).

Following the three steps described above, an area sample of 671 media companies has been created (see Figure 10 for details).

By Segment

By Country

Broadcasting & Cable 25% Print & Publishing 53%

Motion United Kingdom 25%

Figure 10 Sample Distribution

Data from various sources 16 has been used and cross-checked in order to minimize errors in the generation of the sample. In accordance with the one-stage area sampling method, we will study each element within the selected subset (or the full population of the selected subset) without further sampling (Churchill 1999).

16

The full company list has been compiled from three databases (Thompson Financial, Onesource, and Hoover's Online). In an extensive and time-consuming effort, the resulting list has been double-checked against other sources, including industry rankings, industry association listings, and industry-specific databases.

147

Unit of Analysis

Most competitive interaction occurs at the business unit level (Porter 1980). The vast majority of research into the determinants of performance has thus been done at the SBU level. Variance decomposition studies confirm the dominant influence of business unit factors on performance as they have been found to have a far larger share than industry or corporate level factors (Rumelt 1991, Roquebert et al 1996, McGahan / Porter 1997). The primary unit of analysis used in this study is thus the Strategic Business Unit (SBU) with the participating companies also being contacted at the SBU level. The largest part of our questionnaire, including the sections on competitive strategy, firm capabilities, and performance, is targeted at the SBU level. However, a smaller part of the questionnaire contains inquiries relevant to both the corporate and industry levels.

Target Informants

A key informant survey research strategy has been employed in this study (Campbell 1955). While respondents provide information about themselves as individuals (e.g. on their job satisfaction), participants acting as informants report their perceptions and judgments about particular organizational activities and properties (Anderson 1987). In other words, informants provide information at the firm or business unit level. Studying managerial perceptions, rather than the objective characteristics of the organization, is generally accepted as an appropriate approach when studying aspects of firm strategy (Pettigrew 1973, Lyles / Mitroff 1980, Huber / Power 1985). It has been claimed that the key informant should be most knowledgeable as well as willing and able to communicate on the issues of interest (Campbell 1955, Huber / Power 1985). In this study, the key informants were members of the senior management of the SBU. The subjects consisted of three groups: (1) members of the board (mainly CEO or CFO), (2) heads of marketing (VP, EVP, Director or President), and (3) heads of strategy or corporate development (VP, EVP, Director or President). We considered these three groups as most knowledgeable on the broad set of issues covered in our questionnaire. 148

A single key informant was selected for each SBU. The choice between the three groups of informants was made according to (1) the availability of the respective contact information and (2) the size of the organization. In smaller organizations, we preferred board members as informants, but in larger organizations heads of marketing or strategy. This choice reflects the quest for a high willingness and ability among target informants to respond to our enquiry. The contact data was compiled from various sources, including databases and company homepages. The data has been verified and updated through phone calls to the targeted companies. Philips (1981) criticized the key informants survey strategy in his study by proving that various key informant biases reduce the reliability and validity of the obtained information. In response to Philips, Huber and Power (1985) developed a procedure to alleviate potential informant biases. We have applied the most relevant of these measures in this study in order to minimize respondent bias. The first set of measures pertains to the risk of choosing an inappropriate key informant. As described above, we only selected informants occupying a senior management position and with the necessary knowledge to respond to our questions. In addition, we included two screening and four experiential knowledge questions to test and ensure key informant competence. The questions measure the informant's experience in the company and industry, as well as his involvement in strategiC activities within the SBU. 17 The second set of measures regards the risk of key informants' low motivation to respond to the survey. We followed Huber and Power's recommendations to (1) stress the benefit of the study in the cover letter, (2) promise confidentiality and anonymity, and (3) specify the time required to fill out the questionnaire. In addition, we offered to share the results with all the partiCipants. The third set of measures regards the quality of our questionnaire. In order to reduce any risk of misinterpretations by the informants, we used structured questions and conducted a detailed pretest. A question that measured the informant's degree of confidence in answering the survey questions was additionally included.

149

IV.2

Questionnaire Development

This section describes how the questionnaire was developed. First, the pretesting procedure is summarized, followed by the operationalization and measurement of the main constructs of the three models. Seven categories of variables have been taken into account: (1) industry structure, (2) corporate strategy, (3) competitive strategy, (4) firm capabilities, (5) performance, (6) longitudinal items, and (7) informant competency items. The first five categories are later used to test the Integrated and the Complex Model in our framework. The sixth category allows for research into the Evolutionary Model of the framework. Finally, the informant competency category is used to verify the quality of the data collected. For each category, we utilized measures or scales derived from extant literature to capture the respective constructs. In general, well-validated measures that had been reported and tested in previous research were used. With few exceptions, all the constructs in the model were measured with multi-item scales. Most of the variables have also been measured with a seven-point Likert-type scale. Exceptions are explained and justified below.

Pretesting

The purpose of the pretest is to reduce the ambiguity and bias of the questionnaire, or the measurement instruments applied (Churchill 1999). Two pretest phases were conducted with different groups of participants. In the first pretest, a group of academics, experts in strategy or marketing, had been asked to review the questionnaire for inconsistencies. In a personal interview, they had provided valuable comments that were used to refine the questionnaire. Following the first pretest, the questionnaire was sent to fifteen executives in the media industry. The informants had been selected as being personal contacts of the author, but they nevertheless closely matched the profile of the survey's intended final target. Eleven executives responded fully to the questionnaire and were interviewed in a follow-up phone conversation. Their detailed feedback was used to further refine and improve the questionnaire. Most changes concerned the

17

We did not respond to the suggestion to use multiple informants at different levels in the organization due to resource and time constraints.

150

wording, sequence and structure of the questionnaire. Some parts were rearranged or reworded to make them clearer and easier to complete. The final survey questionnaire can be found in Appendix A. It consists of a cover page and five and a half pages of questions. The cover page contains the title and a short description of the study, the seal of the University of Geneva, and our contact details. The questionnaire consists of six sections: Questions on (1) industry structure, (2) corporate strategy, (3) competitive strategy, (4) firm capabilities, (5) longitudinal developments, and (6) performance & informant details. The measures and scales used for these sections are described below:

Industry Structure Measures

Three variables are used to represent industry structure: (I) Industry Concentration, (II) Market Growth Rate, and (III) Barriers to Entry. Bain (1956) developed two different classification systems for the level of industry concentration. The first classification system was based on the eight-firm concentration ratio (CR8), while his second was based on a combination of eight- and four-firm concentration ratio (CR8 / CR4). The majority of the following research in the field of strategic management utilized the four-firm concentration ratio exclusively (Robinson / McDougall 1998, 1088). Industry concentration is thus measured with the following question: 1.

Please estimate the market share currently held by the four largest firms in your industry (in %):

%

The second variable, market growth rate, is measured with the following question (adapted from Narver / Slater 1990): 2.

Please estimate the average annual growth rate of the primary market served by your SBU over the past three years (in %):

151

%

The third variable, barriers to entry, has most often been represented by three issues: economies of scale, capital requirements, and product differentiation (Harrigan 1981, Chappell et al 1983, Rosenbaum / Lamort 1992). All three issues have been measured either by empirical proxies based on archival data (e.g. Harrigan 1981) or by judgments based on expert assessment (e.g. Mann 1966, Weiss 1971). We followed the latter approach and asked the informants to assess entry barriers using a seven-point scale anchored on either end of the spectrum by "not at all" and "a great deal". The three items are represented by the following questions: 1.

Economies of scale play an important role. Market leaders enjoy significantly reduced unit costs due to higher sales volumes.

2.

Our business requires significant capital investment. This represents a considerable challenge to new firms willing to enter the market.

3.

Competing products & services in our industry are highly differentiated. Firms invest strongly in advertising to point out the differences.

Corporate Strategy Measures

Three variables were selected to represent the corporate strategy construct: (I) Product Diversification, (II) International Diversification, and (III) Vertical Integration. Product Diversification: Ramanujam and Varadarajan note: "the extent of energy

devoted to developing measures of diversity is impressive" (1989, 538). Generally, it is possible to distinguish between measures of the degree or extent (i.e. less or more) and the direction or strategy (i.e. related or unrelated) of diversification (Datta et al 1991). The degree of diversification has been operationalized as either a continuous measure or business count approach. Continuous measures have been frequently criticized for their shortcomings in examining curvilinear relationships (Pitts / Hopkins 1982, Datta et al 1991). We thus apply a business count approach to measure the extent of diversification. Some business count approaches rely on a simple counting of the

152

number of businesses (numerical count). These approaches are insensitive to the distribution of revenues to different businesses. Other approaches are focused on the share of the firm's largest business (Rhoades 1974, Rumelt 1974). However, the latter ignores to what extent the remainder of the firm's activities might be diversified. The shortcomings of both approaches can be overcome by an approach applying both independent diversity criteria - the number of individual products and the size of the largest business (Carter 1977, Pitts 1977, Pitts I Hopkins 1982). The number of businesses is measured relative to the major competitors on a seven-point scale anchored on either end of the spectrum by "very few" and "quite a lot":

1.

How would you rate the number of businesses your company operates compared to your main competitors? (i.e. radio, television, cable, books, newspapers, periodicals, printing, movie production, music, retail, services)

The relative size of the largest business is measured with the following question:

2.

What percentage of overall corporate sales does the most important business of your company contribute? 100%

95-99%

70-94%

50-69%

30-49%

10-29%

<10%

While the two questions developed above capture the extent of diversification, they fail to uncover the "under/ying logic" (Rumelt 1974, 54) or relatedness of diversification. In the past, researchers have often used subjective classifications of relatedness, including the most popular categorization by Rumelt (1974, 1982). Despite being widely used, categorical measures of relatedness have been strongly criticized. Prahalad and Bettis (1986) stressed the importance of manager's perception of relatedness. Stimpert and Duhaime (1997) developed a four-item scale of product-market relatedness based on manager perceptions. 18 Based on this concept, we measure relatedness with the following four questions using a seven-point scale anchored on "not at all" and "a great deal":

18

The authors present a multidimensional construct that combines product-market and differentiation relatedness. We selected the first concept as it was highly correlated to other measures of diversification (Herfindahl, Entropy). The latter concept has not been used due to the missing evidence in prior studies and the manufacturing-specific character of this approach.

153

1.

The different businesses in our company share the same customers

2.

The different businesses in our company need the same resources or materials

3.

The different businesses in our company share the same production processes

4.

The different businesses in our company share distribution networks

International Diversification: Studies on international geographic diversification have

focused on aggregate measures, such as the percentage of firms' sales accounted for by foreign sales (Le. Geringer et al 1989). These one-dimensional measures have been criticized for their failure to capture the extent of diversification across foreign countries. However, multidimensional measures have gained little support to date (Hitt et al 1997, 779). Grant (1987) found regional differences and the direction of mUltinational expansion to be insignificant. In line with previous studies, we thus apply the percentage of foreign sales as an aggregate measure of international diversification:

1.

Please estimate the percentage of corporate sales achieved through foreign operations: _ _ _ _ _ %

Vertical Integration: The most commonly used measures of vertical integration are

Adelman's (1955) "value-added to sales index" (VA/S) and Maddigan's (1981) alternative proxy "vertical industry connections" (VIC). These measures produced inconsistent results and have been criticized as oversimplified (Maddigan / Zaima 1985, Harrigan 1986, D'Aveni / Ravenscraft 1994). Harrigan (1986) suggests a multidimensional measure of vertical integration that combines four items: degree of forward integration, degree of backward integration, stages of integration, and breadth of vertical integration. 19 The four items are represented by the following questions as measured on a seven-point scale anchored on "not at all" and "a great deal":

1.

Products and services of our business units are sold internally to other business units.

2.

Input required by our business units are purchased internally from other business units

19

Harrigan includes a fifth item, "ownership form" that is specific to ventures and has been removed.

154

3.

Our company covers several steps of the industry value chain (e.g. Your SBU operates a TV channel, and is also part of a company that is active in TV content production or TV distribution via satellite or cable).

4.

Our company offers a broad range of complementary products and services (e.g. your company publishes a travel magazine and offers in addition travel packages, an online travel agency, or travel insurances).

Competitive Strategy Measures

Competitive Strategy is represented by four generic strategies: (I) Differentiation Strategy, (II) Cost Leadership Strategy, (iii) Focus Strategy, and (IV) Balancer Strategy. Recent research (Kumar et al 1998) has provided conceptual and empirical support for Narver and Slater's (1990) scales to measure Porter's (1980) generic strategies. The scales were based on prior research, including that of Hambrick (1983) and Dess / Davis (1984) and not only found sufficiently reliable, but the authors also provided evidence of their reliability and validity (Narver / Slater 1990). Miller's (1986) scales too have been frequently used in subsequent empirical studies (Le. Miller 1988, Lee / Miller 1996). We combined both established measures and adapted the wording slightly to the targeted industries. Differentiation strategy is measured applying a six-item, seven-point scale, anchored

on "not at all" and "a great deal". The first five items are taken from Narver and Slater (1990) and Kumar et al (1998), while item six is adapted from Miller (1986). Respondents are asked to indicate the extent to which their SBU is concentrated on the following activities: 1.

Introducing new services and products

2.

Differentiating products from competitors'

3.

Offering a broader range of services and products than competitors

4.

Utilizing market research to identify new services

5.

Providing a higher service or product quality than competitors

6.

Advertising products & services and developing a brand image

155

To measure the strategy of overall cost leadership, respondents were asked to indicate on a seven-point scale, anchored on "not at all" and "a great deal", the extent to which their company is focused on the following six activities: 1.

Achieving lower cost of service than competitors

2.

Making services and procedures cost efficient

3.

Improving the time and cost required for coordination of various services

4.

Improving the utilization of available equipment, capacity, and facilities

5.

Performing cost analysis and controlling

6.

Offering lower prices for our products

Items one through five were taken from Kumar et al (1998); item six was taken from Miller (1986). Next, a focus strategy is represented by a three-item scale adapted from Miller (1986). Respondents were asked to indicate on a seven-point scale, anchored on "not at all" and "a great deal", the extent to which their company is focused on the following activities: 1.

Achieving a broader line of products and services than competitors

2.

Achieving a higher breadth of different customer types

3.

Increasing our geographic coverage

Finally, a hybrid or balancer strategy is measured indirectly. Following Wright et al (1991) and Kumar et al (1998), companies that obtain high results in both cost leadership and differentiation are labeled as "balancers".

Firm Capabilities Measures

The fourth construct, firm capabilities, is represented by five variables: (I) Market Orientation, (II) Organizational Learning, (III) Innovation, (IV) Reputation, and (V) Organizational Culture. Market Orientation: Deshpande et al (1993) incorporated prior measures by Kohli and

Jaworski (1990) and Narver and Slater (1990) into a nine-item scale with which to

156

measure market orientation. The scale has been successfully applied in a number of empirical studies (i.e. Deshpande / Farley 1999). Respondents are asked to indicate the extent of their agreement with the following statements as applicable to their business (anchored on "strongly disagree" and "strongly agree"):

1. 2.

We regularly measure customer service. Our product and service development is based on good market and customer information.

3.

We have deep knowledge of our competitors.

4.

We have a good sense of how our customers value our products and services.

5.

We are more customer focused than our competitors.

6.

We compete primarily based on product or service differentiation.

7.

We are strongly focused on our customers' interest

8.

Our products and services are the best in the business.

9.

Customer service is extremely important in our business.

Organizational Learning: The items that make up the learning orientation construct attempt to capture the commitment to learning, shared vision, open-mindedness, and intra-organizational knowledge sharing within the business unit. From a 17-item scale with which to capture learning orientation as developed by Calantone et al (2002), eight items were selected on the basis of their item-to-total correlations. The following items are measured on a seven-point scale anchored on 1 "strongly disagree" to 7 "strongly agree":

1.

Managers agree that our organization's ability to learn is the key to our competitive advantage.

2.

Our SBU sees leaming as a key commodity necessary to guarantee organizational survival.

3.

Employees view themselves as partners in charting the direction of our organization.

4.

Employees in our SBU are committed to the goals of this organization.

5.

We rarely question our own bias about the way we interpret customer information.

6.

Personnel in this SBU realize that the very way they perceive the marketplace must be continually questioned. 157

7.

8.

We have specific mechanisms for sharing lessons learned between SBUs.

We always analyze unsuccessful activities and communicate the lessons learned widely.

Innovation: Firm innovativeness is measured with a six-item scale developed by Calantone et al (2002). The scale integrates prior research and is more detailed than alternative approaches (Le. Deshpande I Farley 1998, Hurley I Hult 1998). Each of the six items is measured with a seven-point Likert-type scale, ranging from 1 "strongly disagree" to 7 "strongly agree,,20:

1.

Our SBU frequently tries out new ideas.

2.

Our SBU seeks out new ways to do things.

3.

Our SBU is creative in its methods of operation.

4.

Our SBU is often the first to market with new products and services.

5.

Our management actively seeks and encourages innovation.

6.

Our new product and service introduction has increased over the last five years.

Reputation: Most empirical studies rely on the measures applied by Fortune Magazine in its annual study on corporate reputation (Le. Fombrun I Shanley 1990, Roberts I Dowling 2002). The Fortune survey captures reputation through an eight-item scale. Respondents are asked to indicate the extent of their agreement with the following statements (anchored on "strongly disagree" and "strongly agree"):

1.

Compared to our major competitors, the quality of management in our SBU is superior.

2.

Our SBU is known for the quality of its products and services.

3.

Our SBU is known for its innovative products.

4.

Our company is perceived as an excellent long-term investment by its shareholders.

5.

Compared to our major competitors, our SBU is in a sound financial position.

6.

Our SBU has the ability to attract and keep talented people.

158

7.

Our SBU is known for its well formulated and executed strategies.

8.

Our company takes on its community and environmental responsibility.

Organizational Culture: The organizational culture scale used in this study had been developed by Deshpande et al (1993) and based on prior research by Quinn (1988). The scale was tested and verified in later studies (Le. Deshpande / Farley 1999). Other recent scales (Le. Ogbonna / Harris 2000) are almost similar to the scale used by Deshpande and colleagues. The scale includes four items and each of these items contains four descriptions or organizations. For each of the items, respondents are asked to distribute 100 pOints among the four descriptions depending on how similar the description is to their business 21

1.

Kind of Organization (Please distribute 100 pOints): (A)

2.

(8)

My organization is very dynamic and entrepreneurial.

(C)

My organization is very formalized with lots of established procedures.

(D)

My organization is very work-oriented. The focus is on getting the job done.

Leadership (Please distribute 100 points): (A)

3.

20 21

My organization is very personal. People share a lot of themselves.

Leaders in my organization are considered mentors.

(8)

Leaders in my organization are considered entrepreneurs or innovators.

(C)

Leaders in my organization are considered coordinators or organizers.

(D)

Leaders in my organization are considered producers or technicians.

What holds your Organization together? (Please distribute 100 points): (A)

Loyalty, tradition, and commitment hold my firm together.

(8)

Commitment to innovation and development hold my firm together.

(C)

Formal rules and policies hold my firm together.

(D)

The emphasis on task and goal accomplishment holds my firm together.

Item number five has been replaced through an item developed by Hurley and Hult (1998). This item has received a particularly high factor loading and is more appropriate to the industry under analysis. The wording of the descriptions has been slightly edited.

159

4.

What is important? (Please distribute 100 points):

(A)

My organization emphasizes human resources and employee morale.

(B)

My organization emphasizes growth and acquiring new resources.

(C)

My organization emphasizes permanence and stability.

(0)

My organization emphasizes competitive actions and goal achievement.

The four resulting scores for the different types of organizational culture are computed by adding the four values allocated to each item.

Performance Measures

Performance is measured with self-evaluations of the (I) Profitability, (II) Market Share, and (III) Sales Growth of a company when compared to the largest competitors in the same market. The measurements and scales have been taken from two often used sets of subjective measures developed by Oess / Robinson (1984) and Buzzell / Gale (1987). For several reasons a subjective measure of performance was preferred to objective data. First, a large share of the targeted media companies is privately owned and is often very reluctant to provide financial data. The alternative, to exclude all privately owned firms from our analysis, would have severely biased the results. Second, objective financial data on the private firms or individual strategic business units of large companies was not publicly available, making the accuracy of any reported financial performance figures impossible to check. Third, subjective performance measures have been used in numerous studies on strategic management and marketing. They are generally accepted as reasonable proxies in the absence of objective measures (Oess / Robinson 1984). The obtained data has been crosschecked as far as possible with secondary sources of data, such as Thomson Financial, Onesource, and Yahoo Finance. The respondents were specifically asked to compare the performance of their SBU against the three largest competitors and assess their performance along the following dimensions on a seven-point scale with the anchors "very low" and "very high":

160

1.

The profitability of the SBU 22

2.

The market share of the SBU

3.

The sales growth of the SBU

Longitudinal Model Measures

The longitudinal development will be researched over a rather short, but particularly troubled period for the media industry. More exactly, we ask respondents to provide information on how key environmental, strategic, and organizational factors developed over the previous two years (2001 and 2002), the present year (2003), and how they expected them to develop further (2004). The analysis thus spans past, present, and future developments of the industry. We did not ask for hard facts, but relied entirely on managers' perception of change. Respondents were asked to indicate how the individual factors would develop in each year in comparison to the previous business year. The direction of change had to be indicated on a scale anchored on 1 (significant reduction) and 5 (significant increase). The individual factors are determined by the following requests to indicate the: 1. 2.

Development of the market share held by the four largest firms in your industry. Development of the annual industry growth rate of the primary market targeted by yourSBU.

3.

Development of the risk of new competitors entering the market.

4.

Development of your company's product diversification (breadth of overall product range).

5.

Development of your company's international business (% of sales achieved abroad).

6.

Development of your company's vertical integration (along the value chain).

7.

Focus of your SBU on product development & differentiation, as well as advertising.

8.

22

Focus of your SBU on cost reduction and efficiency. Respondents are asked to consider different measures of profitability such as return on assets (ROA). return on investment (ROI). return on sales (ROS) or return on net assets as equivalent (in line with Narver and Slater 1990 that applied the same approach).

161

9.

Focus of your SBU on a few core products, customers, and markets.

10.

Focus of your SBU on customers and competitors.

11.

Focus of your SBU on its reputation, service excellence, and attractiveness to both financial and labor markets.

12.

Strength of your organization's particular corporate culture.

13.

Importance

your

SBU

assigns

to

learning,

knowledge

sharing,

and

communication. 14.

Development of your SBU's innovativeness.

15.

Development of your SBU's profitability.

16.

Development of your SBU's market share.

17.

Development of your SBU's sales growth.

Informant Competency

The following questions are used to assess the informant's competency to respond to the questionnaire and thus judge the quality of the collected data: 1.

What is your current position?

2.

How long have you worked in the media industries?

3.

How long have you worked for this company?

4.

How long have you been deployed at your current SBU?

5.

How involved are you in strategic activities in your SBU?

6.

How comprehensible was this questionnaire?

IV.3

Survey Implementation

The mail survey was conducted during the months of September, October, and November 2003. The surveys were addressed to the selected area sample of 671 key informants each representing a single SBU. The mailings to all three target countries were completed within a short period. Following the mailings, 250 non-respondents (roughly one third) were randomly chosen and contacted over by phone. Almost all the respondents were contacted and were asked to participate. The main reasons supplied

162

for not participating in the study were a lack of time and corporate policies of not participating in surveys. Former studies have shown that due to heavy demand on their time, senior managers are less likely to respond to mail surveys (Huber / Power 1985). In order to facilitate response, we utilized a number of recommended techniques (Huber / Power 1985, Dillman 1991): 1.

Used a cover letter and envelope with the seal of the University of Geneva and provided a postage-paid return envelope. Emphasized the academic nature of the study.

2.

Stressed the benefits of the study to the key informants and promised a detailed summary of the results. Acknowledged the informants' distinctive contribution to the success and quality of the study.

3.

Ensured strict confidentiality and anonymity.

4.

Specified the time it would take to complete the questionnaire (ca. 20 minutes).

Response Rate: In total, we received 112 questionnaires. Eight questionnaires were

excluded from analysis due to either incomplete data or insufficient competency as indicated by the informant. The number of questionnaires finally included in the analysis thus amounts to 104. The response rate of this study was calculated as follows (Churchill 1999): Response Rate

= Total

Surveys Returned / (Surveys Sent Out - Surveys

Returned Unopened) x 100. The response rate thus amounted to 16.3% (104/ (671-31) x 100). While response rates of comparably complex surveys vary widely (roughly between 4% and 30%), the majority achieved similar results. The sample size of 104 respondents is particularly significant if we consider that it represents 16.3% of the full population of the media industry segments under consideration.

The distribution of the responses across the different industry segments and countries was largely in line with the targeted overall population. In terms of segment distribution, the Broadcasting & Cable segment was slightly overrepresented, whereas the Motion Pictures segment was somewhat underrepresented. This is due to the fact that some 163

respondents initially classified in the Motion Pictures segment (in particular TV content producers) considered themselves as belonging to Broadcasting & Cable. If this particular exception is taken into account, the segment distribution of the sample matches the distribution of the overall population almost exactly. In terms of distribution by country, a slight overrepresentation of German companies has been found. This can be explained by the University of Geneva and the researchers being better recognized in this country. In the following section it becomes apparent that this overrepresentation did not lead to bias in the overall data.

By Segment

By Country

Print & Publishing 50%

Figure 11

Response Distribution

Non-response Bias: A key concern in mail survey research is the potential presence of

non-response bias. This occurs if significant differences are found between respondents and non-respondents. In such a case, the results of the study are sample specific and do not allow for generalization of the overall population (Churchill 1999). Following the procedures for testing non-response bias described by Armstrong and Overton (1977), we subdivided the sample into first wave (approximately 60% of the sample) and second wave respondents (approximately 40% of the sample). The first SUb-sample mostly contains responses from Germany and the United Kingdom, whereas the second sub-sample mainly consists of responses from the United States. A 164

MANOVA test on the means of the different variables in our study revealed no significant differences between the two groups. An additional chi-square test of covariance matrices of the different variables indicated no significant differences between the two groups. Similar results were achieved on a second test splitting the sample into industry segments. Our tests led us to conclude that our study has no significant non-response bias. Key Informant Competence: Another area of concern is the competence of the key informants used in the survey. As the data on each SBU is acquired from a single respondent. the data quality is highly dependent on the competence of this informant. In our survey we tested key informant competence through a number of questions. First of all. we asked respondents to specify their current position. Almost all participants are part of the senior management team of their SBU. Almost 90% are either members of the Management Board or heads of the Marketing or Strategy departments of their SBU (see Figure 12 for details). In keeping with their position. the participants are highly competent to respond to our questionnaire.

Position

Tenure

SBU

Member of the Board

34%

~nY"IIII"IIII" Media Industry

Marketing

10

34%

Figure 12

15

20

Years

Informant Competence

In addition to their position. we asked all informants for their tenure within the SBU. company. and the media industry. Tenure has been suggested as a proxy for the degree 165

of knowledge on issues concerning the respective company and industry (Seidler 1974, Phillips 1981). On average, the respondents have worked in the media industry for sixteen years. An average experience of eleven years at their current company and of eight years within their respective SBU provides further affirmation of informant competence. Finally, we asked respondents to rate the extent of their involvement in strategic decisions and the confidence they felt in responding to the survey. The average involvement in strategic decisions was rated at 6.1 (on a seven-point scale). The comprehensibility of the survey was 5.5 on a seven-point scale, which is surprisingly high for a questionnaire of significant complexity. Overall, the different measures of respondent competence indicate that all participants are highly able to provide reliable responses to our survey.23

IV.4

Data Analysis

This section is composed of four segments: (I) Reliability, (II) Validity, (III) Statistical Methods, and (IV) Conclusions. In the following, the reliability and validity of the constructs operationalized in the study are verified by a number of statistical tests. Next, we provide an overview of the statistical methods that were applied in the hypothesis testing. We close this section with a few conclusions.

Reliability

A particular challenge for empirical research in social sciences is the necessity to measure constructs that cannot be directly observed. An example is the firm's degree of innovativeness, which we intend to analyze in our study. In order to capture such complex constructs as innovativeness, multiple indicator variables are measured that are collectively expected to represent the underlying construct. In this study, fourteen exogenous constructs and one endogenous construct were operationalized. Of these, 23

Four questionnaires have been excluded from further analysis due to unclear informant competence. In particular, we excluded informants indicating a low overall comprehension of the survey.

166

nine constructs were measured using multiple-item scales. The quality of these multiitem scales is rather doubtful (Fornell I Larcker 1981, Brosius 2002). It is therefore recommended to assess the reliability and validity of these constructs. Reliability analysis has been conducted to verify the quality of the scales and to select the most appropriate items to be retained. Reliability is a measure of internal consistency of a scale. It measures the degree to which measures are free of error and therefore yield consistent results. The most common test for reliability in social sciences is Cronbach's coefficient Alpha (Gerbing I Anderson 1988, Churchill 1999). This measure shows how well the indicators measure each of the constructs. Reliability analysiS based on Cronbach's Alpha was thus conducted for all multiple-item scales using the Statistical Package for the Social Sciences SPSS 11.5. In the literature, Alpha values greater than 0.7 are considered as indicators of sufficient reliability. Values above 0.8 are indicators of strong reliability (Nunnally 1978, Brosius 2002). The individual results of the reliability analysis can be found in Figure 13. Among the industry structure variables, only "Barriers to Entry" was measured applying a multipleitem scale. The reliability of this construct was, at 0.72, the lowest of all variables tested. However, the reliability was sufficient and the scale had been used and validated in multiple prior studies (e.g. Mann 1966, Weiss 1971). "Vertical Integration" was the only multi-item construct among the corporate strategy variables. The construct had a Cronbach Alpha of 0.74 and thus sufficient reliability. All items were retained. The two business strategy variables "Differentiation Strategy" and "Cost Leadership Strategy" had Cronbach Alphas of 0.73 and 0.83 respectively. A single item was removed from both measures in order to improve overall reliability. All five firm capability variables have been measured with multiple-item constructs. All constructs showed high Cronbach Alphas: "Innovativeness" was measured at 0.86, "Market Orientation" at 0.82, "Organizational Learning" at 0.87, "Reputation" at 0.87, and "Organizational Culture" at 0.79. One item was dropped from the "Reputation" and "Organizational Culture" constructs in order to improve overall reliability. 167

In summary, item-to-total correlations greater than 0.4 and Cronbach Alphas greater than 0.7 have been retained for all constructs in the study. It can be concluded that all constructs were reliable at Cronbach Alphas ranging from 0.72 to 0.87.

Construct Barriers of Entry Vertical Integration

Differentiation Strategy

Cost Leadership Strategy

Innovatlveness

Market Orientation

Organizational Learning

Reputation

Corporate

Culture

Figure 13

~em

ENTRBAR1 ENTRBAR2 ENTRBAR3 VERTINT1 VERTINT2 VERTINT3 VERTINT4 DIFSTRA1 DIFSTRA2 DIFSTRA3 DIFSTRA4 DIFSTRA5 COSTLEA1 COSTLEA2 COSTLEA3 COSTLEA4 COSTLEA5 INNOVA1 INNOVA2 INNOVA3 INNOVA4 INNOVA5 INNOVA6 MARKORT1 MARKORT2 MARKORT3 MARKORT4 MARKORT5 MARKORT6 MARKORT? MARKORT8 MARKORT9 ORGLEAR1 ORGLEAR2 ORGLEAR3 ORGLEAR4 ORGLEAR5 ORGLEAR6 ORGLEAR7 ORGLEAR8 REPUTAT1 REPUTAT2 REPUTAT3 REPUTAT4 REPUTAT5 REPUTAT6 REPUTAT7 CORPCUL1 CORPCUL2 CORPCUL3

Item-to-Total

Alpha If ~em Deleted

Alpha

Correlation

0.62 0,55 0,46 0,53 0,64 0,49 0,49 0,42 0,63 0,46 0,53 0,44 0,51 0,59 0,72 0,78 0,52 0,66 0,73 0,71 0,59 0,72 0,54 0,56 0,62 0,46 0.66 0,59 0,43 0,60 0,45 0,41 0,75 0,67 0,68 0,66 0,50 0,62 0,52 0,66 0,67 0,58 0,64 0,70 0,59 0,64 0.61 0.60 0,67 0.61

0.61 0,64 0.67 0,69 0,62 0,71 0,71 0,68 0,62 0,66 0,64 0,67 0,78 0,77 0,74 0,72 0.78 0,84 0,83 0,83 0,85 0,83 0,86 0,80 0,79 0,81 0,79 0,80 0,82 0,80 0,81 0,82 0,85 0,85 0,85 0,86 0,87 0,86 0,87 0,86 0,80 0.81 0,80 0,79 0,81 0,80 0.80 0.66 0.61 0.65

0,72

Reliability Analysis

168

0,74

0,73

0,83

0,86

0,82

0,87

0,87

0,79

Validity

Validity is synonymous with the ability of a construct to accurately measure what it is attempting to measure (Bollen 1989, Churchill 1999). In order to evaluate the validity of the multi-item scales utilized in this study, Confirmatory Factor Analysis (CFA) based on LlSREL 8.3 was performed (Joreskog I Sorbom 2001). As suggested by Gerbing and Anderson (1988), diverse fit indices and normalized residuals were examined. Three types of validity have been considered: (I) content validity, (II) convergent validity, and (III) discriminant validity.

Content Validity: Content validity is a qualitative type validity focusing on the adequacy with which the domain of the construct is captured by the measure (Bollen 1989, Churchill 1999). In this study, content validity has been assured through a detailed review of the respective literature as discussed in Chapter 2. The research domain had been conceptually defined before the related constructs had been established. In addition, all scales used in the empirical study are derived from the literature and have been used in prior research as discussed in Chapter 4.2. Finally, a pretest had been conducted to further clarify the wording of the scales and the general terminology.

