Issuu on Google+


Ferdinand Mahr Aligning Information Technology, Organization, and Strategy


GABLER RESEARCH Markt- und Unternehmensentwicklung / Markets and Organisations Edited by Professor Or. Ores. h. c. Arnold Picot Professor Or. Professor h.c. Or. h.c. Ralf Reichwald Professor Or. Egon Franck Professorin Or. Kathrin Mรถslein

Change of institutions, technoloqy and cornpetition drives the interplay of rnarkets and organisations. The scientific series 'Markets and Organisations' addresses a rnagnitude of related questions, presents theoretic and ernpirical findings and discusses related concepts and rnodels.


Ferdinand Mahr

Aligning Information Technology, Organization, and Strategy Effects on Firm Performance With forewords by Prof. Dr. Dres. h.c . Arnold Picot and Prof. Dr. Tobias Kretsch mer

•

GABLER

RESEARCH


Bibliographie information published by the Deutsche Nationa lbibl iothek The Deutsche Nationalbibliothek usts this cccucencn in the Deutsche Nationatbibüoqrahe : detailed bibliographie oata are evenebie in tne Internet at http://dnb.d-nb.de.

Ooctoral

tnesrs. Ludwig -Ma ximilians-Universität München , 2010

019

1s t Edition 2010 All rights reserved © Gabler Verlag I Springer Fachmed ien Wiesbaden GmbH 2010 Editorial Office: ute Wrasmann I Sabine Schöller Gabler Verlag is a brand 01 Springer Fachmedien . Springer Fachmedien ls pa rt 01 Springer Seience-Business Media . www.gabler.de No part 01 this pubjcanon may be reproduced, stored in a remevar system

or transmitted . in any form or by any means. eiectrcrsc. mecnanlcat. photocopying , recording, or otherwise. without me prior written perrrussion ot the copyright holder. Registered and/or industrial names, trace names, trade descriptions eto. crteo in this publicauco are part 01 the law lor trade -mark proteenon and may not be used Iree in any form or by any means even if tnrs is not ececmceav marked. Cover design: KünkelLopka Medienentwicklung, Heidelberg Printed on aeid-free paper Printed in Germany ISBN 978 -3-8349-2541-1


Foreword Prof. Dr. Dres. h.c, Arnold Picot Information technology has the potential to significantly enhance firm performance. However; firms differ heavily in their capability to leverage the full value of IT. These disparities may be a consequence of elements that strengthen IT's perfonnance impact and that are not present in all finns. The author of this book investigates combinations of IT, organizational structures, human resource management practices, and corporate strategies that are qualified to unlock IT's full potential f端r performance increases. The issue of complementarities between IT, crganization, and strategy is both up to date and of fundamental importance as this concept supports general managers in designing the organization and technology required to implement the strategy they have set. Mter presenting a thoroughly developed theoretical model, the author provides large-scale empirical analyses of two specific phenomena related to complementarities between IT, crganization, and strategy. Based on an extensive review of previous studies on factors that are complementary to IT, the author develops an overarching theoretical framework to guide empirical research on the role of complementarities in the relationship between IT and firm performance. The author highlights that IT has to be aligned with complements both inside the organization such as organizational structure and human resource management and with complements that determine the characteristics of the organization such as a firm's strategy, environment, and culture. For his empirical research, the author sets up two unique datasets. He matches three fully unrelated datasets on IT, firm performance, and organizationjstrategy, because avoiding statistical bias from collecting all information from a single source is allimportant to draw reliable conclusions on complementarities. The data on organization and strategy were collected by two extensive telephone surveys which were specifically designed and conducted for this study. The

author adds a thorough empirical analysis to the centralization versus

decentralization debate that takes place in the research on IT's crganizational effects. Since information is a source of power and IT enhances information transfer and processing, researchers have recurrently asked if IT tends to be related to a shift of decision authority upward or downward the organizational hierarchy. The author's findings suggest a synthesis of both perspectives. Firms that focus on efficiency and


VI

Foreword Prof. Pr. Dres. h.c. Amold Picot

rationalization unlock IT's positive perfonnance effects by combining it with centralized

decision rights that ensure the minimization of slack resources and redundancies. In contrast, finns with a foeus on innovation and differentiation benefit from combining IT

with decentralization which enables autonomous employees to experiment with new ideas.

The author also contributes to another long-standing debate: The question whether hybrid strategies which mix the contradicting efficiency and innovation orientations are

beneficial or detrimental to a firm's performance. The author argues that hybrid strategies may be as beneficial as pure strategies if suitable means to solve organizational tensions are in place. He shows that both a combination of certain organizational structures and human resource management practices as well as a specific composition of IT types are suitable means to mitigate the conflicts that arise from mixing efficiency and innovation orientation. Overall, this book is a remarkable contribution to the research on complementarities between IT, crganization, and strategy on the frontier of the management and economics literature.

Prof. Dr. Dres. h.c. Arnold Picot


Foreword Prof. Dr. Tobias Kretschmer The importance of infonnation technology (IT) for households as well as businesses has grown significantly in the last decades. IT has changed the way firms conduct their business, it has opened up entirely new markets, distribution channels and business

models. But even in "traditional" manufacturing firms, IT has had a tangible impact. Interestingly however, not all finns have benefitted equally from using IT, and earlier studies have found that firms that make IT work well with their organization to support their corporate strategy benefit the most. This is the starting point of Ferdinand Mahr's book, which sets out to shed further light on the interdependencies between IT, strategy and organizational structure, and succeeds brilliantly. Ferdinand Mahr makes both a significant theoretical and an important empirical contribution with his book, Thecretically, he makes use of the theory of infonnation processing to consider the impact of IT on firm success. Information processing theory views an organization as an entity that consists of elements or individuals that send, receive and process information both vertically and horizontally - that is. infonnation can be used to give and receive orders or for cocrdination between hierarchically equal elements. IT has the potential to support both vertical and horizontal infonnation flows, which gives an appealing theoretical proposition aboutthe benefits ofIT. However; while it is undisputed that IT helps transmit and process information, uncovering the real effects of IT and potential complementarities between IT and other firm characteristics poses considerable empirical challenges. Not only is it necessary to gather information at the level of the technology rather than the amount invested - pure investment statistics would not possess the required level of detail to study interactions between IT and specific organizational factors - but even more importantly a competent and relevant study also calls for information on "intangibles" like organizational structure and strategy. Ferdinand Mahr tackles and solves these problems by using a highly sophisticated method of data gathering through semi-structured telephone interviews. This method has the advantage of generating more detailed infonnation than paper questionnaires, but at the same time is sufficiently structured to allow robust statistical analysis with large samples. However; as with most applied research, the method is not helpful if it does not address questions practitioners actually ask, or if the results do not ring true to their potential users. Ferdinand Mahr's work does both - the vast number of IT consultants that align IT systems to strategy and organizational structure demonstrates that this is a problem


VIII

Foreword Prof Dr. Tobias Kretschmer

finns think about and for which they da not have simple, obvious answers. At the same time, Ferdinand Mahr's results also show that the role of IT as part of an organization is

by no means trivial, and that one-size-fits-all solutions can be ineffective at best, and value-destroying at worst: First, more IT is not always beneficial. Even general purpose technologies like personal computers cannot compensate a general misalignment between strategy and organizational structure. In other words, if the non-technological elements of a firm are not aligned, throwing money at technology cannot help overcome this dis advantage. F端r managers, this implies that to make productive use of IT, firms have to make sure their other structural elements are aligned. In a sense, IT can work as an amplijier for the performance of already well-aligned firms. Second, the challenge of finding a balance between the exploitation of existing markets or strengths and the pursuit of new opportunities is made much easier with the support of IT. In fact, firms that attempt this delicate balancing act without IT that helps them collect, process and interpret the vast amount of different Information typically perform worse than if they had focused on either new or existing markets only, while firms that make use of the opportunities presented by IT perform on par with focused firms. For managers in such finns this implies that novel, challenging strategies involving complex tradeoffs are only made possible through successful use of IT. Thus, IT acts as an enabler of such strategies. Ferdinand Mahr dissects and analyzes these two roles of IT in a very thorough and original way. The blend of solid, exhaustive theory and state-of-the-art empirics applied to unique data makes for a series of insightful studies on the complex effects of IT on performance. I am convinced that the questions, methods and insights contained in this book will inspire practitioners and researchers in their werk,

Prof. Dr. Tobias Kretschmer


Preface This dissertation is the result of nearly three years of hard werk, fun, and curiosity, Several persons and institutions have contributed to the completion of this thesis.

First and foremost, I would like to thank my supervisor Tobias Kretschmer f端r his exceptional, continuous. and motivating support. Also, I am

grateful f端r the

opportunities to implement a large-scale telephone survey and conduct my research in London f端r several months. I learned a substantial amount about conducting scientific

research, writing and publishing scientific papers, and academia as a whole. Further; Tobias was always happy to provide advice on good food, venues, and life in general. I am very much indebted to him. Further; I am grateful to Dietmar Harhoff my second advisor, for his excellent support, which began during my graduate studies and continued during my post-graduate 'Master ofBusiness Research' and doctoral studies. Deutsche Telekom Stiftung and, in particular, its chairman Klaus Kinkel generously supported the telephone survey conducted for this dissertation. F旦rderverein Kurt Fordan, German Academic Exchange Service, and LMU Management Alumni have supported me financially at various stages in the process. This help has been gratefully received. The Centre for Economic Performance at the London School of Economics and Political Science offered me to stay for four and a half months and to interact with their faculty. I would like to thank John Van Reenen, Mirko Draca, and Raffaela Sadun for their hospitality, their comments on my research, and their support with the construction of the datasets for this dissertation. Numerous comments for my research were received at the various conferences and seminars I have attended, including the European Academy of Management Conference 2008 in Ljubljana, the 2008 Academy of Management Annual Meeting in Anaheim, the Fifth bi-annual Conference on The Economics ofthe Software and Internet Industries in Toulouse, the DRUm Society Summer Conference 2009 in Copenhagen, the 7th ZEW Conference on the Economics of Information and Communication Technologies in Mannheim, the 2009 Academy ofManagement Annual Meeting in Chicago, and seminars at the University of Southern Denmark in Odense and LMU Munich.


x

Preface

I also benefitted from comments from and conversations with a number of other people, including Chiara Criscuolo. Bo Eriksen, Luis Garicano, Georg von Graevenitz, Karin Hoisl, Fiona Scctt-Morton, Hans Eibe Sarensen, and Nils Stieglitz. They all helped me to clarify my thoughts, and I would like to thank all of them. I am also grateful to Richard Burton, Harold Doty, and Lorin Hitt für making available the details of their surveys. I was lucky to work in an environment of engaged and cooperative colleagues at the Institute für Communication Economics and have made some inspiring friendships there. I would like to thank Esther Almasdi, Melisande Cardona, ]örg Claussen, Christoph Dehne, Hüseyin Doluca, Thorsten Grohsjean, Sibyl Herrmann, Stefan [elinek, Mariana Rösner, Pavlos Symeou, and Leon Zucchini. We have exchanged many ideas about research and, more importantly, life in general. I would also like to thank Sebastian Jacobs and Florian Seliger for their excellent research assistance. In addition, I am grateful to Fiona Bannert, Karol Englert, Luisa Feigenspan. Sebastian [acobs. Vanessa [elinek, Stephanie Leopold, Natalia Ojewksa, Andrea Peschl, Bernhard Rösch, Veronika Ruderer, Werner Skalla, Dorothee Stadler; Anja Staudt, Doris Thalmeier; and Veronika Triphan, who conducted the telephone Interviews for this dissertation. Special thanks are reserved for Marlies. She continuously encouraged me to pursue my professional goals with full vigcr, showed forbearance in my times of mental absence, and reminded me that life is not all about scientific papers and telephone interviews. She has been a partner, friend, and companion for many years, and I hope we "Will share many more experiences. Finally, I would like to express my gratitude to my family, especially to my parents, Herta and Horst, to Stefanie and Eduard, and, above all. to my dear sister Isabel. They support me in every way and have always believed in my abilities. Their advice and their confidence were a great help throughout all the steps I have taken over the years, and I "Will always remain grateful for this.

Ferdinand Mahr


Contents Foreword Prof. Dr. Dres. h,c. Am o ld Picot _ _ _ _ _ _

M

Foreword Prof. Dr. Tobi as Kretschmer...._ _ _ _ _ _ _

•••

M

••••

M

•••

. . . ._

M

M •••M ••••M •••M ••••w

••• _

• •• _

• • •• _

••• _

••••• _

. . ._ ••• _ •••• _ • • • _ •••• •

• •••

V

_.VI I

Preface lX Contents ...._..._..._...._..._..._...._..._...._...__..._...._..._..._...._..._...._..._ _...._..._ _ _...._..._...._..._..._....x:I . M• • • • M • • • M • • • M • • • • _

••• _

•••• _

• •• _

• • • _

• • • • _

• • • •• _

• •• _

• • • • _

• •• _

••• _

• ••• _

••• _

••• _

•••• _

_

•••• _

••• __

_

_

•••• _

••• _

•••• _

••• _

• • • _

• • • •

List 0 f tables .•...•....•...•....•...•...•....•...•.....•....•...•...•....•...•....•...•...•....•...•.....•...•....•...•....•...•...•..XV List of figures _...

M • • • • _

• • • _

• •• _

• • • • _

Chapter 1: lntroduction.,

• • • _

_

• •• _

_

• • • • _

• •• _

• •• _

• ••• _

• • • _

• • • • _

_

_

• • • • M ••• _

••• • • _

••• _

• • •• _

••• _

• ••• _

• • • _

• • • •

XVl]

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 1

1.1

Motivatio ll..... _... _

_

1

1.2

A sUln mary ofthe th re e stu dies ._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

4

1.3

The data sets employe d in t he two empiri cal st udies _ _ _ _ _ _ _ _ _ 8

_

_ _ _ _

_

_

_ _ _

_ _

_ _ _

_ _

_

_ _ _

Chap ter 2: Infonnation t ech nology and rum perfo rmance: An int egrative mod el ofthe role o f complementa rit ies .._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.. 19 2.1

Introd uetion _ _ _ _ _ _ _ _ _

2.2

An overvi ew ofthe empirical research on IT compl em en ts ., _ _ _ _ _ _.. 2 1

2.2.1

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _.. 19

St ud ies on internal or extema llT complem ents _ _ _ _ _ _ _ _ _.. 22

2.2.1.1

Resource-centered pers pective _ _ _ _ _ _ _ _ _ _ _ _ _ _.. 22

2.2.1.2

Contingency pers pective ..._...._..._ _ _ _ _ _ _ _ _ _ _ _ _ _.. 23

2.2.2

St udle s on intem al and externa l IT complem en ts,

2.2.3

The need for an integrative model of IT complemen ts ._ _ _ _ _ _ _.. 24

2.3

Toward an integrative model of IT complements

_ _

_ _ _ _ _ _ _ _.. 23 _ _ _ _ _ _ _.. 26

2.3.1

Fit and comp lement arity _ _ _ _ _ _

2.3.2

Organi zat ional information pro cessing and IT

29

2.3.3

An integrative model ofIT complements

30

2.4 Horizon tal fit between IT and IT complements 2.4.1 The d egree of centralizat ion _ _ _ _ _ _ _

_ _.. 27

33 _ _ _ _ _ _.. 34

2.4.1.1

Centralizat ion..._...._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _. 34

2.4.1.2

Decentralization..._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.35

2.4.1.3 A rie he r defi nit ion of the d egree of eentra lizat ion _ _ _ _ _ _ _.. 36 2.4.2 Sta nda rd izat io n ._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.38 2.4.3

IIRM practices ...._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.40

2.4.4

Infonnat io n technology _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.42

2.4.5

Two ideal configurations of organization design _ _ _ _ _ _ _ _ _.. 44

2.5 Vertical fit between contingencies and organization design ; _ _ _ _ _ _.. 46 2.5.1 Environment ._...._..._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.46 2.5.2

Corporate strategy .._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.48


XI!

Contents

2.5.3

Organi zatio nal culture...._..._...._..._...._..._ _ _ _

M • • •• _

2.5.4 Two idea l ways to ac hieve ve rt ical fit.._ _ _ _ _... 2.6 Conclusio n a nd implicati o ns.._..._..._...._..._...._ _

• •• M

M •••• _

M ••• _

••• _

•••• _

••• _

_

• •• _

••• _

•••• _

••• _

••••• _

_

• •• _

••• • _

• •• _

• •• _

• ••• _

_

••• _

••• _

•••• _

• •• _

• ••• •

.

50

52 _.53 • •

Cha pter 3: Enhancing t he performance effects ofinform ation techno logy t hru ugh de/ centralization: The ro te of corporate explora tion a nd exploitation., _ _ _ _ _.. 57 3.1 Introducticn _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.. 57 3.2

Literature review and hypothesis development..~

~

~

~

~

_ _ _ _ _ _.. 60

H

H

H

H

3.2.1

lnfonnation technology an d finn perform ance;

3.2.2

The infonnati on processing view of organizations and inform ation technology._..._...._..._...._..._..._...._..._....._...._..._...

3.2.3

Inform ation technology, decent ralizati on, and centralization. _ _ _ _. 62

3.2.4

The ro le of organizational lea m ing.,

H •••• H •• • H • ••• H • •• H ••• _

_ _

3.2.5 Hypoth eses 3.3 Empirical approach _ _ _ _ _ _ _ _ _ _

3.3.1

EmpiricaLmodels

3.3.2

Data

3.3.3

Va riab les

3.3.3.1

H

H

~

_ _ _ _ _ _ _ _ _

~

H

H •• • H

_

•••• _

H

••• _

_

_

_

•••• _

_

_ ..

•••• _

_

•••• _

_

_

.

••

60 60

63

~

~

~

~

65 _ _ _ _ _ _. 66

~

~

~

~

_ _ _ _ _ _. 66 68

Prod uctio n function inputs an d ou tputs

~ ~

~

3.3.3.2 Orga nizat iona l leam ing and organization design 3.3.3.3 Control varia bles ._ _ _ _ _ _ _ _ 3.4 Res u1ts H

. H ... H ... _

_

... _

.... _

... _

... _

.... _

_

_

. _

_

_

_

_

_

3.4.1

Dtscuss ion of t he re sults ._ _ _ _ _ _ _

3.4.2

Robu stness of t he res ults

H

H

H

H

H

H

H

H

H

H

H

~

70 _ _ _ _ _ _.. 70

~

_ _ _ _ _ _.. 71

H

_

H

_

H

H

_

_

H

_

_

_

_

_

_.75

_

_

_

_

_

..

_

.

_ _

77

77 80

3.5 Conclusions 85 Chapter 4 : Ilybrid st rategy and firm performance : The mod er atin g role of individual and te chno logical am bidexter ity 4.1

Int roduction

4.2

Theory and hypotheses

4.2.1

89 H

H

H

H

H

H

89 H

91

H

Jlybrid st rategy, organiz atio nal architectu re, and firm perfo rmance in th ecry

4.2.2

91 Hybrid st rategy, organizational architectu re, and firm pe rformance in em ptri ca I stud ies _..._...._..._..._...._..._ _ _ _ ~ ~ ~ ~ ~ _ _ _ _ _ _. 93

_ _ _ _ _ _.. 93

4.2.2.1

Summa ry of e mpirica Lstu dies _ _ _

4.2.2.2

Definitions of a hybri d strate gy's perforrna nce impact _ _ _ _ _.. 95

~

~

~

~

~

4.2.2.3 4.2.3

Organizationa l ambi dexterity as a mod era to r of the hybrid strategyperfonnance rel ationship., _ _ ~ ~ ~ ~ 95 Organizational am bidext erity., _ _ _ 98 H

H

H

H

H

H

H ••• _

••• _

• ••• _

• • • _

_

• •• _

• ••• _

• • • _

• • • • _

_

• •• _

• •

_

••


XIII

Contents 4.2.3.1

Individual ambidexterity

4.2.3.2

Technological ambidexterity

4.2.4 4.3

Hypotheses

Method

4.3.1

99 101 104 106

Models

106

4.3.1.1

Moderation effects model

106

4.3.1.2

Effects-codingmodel

107

4.3.1.3

Dummy-coding models

108

4.3.2

Sample

110

4.3.3

Measures

113

4.4

4.3.3.1

Firm performance

4.3.3.2

Corporate strategy

114

4.3.3.3

Individual ambidexterity

115

4.3.3.4

Technological ambidexterity

118

4.3.3.5

Group indicatcr variables

120

4.3.3.6

Control variables

Results

113

121 124

4.4.1

Moderation effects models

124

4.4.2

Effects- and dummy-coding models

127

4.4.3

Robustness test: Porters definition of the performance impact

4.4.4 4.5

of a strategy

131

Pure strategies

135

Discussion and conclusions

136

4.5.1

Implications for research

137

4.5.2

Implications for managers

138

4.5.3

Limitations and future research

138

Chapter 5: Conclusion

141

5.1

Summary and contributions

141

5.2

Suggestions for future research

146

References .,

149

Appendices

171

Subject index

183


List of tahles Table 1.1 The s pring s urvey a nd th e s umme rsurveY..._....

M •••M ••••M •••M •••• _

Tab le 2.1 Exa mple qu estio ns on th e de gree o f ce ntra lization Table 3.1 Descriptive sta tis tics .._...._..._...._..._....._..._...._..._ _ _ _

M •••M •••• _

Table 3.2 Main esti mations and first ro bustness tests , Table 3.3 Robustness tests;

_ _ _ _ _ _

M ••• _

_

_

••• _

_

•••• _

••• • _

•••• _

• •• _

••• _

••• _

• ••• _

•••• _

••••• _

• •• _

••• _

• • •• _

• •••

• ••• _

• • •

• • • _

• • •

11 38 77

_ _ _ _...._ _..._...._..._..._...._ 78

_ _ _ _ _

84 Table 4.11CT typ es ...._...._..._ _ _ _ _ _._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 119 M •••M •••• _

••• _

_

• ••• _

• • • •• _

• •• _

• •• _

• ••• _

Tabl e 4.2 Cerre lat ion between leT types an d leT use.._ _ _ _ _ _ _ _ _ _ _ _. 120 Table 4.3 Group s in th e individu al and technological ambidexterity samples

_ _ _ 121

Tahle 4.4 Descript ive statistics for t he indivi dual ambidexterity sample..._...._ _ _ _ 123 Tabl e 4.5 Descript ive statistics for the t echnological ambidexterity sa mple.._ _ _ _ 123 Table 4.6 Test of hypothesis 4.1 ._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 125

Table 4.7 Test of hypotheses 4.1 to 4.4 _ _ _ _ _ Table 4.8 Summary of t he res ults .._...._ _ _ _ _ _ _ _

M •••M ••••M •••M

Tab le 4.9 Test of hypoth esis 4.1 with Porters perfo rmance definition

_

_ _ _ _ _ 128 130

• ••• _

M . . .. _

••• _

••• _

•••• _

•••• _

• •••• _

••• _

• • • _

•••• _

132

Tab le 4.10 Test of hyp oth eses 4.1 to 4.4 with Port e r 's pe rfonna nce definitio n., _ _ 133 Tab le 4.11 Summa ry ofthe re sults with Porter s perfo rmance definition.._..._..._ _ _ 135


List of figures Figura 1.1 Survey q uest ion o n t raining int en sity a nd vari ety in t he s pring survey _ _ 13 Figu re 1.2 Survey qu estion o n e mployee ap prai sa l in the summe rs urvey ....._..._...._ _ 14 Figu re 1.3 Correlatio n between in terviewer and supe rviso r scorings in th e s pring sun 'ey ..._..._..._...._....._...._..._..._ _ _ _ _ _ _ _ _ _ _ _ _ 16 Figure 2.1 An integrative model oflT complements

_ _ _ _ _ _ _ _ _ _ _ _ 31

Figure 3.1 Centra lized coordi nati on [left] and decentralized coordination (right)

_ 61

Figura 3.2 Infonnation technology, de j centralizati on, and firm perfonnance under corporate exploration (to p) and corp orat e exploita tion (bottom) ....._..._..._....82 Figure 4.1 Ambid exterity as a moder ato r ofthe hybrid strategy-performance relationship _...._..._...._..._..._...._..._...._..._.... 9 6 Figure 4.2 Hybrid st rat egy as a mediato r ofth e a mbidexterity-perf o rmance reLationship ....._...._..._...._..._..._...._..._..._...._..._...._.... 97 Figure 4.3 Five grou ps of finns s plit by corporate st rategy a nd o rgani zationa l architecture., ..._....._...._..._...._..._..._...._..._..._...._..._....._...._..._ 104 Figure 4.4 The mod erati ng ro le of individ ua l a nd techno logical a rnbidexterity in t he hyb rid st rategy-perfonna nce re lationship _...._..._...._..._..._...._..._...._..._. 126


Chapter 1 Introduction 1.1

Motivation

It is probably a truism that information technology (IT) is increasingly important für all types of business activities. In fact; while IT expenditures accounted für only about eight percent of the German business sector's real expenditures in 1970, the share increased to nearly 42 percent in 2002 (Deutsche Bundesbank, 2004). In 200S, 84 percent of firms in Germany employed IT in their operations, and about 60 percent of employees in Germany regularly used a computer at work (Bauer and Tenz, 2009).1 More importantly, since the 1990s, diverse empirical studies have shown that firms do not only heavily invest in IT; in addition, IT has a substantial, positive effect on firm performance (Anderson et al., 2006; Bharadwaj et al., 1999; Dewan and Kraemer, 2000; Mukhopadhyay et al., 1997). For example, the first firm-level studies in the 1990s consistently estimated that one US dollar of IT capital is associated with more than a 0.6 US dollar increase in revenue each year (Brynjolfsson and Hitt, 1998a; 2000). In contrast, by the late 1980s, the conventional wisdom was that IT was not contributing significantly to productivity. The amount of IT per US employee grew dramatically during the 1970s and 1980s, while the productivity growth rate in the US economy slowed down during the same period (Brynjolfsson and Hitt, 1998a; 2000). Nobel Laureate Robert Solow famously summarized this so-called 'computer productivity paradox' in his quip that "we see the computer age everywhere except in the productivity statistics" (Solow, 1987). The initial problems in detecting any firm performance contribution of IT may be attributed to the high aggregation level of the data that were then available on the industry sector level or the economy level (Brynjolfsson and Hitt, 1998a; Brynjolfsson and Yang, 1996). However; although recent studies show that, on average, IT enhances firm performance, they also indicate that firms vary significantly in their ability to generate returns from IT investments. Some firms with high IT investments are highly productive, while others The 16 percent ofGerman firms t hat do not use IT in their operations for t he most pa rt have less than 20 employees. 97 percent of the firms with at least 10 employees employed IT in their operations in 2008 (Bauer and Tenz, 2009).

1

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6_1, © Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


2

1 Introduction

with similar investments perform poorly (Brynjolfsson and Hitt, 1995; Lichtenberg.

1995). Further; there is evidence that the effects of IT on firm performance are substantially larger when measured over longer time periods. The long-terrn benefits of IT (that Is, five-year to seven-year differences) have been estimated to be two to eight times as much as its short-tenn benefits (that Is, one-year differences) (Brynjolfsson and Hitt, 2003). Finally, there is evidence that IT capital is valued higher by the stock market than other types of capital. While one US dollar of most capital types is valued by the stock market at about one US dcllar; one US dollar of IT capital is valued at about ten US dollars (Brynjolfsson et al., 2002). In summary, there is evidence of a large, unobserved factor in the relationship between IT and firm performance. These findings may indicate that while there exist complementary factors that unlock the value of IT, these factors are not present in all firms, are time consuming to develop, and are often hidden in the balance sheets but emerge in the stock market valuations of finns (Brynjolfsson and Hitt, 1998a; 2000). Two variables are complementary to each other when "doing (more of) one ofthem increases the returns to doing (more of) the other" (Roberts, 2007: 34). In mathematical terms. the marginal benefit of one variable increases in the level of any complementary variable, and vice versa (Roberts, 2007; Stieglitz and Heine, 2007). It is argued that IT is a 'general purpose technology' such as electricity or the steam engine (David, 1990; Rosenberg and Trajtenberg, 2001). A common feature ofthese inventions is that they contribute significantly to overall productivity, but they do not favor all firms and individuals equally (Hempell, 2006). Instead, general purpose technologies "play the role of 'enabling technologies', opening up new opportunities rather than offering complete, final solutions" (Bresnahan and Trajtenberg, 1995: 84). For example, David (1990) describes how electrification at the beginning of the 20th century was initially only used for cost reductions through the replacement of old systems with electrical motors. The full advantages of electricity were not exploited until finns recognized that electric motors allowed them to radically redesign their factories and production processes. that Is, to implement organizational changes. Similarly, higher IT productivity contributions are found in finns that implement organizational changes with regard to organizational structures and human resources. Particularly, structures that allocate the organizational decision authority at a low hierarchical level and human resource management (HRM) practices that support such a power distribution have been identified as complementary to IT (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). Other studies show that IT's


1 1 Motivation

3

potentials for performance enhancements are better exploited under specific corporate strategies. In particular, corporate strategies that focus on the development of new products and the expansion into new markets appear complementary to IT (Li and Ye, 1999; Sabherwal and Chan, 2001). Thus, the effects ofIT on firm performance seem to be substantially dependent on the alignment of IT with the organization and strategy of finns. However, although the existing studies on complementarities among IT, organization, and strategy are a first step to untangle the firm characteristics that are needed to fully exploit IT's potential for performance enhancements, they remain sparse and leave room for improvement. My thesis analyzes several phenomena relating to the complementarities among IT, organization, and strategy. The analysis in this thesis proceeds as follows. In the remainder of Chapter 1, I summarize Chapters 2 to 4 and describe the construction of the datasets used in Chapters 3 and 4. In particular; I describe the telephone survey that was conducted specifically for this thesis that resulted in about 1,500 telephone interviews. In Chapter 2, I review the existing literature on IT complements. I point out issues, potentials for improvement, and develop an integrative model of IT complements that may serve as a theoretical framework for future empirical research on IT complements. In Chapter 3, I analyze the complementarities between IT and the degree of the centralization of organizational decision authority. In particular; Lexamine whether IT and centralization or IT and decentralization are complementary, depending on a firm's corporate strategy. In Chapter 4, Lexamine the complementarities among IT, organizational structures and HRM practices, and hybrid corporate strategies. Hybrid strategies mix distinct strategie activities such as cost leadership and differentiation, result in organizational tensions, and have ambiguous effects on firm performance. I analyze if specific combinations of IT types. organizational structures, and HRM practices are a means to solve the organizational tensions and thus enable a positive performance effect of hybrid strategies. Chapter 5 concludes and highlights areas for future research.


4

1.2

1 Introduction

A summary ofthe three studies

The major contents and findings of Chapters 2 to 4 are as follows. Chapter 2 provides a structured review of the existing empirical literature on IT complements and its drawbacks. For example, two kinds of IT complements are predominantly examined, namely, factors inside a firm's organization design and factors outside the organization design. However, these factors are largely analyzed in isolation, leading to ambiguous results. For example, IT and decentralization are found to be complementary, although prior theoretical and case-based studies argue that combining IT with centralization may also be effective. An explanation is that contingency factors, such as a firm's strategy, that may affect the complementarities between IT and de/centralization are disregarded. However, aligning IT with other elements of the organization design is not very fruitful if this combination is not suitable to support the

firm's corporate strategy. Further theoretical drawbacks of the existing empirical studies on IT complements include a lack of a common theoretical background and a detailed analysis of the mechanisms by which IT and IT complements combine. With regard to methodology, prior studies often employ misleading operationalizations of IT and IT complements. To inspire more sophisticated studies on IT complements, I develop an integrative model of IT complements that addresses the drawbacks of prior studies. In particular; I draw from contingency thecry to argue that IT must be aligned simultaneously with socalled 'horizontal IT complements' inside the organization design as well as with 'vertical IT complements' that determine the characteristics of an organization design. Further; I draw from infonnation processing theory to examine the specific mechanisms that explain why IT is complementary with several horizontal and vertical elements. Based on this integrative framework and existing empirical and theoretical studies on IT complements, I develop testable propositions regarding how to

exploit the

complementarities

vertical

between

IT,

horizontal

IT complements,

and

IT

complements. In particular, I suggest that IT types supporting the information exchange up and down a

firm's hierarchy should be horizontally aligned with the centralization of decision rights. high standardization, and HRM practices that focus on skills and knowledge needed by lower-level employees to implement rules, procedures, and instructions that are imposed on them by central decision makers. This organization design should be vertically aligned with an undemanding environment, a corporate strategy that focuses


1.2 A summarv of the three studies

5

on existing product market domains, and an organizational culture that becomes manifest in a high degree of fonnality. IT types that support Information exchange between a firm's subunits should be horizontally aligned with the decentralization of decision rights, low standardization, and HRM practices that focus on diverse skills and knowledge that are needed by lower-level employees for direct communication with each other and autonomous coordination. This organization design should be vertically aligned with adernanding environment, a corporate strategy that focuses on new product market domains, and an organizational culture that becomes manifest through a low degree of fonnality. In Chapter 3, I empirically analyze one part of the integrative model of IT complements developed in Chapter 2. Particularly, I extend prior empirical studies that find evidence for complementarities between IT and the decentralization of organizational decision authority. Drawing from information processing theory and organizational leaming thecry, I argue that IT and the centralization of organizational decision authority may also be complementary, depending on a firm's corporate leaming type. I argue that IT enhances horizontal infonnation processing, which is the primary mechanism to coordinate a firm's subtasks in decentralized organization designs, but IT may also enhance vertical information processing, which is the primary cocrdination mechanism in centralized organization designs. This implies the existence of a third group of factors that

determines

the

direction

of

the

complementarities

between

IT

and

de/centralization. To address this issue, I refer to organizationallearning theory, which argues that firms must emphasize one of two learning types, namely, exploration or exploitation. Exploration refers to the permanent search for new products and markets, while exploitation indicates the continuous improvement of existing product market domains. Many scholars argue that an exploratory learning type requires enough employee autonomy for constant experimentation and adaptation, making a decentralized organization design most appropriate. In contrast, an exploitative learning type succeeds by minimizing variation and maximizing efficiency and control, thus making a centralized organization design the most beneficial solution. In summary, IT may be complementary with decentralization or centralization, while a firm's corporate learning type determines the benefits of decentralization and centralization. Hence, I argue that it is the corporate learning type that defines the complementarities between IT and de/centralization in a particular firm.


1 Introduction

6

Specifically, I hypothesize that under corporate exploration, IT and decentralization are

complementary, while under corporate exploitation, IT and centralization are complementary. I employ a novel multi-source panel dataset on the computer capital, organization design, and corporate learning type of almost 260 German manufacturing

finns to test the hypotheses. To test f端r complementarities under different corporate learning types, I estimate Cobb-Douglas production functions with an interaction term between IT and centralization f端r firms with an exploratory leaming type as well as f端r

finns with an exploitative learning type. The results support the hypotheses and a variety of robustness checks demonstrates that the qualitative nature of these results is not sensitive to alternative specifications and measures. Chapter 3 focuses on purely exploratory and purely exploitative corporate strategies. In contrast, Chapter 4 examines hybrid strategies that mix exploratory and exploitative strategie activities and, more specifically, the relationship between hybrid strategies and firm perfonnance. It is argued that hybrid strategies produce severe organizational tensions that arise from the simultaneous pursuit of two fundamentally different strategie activities, leading to a negative effect on firm performance. Others argue that hybrid strategies may positively affect firm performance provided that suitable organization arrangements to solve the organizational tensions (that Is, organizational ambidexterity] are in place. In particular, the literature on ambidextrous organizations suggests several suitable organizational arrangements that allow for the pursuit of exploration and exploitation with equal dexterity. Existing empirical studies on the hybrid strategy-performance relationship are inconclusive because their findings range from a positive relationship to no relationship to a negative relationship. I argue that these results have to be interpreted with caution because they do not take into account the organizational arrangements in a way that is consistent with theory, if they take them into account at all. Methodologically, the studies are inconsistent with regard to comparison groups. The average performance of finns with a hybrid strategy is compared to either the sample average performance, the average performance of finns with no particular strategic focus, or the average performance of firms with a purely exploratory or exploitative strategy. I argue that the effect of a hybrid strategy on firm perfonnance is moderated by organizational ambidexterity so that the effect is positive in the presence of organizational

ambidexterity

and

negative

in

the

absence

of

organizational

ambidexterity. Thus, organizational ambidexterity must be considered if analyzing a hybrid strategy's perfonnance effect. I advance several hypotheses on the moderating


1.2 A summarv of the three studies

7

effect of organizational ambidexterity in the hybrid strategy-performance relationship, measuring the performance impact relative to three different comparison groups, that Is, the entire sample, no-emphasis strategies, and pure strategies. Further; I employ two different definitions of a corporate strategy's performance effect. Further; I analyze two different types of organizational ambidexterity, that Is, individual ambidexterity and technological ambidexterity. Individual ambidexterity concerns the potential of specific organizational structures and HRM practices to enable an individual employee to pursue both explcratory and exploitative activities. Technological ambidexterity concems the largely unexplored potential of specific IT types to support the simultaneous pursuit of exploration and exploitation. I employ a novel multi-source dataset on 784 German and Polish manufacturing firms to test the hypotheses. I used moderation effects models to test the central hypothesis that the hybrid strategy-performance relationship is positively moderated by organizational ambidexterity. This hypothesis is supported. I use dummy-coding and effects-coding models to assess the performance differences between firms with a hybrid strategy and several comparison groups. I find evidence that as compared to the sample average, a hybrid strategy positively impacts firm performance in the presence of organizational ambidexterity and negatively impacts firm performance in the absence of organizational ambidexterity. Similarly, I find that a hybrid strategy outperforms no-emphasis strategies in the presence of organizational ambidexterity and matches no-emphasis strategies in the absence of organizational ambidexterity. Also, I find that a hybrid strategy matches pure strategies in the presence of organizational ambidexterity and underperforms pure strategies in the absence of organizational ambidexterity. Chapters 2 to 4 can be read as independent studies on various aspects of the performance effects of aligning IT, organization, and strategy. The autonomy of the chapters is reflected by the fact that each of them contains a review of the literature concerned with the correspondingly relevant topics. I employ different empirical approaches to analyze the effects of aligning IT, organization, and strategy on firm perfonnance. I use an approach more common in the organizational economics literature in Chapter 3 and an approach more common in the strategy and organization theory literatures in Chapter 4. In particular; the production functions analyzed in Chapter 3 relate a firm's output on the left-hand side of the regression models to several input factors on the right-hand side of the regression models. In contrast, in Chapter 4, I use a measure of profitability (that Is, return on total


8

1 Introduction

assets) as the dependent variable. Second, Lexamine IT at different levels of analysis. In Chapter 3, I employ a rather broad measure of IT, that Is, I use the number of personal computers (P端s] and annual average Pf prices to compute a measure of IT capital. In Chapter 4, I consider specific IT types. F端r example, Lexamine if the analyzed firms

employ enterprise resource planning (ERP) applications. workflow applications, groupware applications. and corporate intranets.

1.3

The datasets employed in the two empirical studies

The two empirical analyses in Chapters 3 and 4 are based on two novel multi-source datasets on the IT, organization design, corporate strategy, and perfonnance of German and Polish manufacturing firms. While the data on IT and firm performance stern from two independent third-party sources, data on organization design and corporate strategy are not readily available and were thus gathered specifically for this thesis by means of a large-scale telephone survey that resulted in about 1,500 telephone interviews. Information on IT employed by firms from 2000 to 2008 is obtained from Harte-Hanks' CI Technology Database (CITDB). Harte-Hanks is a market intelligence company that conducts annual telephone surveys to gather information on specific hardware and software types that are used by establishments of more than 10,000 German and 1,800 Polish firms. For example, the CITDB would indicate if Exchange Server by Microsoft versus Lotus Domino by IBM is present in an establishment. Harte-Hanks groups these products into broader dasses of hardware and software products. Exchange Server and Lotus Domino, which are both collaboration software products, would be assigned to the groupware dass. Data are provided on the establishment-level; that Is, Harte-Hanks interviews at least one establishment of each firm in the CITDE. The survey questions concern the IT types that are used by the establishment rather than the IT types that are used by the entire firm. Thus, to make the IT data comparable to the other finn-level data, I aggregate or extrapolate to the firm-level. Harte-Hanks states that the firms in the CITDB are largely representative for the entire populations of German and Polish firms with morethan 100 employees.


1.3 The datasets employed in the two empirical studies

9

Where possible, firm performance information for the years 2000 to 2008 are derived from Bureau Van Dijk's ORBIS database, which provides balance sheets and profit and loss statements from annual company reports. ORBIS contains all Information needed to analyze production functions. The Information on sales serve as a measure of output, and the Information on material costs, tangible fixed assets, and the number of employees serve as measures of the input factors material, capital, and labor, respectively. Further, ORBIS contains all Information needed to compute profitability measures such as return on total assets, which is defined as profit or loss before taxation divided by total assets. Firms that maintain at least one manufacturing establishment and that are covered by both the CITDB and ORBIS formed the pool of potential respondents to two telephone surveys on organizational structure, HRM practices, corporate strategy, and other firm characteristics that were conducted in March 2008 (that Is, the 'spring survey') and from August to September 2008 (that Is, the 'summer survey'). The spring survey differed partly from the summer survey with regard to the research design and survey Instruments. The data from the spring survey are used in Chapter 3, while the data from the summer survey are used in Chapter 4. Due to limited database access at the time of the sampling frame construction, the sampling frame for the spring survey consisted of those German firms for which at least 2004 IT data and 2004 or 2005 firm perfonnance data were available, resulting in a pool of 600 potential respondents. The pool of potential respondents for the summer survey consisted of all 2,369 German and Polish firms that are covered by the CITDB and ORBIS in at least one year from 2000 to 2008 and that were not already interviewed in the spring survey. During the 16 working days of the spring survey, six student research analysts conducted 259 telephone Interviews with production managers or similar managers, representing an overall response rate of 43.2 percent and a response rate of 46.9 percent for the 552 firms that were contacted. During the eight weeks of the summer survey, nine student research analysts conducted 784 telephone Interviews with production managers or similar managers, representing an overall response rate of 3 3.1 percent and a response rate of 43.8 percent for the 1,788 contacted finns. These 784 interviewed finns formed the pool of potential respondents for a second telephone survey on IT use in firms. IT managers or similar managers from 454 of the 784 firms were interviewed, representing a response rate of 57.9 percent. The 48 finns not contacted in the spring survey and the 681 firms not contacted in the summer survey


10

1 Introduction

were not approached because (1) they did not exist anymore, (2) they had no manufacturing establishments in Germany or Poland. (3) because no telephone number could be identified, or (4) the interviewers were not able to get in touch with the firms before the end of the project. In the spring survey, respondents and non-respondents did not significantly differ in terms of firm employees, value added, total assets, or the number of P端s. In the summer survey, the firms of the responding production managers did not significantly differ

from the non-responding finns in terms of firm performance, but they were slightly larger. The firms of the responding IT managers did not significantly differ from the nonresponding firms in terms of firm performance and firm size. The establishments surveyed in the spring survey employed on average 68.5 percent of a firm's total emplcyees, indicating that the obtained measures are largely representative of the firm. Similarly, the establishments surveyed in the summer survey employed on average 70.0 percent of the firm's total employees. On average, the interviewers had to contact a firm 8.0 times to generate one interview in the spring survey and 9.8 times to generate one interview in the summer survey. The interviews in the spring survey lasted on average 44.4 minutes. In the summer survey, the production manager interviews lasted on average 40.9 minutes, and the IT manager Interviews lasted on average 13.4 minutes. The key characteristics of the spring survey and the summer survey are summarized in Table 1.1 (next page). The derivation of information on IT complements, such as organization design and corporate strategy, as well as firm perfonnance from three independent sources was very important to avoid single source bias that may result in unrealistically high correlations between IT and IT complements or between these combinations and firm performance. However; I had to combine two 2000-2008 time-series datasets on IT and firm performance with 2008 cross-sectional datasets on organizational structure, HRM practices, and corporate strategy. This is legitimated by the fact that organization design and corporate strategies are often argued to be subject to inertial forces and to persist


1.3 The datasets employed in the two empirical studies

11

Table 1.1 The spring survey and the summer survey Spring survey

Summer survey

March 2008

August and September 2008

Type of potential respondents

Production managers

Production managers IT managers

Pool of potential respondents

600

2,369

Contacted firms

552

1,788

Interviews

259

784 (production managers) 454 (IT managers)

Response rate - overall (in percent)

43.2

33.1 (productionmanagers) 57.9 (IT managers)

Response rate - contacted finns (in percent)

46.9

43.8 (production managers)

Number of double secrings

55

245 (production managers)

Percentage of double secrings

21.2

31.3 (production managers)

Average percentage of firm employees working in the interviewed establishments

68.5

70.0

Average number of firm contacts per interview

8.0

9.8

Average interview duration (in minutes]

44.4

40.9 (production managers) 13.4 (IT managers)

Survey period

over time. I thus assume that organization design and corporate strategies changed only marginally in the years before the 2008 surveys. Further; the measures for organization design may be interpreted as an indicator of the direction of the crganizational change that took place during the years before the survey. As the complementarities between IT and crganization design can be examined in both causal directions, this does not change the interpretation of the results. For example, the finding that greater IT investments in 2006 and greater decentralization in 2008 are associated with greater firm performance in 2006, 2007, and 2008 would indicate that IT induced greater decentralization over time, which in turn allowed the firm to increasingly exploit IT's potential for performance enhancements. To mitigate the potential for biased corporate strategy measures, respondents were asked to indicate the firm's realized corporate strategy


12

1 Introduction

during the last five years in the case of the spring survey or three years in the case of the summer survey. Subsequently, either in the main regressions or at least in robustness tests. only the three- or five-year periods prior to the 2008 survey were used from the

datasets on IT and firm perfonnance (that Is, the periods 2003-2008 and 200S-200S, respectively]. In the summer survey, respondents were additionally asked f端r strategy changes during the last three years. If a respondent indicated that the corporate strategy strongly or very strongly changed during this period, all observations of this firm are

excluded from analysis in Chapter 4. Manufacturing firms were chosen in order to concentrate on one questionnaire, to avoid issues with interpreting the output of service firms in the production function estimations, and to avoid any bias from differences across organizational structures, HRM practices, corporate strategies. and other firm characteristics that are required for the generation of products versus the generation of services. German and Polish firms were chosen for a practical reason, that Is, the availability of native speakers as interviewers. Production managers were chosen as potential respondents because they are typically in the upper middle management and thus able to deliver insights on both lower-level issues such as organizational structure and HRM practices as well as the more firm-wide issue of corporate strategy. Similarly, IT managers are able to provide a full overview of how IT is used in their firm. Further, through the narrow set of informants, single

Informant bias is held relatively constant. The information was gathered in part using classic Likert scales. For example, respondents were asked to indicate their agreement with several statements on their firm on a scale from 1 CI do not agree at all') to 5 CI stronglyagree'). In addition, an innovative interview method was employed that was developed and first used by Bloom and Van Reenen (2007). This method combines qualitative and quantitative data-gathering methods and thus allows a more detailed insight into the finns than is possible with Likert scale questions. Further, this method avoids the problem of social desirability. This survey interview method was mainly used for HRM practices as they are particularly subject to the problem of social desirability. For example, some respondents might give unrealistically optimistic answers to questions on the amount of employee training if they feel that greater amounts of training are generally considered 'good'.


