A MULTILEVEL FRAMEWORK OF THE ROLE OF ENVIRONMENTAL AND
ORGANIZATIONAL CONTEXT ON RISK TAKING

By
Kalin Kolev

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Strategic Management

2012

ABSTRACT
A MULTILEVEL FRAMEWORK OF THE ROLE OF ENVIRONMENTAL AND
ORGANIZATIONAL CONTEXT ON RISK TAKING
By
Kalin Kolev
In this dissertation I examine one of the critical processes for each organization –
organizational risk taking. To achieve this goal I develop a multilevel framework that explicates
on the antecedents, consequences and moderating conditions of organizational risk taking.
While prior research indicates that risk taking is driven by discrepancy between
performance and aspiration levels, little work has been done on whether all organizations
respond in a similar fashion to poor performance. Using a sample of corporate divisions in
multidivisional corporations (M-firms), I argue that a set of factors from the organizational and
structural context affect the ability and motivation of managers to respond to poor performance
by engaging in risk taking. The structural contingencies of the M-firm also have important
implications for the consequences of divisional risk taking by impacting the ability of divisional
managers to undertake high-quality risky choices. My results suggest that a set of factors
residing at the divisional, corporate, and environmental levels of analysis significantly affects the
risk taking behavior of managers. Overall, this dissertation aims to advance our knowledge on
the embeddeddness of managerial risky choices in a larger organizational and environmental
context and offer insight into how this context impacts organizational risk taking.

DEDICATION

To my lovely mother for all the love and support she gave me and continues to give

iii

ACKNOWLEDGEMENTS
This dissertation marks the end of a long but rewarding period of my life and also serves as a
starting point to an exciting new chapter of being a university professor. Looking back at this 5year journey, I want to thank many people who have helped along the way.
First and foremost, I want to thank my mom Sonya for everything she has done for me.
She has taught me honesty, integrity, work dedication, and courage to overcome challenges in
life. She has been the perfect example of how to deal with difficulties and disappointments and
how to never give up and fight for your dreams. Without her, I would not have achieved
anything in my life.
Second, I want to express my appreciation to my advisor Gerry McNamara. Through
constant advice, mentoring, and constructive comments, Gerry has pushed me to become a better
person and a better scholar. I will always be grateful to Gerry, an excellent scholar and a great
person, for giving me an excellent start in the academic career and I wish every student had such
an excellent advisor as Gerry.
Third, I want to thank my committee members for their excellent job and directions
during the dissertation process. To Bob Wiseman – who has worked with me from the start of the
PhD program and has been an excellent and patient mentor who has taught me a lot and has
supported me in all of my endeavors. To John Haleblian – who is not only a great scholar but
also one of the first people that believed in me and has a major contribution to my first academic
publications. To Margaret Hughes-Morgan – she is an excellent person and friend who gave me
encouragement and advice during the long journey of a PhD student.

iv

Fourth, I would to thank all professors in the Management department for their dedication
and hard work during doctoral seminars and for serving as an example of achieving a successful
academic career.
Finally, I want to thank all fellow doctoral students for those great 5 years in my life and
for making me feel comfortable and appreciated at MSU. In particular, I thank Jonathan Miles
and my cohort mates Seungho Choi and Dustin Sleesman for being excellent and true friends.

v

TABLE OF CONTENTS
LIST OF TABLES ......................................................................................................... viii
LIST OF FIGURES .......................................................................................................... x
INTRODUCTION............................................................................................................. 1
CHAPTER 1: RISK TAKING AT THE DIVISIONAL LEVEL OF M-FIRMS ANTECEDENTS AND MODERATING CONDITIONS ................................... 16
INTRODUCTION........................................................................................................... 16
HYPOTHESES DEVELOPMENT ............................................................................... 18
Risk taking at the divisional level ............................................................................. 21
Intensity of divisional risk taking ............................................................................. 22
Factors affecting intensity of risk taking in divisions ............................................... 26
Structural factors: Level of corporate diversification ............................................... 28
Structural factors: Divisional size ............................................................................. 29
Structural factors: Corporate governance mechanisms............................................. 31
Resource endowment factors: Corporate slack resources ......................................... 36
METHODS ...................................................................................................................... 38
Sample 1.................................................................................................................... 38
Dependent variable ................................................................................................... 39
Independent variables ................................................................................................ 40
Control variables ....................................................................................................... 42
Sample 2 .................................................................................................................... 42
Dependent variable .................................................................................................... 43
Independent variables ................................................................................................ 44
Control variables ....................................................................................................... 46
Analysis..................................................................................................................... 46
RESULTS ........................................................................................................................ 49
DISCUSSION .................................................................................................................. 53
APPENDICES ................................................................................................................. 90
APPENDIX 1 ................................................................................................................... 91
CHAPTER TWO: OUTCOMES OF DIVISIONAL RISK TAKING - HOW THE
STRUCTURAL EMBEDDEDNESS OF DIVISIONS IN ......................................
THE M-FIRM AFFECTS THE QUALITY OF THEIR RISK TAKING....... 102
INTRODUCTION ................................................................................................. 102
HYPOTHESES DEVELOPMENT...................................................................... 106
Do divisions take on “good” or “bad” risks? – The moderating role of
M-firm structure.................................................................................................... 106
Related diversification and “good” risk taking ....................................................... 109
Unrelated diversification and “bad” risk taking ...................................................... 111
The moderating impacts of industry, corporate and divisional characteristics
on the relationship between divisional risk taking and performance ...................... 113
vi

Industry factors affecting the relationship between divisional risk taking and
performance ............................................................................................................ 115
Corporate governance factors affecting the relationship between risk taking and
performance ............................................................................................................ 119
Divisional factors affecting the relationship between risk taking and performance124
METHODS .................................................................................................................... 125
Sample 1 .................................................................................................................. 126
Dependent variable .................................................................................................. 127
Independent variables .............................................................................................. 127
Control variables ..................................................................................................... 129
Sample 2 .................................................................................................................. 129
Dependent variable .................................................................................................. 130
Independent variables .............................................................................................. 130
Control variables ..................................................................................................... 132
Analysis ................................................................................................................... 133
RESULTS……………………………………………………………………………...136
DISCUSSION ................................................................................................................ 139
CONCLUSIONS AND FUTURE DIRECTIONS ...................................................... 146
APPENDICES ............................................................................................................... 167
APPENDIX 2 ................................................................................................................. 168
REFERENCES .............................................................................................................. 175

vii

LIST OF TABLES
TABLE S: Summary of hypotheses and findings ............................................................. 64
TABLE A1: Descriptive statistics and correlation coefficients for the study performance
discrepancies – risk taking ........................................................................................ 66
TABLE B1: Descriptive statistics and correlations for the study performance
discrepancies – risk taking ........................................................................................ 68
TABLE A2: Performance discrepancy-risk taking relationship based on historical
aspirations ................................................................................................................. 69
TABLE A3: Performance discrepancy-risk taking relationship based on social
aspirations ................................................................................................................. 71
TABLE A4: Hypothesis Two – comparing risk taking between single-business firms
and divisions ............................................................................................................. 73
TABLE B2: Relationship between performance discrepancy and risk taking based on
historical aspirations ................................................................................................. 75
TABLE B3: Relationship between performance discrepancy and risk taking based on
social aspirations ....................................................................................................... 77
TABLE B4: Hypothesis Two – comparing risk taking between single banks and
banking divisions (using historical aspirations) ........................................................ 79
TABLE B5: Hypothesis Two – comparing risk taking between single banks and
banking divisions (using social aspirations) ............................................................. 81
TABLE AP1: Correlation table for the reduced sample testing H5a-H5b ....................... 91
TABLE AP2: Relationship between performance discrepancy and risk taking
based on historical aspirations (Compustat sample) ................................................. 93
TABLE AP3: Relationship between performance discrepancy and risk taking based on
social aspirations (Compustat sample) ...................................................................... 95
TABLE AP4: Relationship between performance discrepancy and risk taking based on
historical aspirations (Banking sample) .................................................................... 97
TABLE AP5: Relationship between performance discrepancy and risk taking based on
social aspirations (Banking sample) ......................................................................... 99
TABLE S1: Summary of hypotheses and results ........................................................... 149
viii

TABLE C1: Descriptive statistics and correlation coefficients for the study risk
taking – performance (Compustat sample) ............................................................. 153
TABLE D1: Descriptive statistics and correlations for the study risk
taking – performance (Banking sample) ................................................................. 155
TABLE C2: Relationship between risk taking and performance ................................... 156
TABLE D2: Relationship between risk taking and performance ................................... 157
TABLE AP6: Correlation table for the reduced sample testing H8 ............................... 168
TABLE AP7: Relationship between risk taking and performance testing H8 ............... 170
TABLE AP8: Relationship between risk taking and performance (Banking sample) ... 173

ix

LIST OF FIGURES
FIGURE 1: Main and moderated relationships between attainment discrepancy and
divisional risk taking ............................................................................................... 151
FIGURE 2: Moderated relationships between divisional risk taking and divisional
performance ............................................................................................................ 152
FIGURE A1: Interaction between performance discrepancy and diversification dummy
comparing divisions and single firms (Compustat sample) ...................................... 83
FIGURE B1: Interaction between performance discrepancy and diversification dummy
comparing single banks and banking divisions
(historical aspirations using loan risk) ...................................................................... 84
FIGURE A2: Interaction between performance discrepancy and level of diversification
(Compustat sample) .................................................................................................. 84
FIGURE B2: Interaction between performance discrepancy and level of diversification
based on social aspirations (Banking sample and using loan risk) ........................... 85
FIGURE A3: Relationship between performance discrepancy and divisional relative size
(Compustat sample) .................................................................................................. 85
FIGURE A4: Relationship between performance discrepancy and CEO shares
(Compustat sample) .................................................................................................. 86
FIGURE A5: Relationship between performance discrepancy and CEO options
(Compustat sample) .................................................................................................. 86
FIGURE A6: Relationship between performance discrepancy and pay gap
(Compustat sample) .................................................................................................. 87
FIGURE A7: Relationship between performance discrepancy and potential slack
(Compustat sample) .................................................................................................. 87
FIGURE A8: Relationship between performance discrepancy and free cash flow
(Compustat sample) .................................................................................................. 88
FIGURE B3: Interaction between performance discrepancy and free cash flow
(Banking sample and using loan risk) ....................................................................... 88
FIGURE B4: Interaction between performance discrepancy and free cash flow
(Banking sample and using liquidity risk) ................................................................ 89

x

FIGURE B5: Interaction between performance discrepancy and institutional ownership
(Banking sample and using loan risk) ..................................................................... 101
FIGURE A9: Interaction between risk taking and # related divisions
(Compustat sample) ................................................................................................ 159
FIGURE B6: Interaction between risk taking and # of related divisions (Banking
sample using liquidity risk) ..................................................................................... 159
FIGURE A10: Interaction between risk taking and unrelated diversification
(Compustat sample) ................................................................................................ 160
FIGURE B7: Interaction between risk taking and unrelated diversification (Banking
sample using bank risk) .......................................................................................... 160
FIGURE B8: Interaction between risk taking and unrelated diversification (Banking
sample using loan risk) ........................................................................................... 161
FIGURE A11: Interaction between risk taking and munificence (Compustat sample).. 161
FIGURE B9: Interaction between risk taking and munificence (Banking sample)........ 162
FIGURE A12: Interaction between risk taking and dynamism (Compustat sample) ..... 162
FIGURE A13: Interaction between risk taking and complexity (Compustat sample) ... 163
FIGURE B10: Interaction between risk taking and complexity (Banking sample
using bank risk) ....................................................................................................... 163
FIGURE B11: Interaction between risk taking and complexity (Banking sample
using loan risk) ........................................................................................................ 164
FIGURE A14: Interaction between risk taking and CEO shares (Compustat sample) .. 164
FIGURE A15: Interaction between risk taking and CEO options (Compustat sample). 165
FIGURE A16: Interaction between risk taking and Pay gap (Compustat sample)......... 165
FIGURE A17: Interaction between risk taking and Blockholders (Compustat sample) 166
FIGURE B12: Interaction between risk taking and divisional relative size
(Banking sample) .................................................................................................... 166
FIGURE A18: Interaction between risk taking and divisional shares ............................ 172

xi

INTRODUCTION
Why do organizations take risks? How does risk taking emerge and develop over time?
Over many years scholars in the field of strategic management have tried to answer these
questions and outline the driving factors of organizational risk taking and its consequences for
firm performance. A common framework utilized to study managerial risk taking is provided by
the behavioral theory of the firm (BTOF) (Cyert and March, 1963). It argues that firms have
predetermined aspiration goals which they want to achieve and to which they compare their
current performance. When they are not able to reach these aspirations, e.g. performance falls
short, firms face an attainment discrepancy which urges them to seek alternative solutions and
engage in risky strategies. On the contrary, when firms exceed their aspirations, they are less
likely to undertake risks due to complacency and satisfaction with the status quo. Prior research
has linked performance discrepancies to various managerial decisions, such as factory expansion
(Audia and Greve, 2006), innovations (Greve, 2003a), market change (Greve, 1998), and
likelihood of acquisitions (Iyer and Miller, 2008). In addition, the BTOF has been utilized across
multiple contexts ranging from radio stations (Greve, 1998), to shipbuilding (Greve, 2003a), to
furniture and computer software (Lant, Milliken, and Batra, 1992), to airlines and trucking
(Audia, Locke, and Smith, 2000).
Despite its broad application and fairly well understood foundations, there still remain a
number of under-explained issues central to the theory. First, there has been limited theoretical
discussion of or empirical research on the impact of organizational context, and organizational
structure in particular, on managerial risky behaviors (e.g. Gavetti, 2005; Gavetti, Levinthal, and
Ocasio, 2007). Such a gap is surprising given the fact that managerial decisions and actions are
embedded in and influenced by the organizational context (e.g. Granovetter, 1985; McNamara

1

and Bromiley, 1997; Sutcliffe and McNamara, 2001). Additionally, there has also been little
theorizing on the degree to which the intensity and quality of risk taking varies across
organizations (e.g. Deephouse and Wiseman, 2000). Developing a theory that incorporates
organizational structure with organizational risk taking and that explicates on the former’s role
on the intensity and quality of risky choices will not only extend the BTOF to yet-untested
boundary conditions, but also it will offer a more detailed understanding of organizational risk
taking.
To examine the interactions between organizational structure and risk taking, I utilize the
context of multidivisional (M-form) firms. The latter are firms that operate multiple businesses in
various markets, are comprised of several divisions or subsidiaries with each of them pursuing
their own strategic objectives while also being accountable to a corporate office. Such a context
will not only allow me to distinguish the unique features of risk taking within M-firms from risk
taking in single-business firms, but also it will provide an opportunity to examine whether and
when organizational structure inhibits or enhances risk intensity and risk quality.
The structure of M-firms is associated with several important features which are likely to
influence divisional risk taking and distinguish it from risk taking in single-business firms. First,
the partial isolation of divisions from external market pressure (Bradley et al., 2011), the
corporate protection from failure which they receive (Barnett, Greve, and Parks, 1994; Rajan,
Servaes, and Zingales, 2000), and the existence of low-powered incentives at the divisional level
(Williamson, 1985) reduce the responsiveness of divisional managers to performance shortfalls.
Thus, the M-firm structure appears to lead to a sense of security among divisions (Kraatz and
Zajac, 2001), making them less attentive to existing performance discrepancies and thus less
likely to engage in risk taking compared to single-business firms. However, considering that risk

2

taking is prescribed as a remedy to poor performance (Bromiley, 1991; Cyert and March, 1963;
Lant, 1992) and a means to adjust to environmental shifts (Madsen, 2008), the relative
disinclination and inability of divisions, compared to single-business firms, to undertake the
necessary amount of risky initiatives jeopardizes their chances of successfully closing
performance gaps and avoiding organizational failure.
While these arguments suggest that in contrast to single-business firms M-firm structure
might place divisions at a disadvantage in terms of their sensitivity to performance deficiencies,
it is unlikely that all divisions would be equally unresponsive to such deficiencies. I argue that
within the M-firm there is an additional set of structural and resource endowment characteristics
at the divisional and corporate levels that will strongly influence the motivation and ability of
divisional managers to respond to poor performance. Thus, it would be premature to argue that
inefficiency of divisional risk taking dooms all multidivisional corporations to organizational
failure; instead, certain internal structural arrangements and contingencies could alleviate
divisions’ unresponsiveness to attainment discrepancies and thus partially improve their chances
for appropriate adaptation to unfavorable performance conditions.
Second, the structural embeddedness and hierarchical dependence of divisions generate
strong social influence forces which are expected to affect the outcomes of divisional risk taking.
On one hand, the availability of other divisional counterparts and the common ownership ties
among them increase their interactions and encourage utilization of prior experiences and readymade solutions to problems (e.g. Vissa, Greve, and Chen, 2010). As a result, a focal division has
access to rich and useful corporate information which facilitates successful vicarious learning
and subsequent improvements in performance (e.g. Baum and Ingram, 1998; Beckman and
Haunschild, 2002; Madsen and Desai, 2010). On the other hand, divisions, in contrast to single-

3

business firms, face rigid decision rules on where and how to invest (e.g. Merton, 1968) which
restricts their ability to identify appropriate problem solutions and might lead them to automatic
and suboptimal risky choices (e.g. Sutcliffe and McNamara, 2001). In other words, the M-firm
structure might prevent divisions from undertaking risk taking which is beneficial to
performance. Overall, the above arguments bring to the forefront the following question: Is
organizational structure, and in particular M-firm structure, a key element in explaining the
association between risk taking and performance? I argue that structure is a critical
underexplored contingent factor (e.g. Deephouse and Wiseman, 2000; Wiseman and Bromiley,
1991) that might explain why and how certain organizations are able to engage in valueenhancing risk taking which improves performance (e.g. Miller and Leiblein, 1996) while others
undertake performance-detrimental risks (e.g. Bromiley, 1991; Wiseman and Bromiley, 1996).
The answer to this question is fundamental because risk taking and organizational design
(structure) are critical for achieving organizational effectiveness and survival (e.g. Sutcliffe and
McNamara, 2001; Wiseman and Bromiley, 1996).
Furthermore, utilizing the context of M-firms will allow me to examine the role of
organizational coalitions (Cyert and March, 1963) on risk intensity and risk quality. Despite their
recognition as important factors in organizational decision making and actions (Cyert and March,
1963), organizational coalitions have been overlooked in prior research on organizational risk
taking. More specifically, there has been relative lack of theorizing on how the existence of
coalitions within an organization influences risky initiatives. Instead, prior studies have
implicitly modeled the firm as a single actor and assumed that it pursues a common objective.
However, such an assumption is unlikely to hold within a multidivisional firm where the
divisions could be viewed as separate coalitions with their own goals and interests. In order to

4

fulfill those goals, divisions are expected to pursue diverse initiatives, are likely to respond
differently to performance feedback which subsequently affects the effectiveness of their risk
taking.
Overall, the objective of this dissertation is to build an organizational structure theory of
risk intensity and risk quality and thus to offer a more detailed discussion on the processes
underlying organizational risk choice and its consequences for organizational performance. To
achieve this goal I will structure my dissertation in two chapters. The first chapter will discuss
the characteristics of divisional risk taking in an M-form firm and outline the main reasons why
it differs from the risk taking of single-business companies. The overarching logic of the chapter
is to introduce the idea of coalitions and how their embeddedness in a multidivisional firm
affects risky behaviors. Here I challenge prior studies on organizational risk taking which have
presented the firm as a unified and single actor and have ignored the existence of internal
coalitions characterized by inconsistent interests and pursuing differentiated objectives. I develop
a framework that is built on the premise that the hierarchical nesting of corporate divisions and
the complex political and social relations among them and the corporate office have a major
impeding influence on their risky choices. Furthermore, I argue that the distinctiveness of
divisional risk taking is driven by structural and cognitive contingencies that determine which
external cues and reference targets are salient to divisional managers and how they respond to
them (e.g. Ocasio, 1997). Finally, I investigate how various corporate and divisional factors,
such as diversification, corporate governance variables, corporate slack and divisional size, serve
as moderating conditions to the relationship between divisional attainment discrepancy and
subsequent risk taking.