Convergent Validity: Convergent validity confirms the degree of correspondence between measures and the underlying construct, typically measured according to their correlation (Bollen 1989). The convergent validity of the constructs used in our study was assessed by the significance of the item's loading on its underlying construct (t-values > 1.96). The unidimensionality of the constructs was assessed through the chi-square statistic and additional goodness-of-fit statistics as described above. The detailed results can be found in Figure 14. The "Barriers to Entry" construct was measured by a three-item construct. All items loaded significantly, indicating convergent validity. The CFA analysis indicates a single dimension, which fits the data well (X 2 =1.61, dl=1, p=0.2). Concerning the "Vertical Integration" construct, the modification index of the CFA indicated that two items were highly correlated and negatively impacted the overall

169

model fit. The model was respecified by deleting one item leading to an excellent model fit (X 2 =O.98, dl=1, p=O.32). The remaining three items loaded significantly (t values > 1.96). Construct

Item

ENTRBAR1 ENTRBAR2 ENTRBAR3 VERTINT1 Vertical VERTlNT3 Integration VERTINT4 DIFSTRA2 Differentiation DIFSTRA3 Strategy DIFSTRA4 DIFSTRA5 Cost Leadership CDSTLEA1 COSTLEA2 Strategy COSTLEA3 COSTLEA5 Innovattveness INNOVA1 INNOVA2 INNOVA3 INNOVA4 INNOVA5 MARKORT2 Market MARKORT4 Orientation MARKORT5 MARKORT6 MARKORT7 MARKORT8 MARKORT9 Organizational ORGLEAR1 ORGLEAR3 Learning ORGLEAR4 ORGLEAR5 ORGLEAR6 ORGLEAR8 REPUTAT1 Reputation REPUTAT2 REPUTAT3 REPUTAT5 REPUTAT6 REPUTAT7 CORPCUL1 Corporate CORPCUL2 Culture CORPCUL3 Barriers of Entry

Figure 14

Lambda (St.)

T·Value

Chl2

DF

P~Value

CFI

IFI

RMR(st)

GFI

0.47 0.28 046 0.43 0.70 073 0.70 0.54 0.70 064 0,43 0,85 0.90 060 0.73 0.87 0.83 0,65 076 0.66 0,80 0,72 0.51 0,72 0,58 043 0.76 0.78 0.78 0.57 0,68 069 0.80 0.63 0.72 0,73 0.67 069 0.81 0,83 0.67

NA 4,47 397 NA 4.83 477 NA 4.26 4.98 474 NA 4.26 4.26 380 NA 8.33 7.95 6,27 736 NA 6,42 6,01 4,44 6.00 5.00 382 NA 7,63 7,62 5,50 6.66 675 NA 6.25 7.36 6,34 7,86 699 NA 7,77 5.81

1.61

1

0.20

0.90

0.92

0.047

0.99

0.98

1

0.32

0.99

0.99

0.033

0.99

4.99

2

0.08

0.96

0.96

0.043

0.98

0.98

2

0.61

1.00

1.00

0.019

0.99

5.45

5

0.36

0.99

0.99

0.025

0.98

23.98

14

0.05

0,94

0,94

0.052

0.94

10,46

9

0,31

0.99

0.99

0,034

0,97

20,41

9

0,02

0,95

0,95

0,052

0.94

1.87

1

0.17

0.99

0,99

0,053

0,99

Convergent Validity Analysis

Next, both strategy constructs resulted in four-item scales. One item was excluded from each construct to improve overall fit. The resulting models indicate Single dimensions for both the "Differentiation Strategy" (x 2 =4.99, dl=2, p=O.08) and the "Cost Leadership Strategy" (X 2 =O.98, dl=2, p=O.61) constructs. All remaining items loaded significantly. Finally, the five constructs for firm capabilities resulted in five-, six- and seven-items scales. One or two items were deleted from each scale to improve overall model fit. The respecified models indicate a single dimension and good to excellent model fit. All items 170

differed significantly from their zero values as indicated by their t-values, thereby providing further support for convergent validity. In detail, the results indicate a particularly good fit for "Innovativeness" (X 2 =5.45, dl=5, p=O.36) and "Organizational Learning" (X 2 =10.46, dl=9, p=O.31), and a reasonable fit for "Market Orientation" (X 2 =23.98, dl=14, p=O.05), "Reputation" (X 2 =20.41, dl=9, p=O.02), and "Corporate

Culture" (x 2 =1.87, dl=1, p=O.17). Discriminant Validity: Discriminant validity requires that a measure of a construct not be too highly correlated with other constructs from which it is supposed to differ (Churchill 1999). Following the procedures outlined by Fornell and Larcker (1981), discriminant validity was tested by comparing the average variance extracted from a construct with the squared correlation between this construct and other constructs in the study. Discriminant validity is upheld if the average variance extracted from each construct exceeds the squared correlation between the constructs. As shown in Figure 15 this requ irement has been met for all the multiple-item constructs employed in this study. Barriers of Entry verage Variance Extracted Barriers of Entry Vertical Integration Differentiation 51""eav Cost leadership

0,50

Different. Strategy

0,57

Cost Leader· ship Strategy

0,29

Innovative·

ness

Mat1l:el

Orientation 0,84

0.46

OrganlzaL Learning

0,99

Reputation

Corporate

Culture 0,83

0.17 (0.41) 0.14 (0.38)

51""""Y

Innovative ness

Market Orientation Organizational Leamina Reputation Corporate

Culture

Perfonnance

Figure 15

Vertical Integration

0.16 (0.40)

0.00 (0.03)

0.08 (0.28)

0.02(0.14)

0.00 (0.03)

0.12 (0.35)

0.21 (0.46)

0.07 (0.26)

0.02 (0.15)

0.08 (0.29)

0.44 (0.66)

0.03(0.16)

0.38 (0.61)

0.01 (0.09)

0.21 (0.46)

0.15 (0.39)

0.08 (0.29)

0.68 (0.82)

0.36 (0.60)

0.01 (0.09)

0.21 (0.46)

0.26 (0.51)

0.05 (0.22)

0.75 (0.87)

0.38 (0.61)

0.53 (0.73)

0.02 (0.13)

0.08 (0.28)

0.07 (0.26)

0.05 (0.22)

0.51 (O.71)

0.15(0.39)

0.41 (0.64)

0.41 (0.64)

0.13(0.36)

0.24 (0.49)

024 (0.49)

0.14 (0.37)

0.29 (0.54)

0.22 (0.47)

0.20 (O.4S)

0.70 (0.B4)

0.19 (0.44)

Discriminant Validity Analysis

On the whole, the subsequent data analysis has shown high reliability of the constructs employed in this study. Furthermore, reasonable to high validity has been ascertained for all constructs.

171

Statistical Methods

Before proceeding with hypotheses testing in the following chapter, we briefly outline the methods applied in the evaluation of the three models. Integrated Model: Given the nature of the integrated model with the presence of latent

constructs, structural equations were selected as the most appropriate methodology. Structural equations modeling (SEM) is a statistical approach to testing hypotheses about relations among observed and latent variables (Hoyle 1995). SEM is a multivariate statistical technique combining (confirmatory) factor analysis and econometric modeling to analyze hypothesized relationships among latent (Le. unobserved or theoretical) variables measured by manifest (Le. observed or empirical) indicators (Diamantopoulos / Siguaw 2000). In this study, the LlSREL 8.3 computer program (Joreskog / Sorbom 2001) is used for SEM analysis. A full LlSREL model is typically composed of two sub-models (Bollen 1989, Kelloway 1998). The measurement model describes how each latent construct is measured or operationalized through corresponding manifest indicators. For example, we previously suggested that innovativeness and reputation might be manifest indicators of the latent variable capabilities. The structural model, on the other hand, describes the relationship between the latent variables themselves and indicates the quantity of unexplained variance (Diamantopoulos / Siguaw 2000). In our study, the structural model explains whether the dependent latent variable performance is determined or influenced by independent latent variables such as capabilities or industry structure. Structural equations modeling combines the simultaneous assessment of the quality of observed variables' measurement and the examination of predictive relationships among latent constructs. Structural equation modeling is inherently a confirmatory technique; it seeks to confirm that the relationships hypothesized are indeed consistent with the empirical data at hand (Kelloway 1998). Confirmation (or lack of) is achieved by comparing the computed covariance matrix implied by the hypothesized model to the actual covariance matrix derived from the empirical data (Diamantopoulos / Siguaw

172

2000). When the hypothesized model is close to the observed data, the model is said to "fit" the data (Kelloway 1998).24 Complex Model: The complex model is nothing more than an advanced version of the

integrated model. The difference is that relationships between manifest variables are taken into consideration. In order to test the complex model, we initially draw on findings from the previous structural equations analysis. SEM is particularly appropriate for framing and answering increasingly complex models (Kelloway 1998, 3). LlSREL models allow for evaluation of the existence and importance of relations between manifest variables. Besides the overall evaluation of the complex model through SEM analysis, we apply mediation and moderation tests (Baron / Kenny 1986) as additional statistical techniques for individual hypothesis testing.

1.

c

x

y

M 2.

x Figure 16

Mediation

In order to explain mediation, let us consider a manifest variable X that is assumed to affect another variable Y (see Figure 16). The effect of the initial variable X on the outcome Y may be mediated by an intervening or mediating variable M. Complete mediation is when variable X no longer effects Y after M has been controlled for (path c·=O). Partial mediation is when the path from X to Y is reduced in absolute size but is 24

Please refer to Chapter V.1 for a detailed introduction to LlSREL (or SEM) modeling.

173

still different from zero when the mediator is controlled (Judd / Kenny 1981, Baron / Kenny 1986). Mediation implies a causal sequence among three variables: The independent variable causes the mediator that in turn causes the dependent variable (Shad ish / Sweeney 1991). A variable functions as a mediator when it fulfills the following four conditions (Baron / Kenny 1986): (1) the initial variable significantly affects the outcome (in the absence of the mediator), (2) the initial variable significantly affects the mediator, (3) the mediator has a significant unique effect on the outcome, and (4) the effect of the initial variable on the outcome shrinks on addition of the mediator to the model. For example, we might find that adopting a differentiation strategy causes higher performance (1). We might continue to detect that firms adopting a differentiation strategy also have a significantly high degree of innovativeness (2). Innovativeness, in turn, might positively affect firm performance (3). If the initial effect of differentiation strategy on performance is significantly reduced through the inclusion of innovativeness to the model, (partial) mediation can be stated (4). In simple terms, this means that the effect of differentiation strategy on performance is partially explained by innovativeness. Mediation effects have to be distinguished from moderation or interaction effects (Baron / Kenny 1986). Moderators affect the direction and / or strength of the relationship between an independent and a dependent variable (Holmbeck 1997). Thereby, the moderator is generally uncorrelated with both the independent and the dependent variable. No causal sequence is implied by interaction. Baron and Kenny (1986, 1176) conclude that "whereas moderator variables specify when certain effects will hold,

mediators speak to how or why such effects occur". When a mediator or moderator model involves latent constructs, SEM provides the most appropriate data analysis strategy (Bollen 1989). However, if it involves only measured variables, the basic analysis approach is an ordinary least square regression analysis (Baron / Kenny 1986). In our study, we attempt to test concrete hypotheses on mediation and moderation effects between measured variables. Consequently, we refer to regression analysis and the SPSS computer program to test for these effects. Multiple regression analysis is the preferred method for testing mediation in the present literature and is considered to provide "meaningful tests of hypotheses" (Holmbeck 1997, 600). 174

Evolutionary Model: A multitude of methods for longitudinal analysis has been introduced in the respective literature. Miller and Friesen (1982b) reviewed existing approaches and speak in favor of those methods that consider multiple variables in a large number of organizations. While less explanatory than, e.g., case studies, they allow results that are replicable and more generally applicable to be derived. These multivariate and quantitative methods require panel data. Panel data, also called longitudinal data or cross-sectional time series data, is data where multiple cases (e.g. companies) were observed during two or more time periods (Stock / Watson 2003). In our study, we collected panel data from a representative sample of media companies surveyed for a four-year time period. There are two kinds of information in cross-sectional time-series data: the crosssectional information reflected in the differences between subjects, and the time-series information reflected in the changes within subjects over time. Panel data regression techniques allow advantage to be taken of these different types of information (Stock / Watson 2003). The use of SPSS 11 allows both types of analyses (Brosius 2002). First, we apply cross-correlations analysis to identify within-subject changes. This technique allows analysis of the development of a company's strategies over the period under consideration. Second, we use ordinary regression techniques for analysis of the interrelations between subjects (the cross-sectional part). For instance, we may test the effect of changes in independent variables on business unit performance. Such effects of by independent variables can be tested against both the performance in the same year and the performance in the following years (a "time lag" of one or two years).

Conclusions

The present chapter reviewed the methodology applied to validate the proposed theoretical models. We went to great lengths in order to establish credibility for the research findings to be presented in the following chapter. The thorough establishment of the target sample, the careful selection and identification of qualified target informants, as well as the accurate questionnaire development and testing contributed to the reliability of the acquired data. Moreover, we obtained a highly satisfactory response to our survey. The collected data covers more than sixteen percent of all media companies

175

in the targeted industry segments. Furthermore, the respondents showed a particularly high competence and indicated excellent comprehension of the survey. Data quality has been further confirmed by the high reliability and validity of all constructs employed in this study. In summary, the various sections of this chapter have established a sound basis for hypothesis testing in the following chapter.

176

V.

Research Findings

The purpose of this chapter is to present and discuss the results of the tests of the hypotheses and the alternative models proposed in this study. The chapter is broken down into four sections. The first three sections outline the findings concerning (V.1) The Integrated Model, (V.2) The Complex Model, and (V.3) The Evolutionary Model. In the

final section, based on the different research findings, the most relevant (V.4) Implications are outlined.

V.1

The Integrated Model

The lead hypothesis of the integrated model claims that a combination of insights from different perspectives of strategic management is required to best explain performance. On a more concrete level, a number of hypotheses have been established on the significant

effect

of individual

determinants,

derived

from

various

theoretical

perspectives, on business unit performance. In order to generate Simple and structured findings, the complex task of the hypothesis testing of the integrated model has been divided into three sequential analyses: (1) Determinant Analysis, (2) Submodel Analysis, and (3) Model Analysis. The section

concludes with a brief (4) Discussion of the Research Findings.

Determinant Analysis

The first step in our effort to test the research propositions of the integrated model is concerned with the analysis of the direct effect of the identified determinants on performance. Thereby, we intend to establish the first evidence on the hypothesized significant relationships

between the different determinants and

performance.

177

business unit

Two analyses were conducted for hypothesis testing. First the hypotheses had been tested in a series of simple regression equations. Then the hypotheses were tested in multiple regressions at the sub-model level (e.g., all industry structure determinants together). While all tests were done using L1SREL 8, the results have been confirmed by additional tests using SPSS 11. Simple Regression Analysis: Before discussing the findings, it is important to remember that simple regression analysis simply measures the direct effect of each determinant on performance. The direct effect is the influence of one variable on another that is unmediated by any other variable in the model. However, simple regression fails to capture the indirect effects of a variable, i.e. mediation or intervening effects. The sum of the direct and indirect effect is called the total effect. Exclusion of the indirect effects might thus lead to a misleading impression of the influence of a determinant on performance (Bollen 1989). In addition, the exclusion of other, relevant, variables may ignore spurious effects. Third factors might be correlated to both the determinant and the outcome and thus be the "true" explanation behind the significant relationship. We consider both indirect and spurious effects in the more complex models later in our analysis. However, with these limitations in mind, we consider simple regressions as a good starting point for our analysis. While measuring the pure direct effect fails to provide a basis for final conclusions, it is nevertheless an important (if not the most important) effect in the greater majority of models. Figure 17 summarizes the results of the simple regression analysis based on L1SREL 8. In general, the results confirm most of the hypothesized relationships. The most noteworthy exceptions are the industry structure variables. Both industry concentration and market growth failed to show a significant effect on performance (t-values of 1.894 and 1.325 respectively). The third industry variable, entry barriers, showed a significant and positive effect on performance. However, the effect is rather weak ((3= 0.24; t= 2.204,

25

R2=

0.06).

25

T-values above t= 1.96 are considered as significant at the p< 0.05 level of significance; the variance explained in the dependent variable by the independent variable.

178

R2

indicates

Variable

13sm

T-value

R2

Industry Concentration

0.204

1.894

0.042

Market Growth

0.144

1.325

0.021

Entry Barriers

0.240

2.204

0.058

0

International Diversification

0.236

2.282

0.056

0

Product Diversification

0.485

5.182

0.236

00

Vertical Integration

0.395

3.645

0.156

00

Differentiation Strategy

0.416

4.070

0.173

00

Cost Leader Strategy

0.378

3.725

0.143

00

Balancer Strategy

0.514

5.374

0.264

00

Innovativeness

0.660

6.899

0.436

00

Market Orientation

0.466

4.652

0.217

00

Organizational Learning

0.425

4.226

0.180

00

Reputation

0.516

4.975

0.267

00

Corporate Culture

0.390

3.617

0.152

00

Figure 17

Significance

Simple Regression Analysis (LiSREL)

Among the corporate strategy variables, both product diversification and vertical integration showed a highly Significant and positive effect on performance (t-values of

5.182 and 3.645 respectively).26 The third corporate strategy variable, international diversification, had a significant and positive, but rather weak impact on performance (13=

0.24; t= 2.282; R2= 0.06), a curvilinear relationship could not be confirmed.

26

The relatedness of product diversification was measured in a separate construct (as discussed). However, it did not reveal any Significant influence on the strength or direction of the relationship between product diversification and performance. This might be partly due to the fact that almost none of the analyzed media companies pursued unrelated diversification.

179

All three competitive strategy variables had a highly significant and positive effect on performance. 27 As claimed by strategy theory, the balancer strategy construct indicated the highest significance (13= 0.51; t= 5.374), followed by the differentiation strategy (13= 0.42; t= 4.070) and the cost leadership strategy (13= 0.38; t= 3.725). Finally, all five variables of firm capabilities showed highly significant and positive relationships with performance. Innovativeness had the highest significance of all the variables in the model (13= 0.66; t= 6.899), followed by reputation (13= 0.52; t= 4.975), market orientation (13= 0.47; t= 4.652), and organizational learning (13= 0.43; t= 4.226). The fifth variable, corporate culture, indicated a slightly lower, but nevertheless highly significant effect on performance (13= 0.39; t= 3.617).28 A subsequent SPSS-based simple regression analysis confirmed the findings of the LlSREL-based analysis (see Figure 18 for the detailed results). The highly significant effect of all five firm capability variables, all three competitive strategy variables, as well as of the two corporate strategy variables, product diversification and vertical integration, has been reaffirmed. Slight differences in the absolute values between those of LlSREL and SPSS did not translate into substantial divergence of the results. Similar to that found in LlSREL, the two industry variables (industry concentration and market growth) failed to indicate a significant relationship with performance in SPSS and the third variable, entry barriers, showed a significant but rather weak relationship (13= 0.23; t= 2.365; Sig.= 0.02). The most relevant difference in the SPSS findings concerns the third corporate strategy variable, international diversification. While indicating a significant but rather weak effect in LlSREL (13= 0.24; t= 2.282), the relationship is no longer significant if tested in SPSS (13= 0.13; t= 1.355). Moreover, the relationship between international diversification and performance is the weakest of all in SPSS.

27 28

The fourth strategy type, focus strategy, was measured and tested in a separate construct. It did not reveal a significant effect on performance. In this study corporate culture is represented by the entrepreneurial culture construct as discussed before. The alternative types of corporate culture (competitive, bureaucratic, and consensual) were measured and tested as separate constructs and did not reveal a significant effect on performance. However, the directions of the relationships have been confirmed as predicted by the theory (positive for competitive cultures, negative for bureaucratic and consensual cultures).

180

Variable

i3STD

T-value

Sig.

Industry Concentration

0.186

1.898

0.06

Market Growth

0.162

1.638

0.11

Entry Barriers

0.230

2.365

0.02

International Diversification Product Diversification Vertical Integration Differentiation Strategy Cost Leader Strategy Balancer Strategy

0.134

1.355

0.18

0.384

4.165

0.00

It1It1

0.356

3.812

0.00

It1It1

0.363

3.895

0.00

It1It1

0.310

3.264

0.00

It1It1

0.439

4.888

0.00

It1It1

Innovative ness

0.589

7.296

0.00

It1It1

Market Orientation Organizational Learning

0.390

4.323

0.00

It1It1

0.354

3.781

0.00

It1It1

Reputation

0.476

5.407

0.00

It1It1

Corporate Culture

0.360

3.861

0.00

It1It1

Figure 18

Significance

It1

Simple Regression Analysis (SPSS)

In summary, the findings from both the LlSREL- and SPSS-based simple regression analyses confirmed all the determinants' hypothesized significant and

positive

relationship with performance, with the notable exceptions of the two industry structure variables, industry concentration and market growth. The third industry variable, entry barriers showed a significant but weak effect. The corporate strategy variable, international diversification, showed inconsistent but generally weak results. Multiple Regression Analysis: As described in the model development section of this study (Chapter IliA), the integrated model excludes interactions between the four main perspectives of the model (industry, corporate strategy, competitive strategy, and

181

capabilities). However, interactions between factors of the same perspective (i.e. innovativeness and market orientation) have to be considered. We thus continue to evaluate the effects of all the determinants within one perspective on performance (i.e. all industry variables). These models capture the correlations between the single factors within a perspective. Contrary to the single regression analysis above, indirect effects will be captured. Multiple regression analysis thus provides a more reliable estimate of the total effect (direct plus indirect effects) of each determinant on performance. 29 The industry perspective comprises three measured variables (industry concentration, market growth, and entry barriers) that are considered to be positively related to the latent construct performance. Performance is in turn measured by three manifest indicators (profitability, market share, and sales growth).3o Figure 19 illustrates the respective LlSREL path model for the industry perspective. 31 As the only latent variable, performance is represented in LlSREL by an ellipse. The three industry variables and the three manifest indicators of performance are measured variables and thus represented by a rectangle. Causal relationships between two variables (i.e. between industry concentration and performance) are indicated by an arrow that moves from the independent to the dependent variable. The direction of the arrow shows the assumed direction of the effect between the two variables. Double-headed arrows (i.e. between industry concentration and market growth) represent expected correlations between variables. The figures for each arrow indicate the J3-values (upper figure) and the t-values (lower figure in brackets) of the relationship. The figures to the right (connected with an arrow to the performance indicators) and below the industry variables represent the errors of measurement. 32 Due to measurement errors and third variable effects, variables

29 30

31 32

However, as stated above, the multiple regression models fail to capture potential indirect and spurious effects caused by interactions with variables from other perspectives. These effects will be considered later by the complex model. The three measured industry variables take the form of formative indicators (they cause or influence the latent variable), whereas the three manifest performance indicators are modeled as reflective indicators (they are caused or influenced by the latent variable). This is illustrated by the direction of the arrows in the model (formative: ~; reflective: f-). LlSREL models are explained in more detail in the following section. The description of the path model in this section is based on Kelloway (1998) and Diamantopoulosi Siguaw (2000). The error terms of the three industry variables have been fixed at 1.00

182

can never be measured without error (Bollen 1989, Diamantopoulos / Siguaw 2000).33 An error term is thus added to all variables in a LlSREL model.

0.26 (2.20) 0.86 (---) 0.42 (3.86) 0.23 (2.27)

0.07 (0.67)

0.74 (5.59)

0.17 (1.45)

0.58 (4.96)

-1 ~

Market Share

Sales Growth

r-

r

0.46 (4.38)

0.67 (6.19)

Chi-Square= 6.98, df= 6, P-value= 0.32, CFI= 0.99, IFI = 0.99, GFI= 0.98, SI. RMR= 0.044

Figure 19

Multiple Regression - Industry Perspective

The results of the multiple regression model of the industry perspective confirm our prior findings from the simple regression analysis. None of the relationships between the three industry variables and performance has been found to be significant. Even the weak effect of entry barriers on performance found in simple regression analysis has been lost (13= 0.17, t-value= 1.45). Two additional findings can be ascertained from the industry model. First, the performance construct seems to be well represented by the three manifest indicators. All three indicators showed high t-values (above 4.96) and the overall model fit is

33

It is highly unlikely that the subjective measures widely used in social sciences (as well as in this study) will exhibit perfect validity or reliability. The inherent measurement error should therefore be considered in the model. In addition, it is simply not realistic that, e.g., performance is solely and fully captured by the three indicators in the model. The extent to which other variables (not included in the model) reflect performance will be unexplained variation in the latter. Thus we need to include a residual term to account for those influences on performance that have not been explicitly modeled (Diamantopoulos I Siguaw 2000, 3).

183

excellent. 34 Next, significant correlations have been found between all three industry variables (t-values between 2.27 and 3.86). This indicates that the three industry variables may be better modeled as indicators of a common underlying construct, industry structure, than directly related to performance. We will do so in the following section, turning the focus towards the analysis of a consistent sub-model for each perspective. The corporate strategy perspective contains three variables (international diversification, product diversification, and vertical integration) that are assumed to be significantly related to performance. Figure 20 illustrates the conceptual model and the findings from LlSREL-based analysis.

0.21 (2.05) 0.01 (0.15)

0.14 (1.51)

Profitability

o.~~o"

(4.16) 0.22 (2.12)

0.90 (---)

~

e ormance

0.26 (2.75)

r-

(6.45)

0.54 (5.08)

~

Sales Growth

r-

0.18 (1.85)

0.49 (5.29)

0.71 (6.54)

Chi-Square= 10.94, df= 6, P-value= 0.09, CFI= 0.96, IFI = 0.96, GFI= 0.97, St. RMR= 0.051

Figure 20

Multiple Regression - Corporate Strategy Perspective

The findings confirm the prior results of the simple regression analysis. Both product diversification and vertical integration are significantly and positively related to performance (t-values of 4.16 and 2.75 respectively). In line with prior findings from

34

Questions of model identification, specification, and fit will be discussed in the following section. While these issues playa certain role in the present simple models, they are not essential for following the current discussion. They are a far more relevant and easier to explain for the following, more complex, models.

184

SPSS-based simple regression analysis, international diversification did not show a significant effect on performance. The competitive strategy perspective has been limited to two variables, differentiation strategy and cost leadership strategy. Both variables are assumed to be positively and significantly related to performance. The excluded third variable, balancer strategy, is a composite of the two variables considered in the model. If both variables show significant effects, the balancer strategy is assumed to behave accordingly. The presented model and the detailed findings can be found in Figure 21. As expected, both differentiation strategy and cost leadership strategy indicate positive and significant effects on performance (t-values of 3.68 and 3.26 respectively). This provides indirect evidence for a significant and positive relationship of the balancer strategy variable with performance.

0.17 (1.60) 0.91

0.35

(_._)

(3'68)~ Performance

0.51

0.70

(6.14)

0.31 (3.26)

0.55 (5.06)

"l

Sales Growth

r

(5.31)

0.70

(6.48)

~.

Chi-Square= 0.93, df= 4, P-value= 0.92, CFI= 1.00, IFI = 1.00, GFI= 0.99, St. RMR= 0.016

Fi9ure 21 Multiple Regression - Competitive Strategy Perspective

The fourth perspective, firm capabilities, postulates significant and positive effects of five variables (innovativeness, market orientation, organizational learning, reputation, and corporate culture) on performance. Figure 22 shows the hypothesized model and the detailed findings regarding the firm capabilities perspective.

185

At first glance the findings regarding the firm capabilities perspective are rather surprising. Only innovativeness showed a significant effect on performance (t-value: 4.40). All other variables, despite their high significance in the simple regression analysis, are no longer significant (t-values of 1.59 and below). The causes of this effect can be traced if we examine the correlations between the five capability variables. All five variables showed exceptionally strong correlations (t-values between 3.00 and 5.76). They furthermore seem to share a common core. A small test provides further evidence for this assumption: If we exclude innovativeness from the model, market orientation turns highly significant. If we continue to exclude market orientation from the model, reputation becomes highly significant in turn (and so on). The five variables seem to jointly share a large part of the variance that they cause in the dependent variable. This shared part of the explained variance is attributed to the variable with the strongest overall effect. Once the strongest variable is removed (in this case, innovativeness), the next variable in line (in this case, market orientation) takes over. Consequently, it no longer makes sense to analyze single variables independently. All five variables reflect a common underlying factor and should be modeled accordingly.

0.85

~

(---)

0.75 (7.13)

Profitability

1

0.58 (5.56) ~

~

Market Share

~

Sales Growth

~

0.28 (3.40)

0.44 (5.14)

0.66 (6.39)

Chi-Square= 10.89, df= 10, P-value= 0.37, CFI= 0.99, IFI = 0.99, GFI= 0.97, St. RMR= 0.027

Figure 22

Multiple Regression - Firm Capabilities Perspective

186

Interim Conclusion: On the whole, multiple regression analysis at the sub-model level confirms our prior findings. All three industry variables and international diversification failed to show a significant effect on performance. All three competitive strategy variables and the two corporate strategy variables (product diversification and vertical integration) showed significant and positive relationships with performance. The five firm capability variables are highly correlated. Together, they have a significant and positive effect on performance. 35 However, due to the high correlation, it is difficult to deconstruct the overall effect to single factors. We will thus continue to develop comprehensive models for each perspective. These "submodels" treat the single variables as parts of one underlying construct with a common influence on performance.

Submodel Analysis

The objective of the present section is to establish comprehensive LlSREL submodels for all four perspectives. Before addressing this objective, we need to provide more background on LlSREL models. The first part of this section is thus dedicated to a brief introduction to more complex LlSREL models. LlSREL Models: Despite minor differences, most introductions to Structural Equations Modeling discern six main phases when establishing and testing LlSREL models: (1) model conceptualization, (2) model specification, (3) model identification, (4) model estimation, (5) model fit, and (6) model modification. We will briefly summarize these six steps.36 When establishing a LlSREL model, one has to ensure that the model is characterized by sound model conceptualization. This means that all variables and relationships in the model are based on prior theorizing. The LlSREL model is most effective when applied for confirmatory purposes (Diamantopoulos I Siguaw 2000, 13). In our study, all models reflect the prior theoretical discussion and are meant to confirm established hypotheses.

35 The five variables explained almost 48% of the variance in performance (R2= 0.479) - significantly more than any of the variables on their own.

187

Once a theoretical foundation and testable hypotheses have been established, the next step is model specification. Specification normally requires two steps in LlSREL. First, a path diagram is established that specifies all the variables and relationships of a hypothesized model. In a second step, the relationships are translated into a linear equations system. Whereas the latter is a simple (?) question of programming, the former merits further explanation. In general, a full LlSREL path model is composed of three parts. Let us choose the industry structure submodel as an example (see Figure 23). The measurement model of the latent exogenous variable describes how the latent exogenous variable (in our example, industry structure) is measured by corresponding manifest indicators (in our example, industry concentration, market growth, and entry barriers).37 The measurement model of the latent endogenous variable describes how the latent endogenous variable (in this case, performance) is operationalized through manifest indicators (profitability, market share, and sales growth). The structural model describes the relationship between the different latent variables (here, industry structure and performance). The directional ("causal") relationship between the exogenous and endogenous latent variables is measured by means of a regression coefficient. Each variable is complemented by an error (residual) term. All relationships between the model's individual elements are graphically represented in the path diagram (see Figure 23).38 Once the model is specified in a path diagram, the model identification has to be tested. The problem of identification centers on the question of whether one has sufficient information to generate a unique solution for the parameters to be estimated in the model. To obtain a unique solution, the number of unknowns has to be less or at least equal to the number of equations. In order to test the minimum requirement for identification for our example in Figure 23, the following formula can be applied: DF

=q

(q+1 )/2-p (Bollen 1989, 93). If DF < 0, then the model is unidentified and cannot be

38

37

38

The following explanations are taken from a number of LlSREL manuals and introductory essays. The main references are Bollen (1989) and Diamantopoulos I Siguaw (2000), additional sources include Hoyle (1995), Kelloway (1998), and J6reskog I S6rbom (2001). Manifest indicators in LlSREL modeling take the form of reflective indicators by default, meaning that a latent variable is thought to "cause" the observed indicators. Observable variables are indicators of the underlying cause (i.e. the latent variable). In this (and the following) section, we follow this approach. As can be seen from the path diagram, one of the path coefficients of each latent variable has been fixed at 1.00 (i.e. the path between performance and profitability). These parameters are fixed in order to set the unit of measurement for the respective latent variables. Latent variables have to have a scale in order to make them interpretable (Bollen 1989, 91).

188

estimated. 39 If OF = 0, the model is said to be just-identified, a single solution can be obtained. Ideally, the model is overidentified (OF > 0), it provides several sets of estimates that can be used to both estimate and test the model. In our example, OF = 8, the model is thus overidentified. 4o The identified model can then be estimated by LlSREL.

Measurement Model of the latent exogenous variable Figure 23

Structural Model

Measurement Model of the latent endogenous variable

Example of LlSREL Model

Model estimation generates numerical values for the free parameters in the model. More

exactly, the objective of estimation is to analyze the fit between the input covariance matrix and the corresponding covariance matrix calculated by the model. LlSREL 8 offers seven alternative estimation methods. 41 The default estimation method is maximum likelihood. Maximum likelihood is also the most widely accepted method and performs well under a variety of less-than-optimal analytic conditions including small sample sizes or moderate non-normality (Hoyle I Panter 1995, 163; Oiamantopoulos I Siguaw 2000, 56). All parameters in this study have been estimated by the maximum likelihood method.

39 40

41

The model can, however, be corrected by adding further constraints (i.e. by fixing additional parameters to 0 or setting parameters equal to each other). OF is an abbreviation of degrees of freedom and is automatically calculated by LlSREL. In order to calculate it manually, one has to count all the observed variables (q) and all variables to be estimated (p). In our example, q=6 (3 observed variables for industry and 3 observed variables for performance) and p=13 (6 error covariances for the observed variables, 2 error coefficients for the latent variables, 4 path coefficients between observed variables and latent variables, 1 regression path between the two latent variables). Namely, Instrumental Variables (IV), Two-Stage Least Squares (TSLS), Unweighted Least Squares (ULS), Generalized Least Squares (GLS), Maximum Likelihood (ML), Generally Weighted Least Squares (WLS), and Diagonally Weighted Least Squares (OWLS).