1.3 The datasets employed in the two empirical studies

13

Respondents were asked open questions and encouraged to answer freely. The interviewers were provided with prepared questions to ask. They began with open and quite general questions and continued with more detailed questions. Interviewers were also encouraged to deviate from the prepared questions if needed or suitable, to ask their

0"WIl

questions, and to ask for examples. Thus, a conversation led by the

interviewer developed for each question, which was not ended by the interviewer until she or he had a full picture of the topic in question. Subsequently, the interviewers scored respondent answers on a scale from 1 to 5, while the respondents did not know that their open answers were scored, assuring that the problem of social desirability did not occur. In the spring survey, the interviewers were provided with example answers for the scores of 1, 3, and 5 to facilitate secring. The scores of 2 and 4 were used for intermediate cases. An example question is shown in Figure 1.1.

Prepared guestions Could you describe the training programs that are provided for workers? What are the contents of these training programs? What are the goals of these training programs? How often do these trainin ~ oroerams take nlace? Examnle answer for score 1 Examnle answer for score 3 Examnle answer for score 5 Workers do not participate in Workers participate in In addition to what is regular training programs regular training programs !ctescribed in score 3, workers that upgrade or expand their that are directed at upgrading participate in regular training capabilities. There are no or expanding existing programs that are directed at training sessions, or only capabilities that are needed acquiring additional training sessions that are for each worker's job. capabilities from other areas mandated by law (for Ior general capabilities like example, regarding job leadership capabilities or sosafetV1 take nlace. Icalled 'soft skills'. Figure 1.1 Survey question on trammg mtensrty and vanety m the spnng survey In the summer survey, the interviewers had to answer questions on a scale from 1 ('not important at all') to 5 ('very important'] subsequent to discussions with respondents. An example question is shown in Figure 1.2 (next page). This interview method was passed by the Human Subjects Committee of Stanford University for the study of Bloom and Van Reenen (2007). The lack of respondent awareness of being scored is acceptable because it is necessary to reduce the problem of social desirability, is not risky for the respondents or their finns as the data are held confidential, and is temporary as the respondents are debriefed after the end of the project (Bioom and Van Reenen, 2007).


14

1 Introduction

Employee appraisal Open questions to respondent Please describe howworker performance is appraised. Which criteria are used to appraise worker performance? Da employee talks take place? Ifyes, please describe the typical process and contents of such a talk. Please describe the effect of performance appraisals on wages. How often da performance appraisals take place? Rating questions to interviewer (seale fram 1 ('not irnportantat att] to 5 ('very importantJ) (1) How irnportant is the role of objective criteria, such as quantitative or qualitative individual performance, teamjdepartmentjshift performance, or firm performance, as critera during employee appraisal? (exploitative component) (2) How important are the roles of subjective criteria, such as interpersonal skills, problemsolving skills, and personality traits (for example, for team work andjor autonomous decision-making and acting) or their improvement, as criteria during employee appraisal, and how important is the role of behavior, including suggestions for improvement and the pro-active participation in decision-making processes, as a criterion during employee appraisal? (exploratory component) Figure 1.2 Survey question on employee appraisal in the summer survey While the interview method reduces the problem of social desirability, it may result in bias through the individual characteristics of interviewers or their prejudices with regard to specific firms. As all interviewers ran a sufficient number of interviews, the first problem can be mitigated by controlling for interviewer fixed effects in the multivariate analysis, as shown in Chapter 3. In the spring survey, every individual interviewer conducted between 32 and 52 Interviews. In the summer survey, every individual interviewer conducted between 49 and 117 production manager Interviews and between 14 and 74 IT manager Interviews. The second problem is reduced by the fact that the interviewers were only provided with the name, telephone number, and a short description of the primary industry sector of the firms they contacted. They were not provided with more detailed information that might have led to prejudice (for example, information on firm performance]. As the interview method strongly relies on the interviewer capabilities, several steps were undertaken to ensure a high interview quality. The 15 student Interviewers came from management, economics, and sociology and were chosen in two-step assessment procedures. Selection criteria included prior experience with the topics of the survey


1.3 The datasets employed in the two empirical studies

15

and data gathering in general as well as performance in a simulated interview situation. Interviewers were intensively trained in the surveys' background, method, questions, and software during two-day training periods before the start of the survey periods. Several simulated Interviews were conducted during these trainings. The interview quality was further enhanced through regular team meetings and through interview monitoring by two PhD supervisors that were always present. The supervisors listened to apart of the Interviews in the first week of both survey periods, scored the respondent answers independently from the interviewer, and discussed possible differences between interviewet- and supervisor scores to enhance interview quality. Mcreover; during the spring survey, the supervisors continued to listen to apart of alt Interviews and to assign scores to respondent answers independently from the interviewer (that is. 'double scoring']. In particular, the supervisors listened to 55 interviews, or 21.2 percent ofthe Interviews. In the summer survey, the double scoring was taken over by the student research analysts. They listened to 245 of their colleagues' production manager interviews, or 31.3 percent of all production manager Interviews. The double secrings were used as a test of inter-rater reliability. To analyze

if two raters give similar scores to the answers of the same respondent, the partial correlation of the first interviewer secrings with the second interviewer secrings was calculated. This means that first interviewet- and second interviewer fixed effects were controlled for when the correlation coefficient was calculated. The partial correlation coefficient for the HRM variable used in Chapter 3 is as high as 0.90 and highly significant (see Figure 1.3, next page). The partial correlation coefficients for the six HRM variables used in Chapter 4 range from 0.63 to 0.81 and are highly significant. This implies that this interview method leads to relatively homogenous results even when different interviewers score the same interview. In the multivariate analyses, only the first interviewer secrings are used. Most survey items were derived and adapted from prior studies, as described in more detail below. Prior to the spring survey, the survey instrument was discussed at length with five production managers. This pretest led to adaptations of the survey Instruments. Prior to the summer survey, the survey instrument for production managers was discussed at length with one production manager and then pretested in Interviews with 15 additional production managers. The survey instrument for IT managers was discussed in detail with one IT manager and two PhD students specializing in Information systems research. Subsequently, the survey instrument was pretested in Interviews with ten IT managers. The questionnaires for Polish production


16

1 Introduction

and IT managers were translated from German into Polish by an accredited translator and then rechecked by two Polish managers and two Polish research analysts.

inter-rater reliability HRM variable used in Chapter 3

~ ~

c

"s

•••

.0

.~

E

I.!:!

., • ., • ~

-1

-.5

0

supervisor scoring

.5

Figure 1.3 Correlation between interviewet- and supervisor secrings in the spring survey To achieve high response rates in both survey periods, several steps were undertaken. The student interviewers and the two supervising PhD students were located in a specially equipped telephone studio. The interviewers targeted the switchboards of the finns and asked für the production/I'I' managers or other employees in a similar position. An important factor was to convince the switchboards as well as the potential respondents (and sometimes their assistants) to take part in the survey. The interviewers explicitly introduced the survey as a 'scientific research project' rather than a 'survey', which would often lead to an immediate rejection by switchbcards, assistants, or potential respondents. Interviewers assured potential respondents that they would not ask for financial or any other clearly confidential information and that the respondent answers would remain anonymous. Potential respondents were offered a free summary of the research results. An e-mail with a written description of the research project and endorsement letter from the Chairman of Deutsche Telekom Stiftung and former German Foreign Minister Klaus Kinkel was offered to all potential respondents and sent upon request. To further underline the respectability of the project, individual e-mail-addresses for the interviewers and a website with information on the prcject, the interviewers, and the supervisors were set up. As soon as an


1.3 The datasets employed in the two empirical studies

17

interview appointment had been made, respondents were sent an approval of the appointment and areminder 24 hours prior to the appointment. To minimize interview terminations, the interview started with rather neutral questions (for example, on the number of firm employees), with more critical questions (for example, on the owner of the firm) only occurring at the end of the interview. The main part of an interviewer's pay was based on the number of Interviews conducted by her or him. Additionally, team bonuses were paid to all interviewers as soon as the entire team had conducted certain overall amounts of Interviews. The performance of each interviewer and of the whole team was monitored on a daily basis and made public to the entire team at the end of each week, Together, this led to strong incentives for each individual interviewet- to continuously reassess her or his procedure in making interview appointments and conducting interviews: it also encouraged interviewers to exchange best practices. More formally, best practices were also discussed during regular and spontaneous team meetings moderated by the supervisors. Further; the pools of potential respondents were randomly segmented and subsequently assigned to the interviewers at several predefined times during the survey periods. As the interviewers were thus endowed with only a limited number of potential respondents at all times, they were incentivized to maximize the response rate rather than only the absolute number of contacted and interviewed firms. To maximize efficiency, the whole process of making interview appointments and conducting the Interviews was managed using a Microsoft Access database. This database allowed every interviewet- to protocol her or his attempts to make appointments, manage her or his calendar, and directly protocol the Interviews. Due to increasing experience in how to approach potential respondents and an accumulated stock of interview appointments, the daily average number of Interviews per interviewet- increased from 0.9 in the first week to a maximum of 3.3 in the last week of the spring survey. In the summer survey, the daily average number of production manager Interviews per interviewer increased from 1.8 in the first week to a maximum of 2.7 in the sixth week. The daily average number of IT manager Interviews per interviewet- increased from 0.2 in the first week to a maximum of 1.9 in the seventh week. The dataset used for the analysis in Chapter 3 is constructed from the CITDB, ORBIS, and the spring survey. After removing observations with missing data, the final sample used for the multivariate analysis consists of 182 ftrms, with each firm represented by at least one of 572 observations. These firms do not significantly differ from all other German manufacturing firms included in the CITDB and ORBIS with regard to the number of firm emplcyees, value added, total assets, or the number ofPCs.


18

1 Introduction

The dataset used f端r analysis in Chapter 4 is constructed from the eITDB, ORBIS,and the summer survey. Two separate samples f端r individual ambidexterity and technological ambidexterity are created to minimize the

1055

of observations through missing values.

A sample of 258 firms, with each firm represented by at least one of 457 observations, is available f端r the analysis of individual ambidexterity. A sample of 302 firms, with each firm represented by at least one of 502 observations, is available f端r the analysis of technological ambidexterity. The firms in both samples do not significantly differ from

all other German and Polish manufacturing finns included in the eITDE and ORBISwith regard to firm performance, but they are slightly smaller.


Chapter 2 Information technology and firm performance: An integrative model of the role of complementarities 2.1

Introduction

Postulating a positive impact of information technology (IT) on firm performance has had theoretical appeal fĂźr decades (Leavitt and Whisler, 1958). Empirical research has well documented this relationship (Barua and Mukhopadhyay, 2000; Brynjolfsson and Hitt, 2000; Chan, 2000; Dedrick et al., 2003; Dehning and Richardson, 2002; Kohli and Davaraj, 2003; Melville et al., 2004) after having overcome the problems that initially led to the so-called 'computer productivity paradox' (Baily, 1986; Brynjolfsson and Yang, 1996; Rcach, 1987). However, empirical research suggests that firms differ significantly in their ability to realize performance gains from IT (Brynjolfsson and Hitt, 1995; Loveman, 1994). Further, IT's perfonnance impact was found to increase with longer measurement periods (Brynjolfsson and Hitt, 2003), and the stock market valuations of IT have been found to be high compared to other capital types (Brynjolfsson et al., 2002). These findings have been explained by factors complementary to IT that differ from firm to firm, change slowly, and remain invisible in the investment statistics but appear in the stock market valuations of finns (Brynjolfsson and Hitt, 1998a). Thus, the extent of IT's perfonnance impact has been supposed to depend on a variety of IT complements (Brynjolfsson and Hitt, 2000; Melville et al., 2004). Factors inside a firm's organization design and factors outside the organization design are two different types ofIT complements that have been examined in prior studies. For example, it has been shown that IT payoffs rely on complementary organizational structures and human resource management (HRM) practices, such as decentralized decision authority, team work and broader job classifications (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). Additionally, an open organizational culture was found to enhance IT's performance impact (Powell and DentMicallef 1997). Concerning factors outside the organization, higher IT payoffs have been found under corpcrate strategies that focus on new products and markets as well as in dynamic environments [Sabherwal and Chan, 2001; Li and Ye, 1999).

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6_2, Š Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


20

2 Information technology and firm perfonnance

A major drawback of the empirical studies on IT complements inside and outside the

organization design is that both types of complements are not analyzed in conjunction. This leads to ambiguous results such as the finding that IT and decentralization are

complementary, although prior theoretical and case-based studies have argued that combining IT with centralization may also be effective (George and King, 1991; Robey, 1977). Further, bundles of IT investments and intemal IT complements that do not, f端r example, support a firm's corporate strategy are likely to have limited perfonnance impacts (Oh and Pinsonneault, 2007).

Similarly, it is of little use to know that IT

investments may enhance firm performance in conjunction with certain corporate strategies if it remains unclear how IT must be aligned with other organizational elements to deliver the functionality required by the strategy. Further drawbacks of the existing empirical studies on IT complements include often misleading measurements of IT and IT complements, a lack of a common theoretical background, and a lack of detailed analysis of the mechanisms by which IT and IT complements combine. Summarizing, Melville et al. (2004: 303) state: "What is not understood is the specific nature of complementarities, i.e., what specific resources are complementary to one another, under what conditions, and how are the attributes of complementary resources related to business process and organizational performance impacts?" Empirical research that leaves room for improvement is a reason to seek better theory. To address the criticisms outlined and guide future empirical research, I develop an integrative model of IT complements. I draw on contingency thecry to argue that IT must be aligned simultaneously with so-called 'horizontal IT complements' inside an organization design and 'vertical IT complements' that determine the characteristics of the organization design. While contingency theory provides more general explanations for the need to achieve horizontal and vertical fit, I use information processing theory to examine the specific mechanisms that explain why IT is complementary with several horizontal and vertical elements. Based on this integrative theoretical framework, I develop propositions on how to exploit the complementarities among IT, horizontal IT complements, and vertical IT complements. In particular, I suggest that IT types that support the information exchange up and down a firm's hierarchy should be horizontally aligned with the centralization of decision rights, high standardization, and HRM practices that focus on skills and knowledge needed by lower-level employees in order to implement rules, procedures, and instructions that are imposed on them by central decision makers. This organization design should be vertically aligned with a comparably undemanding environment,


2.2 An overview of the empirical research on IT complements

21

a corporate strategy that focuses on existing product market domains, and an organizational culture that becomes manifest in a high degree of fonnality. IT types that support the information exchange between a firm's subunits should be horizontally aligned with the decentralization of decision rights, low standardization, and HRM practices that focus on diverse skills and knowledge needed by lower-level employees for direct communication with each other and autonomous cocrdination. This organization design should be vertically aligned with ademanding environment, a corporate strategy that focuses on new product market domains, and an organizational culture that becomes manifest in a low degree of formality. This study contributes to the theoretical and empirical research on IT complements in several ways. First, I provide a structured review of existing studies on IT complements, with an emphasis on empirical research. Second, I provide a consistent theoretical framework to guide future research and facilitate the accumulation of comparable knowledge and wisdom on IT complements. Third, I advance testable propositions on complementarities between IT and IT complements that go beyond existing findings. Fourth, I provide suggestions on howto operationalize and measure IT, IT complements. and firm perfonnance in a meaningful way. I begin with abrief overview of the existing literature on intemal and external IT complements and summarize their drawbacks in section 2.2. Based on this, I introduce stepwise the components of an integrative model of IT complements in section 2.3. Section 2.4 provides a detailed analysis of the complementarities between IT and horizontal IT complements, while section 2.5 examines vertical IT complements. I conclude with implications for future research in section 2.6.

2.2

An overview ofthe empirical research on IT complements

In this section, I give a short overview of the key conceptual and methodological characteristics and flaws of existing empirical studies on IT complements to motivate the integrative model of IT complements. A more comprehensive review of these prior findings is offered in sections 2.4 and 2.5.


22

2 Information technology and firm perfonnance

2.2.1 Studies on internal or external IT complements The conditions under which IT positively affects firm performance have been examined from two theoretical perspectives, namely, from a so-called 'resource-centered

perspective' and from a 'contingency perspective' (Oh and Pinsonneault, 2007).

2.2.1.1 Resource-centered perspective The resource-centered perspective contains studies that take a 'production function

view' and mainly stern from the literature on organizational economics and studies that draw on the 'resource-based view' (Oh and Pinsonneault, 2007). The studies employing a production function view foeus on the size of IT investments (Oh and Pinsonneault, 2007) and draw on the economic concept of complementarity. Two elements are complementary if the marginal benefit of the first element increases in the levels of the second element, and vice versa (Milgrom and Roberts, 1990; 1995). The studies show that the positive effect of IT investments on firm performance is enhanced through greater levels of the decentralization of organizational decision rights and supporting HRM practices (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). A detailed review of the production function view is available from Brynjolfsson and Hitt (2000). The studies employing a resource-based view focus less on the size of IT investments and more on their scope and properties (Oh and Pinsonneault, 2007). Following the resource-based view of the firm (Barney, 1986a; 1986b; 1991; Wemerfelt, 1984), it is argued that unique and hard-to-imitate bundles ofIT resources have to be established to achieve superior firm performance (Oh and Pinsonneault, 2007). Physical IT resources. that Is, IT investments, must be combined with human IT resources, such as technical and managerial IT skills or top management commitment to IT, and intangible ITenabled resources, such as a flexible organizational culture (Bharadwaj, 2000; Mata et al., 1995; Powell and Dent-Micallef 1997). Detailed reviews ofthe resource-based view in IT research are available in Melville et al. (2004) and Wade and Hulland (2004). Note that most studies with a resource-centered perspective focus exclusively on IT complements that are part of a firm's organization design, such as organizational structure and human resources, while external factors. such as a firm's corporate strategy or environment, are widely disregarded.


2.2 An overview of the empirical research on IT complements

23

2.2.1.2 Contingency perspective The studies employing a contingency-perspective mainly stern from the information systems literature and draw on the proposition that firm performance is a result of the consistency (or 'fit) among organizational elements. These studies thus argue that IT investments have to be aligned with a firm's corporate strategy and environment to positively affect firm performance (Chan et al., 1997; Croteau and Bergeron, 2001; Li and Ye, 1999; Oh and Pinsonneault, 2007; Sabherwal and Chan, 2001). These studies are largely consistent in their finding that combining IT investments with corporate strategies that focus on the exploration of new product market domains rather than efficiency in existing product market domains (Croteau and Bergeron, 2001; Li and Ye, 1999; Oh and Pinsonneault, 2007; Sabherwahl and Chan, 2001), and that combining IT Investments with dynamic rather than stable environments (Choe, 2003; Li and Ye, 1999), is beneficial to firm perfonnance. An annotated bibliography of the contingency perspective is available in Chan and Reich (2007). Note that all studies with a contingency perspective exclusively focus on IT complements that are not part of a firm's organization design, such as a firm's corporate strategy or environment, while internal factors such as a firm's organizational structure or human resources are disregarded.

2.2.2 Studies on internal and external IT complements Onlya few studies analyze the role of both intemal and external IT complements in the IT-performance relationship. Drawing on the so-called 'strategic alignment model' (Henderson and Venkatraman, 1993), one case-based study and one large-sample study analyze the six possible types of alignment between two intemal factors, namely, a firm's (1) organizational infrastructure and processes and its (2) IT infrastructure and processes, and two external factors. that Is, a firm's (1) business strategy and its (2) IT strategy (Bergeron et al., 2004; Sabherwal et al., 2001). Both studies define particular strategies and structures that fit each other; argue that all six alignment types are equally important for firm performance, and that the overall alignment is high if "the number of alignments that are high exceeds those that are low" [Sabherwal et al., 2001: 182). Bergeron et al. (2001) use six different definitions of fit as proposed by Venkatraman (1989) to separately examine the perfonnance effects of aligning strategie IT


24

2 Information technology and firm perfonnance

management with one intemal IT complement, that Is, the firm's structural complexity,

and two external IT complements, namely, a firm's (1) strategic orientation and (2) environmental uncertainty.

Also drawing on Henderson and Venkatraman's (1993) model, Rivard et al. (2006) separately examine and find positive perfonnance effects of aligning business strategy with the IT strategy (that Is, 'IT support f端r strategy') and the organizational infrastructure and processes with the IT infrastructure and processes (that Is, 'IT support f端r firm assets'). They further show that IT support f端r strategy and IT support

f端r firm assets are complementary; that Is, that both types of alignment have a joint positive effect on firm perfonnance.

2.2.3 The need for an integrative model of IT complements I summarize several conceptual and methodological drawbacks of these studies to motivate the development of an integrative model ofIT complements. First and foremost, one conceptual criticism is that prior studies often omit important factors. The studies with a resource-centered perspective focus largely on intemal IT complements. while the studies with a contingency perspective focus largely on external IT complements. Thus, the former studies indicate the potential for complementarities between IT and certain elements of organization design, but they do not explain which intemal IT complements are important in which settings (Melville et al., 2004). However; bundles of IT investments and intemal IT complements that do not support a

firm's corporate strategy or that are not suitable for a certain organizational environment are likely to be wasted or have limited performance impacts (Oh and Pinsonneault, 2007). One indicatcr for the limited explanatory power of the studies that disregard external factors is their one-sided findings on the relationship between IT investments and the degree of organizational centralization. While prior theoretical and case-based studies plausibly argue that depending on third factors, IT may be related to both centralization and decentralization (George and King, 1991; Robey, 1977), the abovementioned studies find evidence only for an IT-decentralization relationship (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). Analogous reasoning applies to the studies that focus on external IT complements. It is of little practical relevance to know that IT investments may enhance firm perfonnance in conjunction with certain corporate strategies or crganizational environments if it remains unclear how IT must be aligned with other elements of organization design to deliver the functionality required by the strategy or the environment. The problems of


2.2 An overview of the empirical research on IT complements

25

separately analyzing intemal and external IT complements remain largely unsolved in the studies that incorporate both types of complements. One study examines both types of alignment, though never in conjunction (Bergeron et al., 2001). Thus, this study has no advantages as compared to studies on only internal complements as well as compared to the studies on only external complements. Two further studies regard all types of alignment as equally important (Bergeron et al., 2004; Sabherwal et al., 2001). Thus, aligning IT with two intemal complements is regarded as just as well as aligning IT with both one intemal and one external complement. However, as argued above, three horizontally aligned complements are not useful if they are not also vertically aligned. The study by Rivard et al. (2006) examines the joint effect of IT support for firm assets and IT support for strategy and is thus the most sophisticated with regard to analyzingthe joint effect of intemal and external IT complements. However; this study is subject to criticisms with regard to measurement, as outlined below. The second major drawback of existing studies has a methodological nature and concerns the operationalization and measurement of IT and IT complements. Several studies do not operationalize IT as the size or type ofIT investment but directly measure to what extent IT is aligned with its complements (Chan et al., 1997; Choe. 2003; Croteau and Bergeron, 2001; Rivard et al., 2006; Sabherwal and Chan, 2001). For example, Rivard et al. (2006: 39) directly ask respondents "to evaluate the extent to which IT supports each element ofthe business strategy". Similarly, in the study from Chan et al. (1997: 131), each questionnaire item on the firm's corporate strategy (for example, "We are almost always searching for new business opportunities"] is reflected by an item on the firm's IT use (for example, "The systems used in this business unit assist in the identification of new business opportunities"]. It is probably a tautology that a corporate strategy's performance effect is enhanced through IT that is designed to support this strategy. Hence, directly measuring fit disables researchers from drawing conclusions on how to achieve this fit. For example, it remains unclear which specific IT types are complementary to which organizational structures or corporate strategies and how these IT types must be implemented. Thus, finding that a direct measure of fit between IT and other factors positively affects firm perfonnance largely enforces the well-known finding that IT can in general enhance firm performance through supporting several intemal and external factors. However; such findings do little to shed light on how exactly the complementarities between IT and these factors can be exploited. A related criticism is that the level of analysis of IT complements in several studies is too abstract (Bergeron et al., 2001; 2004; Chan et al., 1997; Rivard et al., 2006). For example, several operationalizations of corporate strategy include the extent of both


26

2 Information technology and firm perfonnance

'strategie aggressiveness', that Is, a firm's foeus on new markets. and 'strategic

defensiveness', that Is, a firm's foeus on existing markets. In particular, these studies measure the extent of strategie orientation as the average of strategic aggressiveness, strategie defensiveness, and other factors (Bergeron et al., 2001; 2004; Chan et al.,

1997). Similarly, the measure of so-called IT support f端r strategy in Rivard et al. (2006) incorpcrates both IT support f端r a differentiation strategy and IT support f端r a low-cost strategy. However, strategies that foeus on new markets versus strategies that foeus on existing markets and differentiation strategies versus low-cost strategies are regarded as fundamentally distinct strategies that require distinct organization designs (Miles and Snow, 1978; Porter, 1980). If IT is found to be complementary to such a measure of corporate strategy, one leams that IT can enhance any type of corporate strategy but not which types of corporate strategies and IT are complementary. A third criticism is that several existing studies do not examine the mechanisms through which IT complements enhance IT's perfcrmance effects (Melville et al., 2004). Other studies that try to explain these mechanisms often use distinct theoretical backgrounds, as outlined above. Future studies should at least highlight the common grounds of the production function thecry, the resource-based view, and the contingency theory to increase comparability. Finally, several studies focus on IT complements that are not specific to IT, but that enhance the performance impact of any technology. Examples are technical and managerial IT skills or top management commitment to IT (Bharadwaj, 2000; Mata et al., 1995; Powell and Dent-Micallef 1997). Fruitful studies should focus on those complements that are specifically relevant to IT's performance effects.

2.3

Toward an integrative model ofIT complements

As outlined above, existing empirical studies on IT complements leave room for improvement and thus serve as an impetus to seek better theory. Therefore, I develop an integrative model of IT complements, with elementary theoretical foundations and components presented stepwise below. In particular, while contingency theory provides the general intuition for the need to align IT with factors both inside and outside organization design, information processing thecry provides explanations for the detailed mechanisms by which IT and these factors combine. This novel arrangement of key elements yields an integrative framework to guide future empirical research.


2.3 Toward an inteerative model ofIT complements

27

2.3.1 Fit and complementarity Contingency theory postulates that firm perfonnance depends on the 'fit' (synonymous concepts include match, congruence, consistency, and alignment) among organizational elements, that Is, the effectiveness of individual elements is expected to be contingent on other elements (Van de Yen and Drazin, 1985; venkatraman. 1989). Many early contingency theories hypothesized that firm perfonnance relies on bi-variate alignments between elements inside and elements outside organization design (Van de Yen and Drazin, 1985), for example, between corporate strategy and organizational structure (Chandler, 1962) or between environment and organizational structure (Lawrence and Lorsch, 1967). 端ther theories have highlighted the need to create intemally consistent organization designs to enhance firm performance, for example, by aligning the decentralization of decision rights with specific HRM practices that support autonomous decisions (Bresnahan et al., 2002; Brynjolfsson and Hitt, 1998b; 2000; Brynjolfsson and Mendelson, 1993; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). As outlined above, these two notions of fit focus exclusively on either intemal or external complements and have been used in existing empirical studies on IT complements. However; more recent contingency studies stress the need to achieve both types of fit. First, 'horizontal fit' is achieved by aligning all elements of an organization design such as organizational structure and HRM practices with each other. Second, 'vertical fit' is achieved by matehing this intemally consistent organization design with contingency factors like environment, strategy, or culture (Becker and Huselid, 1998; 2006; Burton and 端bel, 2004; Delery and Doty, 1996; Doty et al., 1993; Drazin and Van de Yen, 1985; Miles and Snow, 1978; Nadler et al., 1992; Porter, 1996; Wright and Snell, 1998). Vertical fit is viewed as directing the elements of the organization design towards the functional demand imposed by a firm's environment, strategy, or culture. Horizontal fit is viewed as a means to efficiently allocate the elements of the organization design that offer this functioning (Andersen and [onsson, 2006; Gresov and Drazin, 1997; Wright and Snell, 1998). Conversely, a highly consistent organization design that does little to provide the functionality required by the contingencies has little value (Becker and Huselid, 2006). Further, it is likely that an organization design is inefficient if one or more elements of the design are only vertically aligned with environment, strategy, or culture, but not horizontally aligned with each other. Contingency theories that propose the need to achieve both horizontal and vertical fit are also termed system or configurational theories, as they often suggest several ideal systems or configurations of


2 Information technology and firm perfonnance

28

elements that are equal alternatives to achieve horizontal and vertical fit (Delery and Doty, 1996; Doty et al., 1993; Drazin and Van de Yen, 1985). It is possible that two or more contingency factors demand different functionalities, such as in the case of so-called 'situational misfit' (Button and 端bel, 2004; Gresov, 1989; Gresov and Drazin, 1997). In such a situation, it is possible to choose a horizontally fitted organization design that displays vertical fit with one contingency factor at the expense of the vertical fit with the other contingency factors. Altematively, it is possible to choose a hybrid organization design that may lack aperfeet horizontal fit but that displays at least partial vertical fit with more than one contingency factor. While the first option should lead to lower firm perfonnance than under situational fit, there is no agreement on the performance implications of the second solution (Gresov, 1989; Gresov and Drazin, 1997). Some authors argue that hybrid organization designs lack horizontal fit and lead to poor performance (Gresov, 1989; Gresov and Drazin, 1997; Porter, 1980; 1985; 1996). Others suggest that it is possible to create horizontally congruent hybrid designs and that the performance of such organization designs is equal to the performance of designs that focus on only one contingency (Miles and Snow, 1978) or even higher (March, 1991; Tushman and O'Reilly, 1996). Several different specifications of fit have been used in the literature, including fit as moderation,

mediation,

matehing.

gestalts,

profile

deviation,

and

covariation

(Venkatraman, 1989). The operationalization of fit as moderation is most fruitful for a model that incorporates horizontal and vertical fit. Acccrding to the moderation view of fit, the impact of a predictor variable (for example, IT) on a criterion variable (e.g., firm performance) is dependent on a third moderating variable (for example, organizational structure). The moderator affects the strength and direction of the relations hip between predictor variable and criterion variable, that Is, the predictor and the moderating variable have a joint or interaction effect on the criterion variable (Venkatraman, 1989). The moderation view of fit is also reflected through the economic concept of complementarity (Besanko et al., 2007; Milgrom and Roberts, 1990); two elements are complementary if the marginal benefit of the first element increases in the levels of the second element,

and vice

versa

(Stieglitz

and

Heine,

2007).

Although

the

complementarity view largely corresponds to the moderation view, it points out more explicitly that to explain an interaction effect of two elements, it does not matter which of the two elements is defined as the predictor variable and which as the moderating variable (Bresnahan et al., 2002). That Is, in the abovementioned example, it does not matter if IT increases the benefits of organizational structure or if organizational structure increases the benefits of IT. Thus, a model that includes horizontal fit and


2.3 Toward an inteerative model ofIT complements

29

vertical fit can be viewed as a 'double' moderation, or complementarity, view of fit. That is. the interaction effect of two or more elements of the organization design (for example, IT and organizational structure) on firm performance is further moderated by one or more contingency factors (for example, corporate strategy).

2.3.2 Organizational information processing and IT While the fit and complementary literatures give general explanations for the need to achieve horizontal and vertical fit, Information processing thecry offers specific explanations for the existence of complementarities between IT and its horizontal and vertical complements. Galbraith (1973; 1974; 1977) described the organization design problem as an information processing problem. To benefit from economies of scale and specialization, finns divide their activities into subtasks. The subunits that are responsible for these subtasks lack a global view of all subtasks, resulting in a reduced ability to conduct the subtasks in line with the firm's overall goals. Thus, coordination must be achieved through information processing. I summarize the three main arguments of infonnation processing theory that are fruitful for an integrative model of IT complements. First, there are two alternative types of information processing that serve the coordination of subunits. The type of infonnation processing is determined by a firm's organizational structure. Under a centralized organizational structure, which allocates decision rights high in the hierarchy, information from the subunits is referred upward the hierarchy to central decision makers that have a better overview of the subtasks. Subsequently, central decision makers make decisions and disseminate instructions to the subunits. Thus, information is processed vertically, up and down the hierarchy between subunits and central decision makers (that is. vertical information processing). Under a decentralized organizational structure, which allocates decision rights low in a hierarchy, the subunits communicate directly with each other. Thus, they are able to themselves decide how to act in line with a firm's overall goals. Here, information is processed horizontally between lateral subunits (that Is, horizontal information processing) (Aoki, 1986; Burton and 端bel, 2004; Galbraith, 1973; 1974; 1977; Premkumar et al., 2005; Tushman and Nadler; 1978). Second, both vertical information processing and horizontal information processing may be enhanced through IT (Burton and 端bel, 2004; Galbraith, 1973; 1974; 1977). IT increases the speed, quantity, and quality of infonnation that can be transferred


30

2 Information technology and firm perfonnance

vertically and horizontally between organizational units (Anand and Mendelson, 1997; Dewett and [ones, 2001; Huber, 1990) and enhances the information processing capabilities of organizational units with regard to speed, quantity, and quality

(Brynjolfsson and Hitt, 2000; Brynjolfsson and Mendelson, 1993; Galbraith, 1974; Leavitt and Whisler, 1958].2 Thus. information processing provides a fruitful explanation f端r the apparent need to achieve horizontal fit between organizational structure and IT. F端r example, a centralized organizational structure, which requires vertical infonnation processing, benefits from IT, because IT enhances vertical information processing. In turn, as IT enhances vertical information processing, it will positively affect firm performance if its capabilities are exploited by a centralized organizational structure. Third, both types of organizational structures as well as information processing have relative strengths and weaknesses with regard to the quantity and quality of information processing that they provide (Aoki, 1986; Galbraith, 1973; 1974; 1977). In turn, contingencies such as a firm's environment or other factors require different quantities and qualities of information processing, that Is, they demand different types of organizational structures and infonnation processing (Burton and 端bel, 2004; Galbraith, 1973; 1974; 1977). Further; a specific organizational structure is more beneficial if the firm pursues a corporate strategy that requires the type of information processing that is offered by this particular organizational structure. Thus, the information processing requirements of contingency factors and the infonnation processing capacity of organizational structure provide a fruitful theoretical explanation forthe need to achieve vertical fit.

2.3.3 An integrative model 01IT complements I develop an integrative model of IT complements (see Figure 2.1, next page) by synthesizing the key findings of existing theoretical and empirical studies on IT complements. the fit and complementarity literatures, and information processing

thecry.

2 Note that Galbraith (1973; 1974; 1977) viewed IT as suitable to support vertical information processing only. However, today, IT is also capable of supporting horizontal information processing, as further outlined below (Burton and 端bel, 2004).


2.3 Toward an inteerative model ofIT complements

31

Contingency factors Environment, strategy, culture

Infonnation processing requirement

~

Vertical fit

t

Infonnation processing capadty

I

Organization design Organizational structure HRM practices

-+

Hcrfzontal fit

1

Information technology

Figure 2.1 An mtegratrve model of IT complements I consider all IT complements that have been analyzed in existing studies, that Is, a firm's organizational structure, HRM practices, environment, strategy, and culture. I do not consider complements such as technical and managerial skills or top management commitment, which are not specific to IT. Note that I distinguish between horizontal and vertical IT complements, that Is, complements that determine an organization design's information processing capacity and complements that determine an organization design's infonnation processing requirement. The terms 'intemal IT complements' and 'external IT complements' used above were only working definitions, as some researchers would define a firm's strategy and its culture as intemal factors. while only the environment would be seen as an external factor. IT has the potential to enhance vertical infonnation processing and horizontal information processing, while organizational structure and HRM practices determine if a

firm's subunits are cocrdinated through vertical information processing or horizontal information processing. As a consequence, IT must be horizontally aligned with organizational structure and HRM practices to positively affect firm performance. For example, a firm with an organizational structure and HRM practices that rely on horizontal information processing does not benefit from IT that is designed to enhance vertical information processing. Different horizontally aligned bundles of IT, organizational structure, and HRM practices may have relative strengths and weaknesses with regard to the quantity and quality of their infonnation processing capacities. In turn, contingency factors such as a firm's environment; its corporate strategy, or its cultures may demand different quantities and


32

2 Information technology and firm perfonnance

qualities of information processing capacities. As a consequence, the organization design and contingency factors have to be vertically aligned. F端r example, a firm with a strategy that demands horizontal infonnation processing does not benefit from an organization design that operates with vertical information processing. Thus, IT has to be matched with horizontal and vertical complements at the same time to positively affect firm perfonnance. It is not fruitful to analyze IT's alignment with

either horizontal complements cr vertical complements in Isolation. Note that this model does not make any statements that horizontal fit and vertical fit have to be achieved in a specific sequence or that either horizontal fir or vertical fit are dominant. On the one hand, it is likely that some elements of the integrative model of IT complements are harder and slower to change than others. On the other hand, firms can ultimately change or influence every component of the model to (re)align it with the other elements. For example, a firm can even influence its environment through lobbying. through disruptive Innovations, or ultimately through completely changing its area of operation.e The criterion variable of the integrative model of IT complements is firm performance, which has been operationalized as overall firm performance based on accounts or stock market valuations in a major part of the studies on the IT-firm performance relationship and the role of IT complements in this relations hip. However, it has been debated as to whether intermediate-level performance measures such as operating time, product quality, or IT user satisfaction are better suited to evaluate the perfonnance effects of IT or bundles of IT, since their perfonnance effects may not translate into overall firm performance (Brynjolfsson and Hitt, 2000; Dehning and Richardson, 2002; Melville et al., 2004). I argue that intermediate-level performance measures are fruitful to examine the alignment of IT with single IT complements. However, if IT's performance effects do not ultimately translate into overall firm performance, this is an indicator of an incomplete alignment with both horizontal and vertical IT complements. For example, IT that is horizontally aligned with other elements of organization design may well enhance operating time. However, this may not translate into overall firm performance if vertical fit is missing. that is. for example, if the firm's corporate strategy does not focus on efficiency but on delivering unique products. Thus, overall firm performance is the right performance indicator for an integrative model of horizontal and vertical IT complements. 3 For example, t he German tourism and logistics company TU! AG was actually founded as the mining company Preussag AG in t he early 1920s.


2.4 Horizontal fit between IT and IT complements

33

Every component of the integrative model ofIT complements is examined in more detail below, with the major goal to motivate future empirical research on IT complements that overcomes the abovementioned drawbacks of existing studies. First, I develop different ideal configurations of IT, horizontal IT complements. and vertical IT complements that positively affect firm performance. Second, I explain the detailed mechanisms by which IT and its horizontal and vertical complements combine. Third, I propose richer definitions of horizontal and vertical IT complements that allow for more reliable conclusions than existing studies.

2.4

Horizontal fit between IT and IT complements

This section provides an analysis of horizontal IT complements and of IT itself Two horizontal IT elements that have been analyzed in prior empirical studies are the degree of the centralization of decision rights and supporting HRM practices. Both components influence the distribution of power in a firm [lasperson et al., 2002; Townley, 1993). However; I argue that to gain a full picture of the complementarities between IT and the power distribution in a firm, more detailed layers of both the degree of centralization and HRM practices must be considered. Further, an additional means of influencing the power distribution in a firm, namely standardization, has to be considered. Drawing from contingency thecry, I develop two ideal configurations of IT and horizontal IT complements, that Is, one configuration that allocates organizational decision authority high in the hierarchy and that operates with vertical information processing, and one configuration that allocates organizational decision authority low in the hierarchy and that operates with horizontal infonnation processing.e

The macro-Ievel organization of a firm, for example, with regard to vertical integration (Dewan et al., 1998; Hitt, 1999), diversification (Chari et al., 2008; Dewan et al., 1998; Hitt, 1999; Tanriverdi, 2006), size (Brynjolfsson et al., 1994), and position on the continuum from hierarchy t o cooperation t o market (Clemons and Row, 1993; Gurbaxani and Whang. 1991; Malone et al., 1987) has also been hypothesized and shown to be interrelated with IT. However, t hese macro-Ievel phenomena can largely be viewed as high-level decisions regarding the degree of centralization and specialization or are explained through a firm's corporate strategy. Thus, t hey are not considered as separate IT complements in the remainder of t his thesis.

4


34

2 Information technology and firm perfonnance

2.4.1 The degree 01centralization The degree of centralization refers to the extent to which the decision rights in a firm are allocated high in the hierarchy (that Is, centralization) or low in the hierarchy (that Is,

decentralization).

2.4.1.1 Centralization As outlined above, under centralized organizational structures, infonnation from subunits is referred upward in a hierarchy to a central decision maker who processes the information, makes decisions, disseminates instructions downward to the subunits,

and monitors their implementation (Aoki, 1986; Galbraith, 1974). However, lower-level employees in subunits often have better 10ca1 infonnation about supplier and customer needs, competitor behavior and production conditions (Anand and Mendelson, 1997; Roberts, 2007). This infonnation is often hard-to-transfer and requires a 'feel', 'judgment', or 'experience' (Anand and Mendelson, 1997; Aoki, 1986; Gurbaxani and Whang, 1991; Leavitt and Whisler, 1958). Further, if much information is referred upward the hierarchy, central decision makers may suffer from infonnation overload, which reduces decision speed and quality (Bolton and Dewatripont, 1994; Galbraith, 1974; Leavitt and Whisler, 1958). Centralization and IT may be complementary, as IT can enhance the benefits from centralization through its abovementioned effects on the speed, quantity, and quality of information transfer and processing. In particular, IT can facilitate the transfer of historically hard-to-transfer local infonnation from subunits upward to central decision makers (Bolton and Dewatripont, 1994; Brynjolfsson et al., 1994; Dewett and [ones, 2001; Galbraith, 1974; Gurbaxani and Whang, 1991; Huber, 1990; Leavitt and Whisler, 1958; Nault, 1998; Pinsonneault and Kraemer; 1997; Robey, 1977). For example, enterprise resource planning (ERP) and workflow systems offer real-time metrics from different functions to central decision makers and thus give them a full picture of what is currently happening in the organization (Bloom et al., 2009; Cardoso et al., 2004; Chari et al., 2008; Hendricks et al., 2007; Hitt et al., 2002). Further; IT can increase the processing power of central decision makers. thus reducing the risk of information overload (Bolton and Dewatripont, 1994; Brynjolfsson et al., 1994; Dewett and [ones, 2001; Galbraith, 1974; Gurbaxani and Whang, 1991; Huber, 1990; Leavitt and Whisler, 1958; Nault, 1998; Pinsonneault and Kraemer, 1997; Robey, 1977). For example, IT systems such as ERP systems, decision support systems, and expert systems support decision makers in analyzing voluminous and complex information (Chari et al., 2008;


2.4 Horizontal fit between IT and IT complements

35

Shim et al., 2002). In addition, IT can facilitate the dissemination of instructions to lower-level employees and their subsequent control by central decision makers (Brynjolfsson et al., 1994; Gurbaxani and Whang, 1991; Pinsonneault and Kraemer, 1997). For example, workflow systems make performance measurement "a matter of examining the system log' (Stohr and Zhao. 2001: 291). However; it is argued that even if vertical information processing is supported by IT, some tacit local information will remain nontransferable and thus not become available to central decision makers. Further; it is argued that even IT is not able to enhance the information processing power of central decision makers without some constraints (Anand and Mendelson, 1997; Brynjolfsson and Mendelson, 1993; Daft and Lengel, 1984; Hitt and Brynjolfsson, 1997). In summary, it is plausible that IT enhances the benefits from centralization. In turn, the benefits from IT increase with the levels of centralization, as increasing levels of centralization exploit IT's information processing capabilities better and better. Thus, IT and centralization may be horizontal complements.

2.4.1.2 Decentralization As outlined above, under decentralized organizational structures, autonomous subunits communicate directly and decide on their activities and coordination without the direction of superior hierarchical Ievels (Aoki, 1986; Galbraith, 1974). However, toplevel units often have more detailed information about firm-wide goals, needs, resources, and problems (Anand and Mendelson, 1997; Roberts, 2007), even when subunits directly communicate and coordinate themselves. Decentralization and IT may be complementary, as IT can enhance the benefits from decentralization through its abovementioned effects on the speed, quantity, and quality of infonnation transfer and processing. IT can help endow lower-level employees in subunits with more information about the firm's overall situation (Anand and Mendelson, 1997; Brynjolfsson, 1994; Brynjolfsson and Mendelson, 1993; Dewett and [ones, 2001; Hitt and Brynjolfsson, 1997; Huber, 1990; Pinsonneault and Kraemer, 1997; Robey, 1977). For example, ERP and workflow systems can offer a real time picture of what is currently happening in the organization not only to central decision makers but also to subunits. Further; IT can help subunits to better cocrdinate their actions with the other subunits on the same hierarchical level (Anand and Mendelson, 1997; Brynjolfsson, 1994; Brynjolfsson and Mendelson, 1993; Dewett and [ones, 2001; Hitt


36

2 Information technology and firm perfonnance

and Brynjolfsson, 1997; Huber, 1990; Pinsonneault and Kraemer, 1997; Robey, 1977). F端r example, groupware applications and corporate intranets enable employees throughout an organization to communicate with each other, share information and

data, coordinate, and work together (Alavi and Leidner; 2001; Bloom et al., 2009; Dewett and [ones, 2001; Ellis et al., 1991; Karre and Alavi, 2007; MOll et al., 2003). Thus, IT enables lower-level employees in subunits to make more and better autonomous

decisions in line with a firm's central goals without the direction of higher hierarchical levels (Brynjolfsson and Mendelson, 1993; Dewett and [ones, 2001; Hitt and

Brynjolfsson, 1997; Huber, 1990). In summary, it is plausible that IT enhances the benefits from decentralization. In turn, the benefits from IT increase with the level of decentralization, as increasing levels of decentralization exploit IT's infonnation processing capabilities better and better. Thus, IT and decentralization may be horizontal complements. This perception is supported by a number of empirical analyses in which IT and decentralization are found to be complementary (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). 2.4.1.3 Arieher definition ofthe degree ofeentralization The degree of centralization is a major mechanism to determine how the power in a firm is distributed (Andersen and [onsson, 2006; Burton and 端bel, 2004; ]asperson et al., 2002). Nonetheless, there are further mechanisms that determine the power distribution in a firm and that may also be complementary with IT (Burton and 端bel, 2004; ]asperson et al., 2002). The few empirical studies on horizontal IT complements have considered not only the allocation of decision rights, but also whether employees with decision rights are supported by specific HRM practices that give them the proper motivation and qualification to make high-quality decisions (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). While this is a first step to analyze the complementarities between the power distribution in a firm and IT, I propose three advancements. First, the allocation of decision rights has to be analyzed in more detail. Second, standardization, as a further structural mechanism that affects the power distribution, has to be considered. Third, the HRM practices that influence the power distribution have to be analyzed in more detail. The existing empirical studies measure the allocation of low-level or operating decisions in the hierarchy, including. for example, the decisions on the pace of work or the method of work of production workers. Further, the allocation of these decisions has been


2.4 Horizontal fit between IT and IT eomplements

37

measured on a rather raw five-point Likert seale from 1 ('exclusively workers') to 5 ('exclusively managers') (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). 端ther authors have proposed more sophisticated notions of the degree of eentralization. Aeeordingly, decision rights eannot only be alloeated upward or downward the hierarchy (that Is, vertical eentralization). Instead, they ean also be eoneentrated in specifie functions or divisions (that Is, horizontal eentralization). Further; specific decision rights may be centralized, while others may be deeentralized (that Is, seleetive eentralization) (Mintzberg, 1979). To aeeount for arieher picture of the degree of centralization, future studies should also measure the alloeation of high-level or strategic decisions. including. for example, the decisions related to the budget, evaluations and rewards, hiring and firing. purehasing supplies and equipment, and establishing programs and projects (Burton and 端bel, 2004). Further, respondents ean be asked to indicate the specific hierarchical level that is responsible for eaeh decision of interest and not only to answer a raw Likert seale. This method allows for a very detailed insight into the alloeation of decision rights. Further; one ean eount the overall number of hierarchical levels in a firm. Thus, it would be possible to ealculate a measure that is easily eomparable aeross fu-ms.s Table 2.1 (next page) provides an example set of questions conceming low-level, intermediate-level, and high-level decisions. This survey instrument was derived and adapted from prior studies on the degree of eentralization (Andersen and Ionsson, 2006; Bloom et al., 2009a; Burton et al., 2002; Colombo et al., 2007; Hanks et al., 1993; Lee and Grover, 2000; Nahm et al., 2003; Pugh, 1973) and pretested as well as used in a telephone survey on 784 German and Polish manufaeturing firms in 2008.6 The assignment of questions to the low-level, intermediate-level, and high-level eategories is derived from a joint exploratcry factor analysis with varimax rotation that clearly showed three distinct factors.