5

The second chapter discusses how divisional risk taking is related to subsequent
performance by examining the critical role of organizational structure as an underlying
mechanism. The central argument is based on the idea that because divisions are not independent
entities, are subject to lower external pressures and weak internal motivation (Kraatz and Zajac,
2001; Barnett, et al., 1994; Williamson, 1996), and enjoy the availability of comparing to and
learning from other divisions (e.g. Greve, 2003b), the quality of their risky choices will be
affected by these contingencies and ultimately will impact their performance outcomes. In
addition, I develop a set of cross-level moderating factors that are predicted to alleviate or
exacerbate the impact of divisional risk taking on subsequent performance. Drawing on past
research on managerial decision making, I argue that environmental-level (munificence,
dynamism, and complexity) (Goll and Rasheed, 1997; Henderson, Miller, and Hambrick, 2006;
Palmer and Wiseman, 1999), corporate-level (diversification, CEO incentive pay, institutional
and blockholder ownership) (Sanders and Hambrick, 2007; Mikkelson and Ruback, 1991;
Palmer and Wiseman, 1999), and divisional-level variables (divisional size and divisional
managers’ incentive pay) (Audia and Greve, 2006) are important contingent factors that might
explain why in some contexts risk taking generates “good” organizational performance and in
other contexts risk taking results in worsened performance.
In summary, my dissertation will extend the boundary conditions of the BTOF by
elaborating on the role of coalitions in organizational risk taking, discussing the role of structural
contingencies on intensity of risky activities, and explicating on the performance implications of
risky behaviors across various structural arrangements. In developing my dissertation I will rely
on a variety of theories and perspectives of managerial behavior, such as BTOF (Cyert and
March, 1963), theories of M-form firms (Hill and Hoskisson, 1987; Hoskisson, Hitt, and Hill,

6

1991, 1993; Hoskisson and Turk, 1990), agency theory (Jensen and Meckling, 1976; Eisenhardt,
1989), theories of environmental dynamism, munificence and complexity (Dess and Beard,
1984; Duncan, 1972; Palmer and Wiseman, 1999), and organizational learning theories (Baum
and Ingram, 1998; Beckman and Haunschild, 2002; Chuang and Baum, 2003; Greve, 1998;
Ingram and Baum, 1997a, 1997b; Madsen and Desai, 2010).

Behavioral Theory of the Firm Revisited
As previously discussed, organizational structure and design have been overlooked in the
BTOF. Instead, the theory’s arguments about organizational risk taking have been developed by
focusing more broadly on the concept of organization and its tendency to engage in risky
activities. Subsequent scholarship on risky choices has continued this line by viewing the firm as
a unified actor without paying recognition to the fact that organizations might be comprised of
multiple coalitions, such as separate units or divisions. Considering that theory suggests the
existence of important links between decision-making and organizational context (McNamara
and Bromiley, 1997; Sutcliffe and McNamara, 2001), it is surprising that little attention has been
paid to how or whether particular organizational structures might exhibit different patterns of
risky activities. To fill this gap I examine how the structural characteristics of multidivisional
firms might affect risk taking and its consequences on performance. Studying M-form firms is
particularly relevant for two main reasons. On one hand, because multidivisional corporations
are wide-spread across industries and represent a large number of all organizational forms, their
relative omission from research on organizational risk taking might lead to misspecification and
underrepresentation of empirical findings. On the other hand, M-firms have unique features, such
as structural hierarchies and social embeddedness of their corporate divisions (e.g. Vissa et al.,

7

2010), with important implications for managerial behavior, particularly risky choices. First, the
M-form firm is characterized by hierarchical structure and relative lack of independence of
divisions from the corporate office – divisions rely on the corporate office for internal resource
allocations. While in general the corporate office is expected to allocate resources to divisions
with strong financial performance and positive investment forecasts (e.g. Thomas III and
Waring, 1999), very often capital allocation is biased with underperforming divisions being
subsidized and not allowed to fail (Lamont, 1997; Rajan et al., 2000; Scharfstein, 1998;
Scharfstein and Stein, 2000; Shin and Stulz, 1998). As a result, instead of achieving optimal
level of efficiency the internal capital allocation system generates flawed incentives (e.g.
disincentives) for divisional managers. They might develop a false sense of security (Kraatz and
Zajac, 2001) and experience lack of urgency in addressing performance problems they face in
their divisions. Also, the corporate structure partly isolates them from the external market
pressure and criticism of investors which further exacerbates their slow responsiveness to
performance shortfalls. Furthermore, the existence of the internal capital allocation system is
directly linked to the identification of reference targets against which performance is evaluated.
Because divisions are evaluated against one another and the allocation system serves as a
mechanism to determine the division’s relative standing in the corporation, this system could be
viewed as an important attention allocation structure (Ocasio, 1997) that redirects managerial
attention internally to salient targets – e.g. other corporate divisions. In other words, the internal
allocation system as a unique feature of the M-form firms plays a major role in where and how
divisions undertake risks. Furthermore, compared to single-business firms that utilize historical
and social comparison targets (e.g. Bromiley, 1991; Cyert and March, 1963), the internal

8

allocation system creates an additional reference target for divisions to judge their performance
against.
Second, M-form firms are composed of multiple divisions with quite distinct and
oftentimes conflicting interests and objectives. Divisions are similar to what Cyert and March
(1963) discuss as organizational coalitions. And while their definition is centered around the
inter-personal (or inter-department) level, the underlying logic is that “participants in a particular
“region” (p. 27) have common interests, pursue specific goals, lay particular claims on the whole
organization and bargain over payments. Since divisions are relatively autonomous units, might
have objectives inconsistent with the overall corporate goals, and compete over internal resource
allocations, they could easily be classified as coalitions. While they are expected to work towards
achieving the overall corporate goals, they are not immune to pursuing objectives which serve
their own interests. In particular, divisions prefer to avoid engaging in uncertain behaviors
because of the inherent risk aversion of their managers and the low-powered incentives these
managers receive (Williamson, 1985). Because divisions possess information asymmetry
advantage over the corporate office they are able to advance their interests even in cases when
those interests are not beneficial to the whole corporation. This informational advantage is based
on divisional managers’ ability to misrepresent information and thus capture more political and
bargaining influence against the corporate office. In other words, divisional managers could
easily manipulate corporate executives, disguise performance failures or at least undermine their
severity, and divert attention away from those problems. As a result, divisional managers are
able to avoid the need to solve those existing performance problems through risk taking and
maintain the status quo.

9

The structural contingencies of the M-firm also have implications for the outcomes of
divisional risk taking. While the corporation serves as a “safety net” to divisions and helps them
survive (e.g. Kraatz and Zajac, 2001), it also discourages them from careful evaluation of risky
projects and thus increases the likelihood that those projects will be suboptimal. Being partially
isolated from the market pressure and expectations of investors makes divisional managers less
diligent and sensitive to the characteristics of their risky initiatives. In other words, they are more
tolerant to suboptimal projects – even if their risk strategies are unsuccessful and result in poor
performance, the M-firm could compensate for the negative results and absorb divisional losses
(e.g. Barnett et al., 1994). In a similar vein, the inherent low powered incentives (Williamson,
1985, 1996) might encourage “bad” risk taking which hurts divisional performance. Finally, the
rigid corporate rules and established procedures on how and where to look for problem solutions
leave little room for divisional non-compliance and modification of risky choices. However,
often those taken-for-granted rules do not fit with a focal division’s context and usually result in
inappropriate risky strategies which are detrimental to performance.
At the same time, the M-firm and its internal capital market offer rich information to
divisional managers and allow them to learn from the strategic choices of other divisions. As a
result, divisional managers could utilize vicarious learning (e.g. Baum and Dahlin, 2007; Baum
and Ingram, 1998; Ingram and Baum, 1997a, 1997b; Madsen and Desai, 2010) which has a
beneficial effect on the quality of divisional risk taking. In summary, the M-firm structure might
inhibit or alleviate divisional performance by enhancing or reducing the quality and effectiveness
of divisional risk taking.

Contingent perspective on divisional risk taking

10

Divisional risk taking is an important part of the decision-making repertoire of managers
and is subject to particular behavioral biases. For example, prior research offers evidence that
various managerial characteristics, such as motivation (Devers, McNamara, Wiseman, and
Arrfelt, 2008; Sanders, 2001; Sanders and Hambrick, 2007) and strategic orientation (Gupta,
1987) play an important role on the likelihood and patterns of risky choices. Thus, it is warranted
to examine their moderating role on divisional risk taking. In addition, managerial behavior does
not happen in “a social vacuum” (Sutcliffe and McNamara, 2001: 486); on the contrary, evidence
exists that the decision making process is contextually dependent and because of its
embeddedness in the organization, it is influenced by different characteristics of the latter (Audia
and Greve, 2006; Granovetter, 1985, 1992; McNamara and Bromiley, 1997; Sutcliffe and
McNamara, 2001). In a similar vein, because divisional risk taking has major implications on the
functioning of the whole multidivisional corporation and because divisions are structurally
situated in and strategically dependent on the M-firm, it is expected that factors from higher
levels in the organization (e.g. the corporate) will play a contingent role on divisional risk taking.
Finally, considering that the M-firm is situated in an environmental context and given the
evidence that the environment might inhibit or facilitate decision making (Goll and Rasheed,
1997; Palmer and Wiseman, 1999), the implications of divisional risk taking are expected to vary
across different environmental contexts. Overall, the hierarchical embeddedness of divisions in
the M-firm and the “multilevel relational contexts” (Langley 1989; O’Reilly and Chatman, 1996)
within which they operate characterize divisional risk taking as a multilevel phenomenon which
is affected by factors situated at the divisional, corporate, and environmental levels of analysis.
Intended Contributions

11

Considering that organizational risk taking represents an important managerial action and
recognizing the influence of organizational structure on decision making processes (Hill and
Hoskisson, 1987; Hoskisson et al., 1991, 1993; McNamara and Bromiley, 1997), it is surprising
that there is a relative lack of research trying to bridge the BTOF with theories on the M-form
firm; in particular, our knowledge is incomplete on how the structural contingencies of the Mfirm affect the risky choices of corporate divisions. To fill this gap I build an organizational
structure theory of risk intensity and risk quality. By developing such a theory, my dissertation
could answer several important questions. First, is divisional risk taking different from risk
taking in a single-unit corporation and if so, how and why? Second, how do organizational
structure and hierarchical embeddedness of divisions impact their propensity to engage in risky
activities? Finally, how do the unique features of divisional risk taking influence subsequent
performance and organizational effectiveness?
There are several major contributions which this dissertation aims to add to existing
research. First, I integrate theories of M-firm structure (Hill and Hoskisson, 1987; Hoskisson,
Hitt, and Hill, 1991, 1993) and organizational search theory (Cyert and March, 1963; Greve,
2003a) to outline the critical role of structure on the intensity of divisional risk taking. In
particular, I argue that the structural arrangements of the M-firm constrain divisional intensity of
risk taking which might contribute to the inefficiency of organizational responses to performance
shortfalls. This allows me to address an implicit and often taken-for-granted assumption that
poor performance should trigger risk taking. Instead, I posit that whether organizations are able
and willing to engage in risky actions following poor performance is dependent on the
facilitating or inhibiting features of their organizational design. Second, by developing a model
that links divisional risk taking to subsequent divisional performance, I contribute to the vast

12

literature on risk-return relationships. More specifically, I integrate organizational learning
(Baum and Ingram, 1998; Beckman and Haunschild, 2002; Chuang and Baum, 2003; Greve,
1998; Madsen and Desai, 2010) with organizational risk taking to argue that successful vicarious
learning is critical for the quality of risky behaviors and is the main reason why some
organizations engage in performance-enhancing risk taking. Third, I discuss the role of coalitions
within multidivisional corporations. Recognizing the existence of coalitions within corporations,
the model challenges the assumption of corporations as unitary economic actors and instead
argues that multiple divisions could often have conflicting interests that affect the internal
efficiencies in the multidivisional corporations. Thus, I extend the theoretical underpinnings of
the BTOF with an analysis focusing on the important but often overlooked role of coalitions.
Fourth, my dissertation contributes to research on internal organizational dynamics and the
complex interrelationships across various actors. More specifically, I explicate on the inherited
conflict of interests between organizational divisions and the corporate office and its impact on
the likelihood of divisions engaging in risky activities. Due to their ability to bridge hierarchical
and divisional boundaries and thus create power asymmetries vis-à-vis the corporate office,
divisional coalitions could ultimately affect the way divisions undertake risky initiatives. Fifth, I
provide a theoretical framework which utilizes organizational structure to more clearly link the
BTOF with the attention based view of the firm. While those theories are internally consistent
and logically connected, less empirical work has been conducted to integrate them in a single
framework. While the BTOF is mainly interested in why and when risk taking is initiated, the
ABV allows answering how and where risky choices are initiated and directed. Sixth, I build one
of the first studies to apply a multilevel approach to the BTOF. By introducing factors from the
divisional, corporate, and environmental level, I present divisional risk taking as the outcome of

13

“a cascading hierarchy of influences” (Sutcliffe and McNamara, 2001: 486) which allows me to
provide a more coherent picture of the risky behaviors in a multidivisional organization. Finally,
my dissertation provides valuable knowledge to practitioners and managers regarding the link
between organizational structure and risk taking. Considering that structure might have an
overarching impact on risk intensity and risk quality and ultimately on firm survival, it is
advisable for managers to consider in advance and avoid suboptimal structural decisions, such as
excessive diversification, which have a detrimental impact on organizational performance
(Palich, Cardinal, and Miller, 2000).
Overall, my dissertation presents a comprehensive model that builds on several
distinctive literatures – organizational risk taking, agency theory, contextual impact on decision
making, and the relationship between risk taking and performance. These literatures are
integrated through the lens of organizational coalitions within a multidivisional firm. Thus, this
model explicates on the underlying mechanisms and contingency factors affecting the intensity
and direction of divisional risk taking, and elaborates on the subsequent impact of divisional risk
taking on divisional performance.
The dissertation is organized as follows. In Chapter One I build the theoretical model of
divisional risk taking. Here I discuss the antecedents to divisional risk taking. I develop
hypotheses elaborating on the factors that are expected to affect the intensity of divisional risk
taking. Further, I hypothesize about the moderating conditions that are expected to influence
risky behaviors. In addition, I present the analytical part of this chapter (data collection,
sampling, estimation methodology, and variable construction) along with empirical results and
discussion on the findings.

14

Chapter Two discusses the implications of divisional risk taking on subsequent divisional
performance. I develop theoretical propositions on the expected relationship between risk taking
and performance at the division level and advance several hypotheses testing the moderating
effects of environmental, corporate, and divisional variables on that relationship. I proceed with
empirical analysis and results, followed by discussion on the empirical findings and their
meaning in the context of my dissertation and for the general literature on organizational risk and
performance. Finally, I conclude with future directions for research and how my dissertation
could be extended into further theoretical contexts and domains.

15

CHAPTER ONE: RISK TAKING AT THE DIVISIONAL LEVEL OF M-FIRMS –
ANTECEDENTS AND MODERATING CONDITIONS

INTRODUCTION
Starting with Cyert and March (1963) and continuing to date, organizational risk taking
has been studied within the context of firms’ attempts to improve on their market position
(Greve, 1998), enhance their innovative capabilities (Greve, 2003a), expand interorganizational
ties (Baum et al., 2005), or augment corporate scope (Iyer and Miller, 2008). The common
denominator among all these initiatives is that they are characterized by uncertainty and
significant chance of failure. For example, undertaking innovations is very risky, puts the firm in
an unknown territory and often ends up with significant operational losses (e.g. Greve, 2003a).
This leads to the question of why firms continue to engage in risky initiatives rather than
maintain their current routines and practices (which they know and feel comfortable managing).
The reason for these moves is often dissatisfaction with current organizational performance.
Organizations establish aspiration levels (both historical and social (e.g. Bromiley, 1991; Cyert
and March, 1963)) against which they judge their current performance. When performance is
below the aspiration level, it is seen by managers as a threat to the viability of the organization
and triggers them to look for alternatives (Bromiley, 1991; Cyert and March, 1963; Tversky and
Kahneman, 1986). In other words, the discrepancy between actual performance and initial
aspirations serves as a “master switch” to encourage more risk taking (Greve, 2003a). Initially,
organizations look for alternatives in the vicinity of their routines and procedures; however, the
further performance deviates from aspiration levels the more willing managers become to
undertake riskier moves.

16

While the empirical examinations on organizational risk taking have produced relatively
consistent findings, there are still important under-explored issues within the BTOF. We know
relatively less about the relationships between risky choices and organizational structure.
Considering that managerial decision making processes do not happen in vacuum and are
structurally embedded in and influenced by organizationaal design (Granovetter, 1985;
McNamara and Bromiley, 1997; Sutcliffe and McNamara, 2001), it is important to examine how
organizational structure affects organizational risk taking. To address this knowledge hole, I will
investigate divisional risk taking within a multidivisional corporation and compare its
characteristics to risk taking of single-business firms.
M-form firms, compared to single-business firms, have at least two important distinctions
which are expected to influence the risky activities of their corporate divisions. First, divisions
(especially struggling ones) are often subsidized by the corporate office (Lamont, 1997; Rajan et
al., 2000) and their survival is partially detached from their operational efficiency. In other
words, the M-firm structure protects divisions from market pressures and reduces their
sensitivity to own performance misfortunes. Second, the M-firm is comprised of multiple
coalitions (e.g. divisions) which have different interests and goals which are not necessarily
consistent with overall corporate goals. In particular, due to their risk aversion, partially driven
by ineffective incentive systems or inherent risk bearing stemming from employment underdiversification (Fama, 1980; Williamson, 1985), divisional managers are less likely to pursue
risky initiatives. By utilizing their informational advantage against the corporate office,
divisional managers could avoid risk taking and maintain the existing status quo.
In summary, by combining two distinctive theoretical frameworks – BTOF (Cyert and
March, 1963) and theories of M-form firms (Hill and Hoskisson, 1987; Hoskisson, Hitt, and Hill,

17

1991, 1993) – this chapter aims to provide a more detailed and nuanced picture on divisional risk
taking and some of the facilitating and/or inhibiting factors that distinguish the patterns of risk
taking across divisions in M-form firms and single-business firms.