189

As stated above, model estimation analyzes the model fit. When talking about fit, we refer to the extent to which a hypothesized model is consistent with the input data. Model fit evaluates how well the model represents the "reality". LlSREL offers a wide range of goodness-of-fit indices to evaluate model fit. There is great controversy in the literature around the question of which fit indices are best. The only consensus is that one should rely on multiple criteria. In line with most recommendations in the literature, we have chosen five fit indices for this study.42 The first three indices (Chi-Square, Standardized RMR, and GFI) evaluate the absolute model fit. Tests of absolute model fit are concerned with the ability of a model to reproduce the input covariance matrix (Kelloway 1998, 24). The first measure is the chisquare statistic (X 2 ), the traditional measure for evaluating overall model fit. Contrary to traditional hypothesis testing, a nonsignificant X2 indicates that the model perfectly fits the data. Small X2 values therefore indicate good model fit. The degrees of freedom (OF) serve as a standard by which to judge whether X2 is large or small (Joreskog / Sorbom 2001, 43).43 The second absolute fit index, the standardized root mean square residual (St. RMR) indicates the square root of the mean of the squared discrepancies between the implied and observed covariance matrices. Low values are taken to indicate good fit. Generally, values less than 0.05 indicate a good fit. The goodness-of-fit index (GFI) is an indicator of the relevant number of variances and covariances accounted for by the model. The GFI values range from 0 to 1, with values exceeding 0.9 indicating a good fit with the data (Kelloway 1998, 27; Diamantopoulos / Siguaw 2000, 87). Due to the problems inherent in assessing the absolute fit of a model, researchers have increasingly turned towards measures of comparative (or relative) fit. These measures indicate whether the model is better than some competing model. 44 The literature recommends two factors (IFI and CFI) that we include in our study (Kelloway 1998, 29ff). Bollen's (1989) incremental fit index (IFI) and Bentler's (1990) comparative fit index (CFI) range between 0 and 1, with values exceeding 0.9 indicating a good fit with the data.

42

43

44

Please refer to Bollen (1986. 256ft), Hu I Bentler (1995. 76ft). Kelloway (1998. 23ft). Diamantopoulos I Siguaw (2000. 82ft) for a detailed presentation and discussion of alternative fit indices. A well-fitting model would ideally be indicated by a "I.: that approximates the degrees of freedom. In practice. X2/df ratios of 2 up to 5 have been used as thresholds. However. these normed values are criticized by diverse scholars (Diamantopoulos I Siguaw 2000. 98). Normally the baseline (or independence) model without relations (Kelloway 1998. 29).

190

Having established and estimated a model, model modification can help to improve the overall model fit. Modification means changes to the initial model specification by adding or deleting certain parameters. Suggestions for model modifications are provided by LlSREL in the so-called modification index. However, all changes have to be justified by theoretical argumentation. Industry Structure Submodel: After this short introduction to LlSREL modeling, we

continue to discuss the findings of the industry structure submodel. The industry model was used above as an example in our discussion of LlSREL. It suggests the significant effect of industry structure (reflected by the manifest indicators industry concentration, market growth, and entry barriers) on performance (reflected by profitability, market share, and sales growth). Figure 24 shows the path diagram and the respective parameter estimates.

~

0.52 (2.48)

0.86 (---)

0.91 (6.67)

-1

Market Growth

0.62 (3.52)

-1

Entry Barriers

~

0.30 (2.04)

r

0.30 (1.89)

0.61 (2.38)

0.74 (5.51) 0.57 (4.90)

Profitability

~

Market Share

~

Sales Growth

~

-1

1

0.26 (2.15)

0.46 (4.30)

0.67 (6.20)

Chi-Square= 7.62, dl= 8, P-value= 0.47, CFI= 1.00, IFI = 1.00, GFI= 0.98, SI. RMR= 0.048

Figure 24

Structural Equations Modeling - Industry Submodel

The industry submodel shows that all three manifest indicators reflect the underlying construct, industry structure, well (t-values of 2.04 and above). However, while the effect of industry structure on performance is positive and considerable

(~=

0.30), it fails to be

significant (t-value of 1.89). All fit indices indicate an excellent model fit. The results suit our prior findings: Neither the three individual variables - industry concentration, market growth, and entry barriers - nor the common underlying construct - industry structure had a significant influence on performance. According to these findings the perspective

191

does not significantly contribute to the explanation of performance and might be excluded from further modeling. Corporate Strategy Submodel: The corporate strategy submodel suggests the

significant effect of corporate strategy (reflected by the manifest indicators international diversification, product diversification, and vertical integration) on performance. Figure 25 shows the detailed findings of the model's estimation. The most important finding is a particularly strong positive (13= 0.86) and significant (t-value= 2.83) effect of corporate strategy on performance. The corporate strategy construct is well reflected by the two indicators product diversification and vertical integration (t-values of 2.90 and 2.18 respectively), whereas international diversification shows only a weak relationship.

~

0.93 (6.85) 0.27 (---) 0.67 (4.61)

0.86 (6.50)

0.58 (2.90)

Profitability

~

0.17 (1.63)

0.91 (---) 0.86 (2.83)

0.38 (2.18)

0.50 (5.39)

0.71 (6.39) 0.54 (5.04)

1

Sales Growth

~

0.71 (6.57)

Chi-Square= 13.66, df= 8, P-value= 0.09, CFI= 0.96, IFI = 0.96, GFI= 0.96, SI. RMR= 0.057

Figure 25

Structural Equations Modeling - Corporate Strategy Submodel

The overall model fit is good as indicated by the X2, the GFI and the two comparative fit indices. A standardized RMR of 0.057 is acceptable, but reflects the negative influence of the weaker indicator, international diversification. If industry diversification is deleted from the model, the overall fit indices improve further and the effect of corporate strategy on performance increases dramatically in significance (t-value= 4.47). We can conclude that corporate strategy has a significant effect on performance. Furthermore, the construct is well reflected by product diversification and vertical integration. However, international diversification does not contribute to the model and should be excluded from further modeling. These findings are consistent with our prior

192

results that showed the significant impact of both product diversification and vertical integration, while failing to do so for international diversification. 45 Competitive Strategy Submodel: The competitive strategy submodel suggests the significant effect of competitive strategy (reflected

by the manifest indicators

differentiation strategy and cost leadership strategy) on performance. 46 Figure 26 shows the path diagram and summarizes the findings. First of all, competitive strategy shows a particularly strong positive

(~=

0.93) and highly significant (t-value= 3.81) effect on

performance. Furthermore, the two manifest indicators, differentiation strategy and cost leadership strategy, reflect the underlying latent variable, competitive strategy, well. The overall model fit is excellent as indicated by all five fit indices. 0.83 (5.80) 0.91

~

(---)

0.93 (3.81)

0.55 (5.05)

0.43

0.81 (5.59)

0.70 (6.13)

(3.54)

Profitability

~

Market Share

~

Sales Growth

~

-1

1

0.17 (1.57)

0.51 (5.32)

0.70 (6.49)

Chi-Square= 1.11, df= 5, P-value= 0.95, CFI= 1.00, IFI = 1.00, GFI= 0.99, St. RMR= 0.024

Figure 26

Structural Equations Modeling - Competitive Strategy Submodel

Based on the findings, we confirm the strong and significant effect of competitive strategy on performance. The findings are in line with our prior results of a significant impact by both differentiation and cost leadership strategy on performance.

45 46

A potential curvilinear relationship has been tested for all models, including corporate strategy variables. No evidence has been found for such a relationship. The third indicator, balancer strategy, is composed of the two other indicators and thus tested implicitly. The path from competitive strategy to differentiation strategy has been set equal to the path from competitive strategy to cost leadership strategy. The variance of competitive strategy has been fixed at 1.00. Both measures are required in a LlSREL model with only two indicators reflecting a latent variable (in order to identify the model).

193

Firm Capabilities Submodel: Earlier in this study, we found that while each of the five

firm capability variables separately showed a highly significant effect on performance, they failed to do so when taken together. We concluded that all five variables represent a common underlying construct with a (partly) shared impact on performance. The firm capabilities submodel is based on these conclusions. Firm capabilities are reflected by five manifest indicators (innovativeness, market orientation, organizational learning, reputation, and corporate culture). We further assume a positive effect of firm capabilities on performance. Figure 27 shows the path model and the detailed estimates for all the parameters. 47 0.41 (4.85) 0.80

(---)

-0.14 (-2.24)

0.60 (6.35)

\

0.38 (3.76)

»

0.08 (1.11)

>- (~~~) -1

0.14 (2.29) '-.... 0.63 (6.39)

r

0.77 (6.97

Reputation

0.61 (6.71)

Chi-Square= 22.35, df= 16, P-value= 0.13, CFI= 0.98, IFI = 0.98, GFI= 0.95, St. RMR= 0.050

Figure 27

Structural Equations Modeling - Firm Capabilities Submodel

From the findings we can conclude that firm capabilities had a positive and highly significant (t-value: 5.71) effect on performance. All five manifest variables significantly reflected the underlying latent construct firm capabilities (particularly high t-values of 6.05 and above). The model shows a very good fit across all five fit indices. In summary, the model confirms the prior assumption that all five indicators represent a common underlying construct and exert a strong common impact on performance.

47

The modification index recommended adding three error covariances to the model in order to improve the overall model fit. As shown in Figure 27, all three paths have been added. They are justified by the literature on firm capabilities reviewed above. Figure 27 shows the estimates for the modified model.

194

Interim Conclusion: Up to this point we can confirm that three constructs, notably

corporate strategy, competitive strategy, and firm capabilities, indicate a positive and significant effect on performance. The fourth construct, industry structure, failed to show a significant effect. At the indicator level, a number of variables have been repeatedly confirmed as relevant. Product diversification and vertical integration showed a significant direct effect on performance, as well as indirectly as constituent elements reflecting the underlying construct corporate strategy. Differentiation strategy, cost leadership strategy and (measured indirectly) balancer strategy had a significant relationship with performance and together represent the underlying latent variable competitive strategy that exerts a Significant impact on performance.

Finally,

innovativeness, market orientation, organizational learning, reputation, and corporate culture reflect the latent construct firm capabilities that, in turn, Significantly affects performance. Based on these findings, we continue to bring the different perspectives together in an overall model.

Model Analysis

The overall integrated model combines the different perspectives on strategic management. However, as discussed above, an important restriction of the model is the exclusion of the interaction effects between the different perspectives. The main objective of this section is thus to show both the merits and the limitations of the integrative perspective. We excluded the industry perspective and one determinant of corporate strategy, namely international diversification, from our analysis for the integrated model. All three industry variables, as well as international diversification, had repeatedly failed to produce a significant result in our prior analysis. These changes lead to a model with three exogenous latent constructs (corporate strategy, competitive strategy, and firm capabilities) expected to exert a significant effect on a single endogenous latent variable (performance). All four latent constructs are reflected by a number of underlying manifest indicator variables (the same as the above). An important particularity of the model is that the paths between the three exogenous latent variables have been set at equal to zero. This has been done to reflect the above mentioned restriction of no interaction

195

effects between the three perspectives. Figure 28 shows the path diagram and the results of the model estimation in LlSREL. The results imply two important sets of findings that mirror the inherent strengths and weaknesses of the integrated model. Let us consider the strengths of the model first. As we can see from Figure 28, all three perspectives show a significant and strong effect on performance. Firm capabilities had a particularly significant effect (t-value: 4.68), followed by corporate strategy (t-value: 2.94) and competitive strategy (t-value 2.55). The strength of the effect was comparable for all three latent constructs, with corporate strategy

(13=

(13=

0.53) slightly ahead of firm capabilities

(13=

0.49) and competitive strategy

0.48). Furthermore, all manifest indicators load significantly and positively on the

assumed underlying latent constructs (t-values between 2.38 and 9.78; betas between 0.33 and 0.87). These findings are perfectly in line with our prior results. The findings strongly support our first lead hypothesis that insights from multiple perspectives on strategic management are jointly required to explain performance. Furthermore, the identified determinants have been confirmed as highly relevant in explaining firm performance.

-+l (~:~~) -+l

(g:;~)

(~.~~) 0.85 (5.31) 0.25 (4.40)

-+l

-+I -+I

PrOduct Diversific.

Vertical

Intearation

rJlfreren~. Slli!t!!9

Cost Lead Strategy

h

0.41 (5.64)

Organtzaf. Learning

Market

(~:~~) -I

Reputation

0.57 (6.45)

Corporate Culture

0.53 (2.94)

0.46 (2.50)

r- &~~)

ness

Orientation

-+I

I*- (2.40) 0.33

Innovahve- ~

(~:~~) -I

-I

I- (g:~~)

j.-

0.87 (---)

~0.27 0.85 (---)

biti

0.68 ~ 0.54 (5.92) Share (5.48)

0.49 (4.68)

0.51 (4.64) ~ 0.74 Growth (6.51)

0.65 (6.82)

r- (~:~~)

r-

r

0.82 (9.78) 0.66 (6.91)

Chi-Square= 123.59, df= 49, P-value= 0.00, CFI= 0.82, IFI = 0.82, GFI= 0.83, SI. RMR= 0.18

Figure 28

(2.73)

Structural Equations Modeling - Integrated Model

196

If we turn our attention towards the fit indices, the inherent shortcomings of the integrated model become apparent. All five fit indices shown above are below the cutoff values generally accepted in the literature. A look at the modification index reveals the causes for the significantly lower fit of the integrated model (compared to the submodels). The modification index strongly recommends adding paths between all three exogenous latent variables. In order to improve overall model fit, interrelations between the three perspectives have to be taken into account. The correlation matrix further supports that need, indicating strong correlations between various determinants from different perspectives. As Hu and Bentler (1995, 96) have stated: "Especially when the latent variables are dependent, none of the fit indexes behaved adequately at small sample sizes".

At this point, we clearly reach the limits of the integrated model. While revealing the different constructs and underlying manifest variables that are important to explain performance, the integrated model fails to capture reality in its entire complexity. However, it establishes a sound foundation for further analysis of the more complex interactions of the model. After some conclusions on the integrated model, we thus continue in the following section with the analysis of the complex model.

Discussion of the Research Findings

The first objective of this section was to identify the determinants that play an important role in explaining firm performance. Fourteen variables had been derived from the strategy literature and hypotheses had been formulated regarding their effect on performance. Subsequently, three statistical methods were applied to test evidence for these hypotheses. All three tests largely led to the same results. All three industry variables (industry concentration, market growth, and entry barriers) and one strategy indicator (international diversification) failed to prove their importance in explaining performance. In contrast, all ten remaining variables repeatedly confirmed their significant effect on performance. The five strategy variables had both a significant direct impact on performance and reflected the underlying latent constructs of corporate (product diversification, vertical integration) and competitive strategy (differentiation, cost

197

leadership,

and

balancer

strategy)

well.

The

five

firm

capabilities

variables

(innovativeness, market orientation, organizational learning, reputation, corporate culture) reflected the underlying latent construct of firm capabilities well. While they indicated a strong direct effect on performance in simple regression analysis, the more sophisticated multiple regression analysis showed that all five variables are strongly correlated and share a common effect on performance. Overall, these findings confirm 10 of the 14 hypotheses of the integrated model. Figure 29 summarizes the findings according to these hypotheses. Variable

Simple Regression

Multiple Regression

Structural Equations

Overall Importance

Industry Concentration Market Growth Entry Barriers

0

International Diversification Product Diversification

00

00

00

00

Vertical Integration

00

00

00

00

Differentiation Strategy

00

00

00

00

Cost Leader Strategy

00

00

00

00

Balancer Strategy

00

00

00

00

Innovative ness

00

00

00

00

Market Orientation

00

0

00

00

Organizational Learning

00

0

00

00

Reputation

00

0

00

00

Corporate Culture

00

0

00

00

Figure 29

Integrated Model - Hypothesis Testing

198

Besides the detailed evaluation of the role of individual determinants in explaining performance, the objective of this section was to confirm the need for an integrated perspective. The lead hypothesis of the integrated model was that performance is best explained by a model that captures the most important determinants from different perspectives of strategic management. Analysis of the integrated model showed that three of the four perspectives confirmed their important role in explaining performance. Firm capabilities, corporate strategy, and competitive strategy all had a significant effect on performance thus confirming the integrated model. While the integrated model clearly represents a step forward compared to simpler models limited to one perspective, the model itself showed some limitations. LlSREL output indicated that a more complex model that considers interrelations between the different perspectives might lead to a better overall fit. This brings us closer to the lead hypothesis of the complex model.

V.2

The Complex Model

The lead hypothesis of the complex model is that performance can be best explained by a model that captures the most important relationships between determinants from different perspectives. The complex model is thus an extension of the integrated model, abolishing the

assumption

of there

being

no

interactions

between

individual

perspectives. Our starting point for an analysis of the complex model is therefore the results of the integrated model presented above. Having established the general importance of the complex model, we continue to analyze the most important interactions between single factors in detail. The section ends with some conclusions on our findings.

The Overall Model

In the previous section, we found that the overall fit of the integrated model could be improved through the addition of paths between the three exogenous latent constructs. Taking these interactions into account takes us from the integrated model to a model that

199

reflects some of the ideas of the complex model. Figure 30 shows the estimates resulting from the integrated model through the simple addition of paths between the latent variables. 48 First of all, the results confirm the expected increase in model fit. All five fit indices are now in line with the generally accepted conditions (X 2 = 65.5, DF= 45, p= 0.03, CFI= 0.94, IFI = 0.94, GFI= 0.90, St. RMR= 0.062). However, while all manifest variables still load significantly on the underlying latent variables, the latter no longer exhibit a significant influence on performance. The t-values for the effects of corporate strategy, competitive strategy, or firm capabilities dropped below one. 0.76 (5.52)

-+I

(g~)~

prOduCt

Diversific.

vertical

Intearation

g~~) ~

r:'il"eren~. Strat!l9

0.81 (6.01)

I:!osf~aa Strat y

-+I

(~:~~) ~

--.J &~~) --.J (g.!~) --.J (~.~)

(~~g) ~

I- (~:~~)

r-

ne

Market

Organlzat. Learning Reputation

Corporate Culture

0.08 (0.07)

I- &~3) I- S!~)

Innov~ve- ~ Orientation

0.44 (3.56)

t---

0.93 (---)

0.73 (7.44)

0.16 (0.31)

0.25 (3.26)

0.87 (---)

~ 0.47 Share

(5.58)

0.58 (5.71) ~ 0.67 Growth (6.49)

0.60 (6.87)

r-- (~:~~)

r---

r

0.75 (9.07) 0.55 (6.24)

Chi-Square= 65.50, df= 45, P-value= 0.03, CFI= 0.94, IFI = 0.94, GFI= 0.90, SI. RMR= 0.062

Figure 30

Structural Equations Modeling - Complex Model (1/3)

These findings might appear surprising at first sight. However, once again, a look at the modification index provides some hints regarding the root causes underlying these findings. The modification index recommends the addition of a number of additional paths between manifest indicators and latent variables of other perspectives. The strict division into three perspectives does not seem to fully represent the empirical findings.

48

Technically, we are not really "adding" paths, but removing the restrictions to the paths established before. However, this is just a technical difference due to LlSREL's default function.

200

Are corporate strategy, competitive strategy, and firm capabilities really distinct constructs as claimed in most of the extant literature? The covariances between the three latent variables indicate a close dependence (tvalues between 3.47 and 4.65). In addition, the correlation matrix shows highly significant correlations between several manifest variables. From the 72 correlations between the remaining nine manifest indicators, 52 turn out to be significant at the 0.05 level, thereof 44 even at the 0.01 level of significance (2-tailed). This provides a strong basis for the assumption that the three theoretical constructs are a lot more interrelated than generally assumed in the literature. We could even speculate that the respective manifest variables reflect a single underlying construct. In order to test the (in)dependence of the latent constructs, we established a second model. This time, we added a fourth exogenous latent variable to the present model. We labeled this construct "competitive advantage". The construct is reflected by all nine of the manifest variables from corporate strategy, competitive strategy, and firm capabilities retained in the model. The competitive advantage construct is supposed to represent the totality of all firm activities and characteristics influencing performance. 49 At the same time, the manifest indicators remain connected to their "traditional" latent variables (corporate strategy, competitive strategy). This way, the model allows the share of the impact that is really unique to these constructs to be detected. Figure 31 shows the results of this second step in modeling the complex model. 5o As can be seen from Figure 31, the former distinctly latent variables no longer have a significant effect on performance once the manifest indicators are relayed to competitive advantage. All nine indicators load highly significant on the new latent variable competitive advantage. The loadings are all superior to the remaining loadings of the former latent constructs. We can take this as strong evidence that all nine manifest indicators are strongly related and reflect a common underlying construct. A large part of the impact on performance is shared between the indicators. A division of the effect into 49

50

It has been argued (among others by the scholars of the efficiency school) that firms actively influence industry factors such as industry concentration. However. most scholars of strategic management agree that this influence is rather indirect and a lot less effective than the manipulation of factors such as firm strategy or resources. Exhibit 31 has been simplified by not specifying the standard errors for the manifest indicators.

201

several latent variables, or distinct perspectives, seems to deteriorate the model rather than improve it.

0.32 (3.13) 0.40 (3.99)

-1 -4

.,

Proauct

Diver~ific.

Verllca~

Intearati n

j... g~;)

fo-

0.46 (1.61)

0.16 (1.20)

;~'45~;:::::;;;;;:;;~;:::::--::-::::----0.45 (4.56) 0.28 (2.70)

0.56 (7.17) \ \ '\ 0.76 (8.70) 0.84 (---)

0.63 (6.76)

~'lfereni trateg

j... (~:~~)

'1

Coe![ead Sir tegy

~ (~:~i)

~

Innov:ve-

~

~

0.88 (---)

~0.72

0.69 ~(7.35) (6.54) 0.50 (1.19)

0.56 (5.55)

n~

~a~e!

Qr.L.n ... tlon

organlzat. Les;!rning

Chi-Square= 106.41, df= 47, P-value= 0.00, CFI= 0.85, IFI = 0.86, GFI= 0.85, SI. RMR= 0.072

Figure 31

Structural Equations Modeling - Complex Model (2/3)

Consequently, in a third (and final step), we continue to establish a model that no longer distinguishes between different perspectives. It truly integrates the different variables and models them as indicators reflecting a single underling construct, competitive advantage. This construct is assumed to represent the sum of the company's activities and characteristics that have a direct impact on performance. Figure 32 shows the final, truly complex model and the results of this model's estimates. 51 The complex model shows the strongly positive (13= 0.75) and highly significant (t-value= 6.86) effect of competitive advantage on performance. All eight indicators load positively (betas of 0.40 and above) and highly significant (t-values between 4.67 and 7.10) on competitive advantage. The fit indices indicate an excellent 51

One additional change has been made: The manifest variable balancer strategy replaced the two separate variables: differentiation strategy and cost leadership strategy. As the variable reflects both constructs and has been repeatedly proven as superior to the sum of its two parts, we preferred this variable for the model at hand. We modeled all alternatives and found this one to perform best in terms of model fil.

202

overall model fit (J(2= 0.054).52

42.7,

DF=

36,

p=

0.21,

CFI=

0.98,

IFI = 0.98, GFI=

0.93,

st.

RMR=

Both in model fit and parameter estimates, the model shows results largely

superior to those of the integrated model and the alternative complex models shown above. 53 Overall, the complex model explained about performance (R2=

_ -0.25 (-3.96)

0.51 (5.11)

0.84 ~~~ 0.62 (6.07) 0.29 (3.90)

of the variance in

0.57).

0.74 ;(6.89) ~~ (-3.60)

57%

--l

I

Balancer Strategy

L

i

0.23 (2.90)

O~ (4:02) 0 61 (5.12) 0.84 ( ) - -0.62 (6.30)

0.75 (6.86)

0.72 (7.23) 0.58 (5.71)

0.69 (6.21) 0.69 (7.10)

0.48 (5.67)

0.67 (6.50)

Chi-Square= 42.68, df= 36, P-value= 0.21, CFI= 0.98, IFI = 0.98, GFI= 0.93, SI. RMR= 0.054

Figure 32

Structural Equations Modeling - Complex Model (3/3)

If all manifest indicators are closely related, it makes sense to further analyze these relationships and to uncover the "common core" of these factors. In the remainder of this section, we continue this line of thought with a further analysis of single relationships between the manifest indicators of competitive advantage.

52 53

The Root Mean Square Error of Approximation (RMSEA) was at 0.042, a highly satisfying score and largely superior to the alternative models discussed above. Hoyle and Panter (1995) stated: "Comparison of adjunct fit indexes, parlicularly incremental fit indexes, should be avoiderJ'. We followed their recommendation and compared differences in X2 (in ratio to OF), as well as in the cross-validation index (CVI), a measure recommended by Oiamantopoulos and Siguaw (2000).

203

Important Interactions

As described above, two types of intervening variables can be distinguished: moderators and mediators (Baron / Kenny 1986).54 Moderation occurs if the causal effect of one predictor variable on an outcome variable changes as a function of a third (moderator) variable. No causal sequence is implied by interaction. For example, a high degree of innovativeness might lead to success in a high growth market but not in low growth markets - an interaction effect. If there is moderation, the intervening variable is generally uncorrelated to both the predictor and the outcome variable (Baron / Kenny 1986, 1174). In our model, the three industry variables might be considered as potential moderators. They fulfill the condition of being uncorrelated to the outcome variable (performance), as we have found above. Furthermore, the industry variables have been found to be uncorrelated to predictor variables of the remaining three perspectives. In addition, we have reviewed numerous theoretical arguments from contingency literature that assume such a moderator effect of industry variables on strategy or capability indicators. 55 However, we did not find any proof for significant moderator effects of the three industry variables in this study. Together with the prior findings, we can conclude that the analyzed industry variables failed to show any direct, indirect (or intervening) effect on performance. The industry perspective has thus been found to contribute very little to the explanation of variance in business unit performance. In the remainder of this section, we thus turn our attention to interrelations between the three remaining groups of variables: corporate strategy, competitive strategy, and capabilities. With the exception of international diversification, all remaining variables have been found to significantly affect the outcome variable (performance). Furthermore, we found them to be closely related and to share a common effect to some degree. These findings point to potential mediation effects. Mediation is said to occur when a causal effect of some predictor variable on an outcome variable is (partially) explained by some intervening variable (Shrout / Bolger 2002). 54

55

For an introduction into moderation and mediation, please refer to Chapter IVA (Statistical Methods). Please refer to Chapter 111.2 for a discussion of the contingency theory and Chapter lilA for the hypotheses derived from the respective literature.

204

Mediation can be identified by the causal step approach that specifies a series of tests of links in a causal chain (Judd I Kenny 1981, Baron I Kenny 1986). In more detail this means that the predictor variable has to significantly affect both the outcome and the mediator variables in simple regressions. Additionally, the mediator has to cause the outcome variable in a simple regression. When taken together in a multiple regression, the initial effect of the predictor on the outcome variable has to shrink (due to the addition of the mediator to the model). The causal step approach has been widely applied and praised for its conceptual clarity, but at the same time criticized for low statistical power (MacKinnon et al 2002). Alternative approaches have been developed to assess mediation more formally. Some of these tests are based on the product of coefficients involving paths in a path model (MacKinnon et al 1995). Three of these tests have been frequently applied in prior research: the Sobel test, the Goodman (I) test, and the Goodman (II) test (Goodman 1960; Sobel 1982, 1988; Baron I Kenny 1986, MacKinnon et al 1995).56 Especially the first two tests performed best (compared to alternative tests) in two Monte Carlo studies (MacKinnon et al 1995, 2002) and converged closely with sample sizes greater than 50. Following recommendations from the applied literature (Holmbeck 1997), we deploy both the causal step approach and the three formal product of coefficients tests for our model. Based on the hypotheses established above, we continue to test for mediation effects in our model. Competitive Strategy - Capabilities: Three mediational relationships have been assumed between competitive strategy and firm capability variables. More exactly, we assumed the effect of a differentiation strategy on business unit performance to be mediated by (1) market orientation, (2) innovativeness, and (3) organizational learning. Figure 33 summarizes the results of both causal step and product of coefficients testing for mediation.

56

The advanced statistician might enjoy the respective formulae for the three tests (MacKinnon et al 1995): (1) Sobel test equation: z-value= a*b/SQRT(b2*s;+a2*sb2); (2) Goodman (I) test equation: zvalue= a*b/SQRT(b2*s;+a2*sb2+sa2*Sb2); (3) Goodman (II) test equation: z-value= a*b/SQRT(b2*s;+a2*sb2-s;*Sb2). a and b are path coefficients. a stands for the unstandardized regression coefficient for the association between the predictor variable and the mediator, b for the association between mediator and outcome variable. s. and So are the standard errors of a and b.

205

1. Without Mediation: 0.36

~

(3.90)~

2. After Mediation:

0.53 (6.29)

0' 0.27 (2.56)

Performance

3. After Mediation:

0.37 (4.03)

0' 0.40 (4.26)

0'

4. After Mediation:

0.29 (3.08)

0.27 (2.87)

0.22 (2.03)

0.22 (2.32)

0.28 (3.01)

Sobel test: 2.37 (p= 0.02) Goodman (I) test: 2.35 (p= 0.02) Goodman (II) test: 2.40 (p= 0.02)

Sobel test: 2.93 (p= 0.Q1) Goodman (I) test: 2.89 (p= 0.01) Goodman (II) test: 2.97 (p= 0.01)

Sobel test: 2.10 (p= 0.04) Goodman (I) test: 2.04 (p= 0.05) Goodman (II) test: 2.16 (p= 0.04)

Figure 33

Mediation Effects (1/3) - Competitive Strategy & Capabilities

The figure shows first the significant effect of differentiation strategy on performance in the absence of any mediation effects

(~=

0.36, t-value= 3.90). If we consider the first

mediator, market orientation, we can confirm a strong mediation effect. The causal step approach shows both the significant effect of differentiation strategy on market orientation and of the latter on performance. Furthermore, once market orientation is taken irito account, the effect of differentiation strategy on performance drops significantly (the beta from 0.36 to 0.22, the t-value from 3.90 to 2.03). All four conditions for mediation are met. The three product of coefficient tests confirm these findings. The test statistics are all highly Significant (t-values larger than ±1.96; p= 0.02). The two other assumed mediational effects have also been confirmed.

Both

innovativeness and organizational learning fulfill all conditions for a mediator and show significant results for all three product of coefficient tests. From these findings, we conclude that the established effect of differentiation strategy on performance is partially explained by the underlying capabilities market orientation, innovativeness, and organizational learning. While each a differentiation strategy and the related capabilities have a unique (and Significant) effect on performance, there is an additional shared effect. This shows the close interrelation between these factors. In other words, the

206

success of a differentiation strategy would be a lot more likely if the company simultaneously enjoys a high degree of market orientation, innovativeness, and organizationallearning. 57 Corporate Strategy - Competitive Strategy: With regard to the interrelations between

the two major levels of firm strategy, we assumed (based on the extant literature) three mediational relationships. First, we hypothesized the effect of product diversification on business unit performance to be mediated by a cost leadership strategy. Second, the relationship between international diversification and business unit performance was expected to be mediated by a differentiation strategy. Finally, a balancer strategy was expected to mediate the relationship between vertical integration and business unit performance. A mediational effect between international diversification can

be

immediately excluded from further analysis. This factor has been found to neither effect performance, nor the (potential) mediator differentiation strategy. 58 Figure 34 outlines the results of the tests for mediation for the two remaining hypotheses. The left half of the figure shows the findings for the mediational effect between product diversification and a cost leadership strategy. The upper box recalls our former finding of a significant mediation effect between product diversification and performance t-value:

4.17).

The figure

below confirms significant effects between

(~=

0.38,

product

diversification, cost leadership strategy, and performance. Inclusion of the mediator cost leadership strategy leads to a reduced effect of product diversification on performance (the beta from 0.38 to 0.31; t-value from 4.17 to 3.30). The four causal steps of mediation are present. However, the product of coefficient tests shows that the mediational effect fails to be significant at the 95% confidence level (values lower than ±1.96). We can conclude that while there is some indicator of a mediational effect, it fails to be

57

58

While not part of our established hypotheses, it is important to state that the same mediational effects (with market orientation, innovativeness. and organizational learning) have also been found for the relationship between a balancer strategy and performance. Testing these effects for a cost leadership strategy revealed much lower mediational effects for the three capabilities (only the effect of innovativeness was significant). It has been argued that Baron and Kenny's (1986) condition of a significant association between predictor and outcome variables should not be a requirement for mediation (i.e. Shrout I Bolger 2002). However. as no significant association has been found between predictor and mediator variables. a mediational effect can be excluded for international diversification. In addition. no evidence has been found for a potential moderation effect.

207

significant. The result is pretty much in line with the limited evidence we found in the present literature.

1a. Without Mediation:

I Diversification Product

0.38 (4.17)

2a. Without Mediation:

---I

Performance

.

I

I

[8]

1b. After Mediation:

Vertical Integration

Performance

.

0'

\

0.31 (3.30)

Performance

0.37 (4.16)

0.25 (2.62)

i

I

I

Vertical Integration

\ 0.26 (2.92)

Sobel test: Goodman (I) test:

1.83 (p= 0.07) 1.78 (p= 0.07)

Sobel test: Goodman (I) test:

2.22 (p= 0.03) 2.17 (p= 0.03)

Goodman (II) test:

1.89 (p= 0.06)

Goodman (II) test:

2.27 (p= 0.03)

Figure 34

I

Balancer Strategy 0.21 (2.15)

i

I Diversification Product

---I

2b. After Mediation:

Cost Leader Strategy 0.33 (3.46)

0.36 (3.81)

Performance

I

Mediation Effects (2/3) - Corporate & Competitive Strategy

The right half of the figure recalls the significant relationship between vertical integration and performance (13= 0.36; t-value= 3.81). The figure below confirms significant relationships between vertical integration, balancer strategy, and performance. Inclusion of the mediator balancer strategy leads to a significant reduction of the effect of vertical integration on performance (the beta drops from 0.36 to 0.26, the t-value from 3.81 to 2.92). All four conditions for mediation are met. Furthermore, all three tests confirm the significance of the mediational effect (at p= 0.03). Based on these findings, we can state that the effect of vertical integration on performance is partially explained by a balancer strategy. In other words, the success of a strategy of vertical integration is to some extent dependent on the firm's simultaneous realization of a balancer strategy. This implies the need for a close reCiprocal adjustment of business-level and corporate-level strategy. 59

208

Corporate Strategy - Capabilities: Similar to the above, we derived three hypotheses on mediational effects between corporate strategy and capability variables from the present literature. Initially, we assumed the effect between product diversification and performance to be mediated by organizational learning. Next, we assumed the relationship

between

vertical

integration

and

performance to

be

mediated

by

innovativeness and organizational learning. Figure 35 shows a significant mediation effect for all three cases, thus confirming the established hypotheses. The four conditions for mediation have been met in all three cases: There are significant relationships between the predictor, mediator, and outcome variables, as well as a significant drop in the effect of the predictor on the outcome variable once the mediator is added to the model. Furthermore, the product of coefficient tests confirm the significance of the effect (at 0.01 < P < 0.05).