The degree of centralization of a particular decision right would be measured as (hierarchicallevel tha t makes the decision / number ofhierarchicallevels in the firm).

5

6

Further information on t his survey is in Chapters 1 and 4 ofthis thesis.


38

2 Information technology and firm perfonnance

Table 2.1 Example questions on the degree of centralization Hierarchicallevel x ofy hierarchicallevels makes the decision on ... the pace ofwork für production workers. ... the methods ofwork für production workers. ... the allocation ofwork to production workers. ... the daily production plans. ... the weekly production plans. ... overtime in production.

low-level decisions

... meeting problems in the production process. ... the reaction to customer problems and complaints. ... the machines and tools used in production. ... the creation of a new full-time position. hiring or not hiring an applicant für an open position in production. ... hiring or not hiring an applicant für an open position in middle management intermediate-level ... the methods of employee selection. decisions ... the criteria for employee appraisal and bonuses. ... promotions. ... wage increases.

... greater changes in production. ... the introduction of a new product. entering a new market. ... product pricing.

high-level decisions

2.4.2 Standardization Standardization is a further structural mechanism that influences the power distribution in a firm (Andersen and [onsson, 2006; Burton and übel, 2004). However, existing studies on IT and horizontal IT complements have not incorpcrated standardization (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). I define standardization as the standardization of employee behavior by written rules and procedures. that Is, by formalization, and by oral rules and procedures. Further, standardization incorpcrates the latitude of employees in complying with rules and procedures (Burton and übel, 2004). Standardization is often seen as means to achieve a centralized power distribution alternative to the centralization of decision rights (Adler and Bcrys, 1996; Adler et al., 1999; Burton and übel, 2004; Pfeffer and Leblebici, 1977). While the behavior of


2.4 Horizontal fit between IT and IT complements

39

employees is controlled personally through instructions in the case of centralized decision rights, it is controlled impersonally through rules and procedures in the case of standardization (Pfeffer and Leblebici, 1977). This type of standardization has been termed 'coercive' (Adler and Borys, 1996) and may enhance a centralized power distribution. However; standardization may not only be used to exercise command and control over employees. Instead, standardization may be used to codify best practices and quickly disseminate them throughout the firm. Thus, employee activities are facilitated and employees are enabled rather than coerced to work efficiently (Adler and Borys, 1996; Adler et al., 1999; Benner and Tushman, 2003; Jansen et al., 2006; Roberts, 2007; Sheremata, 2000). This type of standardization has been tenned 'enabling' (Adler and Bcrys, 1996) and may enhance a decentralized power distribution. Standardization and IT may be complementary, as IT can enhance the benefits of rules and procedures by facilitating the monitoring of their implementation (Alavi and Leidner, 2001; Bloomsfield and Coombs. 1992; Coombs et al., 1992; Dewett and [ones. 2001; Grant and Higgins, 1991; Huber, 1990; Orlikowski and Robey, 1991; Stohr and Zhao. 2001). For example, ERP systems provide a clear view of the relative perfonnance of different parts ofthe organization (Bloom et al., 2009; Cardoso et al., 2004; Chari et al., 2008; Hendricks et al., 2007; Hitt et al., 2002). Further; through workflow systems, "performance measurement becomes a matter of examining the system log' (Stohr and Zhao. 2001: 291). Thus, IT may be complementary with coercive standardization, which supports a centralized power distribution. However; IT may also facilitate the search for the rule or procedure that applies for a specific issue (Dewett and [ones, 2001). One example is the electronic knowledge pools used by strategic consultancies to recuperate best practices from prior projects and facilitate their dissemination throughout the firm (Hansen et al., 1999). Thus, IT may also be complementary with enabling standardization, which supports a decentralized power distribution. As IT may be complementary with distinct forms of standardization that advance either a centralized power distribution or a decentralized power distribution, the complementarities between standardization and IT have to be taken into account if the complementarities between the power distribution and IT are to be adequately examined.


40

2 Information technology and firm perfonnance

2.4.3 HRM practices

As outlined above, the existing empirical studies on horizontal IT complements have

taken into account HRM practices such as self-managed teams. employee involvement groups, broad job specifications, promotions based on team performance, teambuilding mechanisms, pre-employment screenings. and training measures (Bresnahan et al., 2002; Brynjolfsson and Hitt, 1998b; 2000; Brynjolfsson and Mendelson, 1993;

Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). The studies draw on the idea that multiple HRM practices must be adopted to support decentralized decision-making (Ichniowski and Shaw, 2003; Milgrom and Roberts, 1990; 1995). Particularly, "emplcyees must have the opportunity, the incentive and the skills" for decentralized decision-making (Ichniowski and Shaw,

2003:

160). For example, teamwork,

involvement groups, and broad job specifications give employees the opportunity to make use of decision rights. Performance-based pay or promotions give monetary incentives. while teambuilding mechanisms create group pressure, both of which motivate high-quality decisions. Screening and training produce employees with the skills and knowledge needed for interactions with other employees and for autonomous decisions [Ichniowski and Shaw, 2003). The HRM practices that were considered and the survey instruments used to measure them have been derived from the literature on 'high perfonnance work practices' (Bresnahan et al., 2002; Brynjolfsson and Hitt, 1998b; Ichniowski et al., 1997; Osterman, 1994). However; several HRM studies recently have argued that in the past, high performance work practices have been analyzed on the 'architectural level', which is too abstract to allow conclusions on the fit between HRM and contingency factors such as a

firm's corporate strategy. Accordingly, future studies should consider more detailed levels of analysis, that is. they should measure particular HRM 'policies' and 'practices' that concretize the HRM architecture (Becker and Gerhart. 1996; Becker and Huselid, 1998; 2006). For example, a firm's HRM architecture may determine that employee performance is valued and rewarded in a firm, while it is the HRM policies and practices that define if this is realized through performance appraisals, perfonnance-based pay, or performance-based promctions, depending on what is more appropriate to achieve horizontal and vertical fit (Becker and Gerhart. 1996). For example, perfonnance appraisals could be used to identify potentials for further developing employee skills for autonomous decisions, thus supporting a decentralized power distribution. In contrast; performance-based pay and promotions may provide incentives to correctly implement the instructions of central decision makers, which are seen as supportive of a centralized power distribution (Burton and 端bel, 2004).


2.4 Horizontal fit between IT and IT complements

41

As prior studies on complementarities between IT and a decentralized power distribution measure the HRM architecture of firms, it is unclear if the HRM practices under study were actually suitable to support the decentralization of decision rights. For example, asking for the importance of educational background in pre-employment screenings or for the "percentage of production workers [that] received any workrelated training" (Brynjolfsson and Hitt, 1998b: 50) measures the extent of screening and training but not its contents. Several studies have developed distinct systems of HRM practices that fit different corporate strategies (Delery and Doty, 1996; Jackson et al., 1989; Miles and Snow, 1984; Schuler and Iackson, 1987; Youndt et al., 1996). These studies may provide an indication as to how HRM practices must be measured in order to gain a reliable picture of their effect on the power distribution in a firm. Nonselective screening procedures as well as training efforts that focus on "general information, such as company policies and procedures"

or

at

"correcting skill

deficiencies"

are

suggested

for

low-cost

manufacturing strategies (Youndt et al., 1996: 842). Further; results-oriented performance appraisals that enhance error reduction and process standardization are suggested for cost-reduction strategies (Schuler and [ackson, 1987; Youndt et al., 1996). In contrast, "selective staffing and comprehensive training programs that emphasize attracting and developing individuals with superior technical, problem-solving. and interpersonal skills" are proposed for quality and flexibility manufacturing strategies (Youndt et al., 1996: 843). Further, behavior-oriented perfonnance appraisals that enhance employee development are suggested for quality-enhancement and innovation strategies (Schuler and [ackson, 1987; Youndt et al., 1996). The different contents of these screening. training, and perfonnance appraisal measures would have been impossible to detect by simply measuring the extent of screening and training. Note that the HRM practices defined by Schuler and Jackson (1987) and Youndt et al. (1996) not only vertically fit different corporate strategies, but also horizontally fit different degrees of centralization with respect to decision rights and standardization. To find employees that follow the instructions by central decision makers or by rules and procedures, relatively nonselective screening procedures should be sufficient. Training efforts that convey information on rules and procedures and that correct skill deficiencies further enhance the capability to implement top-level decisions. Resultsoriented perfonnance appraisals provide incentives to correctly implement the instructions.

In

contrast,

autonomous

decisions

by

subunits

and

horizontal

communication between these subunits are enhanced by employees with superior technical, problem-solving. and interpersonal skills. Behavior-oriented perfonnance


42

2 Information technology and firm perfonnance

appraisals are better suited to measure the performance of activities that rely on interaction with other employees. Further, these appraisals can offer feedback f端r further skill and knowledge development.

Thus, to correctly analyze complementarities between a firm's power distribution and IT, a firm's HRM system must be analyzed at the level of HRM pclicies and practices ratherthan at the architectural level as in prior studies.

2.4.4 Information technology I have shown that IT may be complementary with a centralized power distribution as well as with a decentralized power distribution. So far, with the exception of some illustrative examples. IT has been treated as a homogenous technology that can facilitate

both vertical and horizontal infonnation processing. Actually, IT consists of different types ofhardware, software, and interconnecting technologies. Some IT types may indeed facilitate vertical and horizontal information processing alike. For example, the studies on the complementarities between IT and the centralization of decision rights have derived their IT measures from the number of personal computers (PCs) that are in place in a firm (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). PCs allow employees on both lower and higher hierarchical levels to feed in, process, and read out Information. Thus. in a firm with many PCs, vertical as well as horizontal information processing is significantly facilitated. In contrast, when there are only a few PCs in place, a significant amount of vertical or horizontal communication has to take place via alternative media that have disadvantages with regard to the speed, quality, and quantity of communication. However; in many cases it is likely that the IT types that are suitable to support vertical information processing differ from the IT types that are suitable to support horizontal information processing. Several studies on the relationship between IT and firm performance distinguish IT types according to their strategic purpose. For example, IT has been divided into three categories, including 'transactional systems' that cut costs by substituting labor and automating processes, 'informational systems' that cut costs and enhance productivity by providing the infonnation backbone of a firm, and 'strategic systems' that increase innovativeness by enabling new products, services, or the entry into new markets (Aral and WeiH, 2007; Gregor et al., 2006; Mirani and Lederer, 1998; WeiH, 1992). Similarly, Dehning et al. (2003) distinguish 'automate IT' used to replace human labor; 'informate IT' used to provide more Information to management,


2.4 Horizontal fit between IT and IT complements

43

emplcyees, and customers, and 'transform IT', which alters traditional business capabilities, facilitates the acquisition of new capabilities, and ultimately leads to a higher market valuation. Though these studies give an indication of how distinct IT types may be categorized, the first four studies have asked survey respondents to indicate what percentage of their IT investments falls into each of the IT categories (Aral and Weill, 2007; Weill, 1992), and the fifth study by Dehning et al. (2003) categorized public IT investment announcements of firms into the IT categories. Both approaches are subject to the criticism outlined above. These operationalizations do not allow conclusions regarding which IT types are complementary to horizontal and vertical factors and how IT can be aligned with horizontal and vertical factors. In contrast, Oh and Pinsonneault (2007) and Pinsonneault and Kraemer (1997) categorized a pool of actual IT applications into one of their categories and asked survey respondents to indicate which of these IT applications were in operation in their firms. This method is most fruitful for operationalizing an integrative model of IT complements, as it allows detailed conclusions on which IT types fit which horizontal complements. Similar to the abovementioned studies, Oh and Pinsonneault (2007) distinguish IT applications that aim at three different strategic priorities, that Is, at cost reducticn/operational efficiency, quality improvement, or revenue growth. Although an IT classification that classifies IT systems with regard to their vertical fit with corporate strategy gives an indication for a fruitful IT classification, the study by Pinsonneualt and Kraemer (1997) is more in line with my thinking on the horizontal fit between IT and other elements of the organization design. They distinguish IT types according to their effect on information processing. Specifically, they make a differentiation between control-oriented, cocrdination-oriented, and efficiency-oriented IT and classify 539 IT applications used in city governments into these three types. Control-oriented IT facilitates vertical information processing along a firm's hierarchy; that Is, decision makers are enabled to more easily disseminate instructions, supervise and monitor the activities of lower-level employees and their performance outcomes. Cocrdinationoriented IT facilitates horizontal information processing between employees or departments on the same hierarchical level; that is. it enhances communication and decision-making on

a lateral level.

Efficiency-oriented IT increases the labor

productivity of middle management, thus allowing the amount of work of middle managers to be held constant or the number of middle managers to be reduced (Pinsonneault and Kraemer, 1997). Thus, control-oriented IT may be complementary with a centralized power distribution, and cocrdination-oriented IT may be complementary with a decentralized power distribution.


2 Information technology and firm perfonnance

44

While some measures f端r the IT component, such as the number of Pcs, may be sufficient to capture IT's potential to facilitate both vertical and horizontal infonnation processing, classifying different IT applications according to their information processing

effects

should

give

more

detailed

insights

into

the

horizontal

complementarities between IT and other elements of the organization design. Though it

is impossible to define an IT classification that fits any empirical setting. the abovementioned studies can serve as a methodological guide f端r future studies.

2.4.5 Two ideal configurations 01 organization design In summary, I have shown that IT's potential to enhance firm performance by facilitating information processing may be equally well exploited by two alternative organization designs. Further; I have shown that to gain a complete picture of the power distribution in a firm, the degree of centralization and the HRM practices in place should be analyzed in more detail in the future. Additionally, the type and degree of standardization should also be considered. First, horizontal fit can be achieved by aligning IT with organizational elements that support a centralized power distribution. Highly centralized decision rights and a high level of coercive standardization require employees in subunits to transfer information upward the hierarchy and to efficiently implement the instructions received from central decision makers or rules and procedures. These activities are supported by screening. training, and perfonnance appraisal measures that endow employees with knowledge on rules and procedures and that give incentives to implement instructions, rules, and procedures. Control-oriented IT applications may be particularly fruitful for this organizational configuration. Second, horizontal fit can be achieved by aligning IT with organizational elements that support a decentralized power distribution. Decentralized decision rights and low levels of coercive standardization require employees in subunits to communicate directly with each other to coordinate their activities. These activities are supported by enabling standardization, screening. training. and performance appraisal measures that endow employees with diverse knowledge that is needed to understand the perspective of employees from other functions, interpersonal skills that are needed for communication, and incentives to perform in teamwork situations. Cocrdination-oriented IT applications may be particularly fruitful for this organizational configuration.


2.4 Horizontal fit between IT and IT complements

45

Obviously, both configurations are ideal types to achieve horizontal fit. The performance of deviations from these ideal configurations should result in performance losses, while there is no agreement on the performance of virtually hybrid configurations, as outlined above. In summary, I advance two propositions on horizontal IT complements.

Proposition 2.1a. Horizontal fit results from combining (control-oriented) information technology with organizational structures and HRM practices that advance a centralized power distribution. Proposition 2.1b. Horizontal fit results from combining (coordination-oriented) information technology with organizational structures and HRM practices that advance a decentralized power distribution. Although both alternative organization designs are intemally consistent, they have relative advantages and dis advantages resulting from their information processing capacities. As different occurrences of contingency factors such as environment, corporate strategy, and culture require different information processing types and outcomes, the pros and cons of the two ideal crganization designs are major pillars for achieving vertical fit. Organization designs with a centralized power distribution are usually regarded as highly efficient. Employees are incentivized to execute their activities without deviations from instructions. rules, and procedures. Thus, slack resources are reduced to a minimum. Further; vertical infonnation processing avoids any unneeded and costly communication. However, the quality of the decisions by central decision makers can be diminished, either because they do not have all the tacit infonnation on the subtasks or because they are overburdened by the huge amount of information that is transferred to them from below. As outlined already, even IT may not fully solve these problems (Anand and Mendelson, 1997; Aoki, 1986; Bolton and Dewatripont, 1994; Galbraith, 1974; Gurbaxani and Whang, 1991; Leavitt and Whisler, 1958). Further, centralized power distributions are viewed as detrimental to Innovation, as employees are not encouraged to get involved in non-routine activities and also lack the skills and knowledgeto do so (Adler and Borys, 1996; Youndt et al., 1996). Organization designs with a decentralized power distribution are usually regarded as highly flexible and innovative. Employees are incentivized to communicate with each other and cocrdinate their activities themselves, and they are also endowed with the


2 Information technology and firm perfonnance

46

skills and knowledge to do so . This allows employees to gain a more complete picture of the firm and thus to come up with ideas f端r problem-solving as well as concepts f端r new products or new services. Through decision authority, they are able to share these ideas and experiment with them. However, autonomous employees may invest

tOD

much in

imagination and trial and error; that Is, they may disregard efficiency or the firm's overall goals (Adler et al., 1999; Benner and Tushman, 2003; Birkinshaw, 2006; Birkinshaw and Gibson, 2004a; 2004b; 2005; Gibson and Birkinshaw, 2004; Jansen et al., 2005a; Raisch et al., 2009; Roberts, 2007; Sheremata, 2000; Tushman and O'Reilly, 1996). The lacking foeus on efficiency is worsened by the facts that the perfonnance and commitment of autonomous employees is much harder to control [Ichniowski and Shaw, 2003) and that horizontal information processing requires more costly communication than vertical information processing (Bolton and Dewatripont, 1994; Radner; 1993; Wyner and Malone, 1996). The lacking focus on the firm's overall goals is worsened by the fact that even though autonomous subunits directly communicate, they lack the global view that central decision makers have (Anand and Mendelson, 1997; Aoki, 1986; Roberts, 2007).

2.5

Vertical fit between contingencies and organization design

In this section, I analyze how to achieve vertical fit between the two ideal organization designs developed above and three important contingency factors, namely, a firm's environment, its corporate strategy, and its organizational culture (Burton and 端bel, 2004; Miles and Snow, 1978; Mintzberg, 1979). Drawing from contingency thecry, achieving vertical fit

is interpreted as

matehing the information processing

requirements of the contingency factors with the information processing capacities of organization design.

2.5.1 Environment A firm's environment may include areas such as customer groups or the merket, competitcrs, technology, the labor market or uruons, the capital market, legislators or regulators, and the socio-cultural environment (Daft et al., 1988; May et al., 2000). The environment is well established as a contingency factor (Burton and 端bel, 2004; Duncan, 1972; Galbraith, 1973; 1974; 1977; Lawrence and Lorsch, 1967). Additionally, the relationships between environment and IT have been analyzed empirically (Lee and Grover, 2000; Li and Ye, 1999).


2 5 y ertj ca! fit between contjn gencies and organjzation desjgn

47

Based on a Iiterature review, Burton an d 端bel (2004) nam e four dimensions that describe a flrm's environment. Environmenta l comp lexity cc ncerns t he number of environmenta l vari ab les t hat affect a firm [Duncan, 1972), w hile enviro nmenta l uncertainty concem s t he lack of knowledge about the values of these variables, that is, the extent to which these variables can tak e a great ran ge of values a nd cha nge quickly (Lawren ce and Lorsch, 1967). Equivocality refers to a lack ofkno wted ge abo ut variables that affect the firm or about their importance (Daft and Lengel, 1986) ; hostili ty refers to the malevolence of the environment, th at is. th reats to a flrm's perfo nnance or existence (Burton and Obel, 2004). An environment's infonnati on pro cessing req uirements change across a11 four dimensions. Complexity increases t he a mount of issues t hat have to be monitored, assessed, a nd addres sed. Thus, t he amount of informa tion processing increases. Uncertainty increa ses t he freq ue ncy in whic h known Issues have to be monitored, assessed, a nd addressed. Purther, it increases the need to forecast changes in environmenta l issues. Equivocality increa ses the need to sca n t he e nviro nment for yet unknown issues. Hostility increa ses the pro bability that t he releva nce of infonnation changes (Burton and Obel, 2004). Drawing on Burton and Obe l (2004 ), I arg ue that an environment that scores low on many or all dimensions (t hat ls, a n undemanding

environment] vertically fits an organization design with a centralized power distrib ution. In contrast, an environment that score s high on many or all dimensions (th at is, a demanding env ironment) vertically fits an orga nization design w it h a decentralized power distribution. In th e first case, the activities of a firm's subunits are affected by a manageable number of enviro nmental vari ables that change only infrequently and whose changes are pr edictable. Cent ral decision makers are w ell able to gat her and pro cess information on such a reduced number of enviro nmental issues without bein g overwhelmed. Rules and pro cedur es are a beneficial means to coordinate subunits if t he environmenta l conditions rarely change. The inform ation in a well-understood enviro nment is not ext re mely complex a nd does not requi re high med ia rtchness, which re nders of Iittle importa nce t he relat ive disadvan tages of control-ori ented IT systems with regard to inform ation pro cessing a nd tacit infonnatio n (Burton a nd Obel, 2004). Thus, a centra lized orga nization design with its cost-saving vertical infonnation processing is a suita ble means to coordinate specialized subunits under undemanding environmenta l conditio ns. Furt he r, an und emanding environment allows emp loyees to foeus on the efficient execution of their activities without having to think about innova tions that answer environmental changes. Thus, an undeman ding envir onment favors a


2 Information technology and firm perfonnance

48

centralized power distribution's relative advantages with regard to efficiency and does not penalize its relative dis advantages with regards to innovation. In the second case, the activities of a firm's subunits are affected by a large number of environmental variables that change frequently and are poorly predictable. In addition, yet unrecognized environmental issues may occur without notice. Central decision makers may be unable to gather and process the information on such a great number of environmental issues without reductions of decision speed and quality. Rules and procedures are likely to become obsolete quickly and even impede adaptations to environmental changes. In such ademanding environment, information comes in different forms, from different sources, and includes tacit infonnation (Burton and 端bel, 2004). A large number of autonomous employees with an understanding of each other's activities is better able to process such large amounts of complex and tacit information and to quickly react to changing environmental conditions. Cocrdination-oriented IT systems can support lateral communication. Thus. a decentralized organization design with its horizontal infonnation processing is a suitable means to coordinate specialized subunits

under

demanding

environmental

conditions.

Further,

ademanding

environment requires employees to be flexible in answering environmental changes. As such, ademanding environment favors a decentralized power distribution's relative advantages with regard to flexibility and innovation and does not penalize relative disadvantages with regard to efficiency. In summary, I advance two propositions on the vertical fit between the contingency factor of environment and organization design.

Proposition 2.2a. Vertical fit results fram combining an undemanding environment with (control-oriented) information technology, organizational structures, and HRM practices that advance a centralized power distribution. Proposition 2.2b. Vertical fit results fram combining ademanding environment with (coordination-oriented) information technology, organizational structures, and HRM practices that advance a decentralized power distribution.

2.5.2 Corporate strategy A corporate strategy is the specification of the business opportunity that is pursued by a firm and how it is exploited (Grant, 2008; Porter, 1996; Roberts, 2007; Saloner et al., 2001). There are many typologies of corporate strategies, but the most often-used ones


2.5 Vertical fit between contingencies and organization design

49

come from March (1991), Miles and Snow (1978), and Porter (1980). Corporate strategy is well established as a contingency factor that requires a certain organization design (Burton and 端bel, 2004; Saloner et al., 2001). Additionally, the relationship between corporate strategy and IT has been analyzed empirically (DeSarbo et al., 2005; Li and Ye, 1999). The three best-known strategy typologies distinguish similar types of corporate strategies (Spanos et al., 2004). March's 'exploiters', Miles and Snows 'defenders', and Porter's finns with a 'cost leadership strategy' focus on a narrow product market domain. They try to deliver products or services efficiently and to further enhance productivity, Further, they try to defend their market share and achieve an even deeper market penetration. In contrast, March's 'explorers', Miles and Snows 'prospectcrs', and Porter's finns with a 'differentiation' strategy constantly search for new and unique business opportunities by experimenting with novel products and markets. The outlined strategy typologies and many other studies in areas such as organization theory (Benner and Tushman, 2003; Bums and Stalker, 1961; Duncan, 1976; Holmqvist, 2004), managerial economics (Ghemawat and Ricart I Costa, 1993), technology and innovation management (Ambos et al., 2008; He and Wong, 2004), and HRM [lackson et al., 1989; Schuler and [ackson, 1987; Youndt et al., 1996) largely agree on the degree of centralization of decision rights, standardization, and HRM practices required for both types of corporate strategy. According to this, a centralized power distribution is best suited for strategies that focus on efficiency in existing product market domains. As outlined above, through centralized coordination in line with a firm's overall goals, coercive standardization, as well as screening. training. and incentives that support the implementation of instructions, rules, and procedures. a centralized power distribution enhances efficient employee behaviour, The vertical information processing required by this power distribution can be supported by control-oriented IT. The relative disadvantages of centralized power distributions with regard to innovativeness are not penalized by a strategy that focuses on efficiency. In contrast, a decentralized power distribution is best suited for strategies that focus on product and market innovations. As outlined above, through decentralized coordination by subunits, low coercive standardization, high enabling standardization, as well as screening, training, and incentives that support diverse skills and knowledge and interaction between emplcyees, a decentralized power distribution enhances creative


50

2 Information technology and firm perfonnance

and innovative employee behaviour. The horizontal information processing required by this power distribution can be supported by coordination-oriented IT. The relative disadvantages of decentralized power distributions with regard to efficiency are not penalized by a strategy that focuses on innovation. In summary, I advance two propositions on vertical fit between the contingency factor of corporate strategy and organization design.

Proposition 23a. Vertical fit results fram combining a corporate strategy that focuses on existing product market domains with (control-oriented) information technology, organizational structures, and HRM practices that advance a centralized power distribution. Proposition 23b. Vertical fit results from combining a corporate strategy that focuses on new product market domains with (coordination-oriented) information technology, organizational structures, and HRM practices that advance a decentralized power distribution.

2.5.3 Organizational culture A firm's organizational culture is understood as the entirety of the shared values and beliefs of the firm's members. These values and beliefs may partly become observable through artifacts such as language, behavicr, or interaction orders (Schein, 1985). An "organization's culture may be one ofthe factors that determine the nature, extent and form of its informaticn/communication" (Brown and Starkey, 1994: 824) and is thus well established as a contingency factor that demands a certain information processing capacity from the organization design (Burton and 端bel, 2004). Also, it thus may be particularly related to IT (Bloomsfield and Coombs, 1992; Dewett and [ones, 2001; George and King, 1991; Orlikowski, 1992; Orlikowski and Robey, 1991; Robey and Boudreau, 1999). The relations hip between organizational culture and IT has been analyzed empirically (Powell and Dent-Micallef 1997). Two key components of organizational culture that are especially relevant to explain a culture's impact on information processing are the level of trust between employees and the level of conflict between employees with regard to goals and beliefs (Burton and 端bel, 2004; Powell and Dent-Micallef 1997). Both may impact the amount and quality of information that is processed horizontally or vertically in a firm. The two components may become explicit through the degree of fonnality with regard to social relations.


2.5 Vertical fit between contingencies and organization design

51

communication, and behavicr, that is. the type of 'interaction order' used in a firm (Diamond, 1991; Morand, 1995; Powell and Dent-Micallef 1997). For example, relatively informal relationships, communications, and behaviors that are characterized by spontaneity, casualness. and interpersonal familiarity (Morand, 1995) may be a sign of high trust and low conflict between employees. In contrast, relatively formal relationships, communications, and behaviors that have a relatively regimented, deliberate, and impersonal nature (Morand, 1995) may be a sign of low trust and high conflict between employees. Drawing on Morand (1995), I argue that an organizational culture with low trust, high conflict, and thus high formality, fits an organization design with a centralized power distribution. In contrast, an organizational culture with high trust, low conflict, and thus low formality, fits an organization design with a decentralized power distribution. Top-level employees that mistrust other firm members will tend to coordinate the subunits by direct instructions and by the monitoring of the implementation of instructions, rules, and procedures. Further, the centralization of decision rights and a high level of coercive standardization ensure that sufficient information for coordination is processed, even though employees on intermediate and lower hierarchical levels mistrust each other and are in conflict. Control-oriented IT with its relatively low media richness is appropriate for the relatively formal communication up and down the hierarchy that does not incorporate too much tacit Information. Top-level employees that trust other firm members will provide real authority to the subunits (Bloom et al., 2009a); that Is, they will endow them with decision rights. impose low coercive standardization on them, and endow them with the suitable skills, knowledge, and incentives to make high-quality decisions. Further, employees at intermediate and lower hierarchical levels that trust each other and are not in conflict will easily exchange large amounts of explicit and tacit infonnation with each other; even without instructions. rules, and procedures. Coordination-oriented IT is thus suitable to support this richer type information processing. In summary, I advance two propositions on the vertical fit between the contingency factor of organizational culture and organization design. As my aim is to guide future empirical research, I restriet the propositions to one observable outcome of the level of trust and conflict in a firm, that is. the degree of formality.


52

2 Information technology and firm perfonnance

Proposition ZAa. Vertical fit results fram combining an organizational culture that becomes manifest in high formality with (control-oriented) information technology, organizational structures, and HRM practices that advance a centralized power distribution. Proposition 2Ab. Vertical fit results fram an organizational culture that becomes manifest in low formality with (coordination-oriented) information technology, organizational structures, and HRM practices that advance a decentralized power distribution.

2.5.4 Two ideal ways to achieve vertical fit I have shown that IT's potential to enhance organization design and thus firm performance may be equally well exploited by two ideal contingency situations. since these different contextual situations vertically fit the relative advantages and disadvantages of the two different ideal configurations of organization design developed in seetion 2.4. First, vertical fit can be achieved by aligning an IT-supported centralized organization design with an undemanding environment, a corporate strategy that focuses on existing product market domains, and an organizational culture that becomes manifest in high fonnality. Second, vertical fit can be achieved by aligning an IT-supported decentralized organization design with ademanding environment, a corporate strategy that focuses on new product market domains, and an organizational culture that becomes manifest in low fonnality. As outlined above, prior empirical studies that have focused exclusively on vertical IT complements have found that the perfonnance impact of IT is enhanced by dynamic environments (Li and Ye, 1999) and that IT positively mediates the relationship between adynamie environment and an organic organizational structure (Lee and Grover, 2000). Further, studies on corporate strategy as a vertical IT complement found that the performance impact of IT is enhanced by corporate strategies that focus on product development and market expansion (Li and Ye, 1999; Sabherwal and Chan, 2001) and that 'prospector' strategies are associated with the highest IT use (DeSarbo et al., 2005). Finally, one study on organizational culture as a vertical IT complement found that the performance impact of IT is enhanced by an organizational culture with trusting


2.6 Conclusion and implicatjons

53

relationships, minimal conflict, and free communication across units (Powell and DentMicallef 1997). In summary, the findings may indicate that IT may have a higher performance enhancing effect in firms with a decentralized power distribution, a demanding environment, a corporate strategy that focuses on new product market domains, and an organizational culture that becomes manifest in low formality. As outlined above, I argue that the findings of empirical studies that do not incorporate both horizontal and vertical IT complements have to be interpreted with caution. However; an explanation might be that these finns operate with horizontal infonnation processing, which is more demanding with regard to the quantity and quality of information exchange. First, horizontal infonnation processing between subunits requires more communication than vertical information processing along a firm's hierarchy. Second. the communication required for the exchange of creative ideas incorpcrates more tacit information than the communication required to notify central decision makers of one's activities and to receive instructions. Thus, these finns might benefit slightly more from IT's capability to enhance information processing than firms that operate with vertical infonnation processing. Obviously, this does not mean that finns with vertical information processing cannot also enhance their perfonnance through IT.

2.6

Conclusion and implications

In this paper; I have presented a structured review of existing empirical studies on the role of complementarities in IT's firm perfonnance impact and pointed out several possibilities for improvement. First, as a methodological drawback, I have pointed out the partly unsuitable operationalization and measurement of IT and IT complements and made suggestions for advancement. Second, as conceptual drawbacks, I have pointed out the separate examination ofhorizontal and vertical IT complements and the lacking examination of the detailed mechanisms by which IT and IT complements combine. To address these issues, I have drawn from contingency theory and information processing theory to develop an integrative model of IT complements. Accordingly, IT must be simultaneously aligned with horizontal IT complements inside the organization design and vertical IT complements that determine the characteristics of an organization design. Subsequently, drawing from prior theoretical and empirical research, I have provided an analysis of every single previously identified horizontal and vertical IT complement as well as of IT itself Summarizing, I have advanced testable propositions on the complementarities among IT, horizontal complements, and vertical


54

2 Information technology and firm perfonnance

complements. Particularly, I suggest that IT types that support the information exchange

along a firm's hierarchy should be horizontally aligned with the centralization of decision rights, high coercive standardization, and HRM practices that foeus on skills and

knowledge needed to implement rules, procedures, and instructions. This organization design should be vertically aligned with an undemanding environment, a corporate strategy that focuses on existing product market domains, and an organizational culture that becomes manifest in high formality. IT types that support the information exchange

between a firm's subunits should be horizontally aligned with the decentralization of decision rights, low coercive standardization, high enabling standardization, and HRM practices that focus on diverse skills and knowledge needed for communication and coordination with other employees. This organization design should be vertically aligned with ademanding environment, a corporate strategy that focuses on new product market domains, and an organizational culture that becomes manifest in low fonnality. This study adds to the research on IT complements in several ways and provides impetus for further advancements of the research on IT complements. Thereby, the advancement of empirical research is one of the main goals of this study. First, the integrative model ofIT complements clearly suggests that that there is no such thing as a so-called 'best practice' in order to fully exploit IT's potentials for perfonnance increases. That is. there are few IT complements inside or outside a firm's organization design that enhance IT's performance effect in any context. Instead, IT is a general purpose technology that may be equally beneficial to firm performance in combination with different organization designs and contingency factors, provided that the full set of IT, organization design, and contingency factors is consistently aligned. Second, I provide a richer view of the degree of decentralization as a horizontal IT complement. Particularly, I argue that the type and level of standardization has to be considered and that the degree of the centralization of decision rights and the HRM practices in place have to be analyzed on a more detailed level than in the past. Future studies should try to identify further horizontal and vertical IT complements to gain a more complete picture on how IT can be beneficially implemented in a firm. For example, there might be other horizontal IT complements that are not connected with the power distribution in a firm. Third, based on contingency thecry and information processing theory, I provide a consistent theoretical framework to guide future research and thus to facilitate the


2.6 Conclusion and implicatjons

55

accumulation of comparable knowledge and wisdom on IT complements. For example, the information processing view can help to derive hypotheses on established IT complements and to operationalize them. Further, the information processing theory may guide the search for and analysis of yet-disregarded horizontal and vertical IT complements. Fourth, I advance testable propositions on complementarities between IT and IT complements that may directly serve as an agenda for future empirical research. Though I provide propositions on two distinct configurations ofIT, horizontal complements, and vertical complements. future studies should analyze the phenomenon of situational misfit in more detail. It would be particularly interesting to examine if IT enhances firm performance in combination with specific horizontal complements. even if the contingency factors

impose different information processing requirements

on

organization design. Fifth, I provide suggestions on how to operationalize and measure IT, IT complements, and firm perfonnance in a meaningful way. Overall, I think that this study is a promising step toward gaining a more complete picture of how IT must be aligned with other factors both inside and outside a firm's organization design to enhance firm performance.


Chapter 3 Enhancing the performance effecls ofinformation technology through dejcentralization: The roJe of corporate exploration and exploitatlon? 3.1

Introduction

Information technology (IT) has been a central issue in research on firm perfonnance and organization design fĂźr more than 50 years (Zammuto et al., 2007). Early research on the relationship between IT and firm performance uncovered a so-called 'productivity paradox' (Brynjolfsson, 1993; Solow, 1987). Subsequent studies found positive returns to IT (Anders on et al., 2006; Bharadwaj et al., 1999; Dewan and Kraemer, 2000; Mukhopadhyay et al., 1997), but also a substantial variation of IT

payoffs across firms (Brynjolfsson and Hitt, 1995; Loveman, 1994). Further results have led to the gruwing consensus that the returns to IT are dependent on the presence of complementary organizational factors (Aral and WeiH, 2007; Bertschek and Kaiser, 2004; Bresnahan et al., 2002; Brynjolfsson et al., 2002; Sabherwal et al., 2006). Particularly, there is solid evidence that IT payoffs are enhanced by a greater decentralization of organizational decision-making authority (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). I extend this line of research by arguing that the effects of IT on firm perfonnance may also be increased by greater centralization, depending on the type of corporate leaming that is pursued by a firm. I use an information processing view of organizations to argue that IT reduces organizational information processing costs and enhances information processing quality (Anand and Mendelson, 1997; Brynjolfsson and Mendelson, 1993; Galbraith, 1973; 1974; 1977; Tushman and Nadler, 1978). Particularly, IT enhances horizontal information processing, which is the primary mechanism to cocrdinate a firm's subtasks in decentralized organization designs. but IT also enhances vertical information processing, which is the primary mechanism to coordinate centralized organization designs (Dewett and [ones, 2001; Huber, 1990; Pinsonneault and Kraemer, 1997). As IT Not e t hat Chapter 3 is co-authored by Tobias Kretschmer, Institute for Communication Economics, Munich School of Managemen t, LMU Munich, Geschwister-Scholl-Platz 1, 80539 Munich, Germany. However, the literature review and t he hypothesis development, the construction of t he dataset, the analysis ofthe data, and the interpretation ofthe results are primarily attributable to Ferdinand Mahr. 7

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6_3, Š Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


3 Enhancjng the performance effects of Information technology

58

is able to facilitate both decentralization and centralization, it is possible that IT's potential to

enhance

firm

perfonnance can be

exploited by

either greater

decentralization or greater centralization. Indeed, earlier research on the relationship between IT and organization design has supported the view that IT may be associated with either decentralization or centralization (Bolton and Dewatripont, 1994; Dewett and [ones, 2001; Galbraith, 1973; 1977; George and King, 1991; Gurbaxani and Whang, 1991; Huber, 1990; Leavitt and Whisler, 1958; Nault, 1998).

These considerations imply the existence of a third group of factors that determine the direction of the complementarities between IT and de/centralization (Robey, 1977). To address this issue, I refer to the thecry of organizational leaming. which argues that finns must emphasize one of two learning types; namely, exploration or exploitation (Gupta et al., 2006; March, 1991). As corporate-level activities, exploration refers to the permanent search for new products and markets, while exploitation indicates the continuous improvement of existing product market domains (Auh and Mengue, 2005; He and Wong, 2004; Lubatkin et al., 2006). Many scholars have argued that both learning types are best pursued with different organization designs (Ancona et al., 2001; Benner and Tushman, 2003; Brown and Eisenhardt, 1998; Burns and Stalker; 1961; Duncan, 1976; Lewin et al., 1999; Roberts, 2007; Tushman and O'Reilly, 1996). Specifically, an exploratory learning type requires enough employee autonomy for constant experimentation and adaptation, making a decentralized organization design most appropriate. In contrast; an exploitative learning type succeeds by minimizing variation and maximizing efficiency and control, thus making a centralized organization design the most beneficial solution (Benner and Tushman, 2003). In summary, IT's potential for perfonnance increases through the reduction of information processing costs can be exploited by either greater decentralization or greater centralization. Meanwhile, the benefits of decentralization versus centralization are determined by a firm's corporate leaming type. Hence, I argue that it is the corporate learning type that defines the complementarities between IT and de/centralization in a particular firm. Specifically, I hypothesize that under corporate exploration, IT payoffs are

enhanced through

greater decentralization. In contrast, under corporate

exploitation, returns to IT are improved by greater centralization. I employ a novel multi-source panel dataset on computer capital, organization design, and corporate learning type for almost 260 German manufacturing firms to test the hypotheses. The results support prior findings that IT positively affects firm performance and that the effect is moderated by the degree of de/centralization.


3.1 Introduction

59

However; in support of the hypotheses, I find no universally valid relationship regarding the complementarities between IT and de/centralization. Instead, decentralization is complementary to IT in firms with an exploratcry leaming type. In contrast, centralization is complementary to IT in firms with an exploitative learning type. A variety of robustness checks demonstrates that the qualitative nature of these results is not sensitive to alternative specifications and measures. This study contributes to several research fields. First, it contributes to the IT payoff literature through further opening the black box containing the organizational factors that increase a firm's ability to generate positive returns from IT. Specifically, rather than testing for complementarities, I identify a contingency factor that affects the complementarities. Second, to the best of my knowledge, I present the first large-sample empirical study in recent times that finds evidence for complementarities between IT and centralization, thus contributing to the long-standing IT and de/centralization debate. Third, this paper makes contributions to the ernerging literatures on how IT can benefit different types of organizational leaming (Kane and Alavi, 2007) and product market strategies (Chari et al., 2008). This research yields important implications for managers. The insight that positive IT payoffs rely on organizational complements is supported. More importantly, the findings imply that there is no such thing as a 'best practice' regarding how to fully exploit IT's potentials through organizational complements. Instead, the findings emphasize the general-purpose properties of IT, that Is, IT's ability to positively affect firm performance in combination with different crganizational complements (Bresnahan and Trajtenberg, 1995; Brynjolfsson and Hitt, 2000; David, 1990; David and Wright, 1999). I thus offer some guidance to managers on how to successfully implement IT in conjunction with organizational complements that fit a firm's product market context. The remainder of this paper is organized as follows. In the next section, 3.2, I develop the hypotheses in the context of relevant previous research. In section 3.3, I describe the research design, data and variables. The results of the analysis are provided in section 3.4. I conclude by discussing the implications of this study for managers and researchers in section 3.5.


60

3.2

3 Enhancjng the performance effects of information technology

Literature review and hypothesis development

3.2.1 Information technology and firm perlormance Much of the early research on the relationship between IT and productivity found no, or negative, relationships at the macro-economic and sector levels; these findings have been summarized as the 'computer productivity paradox' (Brynjolfsson, 1993; Rcach, 1987; Solow, 1987). Beginning in the 1990s, studies with improved data and methodologies found positive returns from IT at the macro-economic level and firmlevel (Anderson et al., 2006; Bharadwaj et al., 1999; Brynjolfsson and Yang, 1996; Dewan and Kraemer, 2000; Mukhopadhyay et al., 1997). As firm-level IT payoffs vary substantially across firms (Brynjolfsson and Hitt, 1995; Loveman, 1994), the ernerging consensus is that returns to IT investments are contingent on the presence of complementary factors (Aral and Weill, 2007; Bertschek and Kaiser, 2004; Bharadwaj, 2000; Chari et al., 2008; Clemons and Row, 1991; Mata et al., 1995; Powell and Dent-Micalleff 1997; Sabherwal et al., 2006; Weill and Aral, 2006). Particularly, there is solid evidence for the existence of complementarities between IT and the decentralization of organizational decision-making authority (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). According to these studies, the contribution of IT to firm performance is increased by complementary investments in the dispersion of decision rights as well as in human resource management (HRM) practices that endow employees that have received more decision rights with the necessary skills and incentives to make high-quality decisions (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). These studies are the first large-sample empirical analyses in recent times that contribute to a long-standing controversy, namely, is IT associated with increased decentralization or with greater centralization (George and King, 1991; Robey, 1977)?

3.2.2 The information processing view of organizations and information technology Organizations have been described as information processors (Galbraith, 1973; 1974; 1977; Hayek, 1945; Tushman and Nadler, 1978). Organization design determines who coordinates a firm's specialized and interdependent subunits by defining how


3.2 Ljterature review and hvpothesis developmen t

61

information required for the coordination of the subunits is processed [Calbraith. 1974; Tushman and Nadler, 1978). Coordination can be attained through two alternative organization designs. First, centralized coordination is achieved through vertical information processing. Information from t he subunits is referred upward in a hierarchy to a centrat decision maker who processes t he information, makas decisions, disseminates instructions downward to the subunits, and monitors their impIementation [see Figure 3.1, left). Second, decentralized coordination can be attained t hrough horizontal Information processing. Autonomous subunits communicate directly and decide on t heir activities and coordination without the direction of superior levels [see Figure 3.1, right) (Aoki, 1986; Calbraith 1974).

Figure 3.1 Centralized coordination (left) and decentralized coordination (right) Both organization designs have relative strengths and weaknesses. centrat decision makers typically lack some of t he hard-to-transfar local information on subtasks that require 'a feeI', 'judgement', or 'experience', for example, regarding customer requirements (Anand and Mendelson, 1997; Aoki, 1986; Gurbaxani and Whang. 1991; Leavitt and Whisler, 1958). Further, under strong centralization, central decision makers ran suffer from information overload, which reduces decision speed and quality (Bolton and Dewatripont, 1994; Galbraith, 1974; Leavitt and Whisler, 1958). In contrast, autonomous subunits lack a global view of all subtasks that reduces their ability to coordinate themselves in compliance with the firm's global task (Aoki, 1986). Further, as decentralized coordination requires more communication t ha n centralized coordination, it is more costly (Bolton and Dewatripont, 1994; Radner, 1993; Wyner and Malone, 1996). IT mcreases t he speed, quantity, and quality of informa tion t hat ran be t ransfer red vertically and horizontally between organizational units (Anand and Mendelson, 1997; Dewett and Iones. 2001; Huber, 1990). IT also enhances t he information processing capabilities of organizational units with regard to speed, quantity, and quality


62

3 Enhancing the performance effects of Information technology

(Brynjolfsson and Hitt, 2000; Brynjolfsson and Mendelson, 1993; Galbraith, 1974;

Leavitt and Whisler, 1958). Consequently, IT is interdependent with organization design, as the latter determines how information is processed in a firm (Anand and Mendelson, 1997; Brynjolfsson and Hitt, 2000; Brynjolfsson and Mendelson, 1993).