HYPOTHESES DEVELOPMENT
Before proceeding with the formal development of hypotheses, I will discuss an
important issue which has been overlooked in prior BTOF research and instead has been
implicitly taken for granted. More specifically, I will elaborate on the motivation and ability of
managers to undertake risky choices in response to poor performance. That is, if managers are to
engage in risky strategies, two conditions should be met: 1) managers want to make those risky
choices; and 2) they have the discretion to do it. First, within the traditional context in which
BTOF has been tested – that is, the firm is viewed as a single actor – it is obvious why managers
are motivated to undertake risk taking. Poor organizational performance translates into lower
compensation, particularly given the fact that significant portion of that compensation is linked
to corporate outcomes (e.g. Jensen and Meckling, 1976; Gomez-Mejia and Wiseman, 1997). In
addition, weakening performance might indicate the inability of managers to deal with corporate
misfortunes and force principals to oust them from the company. As a result, managers are faced
with a significant employment risk which coupled with their over-investment in the firm (Fama,
1980; Jensen and Murphy, 1990) could result in a loss of all potential future income. Thus,
managers are not indifferent to decreases in organizational performance and are willing to
engage in risky activities that might reverse those corporate misfortunes. Second, managers are
hired in the firm to make strategic decisions and undertake actions that result in positive value
for corporate shareholders. In order to fulfill those duties managers have the authority and

18

discretion to manage corporate funds and allocate them in a way they deem most appropriate.
And while an overseeing mechanism – the corporate board – is charged with monitoring and
constraining managerial strategies and the allocation of corporate funds (e.g. Eisenhardt, 1989;
Jensen and Meckling, 1976), there is evidence that the board oversight is not optimal and
managers are able to influence directors and secure their approval and support for investment
decisions (e.g. Westphal, 1998; Westphal and Zajac, 1995; Zajac and Westphal, 1996). In other
words, managers are able to obtain formal and informal discretion in resource allocations.
Within the context of divisions in an M-firm the same two conditions should be met in
order for divisional managers to be able to undertake risky choices. With regard to motivation,
the logic is quite similar to the argument developed in the previous paragraph. If divisions
generate losses and cannot positively contribute to organizational profitability, divisional
managers are faced with the possibility of divisional divestiture (Hoskisson, Johnson, and
Moesel, 1994; Johnson, 1996) and employment termination.
In parallel to managers of single-business firms, divisional managers are also able to
obtain some discretion from the corporate office over resource investment. That is, divisional
managers can secure corporate funding and subsequently manage it. First, because divisional
managers have a better understanding of their divisional needs and possess an asymmetric
informational advantage over corporate executives (Hill and Hiskisson, 1987; Hitt, Hoskisson,
and Ireland, 1990), they could influence the latter and secure access to necessary funds which
subsequently become at their discretion. More specifically, “divisional managers can expend
substantial resources in … internal politics” (Shin and Stulz, 1998: 533) and focus efforts on
“convincing” corporate executives to channel more resources towards their divisions (Meyer,
Milgrom, and Roberts, 1990). Evidence exists that managers could effectively utilize this

19

approach to obtain additional resources – for example, Dougherty and Hardy (1996) show that
managers leverage their positions in the organization to obtain funding support from corporate
executives for otherwise uncertain projects. In a similar vein, Scharfstein and Stein (2000)
elaborate on how divisional managers’ bargaining power allows them to receive “preferential
capital budgeting allocations“(p. 2537). Second, strong empirical evidence from the finance and
economics literature suggests that the corporate office makes inefficient allocations to divisions
and often engages in “socialism” (in the words of Scharfstein and Stein, 1996) where more
resources than needed are given to particular divisions. For example, Scharfstein (1998) offers
evidence for overinvestment (e.g. more allocations to) in divisions with poor investment
opportunities. In other words, those divisions receive more funding than justified based on their
future business prospects. Overall, the empirical results of those papers run counter to the
perspective of efficient internal allocations where resources should be directed to the strongest
and most promising divisions (e.g. Rajan et al., 2000; Thomas III and Waring, 1999). Thus,
inefficiencies in the internal allocation process allow even weak and underperforming divisions
to receive funding and have discretion over it.
Additional evidence that the latter divisions could obtain resources for investing in risky
initiatives (following poor performance) is provided by the escalation of commitment literature.
It argues that managerial reluctance to withdraw from losing projects and businesses (Staw,
1981, 1997) and inability to understand failures (Hayward and Shimizu, 2006) leads managers to
maintain financial commitments and resource allocations with the hope of reversing the fortunes
of those projects and businesses. For example, corporate executives which are ultimately
responsible for divisions’ success or failure might be unwilling to admit personal mistakes and in
the face of public embarrassment (e.g. McNamara, Moon, and Bromiley, 2002) would choose to

20

maintain or even increase financial commitments (Ross and Staw, 1993) to underperforming
divisions.
In summary, the presented theoretical and empirical evidence suggests that divisional
managers have the motivation and ability to obtain funding from the corporate office and
subsequently to have discretion over those resources.

Risk taking at the divisional level
Corporate divisions within a multidivisional corporation are charged with obtaining
particular financial objectives and thus have specific performance aspiration levels. When their
current performance falls below those aspirations, it signals problems for divisions and motivates
them to search for potential solutions (Bromiley, 1991; Cyert and March, 1963). That is, poor
performance induces problemistic search (described as “search that is stimulated by a problem” –
Cyert and March, 1963: 279) and encourages risk taking that is intended to return the division to
its prior performance levels. The further performance negatively deviates from aspiration targets,
the more likely divisional managers will undertake corrective actions to close this attainment
discrepancy (Lant, 1992). Prior research shows that the range of those actions is quite broad –
from R&D initiatives (Greve, 2003a) to changes in market scope (Greve, 1998), to investments
in mergers and acquisitions (Iyer and Miller, 2008). In a similar vein, capital expenditures in new
facilities and projects or expansion of existing capacities could allow divisions to generate
positive cash flows and recover some of their operating losses (e.g. Laughhunn, Payne, and
Crum, 1980). Overall, engaging in risky initiatives is perceived by divisional managers as a
viable alternative to solving performance problems and they are likely to initiate it. Thus, I
propose the following:

21

H1: Divisional negative attainment discrepancy is positively related to divisional risk taking.

Intensity of divisional risk taking
While divisions might try to respond to attainment discrepancies by engaging in risky
behaviors, it is still unclear whether the intensity of risk taking across divisions in an M-form
corporation is the same as the risk taking in single-unit corporations. In other words, while broad
evidence exists when organizations undertake risks (Bromiley, 1991; Cyert and March, 1963;
Greve, 1998; Greve, 2003a; Iyer and Miller, 2008), prior research has paid less attention to
whether all organizations are likely to respond with the same intensity to performance shortfalls.
That is, a research gap exists regarding the potential factors that might influence managerial risk
intensity. One such factor is organizational structure. Extant research argues that structure has
strong implications on bureaucracy and rigidity (Merton, 1968), managerial information
processing (Galbraith, 1977), ability to adapt to and resist change (Aldrich and Auster, 1986) and
inertial tendencies (Hannan and Freeman, 1984; Tushman and Romanelli, 1985) – all of these are
expected to affect managerial decision making and ultimately the ability and motivation of
managers to respond to poor performance. These arguments are in line with recent calls to take a
more-detailed perspective on organizational risk taking and investigate how it is affected by the
organizational context (e.g. Gavetti et al., 2007).
Given the fact that corporate divisions, in contrast to single-business firms, are
hierarchically embedded in M-firms, it is necessary to compare risk intensity across these two
structural forms. I argue that M-firm divisions and single-business firms will show a different
propensity to engage in risk taking with divisions being less sensitive to performance shortfalls.
At the heart of this distinction lie the structural arrangements of the M-firm and the behavioral

22

and cognitive biases of the divisional coalitions. Drawing on the attention based view perspective
(Ocasio, 1997), I posit that those arrangements and biases serve as important attention allocation
mechanisms which reduce the intensity of divisional risk taking. In other words, those attention
allocators lower the sensitivity of divisional managers to performance shortfalls and thus
decrease the urgency to engage in risky initiatives. Overall, I argue that M-firm structure
exacerbates rigidity tendencies, reluctance to change and unresponsiveness of divisional
managers to attainment discrepancies. Below I present four main differences across M-firms and
single-business firms which underlie the negative effect of M-firm structure on divisional risk
intensity: structural isolation, low-powered incentives, information asymmetries, and complexity.
First, the very structure of the multidivisional firm provides an environment where
divisions are relatively insulated from market pressures and do not face directly external
stakeholders (e.g. Bradley et al., 2011). It is the corporate executives and more specifically the
corporate CEO who have to explain to external constituents why performance is unsatisfactory
and how they intend to fix this problem. Thus, the relatively lower level of external pressure
faced by divisional managers diverts their attention from carefully observing market alerts;
instead, divisional managers will focus more internally on day-to-day operations and personal
agendas. Having an internal orientation reduces the likelihood that divisional managers will
1

quickly identify performance shortfalls and intervene with appropriate actions . In addition, the
corporate office is likely to engage in cross-subsidization by moving resources from better
performing divisions to underperforming ones (see Lamont, 1997; Rajan et al., 2000;
Scharfstein, 1998). Thus, divisional managers implicitly develop the expectation that the

1

Usually, division’s performance is evaluated routinely on an annual, semi-annual or quarterly
basis which makes identification of problems slower. On the contrary, market’s reaction to
performance shortfalls is immediate.
23

corporate office will help them overcome difficulties; also, they perceive the whole organization
as a safety net (Kraatz and Zajac, 2001) and generate a sense of security and lack of urgency to
engage in risk taking in response to performance shortfalls.
Second, divisional managers are less likely to respond to attainment discrepancy due to
their low-powered incentives. According to Williamson (1985), low-powered incentives exist
because divisional managers are not able to get claims on the gains from transactions. Hence,
divisional managers have lower motivation to undertake actions which benefit the corporation –
for example, while risk taking might alleviate performance deficiencies, the benefits might not
be captured (or at least not all benefits) by divisional managers. Thus, the latter have less to gain
from engaging in risk taking. In particular, divisional managers are assumed to be risk averse
(Williamson, 1985) and are less willing to engage in risky initiatives which might endanger the
existing status quo. By undertaking risky choices deviate from their division’s current routines
and practices which threatens the current balance of power (Hoskisson et al., 1991). Hence, the
existing incentives (or the lack of high-powered incentives) divert divisional managers’ attention
away from identifying and quickly resolving performance shortfalls.
Third, divisional managers know how to play “the game” and exploit the internal political
system. Due to better knowledge of their divisions which increases the information gap with
corporate executives (Hitt et al., 1990), divisional managers exploit information asymmetries and
gain an influence advantage over the corporate office. Thus, the bargaining power inherent in
their position of divisional managers allows them to easily manipulate corporate officers by
diverting attention from real problems to issues preserving the balance of power in the
organization. In addition, divisional managers could misrepresent information and “dress”
existing underperformance as a minor and temporary issue which does not deserve serious

24

attention. In other words, divisional managers use their informational and bargaining advantages
to divert attention from serious problems and thus avoid the need to have to solve them.
Fourth, the complex organizational structure of M-form corporations inhibits divisional
risky initiatives and instead promotes conformity to existing routines and the status quo
(Dougherty and Hardy, 1996). Due to bureaucracy, multiple hierarchies and often-conflicting
interests across divisions, decision making becomes slower and it takes much longer time
between decision making and actual decision implementation (Shimizu and Hitt, 2005). Thus,
even if divisional managers are urgent to respond to attainment discrepancies, the structural
complexity of M-firms slows down or even suppresses those attempts. As a result, divisional
managers are faced with “tremendous persistence forces that make both the recognition of a need
for change and its implementation difficult” (Lant, Milliken, and Batra, 1992: 586). With time
the persistence of rigid corporate rules (Merton, 1957) and strict adherence to procedures could
become too dominant, suppress divisional managers’ ability to identify and respond to
performance shortfalls and increase unresponsiveness to change (Shimizu and Hitt, 2005).
Furthermore, M-firms are characterized by a hierarchical gap between divisions and the
corporate office with respect to control over whether or how many resources should be invested
in different risky projects. In other words, divisional managers have to obtain corporate approval
and coordinate efforts with upper executives regarding risky initiatives (Greve, 2011). While
obtaining such support is possible, it is based on a lengthy process of extensive politicking and
bargaining which ultimately might discourage divisional managers from initiating necessary
change and risk taking.
In comparison to corporate divisions, single-business companies are stand-alone entities
which face enormous pressure from shareholders and analysts’ expectations to perform

25

according to predetermined goals. Failure to do so sends negative signals to the investor
community and leads to a decrease in firm market value. Contrary to divisional managers, the
executives of single-business firms face lower opportunities for political games and manipulation
tactics against investors – the latter will be less likely to accept excuses for poor performance and
will simply redirect their financial capital to other firms’ assets. Additionally, not achieving
aspirations might be interpreted as inability of executives to perform their duties accurately and
potentially threaten their future employment with the company. All these factors contribute to
single-unit companies building a greater sensitivity and attention to attainment discrepancies
which encourage risky behaviors. Formally stated, this leads to the following hypothesis:
H2: The relationship between divisional negative attainment discrepancy and divisional risk
taking will be weaker in divisions of M-firms than in single-business companies.

Factors affecting intensity of risk taking in divisions
Decision making processes are contextually embedded within the organization
(Granovetter, 1985, 1992) and are systematically impacted by a hierarchy of contingent factors
(e.g. McNamara and Bromiley, 1999; Misangyi, Elms, Greckhamer, and LePine, 2006; Sutcliffe
and McNamara, 2001). That is, because corporate divisions are structurally and hierarchically
embedded in the M-firm, they “are hardly “islands” onto themselves“ (Gavetti et al., 2007: 531)
and their decisions and subsequent actions are likely to be affected by the organizational context.
Since risk taking represents a major strategic decision, it is warranted to investigate how
divisional risk taking is facilitated or inhibited by various contingencies at the divisional and
corporate level. The selection of moderating factors at each level is driven by the idea that those

26

factors could play a significant role as allocators of divisional managers’ attention and affect
their motivation and ability to engage in risk taking.
First, I argue that structural factors will serve as important determinants of divisional
managers’ sensitivity to identify and respond to performance shortfalls (e.g. Hitt et al., 1990;
Merton, 1968; Minzberg, 1979). In particular, corporate diversification has been found to
increase bureaucracy, organizational complexity and difficulties in controlling divisions (Merton,
1968; Hill and Hoskisson, 1987). As a result, the corporate office and divisions become more
vertically separated (Hitt et al., 1990) which reduces the former’s ability to encourage risk averse
divisional managers to initiate risk taking in the face of poor performance. Divisional size is
expected to positively affect the responsiveness to attainment discrepancies. Size leads to greater
visibility and stronger monitoring of larger divisions (e.g. Vissa et al., 2010) and increases the
power of divisions to obtain resources which could be deployed in risky projects. Finally, I
investigate the effect of corporate governance variables on divisions’ responsiveness to poor
performance. Drawing on agency theoretic arguments (Jensen and Meckling, 1976) about the
beneficial role of contingent compensation and external monitoring in aligning the interests of
shareholders and managers and in reducing managerial risk aversion, I argue that managerial
contingent pay and presence of blockholders and institutional owners will encourage divisional
risk taking.
Second, resource endowment factors – corporate slack – have been found to act as
important drivers of organizational change and engagement in uncertain activities (e.g. Palmer
and Wiseman, 1999; Singh, 1986). More specifically, slack could act as a buffer for divisions
and affect their managers’ risk tolerance (Audia and Greve, 2006) by encouraging risk taking in
the face of poor performance (Shimizu, 2007).

27

In summary, by developing a contingent perspective of the factors that are expected to
influence the sensitivity of divisional managers to attainment discrepancies, I am able to present
a more detailed picture of divisional risk taking and some of its key determinants.
Structural factors: Level of corporate diversification
One of the most important characteristics of multidivisional corporations is their level of
diversification. Traditionally, it reflects the degree to which the organization is engaged in
various businesses and operates in multiple industries. I argue that because increasing levels of
diversification generate additional complexity in the M-firm coupled with bureaucracy and
conservatism (Merton, 1968), divisional responsiveness to poor performance will be impaired
(Haveman, 1993). Adding a bigger number and more diverse divisions results in larger “vertical
separation from knowledge associated with operational decision-making” (Hitt et al., 1990: 36)
which inhibits the corporate office’s ability to correctly assess divisions and their initiatives.
Simply stated, higher levels of diversification mean a higher information load for corporate
executives and given their limited cognitive abilities (March and Simon, 1958), it becomes
harder for them to oversee divisions (Hill and Hoskisson, 1987). In other words, there is a certain
threshold beyond which the corporate office “cannot be actively involved in the management of
more than a very small number of businesses without violating the bounded rationality
constraint” (Liebeskind, 2000: 63) – that is, executives are less likely to intervene with corrective
actions. As a result, the M-firm becomes more reliant on financial controls (Hoskisson and Hitt,
1988) which further increase the information gap between divisional managers and the corporate
office. Due to the informational advantage of the former, they are able to more easily engage in
manipulation of the latter. Considering that divisional managers are risk averse and prefer
maintaining the status quo (Williamson, 1986), they will exploit this information asymmetry to

28

protect their interests – e.g. avoid undertaking of risky initiatives. Evidence exists that as
diversification increases and financial controls become more prevalent in organizations,
managers are less committed to innovation (Baysinger and Hoskisson, 1989; Rappaport, 1978)
and become more risk averse (Hayes and Abernathy, 1980; Hitt et al., 1990; Hoskisson, Hitt, and
Hill, 1989). Furthermore, higher levels of diversification lead to more divisional competition for
investment funding in the internal capital market (Liebeskind, 2000) which increases financial
monitoring. However, more monitoring could have a negative effect on divisional initiation of
new risky projects (Garud and van de Ven, 1992; Jelinek and Schoonhoven, 1990). Finally,
diversification increases bureaucracy and reliance on rigid rules and procedures which further
enhance inflexibility and lower divisional managers’ ability to initiate change (Audia, Sorensen,
and Hage, 2001). In summary, diversification leads to rigid and standardized managerial
behavior (Hitt et al., 1990), reduces the ability of the corporate office to control divisional
managers’ opportunism, hinders initiation of risky initiatives and thus creates facilitating
conditions for low sensitivity to performance shortfalls by divisions. Formally stated:
H3: Level of corporate diversification moderates the relationship between divisional negative
attainment discrepancy and risk taking, such that the relationship will be weaker as
diversification increases.

Structural Factors: Divisional size
Divisional size is an important structural characteristic which influences risk taking of
managers. On one hand, size is associated with structural contingencies which are likely to affect
the motivation and ability of managers to initiate necessary actionable initiatives in the face of
adversity (Aldrich and Auster, 1986; Haleblian, McNamara, Kolev, and Dykes, 2012). On the

29

other hand, size leads to power advantages which could be leveraged by divisional managers and
thus impact their likelihood to engage in risky activities.
Small divisions are more flexible and nimble in conducting search initiatives because of
their reduced structural complexity (Mintzberg, 1979), better information processing (Galbraith,
1977) and lower inertial pressures (Hannan and Freeman, 1984; Tushman and Romanelli, 1985),
and lower resistance to change (Aldrich and Auster, 1986). In addition, small divisions appear to
be more vulnerable and less insulated from negative performance shocks which incentivizes
them to be more responsive to performance shortfalls via risky activities (Aldrich and Auster,
1986).
However, since small divisions within the M-firm are likely to play a less important role
for the functioning of the corporation, it is less likely that the corporate office will make serious
attempts to “bail out” these divisions. Simply stated, corporate executives would be reluctant to
allocate additional capital funds to divisions and instead might choose to divest those
underperforming divisions (Hoskisson et al., 1994). Thus, the lack of investment funding would
limit the discretion of small divisions to engage in risky new projects.
On the contrary, large divisions occupy a central place in the M-firm and their activities
are highly visible by corporate executives and other divisions. Thus, the low performance of
these divisions is magnified and their managers are under pressure to respond to performance
shortfalls (e.g. Vissa et al., 2010). Moreover, since large divisions become critical for the
functioning of the whole organization, their managers generate significant power against the
corporate office and could leverage it to obtain additional funds. Ultimately, those funds play a
buffering role and could encourage divisional risk taking. That is, the provision of additional

30

resources to big divisions lowers their survival point and shifts attention away from it towards
closing the performance gap (Audia and Greve, 2006).
Overall, because large divisions are closely scrutinized to close performance gaps and
have the necessary resources to do it, they are more likely to engage in risky efforts.
H4: Divisional size will moderate the relationship between divisional negative attainment
discrepancy and divisional risk taking, such that the relationship will be stronger for larger
divisions.