1a. Without Mediation:

I Diversif. Product

0.38 (4.17)

-1

.

2a. Without Mediation: Perfor-

mance

I

0

1b. After Mediation:

I

Vertical Integration

2b. After Mediation:

Organizational Learning 0.28 (2.97)

,

0

Performance

Perfor-

mance

0.30 (3.10)

I

I

Learning

~ 0.24 (2.64)

Perfor-

mance

2.56 (p= 0.02) Sobel test: Goodman (I) test: 2.52 (p= 0.02)

Goodman (II) test: 2.12 (p= 0.04)

Goodman (II) test: 2.60 (p= 0.01)

0

2e. After Mediation:

0.41 (4.54)

/

Vertical Integr.

I

Organizational

2.06 (p= 0.04) Sobel test: Goodman (I) test: 2.00 (p= 0.05)

Figure 35

59

~ 0.30 (3.32)

--1

Innovativeness 0.27 (2.86)

/

I Diversif. Product

0.36 (3.81)

'~.26

0.38 (4.15)

I

I

(2.5\)

/

Vertical Integr.

Sobel test:

~

0.26 (2.63)

Perfor-

mance

I

2.19 (p= 0.03)

Goodman (I) test: 2.15 (p= 0.03) Goodman (II) test: 2.24 (p= 0.03)

Mediation Effects (3/3) - Corporate Strategy & Capabilities

The same mediational effect has also been tested between vertical integration and a differentiation or a cost leadership strategy. While, contrary to the cost leadership strategy, the differentiation strategy met the four conditions of mediation, both failed the test for significance, We conclude from this that (as hypothesized), only a combination of both strategy types, such as represented by the balancer strategy, has a significant mediational effect on the relationship between vertical integration and performance.

209

The findings confirm that the effect of both product diversification and vertical integration on performance is partly explained by underlying capabilities such as innovativeness or organizational learning. 5o While the effect of both corporate strategy variables on performance remains significant after controlling for mediation, it is nevertheless advisable to take the underlying capabilities when deploying such a strategy. The probability of success can be increased if the resource profile closely matches the chosen corporate strategy.

Discussion of the Research Findings

The main objective of this section was to confirm the need for a complex perspective that captures the most important interrelations between individual determinants. More advanced LlSREL modeling, based on prior results from the integrated model, led to interesting findings. There is strong evidence that the remaining three theoretical perspectives (competitive strategy, corporate strategy, and capabilities) are a lot more correlated than generally assumed in the literature. Our analysis produced highly Significant covariances between the three latent constructs (t-values between 3.47 and 4.65). Moreover, the manifest variables are strongly correlated and reflect a common underlying construct. A large share of the overall effect on performance is shared between the manifest variables. The model that fits best has a single latent construct, competitive advantage, which is reflected by all the remaining variables. All indicators load significantly on the latent construct competitive advantage (betas above 0.40; t-values above 4.00). Competitive advantage has a particularly strong (13= 0.75) and highly Significant (t-value= 6.86) effect on performance. The construct explains close to 57% of the variance in performance (R2= 0.57). A division of this effect between three latent constructs or distinct perspectives (as assumed in large parts of the literature) weakens the model rather than improving it. We can thus conclude that all variables are so closely interrelated that only a complex model that covers these interrelations can truly explain performance.

60

Further testing revealed a significant mediator effect of innovativeness on product diversification. While not having hypothesized this effect. these findings are anything but surprising as both product diversification and vertical integration can be expected to behave largely similar in this regard.

210

Having confirmed the main hypothesis of the complex model, we turned our attention to individual interrelations between variables from different perspectives. We found proof of the large majority of the hypothesized mediational effects between variables from the competitive strategy, corporate strategy, and capabilities perspectives. Figure 36 shows the individual tests and their results.

Predictor Variable Differentiation Strategy Differentiation Strategy Differentiation Strategy Product Diversification

Mediator Variable

Outcome Variable

Market Orientation

Performance

Innovativeness

Perforrnance Performance

International Diversification Vertical Integration Product Diversification

Organizational Learning Cost Leader Strategy Differentiation Strategy Balancer Strategy Organizational Learning

Vertical Integration Vertical Integration

Innovativeness

Performance

Organizational Learning

Performance

Figure 36

Causal steps mediation test

Performance

Product of coefficients mediation tests

0 0 0 0

0 0 0

0 0 0 0

0 0 0 0

Performance Performance Performance

Complex Model- Hypothesis Testing

The findings indicate that the impact of numerous variables on performance is in reality partially explained by underlying mediator variables. For instance, the strong effect of differentiation strategy on performance is partly explained by underlying capabilities, such

as

market orientation,

innovativeness

or organizational

learning.

While

differentiation strategy has a unique effect on performance (even if we control for mediation), there is an additional shared effect explained by related capabilities. In other words, the positive effect of most variables on performance is partially dependent on the presence of other, related variables. The success of a differentiation strategy becomes a lot more likely if the company simultaneously focuses on its market orientation, innovative ness, and organizational learning. The same mediation effects have been found between corporate strategy and business strategy, as well as between corporate strategy and capabilities. This shows the need for a close reciprocal adjustment of all three levels in order to effectively influence business unit performance. Competitive

211

strategy, corporate strategy, and capabilities do all matter - and they are so closely related that any attempt to artificially separate them is doomed to failure. While the complex model clearly represents a major step forward compared to the integrated perspective, the model has one significant limitation. Findings based on crosssectional analysis are valid only for the actual moment in time that the research is realized. While we did not strive for a truly longitudinal analysis, we nevertheless established a third model that analyzes the development over a certain period in time, the evolutionary model.

V.3

The Evolutionary Model

The lead hypothesis of the evolutionary model is that performance can be best explained by a model that captures the development and evolution over time. The evolutionary model complements the integrated and the complex model through the addition of a longitudinal perspective. First, we analyze the acquired longitudinal data for industrywide Patterns of Dynamic Change (1) in key strategic variables underlying performance. Next, we intend to find Interrelations of Dynamic Changes (2) in the different strategic variables and performance over time. The section ends with the Discussion of the Research Findings (3) of the evolutionary model.

Patterns of Dynamic Change

Up to this point all findings in this study have been based on equilibrium assumptions. This means that all relationships identified above are assumed as generally valid, independent of potential developments or changes over time. In contrast, evolutionary approaches to strategy assume continuous dynamic changes regarding the key strategic variables underlying performance, as well as regarding their effect on performance (D'Aveni 1994, Hamel 2000). Earlier in this study we reviewed alternative patterns of change (see Chapter 111.3). Our first objective will thus be to analyze the acquired longitudinal data for potential patterns of dynamic change. Did the strategic variables change over time? Are there consistent patterns of change on an industry-wide level?

212

This effort prepares the groundwork for the later analysis of the potential consequences of these changes. 61 In order to identify the most important changes, we combined two analyses that reflect distinct dimensions of change, namely the persistency of change and the strength of change. The two dimensions together will then allow the classification of all changes into a common matrix. Persistency of Change: Persistency signifies the sustainability of changes. Is there a

sustained change in one direction for the full period under consideration? Or do firms frequently change direction? In the former case we would talk of a persistent (or sustained) change, in the latter case we would confirm a high volatility of change. In order to measure the persistency of change, we applied the cross-correlations technique in SPSS 11 (Brosius 2002). This method allows correlations between single observations to be measured for the same variable over time. For instance, we measured the change in a company's orientation towards organizational learning on a yearly basis over a period of four years. If we find each observation to be significantly correlated to the subsequent observation, we can assume a certain persistency of change. Besides the correlations to the directly succeeding observations, we can also measure crosscorrelations with larger time lags. The highest persistency can be confirmed, if observations are not only significantly related to directly succeeding, but also to later observations. Figure 37 summarizes the cross-correlations for all thirteen independent variables in our study. The first row of each variable indicates the cross-correlations of each observation with the observation for the following year (e.g. cross-correlation between change in industry concentration for 2001 and 2002). The second row shows the cross-correlations between observations with a time lag of two years (2001 and 2003; 2002 and 2004). Finally, the third row of each variable shows cross-correlations with a time lag of three years (2001 and 2004). The strength of the cross-correlations is measured by their

61

As mentioned before, the acquired data is limited to a four-year period (2001-04). However, due to the dramatic cyclical market changes, this period has been particularly dynamic and challenging for media companies. We are thus confident of observing changes in several strategic variables.

213

respective beta values, whereas the asterisk indicates the significance of the correlation (at p< 0.05). Variable Industry

Concentration

Market Growth Entry Barriers Product

Differentiation

International Diversification

Vertical

Integration

Differentiation

Strategy

Figure 37

2001 0.6S" O.SS" 0.40" 0.70"

2002 0.64"

2003 0.69"

0.57"

x

x x

0.73"

0.64"

0.49" 0.30" 0.64" 0.61" 0.55" 0.56" 0.42" 0.20 0.57" 0.12 0.10 O.4S" 0.25" O.lS 0.52"

0.49"

x

x x

0.77"

0.S2"

0.67"

x

x x

0.59"

0.66"

0.30"

x x

0.17 0.19

x 0.45" 0.27"

x 0.49" 0.32"

0.69"

x x 0.71"

x

x x

0.59"

0.72"

0.44"

x x

x

Variable

2004

x x x x x x x x x x x x x x x x x x x x x

Cost Leader

Strategy

Innovativeness

Market Orientation Organizational Learning

Reputation

Corporate

Culture

2001 0.57" 0.05 0.03 0.6S"

2002 0.51" 0.21"

x

x x

0.60"

0.71"

0.46" 0.3S" 0.74" 0.53" 0.50" 0.S3"

0.41"

x x

0.62" 0.65" 0.70" 0.49" 0.3S" 0.67" 0.35" 0.31"

x 0.66" 0.62"

x 0.6S" 0.69"

2003 0.62"

0.91"

x x O.SS"

x

x x

0.74"

0.79"

O.SS"

x x

x 0.64" 0.44"

x

0.76"

x x

2004

x x x x x x x x x x x x x x x x x x

Cross-Correlations

The first finding from the cross-correlation analysis is that significant correlations have been found for all pairs of observations that immediately follow each other. This shows a certain persistency of change for all thirteen variables. In other words, the orientation of a company towards a certain strategic factor has not completely changed within the space of a single year. While there are minor changes every year62 , the fundamental orientation remains the same. However, the picture becomes slightly different if we consider correlations with a time lag larger than a single year. Based on these results, we can distinguish two groups of variables. The first group is composed of all industry and all firm capability variables. 63 All variables in this group showed significant cross-correlations between observations with a time lag of one, two, and even three years. The fundamental trend of each variable remained

62 63

Only a perfect cross-correlation (beta values of 1.00) would indicate the total absence of any change. In detail: industry concentration, market growth, and entry barriers (all industry); innovativeness, market orientation, organizational learning, reputation, and corporate culture (all capabilities).

214

stable over the entire examined period. For instance, if a firm indicated a strong trend towards innovativeness in 2001, this trend remained generally stable until 2004. We can thus substantiate a rather high persistency (or a low volatility) of changes in both industry and capability variables. The second group comprises all corporate and competitive strategy variables. 64 While all variables showed significant cross-correlations between observations with a time lag of one, and still partially with a time lag of two, they completely failed to do so with a time lag of three. The fundamental trend of these variables had changed over the entire examined period. The results of the analyses with a time lag of two indicate that the most important changes occurred over the first half of the observed period. 65 We can thus state a moderate to low persistency (or moderate to high volatility) of changes in both corporate and competitive strategy variables. Strength of Change: While the persistency of change measures the sustainability of a

trend over time well, it fails to account for the actual strength of these changes. Persistency is only measured if changes remain stable, regardless of the nature of these changes. For instance, a high persistency is allocated if a respondent of our survey indicated the same trend for a given variable four times. However, this trend might be a "very strong" or a "strong" increase of the importance assigned to a given variable, or a "very strong" or "strong" decrease, or even "no change" at all. The actual answer does not matter, as long as it remains roughly the same. The second dimension of change has thus to consider the "strength" of change. In our study, the strength of change can be measured by calculation of the mean values of the acquired data. In our survey questionnaire, we asked each respondent to indicate whether there is a "significant increase" (5), "slight increase" (4), "no change" (3), "slight reduction" (2), or "significant reduction" (1) in the development of each variable. Whereas response "three" indicates no change, all deviations from three indicate a change. The strength of this change depends on the importance of this deviation. If the mean value of 64 65

In detail: product diversification, international diversification, and vertical integration (all corporate strategy); differentiation strategy and cost leadership strategy (both competitive strategy). Three of five correlations for the first half (2001 and 2003) failed to be significant, whereas the results for the second half (2002 and 2004) are all significantly correlated.

215

all responses for a given variable is close to three (i.e. between 2.5 and 3.5), there are only weak changes at the industry level. However, if mean values are either lower (i.e. below 2.5) or higher (i.e. above 3.5), there is at least a moderate change at the industry level. Below or above a certain value (i.e. below 2.0 or above 4.0), we could even speak of a strong industry-wide change. Figure 38 shows the mean values for each observation along the thirteen examined variables. The results show only weak changes for all industry (means between 2.52 and 3.52) and corporate strategy (means between 3.23 and 3.51) variables. Two capability variables also exhibited only weak changes, namely corporate culture and reputation (means between 3.26 and 3.57). The remaining capability variables (innovativeness, market orientation, and organizational learning), as well one competitive strategy variable (differentiation strategy) indicated at least a moderate change (means between 3.33 and 3.98). A single variable, cost leadership strategy, showed a particularly high level of change (means between 3.92 and 4.36). Variable

2001

2002

Industry Concentration

3.52

3.42

3.47

3.47

Market Growth

2.52

2.69

2.56

3.37

Entry Barriers

3.07

3.17

3.21

3.21

Product Diversification

3.51

3.45

3.35

3.44

International Diversification

3.39

3.29

3.23

3.40

Vertical Integration

3.27

3.26

3.38

3.36

Differentiation Strategy

3.61

3.60

3.65

3.77

Cost Leadership Strategy

3.92

4.30

4.36

4.09

Innovativeness

3.33

3.59

3.79

3.98

Market Orientation

3.46

3.69

3.90

3.91

Organizational Learning

3.39

3.53

3.68

3.72

Reputation

3.36

3.41

3.55

3.57

Corporate Culture

3.31

3.26

3.40

3.41

Figure 38

Changes in Mean Values

216

2003

2004

In general, most variables have been found to indicate a weak to moderate level of change over the examined period. There is a moderate trend in favor of four variables that have been found to be highly correlated earlier in our study, namely differentiation strategy and the underlying capabilities innovativeness, market orientation, and organizational strategy. However, the strongest trend at the industry level has been found for a cost leadership strategy. The Change Matrix: The combination of our prior results of both the persistency and the

strength of change allows us to classify the thirteen variables according to their specific patterns of change. Along the two dimensions, we can distinguish four patterns of change. The first pattern combines a high persistency with a low strength of change. In this case, we can talk of equilibrium conditions. The low strength indicates the absence of any significant changes; the high persistency confirms the stability of that condition over time. The second pattern combines a high persistency with a moderate to high strength of change. This pattern is best described as evolutionary change. There are significant changes, and they remain highly stable over time in both strength and direction. The third pattern is defined by a low persistency and a moderate to high strength of change. We called this pattern "revolutionary" change. There are significant changes, however, they change or perish as quickly as they occur. Even within a short timeframe, both strength and direction of change vary. The fourth pattern is characterized by both a low persistency and a low strength of change. We labeled this pattern as "punctuated" change. While generally stable, there are individual "outbursts" of change. However, these changes are punctual rather than sustainable over time. Figure 39 shows the patterns of change resulting from our prior observations of the thirteen examined variables. The strength of change has been measured by the average deviation from the middle value three (or "no change") over a four-year period. The

217

consistency of change has been measured by the average beta value for all six crosscorrelations of a variable. 66

1.10

J:

S! J:

1.00 0.90

QI

0.80

Ol

I: 1\1

0.70

.r:

... 0

0.60

0

.r: C,

0.50

2!

0.40

t /)

0.30

I:

•

3:

...J

(

•

•

REPUTAT'~

~ MARKGROW

____

~'

IUilibr~

EqU"~:J

0.10 0.00

~-.....-----.-""

+--~~-~--~--~--+---~-

0.25

U.3

0.35

0.4

LOW

Figure 39

IFF~·

p~ ~"""

0.20

0

PRO

0.45

0.5

_ _ _~_ _~_ _--j

0.55

Persistency of Change

0.6

0.65

0.7

075

HIGH

Patterns of Dynamic Change

The change matrix reveals a number of interesting findings. First of all, we found all three industry variables, as well as corporate culture and reputation as remaining in equilibrium. This is largely consistent with the underlying theoretical assumptions. Industry concentration and entry barriers, as well as corporate culture and reputation, are assumed as rather stable over time. Changes are mostly gradual and take a very long time. Our findings for these four factors make perfectly sense. However, the fifth variable "market growth" shows the limitations of a longitudinal study that is restricted to four years. While this variable remained stable over the past four years (a slight negative growth all along), the findings would certainly have been different if we had considered a longer period (Le. taking the high market growth during the 1990s into consideration).

68

Three cross-correlations with a time lag of one year (2001 & 2002, 2002 & 2003, and 2003 & 2004), two with a time lag of two years (2001 & 2003 and 2002 & 2004), and one with a time lag of three years (2001 & 2004).

218

The second cluster comprises the three corporate strategy variables. While they did not show a significant change at the overall level, there were frequent minor changes in both strength and direction, indicating moves in and out of different corporate strategies. This change pattern fits well with corporate strategy, as all three forms of diversification are mostly characterized by punctual investments (e.g., acquisitions or new ventures). For instance, a company increases its vertical integration through the acquisition of a supplier. These changes are punctual rather than gradual. The overall low level of change might reflect the already high degree of diversification in the media industries. The third cluster consists of both competitive strategy variables. A moderate to strong trend towards both strategy types has been found. The trend towards a cost leadership strategy highly exceeded the change towards a differentiation strategy. However, the low persistency of these changes (even over a short time period) suggests that these trends might be rather short-lived. The peak of the cost leadership trend had already been reached in 2003, while the trend towards differentiation strategy continued to rise. Competitive strategy seems to be characterized by short-term revolutionary changes with rather low sustainability over time. This is consistent with theoretical arguments that assume a high degree of dynamism at the business-unit level of strategizing. Finally, the fourth cluster consists of the three dynamic capability variables. There is a moderate to strong trend towards innovativeness, market orientation, and organizational learning. While this trend is less strong than the trend towards a cost leadership strategy, it seems to be a lot more persistent over time. The upward trend persisted throughout 2004, the final year of observation. Our overall findings thus far indicate a high degree of stability of the examined variables over the four years under examination. However, significant changes have been identified in a number of variables. Some of these changes seem to be rather important for a short time period (Le. changes in the competitive strategy variables), while others might indicate a more fundamental and sustainable change in the marketplace (Le. changes in the dynamic capability variables). From a theoretical perspective, it is interesting that variables underlying a common construct (Le. corporate strategy) showed similar patterns of dynamic change.

219

Interrelations of Dynamic Changes

Earlier in this chapter, we applied the integrated model to an empirical study of the global media industry. Our analysis revealed significant effects of a number of determinants on business unit performance. In the present section of this chapter, we found a number of trends towards some of the very same determinants in the media industries over the past four years. Did these trends have the desired effect on performance? For instance, the integrated model revealed a strong positive effect of innovativeness on performance. In this section, we found a strong and persistent trend towards innovativeness by media companies. Did this orientation have the expected positive effect on performance? As mentioned before, our research approach for the evolutionary model is rather exploratory than confirmatory. Instead of testing hypotheses, we conduct a significant amount of data analYSis and focus on the interpretation of our findings. However, this effort is based on the theoretical foundation of the integrated model assuming significant effects of the very same variables on performance. We applied regression analysis (based on SPSS 11) in order to test the effect of changes in single determinants on performance. Similar to the integrated model, we applied both single and multiple regression at the submodellevel. In addition, due to the strong collinearity of the five firm capabilities, we used Structural Equations Modeling based on LlSREL 8 for this submodel. Figure 40 summarizes the findings of all four years. The first check mark indicates a significant and positive effect in single regression analysis, the second in multiple regression analYSis and I or structural equations modeling. The detailed results can be found in Appendix 2 (figures 1-4). The findings show that a determinant's effect on performance has generally been consistent over time. Beside some smaller deviations, most findings remained stable (either significant or insignificant) over the full period under consideration. Moreover, most deviations occurred in the final year. These inconsistencies may be attributed to the fact that the data for 2004 has been based on the respondent's expectations for the future. The data may therefore be less reliable than for the three time periods covering past events. Apart from these exceptions, the findings confirm our earlier conclusion of

220

the high stability of determinants and their relationships over time (at least over a fouryear period). Variable Industry Concentration Market Growth

2001

2002

2003

2004

0'0' 0'0'

0' 0'0'

0'0' 0'0'

0' 0'0'

0'0'

0'0'

0'0'

0' 0'0'

0'0'

0'0'

0'0'

0'0' 0'0' 0'0' 0'0' 0'0'

0'0' 0'0' 0'0' 0'0' 0'0'

0'0' 0'0' 0'0' 0'0' 0'0'

0'0' 0' 0'0' 0' 0'0'

Entry Barriers Product Diversification International Diversification Vertical Integration Differentiation Strategy Cost Leadership Strategy Innovative ness Market Orientation Organizational Learning Reputation Corporate Culture

Figure 40

Regression Effects on Performance

In addition, the most significant findings had been repeated with performance data for the following period. Changes in important determinants did not only affect performance in the same, but also in the following year.67 These findings further support the (punctuated) equilibrium theory and the assumption of a certain stability over time of most developments that underlie performance. The individual results themselves include both expected and somewhat surprising findings. First, there are several results that mirror our earlier findings in respect of the integrated model. Much as before, a stronger focus on each of the five firm capabilities did lead to superior performance. The same is true for product diversification and a differentiation strategy. And, much as before, both entry barriers and international

221

diversification failed to show a significant effect on performance. While these findings confirm our expectations, there are some less obvious results. Moves into both vertical integration and cost leadership strategies failed to show a significant effect on performance. While both determinants have also been among the weakest in the integrated model, they nevertheless had a significant effect on performance. Even more surprising, changes in industry concentration and market growth revealed a significant effect on performance. Both determinants failed to do so in the integrated model. How can these differences be explained? We can shed some light on these findings, if we consider our earlier analysis of the patterns of change that occurred over the past four years. Figure 41 resumes the change matrix established above. It can be seen that all determinants found to be relevant in explaining changes in performance are located in the upper right hand square of the matrix. Conversely, all determinants outside this zone failed to significantly affect performance. These findings point to the conclusion that changes have to be both strong and persistent in order to have a significant impact on performance. Nor did high strength (Le. cost leadership) nor high persistency (Le. entry barriers) alone prove to be sufficient. Also a combination of low strength and low persistency (Le. international diversification, vertical integration) failed to show an effect. Only a strong and persistent firm orientation towards a determinant seemed to payoff. While missing strength or consistency of changes might well explain the failure of vertical integration and cost leadership to significantly impact performance, it only partially explains the sudden importance of industry concentration and market growth. The results so far suggest that while industry concentration and market growth generally do not affect performance, strong and persistent changes in these factors might indeed have an effect. In other words, it does not make a difference if a firm operates in a high growth or a mature market; however, significant changes in the growth rate within these markets do have an effect.

67

This was true for all findings but with two or three exceptions. For some determinants (i.e. innovativeness, organizational learning), changes did not only affect performance in the same and the following year, but even at higher time lags (two and three years later).

222

::E: C)

i: I» CI

C

-tV

1,10

I

0,90

~

0,80

J

070

J

~

U

I I I I I I I I

COSTLEA

I

1,00

I

~

en

• INN OVA

DIFFSTR

ORGLEAR •

\

0,50

CI

C

•

\

0,60 -j

0

~

• MAAKORT

\

0,40

' , . PROD IFF

j I

• INTO IV.

0,30 -

VERTINT"".....

!

0,20 -

3: 0

..J

0,10

•

REPUTAT I

----------------1 •

CORPCUL

J

ENTBARR

o,ooL 0,25

LOW

Figure 41

•

• MARKGROW

0,3

0,35

0,4

0,45

0,5

0,55

Persistency of Change

0,6

0,65

0,7

0,75

HIGH

Impact of Changes on Performance

While this finding might be surprising at first, it is largely in line with prior theorizing. It has been argued that growth markets represent both opportunities and risks (Aaker I Day 1986). While there are more growth opportunities and less price pressure, growth markets also attract more competition and require higher investments. In the long run, these advantages and shortcomings might cancel each other out. This would explain the failure of market growth to show a significant effect on performance in the integrated model. However, in the short run, the effect might be different. If there is sudden growth, it will take time until new competitors enter the market. Moreover, competitors will only enter the market if the growth appears to be sustainable. For the time being, the companies already present in the market will benefit from the increased market growth. In fact, if market growth suddenly drops, competitors are highly unlikely to immediately leave the market (Le. due to exit barriers). Furthermore, it will take time to adjust the current investments to the lower market demand. These effects will certainly have a negative short-term effect on performance. The example shows that the effect of market growth

223

might be very different if we talk about phases of short-term dynamic change or longterm equilibrium periods. 68

Discussion of the Research Findings

While we did not aim to "test" or "prove" any theory in this section, we can confirm four important findings. First, we identified consistent patterns of change at the industry level for all determinants. Second, the strongest and most consistent changes (and only those) did have a significant effect on performance. Third, most determinants' effects on performance remained stable over a four-year period. Finally, results from dynamic analysis did in some cases depart from prior equilibrium-based findings. These cases indicate that the effect of some determinants during short phases of dynamic change might be different from their effect in long-term equilibrium periods. Our findings provide strong support for our overall research framework and the underlying assumptions. The finding of relative stability of effects on performance over a four-year period provides justification for equilibrium-based theorizing, such as in the two first models in our framework. The findings of characteristic patterns of change and their effects on performance show the importance of complementary evolutionary analysis.

V.4

Implications

The value of a study is largely determined by its contribution in both informing practice and advancing theory. Drawing on the diverse research findings presented above, some important implications for managers and academics are presented below.

68

We do not further discuss industry concentration in this section. This is due to the fact that the results for this factor are less consistent than those for market growth. However, a similar explanation can be provided for industry concentration. In the short run, an increase in market concentration leads to higher performance due to economies of scale and collusion. Whether this advantage can be sustained over time depends on the presence of high barriers of entry. If, as in our study, entry barriers are rather weak, there won't be a significant effect in the long run (Bain 1956, Mann 1966, Scherer I Ross 1980, Jarillo 2003).

224

Managerial Implications

The fifth and final research objective of the present study was to gain a comprehensive view of determinants underlying firm performance in the global media industries. Application of the developed dynamic research framework to the media industries has delivered a multitude of findings. The objective of this section will be to extract a number of essential implications for media practitioners. In total, we derived seven postulates from our research findings that are presented below. 1.

There are no simple success strategies!

The findings from our study draw a far more complex picture of strategy and performance in the media industries than most of the prior theorizing did. The good news is that companies determine their own destiny to a large extent. External factors, such as industry concentration, entry barriers or market growth, played only a minor role. It is a firm's strategy, both at the corporate and the business unit level, as well as its unique capabilities that made the difference. The challenge is that corporate strategies, business strategies, and capabilities are highly interrelated. Not only that all three are relevant, but they need to be combined in the right manner to be successful. Only the right combination leads to success. Even worse, what constitutes the right combination might change over time. Despite this complex and dynamic character, there are some fundamental "rules of the game" that have to be considered. The seven postulates in the present chapter strive to capture these rules. 2.

Capabilities make the difference!

Firm capabilities have been found to be highly related to firm performance, explaining a large share of the variance in the latter. Furthermore, they are the basis for all the firm's strategic activities. Strategic initiatives have a significantly higher effect on performance if the related capabilities are present. Media firms need a high degree of innovativeness, market orientation, and organizational learning. Furthermore, the firm's reputation is important, as well as an entrepreneurial corporate culture. Figure 42 relates an example from the German broadcasting industry, illustrating the importance of capabilities such as innovativeness in the media industry. Prior studies have stressed capabilities as a key

225

source of performance for the media industry (Le. Lopes 1992, Miller / Shamsie 1996, Habann 1999). Much as in other countries, the German Broadcasting industry is characterized by a high degree of imitation and replication. Three major trends in the German television over the last decade, namely daily soaps, quiz shows, and casting shows, serve as an example. The first daily soap Gute Zeiten, Schlechte Zeiten (GZSZ) was introduced in the early 1990s. Due to its success, most of the other channels launched similar programs. It is interesting that GZSZ has remained the market leader to this day. Early imitations, such as Verbotene Liebe or Marienhof have had Significantly less success, while later imitations failed completely and were abandoned. The same applies to both quiz shows and casting shows. The innovators Who wants to be a millionaire? or Deutschland sucht den Superstar (DSDS) remain the market leaders despite countless imitations on competing channels. Even more interesting is the fact that all three innovations mentioned above were produced by the same channel. RTL, a subsidiary of the German media giant Berlelsmann, has frequently been the first to introduce new television formats. The market leader in terms of viewers, RTL seems to satisfy customers' preferences. The development in innovativeness and market orientation translates into higher performance. RTL has always been the most profitable player in the market. While we have chosen the German market as an example, very much the same is true for other markets. BSkyB in the United Kingdom or HBO in the United States are, for example, considered leaders in terms of both innovativeness and performance.

Figure 42

3.

Innovativeness in the German Broadcasting Industry

Differentiation strategy is essential to leverage capabilities!

Differentiation strategy has been found to play an important role in the media industry. Companies that followed a differentiation strategy showed a significantly higher performance. Even more important, a differentiation strategy enables the firm to benefit from its existing resource base. Differentiation strategy and dynamic capabilities complement each

other and

have a common

effect on

performance.

While

innovativeness, market orientation, and organizational learning represent the potential for success, a differentiation strategy is the means with which to realize this potential. This finding is particularly interesting, as it is contrary to the behavior of many media companies. As mentioned above, replication and imitation are far more common than true innovation. The products in most markets are more similar than truly differentiated. Most companies "provide an illusion of differentiation by varying small, nonessential aspects of a producf' (Sherman 1995, 55). For instance, a study of the liberalization of

the Taiwanese television market revealed that competition led to less programming

226

diversity. The networks became more conservative and duplicated one another's programs (Li I Chang 2001). The same has been found in respect of the US video market: competition increased, but the content range offered has not expanded. Customer choice remained the same (Hellman I Soramaki 1994). Our findings confirm earlier studies claiming a critical role for differentiation in the media industry (Sherman 1995, 53). Figure 43 illustrates our findings with the example of the British Pay TV provider BSkyB. BSkyB, part of Rupert Murdoch's News Corporation, is the market leader in the British Pay TV industry. Contrary to other players in the same industry, BSkyB is a true success story. Despite the general crisis in the media industry, BSkyB has its increased Total Return to Shareholder by 62% over the past five years. At the same time, the British index FTSE 100 went down by 27% and the European media index MSCI EMU by as much as 60%. The company is highly profitable and growing steadily. BSkyB explains much of its success through the so-called ABC Strategy. ABC is short for Audience Brand Content. "Audience" represents BSkyB's market and customer orientation. The award-winning customer service and the cutting-edge customer relationship management (CRM) systems explain the particularly low churn rate. "Brand" refers to BSkyB's marketing power and reputation. Second only to the public broadcaster BBC, the company enjoys a brand awareness and reputation that is largely superior to that of its peers. Finally, "contenf' indicates the wide variety and programming choice. BSkyB has an unrivaled content including SkyNews, the European leader in TV news, and SkySports, the market leader in sports. The channels are permanently adding new and innovative services. Large investments have been made in electronic learning solutions to further boost innovativeness. BSkyB's ABC Strategy is an excellent example of a successful differentiation strategy. Strong capabilities, such as market orientation, innovativeness, reputation, and organizational learning are leveraged through a consistent strategy that is directed towards a clearly differentiated range of products that meet customers' interest.

Figure 43

4.

The ABC Strategy of BSkyB

Diversification and vertical integration make sense - If managed correctly!

Both product diversification and vertical integration have been found to be positively related to performance. This confirms former studies that claim a beneficial role for diversification and vertical integration in media companies (Le. Gomery 1993, Sjurts 1996, Wirtz 2000, and Bauder 2002), Diversification allows the multiple deployment of expensive content and enables cross-promotion (Wirtz 2000, 252). Vertical integration leads to reduced cost of sales and increased market access (Gomery 1993).

227

Due to the record industry's panic about illegal file-sharing over the Internet. another major trend of the past years has gained much less attention: the increasing convergence between music and television. It has become increasingly difficult to score a hit without significant television support. More and more hit singles are promoted by casting shows. soaps or serials. For instance, in January 2004. five of the Top 10 singles on the German charts were produced by television stars. Television has become the record industry's most important channel for customer contact. Independent record labels. such as the British major EMI. have to transfer between 20 and 50% of their earnings to television channels in order to promote their products (Der Spiegel 8 I 2004). Both sides are considering further diversification to benefit from this development. For instance, the record majors Universal, Wamer. and BMG have already established TV divisions in order to enter the market. The German television channel RTL is considering the launch of its own record label. despite its close relationship with BMG (both are owned by Bertelsmann). No matter which of the two companies ends up in dominating the relationship. success in the record industry of the future requires a diversified company spanning both record production and television. However. despite the relatively young trend. there are already examples that show the important influence of innovativeness and differentiation on the success of a diversification strategy. Record companies that invested in the stars of the fourth season of the reality show Big Brother failed to repeat the success of earlier seasons. The market had moved on to new formats.

Figure 44

Diversification Makes the Superstar

Further analysis revealed that the success of both diversification and vertical integration is largely dependent on the underlying capabilities and the competitive strategies applied. A high degree of innovativeness and organizational learning significantly increases the effect of both corporate strategies. A competitive strategy that focuses on both costs and a differentiated service offering further affects the impact of diversification and vertical integration on performance. These findings make sense, as diversified and integrated companies have better access to attractive and innovative products. A conglomerate spanning both TV channels and content production companies spreads the risk of innovating new products (Gomery 1993. 66) and makes the channels independent of the increasingly competitive content market (Wirtz 2000, 263). The diversification strategy pays off if two conditions are met. First, additional dynamic capabilities have to be acquired, increasing the firm's current resource base. Second, a competitive strategy has to be implemented that monitors the costs of diversification (or vertical integration) while deploying new capabilities through an increased differentiation.