3.2.3 Information technology, decentralization, and centralization As a consequence, IT facilitates decentralized organization designs. It reduces the costs of the horizontal infonnation flow between subunits and endows them with additional information from above, f端r example, through groupware or database software. Increased processing power, f端r example, through group decision systems, enables subunits to handle the greater amount and complexity of incoming horizontal and vertical Information. Thus, the subunits obtain a more global view, resulting in better decisions that are based on local and global infonnation (Anand and Mendelson, 1997; Brynjolfsson, 1994; Brynjolfsson and Mendelson, 1993; Dewett and [ones, 2001; Hitt and Brynjolfsson, 1997; Huber, 1990; Pinsonneault and Kraemer; 1997; Robey, 1977). IT also facilitates centralized organization designs. It allows the movement of historically hard-to-transfer local infonnation from subunits upward to central decision makers, for example, through executive information systems. Higher processing power allows central decision makers to handle this additional information, for example, through decision support or expert systems (Bolton and Dewatripont, 1994; Brynjolfsson et al., 1994; Dewett and [ones, 2001; Galbraith, 1974; Gurbaxani and Whang, 1991; Huber, 1990; Leavitt and Whisler, 1958; Nault, 1998; Pinsonneault and Kraemer; 1997; Robey, 1977). In addition, IT enhances the dissemination of instructions and their subsequent control (Brynjolfsson et al., 1994; Gurbaxani and Whang, 1991; Pinsonneault and Kraemer; 1997). Some authors argue that IT is more beneficial to decentralization because it would only partly reduce the relative disadvantages of centralized organization designs. Some residual tacit local infonnation may stay nontransferable and not become available to central decision makers. At the same time, the local information that is made transferable and referred upward would overburden central decision makers, as IT would only moderately increase human infonnation processing power (Anand and Mendelson, 1997; Brynjolfsson and Mendelson, 1993; Daft and Lengel, 1984; Hitt and Brynjolfsson, 1997). In opposition to this, recent research maintains that humans are able to 'compensatcry adapt' to the differences between face-to-face and electronic communication (Kock, 2004). These studies emphasize that it is not the medium that


63

3.2 Literature review and bypotbesis development

determines tbe 'richness' of communication but ratber tbe social process surrounding its use. Even electronic media allow rieb communication, including tbe transfer of tacit knowledge (Lee, 1994; Markus, 1994; Ngwenyama and Lee, 1997). Further; present-day IT systems sucb as enterprise resource planning, decision support and expert systems sufficiently support decision makers in analyzing voluminous and complex infonnation (Cbari et al., 2008; Sbim et al., 2002). In summary, I present arguments tbat IT can enbance tbe value of decentralized and centralized organization designs alike by facilitating information processing. At tbe same time, botb greater decentralization and greater centralization require more infonnation processing. Thus, IT's potential for increases in perfonnance can be exploited by eitber greater decentralization or greater centralization. If botb combinations are equally fruitful ceteris paribus, it is plausible tbat bitberto disregarded tbird factors determine if IT is associated witb decentralization or witb centralization in a particular firm (George and King, 1991; Huber, 1990; Robey, 1977). Hence, I believe tbat tbe seminal studies tbat indicate tbat IT payoffs are enbanced tbrougb decentralization sbow only one fruitful combination of IT and organization design (Bresnaban et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). Taking into account contingencies yields a more nuanced picture of tbe complementarities between IT and organization design.

3.2.4 The role olorganizationallearning One

crucial

contingency

factor

tbat

can

detennine

the

direction

of

tbe

complementarities between IT and de/centralization is tbe corporate learning type tbat is pursued by a firm. Marcb (1991) proposes tbat finns bave to make cboices between two fundamentally different types of organizational leaming, namely, exploration and exploitation. As corporate-level activities, exploration refers to the permanent searcb for new products and markets, wbile exploitation relates to tbe ongoing advancement of existing product market domains (Aub and Mengue, 2005; He and Wong, 2004; Lubatkin et al., 2006). Tbe distinction between exploration and exploitation is made explicitly or implicitly in a variety of management researcb areas, including strategie management (He and Wong, 2004; Uotila et al., 2009), organization theory (Holmqvist, 2004), managerial economics (Gbemawat and Ricart I Costa, 1993), tecbnology and innovation management (Benner and Tushman, 2003), and marketing (Kyriakopoulos and Moorman, 2004).


64

3 Enhancing the performance effects of Information technology

It is argued that exploration and exploitation require fundamentally different organization designs because of the distinct activities associated with each of them

(Ancona et al., 2001; Benner and Tushman, 2003; Brown and Eisenhardt, 1998; Bums and Stalker, 1961; Duncan, 1976; Lewin et al., 1999; Roberts, 2007; Tushman and

O'Reilly, 1996). Exploration refers to notions such as adaptability, flexibility, variation, and radical innovation, while exploitation implies alignment, efficiency, refinement; and incremental innovation (Adler et al., 1999; Gibson and Birkinshaw, 2004; March, 1991; Roberts, 2007; Tushman and O'Reilly, 1996). To achieve corporate-level exploration, firms must allow specialized subunits on lower hierarchical levels to constantly look for new opportunities and generate ideas. Subsequently, to be able to innovate, subunits must experiment with ideas in direct communication with other subunits and without delay caused by seeking information or permission from above. Many of these activities are highly judgement-related and creative and thus involve hard-to-transfer local Information. Obviously, central decision makers lack both the tacit local information and the processing power and speed to coordinate such activities in a meaningful way. Further, centralization would restriet the creativity of employees that is crucial for exploration. Thus, decentralized coordination is the superior solution to connect these activities in a way that benefits a firm's overall goal, that Is, the exploration ofnew products and markets (Benner and Tushman, 2003; Galbraith, 1974; Roberts, 2007; Stieglitz and Heine, 2007). Exploration fundamentally depends on both organizational slack, that Is, the time and resources required for imagination and experimentation, and the acceptance of failures. In contrast, corporate-level exploitation requires subunits on lower hierarchical levels to focus on the efficient execution of and incremental improvement in the existing agenda. Thus, timely and efficient execution must be strictly controlled, trial and error must be eliminated as far as possible. Further rationalization requires innovation, but the search for incremental improvements is restricted to the current agenda so that even the development and implementation of process innovations can be regularized and disciplined. It is clear that decentralized subunits lack the overall perspective to achieve such an efficient combination of all subtasks. Further, horizontal information processing may incur excessive costs. Consequently, centralized coordination is the best solution to achieve corporate exploitation (Benner and Tushman, 2003; Roberts, 2007; Stieglitz and Heine, 2007). The notion that firms with an exploratory leaming type require a decentralized organization design, while firms with an exploitative learning type require a centralized


3.2 Literature review and bypotbesis development

65

organization design, is supported by extensive researcb (Ancona et al., 2001; Burns and Stalker, 1961; Duncan, 1976; Lewin et al., 1999; Miles and Snow, 1978; Porter, 1985). Thus, firms witb an exploratory learning type tbat requires decentralization generate performance gains from IT if tbey use IT to strengtben decentralization. In turn, firms with an exploitative leaming type tbat requires centralization unlock IT's full potential for performance gains ifIT is used to enbance centralization.

3.2.5 Hypotheses I bave drawn insigbts from tbe information processing view of organizations as well as tbe IT and de/centralization debate. I bave developed tbe argument tbat tbrougb its capability to facilitate borizontal and vertical information processing, IT enbances botb greater decentralization and greater centralization. I bave furtber noted tbat IT's potential to enbance information processing and tbus firm performance is in turn only exploited if eitber decentralization or centralization is actually enbanced. I bave used organizational leaming tbeory to argue tbat different corporate learning types require different organization designs. Exploration demands decentralization, wbile exploitation calls for centralization. As a consequence, I bave noted tbat it is a

firm's corporate learning type tbat determines tbe direction of tbe complementarities between IT and de/centralization. Corporate exploration requires decentralization. Thus, under corporate exploration, IT's potentials for perfonnance gains tbrougb tbe facilitation of information processing are fully unlocked tbrougb greater levels of decentralization.

Hypothesis 3.1: Under corporate exploration, the impact of information technology on firm performance will be greater for firms with greater levels of decentralization. In contrast, corporate exploitation requires centralization. Thus, under corporate exploitation, IT's potentials for performance gains tbrougb tbe facilitation of infonnation processing are fuHy unlocked tbrougb greater levels of centralization.

Hypothesis 3.2: Under corporate exploitation, the impact of information technology on firm performance will be greater for firms with greater levels of centralization.


66

3.3

3 Enhancjng the performance effects of information technology

Empirical approach

3.3.1 Empirical models Consistent with prior studies on the returns to IT and on the returns to combinations of IT and organization design, I employ the economic theory of production to test the hypotheses (Bresnahan et al., 2002; Brynjolfsson and Hitt, 1996; Dewan et al., 1998; Dewan et al., 2007). I adopt a standard Cobb-Douglas production function, specified in log-linear form as shown in model 3.1 below. Here, for firm j at time t, VA = value added is explained by IT = IT capital, C = non-IT capital, L = labor, and DEC = the degree of decentralization.

Fcllowing

existing

studies,

implement

the

notion

of

complementarities between IT and the degree of decentralization by means of a moderation effects model, that Is, an interaction term ofIT and DEC (Bloom et al., 2009b; Bresnahan et al., 2002; Brynjolfsson et al., 2002; Chari et al., 2008; Colombo et al., 2007; Crespi and Criscuolo. 2006).8 X represents a number of control variables that are described below, and

E

is the error term that is clustered by firm: that Is, it is Huber-

White robust to heteroskedasticity and autocorrelation of unknown form.

(3.1)

The slope coefficients

ßl, ßa, ß3, ß4, and ßs represent the output elasticities with respect

to the inputs IT capital, non-IT capital, labor, decentralization, and the product of IT and decentralization, respectively, that Is, the percent change in output for a one percent change in the quantity of the input when all other inputs are held constant (Brynjolfsson and Hitt, 1995; 1996].9 The hypotheses concern the interaction effect, that Is, the joint effect of IT and decentralization on firm perfonnance. If

ßs

is significantly greater than zero. one can

a A criticism of this approach is t hat there possibly exist unobserved third factors, such as managerial decisions, which could be another explanation for the covariation of IT, DEC, and VA as an alternative to complementarities (Athey and Stern, 2003). Hence, t he evidence presented here must "be considered suggestive conditional correlations ratherthan causal" (Bloom et al., 2009b: 17). A differen t specification could use lagged input factors to account for the possibility that IT requires a period of adjustment to fully affect firm performance. I do not consider lags in the main analysis because capital already includes the effects of prior spending. Moreover, including lagged inputs significantly reduces the sampie size (Brynjolfsson and Hitt, 1995). Nonetheless, I use lagged inputs in one of the robustness tests. 9


3.3 Empirical approach

67

conclude that the effect of IT on VAdepends on the level of DEC, that Is, that IT and DEC are complementary. To test if the complementarities between IT and decentralization are contingent on a

firm's corporate leaming type, I use a common approach in contingency research, subgroup analysis (Venkatraman, 1989). Specifically, I analyze whether the regression coefficient for the interaction term ITxDEC for the group of explorers is different from the corresponding regression coefficient for the group of exploiters. Thus. in regression model 3.2, I interact ITxDEC with the dummy variable XPLOIT. XPLOIT indicates if a firm pursues an exploitative leaming type (XPLOIT=l) or an exploratory leaming type (XPLOIT=O). In addition to the three-way interaction (ITxDEC)xXPLOIT,this model must include the three linear terms IT, DEC, and XPLOIT and the three possible two-way interaction terms composed of these variables, that Is, ITxDEC, ITxXPLOIT, and DECxXPLOIT (Aiken and West, 1991; Jaccard and Turrisi, 2003).

InVAjt = ßo+ ß1lnITjt +ß)nC jt + ß31nLjt +ß41nDECjt +ßsOnI'Ij1x InDEC jt ) + ß6XPLOI'Ijt + ß7 OnI'Ij, xXPLOI'Ij,)+ ß B (lnDECj, xXPLOI'Ij,) + ß9 ((lnI'Ij, x InDEC j,) xXPLOI'Ij,) +

(3.2)

Lx;, +E j,

In this model, ßs indicates the direction and strength of the complementarities between IT and DEC for explcrers, that Is, when XPLOIT equals

o. Explorers are denoted as the

reference group in model 3.2. As I hypothesize decentralization to be a complement to IT under corporate exploration, I expect ßs to be positive and significant in model 3.2. The coefficient for

ßg

equals the slope difference for the regression of ITxDEC onto VA

between exploiters (XPLOIT=l) and explorers (XPLOIT=O). The significance test associated with

ßg is thus a test of the significance of the slope difference.

Model 3.2 thus also yields infonnation about the direction and strength of the complementarities between IT and DEC for exploiters, which can be obtained by summing

ßs

and

ßg

[laccard and Turrisi, 2003; Wooldridge, 2009). A simple way to

cross-check this and obtain information about the significance of ITxDEC for exploiters is to rescore the dummy variable so that an exploratory leaming type equals 1 and an exploitative leaming type equals O. I call this reverse coded dummy variable XPLORE in the remainder of this paper. In the resulting model 3.3, ßs indicates the direction and strength of the complementarities between IT and DEC for exploiters and equals

ßg),

(ßs +

as in model 3.2. I hypothesize centralization to be a complement to IT under

exploitation. By definition, centralization is the inverse of decentralization, and so I expect the coefficient ßs for ITxDEC to be negative and significant in regression model


3 Enhancjng the performance effects of Information technology

68

ßg in model 3.2 equals the ßg in model 3.3, but it is opposite in sign [laccard and

3.3. Obviously, the coefficient für the slope difference

coefficient für the slope difference Turrisi,2003).

InVA jt = ßo + ß1lnITjt +ßzlnC jt + ß)nL jt

+ß4 1nDECjt

+ßs(lnI~t x InDECj,) + ß6 XPLOREjt

+ ß7 (Ion;, xXPLORE j, ) + ßB(lnDEC;, xXPLORE j, ) + ß9(OnITj, x!nDECj, ) xXPLORE1,)+

(3.3)

D;, + Ei'

3.3.2 Data To minimize common method bias. the dataset on German manufacturing finns was

constructed from three independent sources (Podsakoff et al., 2003), namely, (1) a panel of IT capital and other firm characteristics over the period 200D-200S, (2) a panel of further production function inputs and output over the period 2000-2008, and (3) a 2008 cross-section of organization designs and corporate learning types. I obtained measures of IT capital and labor from Harte-Hanks' CI Technology Database (CITDB), a source that has been used in several prior studies on the performance effects of IT (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Dewan et al., 2007; Hitt and Brynjolfsson, 1997).10 Harte-Hanks, a market intelligence company, conducts annual telephone surveys to take stock of specific IT types that are used by the establishments of more than 10,000 German firms. As data are provided on the establishment-levelt t, I aggregated to the firm-level or, if required, extrapolated firm-level measures. HarteHanks states that the finns in the CITDB are largely representative of the entire population of German firms. Where possible, I derived the remaining quantitative data for the production function estimations from Bureau Van Dijk's ORBIS database, which provides balance sheets and profit and loss statements from annual company reports. I also used ORBIS to assign every firm to an industry. The German firms that maintain at least one manufacturing establishment and are covered by both the CITDB and ORBIS formed the pool of 600 potential respondents to a survey on the degree of de/centralization, HRM practices, and corporate learning type. I focused on the manufacturing industry to avoid issues with interpreting the output of 10

This was previously known as the Computer Intelligence InfoCorp (CII) database.

Harte-Hanks interviews at least one establishment of each firm in the CITDB. The survey questions concern the IT types tha t are used by the establishment and not the IT types tha t are used by the entire firm.

11


3.3 Empirical approach

69

service firms in production function estimations and also to be able to use a single questionnaire. I surveyed production managers (or similar managers) because they are typically in upper-middle management and thus are able to deliver insights into both the degree of de/centralization on lower hierarchical levels and the more high-level issue of corporate learning type. Using a narrow set of potential informants, I further held single

Informant bias relatively constant. The surveyed establishments employed on average 70.0 percent of a firm's total employees, indicating that the obtained measures are largely representative of the firms. Six student research analysts conducted 259 telephone interviews within 16 working days in March 2008, representing a response rate of 43.2 percent. I note that respondents and non-respondents do not significantly differ in tenns of firm emplcyees, value added, total assets, and the number of personal computers (P端s). It has been argued that organizational structures and product market strategies are subject to inertial forces and tend to persist over time (Hannan and Freeman, 1984; Miller and Friesen, 1985). I thus assurne that organization design and corporate leaming type have only marginally changed in the years before the survey, in contrast to outputs, IT, and the remaining inputs. The different adjustment speeds make it possible to match 2000-2008 panels on output, IT and other inputs with a snapshot of crganization design and learning type taken at the end of the panel period, which is consistent with other studies (Bresnahan et al., 2002; Brynjolfsson et al., 2002). In addition, in the telephone survey, the respondents were asked about the firm's corporate learning type during the five years before the survey. While this should largely mitigate problems with regard to learning type, the potential for problems with regard to crganization design is reduced by the fact that an alternative interpretation of the organization design measure does not substantively affect the interpretation of the results. That Is, instead of interpreting the measure as an indicator of the organization design during the years before the survey, it can be interpreted as an indicatcr of the direction of the organizational change that has taken place during the years before the survey. As the complementarities between IT and organization design can be examined in both causal directions (Bresnahan et al., 2002), this does not change the interpretation of the results. For example, the finding that greater IT Investments in 2006, 2007, 2008 and greater decentralization in 2008 are associated with greater firm performance in 2006, 2007, and 2008 would indicate that IT has induced greater decentralization over time, which in turn allowed the firm to increasingly exploit IT's potential for performance increases. After removing observations with missing data, I am left with a final sample of 182 ftrms, with each firm represented by at least one of 572 observations that are used in the


70

3 Enhancjng the performance effects of Information technology

analysis. The finns do not significantly differ from all other German manufacturing firms included in the CITDB and ORBIS with regard to the number of firm emplcyees, value added, total assets, and the number of P端s.

3.3.3 Variables 3.33.1 Production function inputs and outputs Following other studies, the IT capital (IT) f端r every firm was computed by multiplying the number of pes by the average value of a Pf (Brynjolfsson and Hitt, 1995; 1996). The number of pes f端r each firm was obtained from the eITDE. As data are reported on the

establishment-level, I aggregated to the firm-level. Where the eITDB does not cover every establishment of a firm, I extrapolated finn-level values. To do so, I calculated the number of PCs per employee for the surveyed establishments and multiplied this figure by the number of firm employees. On average, the establishments covered by the CITDB employed 82.3 percent of firm emplcyees. indicating that the measure is largely representative of the whole firms. Further, by obtaining the average percentage shares of desktop PCs and notebook PCs in the final sample, I was able to compute the number of desktop PCs and notebook PCs for every firm. t? Annual average prices for desktop PCs and notebook PCs were obtained from a market report by Gfl( Group and other market research companies. Specifically, I obtained the average prices for every year from 2005 to 2008 and extrapolated the prices fcr, 2000 to 2004.13 IT capital was then computed as IT = (number of desktop PCs x average desktop PC price] + (number of notebook PCs x average notebook PC price]. Like many other theoretical and empirical studies, I abstract from the fact that IT is not a homogenous technology (Bertschek and Kaiser,

2004; Bresnahan et al., 2002;

Brynjolfsson, 1994; Dewan and Kraemer, 2000; Dewan et al., 2007; Nault, 1998). The IT measure does not include other IT types like peripherals, interconnecting technologies. and software. Further, I do not distinguish IT types according to their abilities to facilitate information processing. For example, Pinsonneault and Kraemer (1997) classify IT applications into control-oriented systems that support vertical information The percen tage shares of desktop PCs and notebook PCs are reported for only 52.1 percent of the observations in the final sampie. For these observations, the mean percen tage share of desktop PCs is 88.0 percent The number of desktop PCs was thus computed as number of desktop PCs = (0.88 x number of PCs), while the number of notebook PCs was computed as number of notebook PCs = (0.12 x number of PCs). 13 Note that t he results are not sensitive to using annual average prices from other sources, such as publicly-available press releases ofmarket research companies.

12


3.3 Empirical approach

71

processing and coordination-oriented systems that enhance horizontal information processing. Kane and Alavi (2007) identify IT types that are better suited for exchanging explicit information and IT types that are more suitable for sharing tacit Information. Nonetheless. I believe that the IT capital measure is sufficient for the purpose of this study. To test the hypotheses, I must proxy the extent to which the IT at hand in every firm has the potential to facilitate both vertical and horizontal information processing. The number of PCs, which is the basis of the IT capital measure, is indeed a satisfactory measure. PCs allow employees on both lower and higher hierarchical levels to feed in, process, and read out Information. Thus, in a firm with many PCs, vertical as well as horizontal information processing is significantly facilitated. In contrast, when there are only a few PCs in place, a significant amount of vertical or horizontal information processing has to take place via alternative media that have disadvantages with regard to the speed, quality, and quantity of information processing. Value added (VA) was calculated as sales minus material costs. Sales and material costs were obtained directly from ORBIS. Non-IT capital (C) was computed as tangible fixed assets minus IT capital. Tangible fixed assets were obtained directly from ORBIS, while IT capital was computed as described above. Labor (L) is measured as the number of emplcyees, which was taken directly from the eITDB.

To allow inter-year comparisons, the nominal values of IT capital, sales, material costs, and tangible fixed assets were converted to constant Euros with 2000 as the base year. As deflators, I used the Producer Price Indices (PPIs) for each three-digit Standard Industrial Classification (SIC) industry, obtained from the Statistical Office of the European Communities (Eurostat). Where PPIs were not available for an industry in a given year, I used the PPI for the entire manufacturing sector.t-

3.3.3.2 Organizationallearning and organization design The group dummies that indicate if a firm pursues corporate exploration (XPLORE) or corporate exploitation (XPLOIT) are derived from a discrete measure of corporate learning (LEARN), which includes exploration and exploitation as its extremes. This reflects March's (1991) interpretation that firms must choose between exploration and 14

Note tha t the results are not sensitive t o using current Euros instead of constan t Euros.


72

3 Enhancing the performance effects of Information technology

exploitation as two ends of a continuum (Gupta et al., 2006; March, 1991). LEARN is the average of four standardized items that are distantly related to items from other studies that interpret exploration and exploitation as corporate-level activities (Auh and

Mengue, 2005; He and Wong, 2004; Lubatkin et al., 2006). Specifically, production managers were asked to indicate on ascale from 1 to 5 if during the last five years their establishment [item 1) and their firm as a whole [item 2) have concentrated on the

improvement of existing products and processes (score 1) or on the introduction of new products (score 5). Further, they were asked to indicate if during the last five years their establishment (item 3) and their firm as a whole (item 4) have concentrated on existing customer groups and markets and their needs (score 1) or on the development ofnew customer groups and markets (score 5). The survey Instrument can be found in the Appendix 3.1. With regard to the considerations to control for bias resulting from combining a cross-section with panels, note again that it was asked for corporate leaming during the last five years. The median of LEARN is used to split up the sample into two groups of nearly equal sizes and to construct the dummy variables XPLORE and XPLOIT.1 5 Overall, 49.7 percent of the observations in the final sample represent firms with an exploratcry leaming type, while the remaining observations stern from firms with an exploitative leaming model. The measure of the degree of decentralization (DEC) reflects the notion developed in prior studies that the decentralization of decision rights must be accompanied by HRM practices that endow autonomous employees with the necessary skills and incentives to make high-quality decisions (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). Thus, DEC is composed of a variable that measures the decentralization of decision rights (DR) together with a measure of suitable HRM practices (HR).16 DR is the average of six standardized items that have largely been derived from prior studies (Bresnahan et al., 2002; Brynjolfsson and Hitt, 1998b; Brynjolfsson et al., 2002; Colombo et al., 2007; Hitt and Brynjolfsson, 1997). In particular, production managers were asked to indicate on a scale from 1 to 5 if the following decisions were made by managers exclusively (score 1), managers mostly (score 2), managers and workers (score 3), workers mostly (score 4) or workers exclusively (score 5): The decision (1) on the delivery time and priority of crders, (2) on the production plans, (3) on the

XPLORE is coded 1 if LEARN<::median(LEARN) and 0, otherwise. Conversely, XPLOIT is coded 1 if LEARN<median(LEARN) and 0, otherwise. 16 Decentralization (DEC) was computed as DEC = (std(DR) + std(HR)) / 2, where std(x) = std(x - 'X) / 0x. 15


3.3 Empirical approach

73

distribution ofwork among production workers, (4) on how exactly work is done (for example, with regards to pace and order), (5) on which machines and tools are used, and (6) on the cocrdination between different stages ofthe production process. The survey instrument can be found in the Appendix 3.2. HR is the average of six standardized items that share similarities with those used in prior studies on HRM practices (Bae and Lawler, 2000; Bresnahan et al., 2002; Brynjolfsson and Hitt; 1998b; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997; Huselid, 1995; Ichniowski et al., 1997; Osterman, 1994; Sun et al., 2007). Endowing workers with decision rights is only fruitful if their daily work allows for a certain degree of autonomy. Production managers were thus asked about two HRM practices that allow some discretion, namely, (1) the use of (self-managed) teamwork and (2) the degree of variety in the daily werk, The other HRM practices in which I was interested should endow workers with the proper information, qualification, and motivation to make fruitful use of decision rights. Thus, I asked about (3) the amount and quality of information given to workers, (4) the amount and contents of training given to them, (5) the use ofpay components that are based on individual performance, and (6) the use of promotion criteria that are coupled with the appropriation of capabilities to make better decisions. While the items for LEARN and DR were measured by means of classic Likert scale items that are rated by the respondents, I employed an innovative survey method that was introduced by Bloom and Van Reenen (2007) to measure the items for HR The respondents were asked open questions about every HRM practice and answered freely. Their open answers were then scored on a scale from 1 to 5 by the interviewers, while the respondents were unaware that their answers were being scored.t? This method provides richer insights into HRM practices than is possible with rather short Likert scale items. Further; the method mitigates the problem of social desirability, which is highly important in the context of HRM practices. For example, some respondents might give unrealistically optimistic answers to questions about the amount of information or training given to workers if they feel that these HRM practices are generally considered 'good'. The interviewers were provided with prepared questions to ask, They began with general questions and continued with more specific questions based on the prior The Human Subjects Committee of Stanford University approved t his method for t he study of Bloom and Van Reenen (2007), since lack of awareness among respondents that they are being scored is necessary t o reduce the problem of social desirability. Moreover, this is not risky for the respondents and t heir employers due t o data confidentiality, and it is t emporary, as the respondents are debriefed afterthe end ofthe projec t (Bloom and Van Reenen, 2007). Note that all respondents in the present study receive a debriefing package. 17


74

3 Enhancing the performance effects of Information technology

answers of respondents. The interviewers were also encouraged to deviate from the

prepared questions if needed. Thus, a conversation led by the interviewer developed regarding every HRM practice. The dialogue was not ended until the interviewer felt that she or he had a full picture of the HRM practice in question and was able to decide on a score from 1 to 5. To facilitate the secring. the interviewers were provided with example answers f端r the scores of 1, 3, and 5. The scores 2 and 4 were used f端r intermediate cases. The survey instrument can be found in the Appendix 3.3. The interviews lasted on average 44.4 minutes. which allowed f端r in-depth insights into HRM practices. As this method strongly relies on interviewer capabilities, several steps were undertaken to ensure a high interview quality. The six student interviewers came from management, economics, and sociology and were chosen using a two-step assessment procedure from almost 80 applicants. Selection criteria included prior experience with the topics of the survey and data gathering experience in general as well as performance in a simulated interview situation. The

interviewers were intensively trained in the survey

background, method, questions. and software during a two-day workshop that also included several simulated interviews. Interview quality was further enhanced through regular team meetings and through interview monitoring by two PhD students. These supervisors listened to 21.2 percent of all Interviews and assigned scores to respondent answers independently from the interviewers.t" In the first week of the project, supervisors and student interviewers discussed possible differences between the interviewer and the supervisor scorings to enhance interview quality. Additionally, the supervisor secrings serve as a test of inter-rater reliability. The partial correlation coefficient (after controlling for interviewer and supervisor fixed effects) of the interviewer and the supervisor secrings for the HR variable is as high as 0.90 and highly significant. This implies that this survey method leads to relatively homogenous results even when different interviewers score the same interview. Corporate leaming is measured at the establishment-level and finn-level. However, I follow other studies and measure the degree of decentralization relating to only nonmanagerial production workers to avoid aggregation problems (Bresnahan et al., 2002; Brynjolfsson and Hitt; 1998b; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997; Osterman, 1994). As this group of employees is large, important, and relatively homogenous in manufacturing firms, I am confident that DECto a large extent captures the overall degree of decentralization in the surveyed firms (Hitt and Brynjolfsson,

1997).

ia

In t he multivariate analysis below, only interviewer scorings are used.


75

3.3 Empirical approach

The alpha reliability of DR is acceptable (a=0.67). Tests for the intemal consistency of LEARN, HR, and DEC are not appropriate, as these measures are designed to capture different aspects of a construct rather than multiple measures of a construct.t" Thus, there are no specific expectations about the correlations between these aspects. They might correlate positively, negatively, or

not at all

(Aral and Weill,

2007;

Diamantopoulus et al., 2008; [arvis et al., 2003).

3.3.3.3 Contral variables To control for industry effects on firm performance, I include industry dummies derived from three-digit SIC core codes in all estimations. Where a firm's core code in ORBISwas not a manufacturing code, I replaced it by the manufacturing code derived from the CITDB.20 The firms in the final sample come from 67 different three-digit SIC code industries and 19 different two-digit SIC code industries. I include a dummy variable to account for the fact that ORBIS contains unconsolidated and consolidated accounting Information. Overall, 79.4 percent of the observations in the final sample are from unconsolidated accounts. As described above, I had to extrapolate the number of PCs in those cases in which not all firm establishments are covered by the CITDB. To control for this fact, I include the percentage of firm employees that is covered by the CITDBas a control variable. I further mitigate the potential for any bias due to combining a cross-section with panels by including a dummy variable that indicates if a firm's owner has changed during the five years prior to the survey. Changes in ownership can be interpreted as a major reason for organizational changes. A takeover took place in firms accounting for 15.2 percent of the observations in the final sample. Note that the procedure of distinguishing between leaming types by using the median of LEARN defines those firms that are relatively dose to the median of LEARN and thus pursue a relatively balanced leaming type as members of one of the two leaming type 19 The different aspects are as folIows: Corporate learning (LEARN) consists ofthe (1) product domain and the (2) market domain. HRM practices t hat support the decentralization of decision rights (HR) must endow workers with (1) opportunities for discretion, (2) information, (3) training, and (4) incentives. The degree of decentralization (DEC) consists of t he (1) decentralization of decision rights (DR) and (2) supportive HRMpractices (HR). 20 lt was reassured in t he course of the t elephone survey tha t all firms in t he sampie have at least one manufacturing establishment. However, it is possible t hat firms were not assigned a manufacturing core code in ORBIS because manufacturing is not the core business of a firm or because of a wrong assignment. Note thatthe results shown below are similar ifthe unadjusted core codes from ORBIS are used.


76

3 Enhancing the performance effects of Information technology

groups. To control f端r this fact, I include a dummy indicating all firms that lie between the 40th and the 60th percentile ofLEARN in all estimations. Overall, 16.4 percent ofthe observations in the final sample represent firms with a relatively balanced leaming type. To further mitigate the potential of bias from the median split procedure, I conduct another robustness test as described below. I include a number of indicators f端r the quality of the survey conducted f端r this study as control variables. To control f端r the representativeness of the survey, I include the percentage of firm employees that are employed in the surveyed establishment as a control variable. Interviewer dummies are included to control for interviewer fixed effects. Apart from potential influences on the respondents and thus respondent-rated Likert scales. systematic differences between interviewers may especially lead to systematically better or worse scorings with regard to the interviewer-rated items related to HRM practices. Possible respondent effects. such as those based on viewpoints related to a respondent's distinct position in a firm's hierarchy or her Ihis different amounts of knowledge on a firm due to her Ihis specific length of tenure, are controlled for by taking into account each respondent's relative position in the firm's hierarchy as well as the respondent's tenure. To control for general interview quality, interview duration is included as a control variable in all analyses. Finally, all estimations include year dummies and dummy variables that indicate if items in the scales LEARN, DR, and HRM are missing. In the case of missing values, I averaged over the remaining items. To increase interpretability, and to avoid multicollinearity problems that may occur in regression models with interaction terms, I mean-centered IT and DEC before the computation ofthe interaction tenn ITxDEC (Aiken and West, 1991; Jaccard and Turrisi, 2003). Descriptive statistics are provided in Table 3.1 (next page).


77

3.4 Results Table 3.1 Descriptive statistics Descrip tion

Variable

Obs.

Mean

Std. dev.

Min.

M",.

Value added"

In(VA)

572

10.6809

1.2 185

7.9590

15.5569

In(IT)

572

-0.0226

1.3530

-5.5412

4.7921

Capital '"

In(C)

572

9.3222

1.7385

3.5207

15.5851

Labor

In(L)

572

6.2577

1.1194

4.0943

10.4341

DEe

572

-0.000 1

0.7804

-2.3698

2.2097

In(IT)xDEC

572

0.3345

1.0671

-2.7914

5.8763

In(DEG)

572

2.2254

0.922 7

0

4.3820

572

0.8231

0.2665

0.0132

1

IT capital (mean centered)

0

0

Decentralization (mean centered)

#

IT capital-Dec. % employees

with degree

#

% of employees covered bythe CITDB

0

% of employees covered

572

0.6995

0.3021

0.0163

1

572

15.9476

10.6932

1

46

Relative hierarchicallevel of respondent (coun ted from bottom of hierarchy) #

572

0.7093

0.1201

0.3333

1

Interview du ration (minu tes)

572

44.4423

9.0956

26

78

by author's survey 0/ # Tenure of respondent (years)

#

#

Source: '" ORIBS, 0 CITDB, # authors' survey. Value added, IT capital, and non-IT capital are in thousand Euros. Ln (VA), In(IT), In(C), In(L), and In(DEG) are in naturallogarithms.

3.4

Results

3.4.1 Discussion ofthe results The results of the crdinary least square (OLS) estimations are provided in Table 3.2 (next page). In the basic production function reported in column 1, t he contribution of non-IT capital and labor to value added is positive and significant, as expected. I further estimate that the elasticity of value added for IT capital is 0.1683 and significant when other inputs are held constant. Note that R2 is relatively high (0.90), which is common to production function estimations and complies with production thecry, according to which the output is fully explained by inputs (Bresnahan et al., 2002; Brynjolfsson and Hitt, 199 6; Dewan et al., 2007). These basic findings hold across all other estimations in Table 3.2 and Table 3.3.


(I) (2) (3) (4) (5) (6) (7) (8) (9) (10) Dependentvariable In(VA) In(VA) In(VA) In{VA) In{VA) In{VA) In(VA) In{VA) In{VA) In(VA) Robustness test Chow test Ernployee skills Decision rights only Reference group Explorers Exploiters Explorers Exploiters Explorers Exploiters Explorers Exploiters In{lT) 0.1683*** 0.1475*** 0.1538*** 0.1682*** 0.1722*** 0.1426** 0.1545*** 0.1696*** 0.1549*** 0.1798*** ITcapital (0.0371J (0.0328J (0.0473J (0.0412J (0.0487J (0.0663J (0.0475J (0.04 10J (0.0479J (0.0435J In{e) 0.2321*** 0.2308*** 0.2299*** 0.2299*** 0.2742*** 0.2777*** 0.2307*** 0.2307*** 0.2486*** 0.2486*** Non-IT capit al (0.0387J (0.0374J (0.0379J (0.0379J (0.0575J (0.0728J (0.038 1J (0.038 1J (0.0378J (0.0378J In{L) 0.5337*** 0.5829*** 0.5670*** 0.5670*** 0.4504*** 0.462 1*** 0.5661*** 0.566 1*** 0.5570*** 0.5570*** Labor (0.0746J (0.0673J (0.0658J (0.0658J (0.0867J (0.1204J (0.0658J (0.0658J (0.0650J (0.0650J DEC 0.0538 0.0808 0.0285 0.0536 0.1571 0.0836 0.0299 0.2306*** 0.1356 Decentralization (0.0469) (0.0643) (0.0608) (0.0702) (0.0991) (0.0659) (0.0632) (0.0820) (0.0877) In{lT)xDEC -0.0255 0.0955*** -0.1262*** 0.1432*** -0.1722*** 0.0949*** -0.1265*** 0.1093* -0.1924*** (0.0385) (0.0355) (0.0463) (0.0338) (0.0460) (0.0355) (0.0461) (0.0596) (0.0668) Complementaritybetween Ir and decentralization [or reierence qroup -0.0459 -0.0479 -0.0775 1.3285 XPLOIT (0.0729) (1.2345) (0.0716) (0.0742) Exploiter dummy -1.3285 0.0459 0.0479 XPLORE 0.0775 Explorer dummy (0.0729J (1.2345J (0.0716J (0.0742J -0.0296 0.015 1 0.0249 In{lT)xXPLOIT 0.0 145 (0.0823J (0.0503J (0.0499J Difference to exploiters (0.0504J -0.0249 In{lT)xXPLORE -0.0145 -0.0151 0.0296 (0.0823J (0.0499J Difference to explorers (0.0504J (0.0503J -0.0536 -0.0950 DECxXPLOIT -0.0523 0.1035 (0.0794) (0.1038) Difference to exploiters (0.0803) (0.12 15) DECxXPLORE 0.0523 -0.1035 0.0950 0.0536 Difference to explorers (0.0803J (0.12 15J (0.1038J (0.0794J -0.3017*** [ln{lT)xDEC]xXPLOIT -0.2217*** -0.3154*** -0.2214*** (0.0604) (0.0571) (0.0606) Difference to exploiters (0.0919) 0.3017*** [ln{lT)xDEC]xXPLORE 0.2217*** 0.3154*** 0.2214*** Difference to explorers (0.0604) (0.0571) (0.0606) (0.0919) In{DEG) -0.0072 -0.0072 % ot emotoveeswith dearee rO.0406l rO.0406l Observations 572 572 572 572 572 572 572 572 572 572 Firrns 182 182 182 182 182 182 182 182 182 182 R2 0.90 0.91 0.92 0.92 0.9 4 0.94 0.92 0.92 0.92 0.92 Ccntrcts: Columns (1) t o (4) include year dummies, industry dummies, % employees covered by t he CITDB, unconsolidated accounts dummy, % employees covered by author's survey. missing item dummies, interviewer dummies, interview duration, respondent's tenure, respondent's relative hierarchical position, recent change of ownership, dummy and balanced learning type dummy. Additionally, column (5) includes interaction terms of In(C), In(L), and all abovementioned controls with XPLOIT. Additionally, column (6) includes inter action t er ms ofln(C), In(L), and all abovementioned controls with XPLORE. Notas. * significan t at 10%; ** significan t at 5%; *** significant at 1%. The ti me period is 2000-2008. The estimation method in all columns is OLS. The standard errors in brackets are clustered by ftrm: t hat is, t hey are Huber-White robust t o heteroskedasticity and autocorrelation of unknown form.

Table 3.2 Main estimations and first robustness tests


3.4 Results

79

Model 3.1 is estimated in column 2. The interaction term ITxDEC is insignificant, implying that the impact of IT on firm performance is not enhanced by greater levels of decentralization for the full sample of 182 firms. Although not formally hypothesized, this finding complies with my theoretical considerations. I hypothesize that the impact of IT on firm performance is enhanced through greater levels of decentralization under corporate exploration, though not for all finns. Nonetheless, the finding appears to contradict studies that found complementarities between IT and decentralization for all finns and without taking into account contingency factors (Bresnahan et al., 2002; Brynjolfsson et al., 2002; Hitt and Brynjolfsson, 1997). An explanation may be differences in the composition of the firms in the samples of these studies and my sample. Specifically, the samples of the earlier studies cover the period from the late 1980s to the mid 1990s. At that time, firms with greater IT investments represented early adopters of IT, as IT was not yet widely used. It is intuitive to assume that these finns were also pioneers in fields other than IT, for example, with regards to the adoption of modern management practices such as the decentralization of decision rights and supportive HRM practices. Thus, the combination of greater IT investments and greater decentralization may have been associated with higher firm perfonnance for two reasons. First, these finns may have had above-average firm perfonnance due to first-mover advantages in the adoption of these technological and organizational innovations. Second, it is likely that those firms with an exploratory leaming type were most likely among the early adopters of both IT and decentralization, as it is very characteristic of explorers to innovations.

Both

alternatives

quickly adopt technological and organizational would

have

resulted

in

strong

estimated

complementarities between IT and decentralization. In the first case, the positive correlation of the IT-decentralization combination with firm performance would have resulted more from first-mover advantages then from actual complementarities between IT and decentralization. In the second case, actual complementarities between IT and decentralization under corporate exploration would have been measured, even though the learningtype was not observed. In the time period from 2000 to 2008, which is covered by this study's sample, IT is much more widespread among firms- t, leading to a greater variety of combinations ofIT, organization design, and firm perfonnance. Thus. positive correlations of high firm performance with specific IT-crganization design combinations should actually indicate complementarities between the elements of these combinations instead of first-mover advantages. Further; because of the greater variety of combinations of IT, organization design, and firm performance, complementarities The average number of Pcs per employee is 0.53 in this study's sampie and 0.33 in Bresnahan et al. (2002).

21


80

3 Enhancing the performance effects of Information technology

can only be identified when taking into account third factors that determine the actual type of complementarities f端r specific subgroups of firms. The results f端r model 3.2, which tests hypothesis 3.1, are presented in column 3. As described above, the interaction tenn ITxDEC tests f端r complementarities between IT and decentralization under corporate exploration, which is the reference group. As ITxDEC is positive and highly significant, the hypothesis is supported that under corporate explcration, the impact of IT on firm performance increases with greater levels of decentralization. The coefficient f端r the three-way interaction term ITxDECxXPLOITrepresents the slope difference for the regression ofITxDEC onto VA between exploiters and explorers. The complementarities between IT and decentralization under corporate exploitation can thus be calculated as the sum of the coefficients for ITxDEC and ITxDECxXPLOIT, that Is, as 0.0955-0.2217=-0.1262. As described above, an easy way to cross-check this and obtain information on the significance of the coefficient -0.1262 is to change the reference group. The results of this cross-check are reported in column 4. As expected, the coefficient for ITxDEC is indeed -0.1262. Note further that the coefficient is highly significant. This finding supports hypothesis 3.2. Under corporate exploitation, the impact of IT on value added decreases with greater levels of decentralization. As the measure of decentralization is an inverse measure of centralization, this implies that the impact of IT on firm performance is greater with greater levels of centralization under corporate exploitation. Note that the coefficients for both ITxDECxXPLOITin column 3 and ITxDECxXPLORE in column 4 are highly significant. Thus, the complementarities between IT and de/centralization are significantly different under corporate exploration and corporate exploitation. As expected, the coefficients for ITxDECxXPLOIT and ITxDECxXPLORE are identical in size but opposite in sign. Further note that the coefficients of all variables that are not allowed to vary between explorers and exploiters (that Is, non-IT capital and labor) are identical in columns 3 and 4, as expected.

3.4.2 Robustness ofthe results Lexamine the robustness of the results to a variety of alternative specifications and measures.


3.4 Results

81

In columns 5 and 6, I employ an alternative specification that is sometimes termed the 'Chow test', Here, not only the variable of primary interest ITxDEC is interacted with the leaming type dummies. Instead, every independent variable, including all control variables, is interacted with either XPLOIT or XPLORE, thus allowing the entire regression function to vary between both corporate leaming types (Chow 1960; Greene 2008; Wooldridge, 2009). Thereby, I mitigate the potential bias from 'fcrcing' ITxDEC to explain more inter-group variation than it actually does. Note that both specifications, namely, the one employed in columns 3 and 4 and the one in columns 5 and 6, are superior to estimating separate regression equations for the group of explorers and for the group of exploiters. Both specifications employed here formally test for the significance of slope differences and include available information from both groups when estimating the error tenns for the significance tests (Auer, 2005; Jaccard and Turrisi, 2003).22 Note that under the Chow test specification in columns 5 and 6, the results presented above hold with regard to

direction and strength for the

complementarities between IT and de/centralization under corporate exploration and exploitation as well as forthe slope differences. A graphical representation of the results for corporate exploration in column 5 is provided at the top ofFigure 3.2 (next page).

22 Note that - other than t he error terms - the coefficients obtained from the Chowtest specification equal t he coefficients that would result from the estimation of two separate regression equations for explorers and exploiters (Auer, 2005).


3 Enhancing the performance effects of Information technology

82

u

~

~ '<t

--------------_ ....

~--......-~-~••••- -....

~..M.-M......-

..·..··....- ..··..---····----

----------------------

N

-5

o

5

information lechnology (In) low decenlralizalion

..._ ....._ .. medium decenlralizalion

high decenlralizalion

~s------------------,

/

/

,'" .....~.....~.

_::.::"~~

/

/

-

/

..

-

N'-----~-----~---------,J -5

5

- - - - - high cenlralizalion

...__..._- medium cenlralizalion

- - - low cenlralizalion

Figure 3.2 Information technology, de/centralization, and firm performance under corporate exploration (top) and corporate exploitation (bottom) The positive and significant coefficient für the linear tenn IT in column 5 indicates that greater IT investments have a positive effect on value added, even für the mean value of decentralization.P This effect is represented by the dotted at the top of Figure 3.2. 23

Note thatthe IT measure has been mean centered.


3.4 Results

83

However; the positive effect of greater IT investments on value added is enhanced by greater levels of decentralization, as indicated by the positive and significant coefficient for ITxDEC in column 5 and as reflected by the solid line. Further, the dashed line shows that for minimal levels of decentralization, that Is, for maximal levels of centralization, the effect of greater IT investments on value added is negative. This can be explained by the fact that in this case, IT facilitates centralization, which is an unfavorable organization design under corporate exploration. The bottom of Figure 3.2, which represents the results for corporate exploitation in column 6, can be interpreted analogously. Theoretical and empirical research on so-called 'skill-biased technical change' argues that new technologies are complementary to higher-skilled workers (Acemoglu, 1998; Autor et al., 1998; Bartel and Lichtenberg. 1987; Berman et al., 1998). Additionally, other studies find complementarities between organizational change and skills (Caroli and Van Reenen, 2001). To mitigate the potential that the results are biased by unobserved correlations between IT, organization design, and skills, I control for the proportion of employees in the surveyed establishments that have eamed a degree. The results are presented in columns 7 and 8. All the known results hold in direction and strength. In columns 9 and 10, I measure the decentralization of organizational decision-making authority by only using the measure of the decentralization of decision rights (DR) without taking into account HRM practices. The significant coefficients for ITxDEC in both columns as well as for the slope differences reassure me that the results are largely independent of alternative measures of decentralization. Further, the loss of significance of the coefficient for ITxDEC in column 9 is not unexpected. Under corporate exploration, the impact of IT on firm performance should be enhanced by greater levels of decentralization of decision rights. Nevertheless, decentralizing decision rights without

introducing supportive

HRM practices

should lead to

less

'actual'

decentralization than doing both. Thus, the sole dispersion of decision rights should not exploit IT's potential for firm performance increases as well as actual decentralization that includes HRM practices. I provide further robustness tests in Table 3.3 (next page). Note that none of the estimations in Table 3.3 uses the full sample of 572 observations either because ofthe specification or due to missing data.