Structural factors: Corporate governance mechanisms
A central feature of the majority of today’s organizations is the separation of control and
management. On one hand, there are professional managers who are hired to devise strategies,
implement them and determine the direction of the organization. On the other hand, principals of
the firm are the legal owners who exercise control over managers. This situation leads to
divergence in interests and objectives between the two groups, generates agency costs and
ultimately could create detrimental effects for corporate wealth (Jensen and Meckling, 1976).
Because managers are assumed to be risk-averse and prefer to avoid risky initiatives (Eisenhardt,
1989; Jensen and Meckling, 1976), classical agency theory prescribes several governance
mechanisms – boards of directors, equity ownership and external monitoring – for constraining
managerial opportunism and self-serving behaviors (Dalton, Daily, Certo, and Roengpitya, 2003;
Jensen and Meckling, 1976; Jensen and Murphy, 1990). However, the effectiveness of those
monitoring mechanisms has been premised on the assumption that managers are able to extend
their control down the organization and manage the actions of their own agents. As a result, prior
research in the corporate governance field has focused “almost exclusively on the first control

31

relationship, that between owners and other stakeholder groups on the one hand, and corporate
management, on the other. It has largely ignored the second control relationship between
corporate management and others in the firm who execute plans and policies” (Child and
Rodrigues, 2003: 339). That is, the existence of double agency has been overlooked.
Double agency exists when there are two sets of control relationships involving agents at
two separate levels. A situation like that is present in M-firms where divisional managers are the
first-level agents and corporate executives are the second-level agents. Double agency generates
additional complexities in multidivisional firms and creates more options for managerial
opportunistic behaviors at the expense of shareholders. In particular, while corporate executives
act as principals to divisional managers, the former are agents to corporate shareholders. As a
result, there exist self-serving divisional managers (e.g. Williamson, 1985) who are monitored by
other self-serving corporate executives. Thus, if shareholder-beneficial strategies are to be
executed at the divisional level, effective governance mechanisms should be placed at both the
divisional and the corporate levels in M-firms. This argument is in line with Hoskisson and
Turk’s (1990) contention that agency costs in M-firms could be curtailed only when adequate
governance controls are implemented at the divisional and corporate level simultaneously. In
other words, while divisional governance structures might be a necessary condition for
implementing value-enhancing strategies, the existence of appropriate corporate-level
governance mechanisms would play a complimentary role.
Divisional-level governance mechanisms. Within the context of a multidivisional firm,
divisional managers (as argued above) are considered risk-averse, try to avoid undertaking risky
projects and prefer to maintain the status quo (Williamson, 1985; Hoskisson et al, 1991).
However, assuming that risky initiatives alleviate performance misfortunes (e.g. Cyert and

32

March, 1963; Miller and Leiblein, 1996), it is clear how divisional managers’ risk aversion might
be detrimental to corporate and shareholder value. In order to alleviate such suboptimal behavior
and align the interests of managers and shareholders, various corporate governance mechanisms
are put in place (Jensen and Meckling, 1976). One such mechanism is stock-based compensation
where part of managerial income is dependent on corporate performance. Because managers
possess shares and/or options in the organization, they have an increased motivation to undertake
strategies, such as initiation of risky projects in response to performance shortfalls, that are
beneficial to the organization (Dalton et al, 2003; Jensen and Murphy, 1990) and themselves. In
other words, stock options and ownership in the corporation alleviate the inherent risk-aversion
of divisional managers (e.g. Sanders, 2001; Sanders and Hambrick, 2007), increase their
responsiveness to poor performance and are expected to play a positive moderating role on the
relationship between attainment discrepancy and subsequent risk taking.
H5a: Divisional managers’ ownership will moderate the relationship between divisional
negative attainment discrepancy and divisional risk taking, such that the relationship will be
stronger with higher levels of ownership.
H5b: Divisional managers’ stock options will moderate the relationship between divisional
negative attainment discrepancy and divisional risk taking, such that the relationship will be
stronger with higher levels of stock options.

Corporate-level governance mechanisms. The multidivisional firm offers an additional
layer of monitoring over divisional managers’ behaviors. It is related to the firm’s CEO incentive
alignment and the presence of institutional and blockholder ownership. These corporate
governance mechanisms represent what I call a second-level monitoring. Take for example the

33

role of CEO incentive alignment (e.g. contingent forms of pay, such as stocks and stock options).
Being awarded ownership in the firm and having his income partially dependent on firm’s
performance, the CEO is likely to be a better protector of shareholder value (Dalton et al, 2003;
Scharfstein, 1998), and in this particular case the CEO will not be indifferent to divisions that are
irresponsive to poor performance. Considering that the CEO is hired to manage the whole
multidivisional firm, to carefully monitor the actions and decisions of divisional managers and to
intervene when those actions are detrimental to shareholder value, contingent compensation will
incentivize the CEO to conduct those duties more diligently. On the contrary, lack of corporate
ownership or stock options might exacerbate the self-serving nature of the CEO, including
maintaining existing political balance (Hoskisson et al., 1991; Lant et al., 1992) and friendship
relationships with divisional managers (e.g. Scharfstein, 1998) at the expense of the interests of
shareholders. In other words, the CEO is likely to protect divisional managers even in situations
when they are less responsive to performance shortfalls and side with them even if this is
detrimental to shareholders.
H6: CEO contingent pay will moderate the relationship between divisional negative attainment
discrepancy and divisional risk taking, such that the relationship will be stronger with higher
levels of contingent pay.

While the above arguments discussed incentive alignment at the divisional and corporate
levels primarily in isolation, it is also necessary to examine their interdependent nature. More
specifically, in addition to divisional managers’ decisions being affected by the level of their
contingent pay, the latter could also be influenced by the relative difference in pay between
divisional managers and corporate CEOs – or what is often referred to as the pay gap across

34

managerial hierarchical positions. Tournament theory arguments suggest that a wider pay gap
serves as a motivating mechanism for lower-level managers to exert extra effort and diligence in
their actions in the hope of advancing up in the career ladder (Lazear and Rosen, 1981; Rosen,
1986). Translating these arguments to the divisional context, I argue that as the gap in pay
between divisional managers and the corporate CEO increases, divisional managers will have
stronger incentives to undertake strategic initiatives which are consistent with the interests of
shareholders (Henderson and Fredrickson, 2001). In particular, when facing performance
shortfalls divisional managers are more likely to engage in risk taking and prevent further
shareholder value losses under conditions of high pay gaps. In other words, the internal
compensation arrangements in M-firms could serve as a positive moderating mechanism that
incentivizes divisional managers to be more responsive to poor performance. Formally stated:
H7: The gap between CEO pay and divisional managers’ pay will moderate the relationship
between divisional negative attainment discrepancy and divisional risk taking, such that the
relationship will be stronger as the pay gap widens.

In addition to managerial incentive alignment, institutional investors and large
blockholders could also serve as effective monitoring devices by increasing the alignment
between managers and shareholders (Brickley, Lease, and Smith, 1988; Hansen and Hill, 1991).
The main mechanism that underlies the positive role of external investors is the consistent
pressure they exercise over corporate executives to undertake only value-enhancing strategies.
Applying this logic to the multidivisional-firm context suggests that the presence of institutional
investors and blockholders will increase divisions’ risk taking under poor performance. This
effect could materialize through two paths. First, external investors could exercise direct pressure

35

on corporate CEOs to conduct better monitoring of corporate divisions. Since CEOs are the main
people to blame for ineffective corporate decisions, they are more likely to be sensitive to
inappropriate divisional strategies (such as risk aversion in the face of poor performance) under
conditions of strong institutional ownership and blockholders. Research by Brickley and
colleagues (1988) and Mikkelson and Ruback (1991) suggests that blockholders could constrain
managerial decisions which are considered detrimental to principals. Second, blockholders and
institutional investors could also directly affect risk taking by divisions. For example, the
presence of strong external monitoring is a very salient cue to divisional managers that their
(lack of) actions and initiatives would not go undetected. Thus, the existence of corporate
governance mechanisms is expected to play a strong attention allocation role for divisional
managers (Ocasio, 1997) and more specifically to shift their attention towards higher sensitivity
and responsiveness to poor performance.
H8a: Institutional ownership will moderate the relationship between divisional negative
attainment discrepancy and divisional risk taking, such that the relationship will be stronger
with higher levels of institutional ownership.
H8b: Blockholders will moderate the relationship between divisional negative attainment
discrepancy and divisional risk taking, such that the relationship will be stronger with larger
number of blockholders.

Resource endowment factors: corporate slack resources
Due to the hierarchical nature of M-form corporations, divisional decision making is
strongly affected by the corporate office. Alternatively, divisional decisions are not fully
independent and might be influenced by higher-order factors in the corporation (e.g. McNamara

36

and Bromiley, 1997; Suttclife and McNamara, 2001). One such factor is the availability of
resources at the corporate level. In other words, the possession of or lack of corporate slack
resources is highly likely to affect the risky behaviors at the divisional level. While each
division’s risk taking is determined mainly by its own performance aspirations, the moderating
role of corporate slack should not be overlooked. I argue that when the corporate office possesses
no slack, this “resource scarcity may lead to less risk taking” (Shimizu, 2007: 1501) by corporate
divisions. In this situation, divisions incur extremely unfavorable conditions – poor performance
and inability to receive financial resources from the corporate office. Hence, they are likely to
face threat rigidity (Staw, Sandelands, and Sutton, 1981). That is, divisional managers are
expected to become conservative, rigid and unwilling to engage in new strategic initiatives
(D’Aveni, 1989; Iyer and Miller, 2007; Staw et al, 1981). Moreover, “with limited slack, unit
performance losses may become more critical” (Shimizu, 2007: 1501) and further discourage
divisional managers to initiate any change to the status quo which ultimately results in
conservative behavior (Palmer and Wiseman, 1999). On the contrary, when the level of corporate
slack is significant, it could serve as a safety net to divisions and alleviate their inherent risk
aversion. Simply stated, the margin for error of divisions increases – even if their future risky
initiatives turn unsuccessful, those additional losses could be covered with the available
corporate slack resources. For example, Shimizu (2007) finds support for the buffering role of
slack on organizational risk taking. More specifically, his study shows that when slack resources
are scarce, managers respond to poor performance in a conservative manner and reduce risky
initiatives (e.g. by divesting poorly performing units). Overall, the above arguments suggest that
poor divisional performance coupled with corporate slack scarcity could lead to freeze in
strategic initiatives (threat rigidity). On the contrary, the availability of corporate resources

37

encourages divisional managers to respond to poor performance through increased engagement
in risky actions.
H9: Corporate slack will moderate the relationship between divisional negative attainment
discrepancy and divisional risk taking, such that the relationship will be stronger with higher
levels of slack.

METHODS
For Chapter One I utilize two separate samples which allow me to test the proposed
relationships in a more robust manner and reach conclusions with higher external validity. More
specifically, Sample 1 includes a wider range of firms and focuses on a broader construct of risk
taking – e.g. divisional capital expenditures. Also, I am able to obtain all variables necessary for
testing the proposed hypotheses. Sample 2 focuses on the banking industry and offers a more
fine-grained examination of banking divisions’ risk taking. In particular, I analyze banking
managers’ decisions to alter the structure of their risk portfolios by changing the relative weights
of various risk-laden categories comprising the portfolio. Due to data limitations, this sample
allows me to test only some of the proposed relationships (see details below).
Sample 1
Data was drawn from several sources: corporate and segment financial data comes from
Compustat, executive compensation and governance measures are accessed through Execucomp,
and divisional managers’ compensation data is obtained through manual search of corporate
annual reports. The sample includes all US-based firms from 1998 to 2009. The sample is
subject to the following screening criteria. First, units with less than $10 million in sales or assets
are excluded (McGahan and Porter, 2003; McNamara, Vaaler, and Aime, 2005). Second, I drop

38

units which are classified as “corporate” or “other” because they are not active units. Third, if the
unit’s four-digit SIC category is defined as “not elsewhere classified” or “nonclassifiable
establishment”, this unit is not included in the sample.
Hypotheses 5a-b are tested on a subsample of Compustat firms. During the period of the
study (1998-2009) I identified all diversified S&P corporations and matched them to the initial
sample. For all corporations that had no missing values on all variables in the model I executed a
manual search of their DEF 14A statements which provide information on the compensation of
the top five highest paid executives. After identifying the compensation of a divisional executive,
I matched it back to the initial sample. One caveat should be recognized – because not all
divisional executives are among the top five highest paid executives during a year, there is a
large underrepresentation in this subsample and it is likely that it is biased and not generalizable
to the entire population of multidivisional firms (the sample size here is about 6 % of the initial
sample). Thus, the results – presented in Appendix 1 – should be interpreted with caution.
Dependent variable
Capital expenditures. My measure of divisional risk taking is represented by the amount
of capital expenditures undertaken by divisions. Capital expenditures are an appropriate measure
of managerial risk taking because: a) they represent difficult-to-reverse, long-term investments
(Audia and Greve, 2006); and b) “their consequences are uncertain and depend partly on difficult
to predict environmental factors” (Desai, 2008: 598). Thus, capital expenditures represent an exante decision of allocating financial resources to projects with uncertain outcomes. Further
support for the use of this measure is found in several prior studies that have utilized capital
expenditures as reflecting managerial risk taking (Devers et al, 2008; Lant and Hurley, 1999;

39

Larker, 1983; Sanders and Hambrick, 2007). Finally, research by Miller and Bromiley (1990)
identifies capital expenditures as one of the key dimensions of strategic risk.
Independent variables
Attainment discrepancy represents the difference between divisional performance
(measured in period t-1) and aspiration levels (measured in period t-2). Divisional performance is
measured as return on assets (ROA) (Iyer and Miller, 2008). I draw on prior research to construct
my aspiration level variables (Baum et al., 2005; Greve, 2003; Iyer and Miller, 2008). Social
aspirations is the mean ROA of all other single-business firms and all other diversified firms’
divisions in the same 4-digit SIC code as the focal division. I exclude ROA of the focal
2

division. Thus,
Social aspirationsit = (∑jPjt)/N
where j is another division (or single-business firm), P is performance, and N is the number of
other divisions and single business firms, j. Historical aspirations are based on the focal
division’s ROA one year prior to past performance.
To account for differences in the slope of attainment discrepancy depending on whether
divisions are above or below the aspiration point (Greve, 2003; Iyer and Miller, 2008), I use a
spline function (Greene, 2003) and create the following variables:
Performance above aspirationsit = ROAit – aspirationsit if ROAit > aspirationsit
= 0 if ROAit ≤ aspirationsit
Performance below aspirationsit = aspirationsit – ROAit if ROAit < aspirationsit
= 0 if ROAit ≥ aspirationsit.

2

I include single-business firms since they represent a salient comparison for a focal division.
40

Corporate slack. It is operationalized as potential slack (inverse ratio of debt to equity)
and free cash flow (Bourgeois, 1981; Haleblian et al., 2012; Iyer and Miller, 2008).
Level of diversification. I argued that the primary impact of diversification is to increase
organizational complexity and inhibit the monitoring ability of corporate executives over
divisional actions. Thus, more complexity results from high diversification which could stem
either from the existence of diverse divisions or from large number of divisions. In other words,
diversification could also be represented by: 1) the number of divisions within the M-firm
(Robins and Wiersema, (2003); and 2) the entropy index (Jacquemin and Berry, 1979). More
specifically, the entropy measure consists of related diversification (DR) and unrelated
diversification (DU) component which sum to total diversification (DT):
Entropy = (∑Piln(1/Pi)
where Pi is the firm’s share of sales in segment i and ln (1/Pi) is the weight of each segment.
To test hypothesis 2 regarding differences among divisions’ and single-business firms’
sensitivity to performance shortfalls, I create a dummy variable which equals 1 if two or more
divisions belong to a corporation and 0 otherwise.
Institutional ownership is calculated as the total number of shares owned by institutional
investors, such as pension funds, mutual funds, and hedge funds, divided by the total number of
outstanding firm shares (Brickley et., 1988).
Blockholders is the number of shareholders with at least 5% stake in the company
(Hoskisson et al., 1994; Sanders and Boivie, 2004).
CEO contingent pay is measured in two ways – CEO option pay and CEO stock
ownership. Instead of combining these two forms of contingent pay together, as classical agency
theory would suggest (Jensen and Meckling, 1976; Jensen and Murphy, 1990), I follow more

41

recent research which argues that stocks and options should be treated separately (Sanders,
2001). CEO stock ownership is measured as the value of shares owned by the CEO (Sanders,
2001; Wright et al., 2002). CEO option pay is measured as the value of options granted during a
year using Black-Scholes method.
Pay gap is calculated as the difference in total compensation between the corporate CEO
and the next highest paid executive (Henderson and Fredrickson, 2001).
Divisional relative size is calculated as the ratio of divisional assets to total corporate
assets.
Divisional managers’ ownership is measured as the value of corporate shares listed in the
annual corporate report.
Divisional managers’ stock options reflect the annual award of options.
Control variables
Drawing on past research I include several control variables. Divisional size is associated
with additional bureaucracy and inertia (Hannan and Freeman, 1984; Tushman and Romanelli,
1985) which might inhibit the ability and diligence of corporate executives to encourage risky
initiatives. Divisional size is calculated as the natural log of divisional assets. Divisional sales
growth is measured as the annual change in sales and is included in the model following prior
evidence that it might impact the engagement of firms in major strategic actions (e.g. Haleblian
et al., 2012). Prior risk taking is measured as capital expenditures in period t-2. All other
independent and control variables are measured at period t-1.
Sample 2
In this sample I utilize data on the banking industry and more specifically, I collect
variables for parent and divisional banking establishments. The data is obtained from Federal

42

Reserve Bank Regulatory (Call Reports) databases. In these databases parent banks are
represented as bank holding companies that are comprised of several commercial banks which
operate like divisional units. For example, Citicorp Holdings is the corporate parent of several
nationally represented commercial banks, including Citibank NV NA, Citibank NY NA,
Citibank South Dakota NA, Citibank Delaware.
The cut-off criterion for inclusion of banks in the sample was $80 million in line with
recommendations from prior research (McNamara, Luce, and Thompson, 2002).
Hypotheses 5a and 5b (moderating role of divisional managers’ incentive pay) will not be
tested in the Banking sample due to data unavailability. Hypotheses 6, 7, and 8a-b will be tested
on a subsample of banks. After estimating the models with the initial available data from the Call
Reports, I identified all bank holding corporations with no missing data on all variables. Then, I
manually searched the ticker symbols of those corporations and used those tickers to obtain data
for CEO compensation and external governance factors from Compustat. After adding those
supplementary variables to the rest of the sample, I reestimated my models. It should be noted
that this subsample is much smaller (around 10% of the initial sample), it is likely to be biased
and not representative of the entire banking population. Thus, the results from this subsample,
presented in Appendix 1, should be interpreted with caution.

Dependent variable
Bank divisional risk taking. Following recommendation by Miller and Bromiley (1990)
and Wiseman and Catanach (1997) regarding the multidimensional nature of risk taking, I use
three types of risk taking – bank, loan, and liquidity risk. Bank risk is an indicator of capital
adequacy and solvency (Marcus, 1983) and reflects the risk position of the banking institution. It

43

is measured as the ratio of risk weighted assets to total assets (Shrieves and Dahl, 1992). Loan
risk represents an operational risk stemming from accepting loans with high probability of
borrower default. It is calculated as sum of non-accrual loans and one half of loans past due 90
days to total loans (Meeker and Gray, 1987; Shrieves and Dahl, 1992) Liquidity risk indicates
the ability of a bank to stay solvent and pay its obligations (c.f. Hambrick and D’Aveni, 1988;
Wiseman and Catanach, 1997). It is measured as the inverse ratio of liquid assets to total assets
(Hambrick and D’Aveni, 1988).
Independent variables
Attainment discrepancy is calculated following the same procedure as in Sample 1.
Performance is based on ROA (net income over total assets). ROA is the most commonly used
measure of profitability in banking research (Deephouse, 1996, 1999; Haveman, 1993; Reger,
Duhaime, and Stimpert, 1992) and is considered as an excellent indicator of earnings efficiency
(McNamara et al., 2002). The only difference from Sample 1 is that social aspirations are based
on comparison with single-banks and banking divisions in the same Metropolitan Statistical Area
(MSA). The choice of MSA is driven by the fact that MSAs consist of a “core urban area, as well
as any adjacent counties that have a high degree of social and economic integration (as measured
by commuting to work) with the urban core” (U.S. Census Bureau, 2012) and thus banks that
operate in the same MSA are likely to compare with each other in that MSA.
Bank divisional slack is operationalized as potential slack and free cash flow. Potential
slack is the inverse ratio of debt to equity (Bourgeois, 1981). Free cash flow is equal to

44

(operating profit – taxes – interest expense – depreciation – preferred dividends – common
3

dividends)/equity (Haleblian et al., 2012) .
Level of diversification is measured as the number of banking divisions within a bank
holding corporation (Robins and Wiersema, 2003).
To test hypothesis 1 regarding differences among banking divisions’ and single banks’
sensitivity to performance shortfalls, I create a dummy variable which equals 1 if two or more
banking divisions belong to a bank holding corporation and 0 otherwise.
Institutional ownership is calculated as the total number of shares owned by institutional
investors, such as pension funds, mutual funds, and hedge funds, divided by the total number of
outstanding firm shares (Brickley et., 1988).
Blockholders is the number of shareholders with at least 5% stake in the company
(Hoskisson et al., 1994; Sanders and Boivie, 2004).
CEO contingent pay is measured in two ways – CEO option pay and CEO stock
ownership. Instead of combining these two forms of contingent pay together, as classical agency
theory would suggest (Jensen and Meckling, 1976; Jensen and Murphy, 1990), I follow more
recent research which argues that stocks and options should be treated separately (Sanders,
2001). CEO stock ownership is measured as the value of shares owned by the CEO (Sanders,
2001; Wright et al., 2002). CEO option pay is measured as the value of options granted during a
4

year using Black-Scholes method .