228

5.

Attention to the bottom line has to be institutionalized!

Our research has shown that a cost leadership strategy has a significant positive effect on performance. Companies that manage their bottom line well showed superior performance. This finding is once again largely contrary to real life strategic behavior in the media industries for much of the last decade. Particularly during the boom phase between 1995 and 2000, media business strategy was dominated by a focus on revenue growth and expansion with much less attention being spent on the establishment of costefficient procedures. For instance, no more than two of the more than 25 channels in the German broadcasting industry showed a profit during the nineties. However, with the downturn at the turn of the millennium, the media companies' strategic behavior suddenly changed. During the examined period (2001-2004), a strong trend towards cost cutting and controlling has been apparent. For instance, two leading German broadcasters, Pro7Sat1 AG and Premiere, have severely cut costs and streamlined operations. After long years of steady losses, both companies have achieved a turnaround. The British publishing company McGraw-Hili is truly an example of an enduring success story. McGraw-Hili has paid dividends without fail since 1937, and 2004 marks the 31 st consecutive year of increased dividends. The company enjoys a strong balance sheet and an excellent free cash flow. Since 1995, shareholder return has averaged 18% annually, largely exceeding the performance of the S&P 500 index, which was up 6.9% for the same period. McGraw-Hill's sustainable performance has been strongly supported by a consistent and balanced competitive strategy. The company has a record for sustained growth in operating earnings and stringent cost control is emphasized in order to improve operating leverage. Selling, general administrative and other expenses (SG&A) are far below that of the competition. Permanent cost initiatives have led to steady margin improvements. The margins of the Education business have, for example, doubled since the 1996 acquisition of the Times Mirror. However, the cost focus is just one side of the story. On the reverse side, McGraw-Hili has also realized steady growth. 2004 is the 13th consecutive year of revenue growth, despite the recent market downturn. Diversity of services has been key to that growth. McGraw-Hili has the broadest product lineup for the core education business in the industry. Outside the core business, the company permanently pursues and grows new opportunities. The Financial Services business, specifically, shows continued double-digit growth while maintaining operating margins at the mid-30% level. The growth is achieved by leveraging the company's strategic capabilities. McGraw-Hili is a brand leader (according to Standard&Poors), a pioneer in technological innovation (particularly in digital services), and invests heavily in customer service. Active portfolio management not only improves margins, but adds new businesses with complementary capabilities to existing operations.

Figure 45

McGraw-Hill's Balanced Strategy

229

Two additional findings from our research related to the cost issues are particularly interesting. First, we found that the trend towards cost leadership in the media industry is rather short-lived and unstable. Moreover, a cost leadership strategy has only been found to be successful when consistently applied over a longer period. In other words, media companies need to institutionalize their cost commitment in order to guarantee long-term performance. Short-term consolidation efforts, even the most radical ones, did not show a significant and sustainable effect. Second, we found that top performers realized a combination of both a cost leadership and a differentiation strategy. A balanced strategy that accounts for both costs and creative and differentiated product offerings leads to the highest performance. Figure 45 illustrates such a balancer strategy through the example of the British publisher McGraw-HilI.

6. Consolidation and scale are overrated! In our study, industry concentration did not show any significant effect on performance. No evidence has been found that concentration or scope makes a difference. While we do not claim that size does not matter, our findings indicate that its importance for the media industry is often overrated. The one-sided obsession with size in the media industry often leads to a disregard of other, more important determinants of firm performance. This finding is particularly surprising. Market consolidation has been the single most important trend in the media industry over the last decade (Picard 1989, Bauder 2002). Concentration has reached very high levels (Albarran 1996, Sjurts 1996). Economies of scale have been traditionally regarded as particularly important in the media industries (Le. Owen et a11974, Cave 1989, Kruse 2000, and Wirtz 2000). However, due to the already high industry concentration, the unbroken quest for size in the media industry has been increasingly questioned. Growing fragmentation of media audiences makes it increasingly difficult to realize scale economies (Kruse 2000). In addition, studies revealed that large media conglomerates quickly lose control of their disparate operations (Demers 1999). Real life observation of the media industry also indicates the negative effect of size on creativity. The market leaders in most industries are strongly dependent on smaller and more creative providers. Examples are the

230

success of the animated movie producer Pixar (versus Disney), or of small independent record labels (versus the four majors). Peter Chernin, President of News Corporation, has pOinted to the drawbacks to size: "All the benefits of size, whether it's leverage,

synergy, or scope, are fundamentally the enemies of creativity" (The Economist, 23/05/2002). Further analysis will be required to untangle the interrelations between size and dynamic capabilities such as innovativeness or creativity. However, one conclusion can already be drawn: While market consolidation or scale economies may playa certain role (at least in some industry segments), they certainly fail to provide a universal strategy for success in the media industries.

7.

Change is inevitable, but it can be managed!

Our additional longitudinal analysis revealed a number of significant changes in the media industries over the past years. The most important trend was towards a more cost-efficient strategic behavior. However, despite these changes, we found a strong stability regarding effects on performance. Most of the factors that were important to explain performance remained relevant over the full period under consideration. In other words, at least for a four-year period (and a particularly troubled one at that), the seven postulates outlined in this section retained their validity. This mid-term stability allows managers to develop and implement consistent strategic initiatives. Although changes have to be carefully monitored and programs have to be adjusted, the fundamental orientation required for success in the media industry remains intact.

Research Implications

The focus and motivation of this study has been to move beyond simplistic and onedimensional success strategies through the development of a more integrative, complex, and dynamic approach to strategic analysis. As may be seen in the previous sections of this chapter, we did find empirical evidence for the majority of our theoretical research hypotheses. What are the implications of these findings for strategy research? First, and foremost, the integrated model has clearly shown that corporate strategy, business strategy, and firm capability variables are all important in explaining firm performance. This is very much in line with prior findings from variance decomposition 231

studies (Le. Roquebert et al 1996, McGahan / Porter 1997, Chang / Singh 2000). Further research should strive to tear down the walls between different perspectives and end the current fragmentation into separate research streams. The complex model has provided further evidence that the different perspectives are too closely related to justify any further fragmentation. On the contrary, the results of the complex model strongly recommend further research into the various interrelations between the different strategy perspectives. The single determinants of business-unit performance are highly interrelated and often have a combined effect on performance. Our findings strongly confirm Porter's (1991, 114) claim that success in dynamic industries requires the interaction of favorable conditions in several of the determinants of performance. The effect of single factors on performance can no longer be ascertained without complementary analysis of their complex relationships with third factors. In terms of dynamism, we found evidence of a certain stability of the most important determinants over time. This finding provides some support for the validity of equilibriumbased theorizing, at least from a short- to mid-term perspective. However, even from a mid-term perspective, some important developments have been identified. There are differences in the effects of certain variables that are dependent on their strength and consistency over time. Furthermore, there seem to be trends towards and away from different determinants of performance. In future there should be more research that combines equilibrium and longitudinal research. In line with prior theorizing (Le. Porter 1991, Black / Farias 2000), we found both perspectives to be complementary rather than exhaustive. Establishing the frame of reference has shown the importance of more fundamental theorizing in strategic management. Without a common theoretical ground, crossfertilization between different strategy perspectives cannot occur. The frame of reference establishes such a common theoretical basis for the first time. Despite its inherent complexity, the frame of reference is flexible enough to provide the groundwork for a multitude of alternative complex and dynamic models.

232

The empirical study has shown that even more complex and dynamic models can be applied and tested. More sophisticated methods, such as multilevel research, longitudinal analysis, or structural equations modeling, have rarely been applied to strategy research. 69 In future, these (and even more complicated) methods will be required to further advance strategy theory. With growing complexity (such as in our study), research will have to rely more on primary data collected in field studies than is currently the case. This requires more resources and time; however, the present study shows that such field studies can indeed provide rich results at an acceptable "cost". Finally, the selection of a dynamic setting, such as the media industry in our empirical study, is contradictory to the strong focus on stable industries, such as manufacturing, in prior strategy research. Our findings provide new insights into numerous subjects that are partially contradictory to findings from stable industry environments. Stable industry conditions are increasingly replaced by more complex and dynamic settings (McNamara et al 2003). More research is required into dynamic industries to better understand the more complex realities faced by managers every day. Strategy theory cannot be limited to simple and convenient conditions that are less and less the standard across most industries. In summary, this chapter outlined the empirical findings of the established dynamic research framework and summarized the most important implications for both practitioners and researchers. The following final section of this study briefly summarizes the overall findings of the study and discusses the limitations of the study, as well as recommendations for future research.

69

There are a number of exceptions including Melin (1992), Craig (1996). and Webb I Pettigrew (1999). but they are marginal in comparison to the multitude of studies applying standard regression analysis.

233

VI.

Conclusions

The final chapter presents a recap of this study's results together with its associated contributions. Initially, the study's different findings are summarized and discussed, after which we briefly discuss its inherent limitations. The next section provides some directions for future research. A brief set of concluding remarks is presented in the last section.

VI. 1

Summary and Implications

The research objective of this study was to develop and empirically validate a comprehensive research framework that bridges rival perspectives and promotes a state-of-the-art understanding of factors underlying corporate performance. The first important step towards this objective was the development of a common frame of reference that harmonized hitherto conflicting theoretical assumptions from different perspectives. Based on this foundation, we proposed a framework for dynamic strategic analysis embracing three sequential models. The integrated model (1) addresses the problem of fragmentation through the integration of the most important competing perspectives in strategic theorizing. The complex model (2) extends the initial model by capturing the most relevant interrelations between the different perspectives. Finally, the evolutionary model (3) overcomes the static character of prior models dealing with some

of the dynamic evolution over time. Having established the dynamic research framework, we attempted to empirically test the proposed models in the context of the global media industries. A mail survey had been sent out to an area sample of 671 key informants each representing a single SBU of a media company. In total, we included 104 responses in our analysis; these are more than sixteen percent of all media companies in the targeted industry segments. No significant non-response bias has been found.

235

Data analysis revealed sufficient reliability with item-to-total correlations greater than 0.4 and Cronbach Alphas greater than 0.7 for all multiple-item constructs in the study. Convergent validity analysis revealed good validity for all multiple-item constructs based on various measures of absolute and relative fit. Content and discriminant validity have been confirmed. Both standard regression analyses based on SPSS 11 and structural equations modeling based on LlSREL 8.3 were then applied to test the established models and hypotheses.

Results from the Integrated Model

The main hypothesis of the integrated model claimed that a combination of different perspectives on strategic management is necessary to explain company performance. On a more concrete level, hypotheses had been established on the significant effect of a number of individual determinants (derived from the different perspectives) on business unit performance. Subsequently, three statistical methods (simple regression, multiple regression, structural equations) were applied in order to test these hypotheses. All three tests by and large led to the same results. The three industry variables (industry concentration, market growth, and entry barriers) and one corporate strategy indicator (international diversification) failed to prove their importance in explaining performance and were subsequently excluded from the model. In contrast, all ten the remaining variables repeatedly confirmed their significant effect on performance. The remaining two corporate strategy variables (product diversification and vertical integration) had both a Significant direct impact on performance and reflected their underlying latent construct of corporate strategy well. The same significant effect was found for all three competitive strategy variables (differentiation, cost leadership, and balancer strategy).70 Finally, all five firm capability variables (innovativeness, market orientation, organizational learning, reputation, and corporate culture) reflected the underlying latent construct of firm capabilities well. While they indicated a strong direct effect on performance in simple regression analysis, the more sophisticated multiple regression analysis revealed that all

70

The "balancer strategy" variable was measured indirectly through a composite of the two other strategy types and thus does not show in the final integrated model as shown in Figure 46.

236

five variables are strongly correlated and share a common effect on performance. The detailed results are presented in Figure 46.

--1 (g:~~) --1 (g:~~)

--1 (g~~) --1 (~~~)

Product

Diversific.

Vertical

Inteoration

Otfferen~. Strateg

Cost Lead Strategy

I- (g:~~) I- &~g)

0.53 (2.94)

I- (g:~g) I- &~~)

~0.27

0.85 (._-)

~ (2.73)

0.68 ~ 0.54 (5.92) ~ (5.48) 0.51 (4.64) ~ 0.74 0.49

(4.68)

~(6.51)

Chi-Square= 123.59, dt= 49, P-value= 0.00, CFI= 0.82, IFI = 0.82, GFI= 0.83, SI. RMR= 0.18

Figure 46

The Integrated Model

The results show that three of the four strategy perspectives confirmed their important role. Firm capabilities, corporate strategy, and competitive strategy have all been found to be important in explaining performance. Only the integrated perspective manages to consider all three effects simultaneously. However, while an integrated perspective clearly represents a step forward, the model revealed limitations. LlSREL output indicated that a model that takes the interrelations between the different perspectives into consideration might lead to far better results. This leads us to the complex model.

Results from the Complex Model

The lead hypothesis of the complex model is that performance can be better explained by a model that captures the most important interrelations between determinants of different perspectives. The complex model is thus an extension of the integrated model,

237

abolishing

the

assumption

that there

are

no

interactions

between

individual

determinants. More advanced LlSREL modeling, based on the prior findings from the integrated model provided interesting insights. There is strong evidence that the remaining three theoretical perspectives (competitive strategy, corporate strategy, and capabilities) are far more interrelated than generally assumed in the literature. Our analysis produced highly significant covariances between the three latent constructs (t-values between 3.47 and 4.65). Moreover, the manifest variables are strongly correlated and reflect a common underlying construct. A large share of the overall effect on performance is divided between the different variables. The results suggest that distinguishing between the three theoretical constructs does not make sense. All remaining determinants seem to represent a single common underlying construct that we refer to as competitive advantage (see Figure 47). 0.74 ;(6.89) -0.26 (-3.60)

0.51 (5.11)

0.84 (7.00)

___ 0.62 --I -0.25 (6.07)"1

(-3.96)

-0.13 (-2.51)

4

Balancer Strategy

L

r

0.29 (3.90) 0.62 (6.67) 0.53 (5.50)

.

-1

Organizat. Learnin

0.23 0.22 3.09) ~ (2.70) :> 0.53 Reputation 0.30 (6.34). .

"- (~~~)

0.23 (2.90)

04 (4:02) 0 61 (5.12) _~_____ 0.75 (6.86)

0.84 (---) 0.62 (6.30) 0.69 (6.21)

0.48 (5.67) 0.58 (5.71)

0.67 (6.50)

0.69 (7.10) -047 (4' 67) .

~ 0.78 (6.88)

Chi-Square= 42.68, df= 36, P-value= 0.21, CFI= 0.98, IFI = 0.98, GFI= 0.93, St. RMR= 0.054

Figure 47

The Complex Model

All indicators load significantly on the latent construct competitive advantage (betas above 0.40; t-values above 4.00). Competitive advantage has a particularly strong

(13=

0.75) and highly significant (t-value= 6.86) effect on performance. The construct explains

238

close to 57% of the variance in performance (R2= 0.57). A division of this effect into three latent constructs, or distinct perspectives (as assumed in large parts of the literature), depreciates the model rather than improving it. We conclude that all variables are so closely interrelated that only a complex model that covers these interrelations can truly explain performance. Having confirmed the main hypothesis of the complex model, we turned our attention to individual interrelations between variables from different perspectives. The findings indicate that the effect of various variables on performance is in reality partially explained by other underlying variables. For instance, the strong effect of differentiation strategy on performance is partly explained by underlying capabilities such as market orientation, innovativeness or organizational learning. 71 While both differentiation strategy and the capabilities have a unique effect on performance, there is an additional shared effect. In other words, the positive effect of most variables on performance is partially dependent on the presence of other, related variables. The success of a differentiation strategy becomes a lot more likely if the company simultaneously focuses on its market orientation, innovativeness, and organizational learning. The same mediation effects have been found between corporate strategy and business strategy, as well as between corporate strategy and capabilities. In detail: we found a strong mediation effect between vertical integration and a balancer strategy, product diversification and organizational learning, and between vertical integration and both innovativeness and organizational learning. The results show the need for close reciprocal adjustment to all three levels in order to effectively influence business unit performance. Competitive strategy, corporate strategy, and capabilities all matter - and they are so closely related that any attempt to artificially separate them will be doomed to failure.

71

Mediation effects have been tested through the causal step approach (Judd / Kenny 1981, Baron / Kenny 1986), the Sobel test, the Goodman (I) test, and the Goodman (II) test (Goodman 1960, Sobel 1982, Baron / Kenny 1986, Sobel 1988, MacKinnon et aI1995).

239

Results from the Evolutionary Model

The main hypothesis of the complex model is that performance can be better explained by a model that also captures the development and evolution over time. The evolutionary model complements the integrated and the complex model with the addition of a longitudinal perspective. Three main findings can be summarized: First, we identified consistent patterns of change at the industry level in respect of the different determinants. While a number of variables remained stable over the full period (Le. corporate culture or industry concentration), others showed a consistent positive trend (Le. innovativeness or organizational learning), or even several changes (Le. cost leadership strategy). Second, the strongest and most consistent changes (and only those) had a significant effect on performance. Changes that were either less consistent (like the trend towards cost leadership) or less strong (like minor changes to entry barriers) failed to make an impact (see Figure 48). Third, most determinants' effects on performance remained stable over a four-year period. However, the results from dynamic analysis did, in some cases, depart from prior equilibrium-based findings. These cases indicate that the effect of some determinants during short phases of dynamic change might be different from their effect in long-term equilibrium periods. Our findings provide strong support for our overall research framework and the underlying assumptions. The finding of the relative stability of effects on performance over a four-year period provides justification for equilibrium-based theorizing such as in the two first models in our framework. The findings of characteristic patterns of change and their effects on performance show the importance of complementary evolutionary analysis.

240

1,10

J:

C)

:i:

I

COSTlEA

I I I I I I I

1,00 0,90

Q)

CI C nI

..c

0,80 0,70

...

-

0,60 j

..c

0,50

UJ

0,30

U

... Q)

:::0 ...J

\ I

.

• INNOVA

DIFFSTR

ORGlEAR •

' , . PROD IFF

0,40 _I

.... ...

I I

• INTDIV.

,

1

VERTINT

..... --. .... :

"

• • REPUTAT • MARKGROW CORPCUl

:

--------------1,

0,20 l

I

ENTBARR

0,10 ~ ,

0,00 I 0,25

0,3

0,35

LOW

Figure 48

•

\

0

CI C

• MARKORT

,

0,4

0,5

0,45

0,55

0,6

Persistency of Change

0,65

0,7

0,75

HIGH

The Evolutionary Model

The value of a study is largely determined by its contribution in both informing practice and advancing theory, Drawing on the diverse research findings presented above, a few important implications for managers and academics are presented,

Managerial Implications

Another research objective of the present study was to gain a comprehensive view of the determinants underlying firm performance in the global media industries. The findings from our study draw a far more complex picture of strategy and performance in the media industries than most prior theorizing. The good news is that companies determine their own destiny to a large extent. Success in the media industry is less a consequence of external factors, such as market concentration, than of the firm's unique strategies and capabilities. The challenge is that corporate strategies, business strategies, and capabilities are highly interrelated. Not only are all three relevant, but they need to be combined in the right manner to be successful. Only the right combination leads to

241

success. Moreover, what constitutes the right combination - while relatively stable in the short run - might change over time. Despite this complex and dynamic character, there are a few fundamental "rules of the game" that have to be considered. A competitive strategy that balances an ongoing control of costs and operating margins with a differentiated and innovative product offering seems to be most beneficial for performance. The success of such a strategy depends largely on the availability of capabilities required for strategy implementation. Dynamic capabilities such as innovativeness, market orientation, and organizational learning have been found to playa particularly important role. Strategic moves at the corporate level, such as product diversification or vertical integration, can also support the firm's competitive strategy. However, it is important to align corporate strategy with competitive strategy. What's more, diversification has to focus on the addition of new capabilities to make a difference. The factors underlying firm performance have been found to be relatively stable over time. Media managers need to monitor changes and adjust their strategies accordingly; however, the fundamental essence for success in media strategies has remained unchanged over the past few years.

Research Implications

In summary, we found strong support for a more complex and dynamic perspective on strategy theory. Variables from different perspectives matter and they are closely related. There are numerous interrelations between single factors and they often share a common effect on performance. Future strategy theory has to surmount the current fragmentation into single perspectives. The focus has to be on a better understanding of the complex network of effects underlying performance. In addition, we found that while most effects are stable over a mid-term period, there are significant developments. Future research should thus combine

both

perspectives.

242

equilibrium-based

and longitudinal

We established a reference frame that provides a flexible theoretical backbone for a multitude of potential complex and dynamic research models. Such models require more sophisticated research methods than currently applied in the strategy debate. Our empirical study shows that even more complex models are testable when new methods to process data acquired in field studies are deployed.

VI.2

Limitations of the Study

As with any research endeavor, there are inherent theoretical and methodological limitations to this study that are highlighted in this section.

Theoretical Limitations

With regard to the integrated model, there is a distinct probability that we failed to include all variables that may affect performance. While the complexity theory and population ecology have argued that a fully comprehensive model can never be achieved, it would certainly be beneficial to select and examine additional variables, such as other capabilities or strategic behaviors. However, our selection has been based on the results of more than forty years of strategy research and a multitude of prior empirical studies. Our selection followed the general consensus on the determinants that are most relevant in this literature. While we are confident that we have covered the most important determinants, we clearly fail to cover all of them. Even in respect of the variables considered in our study, we failed to account for all of their potential effects and interrelations. While the complex model captured some of the interactions among the variables, we failed to discuss all of them. For instance, we excluded feedback effects, such as a potential retroaction of performance on individual determinants. This is partly due to the limited literature and prior empirical research on the subject. Additionally, the possibilities for the empirical validation of such models decrease quickly with increasing complexity. In order to have a testable model, we paid a certain price in terms of how far we capture the full complexity of reality.

243

Methodological Limitations

This study has been undertaken in the global media industry. As such, all generalizations

to other contexts or industries must be made with care. Furthermore, the field study has been limited to some companies and some markets within this industry. Additionally, single key informants were used for data collection. All this might question the validity of our results. However, every effort has been made to guarantee sufficient data quality. The vast majority of the respondents are members of top management teams, have spent over a decade in their respective companies, and are highly involved in strategic decisions. The resulting usable sample size of sixteen percent represents the full population of the targeted industry segments. A major limitation in our empirical study is the relatively short timeframe of our longitudinal analysis. Ideally, longitudinal analysis should take into account at least one, but preferably several, decades. This would require both different data sources and statistical methods. In this case, we did not have the required resources and time to start a completely separate effort. We consider the results of the evolutionary model as both important and reliable; however, a more important longitudinal study would be required in order to truly empirically evaluate the evolutionary model. Other limitations can be found in the survey design and the statistical methods selected. They are mostly generic and not specific to this study. However, one additional limitation should be stressed. While we used multi-dimensional measures for most variables, we used single indicators for a few variables (Le. industry concentration). This step was taken in order to reduce the length of the questionnaire as far as possible. While these measures have been used and tested in prior literature, we would recommend the sole use of multi-dimensional measures for future research. Multi-item measures allow better evaluation of measurement errors and capture more facets of a construct. The use of single indicators in the evolutionary model bears the risk of biased results. While results were in line with prior findings (thus supporting their validity), the use of multiple-item measures could clearly improve the overall data reliability.

244

VI.3

Future Research

One advantage of the framework we developed in this study is its inherent flexibility. For instance, future research might detect additional variables and establish their importance in explaining performance. The integrated model is flexible enough to include additional variables or even new perspectives. Even better, completely new models might be added to the framework. Due to the separation of the overall frame of reference and the research framework, it is possible to replace one (or more) of the existing models with an alternative model. Such a model, for instance, might capture very little dynamism, but a high degree of complexity. It would have the advantage of allowing more in-depth analysis of the interactions of variables underlying performance. Even more interesting would be a model that also captures a higher degree of dynamism. Such a model should turn to qualitative methods for empirical validation, as is done in longitudinal case studies. Based on the findings from our study, further research may aim for a much higher degree of both complexity and dynamism. The more insights that are gained into the "mechanics" underlying performance, the more the individual models in our framework can be developed. Alternatively, it would be beneficial to apply our research framework to other industries or countries. The framework could be particularly useful for analysis of other dynamic industries such as telecommunications or high tech. More and more industries are characterized by a truly dynamic and challenging environment. It is time to replace simple tools with a more realistic approach to analyzing performance in such environments. A comparison of the results of such an analysis with the findings of the present study could be conducive to a higher validity and generalizability of the approach. Another interesting alternative would be to apply the research framework to a crossindustry set of companies. In this study we found little evidence of a Significant impact of industry factors on performance. While this finding is in line with certain prior research, other studies found contrary evidence. A cross-industry analysis might be more appropriate to account for industry effect, particularly if the selected industries vary in terms of environmental characteristics.

245

Finally, a simple repetition of the study in the media industries would provide interesting insights into the stability of our findings over time. We have established a certain stability of findings for a four-year period. However, not much can be said regarding the longterm development of these factors and their interactions. It would be interesting to compare the findings of a later study on the same area sample of media companies with the results of the present study.

VI.4

Concluding Comments

In this study, we stressed the need for a more integrative and dynamic perspective on strategic analysis. Simple success strategies are less and less appropriate and are blind to the market realities and developments in most industries. A principal concern of this study has been to overcome the current chasm between traditional and dynamic approaches to strategy. The established frame of reference and the empirically validated research framework provide a foundation for both camps' future cooperation. It is our hope that this study contributes to further acceptance of more complex and dynamic approaches in mainstream strategy theorizing.

246

References Aaker, D.A.I Day, G.S. (1986): "The Perils of High-Growth Markets"; Strategic Management Journal, Vol. 7 (September-October), pp. 409-21 Aaker, D.A. (1984): Strategic Market Management, John Wiley, New York Aaker, D.A. (1991): Managing Brand Equity: Capitalizing on the Value of a Brand Name; The Free Press, New York I NY Abell, D.F. (1980): Defining the Business. The Starting Point of Strategic Planning, Englewood Cliffs Abernathy, W.J. I Utterback, J.M. (1975): "Innovation and the evolving structure of the firm"; Working Paper HBS 75-18, Harvard Business School Acs, Z.J.I Audretsch, D.B. (1988): "Innovation in large and small firms: An empirical analysis"; American Economic Review, Vol. 78, September, pp. 678-90 Adelman, M.A. (1955): Concept and Statistical Measurement of Vertical Integration; Princeton University Press, Princeton / NJ Alavi, M. I Yoo, Y. I Vogel, D.R. (1997): "Using Information Technology to Add Value to Management Education"; Academy of Management Journal, Vol. 40 (6), pp. 1310-1334 Albarran, A.B. I Porco, J. (1990): "Measuring and analyzing diversification of corporations involved in pay cable"; Journal of Media Economics, Vol. 3 (2), pp. 3-14 Albarran, A.B. (1996): Media Economics: Understanding Markets, Industries and Concepts; Iowa State University Press, Ames Albarran, A.B. I Chan·Olmsted, S.M. (1998): Global Media Economics: Commercialization, Concentration, and Integration of World Media Markets; Iowa State University Press, Ames Albarran, A.B. I Moellinger, T. (2002): "The Top Six Communication Industry Firms: Structure, Performance and Strategy"; Unpublished Paper presented at the 5th World Media Economics Conference, University of Turku, Finland Alexander, A. lOwers, J. I Carveth, R. (1998): Media Economics Theory and Practice; Lawrence Erlbaum Associates, Hillsdale, N.J. Altmeppen, K.D. (1996): Qkonomie der Medien und des Mediensystems; Gabler Verlag, Wiesbaden Amit, R. I Schoemaker, P.J. (1993): "Strategic assets and organizational rent"; Strategic Management Journal, Vol. 14, pp. 33-46 Anderson, C.R. I Paine, F.T. (1978): "PIMS: A reexamination"; Academy of Management Review, Vol. 3, pp.602-611 Anderson, E.W. I Farnell, C. I Lehmann, D.R. (1994): "Customer Satisfaction, Market Share, and Profitability: Findings from Sweden"; Journal of Marketing, Vol. 58 (July), pp. 53-66 Anderson, J.C. (1987): "An Approach for Confirmatory Measurement and Structural Equations Modeling of Organizational Properties", Management Science. Vol. 33, pp. 525-541 Anderson, P. (1999): "Complexity Theory and Organization Science"; Organization Science, Vol. 10 (3), Special Issue, pp. 216-232

247

Andrews, K. (1971): The concepts of corporate strategy, Dow Jones-Irwin, Homewood IlL Ansoff, H.I. (1957): "Strategies for diversification"; Harvard Business Review, Vol. 35 (5), pp. 113-124 Ansoff, H.I. (1965): Corporate Strategy, McGraw-Hili, New York Arag6n-Correa, J.A. I Sharma, S. (2003): "A contingent resource-based view of proactive corporate environmental strategy"; Academy of Management Review, Vol. 28 (1), pp. 71-88 Armour, H.O. I Teece, D.J. (1978): "Organizational Structure and Economic Performance: A Test of the Multidivisional Hypothesis"; Bell Joumal of Economics, Vol. 9, pp. 106-112 Armour, H.O. I Teece, D.J. (1980): "Vertical integration and technological innovation"; The Review of Economics and Statistics, pp. 470-474 Armstrong, S.J. I Overton, T.S. (1977): "Estimating Nonresponse Bias in Mail Surveys"; Joumal of Marketing Research, Vol. 14, pp. 396-402 Atkin, D.J. (1994): "Cable Exhibition in the USA: Ownership Trends and Implications of the 1992 Cable Act"; Telecommunications Policy, Vol. 18 (4), pp. 331-341 Bharadwaj, S. I Varadarajan, P. I Fahy, J. (1993): "Sustainable competitive advantage in service industries: a conceptual model and research propositions"; Journal of Marketing, Vol. 57, pp. 83-99 Bajtelsmit, V.L. I Bouzouita, R. (1998): "Market Structure and Performance in Private Passenger Automobile Insurance"; Journal of Risk & Insurance, Vol. 65 (3), pp. 1-11 Bain, J.S. (1951): "Relation of Profit Rate to Industry Concentration: American Manufacturing 1936-1940"; Quarterly Journal of Economics, Vol. 65 (August), pp. 293-324 Bain, J.S. (1956): Barriers to New Competition, Harvard University Press, Cambridge I MA Bain, J.S. (1959): Industrial Organization; John Wiley & Son, New York Bain, J.S. (1972): Essays in Price Theory and Industrial Organization, Little Brown and Company, Boston I MA Baker, W.E. I Sinkula, J.M. (1999): "The synergistic effect of market orientation and learning orientation on organizational performance"; Journal of the Academy of Marketing Society, Vol. 27, pp. 411-27 Barnett, W.P. I Burgelman, R.A. (1996): "Evolutionary Perspectives on Strategy"; Strategic Management Journal, Summer Special Issue, Vol. 17, pp. 5-19 Barney, J.B. (1986): "Types of competition and the theory of strategy: towards an integrating frarnework"; Academy of Management Review, Vol. 11, pp. 656-65 Barney, J.B.I Hoskisson, R.E. (1990): "Strategic Groups: Untested Assertions and Research Proposals"; Managerial and Decision Economics, Vol. 11, February, pp. 187-198 Barney, J.B. (1991): "Firm resources and sustained competitive advantage", Journal of Management, Vol. 17 (1), pp. 94-120 Barney, J.B. (1992): "Integrating organizational behavior and strategy formulation research: A resourcebased analysis"; in: Shrivastava, P. et al (Eds): Advances in Strategic Management, JAI Press, Greenwich I Conn., Vol. 8, pp. 39-61 Barney, J.B. (1997): Gaining and Sustaining Competitive Advantage; Addison-Wesley, Reading I MA Barney, J.B. (2001): "Resource-based theories of competitive advantage: A ten-year retrospective on the resource-based view"; Journal of Management, Vol. 27, pp. 643-650

248

Barney, J.B. (2002): "Strategic management: From informed conversation to academic discipline"; Academy of Management Executive, Vol. 16 (2), May, pp. 53-57 Baron, R.M. I Kenny, D.A. (1986): "The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations"; Joumal of Personality and Social Psychology, Vol. 51, pp. 1173-1182 Bass, F.M. I Cattin, P. I Wittink, D.R (1977): "Market Structure and Industry Influence on Profitability"; in: Thorelli, H. (ed.): Strategy + Structure = Performance, Indiana University Press, Bloomington Bauder, M. (2002): Der deutsche Free-TV-Markt: Chancen fUr neue Anbieter?; Arbeitspapiere des Instituts fUr Rundfunk6konomie an der Universitat zu K61n, Heft 153 Baumol, W.J. I Panzar, J.C. I Willig, RP. (1982): Contestable Markets and the Theory of Industry Structure, Harcourt, Brace, and Janovich, New York Beard, D.W. I Dess, G.G. (1981): "Corporate Level Strategy, Business-Level Strategy, and Firm Performance"; Academy of Management Journal, Vol. 24 (4), pp. 663-688 Bentler, P.M. (1990): EQS: Structural equations programming manual; Multivariate Software, Encino I CA Berry, C.H. (1974): "Corporate Diversification and Market Structure"; Bell Journal of Economics & Management Science, Vol. 5 (1), pp. 196-205 Bettis, RA. (1981): "Performance differences in related and unrelated diversified firms"; Strategic Management Journal, Vol. 2 (4), pp. 379-393 Bettis, R.A. I Hall, W.K. (1982): "Diversification strategy, accounting determined risk, and accounting determined return"; Academy of Management Journal, Vol. 25, pp. 254-264 Bhagat, S. I Shleifer, A. I Vishny, RW. (1990): "Hostile takeovers in the 1980s: The return to corporate specialization"; in: Baily, M.N. I Winston, C. (eds.): Brookings Papers on Economic Activity, Microeconomics, Brookings Institution, Washington DC, pp. 1-84 Bharadwaj, S.G. (1994): "Industry Structure, Competitive Strategy, and Firm-Specific Intangibles as Determinants of Business Unit Performance: Towards an Integrative Model"; Doctoral Dissertation, Texas A&M University Bierly, P. I Chakrabarti, A. (1996): "Generic Knowledge Strategies in the US Pharmaceutical Industry"; Strategic Management Journal, Vol. 17, Winter Special Issue, pp. 123-135 Black, J.A. (2000): "Fermenting Change: Capitalizing on the inherent change found in dynamic non-linear (or complex) systems"; Journal of Organizational Change Management, Vol. 13 (6), pp. 520-525 Black, J.A. I Farias, G. (2000): "Dynamic Strategies: Emergent Journeys"; Emergence, Vol. 2 (1), pp. 101-113 Bollen, K.A. (1989): Structural Equations with Latent Variables, John Wiley & Sons, New York Boulding, W. I Staelin, R. (1990): "Environment, Market Share, and Market Power"; Management Science, Vol. 36, October, pp. 1160-1177 Bourgeois, L.J. I McAllister, D. I Mitchell, T.R. (1978): "The Effects of Different Organizational Environments upon Organization Design Decisions", Academy of Management Journal, Vol. 21 (3), pp. 508-514 Bowen, H. I Wiersema, M. (1999): "Matching Method to Paradigm in Strategy Research: Limitations of Cross-Sectional Analysis and Some Methodological Alternatives"; strategic Management Journal, Vol. 20, pp.625-636