422 173 0.92

-0.2449*** (0.0675)

-0.0309 (0.09 11)

0.0048 (O.0542J

-0.0400 (0.0794)

rn

0.2449*** (0.0675) 422 173 0.92

0.0309 (O.0911J

-0.0048 (O.0542J

0.0400 (O.0794J

m

In(VA) In(VA) Years 2003-2008 only Explorers Exploiters 0.1630*** 0.1678*** (O.0561J (O.0499J 0.2087*** 0.2087*** (O.0436J (O.0436J 0.5838*** 0.5838*** (O.0768J (O.0768J 0.0667 0.0358 (0.0712) (0.0666) 0.1177** -0.1272*** (0.0473) (0.0439)

423 166 0.92

-0.1861 *** (0.0699)

-0.052 1 (0.08 71)

-0.0439 (O.0595J

-0.0343 (0.0798)

Ă&#x;)

0.1861*** (0.0699) 423 166 0.92

0.0521 (O.0871J

0.0439 (O.0595J

0.0343 0.0798J

~

In(VA) n-n In(VA) (t+1) Lagged independent variables Explorers Exploiters 0.1929*** 0.1490*** (O.0630J (O.0486J 0.2447*** 0.2447*** (O.0395J (O.0395J 0.52 12*** 0.52 12*** (O.0773J (O.0773J 0.0993 0.0472 (0.0678) (0.0657) 0.0830* -0.1031* (0.0446) (0.0529)

297 91 0.92

-0.3167*** (0.0906)

-0.1365 (0.1088)

-0.0602 (O.0816J

-0.1332 (0.0945)

~

0.3167*** (0.0906) 297 91 0.92

0.1365 (O.1088J

0.0602 (O.0816J

0.1332 (O.0945J

~

In(VA) In(VA) Groups by asÂť & 75th percentiles ofLEARN Explorers Exploiters 0. 1708** 0.1107 (O.0716J (0.0707] 0.2862*** 0.2862*** (O.0526J (O.0526J 0.4782*** 0.4782*** (O.0937J (0.0937] 0.0729 -0.0636 (0.1003) (0.0827) 0.1678*** -0.1489* (0.0439) (0.0863)

Notas. * significan t at 10%; ** significan t at 5%; ***significant at 1%. The time period is 2003-2008 in columns (1) and (2) 2000-2008 in columns (3) to (6). The estimation method in all columns is OLS.The standard errors in brackets are clustered by firm; thatis, t hey are Huber-White robust t o heteroskedasticity and autocorrelation of unkn ow n form.

Ccntrcts: Year dummies, industry dummies, % employees covered by the C!TDB, unconsolidated accounts dummy, % employees covered by author's survey. missing item dummies, interviewer dummies, interview duration, respondent's te nure, respondent's relative hierarchical position, recent change of ownership dummy, and balanced learning type dummy.

Dependentvariable Robustness test Reference arcup In(lT) Information technolcoy In{e) Non-IT capit al In(L) Labor DEC Decentralization In(lT)xDEC Complementaritybetween Ir and decentralization for reierenceqroup XPLOIT Exploiter dummy XPLORE Explorer dummy In{lT)xXPLOIT Difference to exploiters In(lT)xXPLORE Difference to explorers DECxXPLOIT Difference to exploiters DECxXPLORE Difference to explorers [ln(lT)xDEC]xXPLOIT Difference to exploiters [ln{lT)xDEC]xXPLORE tntterenceto exotorers Observations Firms R2

Table 3.3 Further robustness tests


3.5 Conclusions

85

In columns 1 and 2 ofTable 3.3, I use the observations from 2003 to 2008 only. As this is the time period that is covered by the survey questions on corporate leamingo. I largely exclude the potential for bias from combining a cross-section on learning type with panels on IT and firm performance. The known results hold in direction and largely in strength. In columns 3 and 4, I control for the possibility that investments in IT need a period of adjustment to take effect on firm perfonnance. All independent variables of models 3.2 and 3.3 that are not obtained from my survey are instrumented by their one-year lags. The results hold in direction. The significance of the coefficients for ITxDEC under corporate exploration and exploitation are reduced, while the difference between both is still highly significant. In columns 5 and 6, I exclude finns with a relatively balanced learning type to avoid any bias from defining firms with only a slight focus on exploration (or exploitation] as explorers (or exploiters]. To do so, I do not split the sample using the median of LEARN. Instead, only the finns with values equal to or greater than the 75th percentile of LEARN are defined as explorers. The firms with values equal to or below the 25th percentile are defined as exploiters. All other finns are dropped. Note that the known results hold in direction and largely in significance. Only the significance of the coefficient for ITxDEC in column 6 is reduced. In summary, hypotheses 3.1 and 3.2 are supported by a variety of different specifications and measures. I find that the impact of IT on firm performance is greater for firms with a greater level of decentralization under corporate exploration. In contrast, I find that the impact of IT on firm performance is greater for firms with greater levels of centralization under corporate exploitation.

3.5

Conclusions

In this chapter; I have examined the performance payoff of IT in the context of organizational complements to IT. In particular, I examined complementarities between IT and the degree of de/centralization and whether these complementarities are contingent on third factors. The analysis was guided by an information processing view of organizations, the IT and de/centralization debate, and theories of organizational 24

As outlined, I asked for corporate Iearning duringthe five years before 2008.


86

3 Enhancing the performance effects of Information technology

learning. Particularly, I argued that IT can facilitate both horizontal and vertical information processing. Thus, IT's potentials f端r performance increases can be exploited

by either greater decentralization, which requires greater horizontal information processing, or by greater centralization, which requires greater vertical infonnation processing. As both alternatives are feasible, I argued that third factors determine which of the two alternatives is appropriate. Particularly, I demonstrated that a firm's corporate learning type determines if a decentralized or centralized organization design is appropriate. I hypothesized that the impact of IT on firm performance is enhanced through greater decentralization for firms that pursue an exploratory leaming type. In contrast, I argued that the impact ofIT on firm performance is enhanced through greater centralization for firms that pursue an exploitative leaming type. I examined a novel panel dataset that was constructed from three independent sources and covers 2000 to 2008. The results support the hypotheses and are robust to a variety of different specifications and measures. Nonetheless. this research is not without limitations. First, the measure of IT is rather narrow and does not consider all IT types, such as interconnecting technologies. peripherals, and software. Although I am confident that this IT measure is sufficient to test the hypotheses, this is a limitation that could be addressed by future research. Second, I needed to combine panel data on IT and firm performance with a cross-section on organization design and corporate learning. Even though I undertook several measures to mitigate the potential for bias from this, further research should use timeseries data on organization design and contingency factors. Finally, while testing for complementarities using a moderation effects model is a standard method, some studies have suggested alternative approaches (Burton et al., 2002; Cassiman and Veugelers, 2006; Miravete and Pemias, 2006; Mohnen and Roller, 2005). Future studies should employ such alternatives to ensure that complementarities in fact explain the correlations between IT, organization design, and firm performance. From the many further implications for research and practice, I highlight three. First, this paper indirectly contributes to the emerging literatures on how IT can benefit different types of organizationallearning (Kane and Alavi, 2007) and product market strategies (Chari et al., 2008). Specifically, the results showthat IT may not only directly impact organizational learning and product market strategies. Instead, future research could examine if IT's influence on organizationallearning and product market strategies is moderated by the organizational arrangements that are in place.


3.5 Conclusions

87

Second, to the best of my knowledge, I have presented the first large-sample empirical study using recent data that finds evidence for an association between IT and centralization, thus contributing to the long-standing IT and de/centralization debate. A particularly interesting avenue for future research in the context of this controversy is to examine if different software applications are associated with decentralization and centralization. For example, groupware applications might better serve horizontal information processing and thus decentralization, while decision support systems might better serve vertical infonnation processing and thus centralization. Finally, I believe that the most important implications for management research and practice stern from the general finding that the complementarities between IT and organization design may strongly depend on contingency factors. Future research should try to identify further contingencies that affect the complementarities between IT and organization design rather than simply test for the existence of complementarities. Subsequent studies should examine if also other complements to IT in addition to organization design are affected by third factors. I further believe that the consideration of contingency factors makes the concept of complementarities more relevant to practitioners. My findings clearly suggest that there is no such thing as a 'best practice' with regard to how to fully explcit IT's potentials for performance gains. Instead, IT is a general-purpose technology that will only benefit firm performance if carefully aligned with both organization design and the contingency factors that determine the appropriate organization design.


Chapter4 Hybrid strategy and firm performance: The moderating role of individual and technological ambidexterity 4.1

Introduction

A central concern of strategic management involves the composition of corporate strategy. Strategy scholars suggest that firms must choose their strategy mix from two distinct types of strategic activities, namely, exploration and exploitation. Exploration is targeted at the development of new and unique product market domains, while exploitation is targeted at efficiency and improvements in existing product market domains (March, 1991; Miles and Snow, 1978; Porter, 1980; 1985). Much discussion has been dedicated to whether the pursuit of a hybrid strategy that combines exploration and exploitation is beneficial or detrimental to a firm's performance in general as well as relative to pure strategies that focus exclusively on exploration or exploitation. Some scholars argue that hybrid strategies negativelyaffect firm performance because of the insurmountable organizational tensions that arise from the simultaneous pursuit of exploration and exploitation (Porter, 1980; 1985; 1996). Others argue that these organizational tensions are solvable, and that hybrid strategies have a positive performance effect equal to (Miles and Snow, 1978) or greater than that of pure strategies (Tushman and O'Reilly, 1996), provided that suitable organizational arrangements are in place. Particularly, the literature on ambidextrous organizations suggests several organizational arrangements aimed at managing the tensions resulting from simultaneously pursuing exploration and exploitation. The empirical findings on the relations hip between hybrid strategy and perfonnance are diverse and range from a positive effect (Spanos et al., 2004; Uotila et al., 2009) to no effect (Kyriakopoulos and Moorman, 2004) to a negative effect (Thomhill and White, 2007). However, these results must be interpreted with caution because studies have largely neglected the role of organizational architecture in the relationship between hybrid strategy and perfonnance. Moreover, a hybrid strategy's perfonnance impact is inconsistently defined in these studies. I address these conceptual and methodological issues in this study. First, following the most influential strategy typologies. I argue that any analysis of the relationship between hybrid strategy and performance must take into account

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6_4, Š Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


90

4 Hybrid strategy and firm performance

organizational architecture, as it moderates the impact of hybrid strategies on firm performance. I argue that an ambidextrous organizational architecture is required to manage the tensions resulting from the simultaneous pursuit of exploration and

exploitation. I analyze two different ways to achieve organizational ambidexterity, that is. through individual and technological ambidexterity. Individual ambidexterity concems the potential of specific organizational structures and human resource management (HRM) practices to enable an individual employee to pursue both exploratory and exploitative activities. Technological ambidexterity concerns the yet largely unexplored potential of infonnation and communication technology (lCT) to support the simultaneous pursuit of exploration and exploitation. Second, I argue that different definitions of the perfonnance impact of hybrid strategies must be employed to draw meaningful conclusions on the relationship between hybrid strategy and performance. On the one hand, different comparison groups must be consulted. The performance impact of hybrid strategies can be evaluated in comparison to all finns or in comparison to finns with a pure strategy, that Is, a strategy with a clear emphasis on either exploration or exploitation. Further, firms with a no-emphasis strategy, that Is, finns that lack any strategic emphasis, can serve as a reference group. On the other hand, a hybrid strategy's performance effect can be defined as the effect on average performance or; more stringently, as the effect on the odds of achieving performance above the industry average. To test my hypotheses, I employ a novel multi-source dataset. I match time-series data on objective firm performance indicators and time-series data on lCT with crosssectional data on corporate strategy, organizational architecture, and other firm characteristics. I find that the impact of a hybrid strategy on firm performance is positive in the presence of organizational ambidexterity, while it is negative in the absence of organizational ambidexterity. Further, I find that hybrid strategies outperform the entire sample and no-emphasis strategies in presence of organizational ambidexterity, but that hybrid strategies underperfonn the entire sample and are equal in

performance to

no-emphasis

strategies

in

the

absence

of organizational

ambidexterity. Finally, I find that hybrid strategies match the performance of pure strategies in the presence of organizational ambidexterity, but that hybrid strategies underperform pure strategies in the absence of organizational ambidexterity. These findings largely hold across two different definitions of a strategy's perfonnance impact. This study contributes to theory and practice in several ways. First, by taking into account the moderating effect of organizational architecture in the relationship between


4.2 Theory and hypotheses

91

hybrid strategy and performance, this study is more theoretically grounded than prior empirical studies. Second, with regard to methodology, by utilizing different definitions of a hybrid strategy's performance effect, the relative advantages and disadvantages of a hybrid strategy to different comparison groups are elaborated in more detail than in prior studies. Third, connections between the largely separated literatures on strategy typologies and organizational ambidexterity are explored. Fourth, the potential ofICT in enhancing the perfonnance effects of hybrid strategies is analyzed by introducing the concept of technological ambidexterity, Overall, the issues of hybrid strategy and organizational ambidexterity are examples of "conceptual and methodological innovations in the study of organization in recent years" that require an analysis of the precise connections between strategy and the design of organizational architecture, as stipulated by Gulati, Puranam, and Tushman (2009: 1). I believe that this study is an important step in this direction. In the following section, 4.2, I present the theoretical and empirical background of this study and develop the hypotheses to be tested. In section 4.3, I describe the methods. data, and measures that are employed in the multivariate analysis. In section 4.4, I discuss the results of the analysis. In the final section, 4.5, I present conclusions for researchers and practitioners as well as directions for future research.

4.2

Theory and hypotheses

4.2.1 Hybrid strategy, organizational architecture, and firm performance in theory The right composition of corporate strategy is the central concem of strategy typologies. The typologies of Porter (1980; 1985), Miles and Snow (1978), and March (1991) have received the most attention; together, the three typologies share two broad similarities. The first similarity is the idea that firms have to choose their strategy mix from two distinct strategic emphases. Porter (1980; 1985) distinguishes between a differentiation and a cost leadership strategy.รถ while Miles and Snow (1978) distinguish between a prospector and adefender strategy. Finally, March (1991) distinguishes between an

25 Porter's (1980; 1985) focus strategy is a differen tiation, or a cost leadership, strategy t hat is directed at a partieular strategie market.


4 Hybrid strategy and firm performance

92

exploration and an exploitation strategy.รถ I use March's (1991) tenns in the following, as his typology has given new momentum to the research on the hybrid strategy-

performance relationship in recent yearsP The second similarity is that the typologies advance hypotheses regarding the performance outcomes of different strategy mixes. Particularly, it is a topic of much discussion if the pursuit of hybrid strategies that entail both exploration and exploitation is beneficial or detrimental to a firm's perfonnance in general, or in

comparison to pure strategies that foeus on either exploration or exploitation. According to Porter (1980; 1985), differentiation and cost leaders hip strategies each require a total commitment to distinct organizational arrangements, making the pursuit of a hybrid strategy a "recipe for below-average performance" (Porter, 1985: 16). Firms with a hybrid strategy suffer from a need for a conflicting set of organizational arrangements and are thus "stuck in the middle" (Porter, 1980: 41). In contrast, Miles and Snow (1978) introduce a hybrid analyzer strategy that "combines the strengths of both the Prospector and the Defender" (Miles and Snow, 1978: 68). The analyzer strategy demands a "delicate balance" between distinct organizational structures and processes and a management that is "continually vigilant" (Miles and Snow, 1978: 68). However, given separate organizational units with different structures and processes for exploration and exploitation, the analyzer strategy is an equal alternative to pure strategies with respect to achieving superior firm performance (Miles and Snow, 1978). March

(1991)

recognizes that

exploration and

exploitation require

different

organizational forms and also that focusing on either exploration or exploitation is only possible at the expense of the other. A pure explorer "will ordinarily suffer from the fact that it never gains the returns of its knowledge" due to a never-ending sequence of unrewarded search and change, while a pure exploiter "will crdinarily suffer from obsolescence" due to a reduced responsiveness to environmental changes (Levinthal and March, 1993: 105). Following this insight, Tushman and O'Reilly (1996) state in their ambidexterity premise that "firms capable of simultaneously pursuing exploitation 26 Originally, March (1991) distinguishes exploratory and exploitative types of organizationallearning bu t states that t he choice between t he two is made in decisions regarding competitive strategy. His distinc tion is thus often interpreted as a strategy typology (Auh and Mengue, 2005; Gibson and Birkinshaw, 2004; He and Wong, 2004; Lubatkin et al., 2006; Thornhill and White, 2007; Uotila et al., 2009; Venkatraman et al., 2007). 27 See the recent studies by Auh and Mengue (2005), Gibson and Birkinshaw (2004), He and Wong (2004), Lubatkin et al. (2006), Thornhill and White (2007), Uotila et al. (2009), and Venkatraman et al. (2007).


4.2 Theory and hypotheses

93

and exploration are more likely to

achieve superior perfonnance than firms

emphasizing one at the expense ofthe other" (Raisch and Birkinshaw, 2008: 392). The capability to leverage a hybrid strategy results from setting up an ambidextrous organization that hosts "multiple contradictory structures, processes. and cultures within the same firm" (Tushman and O'Reilly, 1996: 24). Different ways to achieve organizational ambidexterity

are

suggested,

including structural ambidexterity

(Tushman and O'Reilly, 1996), individual ambidexterity (Adler et al., 1999; Gibson and Birkinshaw, 2004), leadership-based ambidexterity (Lubatkin et al., 2006), and informal ambidexterity (Gulati and Puranam, 2009).28 Tushman and O'Reilly (1996) emphasize that it is difficult to manage the organizational tensions between exploration and exploitation.

These

authors

implicitly argue that for

finns

without suitable

organizational arrangements, the performance effects of a hybrid strategy can be negative (He and Wong, 2004). In summary, while the most influential strategy typologies agree that hybrid strategies produce considerable organizational tensions, they differ with regard to the solvability of these tensions and thus on the performance effects of hybrid strategies. The hypotheses regarding the effect of a hybrid strategy on firm perfonnance range from a negative effect (Porter, 1980; 1985; 1996) to a positive effect equal to that of pure strategies (Miles and Snow, 1978) to a positive effect greater than that of pure strategies (Tushman and O'Reilly, 1996).

4.2.2 Hybrid strategy, organizational architecture, and firm performance in empirical studies 4.2.2.1 Summary 01 empirical studies Much like the theoretical analyses discussed above, the findings of empirical studies on the hybrid strategy-performance relations hip are mixed as well. Porter's (1980; 1985) typology is supported by findings that only pure strategies are related to above-average firm perfonnance (Dess and Davis, 1984; Hall, 1980), that pure strategies outperform hybrid strategies, and that hybrid strategies only perform as well 28 'Organizational ambidexterity' constitutes the ability to solve organizational tensions from a hybrid strategy and thus to successfully pursue a hybrid strategy. An 'ambidextrous organizational architecture' is the antecedent of this ability (Raisch and Birkinshaw, 2008) and thus a good proxy for organizational ambidexterity. Hence, in line with many prior studies (Gibson and Birkinshaw, 2004; Raisch and Birkinshaw, 2008; Raisch et al., 2009), the terms forthe ability and its antecedent are used synonymously to enhance readability.


94

4 Hybrid strategy and firm performance

as no-emphasis strategies (Miller and Friesen, 1986). In contrast, other studies find that pure strategies do not always outperform hybrid strategies (Milles and Dess, 1993) and that hybrid strategies can also lead to above-average firm performance or even outperform no-emphasis strategies [Miller, 1992; Spanos et al., 2004; White, 1986). Some studies even find that hybrid strategies outperform one or both types of pure strategies (Spanos et al., 2004; White, 1986). Miles and Snow's (1978) typology is supported by studies that find a positive performance effect f端r prospector, defender, and analyzer strategies but a negative effect f端r no-emphasis strategies (Moore, 2005). Further support comes from findings that prospectors. defenders, and analyzers perform equally well and outperform noemphasis strategies (Conant et al., 1990). Other studies find that analyzers can even outperform prospectors and defenders in certain contexts (Hambrick, 1983a; Parnell and Wright, 1993). Some studies, using March's (1991) typology, find that hybrid strategies have a positive impact on firm performance (Cao et al., 2009; Gibson and Birkinshaw, 2004; He and Wong, 2004; Lubatkin et al., 2006; Uotila et al., 2009), while others find that hybrid strategies do not significantly affect firm performance (Kyriakopoulos and Moorman, 2004; Venkatraman et al., 2007). Finally, while several studies do not directly test these three strategy typologies, they nevertheless derive similar typologies to one or more of them. Some studies find that the finns in high profit clusters pursue pure strategies (Hambrick, 1983b) and that the degree of strategie purity positively impacts firm performance (Thornhill and White, 2007). Other studies find that hybrid strategies positively correlate with firm performance in some cases (Parnell and Hershey, 200S), lead to above-average firm performance in general, or may even outperform no-emphasis and pure strategies (Parnell, 1997; Parnell et al., 2000). In summary, the findings of the empirical studies range from a negative performance impact of hybrid strategies to no significant impact to a positive impact. Of course, these studies significantly differ with regard to their research designs. I argue that these ambiguous results must be interpreted with caution due to methodological and conceptual reasons.


4.2 Theory and hypotheses

95

4.2.2.2 Definitions ofa hybrid strategy's performance impact With regard to methodology, the empirical studies outlined above employ different definitions of a hybrid strategy's performance impact. The studies using March's (1991) strategy typology concentrate on whether hybrid strategies positively affect firm performance at all. In contrast, most studies on Porter's (1980; 1985) and Miles and Snows (1978) typologies compare the average perfonnance of firms with a hybrid strategy with the average performance of different comparison groups. Specifically, firms with a hybrid strategy are compared with all ftrms, that Is, the entire sample of firms of the respective study. Further, firms with a hybrid strategy are compared with finns that pursue a no-emphasis strategy or with firms that pursue a pure strategy. Further; all studies define a hybrid strategy's performance impact as the effect on average perfonnance. None of the studies employs Porter's original definition of superior firm performance, that Is, above industry average performance (Porter, 1980; 1985). According to the latter; a hybrid strategy's perfonnance impact is defined as the strategy's effect on the odds of achieving a performance that is above the industry average [Campbell-Hunt, 2000). Obviously, this is a more stringent definition of a strategy's performance impact than analyzing whether a strategy positively affects average performance at all. These

methodological inconsistencies make

prior findings

hard to

compare.

Consequently, I employ different definitions of a hybrid strategy's performance impact in this study to allow for comprehensive conclusions on the hybrid strategyperformance relationship.

4.2.2.3 Organizational ambidexterity as a moderator of the hybrid strategy-performance relationship With regard to theory, although the strategy typologies reviewed above emphasize the role of organizational architecture in the hybrid strategy-perfonnance relationship, the empirical studies largely do not factor in organizational architecture, at least not in a way that is consistent with theory. As outlined above, some strategy typologies suggest that hybrid strategies positively affect firm performance, provided that there are suitable organizational mechanisms to manage the tensions from the simultaneous pursuit of exploration and exploitation (that


4 Hybrid strategy and firm performance

96

is. organizational ambidexterity). In contrast, hybrid strategies negatively affect firm performance in the absence of suitable organizational arrangements (He and Wong, 2004). Thus, the effect ofhybrid strategies on firm performance is moderated by organizational ambidexterity (see Figure 4.1). According to the moderation perspective, the impact of a predictor variable (f端r example, hybrid strategy) on a criterion variable (f端r example, firm performance) is dependent on a third moderating variable (f端r example, organizational ambidexterity]. The moderator affects the strength and direction of the relationship between predictor variable and dependent variable (Venkatraman, 1989).

Organizational ambidexterity

Hybrid strategy

1

Firm Performance

Figure 4.1 Ambidexterity as a moderator of the hybrid strategy-performance relationship Thus, the many studies that assume a positive or negative effect of hybrid strategies on firm performance without considering organizational ambidexterity do not distinguish between the level at which the tension between exploration and exploitation is caused (that is. corporate strategy) and the level at which the conflict is resolved (that Is, organizational architecturej.s" Implicitly and counterintuitively, these studies assume that corporate strategies are always supported by the right organizational architecture. This is also reflected by the terminology of many studies using March's (1991) typology. There, the term ambidexterity is used to describe a hybrid strategy (Cao et al., 2009; He and Wong, 2004; Thomhill and White, 2007; Uotila et al., 2009; Venkatraman et al., 2007). In contrast, Tushman and O'Reilly (1996) originally use the tenn exclusively to refer to the organizational arrangements for managing the tensions from a hybrid

See Cao et al. (2009), Conant et al. (1990), Dess and Davis (1984), Hall (1980), Hambrick (1983a; 1983b), He and Wong (2004), Kyriakopoulos and Moorman (2004), Miller (1992), Miller and Dess (1993), Miller and Friesen (1986), Moore (2005), Parnell (1997), Parnell and Hershey (2005), Parnell and Wrigh t (1993), Parnell et al. (2000), Spanos et al. (2004), Thornhill and White (2007), Uotila et al. (2009), and Venkatraman et al. (2007). 29


4.2 Theory and hypotheses

97

strategy; that Is, to refer to a feature of crganizational architecture and an organizational ability rather than to a feature of corporate strategy. In the few empirical studies that incorpcrate organizational ambidexterity, the hybrid strategy is seen as a mediator between organizational ambidexterity and firm performance (see Figure 4.2). Thus, organizational ambidexterity is seen as an antecedent of the hybrid strategy, but it is still the hybrid strategy per se that is assumed to lead to superior firm performance (Gibson and Birkinshaw, 2004; Lubatkin et al.,

2006).

Organizational ambidexterity

Hybrid strategy

Firm Performance

Figure 4.2 Hybrid strategy as a mediator of the ambidexterity-performance relationship The mediation perspective is useful to decompose the effects of organizational ambidexterity on firm perfonnance into direct versus indirect effects. However, the moderation perspective is appropriate to evaluate the differences in the relationship between a hybrid strategy and firm performance across different organizational architectures (Venkatraman, 1989). Thus, the moderation perspective is more suitable to test strategy typologies with respect to the role of organizational ambidexterity in the hybrid strategy-performance relationship. To date, a moderation perspective is only employed for contingencies such as environmental dynamism (Andripoulous and Lewis, 2009; Grosyberg and Lee, 2009; Jansen et al., 2005a; 2006; Mom et al., 2009, Rothaermel and Alexandre, 2009; Uotila et al., 2009) and resource endowment (Cao et al., 2009) that influence a firm's need or capability to pursue a hybrid strategy. To my knowledge, no empirical study has analyzed the role of organizational architecture as a moderator of the hybrid strategyperformance relationship in a theoretically-informed way.ae

[ansen (2005) assumes that structural ambidexterity positively moderates the hybrid strategyperformance relationship. Nonetheless, in the first instance, he assumes t hat a hybrid strategy positively affects firm performance in any case and that a suitable organizational architecture will only strengthen this positive effect [jansen. 2005). Organizational architecture is t hus assumed t o affect the strength bu t not the direction of the hybrid strategy-performance relationship. This does not fully reflect t he strategy typologies tha t suggest t hat organizational architec ture affects both t he strength and direction of the hybrid strategy-performance rela tionship. 30


98

4 Hybrid strategy and firm performance

4.2.3 Organizational ambidexterity The question regarding how to create an ambidextrous organizational architecture that reconciles the contradictions between exploration and exploitation has been answered in different ways (Gupta et al., 2006; O'Reilly and Tushman, 2008; Raisch et al., 2009; Raisch and Birkinshaw, 2008; Simsek, 2009). The strategy typologies described above and many other studies in management research areas such as organization theory (Benner and Tushman, 2003; Bums and Stalker, 1961; Duncan, 1976; Holmqvist, 2004), managerial economics (Ghemawat and Ricart I Costa, 1993), technology and innovation management (Ambos et al., 2008; He and Wong. 2004), and HRM [lackson et al., 1989; Schuler and [ackson, 1987; Youndt et al., 1996) explicitly or implicitly make the distinction between exploration and exploitation: they largely agree on the organizational structures and HRM practices required for both strategy types. Basically, an organic organizational architecture that relies on decentralization, autonomy, and coordination among functions as well as on HRM practices that focus on general skills and subjective performance appraisals are best suited for exploration. In contrast, a mechanistic organizational architecture that relies on centralization, standardization, and monitoring, as well as on HRM practices that focus on specific functional skills and objective performance appraisals. are best suited for exploitation. The original and most obvious solution to combine these contradictory organizational architectures is the creation of separate organizational units (that Is, structural ambidexterity]. Hence, individuals in units with an organic architecture focus entirely on exploration, while individuals in units with a mechanistic architecture focus entirely on exploitation (Benner and Tushman, 2003; Duncan, 1976; Miles and Snow, 1978; Tushman and O'Reilly, 1996). However; recent studies argue that dual structures are difficult to implement and that some finns fail at establishing sufficient linkages between exploratory and exploitative units, leading to the neghgence of either exploration or exploitation. Further, it is argued that it is inefficient to let upper management decide on how to divide up the time of employees between exploration and exploitation. Also, it is argued that structural ambidexterity is better suited for larger firms, while smaller firms, which constitute the majority of finns in most economies, lack the slack resources and number of hierarchical levels to attain structural ambidexterity (Birkinshaw, 2006; Birkinshaw and Gibson, 2004a; 2004b; 2005; Gibson and Birkinshaw, 2004; Lubatkin et al., 2006).


4.2 Theory and hypotheses

99

4.2.3.1 Individual ambidexterity

Consequently, an approach other than structural ambidexterity has focused on how organic and mechanistic features can be combined within every organizational unit throughout the firm to create a behavicral context that enables and encourages individuals to make their own judgments about how to divide their time between exploratory and exploitative activities in the course of their day-to-day work (Adler and Bcrys. 1996; Adler et al., 1999; Birkinshaw, 2006; Birkinshaw and Gibson, 2004a; 2004b; 2005; Jansen et al., 2005b; Sheremata, 2000). What I call 'individual ambidexterity' is achieved at the level of the individual employee and not at the level of organizational units. Individual ambidexterity "emerges from choices made by individuals. not from the strategic decisions made by senior executives" (Birkinshaw, 2006).31 The literature suggests four critical factors to achieve individual ambidexterity, namely, decentralization and standardization, which are structural arrangements, and incentives and skills, which can be attained through certain HRM practices. Decentralization, that Is, the delegation of decision authority to lower hierarchical levels, enables employees to switch between exploratory and exploitative activities in their sole discretion. Employees can pursue their regular production work and at the same time stay alert for

new opportunities, generate new ideas, and subsequently act

spontaneously and experiment with these ideas without seeking permission from above. Decentralization allows employees to directly communicate with and access the resources of employees on the same hierarchical level but from other functions. This allows for an interplay between a variety of perspectives as well as the development of a broad range of skills and knowledge that are crucial for creativity. Thus, on the one hand, decentralization allows employees to divide up their time between exploration and exploitation in their sole discretion, but on the other hand, it clearly favors exploratory activities. In fact, decentralization may even be detrimental to exploitation. Through decentralization, employees may lose the discipline to efficiently pursue their day-to-day werk, They may invest too much time and too many resources in imagination 31 Two further ways t o achieve organizational ambidexterity tha t are suggested are leadership-based ambidex terity and what could be called 'informal ambidexterity'. Leadership-based ambidexterity is obtained by top management teams that think and act ambidextrously due to a high level of behavioral integration (Lubatkin et al., 2006 ; Smith and Tushman, 2005). Leadership-based ambidexterity complements structural or individual ambidexterity, while t he latter two are indispensable t o achieve organizational ambidexterity. Informal ambidexterity stands for personal relationships as weil as the shared values and beliefs of employees that for a limited t ime might compensate for problems emerging when a formal organization is directed to only one of both, exploration or exploitation (Gulati and Puranam 2009). Thus, informal ambidexterity is a t emporal phenomenon. Hence, although both types of organizational ambidexterity can be fruitful, t hey constitute complementary or temporal forms of organizational ambidexterity and are thus not further considered in t his study .


100

4 Hybrid strategy and firm performance

and trial and error (Adler et al., 1999; Benner and Tushman, 2003; Birkinshaw, 2006; Birkinshaw and Gibson, 2004a; 2004b; 2005; Gibson and Birkinshaw, 2004; Jansen et al., 2005a; 2006; Raisch et al., 2009; Roberts, 2007; Sheremata, 2000; Tushman and

Q'Reilly, 1996). Thus, the "specification of clear limits beyond which the individual must no stray" is required (Birkinshaw, 2006). This "tight-loose aspect [...] is crucial f端r ambidextrous organizations" (Tushman and O'Reilly, 1996: 27). Standardization, that Is, the release of oral and written rules and procedures, can guide the efficient execution and continuous enhancement of exploitative day-to-day activities (Adler and Bcrys, 1996; Adler et al., 1999; Benner and Tushman, 2003; Jansen et al., 2006). In addition, standardization can also systematize and thus increase the efficiency of exploratory activities without impairing creativity and flexibility. First, standardization can be used to capture and codify lessons of prior experience with the creation and implementation of exploratory innovations as well as to quickly disseminate these best practices throughout the firm (Adler and Borys, 1996; Adler et al., 1999; Jansen et al., 2005a; Roberts, 2007; Sheremata, 2000). Second, standardization can be used to increase the lateral knowledge exchange needed for creativity and at the same time reduce its costs [lausen et al., 2005a).32 Thus, standardization ensures that employees pay attention to exploitation in two manners. First, standardization enhances classic exploitative activities: that Is, the efficient pursuit of the current product market agenda. Second, standardization also ensures the efficient pursuit of explcratory activities that are targeted at changingthe product market agenda. A further mechanism to make sure that employees balance their exploratory and exploitative activities is to give them incentives for both types of activities through employee appraisal and performance-based pay (Birkinshaw, 2006; Birkinshaw and Gibson, 2004a, 2004b, 2005; Gibson and Birkinshaw, 2004). Exploration requires interaction and collaboration among employees as well as the development of a broad range of skills and knowledge, which are hard to measure objectively. Thus, employee appraisal and perfonnance-based pay should be based on subjective perfonnance measures; that Is, on behavior rather than results and on skills and skill development [lackson et al., 1989; Schuler and [ackson, 1987; Youndt et al., 1996). In contrast,

32 Actually, standardization is often seen as a substitute for centralization and thus as harmful to decentralization. However, if standardization is not only used to exercise command and control over employees bu t in conjunction with decentralization and thus to facilitate activities, employees will welcome standardization and see it as 'enabling' rather than 'coercive' (Adler and Borys. 1996; Adler eta!., 1999).


4.2 Theory and hypotheses

101

exploitation focuses on cost reduction and efficiency improvement. Thus, individual performance contributions to exploitation can easily be attributed. Consequently, performance appraisal that incentivizes exploitation should be based on objective performance measures, that Is, on quantitative results (Schuler and [ackson, 1987; Youndt et al., 1996). Finally, the importance of both employee selection and employee training to enable employees to fulfill exploratory as well as exploitative tasks has been highlighted (Adler et al., 1999; Birkinshaw, 2006; Birkinshaw and Gibson, 2004a, 2004b, 2005; Gibson and Birkinshaw, 2004). To foster explcratory skills and knowledge, employee selection and training should lay stress on a broad range of skills. including problem-solving skills, skills that allow employees to understand the entire production process. and skills that are usable in a variety of positions in the firm. Finally, skills that foster the exchange of knowledge and ideas throughout an organization should be advanced (Schuler and [ackson, 1987; Youndt et al., 1996). To foster exploitative skills and knowledge, employee selection and training should focus on firm-, product-, and function-specific skills and expertise, on the correction of skill deficiencies, and on knowledge about the

firm's rules and procedures (Delery and Doty, 1996; Schuler and [ackson, 1987; Youndt et al., 1996).

In summary, individual ambidexterity can be achieved by a combination of explcratory structures and HRM practices and exploitative structures and HRM practices. Decentralization, employee appraisal based on subjective criteria, and employee selection and training targeted at the development of diverse, soft skills can foster exploratory activities. Standardization, employee appraisal based on objective criteria, and employee selection and training targeted at skills and knowledge needed for function-specific efficiency can foster exploitative activities.

4.2.3.2 Technological ambidexterity The literature has thus far largely disregarded the potential of ICT to enhance organizational ambidexterity and thus

affect

the hybrid strategy-perfonnance

relationship. Yet, ICT can support both structural and individual ambidexterity. ICT has been suggested as a means to support components of both organic and mechanistic organizational architectures (Dewett and [ones, 2001; George and King, 1991; Huber, 1990; Pinsonneault and Kraemer, 1997), that Is, architectures supporting both exploration and exploitation (Benner and Tushman, 2003; Duncan, 1976; Miles and


102

4 Hybrid strategy and firm performance

Snow, 1978). Other studies show that leT directly facilitates the exploratory and exploitative activities of individuals (Karre and Alavi, 2007; MOll et al., 2003). Further studies show that ICl' may serve exploratory goals like product innovation and exploitative goals such as cost reduction, though these studies do not directly investigate the mechanisms through which this is achieved (Aral and WeiH, 2007; Oh and Pinsonneault, 2007). These studies also emphasize that leT is not a homogenous technology but rather that different types of leT support different organizational architectures (Dewett and [ones, 2001; Huber, 1990; Pinsonneault and Kraemer, 1997), employee activities (Karre and Alavi, 2007), and performance dimensions (Aral and Weill, 2007; Oh and Pinsonneault, 2007). Thus, I argue that different lCT types enable exploratory activities (that Is, exploratory lCT) and exploitative activities (that Is, exploitative lCT) and that their combination supports both structural and individual ambidexterity. As outlined above, exploration is fostered through decentralization under both structural and individual ambidexterity. Decentralization allows autonomous employees to communicate and collaborate and consequently to generate a broad range of skills and ideas as well as the capacity to experiment with these ideas. As groupware applications and corporate intranets facilitate the communication and collaboration of emplcyees, they are examples of exploratory lCT. Groupware applications offer functionalities such as e-mail, threaded discussions, instant text messeging. data conferencing, audio and video conferencing, document sharing, group scheduling, and task sharing (Ellis et al., 1991; Grudin, 1994). Similarly, corporate intranets are private computing networks intemal to an organization that are used to offer groupware applications (Curry and Stancich, 2000; Ginsburg and Duliba, 1997). Thus, groupware applications (Alavi and Leidner; 2001; Dewett and [ones, 2001; Ellis et al., 1991; Kane and Alavi, 2007) and corporate intranets (Bloom et al., 2009; Mom et al., 2003) enable employees throughout an organization to communicate with each other, share information and data, cocrdinate, and work together. As a consequence, groupware applications and corporate intranets facilitate decentralization and thus exploration in either structural or individual ambidexterity. Exploitation is fostered through centralization under structural ambidexterity and through standardization and results-based employee appraisal under individual ambidexterity. As enterprise resource planning (ERP) applications and workflow applications facilitate centralization, standardization, and results-based employee appraisal, they are examples of exploitative lCT. Both ERP and workflow applications focus on the automation of business processes and data transfers across the


4.2 Theory and hypotheses

103

organization, but they use different architectural approaches (Cardoso et al., 2004).33 ERP applications consist of prefabricated applications and thus require business processes to be reshaped in conformity with the ERP's design. Instead, workflow applications are a type of middleware that improves already-existing business processes by integrating separate applications. databases. and legacy applications into an organization-wide system (Cardoso et al., 2004; Stohr and Zhao, 2001). ERP and workflow applications integrate information across various functional departments and operations

such

as

supply chain management

(SCM), customer relationship

management (CRM), financial and cost accounting. and human resources into one unified system (Bloom et al., 2009; Cardoso et al., 2004; Hitt et al., 2002). They offerrealtime metrics regarding orders. stocks, production, deliveries, HRM, financials, and other firm characteristics. Hence, a full picture of what is currently happening in an organization and a clear view of the relative performance of different parts of the organization becomes available through such technologies (Bloom et al., 2009; Cardoso et al., 2004; Chari et al., 2008; Hendricks et al., 2007; Hitt et al., 2002). Thus, under structural ambidexterity, ERP and workflow applications can be used to foster centralization by giving faster; better, and more information to the managers that make central decisions (Bloom et al., 2009; Chari et al., 2008; Hendricks et al., 2007; Stohr and Zhao. 2001). Further, ERP and workflow applications can enhance centralization by allowing managers to better monitor the implementation of their decisions. Similarly, under individual ambidexterity, ERP and workflow applications facilitate the monitoring ofthe implementation ofrules and procedures and thus enhance standardization (Alavi and Leidner, 2001; Grant and Higgins, 1991; Orlikowski and Robey, 1991). Further, as "performance measurement becomes a matter of examining the system log" (Stohr and Zhao. 2001: 291), ERP and workflow applications facilitate the use of employee appraisal and performance-based pay that is using objective performance criteria. ERP and workflow applications can thus impose a relatively rigid work environment on employees (Stohr and Zhao, 2001) and foster exploitation under structural and individual ambidexterity. In summary, exploratory ICT such as groupware applications and corporate intranets can enhance decentralization. Exploitative ICT such as ERP and workflow applications can enhance centralization, standardization, and objective criteria-based employee appraisal. Thus, both structural ambidexterity and individual ambidexterity are enhanced through a combination of exploratory and exploitative ICT. 33 Thus, I t ransfer arguments that have been put forward for ERP applications to workflow applications, and vice versa.


4 Hybrid strategy and firm performance

104

4.2.4 Hypotheses In order to develop hypotheses. I draw on the theoretical and empirical research on strategy typologies, organizational ambidexterity, and H'T. I present the argument that the impact of hybrid strategies on firm performance is positively moderated by organizational ambidexterity, because organizational ambidexterity allows f端r the management of tensions that arise from hybrid strategies.

Hypothesis 4.1: Organizational ambidexterity positively moderates the impact ojhybrid strategies on firm performance. With regard to methodology, I further note that a hybrid strategy's perfonnance impact can be defined in different ways. In addition to analyzing the general performance effect of hybrid strategies as in hypothesis 4.1, the performance advantages of hybrid strategies to different comparison groups can be studied. Figure 4.3 shows the five groups of firms that I distinguish according to the two dimensions of corporate strategy and organizational architecture.

Hybrid strategy

OrganizEtional <rub idexterity

Noorganizational arnbidrnerity

No-ernphasis strategy

Pure strategy

Pureorganizat ional architecture

Nopure organizational architecture

Anyolganizational architecture

Figure 4.3 Five groups of firms split by corporate strategy and organizational architecture Building on hypothesis 4.1, I suggest that the average performance of finns with a hybrid strategy and organizational ambidexterity is higher than the average performance of the entire sample of firms in my study. In contrast, the average performance of firms with a hybrid strategy but without organizational ambidexterity is below the average performance of the entire sample of firms. Thus, as shown in Figure 4.3, hypothesis 4.2 compares the average performance of finns in groups 1 and 2 with the average perfonnance of all firms in groups 1 to 5.


4.2 Theory and hypotheses

105

Hypothesis 4.2a (4.2b): On average, in the presence (absence) of organizationaf ambidexterity, hybrid strategies outperform (underperform) the entire sampfe offirms. Based on hypothesis 4.1, I further hypothesize that the average performance of firms with a hybrid strategy and organizational ambidexterity is greater than the average performance of firms with a no-emphasis strategy. In contrast, the average perfonnance of firms with a hybrid strategy but without organizational ambidexterity, matches the rather low average performance of the firms that lack a clear strategie emphasis. Thus, as shown in Figure 4.3, hypothesis 4.3 compares the average performance of the firms in groups 1 and 2 with the average performance of the firms in group 5.

Hypothesis 4.3a (4.3b): On average, in the presence (absence) of organizationaf ambidexterity, hybrid strategies outperform (match) no-emphasis strategies.

The specification of a hypothesis with respect to the average performance of hybrid strategies relative to the average performance of pure strategies requires further consideration. Miles and Snow (1978) hypothesize that firms with hybrid strategies and organizational ambidexterity perform as well as firms with pure strategies, while ambidexterity researchers hypothesize that firms

with hybrid strategies and

organizational ambidexterity even outperform finns with pure strategies (Raisch and Birkinshaw, 2008). However, March (1991) argues that the benefits from exploration become apparent in the long-term, while the benefits from exploitation become apparent in the short-term. Thus, it is argued that hybrid strategies outperform pure strategies in the long-term rather than the short-term (Raisch and Birkinshaw, 2008; Van Looy et al., 2005). Due tothe composition ofthe sample used in this study, I advance a hypothesis that concerns the short-tenn and follows Miles and Snow (1978). In the short-term, finns with a hybrid strategy and with organizational ambidexterity match the average perfonnance of finns with a pure strategy, while firms with a hybrid strategy but without organizational ambidexterity display a lower average performance than firms with a pure strategy. Thus, as shown in Figure 4.3, hypothesis 4.4 compares the average performance of firms in groups 1 and 2 with the average performance of finns in groups 3 and 4.

Hypothesis 4.4a (4.4b): On average, in the presence (absence) of organizationaf ambidexterity, hybrid strategies match

(underperform) pure strategies.


4 Hybrid strategy and firm performance

106

Although I do not formally propose such hypotheses, it makes sense to subclassify the group of firms with a pure strategy into firms with or without pure structures and HRM practices in order to allow für more subtle performance comparisons. It is intuitive to assume that firms with a pure strategy and high levels of pure structures and HRM practices might display a higher average perfonnance than firms with a pure strategy but with low levels of pure structures and HRM practices. Further; though I do not fonnally propose such hypotheses, I "Will reassess hypotheses 4.1 to 4.4 by employing Porter's (1980; 1985) more challenging definition of a corporate strategy's performance effect, that Is, the strategy's effect on the odds of achieving performance above the industry average.

4.3

Method

4.3.1 Models 4.3.1.1 Moderation effects model Hypothesis 4.1 pertains to the interaction effect, that Is, the joint effect of a hybrid strategy and organizational ambidexterity on firm performance. Thus, the following moderation effects model is used to test hypothesis 4.1:

ß i hybrid strategy i(t-i) + ßz organizational ambidexterity i(t-i) (hybrid strategy i(t-i) x organizational ambidexterity m-n] ß4 environmental dynamism i(t-i) + ßsROA i(t-i) ßs-» [controls] i(t-i) + snn

ROAit= ßo

+ + +

+

ß3

The dependent variable is the return on total assets

(ROA)

(4.1)

for firm i at time t. The

independent variables are hybrid strategy, organizational ambidexterity, the interaction term of hybrid strategy and organizational ambidexterity, environmental dynamism, and

ROA.

Further control variables are described below. To account for repeated

observations of the same firm, the error term

E

is clustered by firm; it is Huber-White

robust to heteroskedasticity and autocorrelation of unknown form. All independent variables are lagged one year; that Is, they appear as at time t-1.


4.3 Method Hypothesis 4.1 predicts that

107

Ă&#x;3 is significantly greater than zero, which indicates that the

regression of ROA on hybrid strategy depends on the level of organizational ambidexterity. I include lagged independent variables for theoretical and statistical reasons. First, choosing a strategy and an organizational architecture is linked to long-term investments that might require a certain time period to affect firm performance. Second, instrumenting independent variables by their lagged values mitigates the potential for simultaneity, that Is, the possibility that firm performance and strategy or organizational architecture are jointly determined (Spanos et al., 2004). I include environmental dynamism as a control variable, since prior studies have shown that environmental dynamism can affect both a firm's necessity to pursue a hybrid strategy and a firm's success in doing so (Auh and Mengue, 2005; Jansen et al., 2005a; 2006; Raisch and Hotz, 2010). Consistent with many prior studies on firm performance, I include the lagged value of the dependent variable as a control variable to account for any unobserved firm heterogeneity that causes differences in the dependent variable and that is not captured by other control variables. In other words, by adding ROA at time t-l to the right hand side of the equation, I take into account factors that may have caused differences in ROA in the past and that are difficult to account for in other ways (Burton et al., 2002; Spanos et al., 2004; Thomhill and White, 2007; Uotila et al., 2009).