3

In the Banking sample I measure slack at the divisional level rather than at the corporate level.
First, the availability of data allows me to do this. Second, calculating slack for the bank holding
corporation reduces largely my sample due to underrepresentation of bank holding corporations’
identification variables (RSSD 9348). Third, I do not see any theoretical rationale for a different
impact of divisional versus corporate slack on risk taking.
4
Due to limitations in data availability, I test the moderating impact of blockholders, institutional
ownership and CEO contingent pay on a smaller sample of banks.
45

Pay gap is calculated as the difference in total compensation between the corporate CEO
and the next highest paid executive (Henderson and Fredrickson, 2001).
Bank divisional relative size is calculated as the ratio of divisional assets to bank holding
company assets.
Control variables
I use several control variables. Bank divisional size accounts for bureaucracy and inertia
(Hannan and Freeman, 1984) and is operationalized as log of assets by the banking division.
Bank divisional age has been argued to create additional inertial tendencies (Amburgey and
Miner, 1992; Baum, 1990). It is measured as the time in years since incorporation. Bank
divisional growth is represented as the annual change in deposits. I also control for prior risk
taking in period t-2. All other independent and control variables are measured at time t-1.
Analysis
The data for my dissertation has a panel structure where repeated observations over time
(e.g. divisional risk taking) are nested within divisions, which are nested within corporations.
That is, I have variables (e.g. divisional size, divisional managers’ contingent pay, and sales
growth) which exist at the divisional level and variables (level of diversification, corporate slack,
institutional ownership, blockholder ownership, CEO contingent pay, and corporate size) that
should be modeled at the corporate level. The nesting of lower levels within higher levels leads
to lack of independence between observations and might result in incorrect estimates of the
regression coefficients (Raudenbush and Bryk, 2002) if standard OLS regression is used. Thus, I
use multilevel linear modeling (MLM) to account for the interdependence between lower levels.
MLM allows for the appropriate modeling across time of the relationship between the variables
in my model while simultaneously considering the nesting structure of the data (Bryk and

46

Raudenbush, 1989). Overall, by using MLM I could obtain “unbiased and efficient estimates of
the regression coefficients and their standard errors despite the dependence among observations”
(Fong, Misangyi, and Tosi, 2010: 637; Bryk and Raudenbush, 1989). An additional benefit of
MLM is that by partitioning the variance across levels it “effectively controls for industry
effects” (Fong et al., 2010: 638; Bloom and Milkovich, 1998).
I rely on the statistical software STATA XT MIXED to conduct the panel data estimation
where level-1 equation represents the within-division variance in risk taking and level-2 equation
models the cross-sectional (e.g. between divisions) variance in divisional risk taking.
Below I present the expected model specification with the appropriate equations at each level of
analysis.
(1)

Risk takingtij = π0ij + π1ij (Attainment discrepancy)t-1ij + π2ij (Relative

divisional size)t-1ij + π3ij (Divisional contingent pay)t-1ij + π4ij (Attainment discrepancy*
Relative divisional size)t-1ij + π5ij (Attainment discrepancy* Divisional contingent pay)t-1ij +
π6ij (Sales growth)t-1ij + π7ij Risk takingt-2ij + π8ij (Divisional size)t-1ij + etij
(2)

π0ij = β00j + β01j (Level of diversification)ij + β02j (Corporate slack)ij +
β03j (Institutional ownership)ij + β04j (Blockholders)ij + β05j (CEO contingent
pay)ij + u0ij

(3)

π1ij = β10j + β11j (Level of diversification)ij + β12j (Corporate slack)ij +
β13j (Institutional ownership)ij + β14j (Blockholders)ij + β15j (CEO contingent
pay)ij + u1ij

47

(4a) π2ij = β20j; (4b) π3ij = β30j; (4c) π4ij = β40j; (4d) π5ij = β50j; (4e) π6ij = β60j; (4f) π7ij =
β70j; (4f) π8ij = β80j;
(5a) β00j = γ000; (5b) β01j = γ010; (5c) β02j = γ020; (5d) β03j = γ030; (5e) β04j = γ040;
(5f) β05j = γ050;
(6a) β10j = γ100; (6b) β20j = γ200; (6c) β30j = γ300; (6d) β40j = γ400; (6e) β50j = γ500;
(6f) β60j = γ600; (6h) β70j = γ700; (6i) β80j = γ800;
(7a) β11j = γ110; (7b) β12j = γ120; (7c) β13j = γ130; (7d) β14j = γ140; (7e) β15j = γ150;
Equation (1) represents the time level where risk taking in year t by division i in
corporation j is regressed on attainment discrepancy (year t-1), relative divisional size (year t-1),
divisional contingent pay (year t-1), sales growth (year t-1), the interaction term of attainment
discrepancy and relative divisional size, the interaction term of attainment discrepancy and
divisional contingent pay, and controls. Thus, π0ij is the mean risk taking over time for division i
in corporation j, accounting for the impact of the independent variables.
In Equation (2) I use the intercept of Equation (1), π0ij, as the dependent variable and
regress it on level of diversification, corporate slack, institutional and blockholder ownership,
and CEO contingent pay. Thus, I am able to account for the direct effect of these explanatory
variables on between-division variance in risk taking.
Equation (3) models the slope coefficient, π1ij, as a dependent variable regressed on the
variables (level of diversification, corporate slack, institutional and blockholder ownership, CEO

48

contingent pay) predicted to moderate the relationship between divisional attainment discrepancy
and divisional risk taking. Thus, Equation (3) allows for testing of cross-level interactions.
Finally, Equations 4a-4f, 5b-5h, 6a-6i, and 7a-7e represent the coefficients from
equations (1), (2), and (3) as dependent variables at a higher level of analysis.
All coefficients are modeled as fixed effects.
Equations (1), (2), and (3) have separate error terms – etij is the across-time residuals, and
uij is the between-unit residual.

RESULTS
Tables A1 and B1 provide descriptive statistics and correlation coefficients coefficients
5

for the Compustat and Banking samples, respectively . I calculated VIFs and all of them are
below 3 which is an indication that multicollinearity is not a concern.
Tables A2-4, Tables B2-5 present the MLM regression analyses. To control for the effect of
6

unobserved factors I have included year dummies in the analyses . To conserve space, those
variables are not included in the tables. All variables are standardized with a mean of 0 and
standard deviation of 1. Of the control variables, divisional size, CEO shares, and pay gap
exhibit positive associations with divisional risk taking, while divisional age are negatively
correlated with the dependent variable.

5

Variables have been winsorized at the 99% level to avoid the impact of extreme outliers.
In a supplementary analysis I created a dummy variable to distinguish between bank holding
corporations consisting of only bank-type divisions and bank holding corporations consisting of
multi-purpose divisions (e.g. credit card, services, etc.). The results remained consistent across
these two estimations.
6

49

According to Hypothesis 1, performance discrepancy should encourage divisional risk taking. In
the Compustat sample, the coefficients for social and historical negative attainment discrepancy
are in the expected directions but they are statistically insignificant and do not support the
hypothesis. In the Banking sample (Tables B2 and B3), I find varying results regarding H1– poor
performance exhibits a significant negative relationship with bank risk (p<0.001 – historical
aspirations and p<0.01 – social aspirations) and with liquidity risk (p<0.1 – historical aspirations
and p<0.05 – social aspirations); however, performance discrepancy and loan risk are positively
and statistically significantly related (p<0.001), indicating that as divisional performance drops
below aspirations, divisions take more loans risks. Overall, across the two samples, H1 is not
supported. Hypothesis 2 argued that divisions compared to single business firms will exhibit
lower responsiveness to poor performance. As seen from Models 6 and 8 in Table A4, the
interaction term between diversification dummy and negative attainment discrepancy is positive
(p<0.01 – historical aspirations and p<0.05 – social aspirations). While, the slope of the main
relationship is almost flat for divisions, the slope for single firms is steeper but in an opposite
direction than hypothesized. In the Banking sample, when loan risk and historical aspirations are
utilized, the results support H2 (p<0.001). All remaining interaction terms between performance
discrepancy and diversification dummy in Tables B4 and B5 are insignificant. Thus, across the
two samples, results do not support H2 (see Figures A1 and B1).
Hypothesis 3 proposed that level of diversification will negatively moderate the relationship
between performance discrepancy and risk taking. Figure A2 (based on historical aspirations in
the Compustat sample) provides contrary results to this argument, such that as level of
diversification increases, divisions engage in more risk taking (p<0.001). When social aspirations
are used, I find insignificant results for the interaction term between performance discrepancy

50

and level of diversification. In the Banking sample, results are statistically insignificant with the
exception of the model for loan risk (based on social aspirations); the latter shows that as
diversification increases the relationship between performance discrepancy and risk taking
weakens. Overall, the results across the two samples provide no support for H3 (see Figures A2
and B2).
In Hypothesis 4 I theorized that larger divisions will exhibit a stronger relationship between poor
performance and risk taking. Using the Compustat sample I find opposite results – when faced
with poor performance smaller divisions take more risks than larger divisions (p<0.001 for both
historical and social aspirations). I do not find statistically significant results from the Banking
sample. Thus, H4 is not supported (see Figure A3).
Hypotheses 5a and 5b argued that divisional managers’ shares and options would positively
moderate the relationship between performance discrepancy and risk taking. Results from Tables
AP2 and AP3 are insignificant and thus the hypotheses are not supported. It should be noted that
those hypotheses were tested on a smaller subsample of S&P 500 companies and thus this
sample is likely to be biased and not representative of the general population of firms.
According to Hypothesis 6, CEO contingent pay moderates the relationship between poor
performance and risk taking, such that the relationship is strengthened with higher levels of
contingent pay. Utilizing historical aspirations in the Compustat sample in Table A2, I find
opposite results for CEO shares (p<0.01) and CEO options (p<0.01). Relatively similar results
are found when social aspirations are used in Table A3 – CEO shares play a negative moderating
role (p<0.01) but the interaction term between CEO options and performance discrepancy is
insignificant.

51

In the Banking sample, results for CEO shares based on historical aspirations are similar – in the
models with bank risk (p<0.01) and liquidity risk (p<0.05) findings are opposite to hypothesized.
The results for CEO shares based on social aspirations are insignificant. Regarding the
moderating role of CEO options, when historical aspirations are used and bank risk (p<0.05) and
loan risk (p<0.01) are the dependent variables, there is some support for the hypothesis. Results
for CEO options based on social aspirations are insignificant. Overall, H6 is not supported (see
Appendices 4 and 5, and Figures A4-A5).
Hypothesis 7 developed the argument that pay gap will positively moderate the relationship
between performance discrepancy and risk taking. Results (in the Compustat sample) based on
historical (p<0.001) and social aspirations (p<0.05) are contrary to predictions – as pay gap
increase, the slope of the main relationship turns negative. In the Banking sample, only the
interaction between historical discrepancy and pay gap in the model for loan risk is significant
(p<0.001) and in the opposite direction than hypothesized (see Figure A6).
Hypotheses 8a and 8b introduced institutional ownership and blockholders as moderators which
are expected to strengthen the relationship between poor performance and risk taking. For the
Compustat sample, in both the historical aspiration and social aspiration models, results are
insignificant. In the banking sample, only the interaction between social discrepancy and
institutional ownership in the model with loan risk is significant (p<0.05) and contrary to
predictions. Thus, support is not found for H8a and H8b (see Figure B5).
Finally, in Hypothesis 9 I argued that slack will positively moderate the relationship between
poor performance and risk taking. In the Compustat sample, I find that free cash flow exhibits
positive moderating role in line with the hypothesis (p<0.001 – historical aspirations and ns –
social aspirations). However, potential slack has a negative moderating role (p<0.001 – historical

52

aspirations and ns – social aspirations). In the Banking sample, models for loan risk and liquidity
risk show statistically significant interactions both for historical and social aspirations. The
models for liquidity risk (p<0.05 – historical aspirations and p<0.05 – social aspirations) and
bank risk (p<0.05 – historical aspirations and ns – social aspirations) offer support for H9.
However, when loan risk is used as DV the results are opposite to H9 (p<0.001 – historical
aspirations and p<0.001 – social aspirations). Overall across the two samples, there is only weak
support for Hypothesis 9 (see Figures A7, A8, B3, and B4 about here).

DISCUSSION
In this chapter, I theorized how the embeddedness of risk taking in the broader
organizational context impacts the responsiveness of organizations to performance shortfalls.
More specifically, I developed a multilevel model about the role of organizational context,
including organizational structure and several multi-level organizational factors, on the intensity
of divisional risk taking. Summarized results are presented in Table S.
The starting argument in this chapter is that performance discrepancies encourage risk taking and
motivate divisional managers to seek solutions to poor performance. However, the empirical
findings do not lend support to this claim and are in contrast to the theoretical foundations of the
BTOF. In particular, with the exception of loan risk, poor performance either discouraged risk
taking or showed insignificant associations with risk taking. While contrary to the hypothesis,
these results are in line with an extensive research on threat rigidity (Staw et al., 1981). This is
particularly relevant considering the way risk taking is operationalized in this study. Bank risk,
and capital expenditure in particular, represent major organizational commitments of resources
with highly uncertain outcomes and probability of failure if market conditions change (Desai,
53

2008). In additional, capital expenditures are viewed as largely irreversible investments which
are fixed and difficult to change once implemented. In line with the nature of the risk taking
measures in the dissertation, it is not surprising to find that managers in underperforming
divisions would be hesitant to commit large resources hoping of turning their division’s fortunes.
The threat-rigidity hypothesis argues that when faced with poor performance, managers
experience anxiety and low information processing capabilities (Sitkin and Pablo, 1992; Staw et
al., 1981) which reduce their ability to identify and undertake risky alternatives (Audia and
Greve, 2006; Iyer and Miller, 2008). In such situations managers are faced with the choice of
looking for risky solutions and making large investments of resources, which are already hard to
find in an underperforming organization, or taking “a position of safety by avoiding additional
losses” (Audia and Greve, 2006: 85). My findings are strongly consistent with these arguments
and indicate that divisional managers are unwilling to commit their divisions to major strategic
investments with uncertain outcomes. Another explanation for the difference between my results
and a stream of prior research (Baum et al., 2005; Greve, 1998, 2003; Miller and Chen, 2004) is
that those studies operationalize risk taking in a different way – e.g. R&D investment and
nonlocal partnership ties. The measures utilized by those authors do not necessarily represent
major investment commitments; rather, those measures of risk taking could be viewed as lessbinding strategic options where an organization is making an initial attempt to engage in some
less-known activities. A strong argument could be made that while R&D and partnership ties
involve uncertainty and unpredictability in outcomes, they are considerably less-risky strategies
compared to the measures I have utilized in this dissertation. Thus, it is easier for managers to
engage in these types of behaviors when facing poor performance rather than make fixed
resource commitments in the form of capital investments.

54

The finding that loan risk increases under conditions of poor divisional performance
could be understood when considering what loan risk represents. Loan risk is operationalized as
the sum of non-accrual loans and one half of loans past due 90 days to total loans. Thus, loan risk
represents the realized risk of prior managerial lending choices and reflects the quality of loans
from prior periods. Thus, loan risk should not be seen as a managerial intention to engage in
risky initiatives but as an outcome of prior unsuccessful decisions. In other words, it is not
surprising to see that poorly performing divisions are associated with large loan risks. While in
my regressions I lagged performance discrepancy compared to loan risk, it could not be argued
with certainty that bad loans (associated with high loan risk) did not simultaneously exist with
poor performance.
Regarding the theorizing about how divisions and single firms respond to performance
discrepancies, the majority of my results are insignificant with two exceptions. First, when loan
risk was used as the dependent variable, results conformed to H2. However, the interpretation of
these finding should be tentative considering what actually loan risk represents (as argued
above). Second, when I used capital expenditures in the Compustat sample, single firms
appeared to reduce their risk taking rather than increase it when experiencing performance
shortfalls. One explanation to this finding might relate to the fact that single firms are smaller
than diversified corporations. In a supplementary analysis, I found that single firms have fewer
assets, employees, and financial resources compared to divisions. Prior research argues that
smaller firms are less insulated from poor performance and thus are more likely to experience
threat rigidity in the face of poor performance (Audia and Greve, 2006). Small firms are not
characterized by “extensive financial assets, manufacturing infrastructure, and a large
workforce” (Audia and Greve, 2006: 86; Levinthal, 1991) which limit their chances to

55

“overcome problems that threaten their survivals” (Mitchell, 1994: 577). In other words, smaller
firms do not possess and are limited in obtaining the necessary financial resources (Brüderl and
Schüssler, 1990) to be deployed in major commitments, such as capital expenditures.
In addition to using organizational structure to distinguish between divisional and singlebusiness risk taking, this dissertation introduced a set of multi-level organizational and
contextual factors that were found to significantly moderate the risk taking of corporate
divisions. Overall, I utilized factors from the divisional and corporate levels of analysis that
either enhance or restrict the ability and motivation of divisions to engage in risk taking. At the
divisional level of analysis, I examined the moderating role of divisional size. I expected that
among larger divisions performance deficiencies will encourage more risk taking because larger
divisions have more access to corporate resources. However, the results do not support this
argument. Smaller divisions in the Compustat sample were more responsive to performance
discrepancies. On one hand, their flexibility, lower inertial pressures and lower resistance to
change (Aldrich and Auster, 1986; Hannan and Freeman, 1984; Mintzberg, 1979) generate
stronger incentives to respond to performance shortfalls via risky activities. On the other hand,
smaller divisions could still have access to internal corporate resources and obtain capital
funding from the corporate office in contrast to small stand-alone firms that have no such
options. Overall, across the Compustat and the Banking samples (here all results are
insignificant), the results on the moderating role of divisional size are inconclusive and suggest
that two competing forces might be at play. The flexibility and low inertia of small divisions to
undertake actions are counterbalanced by the greater ability of large divisions to obtain necessary
funding for investing. As a result, the moderating effect of division size is diminished and
reduced to statistically insignificant.