249

Bowman, E.H. I Helfat, C.E. (2001): "Does Corporate Strategy Matter?"; Strategic Management Joumal, Vol. 22, pp. 1-23 Brandenburger, A.M. I Nalebuff, B.J. (1996): Co-opetition; Doubleday, New York I NY Brosius, F. (2002): SPSS 11, mitp Verlag, Bonn Brown, S.L. I Eisenhardt, K.M. (1998): Competing on the Edge: Strategy as Structured Chaos; Harvard Business School Press, Boston IMA Brozen, Y. (1971): "Bain's Concentration and Rates of Return Revisited"; Journal of Law and Economics, Vol. 14 (July), pp. 351-69 Brush, T.H. I Artz, K.W. (1999): "Toward a contingency based theory"; Strategic Management Joumal, Vol. 20 (3), pp. 223-251 Buhner, R. (1987): "Assessing international diversification of West German corporations"; Strategic Management Joumal, Vol. 8 (1), pp. 25-37 Burns, T.I Stalker, G.M. (1961): The management of innovation, Tavistock, London, UK Burt, S.B. et al (1994): "Contingent Organization as a Network Theory: The Culture-Performance Contingency Function"; Acta Sociologica, Vol. 37, pp. 345-370 Buzzell, R.D. I Bradley, T.G. I Sultan, R.G.M. (1975): "Market Share - A Key to Profitability"; Harvard Business Review, January-February, pp. 97-106 Buzzell, R.D. (1983): "Is Vertical Integration Profitable?"; Harvard Business Review, Vol. 61, JanuaryFebruary, pp. 92-102 Buzzell, R.D.I Gale, B.T. (1987): The PIMS principles: Linking Strategy to Performance, The Free Press Cable, J.R. I Dirrheimer, M.F. (1983): "Hierarchies and Markets: An Empirical Test of the Multidivisional Hypothesis in West Germany"; Intemational Joumal of Industrial Organization, Vol. 1, pp. 43-62 Cable, J.R. I Yasuki, M. (1985): "Internal Organization, Business Groups and Corporate Performance: An Empirical Test of the Multidivisional Hypothesis in Japan"; Intemational Joumal of Industrial Organization, Vol. 3, pp. 401-420 Calantone, R.J. I Cavusgil, S.T. I Zhao, Y. (2002): "Learning Orientation, Firm Innovation Capability, and Firm Performance"; Industrial Marketing Management, Vol. 31, pp. 515-524 Calvo, G.A. I Wellisz, S. (1978): "Supervision, loss of control, and the optimum size of the firm"; Joumal of Political Economy, Vol. 86 (5), pp. 943-952 Campbell, D.T. (1955): "The Informant in Quantitative Research"; American Joumal of Sociology, Vol. 60, pp.339-342 Campling, J.T. I Michelson, G. (1998): "A Strategic Choice-Resource Dependence Analysis of Union Mergers in the British and Australian Broadcasting and Film Industries"; Journal of Management Studies, Vol. 35 (5), pp. 597-600 Capon, N. I Farley, J.U. I Hoenig, S. (1990): "Determinants of Financial Performance: A Meta-Analysis", Management Science, Vol. 36 (October), pp. 1143-59 Cappel, S.D. (1990): Strategic Profiles and Performance in Service Organizations: Examining Retail Department, Variety and General Merchandise Stores; Unpublished Doctoral Dissertation, Memphis State University, Memphis, Tennessee

250

Carroll, G.R. (1993): '"A sociological view of why firms differ"; Strategic Management Journal, Vol. 14 (4), pp.237-249 Carroll, G.R. I Harrison, J.R. (1994): '"On the historical efficiency of competition between organizational populations'"; American Journal of Sociology, Vol. 100 (3), pp. 720-750 Carter, J.R. (1977): '"In search of synergy: a structure-performance tesf'; Review of Economics and Statistics, Vol. 59, pp. 279-289 Cave, M. (1989): ,"An Introduction to Television Economics'"; in: Hughes, G.I Vines, D. (eds.): Deregulation and the Future of Commercial Television, Aberdeen, pp. 9-37 Caves, R.E. (1971): '"International corporations: The industrial economics of foreign investment"; Economica, Vol. 38 (February), pp. 1-27 Caves, R.E. I Porter, M.E. (1977): '"From Entry Barriers to Mobility Barriers: Conjectural Decisions and Contrived Deterrence to New Competition'"; Quarterly Journal of Economics, Vol. 91 (May), pp. 421-441 Caves, R.E. I Pugel, T. (1980): Intra-Industry Differences in Conduct and Performance: Viable Strategies in US Manufacturing Industries; Monograph, New York University Caves, R.E. (1981): '"Diversification and seller concentration: Evidence from change"; Review of Economics and Statistics, Vol. 63, pp. 289-293 Caves, R.E. (1982): Multinational enterprise and economic analysis, Cambridge University Press, Cambridge Chakravarthy, B.S. (1982): '"Adaptation: A promising metaphor for strategic management'"; Academy of Management Review, Vol. 25, pp. 735-744 Chamberlin, E.H. (1933): The Theory of Monopolistic Competition, Harvard University Press, Cambridge Chandler, A.D. (1962): Strategy and Structure: Chapters in History of the Industry Enterprise, Doubleday, Garden City, New York Chan-Olmsted, S.M. I Litman, B.R. (1988): '"Antitrust and Horizontal Mergers in the Cable Industry", Joumal of Media Economics, Vol. 1 (2), pp. 3-28 Chan-Olmsted, S.M. (1996): '"Market Competition for Cable Television: Reexamining Its Horizontal Mergers and Industry Competition'"; Journal of Media Economics, Vol. 9 (2), pp. 25-41 Chan-Olmsted, S.M. I Jamison, M. (2001): '"Rivalry through alliances: competitive strategy in the global telecommunications market"; European Management Journal, Vol. 19 (3), pp. 317-331 Chang, S. I Singh, H. (2000): '"Corporate and Industry Effects on Business Unit Competitive Position'"; Strategic Management Journal, Vol. 21, pp. 739-752 Chappell, H.W.I Marks, W.H. I Park, I. (1983): '"Measuring Entry Barriers Using a Switching Regression Model of Industry Profitability"; Southern Economic Journal, Vol. 49 (4), pp. 991-1002 Chatman, J.A. I Jehn, K.A. (1994): "Assessing the Relationship between Industry Characteristics and Organizational Culture: How different can you be?"; Academy of Management Journal, Vol. 37 (3), pp. 522-553 Chatterjee, S. (1991): '"Gains in vertical acquisitions and market power: Theory and Evidence"; Academy of Management Journal, Vol. 34, pp. 436-448 Chatterjee, 5.1 Wernerfelt, B. (1991): ,"The link between resources and type of diversification: Theory and evidence'"; Strategic Management Journal, Vol. 12 (1). pp. 33-48

251

Chen, P.H. (2002): 'Who owns Cable Television? Media Ownership Concentration in Taiwan"; Journal of Media Economics, Vol. 15 (1). pp. 41-56 Christensen, H.K. I Montgomery, C.A. (1981): "Corporate economic performance: Diversification strategy versus market structure"; Strategic Management Journal, Vol. 2 (4). pp. 327-343 Churchill, G.A. (1999): Marketing Research: Methodological Foundations; Harcourt, New York Claver, E. I Molina, J. I Tari, J. (2002): "Firm and Industry Effects on Firm Profitability: A Spanish Empirical Analysis"; European Management Journal, Vol. 20 (3), pp. 321-328 Coase, R. (1950): British Broadcasting: A Study in Monopoly, Longmans Green, London, Harvard University Press, Cambridge I MA Cockburn, I.M. I Henderson, R.M. I Stern, S. (2000): "Untangling the Origins of Competitive Advantage"; Strategic Management Journal, Vol. 21, pp. 1123-1145 Colliner, D.W. I Monz, J. I Conlin, J. (1984): "How Effective is Technological Innovation?"; Research Management, Vol. 27, September·October, pp. 11-16 Collins, R. I Garnham, N. I Locksley, G. (1989): The Economics of Television: The UK case; Sage, London Collis, D.J. (1994): "Research note: how valuable are organizational capabilities?"; Strategic Management Journal, Vol. 15, pp. 143-52 Collis, D.J. I Montgomery, C.A. (1995): "Competing on Resources: Strategy in the 1990s"; Harvard Business Review, July - August, pp. 119-128 Com paine, B.M. I Gomery, D. (2000): Who Owns the Media: Competition and Concentration in the Mass Media Industry; Lawrence Erlbaum, Hillsdale, N.J. (3 rd edition) Conant, J.S. I Mokwa, M.P. I Varadarajan, P.R. (1990): "Strategic types, distinctive marketing competencies and organizational performance: a multiple measures·based study"; Strategic Management Journal, Vol. 11, pp. 365-383 Conner, K.R. I Prahalad, C.K. (1996): "A resource·based theory of the firm: Knowledge versus Opportunism"; Organization Science, Vol. 7, pp. 478-496 Cool, K.O. I Schendel, D. (1987): "Strategic Group Formation and Performance: The case of the US Pharmaceutical Industry"; Management Science, Vol. 33, pp. 1102-1124 Cool, K.O. I Schendel, D. (1988): "Performance Differences among Strategic Group Members"; Strategic Management Journal, Vol. 9 (3), pp. 207-233 Cournot, A. (1838): Researches into the Mathematical Principles of the Theory of Wealth, Homewood Craig, C.S.I Douglas, S.P. (1982): "Strategic Factors associated with market and financial performance"; Quarterly Review of Economics and Business, Vol. 22, pp. 10H 12 Craig, T. (1996): "The Japanese Beer Wars: Initiating and Responding to Hypercompetition in New Product Developmenf'; Organization Science, Vol. 7 (3), pp. 302-322 Crepon, B. I Duguet, E. I Mairesse, J. (1998): "Research, Innovation, and Productivity: An Econometric Analysis at the Firm Level"; The Economics of Innovation and New Technology, Vol. 7, pp. 115-158 Creswell, J.W. (1994): Research design: Qualitative and quantitative approaches; Sage Publications, Thousand Oaks

252

Cubbin, J. I Geroski, P. (1987): "The convergence of profits in the long run: inter-firm and inter-industry comparisons"; Journal of Industrial Economics, Vol. 35, pp. 427-442 Curd, M.I Cover, J.A. (1998): Philosophy of Science, Norton & Company, New York and London Cyert, R.M.I March, J.G. (1963): A behavioral theory of the firm; Prentice-Hall, Englewood Cliffs I NJ Dale, M. (1997): The Movie Game: The Film Business in Britain, Europe and America; Cassell, London and New York Damanpour, F. (1991): "Organizational Innovation: A meta-analysis of effects of determinants and moderators"; Academy of Management Journal, Vol. 34 (3), pp. 555-90 Damanpour, F. (1996): "Organizational complexity and innovation"; Management Science, Vol. 42 (5), pp. 693-716 Daniels, J.D. I Bracker, J. (1989): "Profit performance: Do foreign operations make a difference?"; Management International Review, Vol. 29 (1), pp. 46-56 Datta, O.K. I Rajagopalan, N. I Rasheed, A.M.A. (1991): "Diversification and Performance: Critical Review and Future Directions"; Journal of Management Studies, Vol. 28 (5), pp. 529-558 D'Aveni, R. I lIinitch, A.Y. (1992): "Complex Patterns of Vertical Integration in the Forest Products Industry: Systematic and Bankruptcy Risks"; Academy of Management Journal, Vol. 35 (3), pp. 596-625 D'Aveni, R. (1994): Hypercompetition; Free Press, New York D'Aveni, R. I Ravenscraft, D.J. (1994): "Economies of Integration versus Bureaucracy Costs: Does Vertical Integration improve Performance?"; Academy of Management Journal, Vol. 37 (5), pp. 1167 1206 Day, G.S. I Wensley, R. (1988): "Assessing advantage: a framework for diagnosing competitive superiority"; Journal of Marketing, Vol. 52, pp. 1-20 Day, G.S. (1994): "The capabilities of Market-Driven Organizations", Journal of Marketing, Vol. 58, pp. 3752 Daymon, C. (2000): "Culture Formation in a New Television Station: A Multi-Perspective Analysis"; British Journal of Management, Vol. 11 (2), pp.121-135 Deal, T.E.I Kennedy, A.A. (1982): Corporate cultures, Reading I Mass. DeGeus, A.P. (1988): "Planning as Learning"; Harvard Business Review, Vol. 66 (2), pp. 70-75 Demers, D. (1999): Global Media: Menace or Messiah?; Hampton Press, Cresskill I NJ Demsetz, H. (1973): "Industry Structure, Market Rivalry, and Public Policy"; Journal of Law and Economics, Vol. 16, April, pp. 1-19 Deng, S. I Dart, J. (1994): "Measuring market orientation: a multi-factor, multi-item approach", Journal of Marketing Management, Vol. 10, pp. 725-42 Denison, D.R. (1990): Corporate Culture and Organisation Effectiveness; John Wiley, New York I NY Deshpande, R. I Webster, F.E. jr. (1989): "Organizational culture and marketing: defining the research agenda"; Journal of Marketing, Vol. 53, pp. 3-15 Deshpande, R. I Farley, J.U. I Webster, F.E. jr. (1993): "Corporate culture, customer orientation, and innovativeness in Japanese firms: a quadrate analysis"; Journal of Marketing, Vol. 57, pp. 23-27

253

Deshpande, R. I Farley, J.U. (1999): "Corporate Culture and Market Orientation: Comparing Indian and Japanese Firms"; Journal of International Marketing, Vol. 7 (4), pp. 111-127 Dess, G.G. I Davis, P.S. (1984): "Porter's Generic Strategies as Determinants of Strategic Group Membership and Organizational Performance", Academy of Management Review, Vol. 27, pp. 467-488 Dess, G.G. I Robinson, R.B. (1984): "Measuring organizational performance in the absence of objective measures: The case of privately-held firm and conglomerate business unit"; Strategic Management Journal, Vol. 5, pp. 265-273 Dess, G.G.I Miller, A. (1993): .Assessing Porter's (1980) model in terms of generalizability, accuracy and simplicity"; Journal of Management Studies, Vol. 30 (4), pp. 553-585 Dess, G.G. I Gupta, A. I Hennart, J.F. I Hill, C.W.L. (1995): "Conducting and Integrating Strategy Research at the International, Corporate, and Business Levels: Issues and Directions"; Journal of Management, Vol. 21 (3), pp. 357-393 Devaney, R.L. (1993): An introduction to chaotic dynamical systems, Benjamin I Cummings, Menlo Park De Wit, B.I Meyer, R. (1994): Strategy: Process, Content, Context; West Publishing, New York I NY Diamantopoulos, A. I Siguaw, J.A. (2000): Introducing LlSREL: A Guide for the Uninitiated; SAGE Publications, Thousand Oaks I California Dierickx, I. I Cool, K. (1989): "Asset Stock Accumulation and Sustainability of Competitive Advantage"; Management Science, Vol. 35, pp. 1504-1511 Dillman, D. (1991): "The Design and Administration of Mail Surveys"; in: Scott, R. I Blake, J. (Eds.): Annual Review of Sociology, Annual Reviews Inc, Palo Alto, pp. 225-249 Dimmick, J. I Wallschlager, M. (1986): "Measuring corporate diversification: A case study of new media ventures by television network parent companies"; Journal of Broadcasting and Electronic Media, Vol. 30 (1), pp. 1-14 Dixon, R.D. I Boal, K.B. (1996): "Strategic Groups: Membership, Environmental Stability, and Performance"; Academy of Management Proceedings 1996 Dooley, K.J. I Van de Ven, A.H. (1999): "Explaining Complex Organizational Dynamics"; Organization Science, Vol. 10 (3), Special Issue, pp. 358-372 Dosi, G. (1988): "Sources, Procedures, and Microeconomic Effects of Innovation"; Journal of Economic Literature, Vol. 26, pp. 1120-71 Dosi, G. I Nelson, R.R. (1994): "An introduction to evolutionary theories in economics"; Joumal of Evolutionary Economics, Vol. 4, pp. 153-172 Dowling, G.R. (2001): Creating Corporate Reputations; Oxford University Press, Oxford Drucker, P.F. (1973): Management, Mc Graw-Hill, New York I NY Eaton, B.C. I Lipsey, R.G. (1980): "Exit barriers are entry barriers: the durability of capital as a barrier to entry"; Bell Journal of Economics, Vol. 11, pp. 721-729 Eisenhardt, K.M. I Martin, J.A. (2000): "Dynamic Capabilities: What are they?"; Strategic Management Joumal, Vol. 21, pp. 1105-1121 Elashmawi, F. (2000): "Creating a Winning Corporate Culture. Experience inside the Asian Telecommunications Industry"; European Business Review, Vol. 12 (3), pp. 148-156

254

Eldredge, N.' Gould, S. (1972): "Punctuated Equilibria: An Alternative to Phyletic Gradualism"; in: Schopf, T.J. (ed.): Models in Paleobiology; Freeman, Cooper & Co., San Francisco, pp. 82-115 Evanoff, D.O. , Fortier, D.L. (1988): "Reevaluation of the structure-conduct-performance paradigm in banking"; Journal of Financial Services Research, Vol. 1 (3), pp. 249-260 Ezzamel, M.A. , Hilton, K. (1980): "Divisionalization in British Industry: A Preliminary Study"; Accounting and Business Research, Vol. 38, pp. 197-214 Ezzamel, M.A. , Watson, R. (1993): "Organizational form, ownership structure and corporate performance: a contextual analysis of UK companies"; British Journal of Management, Vol. 4, pp. 161-176 Fagerberg, J. (2002): A Layman's Guide to Evolutionary Economics; Conference Paper "Industrial R&D and Innovation Policy Learning", Asker, April 18'h 2002 Fahy, J. (2000): "The resource-based view of the firm: Some stumbling-blocks on the road to understanding sustainable competitive advantage"; Journal of European Industrial Training, Vol. 24, pp. 94-104 Farjoun, M. (2002): "Towards an Organic Perspective on Strategy"; Strategic Management Journal, Vol. 23, pp. 561-594 Faulkner, 0.1 Johnson, G. (1992): The Challenge of Strategic Management, Kogan Page Ltd, London Fiedler, F.E. (1967): A theory of leadership effectiveness; McGraw-Hili, New York I NY Fiol, C.M. I Lyles, M.A. (1985): "Organizational Learning"; Academy of Management Review, Vol. 10 (4), pp.803-813 Fliegenbaum, A. I Thomas, H. (1990): "Strategic Groups and Performance: The US Insurance Industry 1970-1984"; Strategic Management Journal, Vol. 11, pp.197-215 Fombrun, C. , Shanley, M. (1990): 'What's in a name? Reputation building and corporate strategy"; Academy of Management Journal, Vol. 33, pp. 233-258 Fombrun, C. (1996): Reputation: Realizing Value from the Corporate Image; Harvard Press, Boston / Mass. Ford, G.S. I Jackson, J.D. (1997): Horizontal Concentration and Vertical Integration in the Cable Television Industry; Jackson Publishing Fornell, C. , Larcker, D.F. (1981): "Evaluating Structural Equation Models with Unobservables and Measurement Error," Journal of Marketing Research, Vol. 18, pp. 39-50. Foss, N.J. I Knudsen, C. I Montgomery, C.A. (1995): "An Exploration of Common Ground: Integrating Evolutionary and Strategic Theories of the Firm"; in: Montgomery, C.A. (Ed.): Resource-based and Evolutionary Theories of the Firm: Towards a Synthesis; Kluwer Academics, Boston / Mass Fox, I. , Srinivasan, S. , Vaaler, P. (1997): Some constructive skepticism about the origins of profit; US FTC Business Paper Series, Washington / DC Frazier, G.L. / Howell, R.D. (1983): "Business Definition and Performance"; Journal of Marketing, Vol. 47, Spring, pp. 59-67 Galbraith, J. (1973): Designing Complex Organizations; Addison-Wesley, Readings / MA Galbraith, C. , Schendel, D.E. (1983): "An empirical analysis of strategic types"; Strategic Management Joumal, Vol. 4, pp. 153-173 Gale, B.T. (1978): "Cross-Sectional Analysis: The New Frontier in Planning"; Planning Review, March

255

Gale, B.T. (1980): "Can More Capital Buy Higher Productivity?"; Harvard Business Review, Vol. 30, JulyAugust, pp. 78-86 Gerbing, D.W. I Anderson, J.C. (1988): "An updated paradigm for scale development incorporating unidimensionality and its assets"; Journal of Marketing Research, Vol. 24, pp. 186-192 Geringer, J.M. I Beamish, P.W. I daCosta, R.C. (1989): "Diversification strategy and internationalization: Implications for MNE performance"; Strategic Management Journal, Vol. 10, pp. 109-119 Geroski, P. (1994): Market Structure, Corporate Performance and Innovative Activity; Oxford University Press, Oxford Gersick, C.J.G. (1991): "Revolutionary Change Theories: A Multilevel Exploration of the Punctuated Equilibrium Paradigm"; Academy of Management Review, Vol. 16 (1), pp. 10-36 Ghemawat, P. (1994): Nucor at a crossroads: teaching note; Harvard Business School, Boston I MA Ghoshal, S. (1987): "Global Strategy: An Organizing Framework"; Strategic Management Journal, Vol. 8, pp.425-440 Gilbert, R.A. (1984): "Bank Market Structure and Competition: A Survey"; Journal of Money, Credit, and Banking, Vol. 16 (November), pp. 617-645 Gilbert, X. I Strebel, P. (1989): "Developing competitive advantage"; in: Mintzberg, H I Quinn, J.B. (eds.): The Strategy Process, Concepts, Contexts, Cases; Prentice-Hall, London, pp. 82-93 (2 nd edition) Ginsberg, A. I Venkatraman, N. (1985): "Contingency Perspectives of Organizational Strategy: A Critical Review of the Empirical Research"; Academy of Management Review, Vol. 10 (3), pp. 421-434 Gioia, D.A.I Pitre, E. (1990): "Multiparadigm Perspectives on Theory Building"; Academy of Management Review, Vol. 15 (4), pp. 584-602 Godfrey, P.C. I Hill, C.W. (1995): "The problem of unobservables in strategic management research"; Strategic Management Journal, Vol. 16, pp. 519-33 Gomery, D. (1993): ''The centrality of media economics"; Journal of Communication, Vol. 43 (3), pp. 190198 Gomez, P. I Probst, G. (1987): Vernetz1es Denken im Management: Eine Methodik des Ganzheitlichen Problemltisens; Die Orientierung 89, Schweizerische Volksbank, Bern. Goodman, L.A. (1960): "On the exact variance of products"; Journal of the American Statistical Association, Vol. 55, pp. 708-713 Gordon, G.G. (1991): "Industry Determinants of Organizational Culture"; Academy of Management Journal, Vol. 16 (2), pp. 396-415 Gordon, G.G. I DiTomaso, N. (1992): "Predicting Corporate Performance from Organizational Culture"; Journal of Management Studies, Vol. 29 (6), pp. 783-98 Gort, M. (1962): Diversification and Integration in American Industry; Princeton University Press, Princeton

I NJ Grant, R.M. (1987): "Multinationality and Performance among British Manufacturing Companies"; Journal of International Business Studies; Vol. 18 (Fall), pp. 79-89 Grant, R.M. I Jammine, A. I Thomas, H. (1988): "Diversity, diversification, and profitability among British manufacturing companies, 1972-1984"; Academy of Management Journal, Vol. 31, pp. 771-801

256

Grant, R.M. (1991): ''The resource-based theory of competitive advantage: Implications for strategy formulation." California Management Review, Vol. 33. pp. 114-135 Grant, R.M. (1996): "Towards a knowledge-based theory of the firm"; Strategic Management Journal, Vol. 17 (10), pp. 109-123 Grant, R.M. (2002): Contemporary Strategy Analysis, Blackwell Publishers, Malden I Mass (4'h edition) Greenley, G.E. (1995): "Market orientation and company performance: empirical evidence from UK companies"; British Journal of Management, Vol. 6, pp. 1-13 Greiner, l.E. I Schein, V.E. (1989): Power and Organization Development: Mobilizing Power to Implement Change, Addison-Wesley Publishing Company, USA Gunnigle, P. I Moore, S. (1994): "Linking Business Strategy and Human Resource Management: Issue and Implications"; Personnel Review, Vol. 23 (1), pp. 63-84 Haar, J. (1989): "A comparative analysis of the profitability performance of the largest US, European, and Japanese multinational enterprises"; Management International Review, Vol. 29 (3), pp. 5-18 Habann, F. (1999): Kernressourcenmanagement in Medienunternehmen; Doctoral Thesis, Lohmar, Koln Habann, F. (2000): "Management of Core Resources: The Case of Media Enterprises"; International Journal of Media Management, Vol. 2 (1), pp. 14-24 Habann, F. (2001): "Resource-based View der Unternehmung: Aussagegehalt fOr das Management von Medienunternehmen", in: Medienwirtschaft & Gesellschaft, Vol. 1, St.Gallen Hall, R. (1989): "The management of intellectual assets: a new corporate perspective"; Journal of General Management, Vol. 15, pp. 53-68 Hall, R. (1992): "The strategic analysis of intangible resources"; Strategic Management Journal, Vol. 13, pp.135-144 Hall, R. (1993): "A framework linking intangible resources and capabilities to sustainable competitive advantage." Strategic Management Journal, Vol. 14, pp. 607-618 Hambrick, D.C. (1979): "Environmental scanning, organizational strategy, and executive roles: a study in mature industries", Unpublished doctoral dissertation, Pennsylvania State University, PA Hambrick, D.C. (1983): "Some tests of the effectiveness and functional attributes of Miles and Snow strategic types"; Academy of Management Journal, Vol. 26, pp. 5-26 Hambrick, D.C. I Lei, D. (1985): "Toward an Empirical Prioritization of Contingency Variables for Business Strategy"; Academy of Management Journal, Vol. 28 (4), pp. 763-788 Hamel, G. I Prahalad, C.K. (1989): "Strategic Intent"; HaNard Business Review, Vol. 67 (May I June), pp. 63-77 Hamel, G.I Prahalad, C.K. (1990): "The core competence of the corporation"; HaNard Business Review, Vol. 68 (3), pp. 79-91 Hamel, G. (1991): "Competition for competence and inter-partner learning within strategic alliances"; Strategic Management Journal, Vol. 12, Special Issue, pp. 83-104 Hamel, G. (2000): Leading the Revolution; Harvard Business School Press, Boston I MA Hamermesh, R.G. I Anderson, M.J. I Harris, J.E. (1978): "Strategies for Low Market Share Businesses"; HaNard Business Review, Vol. 56, May-June, pp. 95-102

257

Handy, C. (1994): The Age of Paradox. Harvard Business School Press, Harvard / Mass. Hansen, G.S. / Wernerielt, B. (1989): "Determinants of Firm Performance: The Relative Importance of Economic and Organizational Factors"; Strategic Management Joumal, Vol. 10 (September-October), pp. 399-411 Harrigan, K.R. (1981): "Barriers to entry and competitive strategies"; Strategic Management Journal, Vol. 2 (4), pp. 395-412 Harrigan, K.R. (1983): Strategies for Vertical Integration; Lexington Books, Lexington / MA Harrigan, K.R. (1986): "Quick change strategies for vertical integration"; Planning Review, Vol. 14, pp. 3237 Hawawini, G. / Subramanian, V. / Verdin, P. (2003): "Is Performance Driven by Industry - or FirmSpecific Factors? A new look at the evidence"; Strategic Management Joumal, Vol. 24, pp. 1-16 Hayek, F.A. (1945): "The use of knowledge in society"; American Economic Review, Vol. 35, pp. 519-530 Heeb, R. (2003): "Innovation and Vertical Integration in Complementary Markets"; Joumal of Economics & Management Strategy, Vol. 12 (3), pp. 387-417 Heinrich, J. (1994): Medien6konomie 1: Mediensystem, Zeitung, Zeitschrift, Anzeigenblatt; Westdeutscher Verlag, Opladen Helfat, C.E. (1997): "Know-how and Asset Complementarity and Dynamic Capability Accumulation: The Case of R&D", Strategic Management Joumal, Vol. 18, pp. 339-360 Helfat, C.E. (2000): "The Evolution of Firm Capabilities: Introduction to the Special Issue"; Strategic Management Joumal, Vol. 21, pp. 955-959 Hellman, H. I Soramaki, M. (1994): "Competition and content in the US video marker'; Journal of Media Economics, Vol. 7 (1), pp. 29-49 Henderson, R. I Mitchell, W. (1997): "The interactions of organisational and competitive influences on strategy and performance"; Strategic Management Journal, Vol. 18, pp. 5-14 Hennart, J.F. I Park, Y.R. (1994): "Location, governance, and determinants of Japanese manufacturing investment in the United States"; Strategic Management Journal, Vol. 15, pp. 419-436 Herremans, I.M. I Akathaporn, P. I Mcinnes, M. (1993): "An investigation of corporate social responsibility reputation and economic performance"; Accounting Organizations and Society, Vol. 18, pp. 587-604 Hill, C.W.L. (1988): "Differentiation versus Low Cost or Differentiation and Low Cost A Contingency Framework"; Academy of Management Review, Vol. 13 (3), pp. 401-412 Hill, C.W.L. I Snell, S.A. (1988): "External control, corporate strategy, and firm performance in researchintensive industries"; Strategic Management Journal, Vol. 9 (6), pp. 577-590 Hill, C.W.L. I Snell, S.A. (1989): "Effects of Ownership Structure and Control on Corporate Productivity"; Academy of Management Journal, Vol. 32 (1), pp. 25-47 Hill, C.W.L. I Hansen, G.S. (1991): "A longitudinal study of the cause and consequences of changes in diversification in the US pharmaceutical industry 1977-1986"; Strategic Management Journal, Vol. 12 (3), pp.187-199 Hill, C.W.L. I Deeds, D.L. (1996): "The Importance of Industry Structure for the Determination of Firm Profitability: A Neo-Austrian Perspective"; Joumal of Management Studies, Vol. 33 (4), pp. 429-451

258

Hill, T.I Westbrook, R. (1997): "SWOT analysis: It's time for a product recall"; Long Range Planning, Vol. 30 (1), pp. 46-53 Hitt, M.A. I Ireland, R.D. (1985): "Corporate distinctive competence: strategy, industry, and performance"; Strategic Management Journal, Vol. 6, pp. 273-93 Hitt, M.A. I Ireland, R.D. (1986): "Relationship among Corporate level Distinctive Competencies, Diversification Strategy, Corporate Structure and Performance"; Journal of Management Studies, Vol. 23 (4), pp. 401-417 Hitt, M.A. I Hoskisson, R.E.I Kim, H. (1997): "International Diversification: Effects on Innovation and Firm Performance in Product-Diversified Firms"; Academy of Management Journal, Vol. 40 (4), pp. 767-798 Hofer, C. I Schendel, D. (1978): Strategy Formulation: Analy1ical Concepts; West Publishing Company, SI.Paul1 MN Holmbeck, G.N. (1997): "Toward Terminological, Conceptual, and Statistical Clarity in the Study of Mediators and Moderators: Examples from the Child-Clinical and Pediatric Psychology Literature"; Journal of Consulting and Clinical Psychology, Vol. 65 (4), pp. 599-610 Homburg, C. I Pflesser, C. (2000): "A multiple layer model of market-oriented organizational culture: Measurement issues and performance outcomes"; Journal of Marketing Research, Vol. 37, pp. 449-462 Hoskisson, R.E. (1987): "Multidivisional structure and performance: The diversification strategy contingency"; Academy of Management Journal, Vol. 30, pp. 625-644 Hoskisson, R.E. I Hitt, M.A. (1988): "Strategic control systems and relative R&D investment in large multiproduct firms"; Strategic Management Journal, Vol. 9 (6), pp. 605-621 Hoskisson, R.E. I Hitt, M.A. (1990): ,Antecedents and performance outcomes of diversification: Review and critique of theoretical perspectives"; Joumal of Management, Vol. 16, pp. 461-509 Hoskisson, R.E. I Hill, C.W.L. I Kim, H. (1993): "The multidivisional structure: organizational fossil or source of value?"; Journal of Management, Vol. 19, pp. 269-298 Hoskisson, R.E. et al (1999): "Theory and research into Strategic Management: Swings of a pendulum"; Joumal of Management, Vol. 25 (3), pp. 417-457 Houcken, R. (1999): The International Feature Film Industry: National Advantage and International Strategies for European Film Companies; Doctoral Dissertation, University of SI. Gallen, Switzerland Hoyle, R.H. (1995): Structural Equations Modeling: Concepts, Issues, and Applications; Sage Publications, Thousand Oaks I California Hoyle, R.H. I Panter A.T. (1995): Writing About Structural Equations Models, in: Hoyle, R.H. (Ed): Structural Equations Modeling: Concepts, Issues, and Applications; Sage Publications, Thousand Oaks I California, 158-176 Hu, L. I Bentler, P.M. (1995): Evaluating Model Fit; in: Hoyle, R.H. (Ed): Structural Equations Modeling: Concepts, Issues, and Applications; Sage Publications, Thousand Oaks I California, pp. 76-99 Huber, G.P. I Power, D.J. (1985): "Retrospective reports of strategic-level managers: Guidelines for increasing their accuracy"; Strategic Management Journal, Vol. 6, pp. 171-180 Hull, G.T.H.I Ketchen, D.J. (2001): "Does Market Orientation Matter? A Test of the Relationship between Positional Advantage and Performance"; Strategic Management Journal, Vol. 22, pp. 899-906 Hume, D. (1966): An Enquiry concerning Human Understanding; Open Court Publishing Company