4.3.1.2 Ejfects-coding model Hypothesis 4.2 concerns the performance difference between the group of firms with a hybrid strategy and with (or without] organizational ambidexterity as opposed to the entire sample of finns. An effects-coding model is suitable to test this hypothesis (Hardy, 1993; Spanos et al., 2004). In an effects-coding model, indicator variables for all distinguishable groups of finns are created (see Figure 4.3), setting aside one group of firms as the reference group. The indicatcr variables are coded 1 for the group that is reflected by the indicatcr variable, 0 for the other included groups, and -1 for the reference group. The coefficient of the indicator variable for a particular group indicates the difference between the mean performance of this group and the mean of the mean performances of all groups, including the respective group (Hardy, 1993; Spanos et al., 2004). Although the mean of


4 Hybrid strategy and firm performance

108

the means of all groups does not equal the overall mean of the sample if there are unequal numbers of observations in each group, the regression coefficient of a particular group can still be interpreted "as a measure of the 'eccentricity' or the 'uniqueness' of the specified group in comparison to an 'average' value für the entire sample" (Spanos et

al., 2004: 153).34 The effects-coding model 4.2 includes indicatcr variables für firms with a hybrid strategy and organizational ambidexterity, für finns with a hybrid strategy but without organizational ambidexterity, für firms with a pure strategy and a pure organizational architecture, and für firms with a pure strategy but without a pure organizational architecture.e- Firms with a no-emphasis strategy serve as the reference group.v'

ß i hybrid strategy with organizational ambidexterity i(t- i) hybrid strategy without organizational ambidexterity i(t- i) ß3 pure strategy with pure organizational architecture i(t- i) ß4 pure strategy without pure organizational architecture i(t- i) ßs environmental dynamism i(t- i) + ß6ROA i(t- i) ß7-n [controls] i(t- i) + snn

ROAit= ßo

+ + + + +

+

ßz

Hypothesis 4.2a predicts that

ßi

(4.2)

is significantly greater than zero. which indicates that

the average perfonnance of finns with a hybrid strategy and organizational ambidexterity lies above the average performance of the entire sample. Hypothesis 4.2b predicts that

ßz

is significantly smaller than zero. indicating that the average

performance of firms with a hybrid strategy but without organizational ambidexterity lies belowthe average perfonnance of the entire sample.

4.3.1.3 Dummy-coding models Hypotheses 4.3 and 4.4 concem the performance differences between the group of firms with a hybrid strategy and with (or without] organizational ambidexterity as opposed to two different comparison groups. Dummy-coding models are suitable to test these hypotheses (Hardy, 1993; Spanos et al., 2004).

34

See also Cohen et al. (2002) and Hardy (1993).

35

The explana tions regarding the time lags and control variables in model 4.1 also apply to model 4.2.

Not e tha t the two groups of firms with a pure strategy could also serve as the reference group without changing the results for firms with a hybrid strategy and organizational ambidexterity as weil as for firms with a hybrid strategy bu twithout organizational ambidexterity. 36


4.3 Method

109

In a dummy-coding model, indicator variables for an distinguishable groups of firms are created (see Figure 4.3), setting aside one group of firms as the reference group. The indicator variables are coded 1 for the group that is reflected by the indicatcr variable and 0 for an other groups, including the reference group. The coefficient of the indicatcr variable for a particular group indicates the difference between the mean perfonnance of this group and the mean performance of the reference group (Hardy, 1993; Spanos et al.,2004). Hypothesis 4.3 concerns the performance difference between the group of firms with a hybrid strategy and with (or without] organizational ambidexterity as opposed to the group of firms with a no-emphasis strategy. Thus, the dummy-coding model 4.3 includes indicator variables for an groups of flrms, except for firms with a no-emphasis strategy. Apart from the coding of the indicatcr variables, model 4.3 is equal to model 4.2.

ROAit= ßo + ßi hybrid strategy with organizational ambidexterity i(t-i) + ßz hybrid strategy without organizational ambidexterity i(t-i) + ß3pure strategy with pure organizational architecture i(t- i) + ß4pure strategy without pure organizational architecture i(t-i) + ßs environmental dynamism i(t-i) + ß6ROA i(t-i) + ß7-n [controls] i(t-i) + snn

(4.3)

Hypothesis 4.3a (4.3b) predicts that ßi (ßz) in model 4.3 is significantly greater than zero (not significantly different from zero), which indicates that the average performance of finns with a hybrid strategy and with [without] organizational ambidexterity lies above (is equal to) the average performance of finns with a no-emphasis strategy. Hypothesis 4.4 concerns the performance difference between the group of firms with a hybrid strategy and with (or without] organizational ambidexterity as opposed to the group of firms with a pure strategy. As outlined, I split the group of finns with a pure strategy in two. resulting in the two models 4.4 and 4.5. The dummy-coding model 4.4 includes indicator variables for an groups of ftrms, except for the firms with a pure strategy and a pure organizational architecture. The dummy-coding model 4.5 includes indicator variables for an groups of firms, except for the firms with a pure strategy but without a pure organizational architecture. Apart from the coding of the indicatcr variables and the reference groups, models 4.4 and 4.5 are equal to model 4.2.


4 Hybrid strategy and firm performance

110

ROAit= ßo + ßi hybrid strategy with organizational ambidexterity iet-i) + ßz hybrid strategy without organizational ambidexterity iet-i) + ß3 no-emphasis strategy iet-i) + ß4 pure strategy with pure organizational architecture iet-i) + ßs environmental dynamism iet-i) + ß6ROA iet-i) + ß7-n [controls] iet-i) + snn

(4.4)

ROAit= ßo + ßi hybrid strategy with organizational ambidexterity iet-i) + ßz hybrid strategy without organizational ambidexterity iet-i) + ß3 no-emphasis strategy iet-i) + ß4 pure strategy without pure organizational architecture iet-i) + ßs environmental dynamism iet-i) + ß6ROA iet-i) + ß7-n [controls] iet-i) + snn

(4.5)

Hypothesis 4.4a (4.4b) predicts that

ßi

(ßz) in models 4.4 and 4.5 is not significantly

different from zero (significantly smaller than zero), which indicates that the average performance of firms with a hybrid strategy and [without] organizational ambidexterity is equal to (lies below) the average perfonnance of firms with a pure strategy. To test hypotheses 4.1 to 4.4 using Porter's (1980; 1985) more stringent performance definition, Irerun the regression models 4.1 to 4.4, employing a logistic analysis with the binary dependent variable ABOVE that indicates if firm i's performance at time t lies above the industry average perfonnance at time t (ABOVE=l) or if it is equal to or below the industry average performance (ABOVE=O).

4.3.2 Sampie I employ a novel multi-source panel dataset on corporate strategy, organizational structure, HRM practices, ICT, and firm performance of 784 German and Polish manufacturing firms to test the hypotheses developed above. To avoid single-source bias (podsakoff et al., 2003), I construct the dataset from three independent sources: (1) a panel of the ICT types used in firms over the period 200S-2008, (2) a panel of firm performance information over the period 200S-2008, and (2) a 2008 cross-section of corporate strategy, organizational structure, HRM practices, ICT use, and other firm characteristics. I obtained information on the ICT types employed by firms from Harte-Hanks' CI Technology Database (CITDB). Harte-Hanks is a market intelligence company that conducts at least annual telephone surveys to gather infonnation on specific hardware


4.3 Method

111

and software types that are used by the establishments of more than 10,000 German and 1,800 Polish firms. If an establishment was surveyed more than once in a year; I used the last observation from that year to gain a reliable picture of the ICT in use during the entire year. Harte-Hanks states that the finns in the CITDB are largely representative of the entire populations of German and Polish firms. Where possible, I derived firm performance Information from Bureau Van Dijk's ORBIS database, which provides balance sheets and profit and loss statements based on annual company reports. Several control variables were also derived from ORBIS. All German and Polish firms that maintain at least one manufacturing establishment and that are covered by both the CITDB and ORBIS formed the pool of 2,369 potential respondents to a telephone survey on corporate strategy, organizational structure, HRM practices, ICT use, and other firm characteristics that was conducted in August and September 2008. Nine student research analysts contacted 1,788 firms''?

and

interviewed 784 production managers (or similar managers), representing a response rate of 43.8 percent. The 784 interviewed finns formed the pool of potential respondents for a second telephone survey on ICT use in firms that was conducted during the same time period with the same research analysts. IT managers (or similar managers) from 454 of the 784 firms were interviewed, representing a response rate of 57.9 percent. The finns of the responding production managers do not significantly differ from the non-responding firms in terms of firm performance (that is. return on total assets], but they are slightly larger.e" The firms ofthe responding IT managers do not significantly differ from the non-responding firms in terms of firm performance and firm size. On average, the interviewers had to contact a firm 9.8 times to generate one interview. A production manager interview lasted on average 40.9 minutes, and an IT manager interview lasted on average 13.4 minutes. Both average durations provided enough time to reasonably work through the questionnaires. I chose manufacturing firms in order to concentrate on one questionnaire and thus avoid any bias from differences between corporate strategies. organizational structures, HRM practices, and other firm characteristics that are required for the generation of products versus services. German and Polish firms were chosen for a practical reason, The remaining 581 firms were not contacted either because t hey did not exist anymore, t hey had no manufacturing establishments in Germany or Poland, no telephone number could be identified, or the interviewers were not able to get in t ouch with the firms before the end ofthe projec t 37

38 This is based on a logistic regression with a binary dependent variable that indicates response. For a one-unit increase in t he number of firm employees, t he odds ofbeing a responder increased by a factor of

1.0001.


112

4 Hybrid strategy and firm performance

namely, the availability of native speakers as interviewers. I chose production managers as potential respondents because they are typically in upper middle management and thus are able to deliver insights into corporate strategy as well as organizational structure and HRM practices. Similarly, IT managers are able to provide a full overview of how leT is used in their firm. Further; through the narrow set of informants, singleInformant bias was held relatively constant. The surveyed establishments employed on average 68.5 percent of a firm's total emplcyees. indicating that the obtained measures are largely representative of the firms. Most of the survey items were derived and adapted from prior studies, as described below. The survey instrument for production managers was discussed at length with one production manager and then pretested in Interviews with 15 additional production managers. The survey instrument for IT managers was discussed in detail with one IT manager and two PhD students from the field of infonnation systems research. Subsequently, the survey instrument was pretested in Interviews with ten more IT managers. The pretests led to slight adaptations in the survey instruments. The questionnaires for Pclish production and IT managers were translated from German into Polish by an accredited translator and then checked by two Polish managers and two Polish research analysts. As high-quality data relies on the capabilities of interviewers, several steps were undertaken to ensure data quality. The student interviewers came from management; economics. and sociology and were chosen in a two-step assessment procedure. Selection criteria were prior experience with the topics of the survey and data-gathering in general as well as the performance in a simulated interview situation. The interviewers were intensively trained in survey background, method, questions, and software during a two-day workshop that also included several simulated Interviews. Interview quality was further enhanced through regular team meetings and through interview monitoring by two PhD students. Corporate strategy is available from the 2008 cross-section, which is matched with two time-series datasets that cover the period from 2005 to 2008. As corporate strategy is the key component of this study, two measures are undertaken to avoid any bias from strategy changes between 2005 and 2008. First, respondents were asked to indicate the

firm's realized strategy during the last three years rather than the intended strategy as in several prior studies on ambidexterity (Cao et al., 2009; He and Wong, 2004). Intended strategies are often not fully realized (Mintzberg, 1978) and are thus less useful in assessingwhether a firm's organizational architecture fits its strategy. Second.


4.3 Method

113

respondents were asked for strategy changes in the last three years. If a respondent indicated that the corporate strategy had strongly or very strongly changed during this period, all observations of this firm are excluded from the analysis. Hypotheses 4.1 to 4.4 are tested with regard to individual and technological ambidexterity. I created two separate samples for individual and technological ambidexterity to reduce the loss of observations due to missing values to an acceptable level. A sample of 258 firms, with each firm represented by at least one of 457 observations, is available for the analysis of individual ambidexterity. A sample of 302 ftrms, with each firm represented by at least one of 502 observations, is available for the analysis of technological ambidexterity. The firms in both samples do not significantly differ from all other German and Polish manufacturing firms included in the eITDB and ORBIS with regard to firm performance, but they are slightly smaller.e'' This is notably unproblematic, however, as I am especially interested in the moderating role of individual ambidexterity, which is seen as the appropriate form of organizational ambidexterity for smaller firms (Gibson and Birkinshaw, 2004; Raisch and Birkinshaw, 2008). The firms in the individual ambidexterity sample have on average 423 emplcyees, while the firms in the technological ambidexterity sample have on average 406 employees.

4.3.3 Measures 4.3.3.1 Firm performance To test hypotheses 4.1 to 4.4, I measure firm performance as the return on total assets (ROA), which is one ofthe most widely-used firm performance measures in analyses of corporate strategy and organizational architecture (Burton et al., 2002). I obtain ROA from ORBISand calculate it as (profit or loss before taxation / total assets) x 100. To test hypotheses 4.1 to 4.4 with Porter's (1980; 1985) performance definition, measure firm performance using the dummy variable ABOVE. ABOVE is coded 1 if a

firm's ROAat time t lies above the average ROAof the manufacturing sector in which the firm operates. ABOVE is coded 0 if the firm's ROA at time t is equal to or below the 39 This is based on a logistic regression with a binary dependent variable indicating sam pie inclusion. For a one-unit increase in the number of firm employees, the odds of being in the sampie for individual ambidexterity changed by a factor of 0.9998; that is, the odds decreased. Similarly, for a one-uni t increase in the number of firm employees, the odds of being in the sampie for t echnological ambidexterity changed bya factor ofO.9997.


114

4 Hybrid strategy and firm performance

average ROA of the manufacturing sector. I assign finns to manufacturing sectors based on the two-digit Standard Industrial Classification (SIe) codes provided by ORBIS and the eITDE.

4.33.2 Corporate strategy Measures f端r the extents to which an exploration strategy and an exploitation strategy are pursued are obtained from my survey on production managers. Respondents were asked to indicate on a Likert scale from 1 CI do not agree at alt') to 5 CI stronglyagree')

if several items have been a key part of the respective firm's strategy during the last three years. Note again that the foeus lies on the realized strategy rather than on the intended strategy. I derived and adapted all items from prior surveys on strategy typologies or ambidexterity (Dess and Davis, 1984; Doty et al., 1993; He and Wong, 2004; Kotha and Vadlamani, 1995; Lubatkin et al., 2006; Miller and Toulouse, 1998; Segev, 1989). The items that measure exploration focus on the exploration of new and unique product market domains, while the items that measure exploitation focus on efficiency and improvement in existing product market domains. The exploratcry items and the exploitative items clearly load on different factors in a joint exploratory factor analysis with varimax rotation. Factor loadings are above .53 with cross-loading below .34. Thus, I use the average of the exploratory items to create a measure of exploration. Further; I use the average of the exploitative items to create a measure of exploitation. Cronbach's alpha for both measures is .76 and thus above the cut-off point of .70 (Nunally, 1978). The survey instrument is presented in Appendix 4.1. Measures for a hybrid strategy have been constructed by multiplying (Gibson and Birkinshaw, 2004; Venkatraman et al., 2007), dividing (Uotila et al., 2009), subtracting (He and Wong, 2004; Cao et al., 2009), and adding [lausen et al., 2009; Lubatkin et al., 2006) measures for exploration and exploitation. To test hypothesis 4.1, I choose the additive model for theoretical and practical reasons. Interacting exploration and exploitation incorpcrates the implicit assumption that both are complementary, that Is, mutually reinfcrcing. In contrast, consistent with the outlined strategy typologies. I believe that exploration and exploitation are fundamentally different activities that even require separate or complex organizational arrangements in order to be combined within one firm. Dividing and subtracting exploration and exploitation is based on the idea that emphasizing either exploration or exploitation is only possible at the expense of the other. Though this perspective is valid, achieving a balance between exploration and exploitation is fine-tuning. In the first


4.3 Method

115

instance, a hybrid strategy requires that both exploration and exploitation are pursued to a sufficiently high degree.s'' I interpret exploration and exploitation as orthogonal activities (Gupta et al., 2006). A firm can engage in high levels of both activities at the same time, provided that suitable organizational arrangements are in place. Further; two studies have shown that an additive approach has had greater explanatory power than the other perspectives [lausen et al., 2009; Lubatkin et al., 2006). Consequently, to test hypothesis 4.1, I measure hybrid strategy as the average of exploration and exploitation.s!

4.3.3.3 Individual ambidexterity

All measures for individual ambidexterity are obtained from my survey on production managers. Decentralization is measured as the average of the degree of decentralization of various organizational decision rights, ranging from lower-level decisions such as the pace of the work of production workers to high er-level decisions such as product pricing (a=.88). I derived and adapted all items from several prior studies on decentralization (Andersen and [onsson, 2006; Bloom et al., 2009a; Burton et al., 2002; Colombo et al., 2007; Hanks et al., 1993; Lee and Grover, 2000; Nahm et al., 2003; Pugh, 1973). The survey instrument is presented in Appendix 4.2. For each decision right, the respondents were asked to indicate the hierarchical level that is responsible for making the decision. Respondents were also able to indicate if adecision was collectively made by two hierarchical levels. For example, a decision that is made by the highest hierarchical level was assigned the value 1. Adecision that is jointly made by the highest and the next lower hierarchical level was assigned the value 1.5. If a decision is collectively made, but with unequal weights, the values 1.3 and 1.7 were also possible. As many firms have different numbers of hierarchical levels, the respondents were asked to indicate the total Cao et al. (2009) find tha t if a measure for the combined dimension of exploration and exploitation and a measure for the balance dimension of exploration and exploitation enter the same regression, only the combined dimension positively affects firm performance. They further find t hat an interaction term of both dimensions positively affects firm performance. I interpret this finding as evidence for my assumption that a balance between exploration and exploitation is complemen tary rather than equal t o pursuing high levels ofboth exploration and exploitation as a means to create a successful hybrid strategy.

40

Obviously, if computing the hybrid strategy measure as the average of exploration and exploitation, a firm with an exploration score of 5 and an exploitation score of 1 and another firm with an exploration score of 3 and an exploitation score of 3 are assigned t he same hybrid strategy score, namely 3. However, one could argue tha t the former firm is a pure explorer and that only t he latter firm pursues an actual hybrid strategy. To test for bias from t his, I employ two robustness tests, as described below.

41


116

4 Hybrid strategy and firm performance

number of hierarchical levels in their firm. Subsequently, every item was standardized so that 0 means that the decision is made at the highest hierarchical level of the firm (that Is, fully centralized) and 1 means that the decision is made at the lowest hierarchical level of the firm (that Is, fully decentralizedj.s? Standardization is measured as the average of two five-point Likert scale items that are derived and adapted from prior studies (a=.56) (Andersen and [onsson, 2006; Burton et al., 2002; Hanks et al., 1993; Nahm et al., 2003). Respondents were asked to indicate how many written or oral rules and procedures are in place in their firm. The survey instrument is presented in Appendix 4.3. Employee appraisal, screening, and training are measured by an innovative interview method that is adapted from Bloom and Van Reenen (2007). The method incorporates open questions and open answers, which are then transformed into numeric values from 1 to 5 by the interviewer. This method allows for much more detailed insights into HRM practices than is possible with rather short Likert scale items. Further, as respondents are unaware that their open answers are transformed into numeric values, the method mitigates the problem of social desirability, which is highly important in the context of HRM practices. For example, some respondents might give unrealistically optimistic answers to questions about the amount or contents of employee training if they feel that these HRM practices are generally considered 'goOd.'43 The interviewers were provided with prepared questions; they began with general questions and continued with more specific questions based on the prior answers of respondents. The interviewers were also encouraged to deviate from the prepared questions if needed. Thus. a conversation led by the interviewer took place regarding every HRM practice. The dialeg was not ended until the interviewer felt that she or he had a full picture of the HRM practice in question and was able to decide on a numeric score. In particular, the interviewers were asked to indicate on a scale from 1 ('not Not e that two items measure t he hierarchicallevel at which t he decisions on rules and procedures are made. The participation of employees in the development of rules and procedures is an indicator of t he enabling rather than coercive character of standardization (Adler et al., 1999). The average of the two items on rules and procedures positively correlates with the overall decentralization measure (p=0.5492; p-value=O.OOOO). This indicates that firms with individual ambidexterity, that is, high levels of both decentralization and standardization, allow their employees to participate in the development ofrules and procedures, leadingto enabling rather than coercive standardization.

42

The Human Subjects Committee of Stanford University approved this method for the study of Bloom and Van Reenen (2007), because the unawareness of respondents with respect t o being scored is necessary to reduce t he problem of social desirability; in addition, this lack of awareness is not risky for t he respondents and their employers due t o data confidentiality, and moreover, it is temporary as the respondents are debriefed after the end ofthe projec t (Bloom and Van Reenen, 2007).

43


4.3 Method

117

important at all') to 5 ('very important'] the importance of both the explcratory and the exploitative components with respect to employee appraisal, screening. and training. Thus, six measures were generated, namely, three regarding the importance of exploratory components in employee appraisal, screening, and training and three regarding the importance of exploitative components in employee appraisal, screening. and training. The survey instruments for HRM practices were developed specifically for this study but nevertheless share broad similarities with prior studies on HRM practices that are appropriate for different corporate strategies [lackson et al., 1989; Schuler and [ackson, 1987; Youndt et al., 1996). In short, the survey instruments measured to which extent employee appraisal, screening, and training are targeted at subjective perfonnance criteria and broad skills as opposed to objective performance criteria and specific functional skills. The survey instrument is presented in Appendix 4.4. To test the inter-rater reliability of this survey method, a second interviewer listened to 31.3 percent of all production manager interviews (that Is, 245 interviews] and assigned scores to the respondent answers independently from the first interviewer.s- The partial correlation coefficients-S of the first interviewer and the second interviewer scorings for the six HRM variables created using this interview method range from 0.6292 to 0.8088 and were highly significant (p-values=O.OOOO). This result implies that the survey method leads to relatively homogenous results, even if different interviewers score the same interview. I use the average of the decentralization measure and the three measures for the exploratory components of HRM practices to create a measure for exploratory structures and HRM practices. I use the average of the standardization measure and the three measures for the exploitative components of HRM practices to create a measure for exploitative structures and HRM practices. Finally, I compute individual ambidexterity as [exploratory structures and HRM practices + exploitative structures and HRM practices] /2. Alpha reliabilities for the decentralization and standardization measures are provided above. Tests for the internal consistency of the measures for explcratory structures and HRM practices, exploitative structures and HRM practices, and individual ambidexterity 44

In t he multivariate analyses below, only first interviewer scorings are used.

45

This is controlling for first interviewer and second interviewer fixed effects.


118

4 Hybrid strategy and firm performance

are not appropriate, as these measures are designed to capture different aspects of a construct rather than multiple measures of a construct. Thus. there are no specific expectations about the intercorrelations between the aspects; they might correlate positively, negatively, or not at all (Aral and Weilt, 200?, Diamantopoulus et al., 200S, [arvis et al., 2003).

4.33.4 Technological ambidexterity All measures f端r technological ambidexterity are obtained from the eITDE. The ClTDB provides binary variables that indicate if specific hardware and software products are used in the observed establishments. F端r example, the eITDB would indicate ifExchange Server from Microsoft or Lotus Domino from IBM are in place in an establishment. Harte-Hanks groups these products into broader dasses of hardware and software products. For example, Exchange Server and Lotus

Domino, which are both

collaboration software products, would be assigned to the groupware dass. I use these groupings to obtain measures for the four lCT types that are outlined above. I create a groupware dummy variable that is coded 1 if any collaboration software, office automation software, or other groupware software is in place in an establishment. The corporate intranet dummy is coded 1 if any frame-relay, leased-line, or wide-area network is in place in an establishment, as these hardware types are typically employed for corporate intranets (Bloom et al., 2009). I create an ERP application dummy and a workflow application dummy in a similar manner; as described in Table 4.1 (next page). As all information in the CITDE pertains to the establishment-level, I aggregate the four dummies to the finn-level. For example, if any of the observed establishments of a firm possesses a groupware application, the finn-level groupware dummy is coded 1. Note that the establishments surveyed by Harte-Hanks employed on average 85.6 percent of a

firm's total employees in the individual ambidexterity sample and 85.2 percent of a firm's total employees in the technological ambidexterity sample, indicating that the obtained measures are largely representative of the finns.


4.3 Method

119

Table4.1 JeT types Exploratory ICT Groupware applications

Collaboration software (e-mail, shared calendars, shared document storage and management, etc.) Officeautomation software Other groupware software

Corporate intranets

Frame-relay Leased-line Wide-area network

Exploitative ICT ERPapplications

Accounting software Human resource managemen t software Supply chain managemen t software Customer relationship management software Other ERPsoftware

Workflowapplications

Application server software Content management software Document managemen t software Enterprise application integration software Portal managemen t software

Subsequently, the average of the groupware and corporate intranet dummies is used to create an exploratcry ICT variable. The average of the ERP application and workflow application dummies is used to create an exploitative ICT variable. To support the abovementioned assumptions regarding the use of groupware applications and corporate intranets to support explcratory activities and the use ofERP and workflow applications to support exploitative activities, information from my survey on IT managers is used. However, these data allow me to check my assumptions exclusively with regard to the relationship between ICT and structural ambidexterity. In particular, IT managers were asked to indicate to what extent the ICT in their firm is used to support decentralization and centralization, that Is, the organizational structures needed for separate exploratory and exploitative units. I measure 'ICT use for decentralization' as the average of several items regarding the extent to which ICT is used to enable autonomous decisions by workers (a=.81). 'ICT use for centralization' is measured as the average of several items regarding the extent to which ICT is used to enable better management decisions (a=.76). Both measures have been developed for this study and were validated through pretests, as described above. The survey Instrument is presented in Appendix 4.5. A joint exploratory factor analysis of all items clearly shows two distinct factors. The varimax rotated factor loadings are above.53 with cross-lcadings below.30.


4 Hybrid strategy and firm performance

120

Table 4.2 shows the correlations among the measures für exploratcry leT, exploitative ICl', leT use für decentralization, and leT use für centralization. Note that to be able to infer causality, exploratory leT and exploitative leT are measured für the year 200?, while leT use für decentralization and leT use für centralization are measured für 2 DOS. As shown, the availability of exploratory leT types significantly correlates with leT use für decentralization, while the availability of exploitative leT types significantly correlates with leT use für centralization. However, the correlation coefficients are small and the significance levels low.se Though I find only weak support für my assumptions, it is worth mentioning that the availability of exploratcry ICT types does not significantly correlate with ICT use for centralization and the availability of exploitative ICT types does not significantly correlate with ICT use for decentralization.

Table 4.2 Correlation between ICT types and ICT use ICTuse for decentralization

ICTuse for centralization

Exploratory ICT

0.0876*

-0.0185

Exploitative ICT

0.0773

0.0846*

* p < 0.1; exploratory ICT and exploitative ICT from 2007; ICT use for decentralization and ICT use for centralization from 2008; between 423 and 426 observations

4.3.3.5 Group indicator variables To create effects-coded and dummy-coded group indicatcr variables, I assign all firms in the individual and technological ambidexterity samples to one group; I employ the two dimensions of corporate strategy and organizational architecture to differentiate between the groups. Consistent with other studies, I use the median-eutoff criterion to assign the firms to groups (He and Wong, 2004). In the individual ambidexterity sample, I assign a firm to the group of firms with a hybrid strategy and organizational ambidexterity if (1) the firm's value for the exploration measure is equal to or greater than the median of the exploration measure; (2) the firm's value for the exploitation measure is equal to or greater than the median of the exploitation measure; (3) the firm's value for the measure for exploratory structures and HRM practices is equal to or greater than the median of this measure; and (4) the firm's value for the measure for exploitative structures and HRM practices is Note that the low correlations may result from the fact tha t the survey on IT managers was not originally designed as a cross-check for the exploratory ICT measure and the exploitative ICT measure. Thus, respondents were asked about general ICT use rather than how the specific ICT types of interest (groupware applications, corporate intranets, ERPapplications, and workflow applications) are used.

46


4.3 Method

121

equal to or greater than the median of this measure. In contrast, a firm is assigned to the group of firms with a hybrid strategy but without organizational architecture if (1) and (2) apply but if either (3) or (4) or both do not apply. I assign a firm to the group of firms with a pure strategy and a pure organizational architecture if (1) the firm's value for the exploration (or exploitation] measure is equal to or greater than the median of this measure; (2) the firm's value for the exploitation (or exploration) measure is below the median of this measure; and (3) the firm's value for the measure for exploratcry (or exploitative] structures and HRM practices is equal to or greater than the median of this measure. In contrast, a firm is assigned to the group of firms with a pure strategy but without a pure organizational architecture if (1) and (2) apply but (3) does not. I assign a firm to the group of finns with a no-emphasis strategy if the firm's values for both the exploration measure and the exploitation measure are below the medians of the respective measures. The group assignments in the technological ambidexterity sample are made analogously, as shown in Table 4.3. Table 4.3 Groups in the individual and technological ambidexterity samples Groups in ind. ambid. sam pie

Firms

Obs.

Hybrid strategy and individual ambidexterity

27

45

Firms

Obs.

Hybrid strategy and technological ambid.

53

90

Hybrid strategy

64

109

Hybrid strategy

57

87

Pure strategy and pure structuresjHRM

62

113

Pure strategy and pure lCT

91

166

Pure strategy

49

86

Pure strategy

37

55

No-emphasis strategy

56

104

No-emphasis strategy

68

122

258

457

Total

306

520

w j 0 individual ambidexterity

w j 0 technological ambid .

w j 0 pure structuresjHRM

Total

Groups in techn. ambid. sampie

wjo pure lCT

Note t hat a total of 306 firms for the technological ambidexterity sampie [> 302 as stated in regressions) results from firms tha t have changed groups over time through changes in their lCT adoption

4.3.3.6 Contral variables

The measure for environmental dynamism is obtained from my survey on production managers and constructed in the style of two prior studies (Daft et al., 1988; May et al.,


122

4 Hybrid strategy and firm performance

2000). Respondents were asked to indicate on scales from 1 to 5 the rate of change, predictability, and importance of seven areas of the environment [customers/market, competitcrs, technology, labor market/unions, capital market, legislators/regulatcrs, and the socio-cultural environment]. I defined rate of change as the magnitude and speed of changes in trends, demands, and conditions that occurred during the last three years in the respective area. Predictability was defined as the ability to anticipate these changes and their consequences f端r the firm during the last three years. Importance was defined as the relevance ofthe respective area f端r a firm's goals and performance. The survey instrument is in Appendix 4.6. Following Daft et al. (1988) and May et al. (2000), I calculated the dynamism in each of the seven environmental areas as (rate of change + predictability] x importance. I used the average of the seven resulting dynamism variables to create a measure of the overall environmental dynamism (a=.74). A dummy variable to account for data limitations is included in all regressions. If the measure for environmental dynamism is not available for a firm due to missing values, the environmental dynamism variable is coded 0, while the environmental dynamism missing dummy is coded 1 (Bloom and Van Reenen, 2007; Uotila et al., 2009).47 ROAat time t-1 is measured as outlined above for ROAat time t. Firm-level control variables are firm size, measured as total number of employees, and firm age, measured as years in operation. Both measures have been associated with inertia and several other effects that may impact firm performance, corporate strategy, or organizational architecture (He and Wong, 2004; Lubatkin et al., 2006; Spanos et al., 2004). The measures are obtained from ORBlS and log-normalized to compensate for skewness. Further. a dummy variable to account for the firm's nationality (that Is, German or Polish) is included in all analyses. I use additional dummy variables to account for industry effects on firm performance, corporate strategy, and organizational architecture. Particularly, I control for the firm's core industry membership at the one-digit SIC code level-" and for the firm's manufacturing sector membership at the two-digit SIC code level. The SIC codes are obtained from ORBlS and the CITDB.

The environmental dynamism variable is missing for 27 observations in the individual ambidexterity sampie (5.9 percen t of observations) and for 47 observations in the t echnological ambidexterity sampie (9.0 percent of the observations).

47

As my survey focused on manufacturing firms, manufacturing (one-digit sie codes 2 and 3) is the core industry for most observations in both sampies (80.7 percent of the observations in t he individual ambidexterity sam pie and 81.5 percent ofthe observations in the technological ambidexterity sam pie).

48


4.3 Method

123

Finally, year dummies are included to control for time effects. Descriptive statistics for both samples are provided in Table 4.4 and Table 4.5.

Table 4.4 Descriptive statistics for the individual ambidexterity sample Variable

Firms

Obs.

Median

Mean

Std. dev

Min

M",

ROA (t)

258

457

8.9585

10.0292

14.2045

-33.3887

136.0431

RGA

258

457

7.9589

9.4653

13.3181

-33.7865

87.8834

Firm size (ln)

258

457

5.5134

5.5852

0.9428

1.3863

8.2721

Firm age (ln)

258

457

3.2958

3.2403

0.9473

0

5.2149

Environmental dyn.

258

457

23.5000

22.4032

7.3501

0

39.6250

Hybrid strategy (me)

258

457

-0.0209

0

0.5170

-1.8334

1.2708

lnd. ambid. (me)

258

457

0.0532

0

0.4527

-1.6933

1.1426

Hybrid strategy (me) x ind. ambid. (me)

258

457

0.0106

0.0367

0.2514

-0.6263

1.5418

Firm size (ut)

258

457

248

423.1745

504.0048

4

3,913

Firm age (ut)

258

457

26

37.1422

35.4999

0

183

(ln) naturallogarithm; (me) mean eentered; (ut) untransformed, not used in multivariate analysis; all variables are at time t -1 if not otherwise stated.

Table 4.5 Descriptive statistics for the technological ambidexterity sample Variable

Firms

Obs.

Median

Mean

Std. dev.

Min

M",

ROA (t)

302

520

7.7171

8.6523

12.7380

-36.8692

79.9198

RGA

302

520

7.3865

8.2637

13.3725

-55.8324

87.8834

Firm size (ln)

302

520

5.5134

5.5492

0.9304

1.3863

8.2721

Firm age (ln)

302

520

3.2958

3.2064

0.9410

0

5.2149

Environmental dyn.

302

520

23

21.4178

8.2368

0

39.6250

Hybrid strategy (me)

302

520

0.0201

0

0.5234

-1.8201

1.2840

Teehn. ambid. (me)

302

520

0.1409

0

0.1735

-0.6091

0.3909

Hybrid strategy (me) x teehn. ambid. (me)

302

520

.0064

-.0032

.0887

-.4586

.3981

Firm size (ut)

302

520

248

405.7591

497.0403

4

3,913

Firm age (ut)

302

520

26

36.0231

35.2535

0

183

(ln) naturallogarithm; (me) mean eentered; (ut) untransformed, not used in multivariate analysis; all variables are at time t -1 if not otherwise stated.


124

4.4

4 Hybrid strategy and firm performance

Results

4.4.1 Moderation effects models Hypothesis 4.1 that the hybrid strategy-performance relationship is positively moderated by organizational ambidexterity is tested with the moderation effects model 4.1. To mitigate the potential for multiccllinearity, I mean centered the linear terms of hybrid strategy and organizational ambidexterity before obtaining their product (Aiken and West, 1991; Jaccard and Turrisi, 2003). Multiccllinearity concerns are further diminished because the variance Inflation factors (VIFs) of all regressions for individual ambidexterity are never above 3.44, while the VIFs of all regressions for technological ambidexterity are never above 4.34. Thus, all VIFs are below the cut-off value of 10 (Cohen et al., 2002). The crdinary least square (OLS) regression results for individual ambidexterity are reported in columns (1) and (2) of Table 4.6 (next page). The regression results in column (1) show that a hybrid strategy does not significantly affect firm perfonnance if the hybrid strategy variable enters the regression in Isolation. However, the interaction term of hybrid strategy and individual ambidexterity in column (2) is positive and significant. This supports hypothesis 4.1 that the effect of hybrid strategies on firm performance is positively moderated by organizational ambidexterity. Note that now also the overall marginal effect of a hybrid strategy on firm performance is positive. The results fortechnological ambidexterity in columns (3) and (4) confirm this finding. Thus. hypothesis 4.1 is supported for both individual and technological ambidexterity.


4.4 Results

125

Table 4.6 Test ofhypothesis 4.1 Dependentvariable Sampie

RaA (t)

Individual ambidexterity (1) (2)

Technological ambidexterity (3) (4)

Firm size (ln)

-0.1047 (0.4271)

-0.2742 (0.4293)

0.0045 (0.3963)

0.0713 (0.4108)

Firm age (ln)

0.2832 (0.4111)

0.1965 (0.3980)

0.5854 (0.3921)

0.5641 (0.3993)

0.8051*** (0.0764)

0.8019*** (0.0768)

0.6222*** (0.0762)

0.6198*** (0.0767)

Environmental dynamism

0.0277 (0.0887)

0.0408 (0.0900)

0.0010 (0.0966)

0.0141 (0.0981)

Hybrid strategy

1.1801 (0.8829)

1.2929 (0.8919)

0.7321 (0.8988)

0.7011 (0.8756)

RaA (t-1)

Individual ambidexterity

1.2142 (0.9168)

Hybrid strategy x individual ambid.

2.8382** (1.3339) -0.1827 (2.6043)

Technological ambidexterity

12.6541*** (4.6075)

Hybrid strategy x t echnological ambid. Further controls Firms Observations

Industry, manufactunng sector, and year dummies: coun try dummy; environmental dynamism missing dummy 258

258

457

457

302 520

302 520

R'

0.61

0.62

0.50

0.50

* p < 0.1; ** P < 0.05; *** P < 0.01; time period 2005-2008; aLS regressions used in all columns; standard errors in parentheses clustered by firm and Huber-White robust to heteroskedasticity and autocorrelation of unknown form

A graphical representation of the results in Table 4.6 is provided in Figure 4.4 (next page). The impact of a hybrid strategy on firm performance is negative for low levels of individual and technological ambidexterity (see the dashed lines). Instead, the impact of a hybrid strategy on firm performance is positive for high levels of individual and technological ambidexterity (see the solid lines).


4 Hybrid strategy and firm performance

126

-2

-1

1-----

0 hybridily er slralegy low ind. ambid.

2

- - - high ind. ambid.

~

~

2

=

~o

â&#x20AC;˘c ,Eu;:' 0

;;

., ., 0

~

-2

-1

0

hybridily er slralegy

I-----Iowtechn.ambid.

2

- - - high lechn.ambid.I

Figure 4.4 The moderating role of individual and technological ambidexterity in the hybrid strategy-performance relationship Note that I employ two robustness tests fĂźr hypothesis 4.1 with alternative measures of hybrid strategy, individual ambidexterity, and technological ambidexterity. Obviously, if computing the hybrid strategy measure as the average of exploration and exploitation, a firm with an exploration score of 5 and an exploitation score of 1 and another firm with


4.4 Results

127

an exploration score of 3 and an exploitation score of 3 are assigned the same hybrid strategy sccre, namely 3. However; one could argue that the former firm is a pure explorer and only the latter firm pursues an actual hybrid strategy. The analogous reasoning applies to the individual ambidexterity and technological ambidexterity measures. To mitigate bias from this and ensure that only firms with an actual hybrid strategy, an actual ambidextrous organizational architecture, and actually ambidextrous ICl' are considered in the analysis, I recode the hybrid strategy, individual ambidexterity, and technological ambidexterity measures to 0 if at least one of their respective two components is below its median value. The results presented in Table 4.6 hold for both individual and technological ambidexterity (not reported). Also, using a dummy-coding model, as described above, I employ subgroup analysis as an alternative means to test hypothesis 4.1; that Is, I compare the average perfonnance of firms with a hybrid strategy and organizational ambidexterity with the average performance of firms with a hybrid strategy but without organizational ambidexterity. As outlined above, the group indicator variables for the dummy-coding models are created by the median-eutoff criterion, meaning that only finns with an actual hybrid strategy are coded as such a firm. The results support hypothesis 4.1 for individual and technological ambidexterity, are shown in Table 4.7 (next page), and described in more detail below.

4.4.2 Effects- and dummy-coding models Hypotheses 4.2 to 4.4 are tested with effects- and dummy-coding models 4.2, 4.3, 4.4, and 4.5. The OLSregression results for individual ambidexterity are reported in columns (1) to (5) of Table 4.7, and the results for technological ambidexterity are presented in columns (6) to (10). The model and the reference group employed in every column are indicated at the top of the table. Note that dummy- and effects-coding are alternative methods but mathematically equivalent (Hardy, 1993; Spanos et al., 2004). As the same data are used in columns (1) to (5), they yield an identical R2 and identical coefficients for all variables except for the group indicator variables. This also applies to columns (6) to (10).


4.13 59*** (1.1560)

4.7970*** (1.6420)

2.34 47* (1.351 9) 3.2392** (1.3687]

2.9001** (1.1450)

-1.8968** (0.8615)

0.4478 (0.82 87]

1.3424 (0.830 1)

Hybrid st ra t egy and ind./ t echn . ambidex terity

w 10 ind./techn. ambidexterity

Pure str ategy and pure str uct ures /HRM/ICT

w/o pure struct ures/HRM/IC T

Ind ust 258 45 7 0.63

Fur ther controls Firms Observations R'

-3 .2413*** (1.155 6)

0.8946 (1.2456)

-2.3447* (1.3519)

2.4523 (1.6280)

0.0 588 (0.08 18)

0.7966*** (0.079 1)

-4.1359*** (1 .15 60)

-0.89 46 (1.2456)

-3.2392** (1.3687)

1.5578 (1.6173)

0.0 588 (0.08 18)

0.79 66*** (0.079 1)

0.3067 (0.3793)

-0.4343 (0 .4 11 4)

0.1198 (1 .05 49)

2.0830*** (0.6649)

-1.9979** (0.8548)

-0.2872 (1.2342)

2.1176 (1 .5511)

4.0809*** (1 .1855)

4.0780*** (1.5141)

-0.0 10 1 (0.09 12)

-0.0 10 1 (0.09 12) 2.0802* (1.0617)

0.6 112*** (0.0759)

0.548 2 (0.3935)

-0.2745 (0.4 146)

0.6 112*** (0.0759)

0.548 2 (0.3935)

-0.2745 (0.4 146)

2.4048* (1.3850)

4.3681 *** (1.02 18)

0.2872 (1.2342)

4.3652*** (1.5145)

-0.010 1 (0.09 12)

0.6 11 2*** (0.0759)

0.5482 (0.3935)

-0.27 45 (0.4 146)

-4.368 1*** (1 .021 8)

-1.9632 (1.4009)

-4.0809*** (1.1855)

-0.0028 (1.3311)

-0.0 10 1 (0.09 12)

0.6 11 2*** (0.0759)

0.548 2 (0.3935)

-0.2745 (0 .4146)

30 2 52 0 0.52

-2.4048* (1.3 850)

1.96 32 (1.4009)

-2.1176 (1.5511)

1.9604 (1.8032)

-0.0 10 1 (0.09 12)

0.6 112*** (0.0759)

0.5 48 2 (0.3935)

-0.2745 (0 .4 146)

Technological ambidexteri ty Dumm Dumm Dumm v Dumm Hybrid strat. No-emphasls Pure strategy Pure strategy w 10 t . ambid. stretegy and oure ICT w/o oure ICT 7 8 9 10

,manufact urin sector, and ear d ummies; countr dumm; environmental d narmsm missin du mm 302 302 258 258 258 302 302 45 7 45 7 45 7 52 0 52 0 52 0 52 0 0.63 0.63 0.63 0.52 0.52 0.52 0.52

0.8967 (1.1159)

5.6936*** (1.5888)

0.0 588 (0.08 18)

0.7966*** (0.0791)

0.3067 (0.3793)

-0.4343 (0.4 114 )

Effects Entire sam ule 6

* P < 0.1; ** P < 0.05; *** P < 0.0 1; tim e penod 2005-2 008; OLS regressions used m all columns: standard errors m parent heses clustered by firm and Huber-Whire robust t o het ero ske dast icity and a utocorrel ation of unkn own for m.

-0.8967 (1 .115 9)

No-em phasis str ategy

Pure strategy

Hybrid strategy

258 45 7 0.63

3.2413*** (1.15 56)

0.0588 (0.08 18)

0.0 588 (0.08 18)

Environment al

dy namis m

0.7966*** (0.079 1)

0.79 66*** (0.079 1)

ROA (t-l)

0.3067 (0.3793)

0.3067 (0.3793)

0.3067 (0.3793)

-0.43 43 (0.4 114 )

Firm age (In)

-0.4343 (0 .4 11 4)

-0.4343 (0 .4 11 4)

Entire sam ule 1

Effe-ts

ROA Indi vi d ual ambidexterity Dumm Dumm Dumm Dumm Hybrid strat. No-emphasis Pure stretegy Pure strategy w/oi. ambid. st rateev and oure st r, w/o o ure st r. 2 3 4 5

Firm si ze (In)

Reference gro up

Codin

Sample

De en dentvari able

Table 4.7 Test ofhypotheses 4.1 to 4.4


4.4 Results

129

Column (1) shows the results for effects-coding model 4.2 and pertains to hypothesis 4.2 that firms with a hybrid strategy and with [without] organizational ambidexterity outperform [underperform] the entire sample of firms on average. As shown, the average performance of the group of firms with a hybrid strategy and individual ambidexterity is significantly higher than the sample average performance, that is. the average of the average perfonnances of all groups of finns (2.9001**). Instead, the average perfonnance of firms with a hybrid strategy but without individual ambidexterity is significantly lower than the sample average performance (-1.8968**). These results hold for technological ambidexterity, as shown in column (6). In summary, hypothesis 4.2 is supported for individual and technological ambidexterity. Column (2) employs dummy-coding and uses firms with a hybrid strategy and without organizational

ambidexterity as

the

reference

group. Although

not formally

hypothesized, column (2) gives further support for hypothesis 4.1 on the moderating effect of organizational architecture in the hybrid strategy-performance relations hip. In fact, the subgroup analysis employed in column (2) is an alternative to the moderation effects model 4.1 in Table 4.6 for testing the moderation approach (Venkatraman, 1989). According to column (2), firms with a hybrid strategy and individual ambidexterity have a significantly higher average performance than finns with a hybrid strategy but without individual ambidexterity (4.7970***). Thus, hypothesis 4.1 is again fully supported for individual ambidexterity. The finding holds for technological ambidexterity in column

(7). Column (3) shows dummy-coding model 4.3 and concerns hypothesis 4.3 that firms with a hybrid strategy and with [without] organizational ambidexterity outperform (match) ftrms, with a no-emphasis strategy on average. Indeed, the coefficient for the group of finns with a hybrid strategy and individual ambidexterity is positive and significant (5.6936***), indicating that these firms outperform firms with a no-emphasis strategy on average. In contrast, the coefficient for the group of firms with a hybrid strategy but without individual ambidexterity is insignificant, indicating that on average, there are no significant performance differences to the firms with a no-emphasis strategy (0.8967). These results hold for technological ambidexterity, as shown in column (8). Thus. hypothesis 4.3 is supported for individual and technological ambidexterity. Columns (4) and (5) show models 4.4 and 4.5 and pertain to hypothesis 4.4, which states that firms with a hybrid strategy and with (without] organizational ambidexterity match (underperform) pure strategies on average. As outlined above, the group of firms with a pure strategy is split into firms with or without pure structures and HRM practices to


130

4 Hybrid strategy and firm performance

allow f端r more subtle performance comparisons. Firms with a pure strategy and high levels of pure structures and HRM practices might display a higher average perfonnance

than firms with a pure strategy and with low levels of pure structures and HRM practices. Column (4) uses firms with a pure strategy and pure structures and HRM practices as

the reference group. As shown, firms with a hybrid strategy and individual ambidexterity do not displaya significantly different average performance than the reference group (2.4523). However, firms with a hybrid strategy but without individual ambidexterity have a significantly lower average perfonnance than firms with a pure strategy and pure structures and HRM practices (-2.3447*). These findings hold iffirms with a pure strategy but without pure structures and HRM practices are used as the reference group in column (5). Thus, hypothesis 4.4 is supported for individual ambidexterity. In columns (9) and (1 D), hypothesis 4.4a is fully supported for technological ambidexterity. In contrast; hypothesis 4.4b is only supported in column (9). In column

(lD), the average perfonnance differences between firms with a hybrid strategy and without technological ambidexterity as opposed to firms with a pure strategy and without pure ICT are insignificant. Given the abovementioned assumption that firms with a pure strategy but without pure ICT may, on average, underperform firms with a pure strategy and pure ICT, this finding is not surprising. However, hypothesis 4.4b is only partly supported for technological ambidexterity.