56

At the corporate level, I studied how structural and resource endowment factors impact
intensity of risk taking. The role of level of diversification was primarily non-significant with
two exceptions – when capital expenditures (with historical aspirations) and loan risk (with
social aspirations) were used as a dependent variable. In the Compustat sample, as diversification
increased, divisions were more responsive to performance cues. Since diversification is
represented by the entropy measure, higher levels of diversification indicate more divisions
operating in diverse and distinctive product markets. As a result, more divisions face various
environments which increase their exposure to a variety of events and ideas (Barkema and
Vermeulen, 1998). Ultimately, this could lead to them having the ability and knowledge to
engage in more risk taking when faced with performance discrepancies.
In the Banking sample, diversification played a negative moderating role when loan risk was the
dependent variable. The distinction from the Compustat sample could be attributed to several
factors. First, diversification here is measured as the number of banking divisions and reflects the
geographic rather than product dispersion of division. As the number of divisions increases, it
raises the internal competition among those very similar divisions for resources (Liebeskind,
2000). Thus, it becomes harder for the corporate office to decide which of the underperforming
divisions to allocate resources to and overall this could lead to reduced risk taking of the latter.
Second, it is likely that diversified banking corporations have highly routinized loan decision
processes which encourage consistency in managerial decision making (Sutcliffe and
McNamara, 2001) and thus reduce the likelihood of managers underwriting bad loans and
increasing the loan risk of their divisions. Finally, research shows that banking managers “are
conservative in adjusting to new information” (McNamara and Bromiley, 1999 p. 333; Sutcliffe and
McNamara, 2001) which would limit their ability to see emerging opportunities and engage in

additional risk taking.
57

Internal governance factors, in the form of CEO contingent pay, provided consistent
results on their impact on risk taking intensity. The stronger those factors were, the lower the
degree to which performance discrepancies encouraged risk taking across divisions. While this
finding might seem surprising in terms of the proposed hypotheses, it has a logical and
theoretical explanation. When divisions experience poor performance the only way to engage in
more risk taking is by receiving additional funds from the corporate office. However, this usually
happens at the expense of other (well-performing) divisions (Rajan et al., 2000; Scharfstein and
Stein, 2000). As a result, to a certain degree the engagement of underperforming divisions in risk
taking is conducted at the expense of achieving optimal internal capital allocations. When CEOs
and corporate executives have compensation designs that encourage them to be better monitors
of divisional strategies, those executives appear to reduce internal inefficiencies through
discouraging risk taking by troubled divisions. The gap between CEO pay and divisional
managers’ pay discouraged risk taking in the face of poor performance. While contrary to
hypothesized, this finding is not inconsistent with prior research. It has been argued that instead
of motivating lower-level managers to engage in beneficial decisions, larger pay gaps could
discourage them from taking actions that are appropriate to the organization (e.g. Henderson and
Fredrickson, 2001), such as taking more risks when performance deteriorates.
External governance factors, such as institutional ownership and blockholders, did not
produce statistically significant results. In other words, those factors did not impact the
responsiveness of divisional managers to poor performance. One potential explanation is the
relative distance between those mechanisms and the actions of divisional managers. Usually,
external monitors exercise pressure on the CEO and corporate executives and hold them directly
accountable for the performance of the corporation. Then those executives are expected to

58

oversee divisional managers and their actions. As a result, multidivisional corporations are faced
with a double agency problem where one set of agents (CEOs and corporate executives) need to
monitors lower-level agents (divisional managers) (Child and Rodriguez, 2003). Overall, this
could decrease the strength and efficiency of external monitoring mechanisms. The different
impact (or lack of impact) between internal and external governance raises two important issues.
First, it is possible that different organizational actions residing at different levels of analysis are
better managed through more proximal sets of governance mechanisms (e.g. Devers et al., 2008).
Because CEOs are relatively closer to divisional managers and could oversee them more easily,
the former’s contingent pay may be more effective in influencing divisional responsiveness to
performance shortfalls. Second, a possibility for substitution between external and internal
governance mechanisms exists. In other words, external mechanisms might not be necessary
when strong internal mechanisms exist. To test this argument, I conducted a post-hoc analysis
where I excluded from the model the interaction terms of performance shortfalls and internal
governance factors. In that case, the interaction term between performance discrepancy and
blockholders was negative and significant (p<0.05) indicating that external governance
mechanisms influenced divisional responsiveness to performance deficiencies. Future work
could extend this line of thought and investigate what the optimal combination of internal and
external governance mechanisms might be for aligning the interests of shareholders, corporate
executives, and divisional managers.
Finally, the existence of resource endowments in the form of slack resources exhibits
some interesting interactions with attainment discrepancy. The results indicate that potential
slack and free cash flow have contrasting moderating effects. More specifically, low levels of
potential slack and high levels of free cash flow enhance the responsiveness of divisions to

59

attainment discrepancies. In other words, these two types of slack conform to two different
perspectives on slack – the hunger-driven view (Bromiley, 1991; Cyert and March, 1963) and
experimentation view (Sigh, 1986). While free cash flows are readily available resources to be
deployed when divisions face poor performance, potential slack indicates only ability to access
external funding and attempts to obtain such resources would be made only when the corporation
is deprived of other available resources. Furthermore, slack (free cash flow) had a different
moderating role on loan and liquidity risk. For liquidity risk, slack enhances the responsiveness
of banking divisions to low performance via engaging in more liquidity risks. Liquidity risk
represents ability to meet current obligations (Wiseman and Catanach, 1997) and the more free
cash flow available, the more likely managers are to pay those obligations even when
performance is poor. When free cash flow interacts with attainment discrepancy, the combined
effect is reduction in loan risks because some of those available funds could be used to recover
bad loans and ultimately reduce the exposure to loan risks.
Overall, this chapter extended research on the organizational responsiveness to poor
performance by developing a model where the ability and motivation of divisional managers to
address performance shortfalls is dependent on factors residing in the broader organizational and
structural context. Аs a result, this chapter contributes to existing research on risk taking by
outlining the contingencies under which performance deficiencies encourage or discourage
additional risk taking. Moreover, the results presented here help to resolve some of the
conflicting findings in prior research (Greve, 1998, Greve, 2008; Iyer and Miller, 2008;
Wiseman and Bromiley, 1996) by highlighting that the relationship between poor performance
and risk taking is more complex than previously thought. In addition to the moderating factors
impacting this relationship, the type of risk choices also affected the sign and strength of that

60

relationship. I found that when risk taking is represented by major fixed commitments, such as
capital expenditures, poor performance makes managers more hesitant and reluctant to make
those commitments due to fear of generating further losses. This finding offers potential avenues
for future research by pointing that further efforts need to be made to understand the
multidimensional nature of risk taking and the complex interactions between those different
types of risks. Furthermore, while prior research has focused on how the distance from an
aspiration level encourages risk taking, my dissertation presents evidence that the degree of
commitment inherent in a risky choice (e.g. high degree of commitment in large fixed
investments, such as capital expenditures versus less-binding obligations across R&D
investments and partnership ties) might play an important role in how poor performance affects
managerial risk taking.
One interesting finding across the two samples and multiple regression analyses is the
relatively weaker results when social aspirations were utilized. While those findings are
consistent with prior research (e.g. Audia and Greve, 2006; Greve, 1998; Iyer and Miller, 2008),
it still unclear why that is the case. Prior studies have focused more on how to construct
aspirations – separate historical and social levels, combination of the two levels (e.g. Greve,
2003) or substitution between the two aspirations depending on where performance is against the
aspiration (e.g. Bromiley, 1991) – rather than develop theoretical frameworks whether and why
social and historical aspirations generate different results. It might be that managers are more
familiar with historical aspirations, receive more consistent cues from the latter and are willing to
pay more attention to those aspiration levels rather than to social ones; it might be that managers
selectively choose aspiration levels which depict their organizations in a positive light vis-à-vis
shareholders (e.g. Audia and Brion, 2005). However, it is also possible that the surrounding

61

context influences where managers look for performance cues; in other words, various factors
from that context could serve as important attention allocation mechanisms (Ocasio, 1997). One
such factor could be organizational structure and more specifically the structural embeddedness
of divisions in a multidivisional corporation. This chapter presents some tentative evidence that
organizational structure leads to prioritizing of historical aspirations over social aspirations.
Because the multidivisional firm partially isolates divisions from the external market (Bradley et
al., 2011) and utilizes different financial controls and monitoring compared to the free market,
divisional managers become primed to focus internally on their own historical performance. In
addition, this quasi isolation from external markets could create the perception among divisional
managers that other firms in the market are more distinct than their divisions and thus managers
could overlook social comparisons (e.g. Greve, 2003b). Future research should delve deeper into
other factors, such as firm characteristics, resource endowments, and market conditions that
make firms focus more on social or historical aspirations.
As any study, this one is not void of limitations. All the proposed relationships relied on
secondary data where the behavior of managers is implied. Although the arguments and
theorizing are based on well-established theoretic frameworks, it is hard to disclose the real
motivations and perceptions of managers unless surveys and observations of the latter are
conducted. The choice of the banking sample and its risk measures could be questioned due to
the timing of the study. Some could argue that during this period the banking industry was
expanding and bank managers might have perceived their investments and risky choice as less
uncertain. However, this should create a common bias among all managers and thus should not
influence the variability in amount of risk taken and the quality of risk across multiple banks. In
addition, if such a bias exists, it would make it more difficult to find statistically significant

62

results; the fact that I find significant results indicates that this is a conservative estimation of the
proposed relationships. Using measures that are specific to an industry is often a challenge and
might raise questions regarding the validity of those measures. Although my bank risk measures
relied on prior research (Shrieves and Dahl, 1992; Wiseman and Catanach, 1997), it could be
argued that those measures, and loan risk in particular, represent realized risks and reflect prior
decisions by managers rather than future risky choices.
In addition, liquidity risk is very similar to available slack – the measurement of this risk
indicates that lower liquid resources reflect higher risk of not covering obligation. While this is
definitely a risk for a bank, it is very different from investment and/or market risks.
The selection of capital expenses could also be put under question. However, I relied on prior
theorizing and research to utilize this measure as reflecting risk taking. Capital expenditures are
major capital outlays with uncertain outcomes and once the organization commits to them it is
hard to reverse them; thus, changes in the external conditions could put such an organization in a
position to experience significant losses (Desai, 2008).
In conclusion, this chapter examined the embeddedness of organizational risk taking in a
larger organizational context and the impact of the latter on the intensity of risk taking. The
chapter also suggests that to better understand organizational risk scholars should 1) consider the
impact of the organizational structure and context; 2) examine how multi-level factors enhance
or inhibit the intensity of risk taking; and 3) distinguish between various types of organizational
risks.

63

Table S: Summary of hypotheses and findings for Chapter 1
Hypothesis
H1: Divisional negative attainment discrepancy is positively
related to divisional risk taking.

Compustat findings
No support – NS

H2: The relationship between divisional negative attainment
discrepancy and divisional risk taking will be weaker in
divisions of M-firms than in single-business companies.
H3: Level of corporate diversification moderates the
relationship between divisional negative attainment discrepancy
and risk taking, such that the relationship will be weaker as
diversification increases.
H4: Divisional size will moderate the relationship between
divisional negative attainment discrepancy and divisional risk
taking, such that the relationship will be stronger for larger
divisions.
H5a: Divisional managers’ ownership will moderate the
relationship between divisional negative attainment discrepancy
and divisional risk taking, such that the relationship will be
stronger with higher levels of ownership.
H5b: Divisional managers’ stock options will moderate the
relationship between divisional negative attainment discrepancy
and divisional risk taking, such that the relationship will be
stronger with higher levels of stock options.

No support – single-firms
reduce risk taking (HA
and SA)
No support – results are in
the opposite direction
(HA)

64

Banking findings
Support for loan risk as DV (HA
and SA)
Opposite results for liquidity risk
as DV (SA)
Opposite results for loan risk as
DV (HA and SA)
Support only for loan risk as DV
(HA)
Support for loan risk as DV (SA)

No support – results are in No support – NS
the opposite direction (HA
and SA)
No support – NS

NA (not tested)

No support – NS

NA (not tested)

Table S (cont’d)
H6: CEO contingent pay will moderate the relationship between
divisional negative attainment discrepancy and divisional risk
taking, such that the relationship will be stronger with higher
levels of contingent pay.

No support – results are in * No support – opposite results
the opposite direction (HA for CEO shares when using bank
and SA)
risk and liquidity risk as DV
(HA)
* Support for CEO options when
using bank risk and loan risk as
DV (HA)

H7: The gap between CEO pay and divisional managers’ pay
will moderate the relationship between divisional negative
attainment discrepancy and divisional risk taking, such that the
relationship will be stronger as the pay gap widens.
H8a: Institutional ownership will moderate the relationship
between divisional negative attainment discrepancy and
divisional risk taking, such that the relationship will be stronger
with higher levels of institutional ownership.
H8b: Blockholders will moderate the relationship between
divisional negative attainment discrepancy and divisional risk
taking, such that the relationship will be stronger with larger
number of blockholders.
H9: Corporate slack will moderate the relationship between
divisional negative attainment discrepancy and divisional risk
taking, such that the relationship will be stronger with higher
levels of slack.

No support – results are in * No support – opposite results
the opposite direction (HA for loan risk as DV (HA)
and SA)
No support – NS

* No support – opposite results
for loan risk (only SA)

No support – NS

No support – NS

Support for free cash flow
(HA)
Opposite results for
potential slack (only HA)

Support for liquidity and bank
risks as DVs (HA and SA)
Opposite results for loan risk as
DV (HA and SA)

DV – dependent variable; HA – historical aspirations; SA – social aspirations

65

* tested on a smaller sample

TABLE A1: Descriptive statistics and correlation coefficients for the study performance discrepancies – risk taking
M
1. Risk taking
143.3
2. Sales growth
0.07
3. Div. size
6.1
4. PBH
0.05
5. PAH
0.04
6. PBS
0.05
7. PAS
0.11
8. Relative size
0.35
9. Diversification
0.51
10. Potential slack 1.1
11. Free cash flow 0.05
12. CEO shares
1838
13. CEO options
2197
14. Pay gap
7.6
15. Inst. ownership 0.55
16. Blockholders
1.68

SD
357.1
0.25
1.85
0.19
0.19
0.18
0.27
0.26
0.41
1.35
0.39
6257
4120
1.3
0.31
1.6

1
1.00
0.06
0.53
-0.04
-0.02
-0.05
-0.04
0.23
0.05
-0.08
0.02
0.18
0.18
0.24
-0.06
-0.13

2

3

4

5

6

7

8

1.00
0.07
-0.15
0.05
-0.08
0.03
0.05
-0.05
0.02
0.10
0.02
0.01
0.05
0.10
-0.01

1.00
-0.10
-0.14
-0.12
-0.19
0.34
0.21
-0.32
0.03
0.16
0.32
0.46
0.01
-0.19

1.00
-0.07
0.60
0.05
-0.05
-0.05
0.03
0.05
0.00
0.00
-0.02
0.00
0.03

1.00
0.01
0.28
-0.07
-0.02
0.04
0.00
-0.02
0.00
0.00
0.00
0.04

1.00
-0.13
-0.08
-0.07
-0.02
0.01
0.02
-0.02
-0.02
0.02
0.06

1.00
-0.03
-0.06
0.17
0.10
-0.02
0.06
0.05
0.05
0.01

1.00
-0.26
0.08
-0.01
0.00
-0.06
-0.10
0.02
0.03

66

TABLE A1 (cont’d)
9
1. Risk taking
2. Sales growth
3. Div. size
4. PBH
5. PAH
6. PBS
7. PAS
8. Relative size
9. Diversification
10. Potential slack
11. Free cash flow
12. CEO shares
13. CEO options
14. Pay gap
15. Inst. ownership
16. Blockholders
N=4982

10

11

12

13

1.00
-0.17
0.02
0.02
0.16
0.21
-0.10
-0.15

1.00
0.03
0.03
-0.04
-0.21
-0.02
0.05

1.00
-0.01
0.04
0.10
0.07
-0.09

0.15
0.10 0.58
-0.09 0.03 0.20
-0.07 -0.15 -0.07 0.58

Correlations larger than 0.03 and smaller than -0.02 are significant at p<0.05

67

14

TABLE B1: Descriptive statistics and correlations for the study performance discrepancies – risk taking

1. Bank risk
2. Loan risk
3. Liquidity risk
4. Sales growth
5. Divisional size
6. Divisional age
7. PBH
8. PAH
9. PBS
10. PAS
11. Relative size
12. Slack
13. Diversification
N= 8454

M
0.69
0.01
0.91
0.08
12.97
63.8
0.003
0.002
0.002
0.003
0.29
-0.01
7.8

SD
0.14
0.02
0.10
0.15
1.58
44.6
0.004
0.004
0.003
0.004
0.27
0.09
8.9

1
1.00
0.08
0.35
0.07
0.14
-0.06
0.03
0.01
0.01
-0.01
0.10
0.03
0.03

2

3

4

5

6

7

8

9

10

11

12

1.00
-0.03
-0.03
0.06
-0.07
0.24
-0.07
0.23
-0.04
0.05
-0.17
-0.03

1.00
0.03
-0.05
-0.01
-0.03
0.01
0.00
-0.03
-0.01
0.02
-0.02

1.00
0.15
-0.11
-0.09
0.08
-0.04
0.04
0.09
0.15
-0.02

1.00
0.16
0.03
0.02
-0.01
0.07
0.35
-0.03
0.07

1.00
-0.04
-0.03
-0.05
-0.04
0.23
-0.09
-0.07

1.00
-0.36
0.50
-0.25
-0.01
-0.44
0.02

1.00
-0.24
0.51
-0.03
0.27
0.02

1.00
-0.37
-0.01
-0.42
-0.01

1.00
-0.04
0.27
0.04

1.00
0.05
-0.51

1.00
-0.06

Correlations larger than 0.02 and smaller than -0.02 are significant at p<0.05

68

TABLE A2: Performance discrepancy-risk taking relationship based on historical aspirations

Variables
Constant
Sales growth
Divisional size
Risk taking t-2
Historical attainment discrepancy <0
Historical attainment discrepancy > 0
Relative divisional size
Diversification
Potential slack
Free cash flow
CEO shares
CEO options
Pay gap
Institutional ownership
Blockholders

Model 1
Model 2
Coeff
SE
Coeff
SE
136.4***
(10) 134.2***
(9.9)
7.5***
(2.1)
6.74**
(2.1)
67.5***
(5.3)
70.6***
(5.4)
195.5***
(4.55)
192***
(4.5)
3.7
(2.5)
2.8
(3.5
5.3†
(3.2)
6.7
(5.3)
4.96
(3.7)
4.4
(3.7
-6.4†
(3.8)
-5.2
(3.8)
7.1†
(4.2)
7.4†
(4.2)
4.2
(3)
4.4
(3.1)
8.5*
(4.2)
8.3†
(4.5)
-3.7
(3.2)
-4.7
(3.2)
6.98*
(3.22)
5.9†
(3.2
-5.6†
(4.7)
-7.2†
(4.7)
-3.04
(3.1)
-2
(3.1)

Relative divisional size X
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Diversification
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Potential slack X
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Free cash flow X
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
CEO shares
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
CEO options
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Pay gap
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Institutional ownership
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0

-23.3***
-2.7
10.99***
0.3

(2.5)
(2.6)

-13***
-4.6

(2.9)
(4.5)

8***
1.8

(1.6)
(4.7)

-11.3**
4.6

(3.7)
(11.2)

-17.9**
21.4***

(6.1)
(6)

-24.1***
-8.5

(4.2)
(5.1)

-8.3†
1.51

69

(2.9)
(4.6)

(4.3)
(6.9)

TABLE A2 (cont’d)
Blockholders
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0

-3.8
2.7

N
2
Wald χ

5583
3427
0.000

0.000

0.54
0.31
0.15

2

Prob > χ
Variance within divisions
Variance between divisions
Variance between firms

5583
3472
0.51
0.33
0.16

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

70

(3.8)
(3.8)

TABLE A3: Performance discrepancy-risk taking relationship based on social aspirations

Variables
Constant
Sales growth
Divisional size
Risk taking t-2
Social attainment discrepancy < 0
Social attainment discrepancy > 0
Relative divisional size
Diversification
Potential slack
Free cash flow
CEO shares
CEO options
Pay gap
Institutional ownership
Bockholders

Model 3
Model 4
Coeff
SE
Coeff
SE
140.3***
(10.6) 140.3***
(10.7)
7.5***
(2.1)
7.2**
(2.2)
75.2***
(5.8)
77.4***
(5.8)
194.9***
(4.7)
192.8***
(4.7)
1.8
(2.4)
4.3
(3.7)
10.7**
(3.97)
16.4*
(6.6)
4.2
(3.95)
4.1
(3.99)
-6.5
(4.04)
-6.1
(4.1)
7.6†
(4.5)
7.9†
(4.5)
4.04
(3.2)
4.99
(3.3)
9.5*
(4.3)
7.6†
(4.4)
-5.8†
(3.3)
-4.99
(3.3)
9.04**
(3.4)
7.1*
(3.4)
-7.1
(4.9)
-8.2†
(4.96)
-1.2
(3.3)
-0.9
(3.3)

Relative divisional size X
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0
Diversification
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0
Potential slack X
x Social attainment discrepancy < 0
x h Social attainment discrepancy > 0
Free cash flow X
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0
CEO shares
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0
CEO options
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0
Pay gap
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0
Institutional ownership
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0

-11.1***
11.4*
-2.5
5.2

(2.95)
(4.6)

4.5
-0.7

(4.6)
(3.9)

0.5
-0.03

(1.7)
(3.2)

-7.7**
-13.9†

(2.8)
(7.4)

5.3
2.96

(5.2)
(4.5)