259

Hurley, R.F. I Hult, G.T.M (1998): "Innovation, market orientation, and organizational learning: An integration and empirical examination"; Journal of Marketing, Vol. 62, pp. 42-54 llinitch, A. y, I O'Aveni, R,A, I Lewin, A.Y. (1996): "New Organizational Forms and Strategies for Managing Hypercompetitive Environments"; Organization Science, Vol. 7, pp. 211-220 Ingham, H. (1992): "Organizational Structure and Firm Performance: An Intertemporal Perspective"; Journal of Economic Studies, Vol. 19 (5), pp. 19-35 ltami, H. (1987): Mobilizing invisible assets; Harvard University Press, Cambridge I MA Jacobson, R.I Aaker, O.A. (1985): "Is Market Share all that it's cracked up to be?"; Journal of Marketing, Vol. 49, Fall, pp. 11-22 Jacobson, R. (1988): "Distinguishing Among Competing Theories of the Market Share Effecf'; Journal of Marketing, Vol. 52, October, pp. 68-80 Jacobson, R. (1992): "The Austrian School of Strategy"; Academy of Management Review, Vol. 17 (4), pp.782-807 Jacobson, R. I Aaker, O.A. (1993): "Composite dependent variables and the market share effecf'; Marketing Science, Vol. 12 (2), Spring, pp. 209-212 Jarillo, J.e. (1987): Dirreci6n estrategica; McGraw Hill, Madrid. Jarillo, J.C. (2003): Strategic Logic, Palgrave Macmillan, New York Jaworski, B.J. I Kohli, A.K. I Sahay, A. (2000): "Market-driven versus driving markets"; Journal of the Academy of Marketing Science, Vol. 28 (1), pp. 119-35 Jemison, O.B. (1981): "The Importance of an Integrative Approach to Strategic Management Research"; Academy of Management Review, Vol. 6 (4), pp. 601-608 Johnson, G. I Thomas, H. (1988): Strategic Groups and Financial Performance: A critical examination; Working Paper, Manchester Business School Jones, G.R. I Hill, C.W.L. (1988): "Transaction cost analysis of strategy-structure choice"; Strategic Management Journal, Vol. 9, pp. 159-172 Joreskog, K. I Sorbom, O. (2001): LlSREL 8: Structural Equation Modeling with the SIMPLIS Command Language, SSI Scientific Software International, Hillsdale I NJ Jose, M.L.I Nichols, L.M.I Stevens, J.L. (1986): "Contributions of diversification, promotion, and R&D to the value of multiproduct firms: A Tobin's q approach"; Financial Management, winter, pp. 33-81 Judd, C.M. I Kenny, O.A. (1981): "Process analysis: Estimating mediation in evaluation research", Evaluation Research, Vol. 5, pp. 602-619 Kay, J. (1993): "The structure of strategy"; Business Strategy Review, Vol. 4, pp. 17-37 Kelloway, Kevin E. (1998): Using LlSREL for Structural Equations Modeling: A Researcher's Guide, SAGE Publications, Thousand Oaks Kelly, M.T. (2000): Competition versus Regulation in CATV: Are local Cable TV Markets natural Monopolies?; Ph. D. Dissertation, University of Delaware Kim, L. I Lim, Y. (1988): "Environment, generic strategies, and performance in a rapidly developing country: A taxonomic approach"; Academy of Management Journal, Vol. 31, pp. 802-827

260

Kim, W.C. I Hwang, P. I Burgers, W.P. (1989): "Global diversification strategy and corporate profit performance"; Strategic Management Joumal, Vol. 10, pp. 45-57 Kim, W.C.I Hwang, P.I Burgers, W.P. (1993): "Multinationals' diversification and the risk-return tradeoff'; Strategic Management Joumal, Vol. 14, pp. 275-286 Kirzner, I,M. (1973): Competition and Entrepreneurship; The University of Chicago Press, Chicago Kirzner, I.M. (1979): "Comment: X-inefficiency, error and the scope for entrepreneurship"; in: Rizzo, M.J. (Ed.): Time, Uncertainty and Disequilibrium; Lexington Books, Lexington I MA Klein, B. I Leffler, K. (1981): "The role of Market Forces in assuring contractual Performance"; Joumalof Political Economy, Vol. 89, pp. 615-641 Klomp, L. I Van Leeuwen, G. (2001): "Linking Innovation and Firm Performance: A New Approach"; International Journal of the Economics of Business, Vol. 8 (3), pp. 343-364 Koeller, C.T. (1995): "Innovation, Market Structure, and Firm Size: A Simultaneous Equations Model"; Managerial and Decision Economics, Vol. 16, pp. 259-269 Kogut, B. (1985): "Designing global strategies: Comparative and competitive value added chains"; Sloan Management Review, Vol. 27 (Summer), pp. 15-28 Kohli, A.K. I Jaworski, B.J. (1990): "Market orientation: Antecedents and Consequences"; Journal of Marketing, Vol. 54, pp. 1-18 Kotter, J.P.I Heskett, J.L. (1992): Corporate Culture and Performance; The Free Press, New York I NY Kruse, J. (1988): "Strategische Markteintrittsbarrieren gegen neue ProgrammanbieterT, in: HoffmannRiem, W. (ed.): Rundfunk im Wettbewerbsrecht: Der offenllich-rechtliche Rundfunk im Spannungsfeld zwischen Wirtschafisrecht und Rundfunkrecht; Baden-Baden, pp. 132-146 Kruse, J. (2000): Okonomische Perspektiven des Fernsehens in Deutschland; Verlag Reinhard Fischer, MOnchen Kuhn, T. (1962): The Structure of Scientific Revolutions, University of Chicago Press, Chicago Kumar, M.S. (1984): Growth, acquisition and investment; Cambridge University Press, Cambridge Kumar, K. et al (1998): "Examining the Market Orientation-Performance Relationship: A context-specific study"; Journal of Management, Vol. 24 (2), pp. 201-234 Kumar, K. I Subramanian, R. I Strandholm, K. (2002): "Market Orientation and Performance: Does Organizational Strategy Matter?"; Journal of Applied Business Research, Vol. 18 (1), pp. 37-50 Kung, L. (2000): "Exploring the Link between Culture and Strategy in Media Organizations: The Cases of the BBC and CNN"; International Journal of Media Management, Vol. 2 (2), pp. 100-109 Lacy, S. I Davenport, L. (1994): "Daily Newspaper Market Structure, Concentration, and Competition"; Journal of Media Economics, Vol. 7 (3), pp. 33-46 Landon, S. I Smith, C.E. (1997): "The use of quality and reputation indicators by consumers: the case of Bordeaux wine"; Journal of Consumer Policy, Vol. 20, pp. 289-323 Lang, L. I Poulsen, A. I Stulz, R. (1995): ,Asset sales, firm performance, and the agency costs of managerial discretion"; Journal of Financial Economics, Vol. 37, pp. 3-37 Langley, A (1999): "Strategies for theorizing from process data"; Academy of Management Review, Vol. 24(4), pp.691-710

261

lawrence, P.R. I lorsch, J.W. (1967): Organization and Environment, Harvard Business School Press, Boston I Mass. lawson, B. I Samson, D. (2001): "Developing Innovation Capability in Organizations: A Dynamic Capabilities Approach"; International Journal of Innovation Management, Vol. 5 (3), pp. 377-400 learned, E.P. et al (1965): Business Policy: Text and Case, Richard D. Irwin, Homewood IlL lee, J. I Miller, D. (1996): "Strategy, Environment and Performance in Two Technological Contexts: Contingency Theory in Korea"; Organization Studies, Vol. 17 (5), pp. 729-750 lei, D. I Hitt, M.A. I Bettis, R. (1996): "Dynamic Core Competences through Meta-Learning and Strategic Context'; Journal of Management, Vol. 22 (4), pp. 549-569 lenz, R.T. (1980): "Environment, Strategy, Organization Structure and Performance: Patterns in One Industry"; Strategic Management Journal, Vol. 1 (3), pp. 209-226 leonard-Barton, D. (1992): "Core capabilities and core rigidities: A paradox in managing new product development"; Strategic Management Journal, Vol. 13, pp. 111-125 levinson, D.J. (1986): "A conception of adult development"; American Psychologist, Vol. 41, pp. 3-13 levinthal, D. I Myatt, J. (1994): "Co-Evolution of Capabilities and Industry: The Evolution of Mutual Fund Processing"; Strategic Management Journal, Vol. 15, pp. 45-62 levinthal, D. (1995): "Strategic management and the exploration of diversity", in: Montgomery, C (ed.): Resource-based and Evolutionary Theories of the Firm: Toward a Synthesis; Kluwer, Boston I MA, pp. 1942 levy, D. (1994): "Chaos Theory and Strategy: Theory, Application, and Managerial Implications"; Strategic Management Journal, Vol. 15, pp. 167-178 Li, S.-C.S. I Chiang, C.-C. (2001): "Market Competition and Programming Diversity: A Study on the TV Market in Taiwan"; Journal of Media Economics, Vol. 14 (2), pp. 105-119 Li-Chuan, M. (2001): An Analysis of Corporate Culture and Business Strategy in Taiwanese CATV Channel Companies, Ph.D. Dissertation, University of Westminster, London Lippman, SA I Rumelt, R.P. (1982): "Uncertain imitability: an analysis of interfirm differences in efficiency under competition"; Bell Journal of Economics, Vol. 13, pp. 418-38 Litman, B.R. (1985): "Economic methods of broadcasting research", in: Dominick, J.R. I Fletcher, J.E. (eds.): Broadcasting research methods; Allyn & Bacon, Boston, pp. 107-122 litman, B.R. (2000): "The Structure of the Film Industry - Windows of Exhibition", in: Greco, A. (ed.): The Media and Entertainment Industries, Boston, pp. 99-121 Littlechild, S.C. (1978): The fallacy of the mixed economy; The Institute of Economic Affairs, Sussex I UK lopes, P,D. (1992): "Innovation and diversity in the popular music industry: 1986 to 1990"; American Sociological Review, Vol. 57, pp. 56-71 lowendahl, B, I Revang, 0, (1998): "Challenges to existing strategy theory in a post industrial society"; Strategic Management Journal, Vol. 19 (8), pp. 755-774 low, Linda (2000): The Economies of Information Technology and the Media; Singapore University Press lubatkin, M,H, I Chatterjee, S, (1994): "Extending modern portfolio theory into the domain of corporate diversification: Does it apply?"; Academy of Management Journal, Vol. 37, pp. 109-136

262

Lunn, J. (1989): "R&D, concentration and advertising: a simultaneous equations model"; Managerial and Decision Economics, Vol. 10, pp. 101-105 Luthans, F. I Stewart, T.I. (1977): ,A General Contingency Theory of Management"; Academy of Management Review, Vol. 2 (2), pp. 181-195 Lyles, M.A. I Mitroff,l.I. (1980): "Organizational Problem Formulation: An Empirical Study"; Administrative Science Quarterly, Vol. 25, pp. 102-119 MacDonald, J.M. (1985): "R&D and the direction of diversification"; Review of Economics and Statistics, Vol. 67, pp. 583-590 MacKinnon, D.P. I Warsi, G. I Dwyer, J.H. (1995): "A simulation study of mediated effect measures"; Multivariate Behavioral Research, Vol. 30, pp. 41-62 MacKinnon, D.P. et al (2002): "A comparison of methods to test mediation and other intervening variable effects"; Psychological Methods, Vol. 7, pp. 83-104 Maddigan, R.J. (1981): "The measurement of vertical integration"; Review of Economics and Statistics, Vol. 63, pp. 328-335 Maddigan, R.J. I Zaima, J.K. (1985): "The Profitability of Vertical Integration"; Managerial and Decision Economics, Vol. 6 (3), pp. 178-179 Mahoney, J.T. (1992): "The Choice of organizational form: Vertical Financial Ownership versus other Methods of Vertical Integration"; Strategic Management Journal, Vol. 13, pp. 559-584 Mahoney, J.T. I Pandian, J.R. (1992): "The resource-based view within the conversation of strategic management"; Strategic Management Journal, Vol. 13, pp. 363-380 Mai, L.C. (2002): "Company Size, Operational Type, Ownership Structure and Business Strate~y - An Analysis of Taiwanese Satellite Channel Companies"; Unpublished Paper presented at the 5' World Media Economics Conference in Turku, Finland, May 2002 Makadok, R. (2001): "Toward a synthesis of the resource-based and dynamic-capability views of rent creation"; Strategic Management Journal, Vol. 22, pp. 387-401 Malik, F. I Probst, G. (1982): "Evolutionary Managemenf'; Cybernetics and Systems, Vol. 13, pp. 153174. Mann, M.H. (1966): "Seller Concentration, Barriers to Entry, and Rates of Return in 30 Industries, 19501960"; Review of Economics and Statistics, Vol. 48 (August), pp. 296-307 March, J.P.I Olsen, J.G. (1989): Rediscovering Institutions, Free Press, New York I NY Marcoulides, G.A.I Heck, R.H. (1993): "Organizational Culture and Performance: Proposing and Testing a Model"; Organization Science, Vol. 4 (2), pp. 209-225 Markham, J.W. (1973): Conglomerate Enterprise and Economic Performance; Harvard University Press, Cambridge / MA Markides, C.C. (1992): "Consequences of corporate refocusing: Ex ante evidence"; Academy of Management Journal, Vol. 35, pp. 398-412 Markides, C.C. I Williamson, P.J. (1994): "Related diversification, core competencies and corporate performance"; Strategic Management Journal, Summer Special Issue, Vol. 15, pp. 149-165 Markides, C.C. (1995): "Diversification, restructuring and economic performance"; Strategic Management Journal, Vol. 16 (2), pp.101-118

263

Marlin, D. I Hoffman, J.J. I lamont, B.T. (1994): "Porter's Generic Strategies, Dynamic Environments, and Performance: A Profile Deviation Fit Perspective"; The International Journal of Organizational Analysis, Vol. 2, April, pp. 155-175 Marshall, A. (1948): Principles of Economics, Macmillan, London Martin, S. (1979): "Entry Barriers, Concentration and Profitability", Southern Economic Journal, Vol. 46 (October), pp. 471-88 Marx, K. (1887): Capital: A Critique of Political Economy; Progress Publishers, Moscow Mason, E.S. (1939): "Price and Production Policies of Large-Scale Enterprise"; American Economic Review, Vol. 29 (March), pp. 61-74 Matsuno, K.I Mentzer, J.T. (2000): "The effects of Strategy Type on the Market Orientation-Performance Relationship"; Journal of Marketing, Vol. 64 (4), pp. 1-17 Mauri, A. I Michaels, M. (1998): "Firm and industry effects within strategic management: an empirical examination"; Strategic Management Journal, Vol. 19, pp. 211-219 McCabe, D.L. (1990): 'The assessment of perceived environmental uncertainty and economic performance"; Human Relations, Vol. 43, pp. 1203-1218 McCutcheon, B.J. (1991): What caused conglomerate formation? An examination of managerial behavior and internal capital markets in the 1960s conglomerates; Doctoral Dissertation, University of Chicago McDaniel, S.W. I Kolari, J.W. (1987): "Marketing strategy implications of the Miles and Snow strategic typology"; Journal of Marketing, Vol. 51 (4), pp. 19-30 McFayden, S.I Hoskins, C. I Gillen, D. (1980): Canadian Broadcasting: Market Structure and Economic Performance; Institute for Research on Public Policy, Montreal McGahan, A.M. I Porter, M. (1997): "How much does industry matter, really?"; Strategic Management Journal, Vol. 18, pp. 15-30 McGahan, A.M. I Porter, M. (2002): 'What do we know about variance in accounting profitability?"; Management Science, Vol. 48 (7), pp. 834-851 McGee, J. I Thomas, H. (1986): "Strategic Groups: Theory, Research and Taxonomy"; Strategic Management Journal, Vol. 7, pp. 141-160 McGuire, J.B. I Schneeweis, T. I Branch, B. (1990): "Perceptions of Firm Quality: A cause or result of firm performance"; Journal of Management, Vol. 16, pp. 167-180 McKelvey, B.I Aldrich, H.E. (1983): "Populations, Natural Selection and Applied Organizational Science"; Administrative Science Quarterly, Vol. 28, pp. 101-128. McKelvey, B. (1999): "Complexity Theory in Organization Science: Seizing the Promise or Becoming a Fad?"; Emergence, Vol. 1 (1), pp. 5-32 McNamara, G. I Vaaler, P.M. I Devers, C. (2003): "Same as it ever was: The Search for Evidence of Increasing Hypercompetition"; Strategic Management Journal, Vol. 24, pp. 261-278 McQuivey, J.l. I McQuivey, M.K. (1998): "Is it a small publishing world after all?: Media Monopolization of the Children'S Book Market, 1992-1995"; Journal of Media Economics, Vol. 11 (4), pp. 35-48 McWilliams, A. I Smart, D.L. (1993): "Efficiency v. Structure-Conduct-Performance: Implications for Strategy Research and Practice"; Journal of Management, Vol. 19 (March), pp. 63-78

264

Melcher, A.J. I Melcher, S.H. (1980): "Towards a Systems Theory of Policy Analysis: StatiC versus Dynamic Analysis"; Academy of Management Review, Vol. 5 (2), pp. 235-247 Melin, L. (1992): "Internationalization as a Strategy Process"; StrategiC Management Journal, Vol. 13, Winter Special Issue, pp. 99-118 Meyer, A.D. (1982): ,,Adapting to environmental jolts"; Administrative Science Quarterly, Vol. 27, pp. 51537 Meyer, M.P.I Milgrom, P.I Roberts, J. (1992): "Organizational prospects, influence costs, and ownership changes"; Journal of Economics and Management Strategy, Vol. 1, pp. 9-35 Michel, A.I Shaked, I. (1984): "Does business diversification affect performance"; Financial Management, Vol. 13, pp. 18-24 Michel, A. I Shaked, I. (1986): "Multinational corporations vs. domestic corporations: Financial performance and characteristics"; Journal of International Business Studies, Vol. 16 (Fall), pp. 89-106 Miles, R.E.I Snow, C.C. (1978): Organizational Strategy, Structure, and Process, N.Y I SI. Louis Miles, R.E. I Cameron, K. (1982): Coffin Nails and Corporate Strategies, Englewood Cliffs, Prentice Hall, New Jersey Milgrom, P. I Roberts, J. (1986): Price and advertising signals of product quality"; Journal of Political Economy, Vol. 94, pp. 796-821 Miller, D. (1986): "Configurations of Strategy and Structure: Towards a Synthesis"; Strategic Management Journal, Vol. 7, No.3, pp. 233-249 Miller, D. (1988): "Relating Porter's Business Strategies to Environment and Structure: Analysis and Performance Implications"; Academy of Management Journal, Vol. 39 (3), pp. 280-308 Miller, D. (1992): "The generic strategy trap"; Journal of Business Strategy, January-February, pp. 37-41 Miller, D. I Friesen, P.H. (1980): "Momentum and Revolution in Organizational Adaptation"; Academy of Management Journal, Vol. 23 (4), pp. 591-614 Miller, D. I Friesen, P.H. (1982a): "Innovation in conservative and entrepreneurial firms: Two models of strategic momentum"; Strategic Management Journal, Vol. 3, pp. 1-25 Miller, D. I Friesen, P.H. (1982b): "The Longitudinal Analysis of Organizations: A methodological perspective"; Management Science, Vol. 28 (9), pp. 1013-1034 Miller, 0.1 Friesen, P.H. (1984): Organizations: A quantum view; Prentice-Hall, Englewood Cliffs I NJ Miller, D. I Shamsie, J. (1996): "The Resource-based View of the Firm in two environments: The Hollywood Film Studios from 1936 and 1965", Academy of Management Journal, Vol. 39 (3), pp. 519-543 Miller, R.A. (1973): "Concentration and marginal concentration, advertising, and diversity: three issues in structure-performance tests"; Industrial Organization Review, Vol. 1, pp. 15-24 Miller, R. (1981): "Strategic pathways to growth in retailing"; Journal of Business Strategy, Vol. 1 (3), pp. 16-29 Mintzberg, H. (1991): "Learning 1, Planning 0: Reply to Igor Ansoff'; StrategiC Management Journal, Vol. 12 (6), pp. 463-466 Mintzberg, H. (1993): The Rise and Fall of Strategic Planning, Macmillan Free Press, New York

265

Mitchell, E.J. (1976): Vertical Integration in the Oil Industry; AEI National Energy Study No. 11, American Enterprise Institute for Public Policy Research, Washington DC Molyneux, P. I Forbes, W. (1995): "Market structure and performance in European banking"; Applied Economics, Vol. 27, pp. 155-159 Montgomery, C.A. (1985): "Product-market diversification and market power"; Academy of Management Journal, Vol. 28, pp. 789-798 Montgomery, C.A. I Wernerfelt, B. (1991): "Sources of Superior Performance: Market Share versus Industry Effects in the US Brewing Industry"; Management Science, Vol. 37 (8), pp. 954-958 Morck, R. I Yeung, B. (1991): 'Why investors value multinationality"; Joumal of Business, Vol. 64 (2), pp. 165-188 Morgan, R.E. I Strong, C.A. (1998): "Market orientation and dimensions of strategic orientation"; European Joumal of Marketing, Vol. 32 (11/12), pp. 1051-73 Murray, A.I. (1988): "A Contingency View of Porter's 'Generic Strategies"'; Academy of Management Review, Vol. 13 (3), pp. 390-400 Narver, J.C. I Slater, S. (1990): "The effect of marketing orientation on business profitability"; Joumal of Marketing, Vol. 54, pp. 20-35 Nayyar, P. (1992): "On the measurement of corporate diversification strategy: Evidence from large U.S. service firms"; Strategic Management Journal, Vol. 13 (3), pp. 219-235 Neiva, E.M. (1996): "Chain Building: The Consolidation of the American Newspaper Industry, 1953-1980"; Business History Review, Vol. 70, pp. 1-42 Nelson, R.R. I Winter, S. (1982): An evolutionary theory of economic change; Harvard University Press, Cambridge Nelson, R.R. (1991): 'Why do firms differ and how does it matter?"; Strategic Management Journal, Vol. 12, pp. 61-74 Nelson, R.R. (1995): "Recent Evolutionary Theorizing About Economic Change"; Journal of Economic Literature; Vol. 33 (March), pp. 48-90 Neumann, I. (1998): PayTV in Deutschland - Markteintritts- und Wettbewerbsbedingungen fOr neue Anbieter; Wiesbaden Nunnally, J.C. (1978): Psychometric Theory, McGraw-Hili, New York Ogbonna, E. I Harris, L.C. (2000): "Leadership style, organizational culture and performance: Empirical evidence from UK companies"; International Journal of HR Management, Vol. 11 (4), pp. 766-788 Okasha, S. (2002): Philosophy of Science: A Very Short Introduction; Oxford University Press, Oxford / UK Ouchi, W.G. (1981): Theory Z; Addison-Wesley, Reading / Mass. Owen, B.M. I Beebe, J.H. / Manning, W.G. (1974): Television Economics, Heath, Lexington, MA Owen, B.M. / Wildman, 5.5. (1992): Video Economics, Cambridge / Mass. and London Patzold, U. I Roper, H. (1999): "Fernsehproduktionsvolumen in Deutschland 1998"; Media Perspektiven, Vol. 9, pp. 447-468

266

Palepu, K. (1985): "Diversification strategy, profit performance, and the entropy measure"; Strategic Management Joumal, Vol. 6 (3), pp. 239-255 Palich, L.E. I Cardinal, L.B. I Miller, C.C. (2000): "Curvilinearity in the Diversification-Performance Linkage: An Examination of over three Decades of Research"; Strategic Management Joumal, Vol. 21, pp. 155-174 Parnell, J.A. I Wright, P. (1993): "Generic Strategy and Performance: an Empirical Test of the Miles and Snow Typology"; British Joumal of Management, Vol. 4, pp. 29-36 Parnell, J.A. I Wright, P. I Tu, H.S. (1996): "Beyond the strategy-performance linkage: The Impact of the Strategy-Organization-Environment Fit on Business Performance"; American Business Review, June, pp. 41-50 Parnell, J.A. (1997): "New Evidence in the Generic Strategy and Business Performance Debate: A Research Note"; British JoumalofManagement, Vol. 8, pp. 175-181 Parnell, J.A. I Lester, D.L. I Menefee, M.L. (2000): "Strategy as a response to organizational uncertainty: an alternative perspective on the strategy-performance relationship"; Management Decision, Vol. 38 (8), pp.520-530 Pascale, R.T.I Athos, A.G. (1981): The art of Japanese Management; Simon & Schuster, New York / NY Pelham, A.M. (1997): "Market Orientation and Performance: The Moderating Effect of Product and Customer Differentiation"; Joumal of Business and Industrial Marketing, Vol. 12 (5), pp. 276-296 Pennings, J.M. / Barkema, H.I Donma, S. (1994): "Organizational learning and diversification"; Academy of Management Joumal, Vol. 37 (3), pp. 608-640 Penrose, E. (1959): The Theory of Growth of the Firm, Blackwell, Oxford Peteraf, M.A. (1993): "The cornerstones of competitive advantage: a resource-based view"; Strategic Management Joumal, Vol. 14, pp. 179-91 Peters, T.J. / Waterman, R.H. (1982): In search of excellence. Lessons from America's best run companies. Harper & Row, New York / NY Petroni, A. (1998): "The analysis of dynamic capabilities in a competence-oriented organization"; Technovation, Vol. 18 (3), pp. 179-189 Pettigrew, A.M. (1987): "Context and Action in the Transformation of the Firm"; Journal of Management Studies; Vol. 24 (6), pp. 649-670 Pettigrew, A.M. / Thomas, H. / Whittington, R. (2002): "Strategic Management: The Strengths and Limitations of a Field"; in: Pettigrew, A. et al (eds.): Handbook of Strategy and Management; Sage, London Phelan, S.E. (2001): ''What is Complexity Really?"; Emergence, Vol. 3 (1), pp. 120-136 Phelps, R. (2001): "JBR Special Issue on the dynamics of strategy: Introduction"; Journal of Business Research, Vol. 51 (3), pp. 167-168 Phillips, L.W. (1981): "Assessing Measurement Error in Key Informant Reports"; Journal of Marketing Research, Vol. 18, pp. 395-415 Phillips, L.W. / Chang, D.R. I Buzzell, R.D. (1983): "Product Quality, Cost Position, and Business Performance. A Test of Some Key Hypotheses"; Journal of Marketing, Vol. 47, Spring, pp. 26-43 Picard, R.G. (1988): "Measures of concentration in the daily newspaper industry"; Journal of Media Economics, Vol. 1, pp. 61-74

267

Picard, R.G. (1989): Media Economics: Concepts and Issues; Sage Publications. Thousand Oaks. California Picard, R.G. I Rimmer, T. (1999): 'Weathering a Recession: Effects of Size and Diversification on Newspaper Companies"; Journal of Media Economics, Vol. 12 (1). pp. 1-18 Pisano, G.P. (1994): "Knowledge. Integration. and the Locus of Learning: An Empirical Analysis of Process Developmenf'; Strategic Management Journal. Vol. 15. pp. 85-100 Pitts, R.A. (1977): "Strategies and structures for diversification"; Academy of Management Journal. Vol. 20, pp. 197-208 Pitts, R.A. I Hopkins, H.D. (1982): "Firm diversity: conceptualization and measurement"; Academy of Management Review, Vol. 7. pp. 620-627 Podolny, J.M. (1993): .A status-based model of market competition"; American Journal of Sociology, Vol. 89. pp. 829-872 Popper, K. (1959): Logic of Scientific Discovery; Hutchinson, London Porter, M.E. (1976): "Please note location of nearest exit: Exit barriers and planning"; California Management Review, Vol. 19, Winter, pp. 21-33 Porter, M.E. (1979): "The structure within industries and companies' performance"; Review of Economics & Statistics, Vol. 61, May, pp. 214-227 Porter, M.E. (1980): Competitive strategy: Techniques for analyzing industries and competitors, Free Press, New York (10th ed.) Porter, M.E. (1981): "The Contributions of Industrial Organization to Strategic Management", Academy of Management Review, Vol. 6, pp. 609-620 Porter, M.E. (1985): Competitive Advantage, The Free Press, New York I NY Porter, M.E. (1987): "From Competitive Advantage to Corporate Strategy"; Harvard Business Review, Vol. 65, May-June, pp. 43-59 Porter, M.E. (1991): "Towards a dynamic theory of strategy"; Strategic Management Journal, Vol. 12, Winter Special Issue, pp. 95-117 Powell, J.H. I Wakeley, T.M. (2003): "Evolutionary Concepts and Business Economics: Towards a normative approach"; Journal of Business Research, Vol. 56, pp. 153-161 Prabhu, S. I Harrison, J.S. (1992): "Vertical Integration Strategies in Growing Industries: An Empirical Study"; American Business Review, January, pp. 17-22 Prahalad, C.K. I Bettis, R.A. (1986): "The dominant logic: a new linkage between diversity and performance"; Strategic Management Journal, Vol. 7, pp. 485-501 Prahalad, C.K. I Hamel, G. (1990): "The Core Competence of the Corporation"; Harvard Business Review, Vol. 68 (3), pp. 79-91 Prahalad, C.K. I Hamel, G. (1990): "Strategy as a field of study: why search for a new paradigm?"; Strategic Management Journal, Special Issue, Vol. 15, pp. 5-16 Prescott, J.E. I Kohli, A.K. I Venkatraman, N. (1986): "The Market Share-Profitability Relationship: An Empirical Assessment of Major Assertions and Contractions"; Strategic Management Journal, Vol. 7, JulyAugust, pp. 377-394

268

Priem, R.L. I Butler, J.E. (2001): "Is the Resource-based "view" a useful perspective for strategic management research?"; Academy of Management Review, Vol. 26 (1), pp. 22-40 Probst, G.J.B.I Buchel, B. I Raub, S.P. (1998): "Knowledge as a strategic resource", in: v. Krogh, G. / Roos, J. / Kleine, D. (eds.): Knowing in Firms; Sage Publications, London, pp. 240-252 Probst, G.J.B.I Raisch, S. (2004): "Die Logik des Niedergangs"; Harvard Business Manager, March, pp. 37-45. Putnam, H. (1991): "The 'corroboration' of theories"; in: Boyd, R. et al. (eds.): The Philosophy of Science, MIT Press, Cambridge / MA Quinn, R.E. (1988): Beyond Rational Management; Jossey-Bass, San Francisco / CA Ramanujam, V. / Varadarajan, P. (1989): "Research on Corporate Diversification: A Synthesis"; Strategic Management Journal, Vol. 10, pp. 523-551 Ravenscraft, D.J. I Scherer, F.M. (1982): ''The Lag Structure of Returns to Research and Developmenf'; Applied Economics, Vol. 14. December, pp. 603-620 Ravenscraft, D.J. I Scherer, F.M. (1987): Mergers, Sell-Offs and Economic Efficiency; The Brookings Institution, Washington / DC Reed, R. I DeFillippi, R.J. (1990): "Causal Ambiguity, Barriers to Imitation, and Sustainable Competitive Advantage"; Academy of Management Review, Vol. 15, pp. 88-102 Reed, R. I Fronmueller, M.P. (1990): 'Vertical Integration: A Comparative Analysis of Performance and Risk"; Managerial and Decision Economics, Vol. 11, pp. 177-185 Rhoades, S.A. (1973): "The effect of diversification on industry profit performance in 241 manufacturing industries: 1963"; Review of Economics and Statistics, Vol. 55 (2), pp. 146-155 Rhoades, S.A. (1974): "A further evaluation of the effect of diversification on industry profit performance"; Review of Economics and Statistics, Vol. 56 (4), pp. 557-559 Rindova, V.P. I Fombrun, C.J. (1999): "Constructing competitive advantage: the role of firm-constituent interactions"; Strategic Management Journal, Vol. 20 (8), pp. 691-710 Roberts, P.W. I Dowling, G.R. (2002): "Corporate Reputation and Sustained Financial Performance"; Strategic Management Journal, Vol. 23, pp. 1077-1093 Robinson, K.C. I McDougall, P.P. (1998): ''The Impact of Alternative Operationalizations of Industry Structural Elements on Measures of Performance for Entrepreneurial Manufacturing Ventures"; Strategic Management Journal, Vol. 19, pp. 1079-1100 Roper, H. (2000): "Zeitungsmarkt 2000: Konsolidierungsphase beendet?", in: Media Perspektiven, Vol. 7, pp.297-309 Roper, H. (2002): "Formationen Deutscher Medienmultis: Entwicklungen und Strategien der groBten deutschen Medienunternehmen", Media Perspektiven, Vol. 9, pp. 406-432 Romanelli, E. I Tushman, M.L. (1994): "Organizational Transformation as Punctuated Equilibrium: An Empirical Tesr'; Academy of Management Journal, Vol. 37 (5), pp. 1141-1166 Roquebert, J.A. I Phillips, R.l. I Westfall, P.A. (1996): "Markets vs. Management: What Drives Profitability?"; StrategiC Management Journal, Vol. 17 (8). October, pp. 653-664 Rosenbaum, 0.1. I Lamort, F. (1992): "Entry, Exit, Barriers and Sunk Costs: an analysis"; Applied Economics. Vol. 24, pp. 297-304