Table 4.8 Summary of the results Hypotheses Hypothesis 4.1: Organizational ambidexterity positively moderates the impact of hybrid strategies on firm performance. Hypothesis 4.2a (4.2b): On average, in the presence (absence) of organizational ambidexterity, hybrid strategies outperform (underperform) the entire sample offirms. Hypothesis 4.3a (4.3b): On average, in the presence (absence) of organizational ambidexterity, hybrid strategies outperform (match) no-emphasis strategies. Hypothesis 4.4a (4.4b): On average, in the presence (absence) of organizational ambidexterity, hybrid strategies match (underperform) pure strategies. ./ supported; 10 partly supported

Individual Technological ambid. ambid.


4.4 Results

131

4.4.3 Robustness test: Porter's definition ofthe performance impact of a strategy I argue that Porters (1980; 1985) original and more challenging definition of a strategy's performance impact, that Is, the effect on the odds of achieving perfonnance above the industry average, is largely neglected in the empirical literature. Thus, I reassess hypotheses 4.1 to 4.4 by employing a logistic analysis with the binary dependent variable ABOVE that indicates if a firm's performance lies above the industry average performance. Note that in both the individual ambidexterity sample and the technological ambidexterity sample, several observations are lost in the logistic regressions because some manufacturing sectors are only represented by one firm, prohibiting the calculation of the industry average performance. All coefficients are exponentiated and thus interpreted as odds ratios. The logistic regression results for hypothesis 4.1 in Table 4.9 (next page) show that the findings described above hold. For a one unit increase in the hybrid strategy variable, the odds of performing above the industry average are not significantly affected in

columns (1) and (3). The interaction terms in columns (2) and (4) are positive and significant. However, in logistic regressions, one cannot "interpret a positive (negative) sign on an interaction term as meaning that there is always an enhancing [diminishing] relationship between the two variables" (Hoetker, 2007: 337). Instead, the magnitude and the significance of the interaction effect can differ across observations (Ai and Norton, 2003; Hoetker, 2007; Norton et al., 2004). The STATAroutine 'inteff provided by Norton et al. (2004) is used to calculate the magnitude and significance of the interaction effect over the entire sample of observations (not reported). According to this, the interaction effects in columns (2) and (4) are positive over almost the entire range of observations but are not significant for all observations. Thus, hypothesis 4.1 is partly supported for individual and technological ambidexterity.


132

4 Hybrid strategy and firm performance

Table 4.9 Test ofhypothesis 4.1 with Porters performance definition Dependentvariable Sam pie

ABOVE (t) Individual ambidexterity (1) (2)

Technological ambidexterity (3) (4)

Firm size (ln)

1.0314 (0.1226)

0.9795 (0.1199)

1.1317 (0.1298)

1.1390 (0.1379)

Firm age (ln)

1.2939** (0.1504)

1.2827** (0.1481)

1.3254*** (0.1426)

1.3229** (0.1443)

ROA (t-l)

1.2010*** (0.0363)

1.2075*** (0.0388)

1.1520*** (0.0316)

1.1505*** (0.0317)

Environmental dynamism

1.0275 (0.0267)

1.0367 (0.0271)

1.0110 (0.0262)

1.0144 (0.0278)

Hybrid

1.3280 (O.3549)

1.3385 (0.3589)

1.2321 (0.2938)

1.2023 (0.2950)

strategy Individual ambidexterity

1.4544 (0.9168)

Hybrid strategy x individual ambid.

3.0893** (1.5973) -0.1827 (2.6043)

Technological ambidexterity

19.0396** (4.6075)

Hybrid strategy x technological ambid. Further controls Firms Observations Pseudo-Ra

Industry, manufactunng sector, and year dummies: country dummy; environmental dynamism missing dummy

254 450

254 450

298 513

0.61

0.62

0.50

298 513 0.50

* p < 0.1; ** P < 0.05; *** P < 0.01; time period 2005-2008; logistic regressions used in all columns; standard errors in parentheses clustered by firm and Huber-White robust to heteroskedasticity and autocorrelation of unknown form.

Further logistic regression results are reported in Table 4.10 (next page). I confine the discussion of the results in Table 4.10 to pointing out the major similarities and differences with respect to Table 4.7. Hypothesis 4.2 is only partly supported. As shown in column (1), the odds of firms with a hybrid strategy and individual ambidexterity to display above industry average performance are on average significantly higher than the odds of the entire sample. However; this is not supported for technological ambidexterity in column (6). Further, the odds of firms with a hybrid strategy but without individual or technological ambidexterity to achieve above industry average performance are, on average, not significantly lower than the odds of the entire sample, as shown in columns (1) and (6).


1.9267* (0.6538) 1.0911 (0.4 113)

0.7569 (0.1666)

1.45 82 * (0.31 96)

0.8258 (0.2 176)

w 10 ind./techn. ambidexterity

Pur e str at egy and pure str uct ures /HRM/ICT

w/o pure struct ures/HRM/IC T

0.39

0.39

254 450

0.39

254 450

0.39

254 450

0.30

0.30

513

0.9013 (0.3379)

1.9 128 (0.8 155)

1.8408* (0.638 2)

2.1318* (0.8241)

1.0079 (0.025 7)

1.15 09*** (0.0322)

1.348 4*** (0. 1486)

1.098 7 (0 .13 11)

513

2.1223** (0.7740)

2.0425*** (0.5 57 5)

1.1096 (0.4160)

2.3654** (0.8393)

1.00 79 (0.0 257)

1.1509 *** (0.0322)

1.3484*** (0.1486)

1.09 8 7 (0.13 11)

0.30

0.30

513

0.4896*** (0.1336)

1.0391 (0.3682)

0.5432* (0.1883)

1.1581 (0.3740)

1.0079 (0.0 25 7)

1.1509 *** (0.0322)

1.34 84*** (0.1486)

1.09 8 7 (0.13 11)

dumm; environmental d narmsm missin du mm 298 298 298 298

513

1.3050 (0.3 480)

1.2559 (0.2225 )

0.6822 (0.1738)

1.4544 (0.3532)

1.00 79 (0.0 25 7)

1.15 09*** (0.0322)

1.348 4*** (0.1486)

1.09 8 7 (0.13 11)

298

0.30

513

0.471 2** (0.1718 )

0.96 24 (0.3411)

0.5228 (0.2229)

1.1145 (0.4764)

1.0079 (0.0 25 7)

1.1509*** (0.0322)

1.3484 *** (0.1486)

1.09 87 (0.13 11)

Technological amb idext erity Dumm Dumm Dumm v Dumm Hybrid strat. No-emphasis Pure strategy Pure stretegy w/o t. ambtd. st rateev and eure org. w 10 oure ora. 8 9 10

* P < 0.1; ** P < 0.05; *** P < 0.01; ti m e period 2005-2008; logistic regresstons used in all columns: standard errors in parentheses clustered by firm and HuberWhite robust t o heteroskedasticity and autocorrelation of unknown form.

rseooo-na

254 450

Ind ustr ,manufact urin sectcr. and ear d ummies; count

0.6779 (0.2656 )

Fur ther controls Firms Observations

0.3839*** (0. 135 5)

0.5663 (0. 209 2)

0.9165 (0.3455)

0.5 190* (0.1761) 1.76 59 (0.65 22)

2.3734* (1.1671)

1.0253 (0.0260)

1.2 105*** (0.0399)

1.3440 (0.5867)

1.0253 (0.0260)

1.21 05*** (0.0399)

1.3398** (0.1608)

0.9 433 (0.11 43)

0.7397 (0.2612 )

1.475 1 (0.5780)

3.5011 *** (1.6307)

1.0253 (0.0260)

1.21 05*** (0.0399)

1.3398** (0.1608)

0.9433 (0 .1143)

No-em phasis str ategy

Pure strategy

Hybrid strategy

0.39

2.6049*** (0.91 96)

2.5896** (1.1194)

1.9599** (0.6257)

Hybrid st ra t egy and ind./ t echn . ambidex terity

254 450

1.3520 (0.4774)

1.0253 (0.0260)

1.0253 (0.0260)

Environment al

dy namis m

1.21 05*** (0.0399)

1.21 05*** (0.0399)

ROA (t -1)

1.3398** (0.1608)

1.3398** (0.1608)

1.3398** (0.1608)

Firm age (In)

0.9433 (0 .1143)

0.9433 (0.1143)

0.9433 (0 .1143)

1

Effects Entire samnle

ABOVE Indi vi d ual ambidexterity Dumm Dumm Dumm Dumm Effects Hybrid strat. No-emphasis Pure strategy Pure strategy Entire w/o i.ambid. st ra tegy and eure org. w 10 oure or . sam nle 2 3 4 5 6

Firm si ze (In)

Reference gro up

De en dentvari able Sam ple Codin

Table 4.10 Test of hypotheses 4.1 to 4.4 with Porter's performance definition


134

4 Hybrid strategy and firm performance

Columns (2) and (7) fully support hypothesis 4.1 for individual and technological ambidexterity. Hypothesis 4.3 is fully supported for individual and technological ambidexterity, as shown in columns (3) and (8). Hypothesis 4.4a is supported for individual and technological ambidexterity in columns (4), (9), and (10). However; in column (5), for individual ambidexterity, the odds of firms with a hybrid strategy and individual ambidexterity to achieve above industry average performance are, on average, higher than the odds of firms with a pure strategy and without pure structures and HRM practices. However, this finding has a low significance level. Again, this is not surprising given the abovementioned assumptions about the average performance differences between the two subgroups of pure strategy firms. Hypothesis 4.4b is partly supported for individual and technological ambidexterity. The odds of finns with a hybrid strategy but without individual or technological ambidexterity to achieve above industry average performance are, on average, lower than the odds of firms with a pure strategy and with pure structures and HRM practices (or lCT) in columns (4) and (9). However; the odds of firms with a hybrid strategy but without individual or technological ambidexterity to achieve above industry average performance are, on average, equal to the odds of firms with a pure strategy but without pure structures and HRM practices (or lCT) in columns (5) and (10). Again, this is not surprising given the abovementioned assumptions about the average performance differences between the two subgroups of pure strategy firms. Table 4.11 (next page) provides a summary of the findings conceming Porter's performance definition. All relationships hypothesized above are reconfinned with regards to their sign. However; as expected given this more stringent definition of a strategy's performance impact, several findings do not hold in significance.


4.4 Results

135

Table 4.11 Summary of the results with Porters perfonnance definition Hypotheses Hypothesis 4.1: Organizational ambidexterity positively moderates the impact of hybrid strategies on the odds of achieving above industrv average performance. Hypothesis 4.2a (4.2b): On average, in the presence (absence) of organizational ambidexterity, the odds that a hybrid strategy achieves an above industry average performance are higher (lower) than the odds of the entire sample of firms. Hypothesis 4.3a (4.3b): On average, in the presence (absence) of organizational ambidexterity, the odds that a hybrid strategy achieves an above industry average performance are higherthan (as high as) the odds of no-emphasis strategies. Hypothesis 4.4a (4.4b): On average, in the presence (absence) of organizational ambidexterity, the odds that a hybrid strategy achieves above industry average perfonnance are as high as (lowerthan) the odds ofpure strategies.

Individual Technological ambid. ambid.

0(0)

./ supported; 10 partly supported; x not supported 4.4.4 Pure strategies Although I did not formally advance hypotheses on pure strategies, I briefly discuss some interesting insights into the relationship between pure strategies and firm performance. Firms with a pure strategy and pure structures and HRM practices (or with pure ICT) consistently outperform firms with a no-emphasis strategy, as shown in columns (3) and (8) of Table 4.7 and Table 4.10. As shown in the same columns, even finns with a pure strategy but without pure structures and HRM practices (or without pure ICT) largely outperform firms with a no-emphasis strategy, with one exception in column (3) in Table 4.10. However; finns with a pure strategy are unable to outperform the entire sample in most cases, as shown in columns (1) and (6) ofTable 4.7 and Table 4.10. In addition to the findings with regard to hypothesis 4.4, these findings may provide a more nuanced assessment of the relative performance advantages and disadvantages of pure versus hybrid strategies. As compared to firms with a hybrid strategy, firms with a pure strategy seem to have the disadvantage of being less able to achieve perfonnance above the sample average. In contrast, as compared to finns with a hybrid strategy, finns with a pure strategy seem to have the advantage ofbeing less reliant on high levels of appropriate structures and HRM practices (or appropriate ICT) for achieving a higher average performance than finns with a no-emphasis strategy. Other than firms with a


4 Hybrid strategy and firm performance

136

hybrid strategy but without organizational ambidexterity, finns with a pure strategy but

without pure structures and HRM practices (or without pure lC'I'] are largely able to outperform firms with a no-emphasis strategy.

This may result from the fact that firms with a pure strategy do not have to organizationally resolve any tensions resulting from corporate strategy. F端r example, f端r

finns that exclusively emphasize exploration, modest levels of crganic organizational arrangements may still be enough to outperform finns with a no-emphasis strategy in most cases. Instead, firms with a hybrid strategy that neglect either organic or mechanistic organizational arrangements are unable to resolve the tensions from the hybrid strategy and thus suffer from reduced performance.

4.5

Discussion and conclusions

In this study, I examined the role of organizational architecture as a moderator of the hybrid strategy-performance relationship. Particularly, I focused on the roles of individual and technological ambidexterity in relatively small firms. Further, I examined the performance impact of hybrid strategies as compared to different reference groups and employed two different definitions of a strategy's performance impact. To test my hypotheses. I employed a novel dataset from three independent sources and with objective perfonnance data from annual firm accounts. I employed moderation effects. effects-coding, and dummy-coding models, and I controlled for a variety of factors such as past firm performance and environmental dynamism. I found evidence that both individual and technological ambidexterity positively moderate the hybrid strategy-performance relationship. Further; I found hybrid strategies to outperform the entire sample and no-emphasis strategies in the presence of organizational ambidexterity, but to underperform the entire sample and match the performance of firms with no-emphasis strategies in the absence of organizational ambidexterity. Finally, I found hybrid strategies to match the performance of pure strategies in the presence of organizational ambidexterity but to underperfonn them in absence of organizational ambidexterity. These findings largely hold across two different definitions of a strategy's performance impact and raise important issues for both theory and practice.


4.5 Discussion and conclusions

137

4.5.1 Implications for research First and foremost, this study provides a conceptual advancement. To the best of my knowledge, this study is the first large-sample empirical analysis of the role of organizational architecture as a moderator of the hybrid strategy-perfonnance relationship. Thus, this study is a first step to closing a gap between theoretical and empirical research on the performance impact of hybrid strategies. Particularly, the finding that the presence of an ambidextrous organizational architecture does not only change the strength but also the direction of a hybrid strategy's performance effect might offer an explanation for the fact that both positive and negative perfonnance effects of hybrid strategies have been found in prior studies which disregarded organizational architecture. Second, to my knowledge, this study provides the first analysis of the potential of leT to

achieve crganizational ambidexterity. Third, this study provides a methodological advancement by employing a variety of performance definitions. The fact that a hybrid strategy's performance impact is consistent across this variety of performance definitions if organizational architecture is incorporated into the research design is further support for the notion that any empirical study on the hybrid strategy-perfonnance relationship must incorpcrate organizational architecture. Fourth, this study is the first example to show that it can be fruitful to combine the largely disconnected theoretical and empirical research on classic typologies of corporate strategy as well as on ambidexterity. For example, classic strategy typologies are more sophisticated with respect to the theoretical differentiation between corporate strategy and organizational architecture, while ambidexterity researchers provide more detailed suggestions with regard to the organizational arrangements that are suitable to manage the tensions arising from a hybrid strategy. Finally, in addition to the theoretical and empirical advancements of this study, there is also a contribution with respect to the data-gathering methods used. I showed that welldesigned telephone surveys can generate a large amount of high-quality observations that are easy to match with other datasets in a limited period of time. Further, telephone surveys provide the possibility of mixing qualitative and quantitative data-gathering methods. resulting in information that is very detailed and unbiased and that can be evaluated statistically at the same time.


138

4 Hybrid strategy and firm performance

4.5.2 Implications for managers With regard to managerial implications, this study draws a more differentiated picture of the performance effects of hybrid strategies than prior studies. First, f端r managers that pursue the exploratory as well as the exploitative success of their firms. this study confirms that hybrid strategies can indeed positively affect firm performance if suitable organizational arrangements are in place. Second, these organizational arrangements are less difficult to establish than prior studies have argued. Instead, a sophisticated combination of well-known and wellunderstood organizational structures and HRM practices is sufficient to provide organizational ambidexterity. Third, managers in relatively small firms should especially consider individual ambidexterity as the appropriate organizational solution for hybrid strategies. Fourth, managers should also draw their attention to the potential of lCT to achieve organizational ambidexterity. Fifth, although firms with hybrid strategies and organizational ambidexterity have relative advantages as compared to firms with pure strategies in achieving perfonnance that is above the sample average, managers should also keep in mind that pure strategies have relative advantages with regard to the crganizational demands required to at least outperform no-emphasis strategies. Also, redesigning ambidextrous organizations may be much more costly than redesigning pure organizational architectures. Thus, the optimal choice between a hybrid and a pure strategy might also depend on the organizational and managerial capabilities of a firm.

4.5.3 Lirnitations and future research 端f course, this study is not without limitations. Although the dataset is a major strength of this study and though several measures were undertaken to avoid any blas, I had to combine cross-sectional data on corporate strategy, organizational structures, and HRM practices with time-series data on firm performance and lCT types. Future studies could employ actual time-series data on corporate strategy and organizational architecture.


4.5 Discussion and conclusions

139

Further; to make firms within the study comparable and in order to focus on one questionnaire, my survey focused on manufacturing firms. To be able to conduct telephone interviews, may survey was limited to German and Polish firms. Finally, to fruitfully analyze the role of individual ambidexterity, I chose relatively small firms as subjects of the study. In sum, future studies could verify if the findings of this study are generalizable to other industries, other countries, and larger firms. Also, general method of moments (GMM) regressions could be employed to mitigate the potential for bias through the inclusion of a lagged dependent variable in the models used to analyze the panel dataset (Roodman, 2006). Further; although the dataset allowed the assessment of objective perfonnance data with a one-year lag, future studies should analyze the long-term perfonnance impacts of hybrid strategies (3, 5, or 10 years). Particularly, it would be interesting to see if hybrid strategies outperform pure strategies in the long tenn. Also, it would be interesting to compare the performance variation within the groups of finns with a hybrid strategy and with (or without] organizational ambidexterity to the performance variation within other groups of finns at a certain point in time (He and Wong, 2004) or over time (Levinthal and March, 1993). It would also be interesting to test if the moderating effect of organizational ambidexterity in the hybrid strategy-performance relationship can be verified for structural, leadership-based, and informal ambidexterity. In addition, the potential of ICT to enhance organizational ambidexterity should be analyzed in more detail. Specifically, it would be interesting to examine whether ICT is particularly fruitful for enhancing specific types of organizational ambidexterity, or if it even enables new fonns of organizational ambidexterity. In conclusion, I view this study as a promising step to understanding the role of organizational ambidexterity in the hybrid strategy-performance relations hip.


Chapter 5 Conclusion 5.1

Summary and contributions

I presented one thecretical and two empirical studies on the effects of aligning

information technology (IT), organization, and strategy on firm performance in Chapters 2 to 4.

The empirical analyses in Chapters 3 and 4 are based on two novel multi-source datasets on IT, organization design, corporate strategy, and perfonnance indicators fĂźr German and Polish manufacturing firms. While the data on IT and firm performance stern from two independent third-party sources, the data on organization design and corporate strategy were gathered specifically fĂźr this dissertation by means of a large-scale

telephone survey resulting in about 1,500 telephone interviews. In Chapter 2, I presented a structured review of existing empirical studies on the role of complementarities in IT's performance impact. These studies come from three distinct theoretical backgrounds; namely, organizational economics, resource-based view, and contingency theory. While the former two focus on the examination of horizontal IT complements inside the organization design, the latter focuses on the examination of vertical IT complements that require a certain organization design. I pointed out the separate examination of horizontal and vertical IT complements and the lacking examination of the mechanisms by which IT and IT complements combine as major theoretical drawbacks. Drawing from contingency theory and information processing theory, I developed an integrative model ofIT complements to address these and further issues. Accordingly, IT must be simultaneously aligned with horizontal and vertical IT complements. While horizontal fit is needed to create an efficient organization design, vertical fit ensures that organization design is targeted at the requirements associated with contingency factors such as a firm's corporate strategy, environment, or culture. Information processing has been presented as the mechanism that explains the demand for both vertical and horizontal fit. Accordingly, contingency factors impose specific Information processing requirements on a firm's organization design that must be matched by the information processing capacities of crganization design. To offer suitable information processing

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6_5, Š Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


142

5 Conclusion

capacities, IT that facilitates infonnation processing and organizational complements that determine the direction of the information flow have to be aligned. Subsequently, drawing from prior theoretical and empirical research, I provided an analysis of a variety of horizontal and vertical IT complements as well as of IT and firm performance. In particular, I provided arieher view of decentralization as a horizontal IT complement, suggesting the examination of the power distribution in a firm rather than only the examination of the degree of the centralization of decision rights. I argued that the type and level of standardization may also influence the power distribution and that the degree of the centralization of decision rights and the human resource management (HRM) practices in a firm have to be analyzed on a more detailed level. With regard to methodology, I argued that IT should be analyzed at the level of IT applications. which allows detailed conclusions on what IT types are complementary with other elements inside or outside the organization design. I further elaborated that when measuring IT complements, researchers must be careful to not to directly measure 'fit', as this would disable them from drawing conclusions on how to achieve this fit. I also argued that the overall firm performance rather than any intermediate perfonnance measure is the right criterion variable when analyzing the complementarities between IT and other elements. If the alignment between IT and other elements does not transfer into overall firm performance, this indicates that not all necessary types of fit have yet been achieved. Finally, I advanced testable propositions on the complementarities between IT, horizontal, and vertical complements. Particularly, I suggested that IT types that support the information exchange along a firm's hierarchy should be horizontally aligned with the centralization of decision rights, high coercive standardization, and HRM practices that focus on skills and knowledge needed to implement rules, procedures, and instructions. This organization design should be vertically aligned with an undemanding environment, a corporate strategy that focuses on existing product market domains, and an organizational culture that becomes manifest in high fonnality. IT types that support the infonnation exchange between a firm's subunits should be horizontally aligned with the decentralization of decision rights, low coercive standardization, high enabling standardization, and HRM practices that focus on diverse skills and knowledge needed for communication and coordination with other employees. This organization design should be vertically aligned with ademanding environment, a corporate strategy that focuses on new product market domains, and an organizational culture that becomes manifest in low formality.


5.1 Summary and contributions

143

In summary, Chapter 2 offers a consistent theoretical framework and a variety of suggestions to spur more fruitful studies on the role of complementarities in the ITperformance relationship. In Chapter 3, I empirically examined one part of the framework presented in Chapter 2. In particular, I analyzed the complementarities between IT and the degree of de/centralization and whether these complementarities are contingent on third factors. The analysis was guided by an information processing view of organizations, the IT and de/centralization debate, and the organizational learning literature. I argued that IT can facilitate both horizontal information processing between lateral organization units and vertical infonnation processing upward and downward a firm's hierarchy. Thus, I argued that IT's potential for performance increases can be exploited by either greater decentralization or greater centralization. Greater decentralization requires greater horizontal infonnation processing for the autonomous communication and cocrdination of autonomous subunits. Greater centralization requires greater vertical information processing to bring information on the subtasks upward to central decision makers and to disseminate their instructions downward to the subunits. In addition, I argued that greater decentralization must be accompanied by suitable HRM practices that endow the autonomous employees with the necessary skills and incentives to make high-quality decisions. As both alternatives, that Is, complementarities between IT and decentralization and complementarities between IT and centralization, are feasible, third factors may determine which of the two alternatives is appropriate. In particular, I demonstrated that a firm's corporate leaming type determines if a decentralized or centralized organization design is appropriate. I advanced the hypothesis that the impact of IT on firm performance is enhanced through greater decentralization for finns that pursue an explcratory leaming type, because the exploratory search for new products and markets requires autonomous employees that constantly experiment and adapt. In contrast, I hypothesized that the impact of IT on firm performance is enhanced through greater centralization for firms that pursue an exploitative leaming type, because efficiency and continuous improvement in existing product market domains require finns to control their employees to minimize variation and slack resources. I employed a novel multi-source panel dataset on computer capital, organization design, and corporate leaming type for almost 260 German manufacturing firms to test the hypotheses. The dataset is constructed from three independent sources and covers the


144

5 Conclusion

years 2000 to 2008. I estimated Cobb-Douglas production functions with an interaction term of IT and de/centralization to test f端r complementarities. To test if these complementarities are contingent on a firm's corporate leaming type, I used subgroup analysis; that Is, the interaction term of IT and de/centralization was further interacted

with a dummy variable indicating the corporate leaming type of a firm. The results support the hypotheses that decentralization is complementary to IT in finns with an explcratory leaming type, while centralization is complementary to IT in finns with an exploitative leaming type. I employed a variety of robustness checks to demonstrate that the qualitative nature of these results is not sensitive to alternative specifications and measures. In summary, to the best of my knowledge, I presented the first large-sample empirical study that finds evidence for an association between IT and centralization. Thus, Chapter 3 significantly contributes to the long-standing debate on whether IT is complementary to centralization or decentralization. In Chapter 4, I examined the role of organizational architecture as a moderator of the hybrid strategy-performance relationship. Hybrid strategies mix distinct strategic activities such as exploration of new product market domains and exploitation in existing product market domains: thus, they suffer from severe organizational tensions. Hence, I hypothesized that hybrid strategies positively affect firm performance only if suitable organizational arrangements to solve these organizational tensions are in place. In contrast, in absence of suitable organizational arrangements, hybrid strategies negatively affect firm perfonnance. Drawing on the literature on ambidextrous organizations, I analyzed the moderating roles of two types of organizational ambidexterity that allow the pursuit of distinct strategie activities with equal dexterity. Individual ambidexterity combines the decentralization of organizational decision rights with enabling standardization and several HRM practices to enable the individual employee to pursue both exploratory and exploitative activities. Technological ambidexterity concerns the yet largely unexplored potential of IT to support the simultaneous pursuit of exploration and exploitation. According to that, employing IT types such as groupware applications and corporate intranets that support exploratory activities and IT types such as enterprise resource planning (ERP) and workflow applications that support exploitative activities in one firm enhances the ability of employees to think and act ambidextrously.


145

5.1 Summary and contributions

Further; I advanced hypotheses conceming the relative firm performance of firms with a hybrid strategy with (or without] organizational ambidexterity as compared to the entire sample of firms, firms with a no-emphasis strategy, and firms with a pure strategy. In particular, I hypothesized that, on average, hybrid strategies outperform the entire sample in the presence of crganizational ambidexterity and underperform the entire sample in the absence of organizational ambidexterity. Also, I argued that, on average, hybrid strategies outperform no-emphasis strategies in the presence of organizational ambidexterity and match no-emphasis strategies in the absence of organizational ambidexterity. Further, I argued that, on average, hybrid strategies match pure strategies in the presence of organizational ambidexterity and underperform pure strategies in the absence of organizational ambidexterity. I employed a dataset on 784 German and Polish manufacturing firms to test the hypotheses. I used moderation effects models to test the central hypothesis that the hybrid strategy-performance relationship is positively moderated by organizational ambidexterity. This hypothesis was supported. I used dummy- and effects-coding models to assess the performance differences between firms with a hybrid strategy and different comparison groups, and I found evidence for the hypotheses outlined above. Although not fonnally hypothesized, I reassessed all hypotheses employing a more challenging definition of a corporate strategy's impact on performance, that Is, the odds of a corporate strategy to induce above industry average perfonnance. Many of the hypotheses were reconfinned. I further examined several findings that may provide a more nuanced assessment of the relative performance advantages and dis advantages of pure versus hybrid strategies. While firms with a hybrid strategy seem to be able to achieve above sample average performance with ease, firms with a pure strategy seem to be less reliant on their organizational architecture for outperforming finns with a no-emphasis strategy. Thus, while hybrid strategies may be better suited to achieve very high firm performance, pure strategies may offer an easier way to achieve satisfying firm performance, because other than finns with hybrid strategies, firms with pure strategies do not have to organizationally solve any tensions from their corporate strategy. In summary, Chapter 4 contains to my knowledge the first large-sample empirical analysis of the role of organizational architecture as a moderator of the hybrid strategyperformance relationship. The

finding that the presence of an ambidextrous

organizational architecture does not only change the strength but also the direction of a


5 Conclusion

146

hybrid strategy's perfonnance effect might offer an explanation f端r the fact that both positive and negative performance effects of hybrid strategies were found in prior studies which disregarded the role of organizational architecture. Also, this study

suggests that it may be fruitful to combine the largely disconnected literatures on classic strategy typologies and organizational ambidexterity. Further, the study provides the

first analysis of the role ofIT in achieving organizational ambidexterity.

5.2

Suggestions f端r future research

Apart from the present findings, this dissertation also opens up

interesting

opportunities f端r future studies. F端r example, theoretical and empirical analyses may benefit from further integrating the information systems literature and the organizational economics literature on IT complements. Chapter 2 has provided a first step in this direction. Also, future studies should identify further contingency factors in addition to strategy, environment, and culture that may affect the complementarities between IT and other elements of organization design. Contingency theory and information processing thecry may provide a fruitful theoretical background to identify those contextual factors and to explain the mechanisms by which they combine. Situational misfit, that Is, the situation where two or more contingency factors impose different information processing requirements on an organization design, has been briefly addressed in Chapter 2. Chapter 4 on hybrid strategies provided a detailed analysis of the situation in which different information processing capacities were demanded by the contingency factor corporate strategy. However, future research should further investigate the cases where different contingency factors such as a firm's environment and its culture demand different information processing capacities. In particular, IT's supporting role in achieving vertical fit with more than one contingency factor should be examined in detail. Although Chapter 4 has provided first ideas and findings. it is of great importance to analyze in more detail which specific IT types are complementary with other organizational elements. Mter all, the findings of future empirical studies on IT complements are managerially relevant only if they provide insights on which IT types are necessary to exploit specific complementarities.


5.2 Suggestions for future research

147

An interesting topic for future research is to analyze the moderating roles of other types of organizational ambidexterity in the hybrid strategy-performance relationship. The findings of Chapter 4 on individual and technological ambidexterity should be reconfirmed for structural, leadership-based, and informal ambidexterity. In addition, although I argued that IT moderates the hybrid strategy-performance relationship through its support for individual and structural ambidexterity, I did not fonnally examine these intermediating mechanisms. Future studies could assess if IT is actually suitable to support individual and structural ambidexterity and if IT may also support other types of organizational ambidexterity. Further, it would be very interesting to examine if IT has relative advantages in supporting specific types of organizational ambidexterity. Although the datasets used in Chapters 3 and 4 are a major strength of this thesis and while I undertook a variety of measures to achieve high data quality and avoid any bias. I present several suggestions to further exploit the potentials of these datasets and to gather data of even higher quality in the future. The potential of the dataset used in Chapter 4 to assess similarities and differences ofIT, corporate strategies, organization designs, and other firm characteristics for German and Polish firms has not yet been fully exploited. In particular, Poland became a democracy with a free-market economy about 50 years after Germany. Even 20 years after the socio-pclitical tumaround in Poland, this may result in interesting differences with regard to strategies, IT, organizational structure, and HRM practices. Similarly, although contained in both datasets. the data on the nationality and type of firm ownership, and specifically on owner-managers, are not yet exploited. For example, owner-managed firms may manage their IT investments better or worse or perform better or worse in exploiting complementarities between IT and other organizational elements. Similarly, Ug-owned firms were found to manage IT better than finns with owners of other nationalities (Bloom et al., 2009b). While I measured the realized levels of exploration and exploitation over the last years in Chapters 3 and 4 and thus made an effort to gain a realistic picture of the exploratory and exploitative activities of firms before 200S, patent applications data could be used to create areal time-series dataset on exploration and exploitation. For example, Katila and Ahuja (2002) used the share of patent citations that are used for the first time as a measure of exploration. The share of patent citations that were repeatedly used during the preceding years is used as a measure of exploitation. Another idea is to measure the


148

5 Conclusion

similarity of patent classes. The share of patent applications in classes that are dissimilar to the classes in which the firms already hold patents may be used as a measure of exploration. The share of patent applications in classes that are similar to the classes in which the firms already hold patents may be used as a measure of exploitation.

More generally, real time-series data could be created by repeating the 2008 survey. I described in detail that the telephone survey method that has been used f端r this thesis allows gathering a large number of data in a limited time period.

Follow-up telephone surveys may also incorporate service firms to test if the findings presented above f端r manufacturing firms are generalizable.

Further; the datasets used in Chapters 3 and 4 "Will become more valuable if they are combined with firm performance data from the years following 2008. This is easily achieved by further extractions from the ORBIS database and allows the analysis of the medium- and long-term impacts of combinations of IT, crganization, and strategy on firm perfonnance. In particular, it would be interesting to see if hybrid strategies outperform pure strategies in the medium- or long-term, as discussed above. Finally, although I undertook two attempts to conduct reliable tests of complementarity, future empirical studies on complementarities among IT, organization, and strategy could use even more sophisticated approaches. First, I used a performance approach rather than an adoption approach in Chapters 3 and 4. According to the adoption approach used in many earlier studies on complementarity, the simultaneous occurrence of two elements in a firm is a sufficient indication of complementarity, because finns have no other reasons to combine elements than to exploit the complementarities between them. Instead, I derived complementarity between two elements from the finding that the combination of these elements actually positively affects firm performance. Second, I accounted for a large number of third factors in all empirical analyses and thus reduced the possibility that there are remaining unobserved factors that might offer an explanation for

the simultaneous occurrence of

organizational elements alternative to complementarity. However; further approaches are proposed to account for unobserved heterogeneity that could be employed in future analyses of the complementarities among IT, organization, and strategy (Burton et al., 2002; Cassiman and Veugelers, 2006; Miravete and Pemias, 2006; Mohnen and Roller; 2005).


References Acemoglu D. 1998. Why Do New Technologies Complement Skills? Directed Technical Change and Wage Inequality. Quarterly journal ofEconomics 113(4): 1055-1089. Adler PS, Bcrys B. 1996. Two Types of Bureaucracy: Enabling and Coercive.

Administrative Science Quarterly 41(1): 61-89. Adler PS, Goldoftas B, Levine D. 1999. Flexibility versus efficiency? A case study of model

changeovers in the Toyota production system. Organization Science 10(1): 43-68. Ai C, Norton EC. 2003. Interaction terms in logit and logit models. Economics Letters 80(1): 123-129. Aiken LS,West SG. 1991. Multiple Regression: Testing and lnterpreting lnteractions. Sage: Newbury Park, CA, USo

Alavi M, Leidner DE. 2001. Review:

Knowledge Management and Knowledge

Management Systems: Conceptual Foundations and Research Issues. MIS Quarterly 25(1): 107-136. Ambos TC, Mäkelä K, Birkinshaw J, D'Este P. 2008. When Does University Research Get Commercialized? Creating Ambidexterity in Research Institutions. journal of

ManagementStudies 45(8): 1424-1447. Anand KS, Mendelson H. 1997. Information and Organization for Horizontal Multimarket Cocrdination. Management Science 43(12): 1609-1627. Ancona DG,Goodman PS, Lawrence BS, Tushman ML. 2001. Time: A New Research Lens.

Academy ofManagementReview 26(4): 645-663. AndersenJA, Jonsson P. 2006. Does organizational structure matter? On the relationship between the structure, functioning and effectiveness. International journal of

Innovation and Technology Management 3(3): 237-263. Anderson MC, Banker RD, Ravindran S. 2006. Value Implications of Investments in Information Technology.Management Science 52(9): 1359-1376. Andriopoulos C, Lewis MW. 2009. Exploitation-Exploration Tensions and Organizational Ambidexterity: Managing Paradoxes of Innovation. Organization Science 20(4): 696-717. Aoki M. 1986. Horizontal vs. Vertical Information Structure of the Firm. American

Economic Review 76(5): 971-983.

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6, © Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


References

150

Aral 5, WeiH P. 2007. IT Assets, Organizational Capabilities, and Firm Performance: How Resource Allocations

and Organizational Differences

Explain Performance

Variation. Organization Science 18(5): 763-780. Athey 5, Stern S. 2003. An empirical framework für testing theories about complementarity in organizational design. Working Paper, National Bureau of Economic Research, Cambridge, MA, USo Auer 1. 2005. Ökonometrie Eine Einführung. Springer: Berlin, Germany. Auh 5, Mengue B. 2005. Balancing Exploration and Exploitation: The Moderating Role of Competitive Intensity.}ournal 01Business Research 58(12): 1652-1661. Autor DH, Katz LF, Krueger AB. 1998. Computing Inequality: Have Computers Changed the Labor Market? Quarterly journal ofEconomics 113(4): 1169-1213. Bae J, Lawler

JJ.

2000. Organizational and HRM strategies in Korea: impact on firm

perfonnance in an emerging economy. Academy of Management journal 43(3): 502-517. Baily MN. 1986. What Has Happened to Productivity Growth? Science 234(4775): 443451. Barney JB. 1986a. Factor Markets: Expectations, Luck and Business Strategy.

Management Science 32(10): 1231-1241. Barney JB. 1986a. Organizational Culture: Can It be a Source of Sustained Competitive Advantage? Academy ofManagement Review 11(3): 656-665. Barney JB. 1991. Firm Resources and Sustained Competitive Advantage. journal of

Management 17(1): 99-120. Bartel AP, Lichtenberg FR. 1987. The Comparative Advantage of Educated Workers in Implementing New Technology. Review ofEconomics and Statistics 69(1): 1-11. Barua A, Mukhopadhyay T. 2000. Information Technology and Business Performance: Past; Present and Future. In Framing the Domains ofIT Management: Projecting the

Future Through the Past, Zmud RW (ed.). Pinnaflex Educational Resources: Cincinnati, OH, USo Bauer

0,

Tenz

B. 2009.

Informations-

und

Kommunikationstechnclogien

in

Unternehmen. Wirtschaft und Statistik 1: 47-54. Becker B, Gerhart B. 1996. The Impact of Human Resource Management on Organizational Performance: Progress and Prospects. Academy of Management

JournoI39(4): 779-801.


References

151

Becker BE, Huselid MA. 1998. High performance work systems and firm perfonnance: A synthesis of research and managerial implications. In Research in Personnel and Human Resource Management, Fenis GR (ed.). JAI Press: Greenwich, CT,USo Becker BE, Huselid MA. 2006. Strategie Human Resource Management: Where Do I Go From Here? journal ofManagement 32(6): 898-925. Benner MJ, Tushman ML. 2003. Exploitation, Exploration, and Process Management: The Productivity Dilemma Revisited. Academy ofManagement Review 28(2): 238-256. Bergeron F, Raymond L, Rivard S. 2001. Fit in strategic information technology management research: an empirical comparison of perspectives. Omega 29(2): 125-142. Bergeron F, Raymond L, Rivard S. 2004. Ideal patterns of strategic alignment and business performance. Information and Management 41(8): 1003-1020. Bennan E, Bound J, Machin S. 1998. Implications of Skill-Biased Technological Change: International Evidence. Quarterly journal ofEconomics 113(4): 1245-1279. Bertschek I, Kaiser U. 2004. Productivity Effects

of Organizational Change:

Microeconometric Evidence. Management Science 50(3): 394-404. Besanko D, Dranove D, Shanley M, Schaefer S. 2007. Economics ofstrategy. John Wiley & Sons: NewYork, USo Bharadwaj AS. 2000. A Resource-Based Perspective on Information Technology Capability and Firm Performance: An Empirical Investigation. MIS Quarterly 24(1): 169-196. Bharadwaj AS, Bharadwaj SG, Konsynski BR 1999. Information Technology Effects on Firm Performance as Measured by Tobin's q. Management Science 45(7): 10081024. Birkinshaw JM. 2006. The ambidextrous organisation - driving leadership. Mannaz

Leadership

Newsletter

June

2006.

mannaz.comjMail.asp?MailID=102&TopicID=1160 [6 October 2009]. Birkinshaw JM, Gibson C. 2004a. Building Ambidexterity into the Organization. Sloan

ManagementReview45(4): 47-55. Birkinshaw JM, Gibson C. 2004b. Building an Ambidextrous Organisation. Working Paper, Advanced Institute ofManagement Research, London, UK. Birkinshaw JM, Gibson C. 2005. The Ambidextrous Organisation. Executive Briefing, Advanced Institute ofManagement Research, London, UK.


References

152 Bloom N, Garicano L, Sadun R, Van Reenen

J.

2009. The Distinct Effects of Information

Technology and Communication Technology on Firm Organization. Working Paper, Centre F端r Economic Performance, London School of Economics and Political Science, London, UK. Bloom N, Sadun R, Van Reenen

J.

2009a. Measuring and Explaining Decentralization

Across Firms and Countries. Working Paper, Centre f端r Economic Performance,

London School ofEconomics and Political Science, London, UK. Bloom N, Sadun R, Van Reenen J. 2009b. Americans Do I.T. Better: US Multinationals and the Productivity Miracle. Working Paper, Centre f端r Economic Performance, London School ofEconomics and Political Science, London, UK. Bloom N, Van Reenen

J.

2007. Measuring and explaining management practices across

firms and countries. Quarterly journal ofEconomics 122(4): 1351-1408. Bloomsfield BP, Coombs R. 1992. Information Technology, Control and Power: The Centralization and Decentralization Debate Revisited. journal of Management

Studies 29(4): 459-484. Bolton P, Dewatripont M. 1994. The firm as a communication network, Quarterly journal

ofEconomics 109(4): 809-839. Bresnahan TF, Brynjolfsson E, Hitt LM. 2002. Information Technology, Workplace Organization, and the Demand for Skilled Labor: Firm level Evidence. Quarterly

journal ofEconomics 117(1): 339-376. Bresnahan TF, Trajtenberg M. 1995. General purpose technologies 'Engines of growth'?

journal ofEconometrics 65(1): 83-108. Brown SL, Eisenhardt KM. 1998. Competing on the Edge: Strategy as Structured Chaos. Harvard Business School Press: Boston, MA, USo Brown AD, Starkey K. 1994. The Effect of Organizational Culture on Communication and

Information.journal ofManagement Studies 31(6): 807-828. Brynjolfsson

E. 1993. The productivity paradox Communications oftheACM 36(12): 66-77.

of

infonnation

technology.

Brynjolfsson E. 1994. Information Assets, Technology, and Organization. Management

Seienee 40(12): 1645-1662. Brynjolfsson E, Hitt LM. 1995. Information Technology as a Factor of Production: The Role of Differences Among Firms. Economics of Innovation and New Technology 3(3): 183-199.


References

153

Brynjolfsson E, Hitt LM. 1996. Paradox Lost? Firm-Level Evidence on the Returns to Information Systems Spending. Management Scienee 42(4): 541-558. Brynjolfsson E, Hitt LM. 1998a. Beyond the Productivity Paradox. Communications 01the ACM: 41(8): 49-55. Brynjolfsson E, Hitt LM. 1998b. Information Technology and Organizational Design: Evidence from Micro Data. Working Paper, Massachusetts Institute of Technology Sloan School, Cambridge, MA, USo Brynjolfsson E, Hitt LM. 2000. Beyond Computation: Information Technology, Organizational Transformation and Business Performance. journal 01 Eeonomie Perspeetives 14(4): 23-48. Brynjolfsson E, Hitt LM. 2003. Computing Productivity: Firm-level Evidence. Review 01

Economies and Statisties 8S(4): 793-808. Brynjolfsson E, Hitt LM, Yang S. 2002. Intangible Assets: Computers and 端rganizational Capital. Brookings Papers on Eeonomie Aetivity 1(1): 137-199. Brynjolfsson E, Malone TW, Gurbaxani V, Kambil A. 1994. Does Information Technology Lead to Smaller Firms? Management Scienee 40(12): 1628-1644. Brynjolfsson E, Mendelson H. 1993. Information Systems and the Organization of Modem Enterprise. journal 01 Organizational Computing and Eleetronie Commeree 3(3): 245-255. Brynjolfsson E, Yang S. 1996. Information Technology and Productivity: A Review of the Literature. In Advanees in Computers, Zelkowith M (ed.). Elsevier Academic Press: San Diego. CA, USo Bums T, Stalker GM. 1961. The Manaqement oflnnovation. Tavistock: London, UK Burton RM, Lauridsen J, 端bel B. 2002. Return on Assets Loss from Situational and Contingency Misfits. Management Scienee 48(11): 1461-1485. Burton RM, 端bel B. 2004. Strategie Organizational Diagnosis and Design. Springer: New York, USo Campbell-Hunt C. 2000. What Have We Learned about Generio Competitive Strategy? A Meta-Analysis. Strategie Management journal 21(2): 127-154. Cao Q, Gedajlovic E, Zhang H. 2009. Unpacking 端rganizational Ambidexterity: Dimensions, Contingencies, and Synergistic Effects. Organization Scienee 20(4): 781-796.