-10.4*
-6.4

(4.2)
(4.7)

-3.7
0.9
71

(2.9)
(4.9)

(3.7)
(7.6)

TABLE A3 (cont’d)
Blockholders
x Social attainment discrepancy < 0
x Social attainment discrepancy > 0

-2.6
-3.7

N
2
Wald χ

5073
3258
0.000

0.000

0.49
0.36
0.15

2

Prob > χ
Variance within divisions
Variance between divisions
Variance between firms

5073
3254
0.49
0.36
0.15

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

72

(3.4)
(4.8)

TABLE A4: Hypothesis Two – comparing risk taking between single-business firms and divisions

Variables
Constant
Sales growth
Divisional size
Risk taking t-2
Historical attainment discrepancy <0
Historical attainment discrepancy > 0
Social attainment
discrepancy <0
Social attainment
discrepancy > 0
Diversification dummy
Potential slack
Free cash flow
CEO shares
CEO options
Pay gap
Institutional ownership

Model 5
Model 6
Model 7
Model 8
Coeff
SE
Coeff
SE
Coeff
SE
Coeff
SE
128.3***
(7.8) 126.4**
(7.9) 128.4***
(8.4) 126.2***
(8.4)
12.31*** (1.90)
11.7*** (1.9)
12.00*** (1.96)
11.88*** (1.96)
66.27*** (3.67)
66.6*** (3.7)
71.15*** (3.96)
71.20*** (3.96)
229.23*** (2.72)
228.2*** (2.7)
230.49*** (2.81)
230.29*** (2.82)
2.27 (2.01)
-16.9** (6.3)
4.16† (2.49)
-1.13 (5.7)
1.13

(5.51)
(2.70)
(2.77)
(2.93)
(2.54)
(2.65)
(3.31)

-4.07
1.8
5.05†
3.2
-2.2
8.4**
-5.13

21.1**
6.11

Diversification dummy
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Diversification dummy
x social attainment discrepancy < 0
x social attainment discrepancy > 0
N
2
Wald χ

(5.6)
(2.7)
(2.8)
(2.9)
(2.5)
(2.7)
(3.3)

-12.94†

(6.75)

8.42**
-5.13
1.99
4.20
3.31
-3.69
8.66**
-5.36

(2.65)
(5.84)
(2.83)
(3.01)
(3.08)
(2.62)
(2.79)
(3.48)

4.11
-2.50
1.77
3.42
3.31
-3.48
8.57**
-5.84†

(5.44)
(5.96)
(2.84)
(3.03)
(3.08)
(2.62)
(2.79)
(3.49)

16.04*
5.31

-6.52
1.87
6.07*
3.25
-2.48
8.22**
-4.73

(2.42)

(7.19)
(6.03)

6.6
6.3

9145
12614
2

Prob > χ

9145
12458

8403
12132

8403
12134

0.000

0.000

0.000

0.000

73

TABLE A4 (cont’d)
Variance within divisions
Variance between divisions
Variance between firms

0.72
0.14
0.13

0.72
0.14
0.14

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

74

0.71
0.17
0.12

0.71
0.17
0.12

TABLE B2: Relationship between performance discrepancy and risk taking based on historical aspirations
Model 1 –
Bank Risk
Variables
Constant
Sales growth
Divisional size
Divisional age
Risk taking t-2
PBH
PAH
Relative
divisional size
Slack
Diversification
PBH * relative
size
PAH * relative
size
PBH *
diversification
PAH *
diversification
PBH * slack
PAH * slack

Model 2 –
Bank Risk

Model 3 –
Loan Risk

Model 4 –
Loan Risk

Coeff
SE
Coeff
SE
Coeff
SE
Coeff
71.1*** (0.33) 71.1*** (0.33) 0.84*** (0.06)
0.80***
-0.34*** (0.07) -0.33*** (0.07)
-0.03* (0.01)
-0.04**
0.89*** (0.19) 0.90*** (0.19) 0.14*** (0.03)
0.13***
-0.87*** (0.14) -0.88*** (0.15) -0.09*** (0.02) -0.09***
7.83*** (0.12) 7.80*** (0.12) 0.49*** (0.02)
0.49***
-0.27*** (0.08)
-0.19* (0.09) 0.17*** (0.02)
0.06**
0.04 (0.07)
0.02 (0.08)
0.02 (0.01)
-0.01
0.82***
0.22**

(0.17)
(0.08)

0.82***
0.18*

0.62**

(0.23)

Model 5 –
Liquidity risk

SE
Coeff
(0.06) 92.7***
(0.01)
-0.03
(0.03)
0.02
(0.02) -0.16**
(0.02) 7.77***
(0.02)
-0.10†
(0.02)
0.03

Model 6 –
Liquidity risk

SE
Coeff
(0.18) 92.7***
(0.05)
-0.03
(0.08)
0.03
(0.06) -0.16**
(0.06) 7.78***
(0.06)
-0.05
(0.05)
0.03

(0.17)
0.01
(0.08) -0.08***

(0.03)
(0.02)

0.01 (0.03)
-0.05** (0.02)

0.61**

(0.23)

(0.04)

0.17*** (0.04)

0.02

(0.08)

0.02 (0.02)

0.02 (0.06)

-0.04

(0.08)

0.01 (0.02)

0.02 (0.06)

0.08

(0.07)

0.01 (0.02)

0.09† (0.05)

0.14†
0.10*
0.13†

(0.08)
(0.04)
(0.07)

0.01 (0.02)
-0.10*** (0.01)
-0.04* (0.01)

-0.03 (0.06)
0.06* (0.03)
0.07 (0.05)

0.16***

75

-0.03 (0.08)
-0.05 (0.06)

SE
(0.18)
(0.05)
(0.08)
(0.06)
(0.06)
(0.07)
(0.06)

-0.27*

0.12

-0.03 (0.08)
-0.07 0.06)
-0.28* (0.12)

TABLE B2 (cont’d)

N
2
Wald χ

10361
5813
2

Prob > χ
Within division
variance
Between div.
variance
Between firm
variance

10361
5798

10327
5590

10327
5804

10225
15764

10225
15785

0.000

0.000

0.000

0.000

0.000

0.000

0.48

0.48

0.67

0.68

0.92

0.9

0.26

0.27

0.06

0.06

0.00

0.00

0.26

0.26

0.27

0.26

0.08

0.08

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

76

TABLE B3: Relationship between performance discrepancy and risk taking based on social aspirations
Model 1 –
Bank Risk
Variables
Constant
Sales growth
Divisional size
Divisional age
Risk taking t-2
PBS
PAS
Relative
divisional size
Slack
Diversification
PBS * relative
size
PAS * relative
size
PBS *
diversification
PAS *
diversification
PBS * slack
PAS * slack

Model 2 –
Bank Risk

Model 3 –
Loan Risk

Model 4 –
Loan Risk

Model 5 –
Liquidity risk

Model 6 –
Liquidity risk

Coeff
SE
Coeff
SE
Coeff
SE
Coeff
SE
Coeff
SE
Coeff
71.1*** (0.35) 71.2*** (0.36) 0.82*** (0.07) 0.77*** (0.07) 92.6*** (0.20) 92.6***
-0.39*** (0.07) -0.38*** (0.07)
-0.04* (0.02)
-0.04** (0.02)
-0.03 (0.05)
-0.02
0.78*** (0.20) 0.78*** (0.20) 0.14*** (0.03) 0.13*** (0.03)
-0.05 (0.08)
-0.05
-0.85*** (0.16) -0.86*** (0.16) -0.11*** (0.02) -0.10*** (0.02)
-0.16* (0.06)
-0.17*
7.77*** (0.13) 7.77*** (0.13) 0.48*** (0.02) 0.48*** (0.02) 7.83*** (0.07) 7.83***
-0.24** (0.09)
-0.22† (0.12) 0.18*** (0.02)
0.04 (0.02)
-0.14* (0.06)
-0.03
0.04 (0.09)
0.05 (0.10)
0.01 (0.02)
-0.04† (0.02)
-0.12* (0.06)
-0.10
0.84***
0.24**

(0.19)
(0.09)

0.84***
0.24**

0.61*

(0.24

SE
(0.20)
(0.05)
(0.08)
(0.06)
(0.07)
(0.09)
(0.07)

(0.19)
0.03
(0.09) -0.09***

(0.03)
(0.02)

0.03
-0.07**

(0.03)
(0.02)

-0.01
0.03

(0.08)
(0.06)

-0.01 (0.08)
0.00 (0.06)

0.61*

(0.24)

(0.05)

0.20***

(0.05)

-0.22†

(0.13)

-0.21 (0.13)

0.16†

(0.09)

0.03

(0.02)

0.06 (0.06)

0.06

(0.09)

0.01

(0.02)

0.03 (0.07)

0.04

(0.09)

-0.04*

(0.02)

0.12† (0.06)

0.05
0.02
0.01

(0.10)
(0.04)
(0.07)

0.01
-0.08***
-0.01

(0.02)
(0.01)
(0.01)

0.10 (0.07)
0.08* (0.03)
0.06 (0.05)

0.20***

77

TABLE B3 (cont’d)

N
2
Wald χ

8599
4825
2

Prob > χ
Within division
variance
Between div.
variance
Between firm
variance

8599
4832

8565
4733

8565
4889

8486
13816

8486
13854

0.000

0.000

0.000

0.000

0.000

0.000

0.45

0.45

0.66

0.67

0.93

0.93

0.29

0.29

0.07

0.07

0.00

0.00

0.26

0.26

0.27

0.26

0.07

0.07

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

78

TABLE B4: Hypothesis Two – comparing risk taking between single banks and banking divisions (using historical aspirations)

Variables
Constant
Sales growth
Divisional size
Divisional age
Risk taking t-2
PBH
PAH
Slack
Diversification
dummy

Model 7 –
Model 8 –
Model 9 –
Model 10 –
Bank Risk
Bank Risk
Loan Risk
Loan Risk
Coeff
SE
Coeff
SE
Coeff
SE
69.5*** (0.14) 69.4*** (0.14) 0.92***
(0.03) 0.93***
(0.03)
-0.23*** (0.03)
-0.23** (0.03) -0.05*** (0.01) -0.05***
(0.01)
0.60*** (0.07) 0.59*** (0.07) 0.23*** (0.01) 0.23***
(0.01)
-1.24*** (0.08) -1.24*** (0.08)
-0.23* (0.01) -0.23***
(0.01)
6.45*** (0.05) 6.45*** (0.05)
0.54 (0.01) 0.54***
(0.01)
-0.25*** (0.03) -0.23***
(0.0) 0.20*** (0.01) 0.22***
(0.01)
0.02 (0.03)
0.02 (0.04)
0.05** (0.01) 0.05***
(0.01)
0.21*** (0.04) 0.21*** (0.00) -0.12*** (0.01) -0.12***
(0.01)

92.7***
0.11***
-0.02
0.05*
7.56***
-0.09***
-0.02
-0.03

Model 12 –
Liquidity Risk

(0.08) 92.7***
(0.02) 0.11***
(0.02)
-0.02
(0.02)
0.05*
(0.02) 7.56***
(0.03) -0.09**
(0.02)
-0.03
(0.02)
-0.03

(0.08)
(0.02)
(0.02)
(0.02)
(0.02)
(0.03)
(0.03)
(0.02)

0.23

(0.14)

-0.04

(0.03)

0.01

(0.06)

-0.04

(0.07)

-0.06***

(0.02)

0.02

(0.05)

0.02

PBH *
diversification
PAH *
diversification
N
2
Wald χ

Model 11 –
Liquidity Risk

(0.07)

-0.02

(0.02)

0.05

(0.05)

41075
22189
2

Prob > χ

41075
22178

41031
22784

41031
22811

40706
106754

40706
106652

0.000

0.000

0.000

0.000

0.000

0.000

79

TABLE B4 (Cont’d)

Within division
variance
Between div.
variance
Between firm
variance

0.47

0.47

0.71

0.71

0.97

0.93

0.34

0.34

0.06

0.06

0.00

0.00

0.2

0.19

0.23

0.23

0.03

0.03

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

80

TABLE B5: Hypothesis Two – comparing risk taking between single banks and banking divisions (using social aspirations)

Variables
Constant
Sales growth
Divisional size
Divisional age
Risk taking t-2
PBS
PAS
Slack
Diversification
dummy

Model 7 –
Model 8 –
Model 9 –
Model 10 –
Bank Risk
Bank Risk
Loan Risk
Loan Risk
Coeff
SE
Coeff
SE
Coeff
SE
69.5*** (0.16) 69.5*** (0.16) 0.92***
(0.04) 0.94***
(0.04)
-0.23*** (0.03) -0.23*** 0.03) -0.06*** (0.01) -0.06*** (0.01)
0.52*** (0.08) 0.52*** 0.08) 0.25*** (0.01)
0.25*** (0.01)
-1.23*** (0.08) -1.23*** 0.09) -0.25*** (0.01) -0.25*** (0.01)
6.40*** (0.06) 6.40*** 0.06) 0.52*** (0.01)
0.52*** (0.01)
-0.17*** (0.04)
-0.16** 0.05) 0.18*** (0.01)
0.18*** (0.01)
-0.09* (0.04)
-0.12* 0.05) 0.06*** (0.01)
0.07*** (0.01)
0.33*** (0.04) 0.33*** 0.04) -0.15*** (0.01) -0.15*** (0.01)

92.5***
0.12***
-0.05†
0.05*
7.6***
-0.05
-0.13***
0.05†

Model 12 –
Liquidity Risk

(0.09) 92.7***
(0.02)
0.12***
(0.03)
-0.05*
(0.03)
0.05*
(0.03)
7.6***
(0.03)
-0.03
(0.03) -0.14***
(0.03)
0.05†

(0.08)
(0.02)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)
(0.03)

0.21

0.16)

-0.07†

(0.04)

0.05

(0.07)

-0.05

(0.08)

-0.03

(0.02)

-0.07

(0.06)

0.07

PBS *
diversification
PAS *
diversification
N
2
Wald χ

Model 11 –
Liquidity Risk

(0.08)

-0.03

(0.02)

0.01

(0.05)

33785
18015
2

Prob > χ

33785
18001

33741
19124

33741
19137

33471
91964

33471
91810

0.000

0.000

0.000

0.000

0.000

0.000

81

TABLE B5 (cont’d)

Within division
variance
Between div.
variance
Between firm
variance

0.46

0.46

0.7

0.7

0.98

0.98

0.36

0.36

0.07

0.07

0.00

0.00

0.19

0.19

0.24

0.24

0.02

0.02

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

82

FIGURE A1: Interaction between performance discrepancy and diversification dummy
comparing divisions and single firms (Compustat sample) *

160

140
120
100

Risk
taking

80
60

40
20
0
_

low

Performance
Performance discrepancy

single firms

divisions

0
0

high/low +1/-1 SD

* For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this dissertation

83

FIGURE B1: Interaction between performance discrepancy and diversification dummy
comparing single banks and banking divisions (historical aspirations using loan risk)

1.4
1.2
1
Risk
taking

0.8

0.6
0.4
0.2
0
_

low

Performance
Performance discrepancy
single banks
banking divisions

0
0

high/low +1/-1 SD

FIGURE A2: Interaction between performance discrepancy and level of diversification
(Compustat sample)

145
140
135

Risk
taking

130
125
120
115
_

low

Performance
Performance discrepancy
low diversification
high diversification

84

0
0

high/low +1/-1 SD

FIGURE B2: Interaction between performance discrepancy and level of diversification based on
social aspirations (Banking sample and using loan risk)

1.2
1
0.8

Risk
taking

0.6
0.4
0.2
0
_

low

Performance
Performance discrepancy
low diversification
high diversification

0
0

high/low +1/-1 SD

FIGURE A3: Relationship between performance discrepancy and divisional relative size
(Compustat sample)

180
160
140
120

Risk
taking

100

80
60
40
20
0
_

low

Performance
Performance discrepancy
low relative size
high relative size

85

0
0

high/low +1/-1 SD

FIGURE A4: Relationship between performance discrepancy and CEO shares (Compustat
sample)

150
145
140
Risk
taking

135

130
125
120
115
_

low

Performance
Performance discrepancy
low CEO shares
high CEO shares

0
0

high/low +1/-1 SD

FIGURE A5: Relationship between performance discrepancy and CEO options (Compustat
sample)

160

140
120
100

Risk
taking

80
60

40
20
0
_

low

Performance
Performance discrepancy
low CEO options
high CEO options

86

0
0

high/low +1/-1 SD

FIGURE A6: Relationship between performance discrepancy and pay gap (Compustat sample)

180
160
140
120

Risk
taking

100

80
60
40
20
0
_

low

Performance
Performance discrepancy

low pay gap

high pay gap

0
0

high/low +1/-1 SD

FIGURE A7: Relationship between performance discrepancy and potential slack (Compustat
sample)

150

145
140
135

Risk
taking

130
125

120
115
110
_

low

Performance
Performance discrepancy
low potential slack
high potential slack

87

0
0

high/low +1/-1 SD

FIGURE A8: Relationship between performance discrepancy and free cash flow (Compustat
sample)

155
150
145
140

Risk
taking

135

130
125
120
115
110
_

low

Performance
Performance discrepancy
low free cash flow
high free cash flow

0
0

high/low +1/-1 SD

FIGURE B3: Interaction between performance discrepancy and free cash flow (Banking sample
and using loan risk)

1.2
1
0.8

Risk
taking

0.6
0.4
0.2
0
_

low

Performance
Performance discrepancy
low free cash flow
high free cash flow

88

0
0

high/low +1/-1 SD

FIGURE B4: Interaction between performance discrepancy and free cash flow (Banking sample
and using liquidity risk)

92.9

92.85
92.8
92.75

Risk
taking

92.7
92.65

92.6
92.55
92.5
_

low

Performance
Performance discrepancy
low free cash flow
high free cash flow

89

0
0

high/low +1/-1 SD

APPENDICES

90

APPENDIX 1
In this appendix I present supplementary analyses using hand-collected divisional managers’ compensation data (for the Compustat
sample) and internal and external governance data (for the Banking sample).