269

Rosenberg, A. (2000): Philosophy of Science: A Contemporary Introduction; Routledge, London Roth, K. I Morrison, A.J. (1992): "Business-level Competitive Strategy: A Contingency link to Internationalization"; Journal of Management, Vol. 18 (3), pp. 473-488 Rueckert, R.W. (1992): "Developing a market orientation: an organisational strategy perspective"; International Journal of Research in Marketing, Vol. 9, pp. 225-45 Rugman, A.M. (1979): "International diversification and the multinational enterprise"; Lexington Books, Lexington I MA Rugman, A.M. (1981): Inside the Multinationals: The Economics of International Markets, Croom Helm, London Rumelt, R.P. (1974): Strategy, Structure, and Econornic Performance; Harvard University Press, Cambridge / MA Rumelt, R.P.I Wensley, R. (1981): "In the Search of the Market Share Effect"; Proceedings of the Annual Meeting of the Academy of Management, Detroit Rumelt, R.P. (1982): "Diversification strategy and profitability"; Strategic Management Journal, Vol. 3 (4), pp.359-369 Rumelt, R.P. (1984): "TowardS a strategic theory of the firm"; in: Lamb, R.B. (ed.): Competitive Strategic Management, Prentice-Hall, Englewood-Cliffs / NJ, pp. 566-70 Rumelt, R.P. (1987): "Theory, strategy and entrepreneurship", in: Teece, D.J. (ed.): The Competitive Challenge, Harper & Row, New York / NY, pp. 137-58 Rumelt, R.P. (1991): "How much does Industry matter?", Strategic Management Journal, Vol. 12, pp. 16785 Salmon, W. (1989): "Four decades of Scientific Explanation"; in: Kitcher, P. and Salmon, W. (eds.): Scientific Explanation, Minnesota Studies in the Philosophy of Science, Vol. 13, Minneapolis University of Minnesota Press Sanchez, R. I Heene, A. (1997): "Reinventing Strategic Management: New Theory and Practice for Competence-based Competition"; European Management Journal, Vol. 15 (3), pp. 303-317 Schein, E.H. (1992): Organizational Culture and Leadership; Jossey-Bass, San Francisco Schendel, D. I Hatten, K. (1972): "Business policy or strategic management: A broader view for an emerging discipline"; Academy of Management Proceedings, August 1972 Schendel, D.I Hofer, C.W. (1979): Strategy Formulation; West Publishing Co, S!.Paul Schendel, D. (1994): "Competitive organizational behavior: toward an organizationally-based theory of competitive advantage"; Strategic Management Journal, Winter Special Issue, Vol. 15, pp. 1-4 Scherer, F.M.I Ross, D. (1980): Industrial Market Structure and Economic Performance; Rand McNally & Co Schmalensee, R. (1985): "Do Markets differ Much?"; American Economic Review, Vol. 75, pp. 341-351 Schoeffler, S. I Buzzell, R.D. I Heany, D.F. (1974): "Impact of Strategic Planning on Profit Performance"; Harvard Business Review, Vol. 52, March-April, pp. 137-145 Schoeffler, S. (1977): "Cross-Sectional Study of Strategy, Structure, and Performance: Aspects of the PIMS Program"; in: Thorelli, H.B. (ed.): The Strategic Planning Imperative, Indiana University Press, Bloomington

270

Schoemaker, P.H.J I Amit, R.H. (1994): "Investment in Strategic Assets: Industry and Firm-level Perspectives"; in: Shrivastava, P. et al (eds.): Advances in Strategic Management: Resource-based View of the Firm; JAI Press Inc, Greenwich ICT Schum peter, J.A. (1939): Business Cycles: A Theoretical, Historical, and Statistical Analysis of the Capitalist Process; McGraw-Hili, New York I NY Schumpeter, J.A. (1942): Capitalism. Socialism, and Democracy; George Allen & Unwin, London Schutz, W.J. (2000): "Deutsche Tagespresse 1999"; Media Perspektiven, Vol. 1, pp. 8-29 Seale, C. (1999): The quality of qualitative research; Sage, London Seidler, J. (1974): "On Using Informants: A Technique for Collecting Quantitative Data and Controlling for Measurement Error in Organizational Analysis"; American Sociological Review, Vol. 39, pp. 816-831 Senge, P.M. (1990): The Fifth Discipline: The Art & Practice of the Learning Organization; Doubleday, New York Servaes, H. (1996): "The value of diversification during the conglomerate merger wave"; Journal of Finance, Vol. 51, pp. 1201-1225 Seth, A. (1990): "Value creation in acquisitions: A reexamination of performance issues"; Strategic ManagementJourna/, Vol. 11 (2), pp. 99-115 Seth, A. I Thomas, H. (1994): "Theories of the firm: implications for strategy research"; Journal of Management Studies, Vol. 31 (2), pp. 165-191 Shadish, W.R.I Sweeney, R.B. (1991): "Mediators and moderators in meta-analysis: There's a reason we don't let dodo birds tell us which psychotherapies should have prizes"; Journal of Consulting and Clinical Psychology, Vol. 59, pp. 883-893 Shapiro, B.P. (1988): "What the hell is market oriented?"; Harvard Business Review, NovemberDecember, pp. 119-125 Shay, J.P. I Rothaermel, F.T. (1999): "Dynamic Competitive Strategy: Towards a Multi-perspective Conceptual Framework"; Long Range Planning, Vol. 32 (6), pp. 559-572 Sherman, B.L. (1995): Telecommunications Management: Broadcasting-Cable & The New Technologies; McGraw-Hili, New York Shrout, P.E. I Bolger, N. (2002): "Mediation in Experimental and Nonexperimental Studies: New Procedures and Recommendations"; Psychological Methods, Vol. 7 (4), pp. 422-45 Silva, G. I Lisboa, J. I Yasin, M. (2000): "Effectiveness of Business Strategies in the Portuguese Culture: An Empirical Investigation"; Cross Cultural Management, Vol. 7 (4), pp. 33-40 Sjurts, I. (1996): Die deutsche Medienbranche - Eine unternehmensstrategische Analyse; Gabler Verlag, Wiesbaden Slater, S.F. I Narver, J.C. (1994): "Does competitive environment moderate the market orientationperformance relationship?"; Journal of Marketing, Vol. 58, pp. 46-55 Slater, S.F.I Narver, J.C. (1995): "Market orientation and the learning organization"; Journal of Marketing, Vol. 59, pp. 63-74 Slater, S.F. I Narver, J.C. (1999): "Market-oriented is more than being customer-led"; Strategic Management Journal, Vol. 20, pp. 1165-1169

271

Slater, S.F. I NaIVer, J.e. (2000): "The Positive Effect of a Market Orientation on Business Profitability: A Balanced Replication"; Journal of Business Research, Vol. 48, pp. 69-73 Smirlock, M. (1985): "Evidence on the (non-) relationship between concentration and profitability in banking", Journal of Money. Credit, and Banking, Vol. 17, pp. 69-83 Smith, A. (1991): The Age of Behemoths: The Globalization of Mass Media Firms; Priority Press, New York Snow, C.C. I Hrebiniak, L.G. (1980): "Strategy, distinctive competence, and organizational performance"; Administrative Science Quarterly, Vol. 25, pp. 317-336 Sobel, M.E. (1982): "Asymptotic Confidence Intervals for Indirect Effects in Structural Equations Models"; in: Leinhardt, S. (Ed): Sociological Methodology, Jossey-Bass, San Francisco, pp. 290-312 Sobel, M.E. (1988): "Direct and indirect effects in linear structural equations models"; in: Long, J.S. (Ed): Common problems I Proper solutions, Sage Publications, Beverly Hills I CA, pp. 46-64 Sorensen, J.B. (2002): "The Strength of Corporate Culture and the Reliability of Firm Performance"; Administrative Science Quarterly, Vol. 47, pp. 70-91 Sorenson, O. (2003): "Interdependence and Adaptability: Organizational Learning and the Long-Term Effect of Integration"; Management Science, Vol. 49 (4), pp. 446-463 Sparks I Sparks (1995): "Concentration and Market Entry in the UK National Daily Press"; European Journal of Communication, Vol. 10 (2), June, pp. 179-206 Spender, J.C. (1989): Industry Recipes: An Enquiry into the Nature and Sources of Managerial Judgment, Basil Blackwell, Oxford Spender, J.C. (1996): "Making Knowledge the Basis of a Dynamic Theory of the Firm"; Strategic Management Journal, Vol. 17, Winter Special Issue, pp. 45-62 Stacey, R.D. (1995): "The Science of Complexity: An Alternative Perspective for Strategic Change Processes"; Strategic Management Journal, Vol. 16 (6), pp. 477-95 Steer, P.I Cable, J.R. (1978): "Internal Organization and Profit: A Note"; Journal of Industrial Economics, Vol. 27, pp. 13-30 Sterman, J.D. (2001): "System Dynamics Modeling: Tools for Learning in a Complex World"; California Management Review, Vol. 43 (4), pp. 8-25 Stigler, G.J. (1962): "Information in the labor market"; Journal of Political Economy, Vol. 70, pp. 49-73 Stigler, G.J. (1963): Capital and Return in Manufacturing Industries, Princeton University Press, Princeton Stimpert, J.L. I Duhaime, I.M. (1997): "Seeing the Big Picture: The Influence of Industry, Diversification, and Business Strategy on Performance"; Academy of Management Journal, Vol. 40 (3), pp. 560-583 Stock, J.H.I Watson, M.W. (2003): Introduction to Econometrics; Pearson Educations, New York Stulz, R. (1990): "Managerial Discretion and Optimal Financing Policies"; Journal of Financial Economics; Vol. 26, pp. 3-27 Summer, C.E. et al (1990): "Doctoral education in the field of business policy and strategy"; Journal of Management, Vol. 16, pp. 361-398 Suppe, F. (1977): The Structure of Scientific Theories; University of Illinois Press, Chicago

272

Tallman, S. I Li, J. (1996): "Effects of international diversity and product diversity on the performance of multinational firms"; Academy of Management Journal, Vol. 39, pp. 179-196 Teece, D.J. (1981): "Internal Organization and Economic Performance: An Empirical Analysis of Large UK Companies"; Journal of Industrial Economics, Vol. 30, pp. 173-200 Teece, D.J. (1984): "Economic Analysis and Strategic Management"; California Management Review, Vol. 26 (Spring), pp. 87-110 Teece, D.J. I Pisano, G. (1994): "The dynamic capabilities of firms: an introduction"; Industrial and Corporate Change, Vol. 3, pp. 537-556 Teece, D.J. I Pisano, G.I Shuen, A. (1997): "Dynamic Capabilities and Strategic Management"; Strategic Management Journal, Vol. 18 (7), pp. 509-533 Teece, D.J. (1998): "Capturing Value from Knowledge Assets: The New Economy, Markets for Know-How, and Intangible Assets"; California Management Review, Vol. 40 (3), pp. 55-80 Thil!rart, R.A. I Forgues, B. (1995): "Chaos Theory and Organization"; Organization Science, Vol. 6 (1), pp.19-31 Thomas, H. I Venkatraman (1988): "Research on Strategic Groups: Progress and Prognosis"; Journal of Management Studies, Vol. 25 (6), pp. 537-553 Thomas, L.G. (1996): "The Two Faces of Competition: Dynamic Resourcefulness Hypercompetitive Shift"; Organization Science, Vol. 7 (3), Special Issue, pp. 221-242

and

the

Thompson, J. (1967): Organizations in Action; McGraw-Hill, New York I NY Thompson, J. (1993): Strategic Management: Awareness and Change, Chapman & Hall, London Thompson, R.S. (1981): "Internal Organization and Profit: A Note"; Journal of Industrial Economics, Vol. 30, pp. 201-11 Thompson, V.A. (1965): "Bureaucracy and Innovation"; Administrative Science Quarterly, Vol. 5, pp. 1-20 Tidd, J. (2001): "Innovation management in context: environment, organization, and performance"; International Journal of Management Reviews, Vol. 3 (3), pp. 169-183 Tunstall, J.I Palmer, M. (1991): Media Moguls, Routledge, London Turner, G. I Spencer, B. (1997): "Understanding the Marketing Concept as an Organizational Culture"; European Journal of Marketing, Vol. 31 (2), pp. 110-121 Tushman, M.L. I Anderson, P. (1986): "Technological Discontinuities and Organizational Environments"; Administrative Science Quarterly, Vol. 31 (3), pp. 439-465 Tushman, M.L. I Newman, W.H. I Romanelli, E. (1986): "Convergence and Upheaval: Managing the unsteady pace of organizational evolution"; California Management Review, Vol. 29 (1), pp. 29-44 Utterback, J.M. (1994): "Radical Innovation and Corporate Regeneration"; Research Technology Management, Vol. 37 (4), pp. 10-18 Van der Post, W.Z. I de Coning, T.J. (1998): "The relationship between organizational culture and financial performance: Some South African Evidence"; South African Journal of Business Management, Vol. 29 (1), pp. 30-41 Van Egeren, M.I O'Connor, S. (1998): "Drivers of Market Orientation and Perfonmance in Service Firms"; Journal of Service Marketing, Vol. 12 (1), pp. 39-58

273

Vazquez, R. I Santos, M.L. I Alvarez, L.1. (2001): "Market Orientation, Innovation, and Competitive Strategies in Industrial Firms"; Journal of Strategic Marketing, Vol. 9, pp. 69-90 Vernon, R. (1966): "International investment and international trade in the product cycle"; Quarterly Journal of Economics, Vol. 81, pp. 190-207 Vernon, R. (1971): Sovereignty at Bay: The Multinational Spread of US Enterprises; Basic Books, New York and London Vester, F. (1988): Leitmotiv Vernetztes Denken; Heine Verlag, MOnchen. Waterman, 0.1 Weiss, A.A. (1997): Vertical Integration in Cable Television, MIT Press & AEI Press Webb, 0.1 Pettigrew, A. (1999): "The Temporal Development of Strategy: Patterns in the U.K. Insurance Industry"; Organization Science, Vol. 10 (5), pp. 601-621 Weir, C. (1995): "Organizational Structure and Corporate Performance: An Analysis of Medium and Large UK Firms"; Management Decision, Vol. 33 (1). pp. 24-32 Weiss, L.W. (1971): "Quantitative Studies of Industrial Organization", in: Intriligator, M.D. (ed.): Frontiers of Quantitative Economics, North-Holland, Boston, pp. 362-403 WeiSS, L.W. (1974): "The Concentration-Profit Relationship and Antitrusf', in: Goldschmidt, H. et al. (eds): Industrial Concentration: The New Learning, Little Brown and Company, Boston, pp. 185-223 Weiss, A.M. I Anderson, E. I Macinnis, D.J. (1999): "Reputation management as a motivation for sales structure decisions"; Journal of Marketing, Vol. 63, pp. 74-89 Wernerfelt, B. (1984): "A resource-based view of the firm"; Strategic Management Joumal, Vol. 5, pp. 171-180 Wernerfelt, B. I Montgomery, C. (1986): 'What is an attractive industry?"; Management Science, Vol. 32 (10), pp. 1223-30 Wernerfelt, B. I Montgomery, C. (1988): "Tobin's q and the importance of focus in firm performance"; American Economic Review, Vol. 78, pp. 246-250 Wernerfelt, B. (1989): "From critical resources to corporate strategy"; Journal of General Management, Vol. 14, pp. 4-12 White, R.E. (1986): "Generic Business Strategies, Organizational Context and Performance: An Empirical Investigation", Strategic Management Journal, No.7. pp. 217-231 Whittington, R.I Mayer, M.I Curto, F. (1999): "Chandlerism in Post-war Europe: Strategic and Structural Change in France, Germany, and the UK, 1950-1993"; Industrial and Corporate Change 8, pp. 519-550 Wiener, N. (1961): Cybernetics; MIT Press, Cambridge. Williams, J.R. (1992): "How sustainable is your competitive advantage"; California Management Review, Vol. 34. pp. 29-51 Williamson, O.E. (1967): "Hierarchical Control and Optimum Firm Size"; Journal of Political Economy, Vol. 75, pp. 123-138 Williamson, O.E. (1971): "The vertical integration of production: Market failure considerations"; American Economic Review, Vol. 61, pp. 112-123 Williamson, O.E. (1999): "Strategy Research: Governance and Competence Perspectives"; Strategic Management Journal, Vol. 20 (12), pp. 1087-1108

274

Winter, S. (1987): "Skill and knowledge as strategic assets"; in: Teece, D. (ed.): The Competitive Challenge; Ballinger, Cambridge I MA, pp. 159-184 Wirth, M.O. (1986): "Economic barriers to entering Media industries in the United States", in: Mc Laughlin, M. (ed.): Communication Yearbook; Vol. 9, SAJE, Beverly-Hills, CA, pp. 423-442 Wirth, M.O. I Bloch, H. (1995): "Industrial Organization Theory and Media Industry Analysis"; Journal of Media Economics, Vol. 8 (2), pp. 15-26 Wirtz, B.W. (1994): Neue Medien, Unternehmensstrategien und Wetlbewerb im Medienmarkl: eine wetlbewerbstheoretische und -politische Analyse; Verlag Peter Lang, Frankfurt am Main Wirtz, B.W. (2000): Medien- und Internetmanagement; Gabler Verlag, Wiesbaden Woo, C.Y.I Cooper, A.C. (1981): "Strategies of Effective Low Share Businesses"; Strategic Management Journal, Vol. 2, July-September, pp. 301-318 Woo, C.Y.I Cooper, A.C. (1982): "The Surprising Case of Low Market Share"; Harvard Business Review, Vol. 60, November-December, pp. 106-113 Woo, C.Y. (1984): "Market Share Leadership - Not Always So Good"; Harvard Business Review, Vol. 62, January-February, pp. 50-54 Wright, P. I Kroll, M. I Pringle, C. I Johnson, J. (1990): "Organizational types, conduct, profitability and risk in the semiconductor industry"; Journal of Management Systems, Vol. 2 (2), pp. 33-48 Wright, P. I Kroll, M. I Tu, H. I Helms, M. (1991): "Generic strategies and business performances: An empirical study of the screw machine products industry"; British Journal of Management, Vol. 2, pp. 57-65 Wrigley, L. (1970): Divisional autonomy and diversification; Doctoral Dissertation, Harvard Business School Yamin, S. I Gunasekaran, A. I Mavondo, F.T. (1999): "Relationship between generic strategies, competitive advantage and organizational performance: an empirical analysis"; Technovation, Vol. 19, pp. 507-518 Young, S. (1987): "Business Strategy and the Internationalization of Business: Recent Approaches"; Managerial and Decision Economics, Vol. 8, pp. 31-40 Young, G. I Smith, K.G. I Grimm, C.M. (1996): "Austrian and Industrial Organization Perspectives of Firm-level Competitive Activity and Performance"; Organization Science, Vol. 7 (3), pp. 243-255 Zajac, E. (1992): "Relating economic and behavioral perspectives in strategy research"; Advances in Strategic Management, Vol. 8, JAI Press I Conn., pp. 69-96 Zajac, E. I Kraatz, M.S. I Bresser, R.K.F. (2000): "Modeling the dynamics of strategiC fil: A normative approach to strategic change"; StrategiC Management Journal, Vol. 21, pp. 429-453 Zollo, M. I Winter, S.G. (2002): "Deliberate Learning and the Evolution of Dynamic Capabilities"; Organization Science, Vol. 13 (3), pp. 339-351 Zott, C. (2003): "Dynamic Capabilities and the Emergence of Intra industry Differential Firm Performance: Insights from a Simulation Study"; Strategic Management Journal, Vol. 24 (2), pp. 97-127

275

Appendix A: Mail Survey Questionnaire

277

THE GLOBAL MEDIA INDUSTRIES STRATEGIC ANALYSIS OF DETERMINANTS OF FIRM PERFORMANCE

This study is designed to produce a better understanding of how media company performance is impacted by the dynamic interaction between a firm's unique capabilities, its strategies at both the corporate and business levels, and the industry environment in which it operates. The findings are intended to support media companies in their future strategic decisions.

ALL INFORMATION WILL BE TREATED AS STRICTLY CONFIDENTIAL. DATA WILL BE ANALYZED ON INDUSTRY LEVEL ONLY.

Please return the completed questionnaire in the enclosed postage paid envelope to: University of Geneva HEC Management Studies Sebastian Raisch Bd du Pont-d'Arve 40 1211 Geneva 4 Switzerland

If you are interested in a summary of the findings, please feel free to contact us or provide your name and email contact directly:

THANK YOU VERY MUCH FOR TAKING THE TIME TO COMPLETE THIS SURVEY!

279

Please answer each question in the survey as complete as possible. If you cannot provide the exact answer to some questions, please provide an estimate. 1/6 The purpose of this section is to get an overview of the industry conditions in which your Strategic Business Unit (SBU) operates.

a)

Please estimate the market share currently held by the four largest firms in your industry (in %):

%

b) Please estimate the average annual growth rate of the primary market served by your SBU over the past three years (in %):

%

Please circle your response to each question: c) Economies of scale play an important role. Market leaders enjoy significantly reduced unit costs due to higher sales volumes.

Not At All 1 2

3

4

5

A Great Deal 6 7

d) Our business requires significant capital investment. This represents a considerable challenge to firms willing to enter the market.

1

2

3

4

5

6

7

e) Competing products & services in our industry are highly differentiated. Firms strongly invest in advertising to point out the differences.

1

2

3

4

5

6

7

216 The purpose of this section is to gain an understanding of the corporate strategy pursued by the company your SBU belongs to. Quite A Lot

a) How would you rate the number of businesses your company operates compared to your main competitors?

Very few

b) What percentage of overall corporate sales does the most important business of your company contribute?

100

Please circle your response to each question:

Not At All 1 2

3

4

5

d) The different businesses in our company need the same resources or materials.

1

2

3

4

5

6

7

e) The different businesses in our company

1

2

3

4

5

6

7

f)

1

2

3

4

5

6

7

1

2

3

95-99

70-94

4

5

6

7

(i.e. radio. television, cable, books, newspapers, periodicals, printing, movie production, music, retail, services)

c) The different businesses in our company

50-69

30-49

10-29

<10

A Great Deal 6 7

share the same customers.

share the same production processes.

The different businesses in our company share distribution networks.

g) Please estimate the percentage of corporate

%

sales achieved through foreign operations?

280

Please circle~~ur response to each guestion: h) Products and services of our business units are sold internally to other business units.

Not At All 1 2

3

4

5

A Great Deal

6

7

i)

Input required by our business units are purchased internally from other business units.

1

2

3

4

5

6

7

j)

Our company covers several steps of the industry value chain

1

2

3

4

5

6

7

k) Our company offers a broad range of com ple-

1

2

3

4

5

6

7

(Le. Your SBU operates a 1V channel. and is also part of a company that is active in TV content production or TV distribution via satellite or cable)

mentary products and services

(Le. Your company publishes a travel magazine. and offers in addition travel packages, an online travel agency, or travel insurances)

3/6 The purpose of this section is to understand the competitive strategy followed by your SBU. Please indicate the extent to which your SBU is concentrated on the following activities:

Please circle your response to each question: a) Introducing new services and products

Not At All 1 2

3

4

5

6

7

b)

Differentiating products from competitors

1

2

3

4

5

6

7

c)

Offering a broader range of services and products than competitors

1

2

3

4

5

6

7

d)

Utilizing market research to identify new services

1

2

3

4

5

6

7

e)

Providing a higher service or product quality than competitors

1

2

3

4

5

6

7

f)

Advertising products & developing a brand image

g) Achieving lower competitors

cost

A Great Deal

services

and

1

2

3

4

5

6

7

service

than

1

2

3

4

5

6

7

of

h)

Making services & procedures cost efficient

1

2

3

4

5

6

7

i)

Improving the time and cost required for coordination of various services

1

2

3

4

5

6

7

j)

Improving the utilization of equipment, capacity, and facilities

1

2

3

4

5

6

7

available

k)

Performing cost analysis and controlling

1

2

3

4

5

6

7

I)

Offering lower prices for our products

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

m) Achieving a broader line of products and services than competitors n) Achieving a higher customer types

breadth

of

0) Increasinq our qeoqraphic coverage

different

281

The purpose of this section is to understand the resources and capabilities of your SBU. Please indicate the extent to which the following statements describe the situation in your SBU:

4/6

Market Orientation a) We regularly measure customer service.

Strongly Disagree 1 2 3

b) Our product & service development is based on good market and customer information.

1

2

We have deep knowledge of our competitors.

1

d) We have a good sense of how our customers value our products and services.

Stronaly Aaree 7 6

4

5

3

4

5

6

7

2

3

4

5

6

7

1

2

3

4

5

6

7

e) We are more customer focused than our competitors.

1

2

3

4

5

6

7

f}

We compete primarily based on product or service differentiation.

1

2

3

4

5

6

7

g) We are strongly focused on our customer's interest.

1

2

3

4

5

6

7

h) Our products and services are the best in the business.

1

2

3

4

5

6

7

Customer service is extremely important in our business.

1

2

3

4

5

6

7

4

5

c)

i)

Organizational Learning a) Managers agree that our firm's ability to learn is the key to our competitive advantage.

Strongly Disagree 1 2 3

b) Our SBU sees learning as a key commodity to guarantee organizational survival.

1

2

3

4

5

6

7

c)

Employees view themselves as partners in charting the direction of our organization.

1

2

3

4

5

6

7

d)

Employees in our SBU are committed to the goals of this organization.

1

2

3

4

5

6

7

e) We rarely question our own bias about the way we interpret customer information.

1

2

3

4

5

6

7

Personnel in this SBU realize that the very way they perceive the marketplace must be continually questioned.

1

2

3

4

5

6

7

g) We have specific mechanisms for sharing lessons learned between SBUs.

1

2

3

4

5

6

7

h) We always analyze unsuccessful activities and communicate the lessons learned widely.

1

2

3

4

5

6

7

Innovativeness a) Our SBU frequently tries out new ideas.

Strongly_Disagree 2 1 3

4

5

b) Our SBU seeks out new ways to do things.

f)

Strongly Agree 7 6

Strongly Agree 7 6

1

2

3

4

5

6

7

Our SBU is creative in its methods of operation.

1

2

3

4

5

6

7

d) Our SBU is often the first to market with new products and services.

1

2

3

4

5

6

7

c)

282

and

1

2

3

4

5

6

7

Our new product and service introduction has increased over the last five years.

1

2

3

4

5

6

7

4

5

6

7

7

e) Our management actively encourages innovation. f)

seeks

Strongly Agree

Reputation a) Compared to our major competitors, the quality of management in our SBU is superior.

Strongly Disagree 2 3 1

b)

Our SBU is known for the quality of its products and services.

1

2

3

4

5

6

c)

Our SBU is known for its innovative products.

1

2

3

4

5

6

7

d)

Our company is perceived as excellent longterm investment by its shareholders.

1

2

3

4

5

6

7

e)

Compared to our major competitors, our SBU is in a sound financial position.

1

2

3

4

5

6

7

f)

Our SBU has the ability to attract and keep talented people.

1

2

3

4

5

6

7

g)

Our SBU is known for its well formulated and executed strategies.

1

2

3

4

5

6

7

h)

Our company takes on its community and environmental responsibility.

1

2

3

4

5

6

7

Corporate Culture Each question contains four descriptions of organizations. Please distribute 100 points among the four descriptions depending on how similar the description is to your company. For each question, please use all 100 points. You may divide the points in any way you wish; most common is a mix of answers. 1. a) b) c) d)

Kind of Organization (Please distribute 100 points): My organization is very personal. People share a lot of themselves. My organization is very dynamic and entrepreneurial. My organization is very formalized with lots of established procedures. My organization is very work-oriented. The focus is on getting the job done.

2. a) b) c) d)

Leadership (Please distribute 100 points): Leaders in my organization are considered Leaders in my organization are considered Leaders in my organization are considered Leaders in my organization are considered

3. a) b) c) d)

What holds your Organization together? (please distribute 100 points): Loyalty, tradition, and commitment hold my firm together. Commitment to innovation and development hold my firm together. Formal rules and poliCies hold my firm together. The emphasis on task and goal accomplishment holds my firm together.

4. a) b) c) d)

What is important? (Please distribute 100 points): My organization emphasizes human resources and employee morale. My organization emphasizes growth and acquiring new resources. My organization emphasizes permanence and stability. My organization emphasizes competitive actions and goal achievement.

as as as as

mentors. entrepreneurs or innovators. coordinators or organizers. producers or technicians.

283

5/6

The purpose of this section is to understand the major changes and developments in your business over the last years. The development is analyzed for the last two years (2001, 2002), the current year (2003) and following year (2004). Please indicate for each year how the particular factor changed compared to the previous year. Please use the following scales: ,j; (Significant Reduction), ~ (Slight Reduction), ~ (Unchanged), " (Slight Increase), l' (Significant Increase) For the future (2004) please forecast based on your expectations.

Please evaluate the development in comparison to the previous year based on the arrows described above: a) Development of the market share held by the four largest firms in your industry? b)

Development of the annual growth rate of the primary market targeted by your SBU?

c)

Development of the risk of new competitors entering the market?

d)

Development of your company's product diversification (breadth of overall product range)?

e)

Development of your company's international business (% of sales achieved abroad)?

f)

Development of your company's integration (along the value chain)?

vertical

g) Focus of your SBU on product development & differentiation, as well as advertising? h) Focus of your SBU on cost reduction and efficiency? i)

Concentration of your SBU on a few core products, customers, and markets?

j)

Focus of your competitors?

k)

Focus of your SBU on its reputation, service excellence, and attractiveness to both financial and labor markets?

I)

Strength of your corporate culture?

SBU

on

customers

organization's

and

particular

m) Importance your SBU assigns to learning, knowledge sharing, and communication? n) Development of your SBU's innovativeness?

0) Development of your SBU's profitability? p)

Development of your SBU's market share?

q)

Development of your SBU's sales growth?

284

2001

2002

2003

2004

6/6

The last section is limited to a few questions on SBU performance and on your personal background. The information is anonymous and will be kept strictly confidential.

A.

Strategic Business Unit Performance Very Hlgh

Compared to our three largest competitors ... a) the Profitability of this SBU is:

VeryLow 1 2

3

b) the Market Share of this SBU is:

1

2

3

4

c)

1

2

3

4

the Sales Growth of this SBU is:

4

6

7

5

6

7

5

6

7

5

The largest share of your SBU's revenues comes from (Please check): Broadcasting & Cable 0

Motion Pictures 0

Print & Publishing 0

Music 0

The headquarters of your SBU are located in (Please check): Germany 0

United Kingdom 0

United States 0

Other 0

B

Personal Information

a.

What is your current position?

b.

How long have you worked in the media industries?

_ _ _ _ _ _ _ _ _ _ _(years)

c.

How long have you worked for this company?

_ _ _ _ _ _ _ _ _ _ _(years)

d.

How long have you been deployed to your current SBU?

_ _ _ _ _ _ _ _ _ _ _(years)

Please circle the most appropriate: How involved are you in strategic activities in your SBU?

Not At All 1 2

Please circle the most a ro riate: How understandable was this questionnaire?

Not At All

1

285

2

A Great Deal

3

4

5

3

4

5

6

7

A Great Deal

6

7

Appendix B: Evolutionary Model - Regression Charts

287

Variable

Simple Regression

Multiple Regression

SEMI LISREL

Overall Significance

IISTO

T-value

IISTO

T-value

Industry Concentration

0.405

4.425

0.262

2.821

00

Market Growth

0.479

5.454

0.379

4.098

00

Entry Barriers

0.012

0.121

0.037

0.430

Product Diversification

0.421

4.638

0.355

3.506

International Diversification

0.134

1.347

0.102

1.096

Vertical Integration

0.261

2.704

0.127

1.259

Differentiation Strategy

0.316

3.327

0.306

3.218

Cost Leader Strategy

0.130

1.306

0.101

1.065

Innovativeness

0.414

4.543

0.239

2.417

Market Orientation

0.355

3.797

0.068

0.726

Organizational Learning

0.505

5.847

0.132

1.088

Reputation

0.490

5.621

0.292

3.065

Corporate Culture

0.407

4.453

0.146

1.481

00

0 00 0 0 0

00 00 00

0 0

00 00

Appendix B - (1/4) Regression Effects 2001

Variable

Simple Regression

Multiple Regression

SEMI LlSREL

Overall Significance

IISTO

T-value

IISTO

T-value

Industry Concentration

0.208

1.991

0.Q15

0.161

0 00

Market Growth

0.536

6.346

0.531

5.756

Entry Barriers

0.151

1.527

0.053

0.610

Product Diversification

0.384

3.914

0.350

3.659

International Diversification

0.030

0.298

0.008

0.084

Vertical Integration

0.143

1.444

0.075

0.783

Differentiation Strategy

0.216

2.212

0.216

2.199 0.002

Cost Leader Strategy

0.009

0.090

0.000

Innovativeness

0.384

3.906

0.135

1.388

Market Orientation

0.358

3.831

0.215

2.375

Organizational Learning

0.446

4.983

0.180

1.668

Reputation

0.387

4.199

0.093

0.893

Corporate Culture

0.423

4.668

0.191

1.799

Appendix B - (2/4) Regression Effects 2002

289

00

00 0 0 0

00 00 00

0 0

00 00

Variable

Simple Regression

Multiple Regression

SEMI LISREL

Overall Significance

IISTO

T-value

IISTO

T-value

Industry Concentration

0.208

1.991

0.015

0.161

0

Market Growth

0.536

6.346

0.531

5.756

00

Entry Barriers

0.151

1.527

0.053

0.610

Product Diversification

0.364

3.914

0.350

3.659

International Diversification

0.030

0.298

0.008

0.064

Vertical Integration

0.143

1.444

0.075

0.783

Differentiation Strategy

0.216

2.212

0.216

2.199

Cost Leader Strategy

0.009

0.090

0.000

0.002

Innovative ness

0.364

3.906

0.135

1.388

Market Orientation

0.358

3.831

0.215

2.375

Organizational Learning

0.446

4.983

0.180

1.668

Reputation

0.387

4.199

0.093

0.893

Corporate Culture

0.423

4.668

0.191

1.799

00

00 0 0 0

00 00 00

0 0

00 00

Appendix B - (3/4) Regression Effects 2003

Variable

Simple Regression

Multiple Regression

SEMI LlSREL

Overall Significance

IISTO

T-value

IIsTO

T-value

0.204

2.064

0.081

0.847

0

Market Growth

0.431

4.770

0.398

4.168

00

Entry Barriers

-0.147

-1.464

-0.132

-1.457

Product Diversification

0.098

0.987

0.108

0.967

International Diversification

0.067

0.674

0.069

0.686

Vertical Integration

0.025

0.253

-0.018

-0.160

Differentiation Strategy

0.334

3.548

0.337

3.607 -1.588

Industry Concentration

Cost Leader Strategy

-0.141

-1.429

-0.149

Innovativeness

0.319

3.363

0.107

1.033

Market Orientation

0.088

0.680

-0.082

-0.854

Organizational Learning

0.439

4.864

0.367

3.442

Reputation

0.105

1.060

-0.071

-0.738

Corporate Culture

0.323

3.408

0.224

2.283

Appendix B - (4/4) Regression Effects 2004

290

00 0 0 0

00 0 00

0 0

0 00