References

154 Cardoso

J,

Bostrom RP, Sheth A. 2004. Workflow Management Systems and ERP

Systems: Differences, Commonalities, and Applications. Information Technology

and Management 5(3-4): 319-338. Caroli E, Van Reenen J. 2001. Skill-Biased Organizational Change? Evidence from a Panel

of British and French Establishments. Quarterly journal 01 Economics 116(4): 1449-1492. Cassiman B, Veugelers R. 2006. In Search of Complementarity in Innovation Strategy: Internal R&D and External Knowledge Acquisition. Management Science 52(1): 6882. Chan YE. 2000. IT Value: The Great Divide Between Qualitative and Quantitative and

Individual and Organizational Measures. journal 01 Management Information Systems 16(4): 225-261. Chan YE, Huff 5L, Barclay DW, Copeland DG. 1997. Business Strategie Orientation, Information Systems Strategie Orientation, and Strategie Alignment. Information Systems Research 8(2): 125-150. Chan YE, Reich BH. 2007. IT alignment: an annotated bibliography. journal of

Information Technology 22(3): 316-396. Chandler AD. 1962. Strategy and structure: Chapters in the history ofAmerican enterprise. MIT Press: Cambridge, MA. Chari MDR, Devaraj S, David P. 2008. Research Note: The Impact of IT Investments and Diversification Strategies. Management Science 54(1): 224-234. Choe J-M. 2003. The effect of environmental uncertainty and strategic applications of IS on a firm's performance. Information and Management 40(4): 257-268. Chow Ge. 1960. Tests of Equality Between Sets of Coefficients in Two Linear Regressions. Econometrica 28(3): 591-605. Clemons EK, Row Me. 1991. Sustaining IT Advantage: The Role of Structural Differences.

MIS Quarterly 15(3): 275-292. Clemons EK, Row Me. 1993. Limits to interfirm cocrdination through information technology: results of a field study in consumer packaged goods distribution.

journal ofManagementlnformation Systems 10(1): 73-95.


References

155

Cohen J, Cohen P, West SG. 2002. Applied Multiple Regression/Correlation Analysis for the

Behavioral Scienees. Lawrence Erlbaum: Hillsdale, NJ,USo Colombo MG, Delmastro M, Rabbiosi L 2007. "High performance" work practices, decentralization, and profitability: evidence from panel data. Industrial and Corporate Change 16(6): 1037-1067. Conant JS, Mokwa MP, Varadarajan PR 1990. Strategie Types, Distinctive Marketing Competencies and Organizational Performance: A Multiple Measures-Based Study.

Strategie Management journal 11(5): 365-383. Coombs R, Knights D, Willmott He. 1992. Culture, Control and Competition: Towards a Conceptual Framework for the Study of Information Technology in Organizations. Organization Studies 13(1): 51-72. Crespi G, Criscuolo C. 2006. Information Technology, Organisational Change and Productivity Growth: Evidence from UK Firms. Working Paper, Centre for Economic Performance, London School of Economics and Political Science, London, UK.

Croteau A-M, Bergeron F. 2001. An information technology trilogy: business strategy, technological deployment and organizational perfonnance. journal of Strategie Information Systems 10(2): 77-99. Curry A, Stancich 1. 2000. The intranet - an intrinsic component of strategic infonnation management? International journal ofInformation Management 20(4): 249-268. Daft RL, Lengel RH. 1984. Information Richness: A New Approach to Managerial Behavior and Organization Design. In Research in Organizational Behavior, Staw BM,Cummings LL (eds.). JAI Press Inc: Greenwich, CT,USo Daft RL, Lengel RH. 1986. Organizational Information Requirements, Media Richness and Structural Design. Management Scienee 32(5): 554-571. Daft RL, Sormunen J, Parks D. 1988. Chief Executive Scanning, Environmental Characteristics. and Company Performance: An Empirical Study. Strategie ManagementjournaI9(2): 123-139. David PA. 1990. The Dynamo and the Computer: An Historical Perspective on the Modem Productivity Paradox. TheAmerican Economie Review80(2): 355-361. David PA, Wright G. 1999. General Purpose Technologies and Surges in Productivity: Historical Reflections on the Future of the ICT Revolution. Working Paper, University of Oxford, Oxford, UK.


References

156

Dedrick J, Gurbaxani V, Kraemer KL. 2003. Information Technology and Economic Performance: A Critical Review of the Empirical Evidence. ACM Computing Surveys 35(1): 1-28. Dehning B, Richardson

VJ. 2002. Returns on Investments in Information Technology: A

Research Synthesis.}ournal oflnformation Systems Management 16(1): 7-30. Dehning B, Richardson

VJ,

Zmud RW. 2003. The Value Relevance of Announcements of

Transfonnational Information Technology Investments. MIS Quarterly 27(4): 637656. Delery JE, Doty DH. 1996. M端des of Theorizing in Strategie Human Resource

Management:

Tests

of

Universalistic,

Contingency,

and

Configurational

Performance Predictions. Academy ofManagement journal39( 4): 802-835. DeSarbo WS, Di Benedetto CA, Song M, Sinha 1. 2005. Revisiting the Miles and Snow strategic framework: uncovering interrelationships between strategic types, capabilities,

environmental

uncertainty,

and

firm

performance.

Strategie

ManagementjournaI26(1): 47-74. Dess GG, Davis PS. 1984. Porter's (1980) Generic Strategies as Determinants of Strategie Group Membership and Organizational Performance. Academy oJ Management

JournoI27(3): 467-488. Deutsche Bundesbank. 2004. The significance of information and communication technology. Monthly Report April 2004, Deutsche Bundesbank, Frankfurt am Main, Germany. Dewan S, Charles S, Gurbaxani V. 2007. Investigating the Risk-Return Relationship of Information Technology Investment: Firm-Level Empirical Analysis. Management

Science 53(12): 1829-1842. Dewan S, Kraemer KL. 2000. Information Technology and Productivity. Evidence from Country-Level Data. ManagementScience 46(4): 548-562. Dewan S, Michael SC, Min C. 1998. Firm Characteristics and Investments in Information Technology: Scale and Scope Effects.lnJormation Systems Research 9(3): 219-232. Dewett T, Jones GR. 2001. The role of information technology in the organization: a review, model, and assessment.journal oJManagement 27(3): 313-346. Diamantopoulos A, Riefler P, Roth KP. 2008. Advancing formative measurement models.

journal oJBusiness Research 61(12): 1203-1218.


References

157

Diamond MA 1991. Dimensions of Organizational Culture and Beyond. Political Psychology 12(3): 509-522.

Doty DH, Glick WH, Huber GP. 1993. Fit, Equifinality, and Organizational Effectiveness: A Test of Two Configurational Theories. Academy of Management journal 36(6): 1196-1250. Drazin R, Van de Yen AH. 1985. Alternative Forms of Fit in Contingency Thecry.

Administrative Science Quarterly 30(4): 514-539. Duncan RB. 1972. Characteristics of Organizational Environments and Perceived Environmental Uncertainty.Administrative Science Quarterly 17(3): 313-327. Duncan RB. 1976. The ambidextrous organization: Designing dual structures for innovation. In The Management of Organization Design, Kilmann RH, Pondy LR, Slevin DP (eds.). Elsevier: NewYork, USo Ellis CA, Gibbs SJ, Rein GL 1991. Groupware: Some issues and experiences.

Communications ofthe ACM 34(1): 38-58. Galbraith JR 1973. Designing Complex Organizations. Addison-Wesley: Reading, MA,USo Galbraith JR 1974. Organization Design: An Information Processing View. Interfaces 4(5): 28-36. Galbraith JR 1977. Organization design. Addison-Wesley: Reading, MA, USo George JF, King J1. 1991. Examining the computing and centralization debate.

Communications ofthe ACM 34(7): 62-72. Ghemawat PJ, Ricart I Costa JE. 1993. The organizational tension between static and dynamic efficiency. Strategic Management journal 14(Winter Special Issue): 59-73. Gibsen C, Birkinshaw J. 2004. The antecedents. consequences, and mediating role of organizational ambidexterity.Academy ofManagement journaI47(2): 209-226. Ginsburg M, Duliba K. 1997. Enterprise-Level Groupware Choices: Evaluating Lotus Notes and Intranet-Based Solutions. journal of Collaborative Computing 6(2-3): 201-225. Grant RM. 2008. Contemporary Strategy Analysis. Blackwell Publishing: Oxford, UK Grant RA, Higgins C. 1991. The Impact of Computerized Performance Monitoring on Service Work: Testing a Causal Model. Information Systems Research 2(2): 116142. Greene WH. 2008. Econometric Analysis. Pearson Education: London, UK.


References

158

Gregor 5, Martin M, Fernandez W, Stern 5, Vitale M. 2006. The transfonnational dimension in the realization of business value from information technology. journal ofStrategic Information Systems 15(3): 249-270. Gresov C. 1989. Exploring Fit and Misfit with Multiple Contingencies. Administrative

Science Quarterly 34(3): 431-453. Gresov C, Drazin R 1997. Equifinality: Functional Equivalence in Organization Design.

Academy ofManagementReview 22(2): 403-428. Groyberg B, Lee L-E. 2009. Hiring Stars and Their Colleagues: Exploration and Exploitation in Professional Service Firms. Organization Science 20(4): 740-758.

Grudin

J.

1994. Groupware and social dynamics: eight challenges f端r developers.

Communications oftheACM 37(1): 92-105. Gulati

R, Puranam P. 2009. Renewal Through Reorganization:

The Value

of

Inconsistencies Between Formal and Informal Organization. Organization Science 20(2): 422-440. Gulati R, Puranam P, Tushman M. 2009. Strategy and the Design of Organizational Architecture, Call for Papers for a Special Issue of the Strategie Management Journal, Strategie Management Society, Chicago, IL, USo Gupta AK, Smith KG, Shalley CE. 2006. The interplay between exploration and exploitation. Academy ofManagement journal49( 4): 693-706. Gurbaxani V, Whang S. 1991. The Impact of Information Systems on Organizations and Markets. Communications oftheACM 34(1): 59-73. Hall WK. 1980. Survival strategies in a hostile environment. Harvard Business Review

58(5): 75-85. Hambrick DC. 1983a. Some Tests of the Effeetiveness and Funetional Attributes of Miles and Snow's Strategie Types. Academy ofManagementjournaI26(1): 5-26. Hambrick DC. 1983b. High Profit Strategies in Mature Capital Goods Industries: A Contingency Approach. Academy ofManagement journal26( 4): 687-707. Hanks SH, Watson CJ, Jansen E, Chandler GN. 1993. Tightening the Life-Cycle Construct: A Taxonomie Study of Gruwth Stage Configurations in

High- Teehnology

Organizations. Entrepreneurship: Theory and Practice 18(2): 5-29. Hannan MT, Freeman JH. 1984. Struetural inertia and organizational change. American

Sociological Review 49(2): 149-164.


References

159

Hansen MT, Nohria N, Thiemey T. 1999. What's Your Strategy for Managing Knowledge?

Harvard Business Review 77(2): 106-118. Hardy M. 1993. Regression with Dummy Variables. Sage: Thousand Oaks, CA, USo Hayek FA. 1945. The use of knowledge in society, American Economie Review 35(4): 519-530. He Z-L, Wong P-K. 2004. Exploration vs. exploitation: An empirical test of the ambidexterity hypothesis. Organization Scienee 15(4): 481-531. Hempell T. 2006. Computers and Produetivity. Physica-Verlag: Heidelberg, Germany. Henderson JC, Venkatraman N. 1993. Strategie alignment: Leveraging information technology for transforming organizations.IBM Systems journaI32(19): 472-484. Hendricks KB, Singhal VR, Stratman JK. 2007. The impact of enterprise systems on corporate perfonnance: A study of ERP, SCM, and CRM system implementations. journal ofOperations Management25(1): 65-82. Hitt LM. 1999. Information Technology and Firm Boundaries: Evidence from Panel Data.

Information Systems Research 10(2): 134-149. Hitt LM, Brynjolfsson E. 1997. Information Technology and Internal Firm Organization: An Exploratory Analysis. journal of Management Information Systems 14(2): 81101. Hitt LM, Wu DJ, Zhou X. 2002. Investment in Enterprise Resource Planning: Business Impact and Productivity Measures. journal of Management Information Systems 19(1): 71-98. Hoetker G. 2007. The use of logit and probit models in strategic management research: critical issues. Strategie Management journal28( 4): 331-343. Holmqvist M. 2004. Experiential Leaming Processes of Exploitation and Exploration Within and Between Organizations: An Empirical Study of Product Development.

Organization Scienee 15(1): 70-81. Huber GP. 1990. A Thecry of the Effects of Advanced Information Technologies on Organizational Design, Intelligence, and Decision Making. Academy ofManagement Review15(l): 47-71. Huselid MA. 1995. The impact of human resource management practices on tumover, productivity, and corporate financial performance. Academy of Management JournoI38(3): 635-672.


References

160

Ichniowski C, Shaw K. 2003. Beyond Incentive Pay: Insiders' Estimates of the Value of

Complementary Human Resource Management Practices. journal 01 Economic Perspectives 17(1): 155-180. Ichniowski C, Shaw K, Prennushi G. 1997. The Effects of Human Resource Management

Practices on Productivity: A Study of Steel Finishing Lines. American Economic Review 87(3): 291-313. Jaccard

JJ,

Turrisi R 2003. Interaction Effects in Multiple Regression. Sage: Thousand

Oaks, CA,USo Jackson SE, Schuler RS, Rivero

Je.

1989. Organizational characteristics as predictors of

personnel practices. Personnel Psychology 42(4): 727-786. Jansen JJP. 2005. Ambidextrous Organizations: A Multiple-level Study of Absorptive Capacity. Exploratcry and Exploitative Innovation, and Performance. Unpublished Dissertation, ERIM, Rotterdam School of Management, Erasmus University, Rotterdam, Netherlands. Available at: ep.eur.nlfhandle/1765/6774 [7 October 2009]. Jansen JJP, Tempelaar MP, Van den Bosch FAJ, Volberda HW. 2009. Structural Differentiation and Ambidexterity: The Mediating Role of Integration Mechanisms. Organization Science 20(4): 797-811. Jansen JJP, Van den Bosch FAJ,Volberda HW. 2005a. Exploratory Innovation, exploitative Innovation, and ambidexterity: The impact of environmental and organizational antecedents. Schmalen bach Business Review 57(4): 351-363. Jansen JJP, Van den Bosch FA], Volberda HW. 2005b. Managing potential and realized absorptive capacity: How do organizational antecedents matter? Academy of ManagementjournaI48(6): 999-1015. Jansen JJP, Van den Bosch FA], Volberda HW. 2006. Exploratory Innovation, Exploitative Innovation,

and Performance:

Effects

of Organizational Antecedents

and

Environmental Moderators. Management Science 52(11): 1661-1674. [arvis CB, MacKenzie SB, Podsakoff PM. 2003. A Critical Review of Construct Indicators and Measurement Model Misspecification in Marketing and Consumer Research. journal ofConsumer Research 30(2): 199-218. Jasperson JS, Carte TA, Saunders CS,Butler BS, Croes HJP, Zheng W. 2002. Review: Power and Information Technology Research: A Metatriangulation Review. MIs Quarterly 26(4): 397-459.


References

161

Kane GC, Alavi M. 2007. Information Technology and Organizational Leaming: An Investigation of Exploration and Exploitation Process. Organization Scienee 18(5): 796-812. Katila R, Ahuja G. 2002. Semething old, something new: a longitudinal study of search behavior and new product introduction. Aeademy of Management journal 45(6): 1183-1194. Kock N. 2004. The Psychobiological Model: Towards a New Theory of ComputerMediated Communication Based on Darwinian Evolution. Organization Scienee 15(3): 327-348. Kohli R, Devaraj S. 2003. Measuring Information Technology Payoff: A Meta-Analysis of Structural Variables in Firm-Level Empirical Research. Information Systems Research 14(2): 127-145.

Kotha S, Vadlamani BL. 1995. Assessing generic strategies: an empirical investigation of two competing typologies in discrete manufacturing industries. Strategie

Management journal 16(1): 75-83. Kyriakopoulos K, Moorman C. 2004. Tradeoffs in marketing exploitation and exploration strategies: The overlooked role of market orientation. International journal of Research in Marketing 21(3): 219-240. Lawrence PR, Lorsch JW. 1967. Organization and environment. Harvard University Press: Boston, MA,USo Leavitt HJ,Whisler TL. 1958. Management in the 1980s. Harvard Business Review 36(6): 41-48. Lee SL. 1994. Electronic Mail as a Medium for Rich Communication: An Empirical Investigation Using Hermeneutic Interpretation. MIS Quarterly 18(2): 143-157. Lee CC, Grover V. 2000. Exploring Mediation Between Environmental and Structural Attributes: The Penetration of Communication Technologies in Manufacturing

Organizations.journal ofManagement Information Systems 16(3): 187-217. Levinthal DA, March JG. 1993. The Myopia of Leaming. Strategie Management journal, 14(Winter Special Issue): 95-112. Lewin AY, Long CP, Carroll TN. 1999. The Coevolution of New Organizational Forms.

Organization Scienee 10(5): 535-550.


References

162

Li M, Ye LR 1999. Information technology and firm perfonnance: Linking with

environmental, strategie and managerial contexts. Information and Management 35(1): 43-51. Lichtenberg FR. 1995. The Output Contributions of Computer Equipment and Personal: A Firm-level Analysis. Economics 01 Innovation and New Technology 3(3-4): 201217.

Loveman GW. 1994. An Assessment of the Productivity Impact of Information Technologies. In Information Technology and the Corporation 01 the 19905:

Research Studies, Allen Tl, Molton SMS [eds.]. Oxford University Press: Oxford, UK. Lubatkin MH,Simsek Z, Ling Y,Veiga JF. 2006. Ambidexterity and Performance in Smallto Medium-Sized Firms: The Pivotal Role of Top Management Team Behavioral

Integration.journal ofManagement 32(5): 646-672. Malone TW, Yates J, Benjamin RI. 1987. Electronic Markets and Electronic Hierarchies,

Communications oftheACM 30(6): 484-497. March JG. 1991. Exploration and exploitation in organizational leaming. Organization

Science2(1): 71-87. Markus ML. 1994. Electronic Mail As the Medium of Managerial Choice. Organization

Science 5(2): 502-527. Mata FJ, Fuerst WL, Bamey JB. 1995. Information Technology and Sustained Competitive Advantage: A Resource-Based Analysis. MIS Quarterly 19(4): 487-505. May RC, Stewart WH, Sweo R. 2000. Environmental scanning behavior in a transitional economy: Evidence from Russia.Academy ofManagementjournaI43(3): 403-427. Melville N, Kraemer KL, Gurbaxani V. 2004. Review: Information Technology and Organizational Performance: An Integrative Model of IT Business Value. MIS

Quarterly 28(2): 283-322. Miles RE, Snow Ce. 1978. Organizational strategy, structure, and process. McGraw-Hill: NewYork, USo Miles RE, Snow Ce. 1984. Designing strategie human resources systems. Organizational

Dynamics 13(1): 36-52. Milgrom P, Roberts J. 1990. The Economics of Modern Manufacturing: Technology, Strategy, and Organization. American Economic Review 80(3): 511-528.


References

163

Milgrom P, Roberts J. 1995. Complementarities and fit strategy, structure, and organizational change in manufacturing. journal of Accounting and Economics 19(2-3): 179-208. Miller A, Dess GG. 1993. Assessing Porter's (1980) model in terms of its generalizability, accuracy and simplicity.journal ofManagement Studies 30(4): 553-585. Miller D. 1992. The Generio Strategy Trap.journal ofBusiness Strategy 13(1): 37-41. Miller D, Friesen PH. 1985. A Longitudinal Study of the Corporate Life Cycle.

ManagementScience 30(10): 1161-1183. Miller D, Friesen PH. 1986. Porter's (1980) Generio Strategies and Performance. An Empirical Examination with American Data: Part II: Performance Implications.

Organization Studies 7(3): 255-261. Miller D, Toulouse JM. 1998. Quasi-rational Organizational Responses: Functional and Cognitive Sources of Strategie Simplicity. Canadian journal of Administrative Seienees 15(3): 230-244. Mintzberg H. 1978. Patterns in Strategy Formation. Management Science 24(9): 934948. Mintzberg H. 1979. The structuring of organizations: A synthesis of the research. New York: Prentice-Hall. Mirani R, Lederer AL.1998. An instrument for assessingthe organizational benefits of IT projects. Decision Seiences 29(4): 803-838. Miravete EJ, Pemias Je. 2006. Innovation complementarity and scale of production.

journal oflndustrial Economics 54(1): 1-29. Mohnen P, Rรถller L-H. 2005. Complementarities in innovation policy. European Economic

Review49(6): 131-1450. Mom TJM,Van Den Bosch FAJ, Volderba HW. 2003. Managing concurrently the process of knowledge exploration and exploitation: the influence of knowledge exchange determinants and the introduction of an Intranet. Working Paper, Department of Strategie Management and Business

Environment,

Rotterdam

School

of

Management, Erasmus University, Rctterdam, Netherlands. Mom TJM, Van Den Bosch FA], Volderba HW. 2009. Understanding Variation in Managers' Ambidexterity: Investigating Direct and Interaction Effects of Formal Structural and Personal Cocrdination Mechanisms. Organization Science 20(4): 812-828.


164

References

Moore M. 2005. Towards a Confirmatory Model of Retail Strategy Types: An Empirical Test ofMiles and Snow. journal ofBusiness Research 58(5): 696-704. Morand DA. 1995. The Role of Behavioral Formality and Informality in the Enactment of Bureaucratic versus Organic Organizations. Academy ofManagement Review 20(4): 831-872.

Mukhopadhyay T, Rajiv S, Srinivasan K. 1997. Information Technology Impact on Process Output and Quality. ManagementScience 43(12): 1645-1659. Nadler DA, Gerstein MS, Shaw RB. 1992. Organizational Architecture. Jossey-Bass: San Francisco, CA. Nahm AY, Vonderembse MA, Koufteros XA. 2003. The impact of organizational structure on time-based manufacturing and plant perfonnance. journal of Operations

Management21(3): 281-306. Nault BR. 1998. Information Technology and Organization Design: Locating Decisions and Information. Management Science 44(10): 1321-1335. Ngwenyama OK, Lee AS. 1997. Communication Richness in Electronic Mail: Critical Social Theory and the Contextuality ofMeaning. MIS Quarterly 21(2): 145-167. Norton EC, Wang H, Ai C. 2004. Computing interaction effects and standard errors in logit and probit models. The StatajournaI4(2): 154-167. Nunally J. 1978. Psychometrie theory. McGraw-Hill: New York, USo Oh W, Pinsonneault A. 2007. On the Assessment of the Strategie Value of Information Technologies: Conceptual and Analytical Approaches. MIS Quarterly 21(2): 239265. O'Reilly C, Tushman ML. 2008. Ambidexterity as a Dynamic Capability: Resolving the Innovator's Dilemma. In Research in Organizational Behavior, Staw BM (ed.). JAI Press: Greenwich, CT,USo Orlikowski WJ. 1992. The Duality of Technology: Rethinking the Concept of Technology in Organizations. Organization Science 3(3): 398-427. Orlikowski WJ, Robey D. 1991. Information Technology and the Structuring of

Organizations.lnformation Systems Research 2(2): 143-169. Osterman P. 1994. How Common is Workplace Transformation and Who Adopts it?

lndustrial and Labor Relations Review 47(2): 173-188.


References

165

Parnell JA 1997. New Evidence in the Generic Strategy and Business Performance Debate: A Research Note. British journal ofManagement 8(2): 175-81. Parnell JA, Hershey 1. 2005. The Strategy-perfonnance relationship revisited: The blessing and curse of the combination strategy.International journal of Commeree and Management 15(1): 17-33. Parnell JA, Lester DL, Menefee M1. 2000. Strategy as a response to organizational uncertainty: an alternative perspective on the strategy-performance relationship.

ManagementDecision 38(8): 520-530. Parnell JA, Wright P.1993. Generic Strategy and Performance: an Empirical Test ofthe Miles and Snow Typology. Britishjournal ofManagement4(1): 29-36. Pfeffer J, Leblebici H. 1977. Information Technology and Organizational Structure.

Paeifie Sociological Review 20(2): 241-261. Pinsonneault A, Kraemer KL. 1997. Middle Management Downsizing: An Empirical Investigation of the Impact of Information Technology. Management Scienee 43(5): 659-679. Podsakoff PM, MacKenzie SB, Podsakoff NP. 2003. Common Method Biases in Behavioral Research: A Critical Review of the Literature and Recommended Remedies.journal

0/Applied Psychology 88(5): 879-903. Porter ME. 1980. Competitive Strategy. Free Press: New York, USo Porter ME. 1985. Competitive Advantage. Free Press: New York, USo Porter ME.1996. What is astrategy? Harvard Business Review 74(6): 61-78. Powell TC, Dent-Micallef A 1997. Information Technology as Competitive Advantage: The Role of Human, Business, and Technology Resources. Strategie Management

JournoI18(5): 375-405. Prekumar G, Ramamurthy K, Saunders CS. 2005. Information Processing View of Organizations:

An

Exploratcry

Examination

of

Fit

in

the

Context

of

Interorganizational Relationships. journal of Management Information Systems 22(1): 257-294. Pugh DS. 1973. The Measurement of Organizational structures. Does context determine form? Organizational Dynamies 1(4): 19-34. Radner R

1993. The Organization of Decentralized Information Processing. Econometrica 61(5): 1109-1146.


References

166

Raisch 5, Birkinshaw J. 2008. Organizational Ambidexterity: Antecedents, Outcomes, and

Moderators.Journal ofManagement 34(3): 375-409. Raisch 5, Birkinshaw

J,

Probst G, Tushman ML. 2009. Organizational Ambidexterity:

Balancing Exploitation and Exploration f端r Sustained Performance. Organization

Science 20(4): 685-695. Raisch 5, Hotz F. 2010. Shaping the context f端r leaming: Corporate alignment initiatives, environmental munificence, and firm perfonnance. In Strategie reconfigurations:

Building dynamic capabilities in rapid-innovation-based industries, Wall 5, Zimmermann C, Klingebiel R, Lange D (eds.). Edward Elgar: Cheltenham, UK. Rivard 5, Raymond L, Verreault D. 2006. Resource-based view and competitive strategy:

An integrated model of the contribution of infonnation technology to firm performance. journal ofStrategic Information Systems 15(1): 29-50. Roach SS. 1987. America's Technology Dilemma: A Profile ofthe Information Economy. Special Economic Study, Morgan Stanley, New Ycrk, USo Roberts

J.

2007. The modern firm: Organizational design for performance and growth.

Oxford University Press: New York, USo Robey D. 1977. Computers and Management Structure: Some Empirical Findings Reexamined. Human Relations 30(1): 963-976. Robey D, Boudreau M-C. 1999. Accounting for the Contradietory Organizational Consequences

of

Information Technology: Theoretical

Directions

and

Methodological Implications.Information Systems Research 10(2): 167-185. Roodman D. 2006. How to Do xtabon2: An Introduction to "Difference" and "System" GMM in Stata. Working Paper, Center for Global Development; Washington DC, USo Rosenberg N, Trajtenberg M. 2001. A General Purpose Technology at Work: The Corliss

Steam Engine in the Late 19th Century. Discussion Paper 3008, Centre for Economic Policy Research, London, UK. Rothaermel FT, Alexandre MT. 2009. Ambidexterity in Technology Sourcing: The Moderating Role of Absorptive Capacity. Organization Science 20(4): 759-780. Sabherwal R, Chan YE. 2001. Alignment Between Business and IS Strategies: A Study of Prospectors, Analyzers, and Defenders.Information Systems Research 12(1): 11-33. Sabherwal R, Hirschheim R, Goles T. 2001. The Dynamics of Alignment: Insights from a Punctuated Equilibrium Model. Organization Science 12(2): 179-197.


References

167

Sabherwal R, Jeyaraj A, Chowa C. 2006. Information System Success: Individual and Organizational Determinants. Management Scienee 52(12): 1849-1864. Saloner G, Shepard A, Podolny J. 2001. Strategie Management. John Wiley and Sons: New York, USo

Schein EH. 1985. Organizational Culture and Leadership. Jossey-Bass: San Fransico, CA, USo

Schuler RS, Jackson SE. 1987. Linking Competitive Strategies with Human Resource Management Practices. Academy ofManagement Exeeutive 1(3): 207-219. Sheremata WA. 2000. Centrifugal and Centripetal Forces in Radical New Product Development under Time Pressure. Academy of Management Review 25(2): 389408. Shim JP, Warkentin M, Courtney JF, Power DJ, Sharda R, Carlsson C. 2002. Past, present, and future of decision support technology. Decision Support Systems 33(2): 111-

126. Simsek Z. 2009. Organizational Ambidexterity: Towards a Multilevel Understanding.

journal ofManagementStudies 46(4): 597-624. Smith WK, Tushman ML. 2005. Managing Strategie Contradictions: A Top Management Model for Managing Innovation Streams. Organization Scienee 16(5): 522-536. SolowR. 1987. We'd BetterWatch Out. New York Times Book Review. July 12, 36. Spanos YE, Zaralis G, Lioukas S. 2004. Strategy and industry effects on profitability: evidence from Greece. Strategie Management journal 25(2): 139-165. Stieglitz N, Heine K. 2007. Innovations and the role of complementarities in a strategie theory of the firm. Strategie Management journal 28(1): 1-15. Stohr EA, Zhao JL 2001. Workflow Automation: Overview and Research Issues.

Information Systems Frontiers 3(3): 281-296. Sun L-Y, Aryee S, LawKS. 2007. High-Performance human resource practices, citizenship behavior, and organizational performance: A relational perspective. Academy of

Management journal 50(3): 558-577. Tanriverdi H. 2006. Performance Effects of Information Technology Synergies in Multibusiness Firms. MIS Quarterly 30(1): 57-77. Thomhill S, White RE. 2007. Strategie purity: A multi-industry evaluation of pure vs. hybrid business strategies. Strategie Management journal 28(5): 553-561.


168

References

Townley B. 1993. Power/Knowledge, and Its Relevance for Human Resource Management. Academy ofManagement Review 18(3): 518-545. Tushman ML, Nadler DA. 1978. Information Processing as an Integration Concept in Organizational Design. Academy ofManagement Review 3(3): 613-624. Tushman ML, O'Reilly C. 1996. Ambidextrous organizations: Managing evolutionary and revolutionary change. California Management Review 38(4): 8-30. Uotila

J,

Maula M, Keil T, Zahra SA. 2009. Exploration, exploitation, and financial

performance: analysis of S&P 500 corporations. Strategie Management journal 30(2): 221-231. Van de Yen AH, Drazin R. 1985. The Concept ofFit in Contingency Theory. In Research in

Organizational Behavior, Staw BM, Cummings LL (eds.). HAI Press: Greenwich, CT,

us Van Looy B, Martens T, Debackere K 2005. Organizing for Continuous Innovation: On the Sustainability of Ambidextrous Organizations. Creativity and Innovation

Management 14(3): 208-221. Venkatraman N. 1989. The Concept of Fit in Strategy Research: Toward Verbal and Statistical Correspondence. Academy ofManagement Review 14(3): 423-444. Venkatraman N, Lee C-H, Iyer B. 2007. Strategie Ambidexterity and Sales Growth: A Longitudinal Test in the Software Sector. Working Paper, Boston University School of Management, Boston, MA,USo Wade M, Hulland

J.

2004. Review: The Resource-Based View and Information Systems

Research: Review, Extension, and Suggestions for Future Research. MIS Quarterly 28(1): 107-142. WeiH P. 1992. The Relationship Between Investment in Information Technology and Firm Performance: A Study of the Valve Manufacturing Sector.Information Systems Research 3(4): 307-333.

WeiH P, Aral S. 2006. Generating premium returns on your IT investments. Sloan ManagementReview 47(2): 39-48. Wernerfelt B. 1984. A Resource-Based View of the Firm. Strategie Management journal 5(2): 171-180. White RE. 1986. Generic Business Strategies, Organizational Context and Performance: An Empirical Investigation. Strategie Management journal 7(3): 217-231.


References

169

Wooldridge JM. 2009. Introductory Econometrics. Thomson South-Western: Mason, Australia. Wright PM, Snell SA. 1998. Toward a Unifying Framework for Exploring Fit and Flexibility in Strategie Human Resource Management. Academy 01 Management Review23(4): 756-772. Wyner GM, Malone TW. 1996. Cowboys or Commanders: Does Information Technology Lead to Decentralization. Conference Proceedings. International Conference on Information Systems, Columbus, OH, USo Youndt MA, Snell SA, Dean JW, Lepak D. 1996. Human Resource Management, Manufacturing Strategy, and Firm Performance. Academy 01 Management journal 39(4): 836-866. Zammuto RF, Griffith TL, Majchrzak A, Dougherty J, Faraj S. 2007. Information Technology and the Changing Fabric of Organization. Organization Science 18(5): 749-762.


Appendices Appendix 3.1: Survey instrument for the corporate learning variable (LEARN) Survey items on products (1) In the last five years, has yaur establishment concentrated more on upgrading existing products and production processes or on introducing new products? (2) In the last five years, has the rest of yaur firm concentrated more on upgrading existing products and production processes or on introducing new products?

Seale fram 1 ('existing productsJ to 5 ('new productsJ. Survey items on customer groups(markets (3) In the last five years, has yaur establishment concentrated more on existing customer groups and markets as weil as their needs or on opening up new customer groups and markets? (4) In the last five years, has the rest of yaur firm concentrated more on existing customer groups and markets as weil as their needs or on opening up new customer groups and markets?

Seale from 1 ('existing eustomer groupsjmarkets] to 5 ('new eustomer groupsjmarkets].

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6, Š Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


1 72

Appendices

Appendix 3.2: Survey instrument tor the decentralization of decision rights variable (DR) Survey items (1) Who (2) Who (3) Who (4) Who (S) Who (6) Who

decides on delivery time and priority of orders? issues production plans? distributes work among workers? decides howwork is done (for example, pace and order)? decides which machines and tools are used? coordinates different production steps?

Seale from 1 ('exclusively managers] ta 5 ('exclusively workers].

Alpha .67


173

Appendices

Appendix 3.3: Survey instrument for the URMpractices variable (UR) (1) Intensity of (self-managed) teamwork Prepared guestions Could you describe the tasks thatworker teams fulfill? Is teamwork used routinely or irregularly? How much discretion do teams have in their daily work? Which roles do manaeers lav for worker teams? Examnle answer for score 1 Examnle answer for score 3 Teams are used irregularly or eams are used routinely. only ifrequired. rrhe teams have no discretion. heir work is regulated by manazers or bv other means.

Examnle answer for score 5 eams are used routinely. rrhe teams are self-managed. Managers rather playa sunnortive role.

(2) Variety of dailywork Prepared guestions Could you describe how many differenttasks a worker is usually qualified to undertake? To what extent do these tasks differ from each other? Why do workers rotate between these tasks? How often do workers switch between these tasks? Examule answer for score 1 Examule answer for score 3 Examule answer for score 5 lWorkers are qualified for Workers are qualified for one lWorkers are qualified for task and only pursue this lightly different tasks within clearly differenttasks and task one step of the production regularly switch between process. They switch between ~~m. hese tasks from time to time or if reouired. (3) Amount and quality of information Prepared guestions On which topics do you make sure that information is always provided for workers? Why should workers be informed on these topics? How do workers receive this information? How often do workers receive this information? xample answer for score 5 Example answer for score 1 Example answer for score 3 Workers do not receive fWorkers receive information n addition to what is information from the firm or actively, regularly, and in an ~escribed in score 3, workers the firm does not actively, accessible way. This ctively, regularly, and regularly, andjor in a information is needed for the ccessibly receive accessible way inform functioning of each worker's information on other sections bf the establishment or firm workers. own job (for example, information regarding the i n order to coordinate with hese sections. quantity and quality produced in each worker's broduction steu1.


174

Appendices

(4) Training intensity and variety Prepared guestions Could you describe the training programs that are provided for workers? What are the contents of these training programs? What are the goals of these training programs? How often da these trainina urozrams take ulace? Example answer for score 1 Example answer for score 3 Example answer for score 5 Workers da not participate in Workers participate in In addition to what is regular training programs regular training programs ~escribed in score 3, workers that upgrade or expand their that are directed at upgrading participate in regular training capabilities. There are no or expanding existing programs that are directed at training sessions, or only capabilities that are needed acquiring additional training sessions that are for each worker's job. capabilities from other areas mandated by law (for or general capabilities like example, regarding job leadership capabilities or sosafe~) take nlace. called 'soft skills '. (5) Performance review and pay Prepared guestions Couldyou describe howthe individual performance ofworkers (or the performance of their workspace or team) is reviewed? How often do performance reviews take place? What are the conseauences of the nerformance reviews for worker nav? Examnle answer for score 1 tExamnle answer for score 3 Examnle answer for score 5 Individual, workspace, or ndividual, workspace, or Individual, workspace, or team performance is not eam performance is eam performance is reviewed regularly. Workers eviewed regularly. reviewed regularly. receive a fixed pay. lPerformance reviews do not Performance reviews imoact oav. stronelv imnact oav. (6) Promotion criteria Prepared guestions Imagine that a higher position is free and that one of &0 candidates for the position has been working for your company five more years than the other candidate. Who will be promoted? Aside from leadership skills, which kind ofknowledge and background is needed for nositions at hiaher hierarchicallevels in vour comuanv? xamole answer for score 5 Examule answer for score 1 Examule answer for score 3 Employees are exclusively Employees are promoted on mployees are mainly IPromoted on the basis of promoted on the basis of fthe basis of special tenure andjor age. knowledge and experience in ~eterogeneous knowledge fthe area to which they are nd experiences (for example, promoted. ~ifferent areas of expertise, ~epartments, functions, stablishments, and so on1.


175

Appendices

Appendix 4.1: Survey instrument for the corporate strategy variables Survey items: During the last three years, a key part ofthe flrm's strategy was

Factor 1 Factor 2 Alpha

Exploratory strategy ... to produce goods incorporating the newest technology. ... to produce goods for high-price markets. ... to produce unique and special goods. ... to develop new products. ... to enter new technological fields. ... to serve market niches. ... to serve diverse markets and customer groups. ... to enter new markets and customer groups.

.23 -.01 -.04 .33 .31 -.06 .16 .31

.58 .65 .66 .57 .54 .66 .55 .54

Exploitative strategy efficiently manage inventory. increase production automation. reduce costs. increase productivity. strictly control quality. increase quality and reliability of existing products. keep deliverytimes short gather customer information. measure customer satisfaction.

.63 .54 .60 .61 .60 .56 .54 .55 .59

.06 -.02 .01 .09 .12 .27 -.01 .23 .15

Scale fram 1 ('I do not agree at all') to 5 ('I strangly agree').

.76

.76


176

Appendices

Appendix 4.2: Survey instrument for the decentralization variable Survey items: Hierarchicallevel x ofy hierarchicallevels makes the decision on ...

Alpha

... the pace ofwork for production workers. .88 ... the methods ofwork for production workers. ... the allocation ofwork to production workers. ... the daily production plans. ... the weekly production plans. ... overtime in production. ... meeting problems in the production process. ... the reaction to customer problems and complaints. ... the machines and tools used in production. ... the creation of a new full-time position. ... hiring or not hiring an applicant for an open position in production. ... hiring or not hiring an applicant for an open position in middle management. ... the methods of employee selection. ... the criteria for employee appraisal and bonuses. ... promotions. ... wage increases. ... the layoff of an employee. ... the establishment of rules and procedures for managers. ... the establishment of rules and procedures for workers. ... investments greater than one percent of the yearly firm-wide investment budget ... greater changes in production. ... the introduction of a new product. entering a new market. ... product pricing.


Appendices

177

Appendix 4.3: Survey instrument for the standardization variable Survey items: In your firm, ... ... there are manywritten or oral rules and procedures for workers. ... there are manywritten or oral rules and procedures for managers.

Scale fram 1 ('I do not agree at all') to 5 ('I strangly agree').

Alpha

.56


178

Appendices

Appendix 4.4: Survey instruments for the HRM practices variables Employee appraisal Open questions to respondent Please describe howworker performance is appraised. Which criteria are used to appraise worker performance? Da employee talks take place? Ifyes, please describe the typical process and contents of such a talk. Please describe the effect of performance appraisals on wages. How often da performance appraisals take place? Rating questions to interviewer (sealefram 1 ('not importantat att] to 5 ('very important]J (3) How important is the role of objective criteria, such as quantitative or qualitative individual performance, teamjdepartmentjshift performance, or firm performance, as critera during employee appraisal? (exploitative component) (1) How important are the roles of subjective criteria, such as interpersonal skills, problemsolving skills, and personality traits (for example, for team work andjor autonomous decision-making and acting) or their improvement, as criteria during employee appraisal, and how important is the role of behavior, including suggestions for improvement and the pro-active participation in decision-making processes, as a criterion during employee appraisal? (exploratory component)

Scale fram 1 ('not important at all') to 5 ('very important').


Appendices

179

Employee selection Open questions to respondent Please describe how and where your firm searches for candidates for open production positions. Please describe the selection process; that is, how does your firm proceed to decide for or against a candidate? Which criteria are used to decide for or against a candidate? Which knowledge, abilities and characteristics are most important? Which further knowledge, abilities, and characteristics are important? Who participates in the selection talks? Rating questions to interviewer (1) How important are the roles of the specific professional and functional qualifications as decision criteria during employee selection? (exploitative component) (2) How important are the roles of general abilities like interpersonal skills, problemsolving skills, and personality traits (for example, for team work andjor autonomous decision-making and acting) as decision criteria during employee selection, and how important is the role of the long-term potential of the candidate as adecision criterion during employee selection? (exploratDrycomponent)

Scale fram 1 ('not important at all') to 5 ('very important').


180

Appendices

Employee training Open questions to respondent Please describe the training measures in which workers participate. What are the most important contents and goals of these trainings measures? What existing knowledge and abilities need to be refreshed and improved? What new knowledge and abilities need be learned? How often da workers participate in such training measures? Rating questions to interviewer (1) How important are the roles of training measures with respectto (a) the goal of building or refreshing and improving specific functional abilities and knowledge of rules and procedures, (b) the goal of downtime and scrap rate reductions, and Ce) the goal of quality improvements? (exploitative component) (2) How important are the roles of training measures with respeet to (a) the goal of building an inclusive understanding of the produetion proeess and the firm, (b) the goal of building knowledge and abilities for different tasks and higher positions, and (e) the goal of building or refreshing and improving general abilities like interpersonal skills, problem-solving skills, and personality traits (for example, for team work andjor autonomous decision-making and aeting)? (exploratory component)

Scale fram 1 ('not important at all') to 5 ('very important').


Appendices

181

Appendix 4.5: Survey instrument for the ICTuse variables Survey items: In your firm, ICTis used to

ICT use for decentralization endowworkers with more and better information. facilitate autonomous coordination ofworkers from different departments, functions, divisions, andjor plants. support the work of autonomous worker teams. increase worker participation in decision processes. collect ideas from workers for improvement ICT use for centralization endow managers with more and better information. enable managers to make more and better decisions. increase managerial span of contro!.

Scale fram 1 ('I do not agree at all') to 5 ('I strangly agree').

Factor 1 Factor 2 Alpha .81 .69

.23

.81

.13

.82 .74 .54

.18 .30 .28

.18 .24 .12

.82 .86 .74

.76


Appendices

182

Appendix 4.6: Survey instrument for the environmental dynarnism variable Far each of the following seven areas [customers/market, competitors, technology, labor market/untons, capital market, legislators/regulators, and socio-cultural environment), please tell me .

how much and how fast have trends, demands, and conditions changed during the last three years?

1:1 [ (rate of change Scale fram 1 ['very tittle and slowJ to 5 ['very much and fast').

+

... how weil were you able to anticipate the emergence of these changes and their consequences for your firm duringthe last three years?

how irnportant for your firm's goals and thus performance is this area?

predictabilityJ

irnportance ]

Scale fram 1 ['very poorlyJ to 5 ['very qood'].

Scale fram 1 ['not irnportant at allJ to 5 ['very

important').


Subj ect index Amb idexterity ....•...•...•....•...•....•...•... 9 2,95 Centra lizat ion .... 29,34, 62, 72

Human resource management practices

CI Tech nology Dat ab ase; _ _B, 68, 110

Hyb rid stra tegy 89, 91, 9 3 Individual a mbidexterity _ _ 99, 115

w . . ._ •••• _ ••• _ ••••

Coercive sta ndardization _ _ 39, 100 Complementa rity .._..._...._ _ _ _ 28,3 0 Computer productlvtty paradox., _..._19 Contingen c.y theory _ _ _ 23,27, 146 Contro l-ori ented lT Coordin ation-ori ent ed IT

43 ..43

Corporate strategy Dateset.•....

M •••M •• • • M • • • _

•• • _

•••• • ••• • ••• •

. ..... .... ...............................40, 73, 100, 116

Inform ation pro cessin g theory _ 29, 60 lnfonnation technolo gy._..._...._42,61,70

lntranet , ... IT complement s, IT expen ditures N .. .. N ... N

N

N

N .... _

N

_

•••• _

_

••• _

N

••• _

•••• _

••• _

••••

•••• _

••• _

••

102, 118

_19, 21, 30 •••• _

••• _

•••• _

••

1

48, 114

Moderati on effects model;., _.. 106, 124

_.8, 68, 110

ORBIS dat ab ase 9,68, 111 Organizat ion design .._ _ _ _ _...._..._..._

Decentralization...._..._...._..._..._...._..._...._..... .. .... ..... .... ..... ..... .... ...29, 35, 62, 72, 99 , 115 Dum my-codi ng rnodet..., _ _ 1OB, 127

. ..... ..... .... ....29, 34, 60, 72, 9 1, 9 3, 98, 115 Organizat iona l a mbidexterity _ 95, 98 Organizat iona l culture..._ _ _ _ _..._50

Effects-eoding modeL l07, 127 Enab ling standa rdization _ _ 39, 100

Orga nizat iona l learning , _ _ _ 63, 71 Produ ctivity .. _.1

Euterprise resou rce planning softwa re

Res ource -based view _...._..._..._ _ _

...... ..... .... .... ..... .... ...................... ... 102, 118 Enviro ument 46

Resource-cente red perspective _ _ 22 Software.... 10 2

Exploitation

_.. 49, 63, 64, 114, 13 5

Standardization

Exploration

49, 63, 64, 114, 135

Demanding envtrc nrnent..., _ _..._.....48

N

N

N ... _

•••• _

••• _

••• _

•••• _

••• _

•••• _

Firm performance . .................................... 32, 60, 71, 95, 113, 131 Fit 27, 30 General purpose tec hnology _ _ _..._.2 Groupware software.._..._...._ _ 102, 118 Hori zontal flt ., ...

N .... _

• •• _

• • •• _ • ••

27,30,33, 44

N . . .N . . .N . . .. _

N .... N ... N

N

N

••• _

_

• ••• _

•••• _

••• _

••• _

• •• _

••• _

• ••• _

•••• _

• ••••

22

••

_.38, IOD, 11 6

Technological amb idexte rity Telephone survey

101, 118 9, 68, 111

Theory of producnon 22, 66 Undemanding environment _ _ _..... 47 Vertical fit

N .... N ... N .... N .. ... .. .. _

••••

27, 3 D, 46, 52

Vert ical inform ation proce ssing ._..._....29 Workflow software ...._..._...._..._... 102, 118

Hori zontal inform at ion pro cessing, .....29

F, Mahr, Aligning Information Technology, Organization, and Strategy, DOI 10.1007/ 978-3-8349-6072-6, © Gabler Verlag I Springer Fachmedien Wiesbaden GmbH 2010


Aligning Information Technology, Organization, and Strategy