TABLE AP1: Correlation table for the reduced sample testing H5a-H5b

1. Risk
2. Sales growth
3. Div. size
4. PBH
5. PAH
6. PBS
7. PAS
8. Relative size
9. Divisional shares
10. Divisional options
11. Diversification
12. Potential slack
13. Free cash flow
14. CEO shares
15. CEO options
16. Pay gap
17. Inst. ownership
18. Blockholders

M
143.3
0.07
6.1
0.05
0.04
0.05
0.11
0.35
624.6
1180.1
0.51
1.1
0.05
1838
2197
7.6
0.55
1.68

SD
357.1
0.25
1.85
0.19
0.19
0.18
0.27
0.26
1131.3
1683.6
0.41
1.35
0.39
6257
4120
1.3
0.31
1.6

1
1.00
0.13
0.67
-0.07
-0.10
-0.09
-0.10
0.55
0.38
0.08
-0.24
0.19
0.01
0.51
0.00
0.14
-0.22
-0.14

2

3

4

5

6

7

8

9

10

1.00
0.16
-0.12
0.23
-0.21
0.10
0.10
0.20
0.07
-0.02
-0.02
0.13
0.09
-0.05
-0.02
0.14
0.02

1.00
-0.15
-0.21
-0.13
-0.19
0.59
0.41
0.18
-0.08
-0.04
0.04
0.27
0.14
0.31
-0.17
-0.29

1.00
-0.06
0.31
0.09
-0.03
-0.03
-0.04
-0.11
0.16
0.03
-0.04
0.00
-0.05
0.07
0.08

1.00
-0.06
0.52
-0.05
-0.06
-0.04
0.02
-0.11
0.02
-0.05
-0.04
-0.05
0.09
0.15

1.00
-0.15
0.03
-0.04
-0.02
-0.08
0.01
0.07
-0.04
0.00
-0.10
0.09
0.06

1.00
-0.04
-0.02
-0.10
0.04
0.03
0.08
-0.07
0.06
0.03
0.01
0.06

1.00
0.15
0.09
-0.34
0.28
0.07
0.38
0.07
0.09
0.02
-0.11

1.00
0.28
-0.02
-0.01
0.04
0.19
0.23
0.29
-0.20
-0.20

1.00
0.24
-0.02
-0.02
0.11
0.47
0.33
-0.04
-0.14

91

TABLE AP1 (cont’d)
11
1. Risk
2. Sales growth
3. Div. size
4. PBH
5. PAH
6. PBS
7. PAS
8. Relative size
9. Divisional shares
10. Divisional options
11. Diversification
12. Potential slack
13. Free cash flow
14. CEO shares
15. CEO options
16. Pay gap
17. Inst. ownership
18. Blockholders

N=292

12

13

14

15

16

17

1.00
-0.27 1.00
0.07 0.04 1.00
-0.15 0.29 -0.01 1.00
0.10 0.29 -0.08 0.04 1.00
0.11 0.11 -0.08 -0.07 0.64 1.00
-0.13 0.01 0.18 -0.34 -0.15 0.00 1.00
-0.03 -0.06 0.10 -0.05 -0.26 -0.16 0.46

Correlations larger than 0.11 and smaller than -0.11 are significant at p<0.05

92

TABLE AP2: Relationship between performance discrepancy and risk taking based on historical
aspirations (Compustat sample)

Variables
Constant
Sales growth
Divisional size
Risk taking t-2
Historical attainment discrepancy <0
Historical attainment discrepancy > 0
Relative divisional size
Divisional managers’ shares
Divisional managers’ options
Diversification
Potential slack
Free cash flow
CEO shares
CEO options
Pay gap
Institutional ownership
Bockholders

Model 1
Coeff
139.1
5.4
196.9***
176.2***
-2.2
4.5
4.3
53.4**
5.0
-48.7†
-3.6
-5.7
54.3*
-45.1
-29.1
-78.5
0.4

Relative divisional size X
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Divisional managers’ shares
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Divisional managers’ options
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Diversification
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Potential slack X
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Free cash flow X
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
CEO shares
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0

Model 2
SE
Coeff
SE
(101.2)
241.8*
(110.4)
(14.7)
7.5
(15.8)
(44.5)
144.3**
(45)
(16.6)
216.8***
(17.9)
(7.9)
17.7
(131.7)
(31.6)
65.8
(246.9)
(26.1)
4.59
(27.9)
(15.4)
75.5**
(23.6)
(24.7)
0.58
(35.1)
(29.0)
-57.7†
(32.98)
(63.4)
-6.3
(85.2)
(20.8)
-84.98
(62.3)
(25.3)
78.6
(54.4)
(28.2)
-128.5**
(37.3)
(33.4)
-9.9
(42.5)
(65.7)
-107.3
(80.2)
(27.3)
11.3
(34.7)

11.4
-34.3
-26.8
231.1*

(64.97)
(104.1)

12.1
-103.3

(92.96)
(135.2)

-70.4
12.4

(58.6)
(80.7)

-56.1
181.6

(89.8)
(282.9)

8.1
-393.4

(48.8)
(279.4)

14.1
192.3

93

(34)
(92.7)

(148.4)
(163.5)

TABLE AP2 (cont’d)
CEO options
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Pay gap
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Institutional ownership
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0
Blockholders
x historical attainment discrepancy < 0
x historical attainment discrepancy > 0

-17.3
-635.1***
-5.5
198.7

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

94

(183.3)
(228.4)

103
-63.1
321
390
0.000
0.52
0.16
0.33

(111)
(152.3)

-129.9
205.97

N
Wald χ2
Prob > χ2
Variance within divisions
Variance between divisions
Variance between firms

(75.6)
(156.95)

(103.8)
(96.3)

321
475
0.000
0.57
0.06
0.38

TABLE AP3: Relationship between performance discrepancy and risk taking based on social
aspirations (Compustat sample)

Variables
Constant
Sales growth
Divisional size
Risk taking t-2
Social attainment discrepancy <0
Social attainment discrepancy > 0
Relative divisional size
Divisional managers’ shares
Divisional managers’ options
Diversification
Potential slack
Free cash flow
CEO shares
CEO options
Pay gap
Institutional ownership
Bockholders

Model 1
Model 2
Coeff
SE
Coeff
SE
107.6
(105.7) 160.1
110.5
4.4 (14.4)
7.4 (16.5)
233.7*** (48.1)
217.9*** (57.5)
170.0*** (17.3)
171.7*** (19.5)
14.3 (21.1)
-14.2 1(37.6)
27.8 (29.7)
174.9 1(09.6)
-8.1 (28.0)
-16.2 (31.5)
52.4** (16.0)
52.0* (24.6)
12.5 (26.0)
-26.6 (34.9)
-55.1† (28.8)
-21.9 (37.8)
8.9 (65.7)
-36.0 (75.5)
-9.6 (22.9)
-37.8 (32.7)
56.5* (25.9)
-28.8 (57.7)
-52.6† (29.7)
-66.4† (39.1)
-32.3 (35.0)
-38.6 (48.9)
-80.4 (67.6)
-124.5 (78.2)
-0.5 (29.4)
8.8 (38.1)

Relative divisional size X
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Divisional managers’ shares
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Divisional managers’ options
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Diversification
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Potential slack X
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Free cash flow X
x social attainment discrepancy < 0
x social attainment discrepancy > 0
CEO shares
x social attainment discrepancy < 0
x social attainment discrepancy > 0

-30.2
118.7**

(51.8)
(41.5)

74.5
-101.8*

(85.8)
(46.1)

-162.2†
-76.5

(90.7)
(69.6)

182.5†
110.7*

(99.6)
(50.3)

222.8 (139.0)
198.9† (108.6)
0.2 (79.8)
-103.6 1(14.4)
-352.0† (199.5)
-47.8 (85.6)

95

TABLE AP3 (cont’d)
CEO options
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Pay gap
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Institutional ownership
x social attainment discrepancy < 0
x social attainment discrepancy > 0
Blockholders
x social attainment discrepancy < 0
x social attainment discrepancy > 0

-92.0 (137.2)
-173.5** (67.7)
-24.5 (159.8)
64.4 (88.7)
9.5 (183.2)
-192.5† (108.5)
-12.5 (107.3)
-11.3 (46.6)

N
Wald χ2
Prob > χ2
Variance within divisions
Variance between divisions
Variance between firms

296
380
0.000
0.52
0.21
0.27

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

96

296
370
0.000
0.43
0.22
0.35

TABLE AP4: Relationship between performance discrepancy and risk taking based on historical aspirations (Banking sample)
Model 1 –
Bank Risk
Variables
Constant
Sales growth
Divisional size
Divisional age
Risk taking t-2
PBH
PAH
Relative div. size
Slack
Diversification
CEO shares
CEO options
Pay gap
Institutional ownership
Blockholders
PBH * relative size
PAH * relative size
PBH * diversification
PAH * diversification
PBH * slack
PAH * slack
PBH * CEO shares
PAH * CEO shares
PBH * CEO options

Coeff
76.1***
-0.62*
-0.66
0.12
12.04***
0.74*
0.58†
0.47
0.70*
0.73
1.14*
-1.06**
0.34
0.68
-1.04**

Model 2 –
Bank Risk

Model 3 –
Loan Risk

Model 4 –
Loan Risk

Model 5 –
Liquidity risk

Model 6 –
Liquidity risk

SE
Coeff
SE
Coeff
(1.2) 76.1*** (1.2) 0.68**
(0.28)
-0.57* (0.28)
-0.01
(0.63)
-0.84 (0.64)
0.15†
(0.41)
0.17 (0.41)
-0.13*
(0.36) 12.18*** (0.37) 0.57***
(0.32)
0.65† (0.39)
0.06
(0.30)
0.62 (0.39)
0.08†
(0.57)
0.61 (0.58)
0.06
(0.30)
0.61† (0.31)
-0.08†

SE
Coeff
SE
Coeff
(0.15) 0.75** (0.16) 94.6***
(0.04)
-0.02 (0.04)
0.01
(0.09)
0.16† (0.09)
-0.18
(0.06)
-0.14* (0.06)
-0.19
(0.04) 0.55*** (0.04) 8.19***
(0.05)
0.06 (0.05)
0.21
(0.04)
0.07 (0.05)
0.05
(0.08)
0.04 (0.08)
-0.24
(0.04)
-0.03 (0.04)
-0.10

SE
Coeff
SE
(0.7) 94.6*** (0.71
(0.21
-0.03 (0.21)
(0.36
-0.29 0.36)
(0.24
-0.17 (0.24)
(0.21 8.11*** (0.21)
(0.24
0.17 (0.29)
(0.22
0.04 (0.29)
(0.33
-0.08 (0.34)
(0.22
-0.08 (0.23)

(0.63)
(0.46)
(0.37)
(0.51)
(0.62)
(0.39)

(0.08)
(0.06)
(0.05)
(0.07)
(0.08)
(0.05)

(0.32
(0.31
(0.27
(0.30
(0.32
(0.28

0.67
1.30**
-1.30**
0.28
0.95
-1.22**

(0.63)
(0.47)
(0.41)
(0.52)
0.63)
(0.40)

0.34
0.72†
0.38
-0.10
0.13
0.47
-1.24**
-0.52*
0.73*

0.10
0.05
0.04
-0.15*
0.14†
0.12*

0.09
0.03
0.01
-0.10
0.16†
0.15**

(0.09)
(0.06)
(0.06)
(0.07)
(0.09)
(0.05)

(0.39)
(0.42)
(0.33)
(0.34)
(0.22)

0.05
-0.03
0.02
0.01
0.03

(0.34)
(0.45)
(0.21)
(0.33)

-0.01
-0.11†
-0.01
0.13**
97

-0.15
0.43
-0.34
-0.33
-0.11
-0.001

-0.22
0.38
-0.35
-0.33
-0.01
-0.02

(0.32)
(0.32
(0.30)
(0.31)
(0.32)
(0.28)

(0.05)
(0.06)
(0.05)
(0.05)
(0.03)

0.23
0.70*
0.002
-0.04
0.07

(0.29)
(0.31)
(0.25)
(0.25)
(0.16)

(0.05)
(0.06)
(0.03)
(0.04)

0.46†
-0.80*
0.11
0.38

(0.25)
(0.32)
(0.16)
(0.25)

TABLE AP4 (cont’d)

PAH * CEO options
PBH * Pay gap
PAH * Pay gap
PBH * Inst. ownership
PAH * Inst. ownership
PBH * Blockholders
PAH * Blockholders
N
Wald χ2
Prob > χ2
Within division
variance
Between division
variance
Between firm variance

0.38
0.37
-0.09
0.46
0.90*
-0.25
0.30

(0.41)
(0.31)
(0.38)
(0.41)
(0.40)
(0.33)
(0.33)

0.02
-0.24***
-0.15**
0.05
-0.01
-0.05
-0.09†

(0.06)
(0.04)
(0.05)
(0.06)
(0.05)
(0.05)
(0.04)

-0.39
0.01
-0.05
0.32
-0.23
0.47†
0.61*

(0.30)
(0.23)
(0.28)
(0.30)
(0.29)
(0.25)
(0.25)

959
1359
0.000
0.8

959
1378
0.000
0.79

950
767
0.000
0.8

950
837
0.000
0.75

959
1753
0.000
0.98

959
1820
0.000
0.98

0.1

0.11

0.14

0.17

0.00

0.00

0.09

0.1

0.06

0.08

0.02

0.02

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

98

TABLE AP5: Relationship between performance discrepancy and risk taking based on social aspirations (Banking sample)
Model 1 –
Bank Risk
Variables
Constant
Sales growth
Divisional size
Divisional age
Risk taking t-2
PBS
PAS
Relative div. size
Slack
Diversification
CEO shares
CEO options
Pay gap
Institutional
ownership
Blockholders
PBS * relative size
PAS * relative size
PBS * diversification
PAS * diversification
PBS * slack
PAS * slack
PBS * CEO shares
PAS * CEO shares
PBS * CEO options

Coeff
75.2***
-0.82**
-0.70
0.12
9.54***
-0.21
-0.34
0.71
0.67*
1.34†
1.08*
-0.82*
0.37

Model 2 –
Bank Risk

SE
Coeff
(1.4) 75.1***
(0.28) -0.74**
(0.84)
-0.96
(0.58)
0.15
(0.46) 9.67***
(0.34)
-0.71
(0.38)
-0.52
(0.72)
0.81
(0.30)
0.31
(0.70)
(0.45)
(0.35)
(0.55)

0.84 (0.70)
-0.99* (0.39)

1.30†
0.96†
-0.96*
0.37

Model 4 –
Loan Risk

Model 5 –
Liquidity risk

Model 6 –
Liquidity risk

SE
Coeff
(1.4) 0.73**
(0.28
-0.02
(0.85
0.14
(0.58
-0.13*
(0.47 0.57***
(0.47)
0.01
(0.42)
0.03
(0.73)
0.05
(0.31)
-0.05

SE
Coeff
(0.16) 0.71***
(0.04)
-0.04
(0.10)
0.14
(0.06)
-0.12†
(0.05) 0.58***
(0.05)
0.03
(0.05)
0.03
(0.08)
0.06
(0.04)
-0.06

SE
Coeff
SE
Coeff
(0.16) 94.3*** (0.69) 94.3***
(0.04)
0.04 (0.21)
0.07
(0.10)
-0.19 (0.36)
-0.16
(0.06)
-0.15 (0.25)
-0.18
(0.05) 8.20*** (0.21) 8.11***
(0.07)
-0.12 (0.25)
-0.29
(0.06)
-0.38 (0.26)
-0.52†
(0.08)
-0.26 (0.33)
-0.17
(0.05)
-0.06 (0.22)
-0.10

SE
(0.69)
(0.22)
(0.36)
(0.25)
(0.22)
(0.35)
(0.30)
(0.34)
(0.24)

(0.72)
(0.51)
(0.38)
(0.55)

(0.09)
(0.06)
(0.05)
(0.07)

(0.09)
(0.07)
(0.05)
(0.07)

-0.07
0.35
-0.27
-0.31

(0.31)
(0.30)
(0.27)
(0.30)

-0.02
0.07
-0.22
-0.20

(0.30)
(0.34)
(0.29)
(0.30)

0.08 (0.09)
0.19** (0.06)

-0.08
-0.04

(0.30)
(0.29)

0.06 (0.30)
-0.05 (0.30)

1.30† (0.71)
-1.22** (0.40)
-0.27
0.38
-0.72
-0.14
0.09
1.20***
-0.09
0.00
-0.21

Model 3 –
Loan Risk

0.09
0.05
0.03
-0.10

0.13 (0.08)
0.16** (0.06)

(0.39)
(0.40)
(0.46)
(0.39)
(0.21)
(0.25)
(0.30)
(0.28)
(0.38)

0.12
0.02
0.05
-0.10

0.08
-0.02
0.16*
0.03
-0.03
0.03
0.02
0.04
0.03
99

(0.06)
(0.06)
(0.07)
(0.06)
(0.05)
(0.04)
(0.04)
(0.04)
(0.06)

0.27
0.64*
-0.19
0.29
-0.02
0.17
0.41†
0.31
-0.36

(0.30)
(0.29)
(0.35)
(0.28)
(0.17)
(0.20)
(0.22)
(0.21)
(0.29)

TABLE AP5 (cont’d)

PAS * CEO options
PBS * Pay gap
PAS * Pay gap
PBS * Inst. ownership
PAS * Inst. ownership
PBS * Blockholders
PAS * Blockholders
N
Wald χ2
Prob > χ2
Within division
variance
Between division
variance
Between firm variance

-0.33
-0.34
-0.80†
0.74
0.69
-0.78
-0.27

(0.41)
(0.30)
(0.47)
(0.46)
(0.46)
(0.49)
(0.33)

-0.04
0.06
-0.03
-0.17*
-0.06
0.13†
-0.06

(0.06)
(0.06)
(0.07)
(0.07)
(0.07)
(0.07)
(0.05)

-0.50
-0.25
0.26
0.43
0.27
0.01
-0.15

(0.32)
(0.23)
(0.35)
(0.35)
(0.34)
(0.37)
(0.26)

872
644
0.000
0.54

872
683
0.000
0.54

863
590
0.000
0.77

863
619
0.000
0.77

872
1782
0.000
0.99

872
1824
0.000
0.99

0.33

0.33

0.16

0.16

0.00

0.00

0.13

0.14

0.07

0.07

0.01

0.01

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

100

FIGURE B5: Interaction between performance discrepancy and institutional ownership (Banking
sample and using loan risk)

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

Risk
taking

_

low

Performance
Performance discrepancy
low inst. ownership
high inst. ownership

101

0
0

high/low +1/-1 SD

CHAPTER TWO: OUTCOMES OF DIVISIONAL RISK TAKING – HOW THE
STRUCTURAL EMBEDDEDNESS OF DIVISIONS IN THE M-FIRM AFFECTS THE
QUALITY OF THEIR RISK TAKING

INTRODUCTION
The relationship between organizational risk taking and subsequent performance plays an
important role in the field of strategic management. The overall conclusion is that risk has
important implications for organizational outcomes (e.g. Bowman, 1980, 1982, 1984; Bromiley,
1991; Fiegenbaum and Thomas, 1985; 1986; 1988; Miller and Bromiley, 1990; Miller and
Leiblein, 1996; Wiseman and Bromiley, 1996) and thus represents one of the central decision
making processes for organizations. Despite its long history, research on that relationship is
inconclusive, with many questions still remaining unanswered.
On one hand, majority of research reports a negative association between risk and
performance where managers undertake risky choices which are detrimental to performance and
do not result in higher returns (Bowman, 1980, 1982; Bromiley, 1991; Fiegenbaum and Thomas,
1988). Alternatively, continuous engagement in risky initiatives leads to a downward spiral in
organizational performance (Wiseman and Bromiley, 1996). On the other hand, there are studies
finding a positive link between risky choices and subsequent performance. For example, Miller
and Leiblein (1996) offer evidence that managers will engage in and persists with risky efforts
until they encounter solutions with high-enough returns.
Overall, these findings are quite puzzling and necessitate the question: Why do some
firms undertake “good”, value-enhancing risky choices (e.g. Miller and Leiblein, 1996) and other
7

organizations conduct “bad” and detrimental risky choices? In other words, it is critical to

7

I define “good” risks as risks leading to improvements in performance; on the contrary, “bad”
risks are associated with performance declines (e.g. Wiseman and Bromiley, 1996).
102

understand how some managers are able to recognize and implement organizationally-beneficial
risky strategies and other managers are not.
Another question that needs an answer is: For managers who “fall into a trap of taking
unprofitable risks” (Wiseman and Bromiley, 1996: 524), what is their motivation to continue
with risky choices and can they learn to avoid such unsuccessful risky efforts?
The purpose of this chapter is to answer these questions and offer more insights about
when and why risk taking might be positively or negatively related to subsequent performance.
To achieve this objective, I argue that it is necessary to investigate unexplored contingencies and
varying moderating conditions, and more specifically corporate and industry factors (Deephouse
and Wiseman, 2000; Wiseman and Bromiley, 1991), that might reconcile previously
inconclusive results. Although prior research has examined some potential moderators (see
Bromiley, 1991; Fiegenbaum and Thomas, 1988), those studies have not emphasized differences
across types of firms – e.g. they have viewed firms as stand-alone entities that function
independently. In contrast, in this chapter I distinguish between divisions across M-firms and
single-business firms and I argue that important structural contingencies will affect the
relationship between risk taking and performance. That is, since divisional managers’ decision
making is contextually situated (Granovetter, 1985, 1992), I examine how “a cascading hierarchy
of influences” (Sutcliffe and McNamara, 2001: 486) will moderate the relationship between their
risk taking and performance. In particular, I present how environmental-level (Goll and Rasheed,
1997; Henderson et al., 2006; Palmer and Wiseman, 1999), corporate-level (Sanders and
Hambrick, 2007; Mikkelson and Ruback, 1991; Palmer and Wiseman, 1999), and divisionallevel variables (Audia and Greve, 2006) impact the degree to which risk taking is beneficial or
detrimental to subsequent performance.

103

Furthermore, I posit that it is necessary to take a deeper look at the facilitating and/or
inhibiting mechanisms that distinguish between “good” and “bad” risk taking. More specifically,
the existence of structural and cognitive contingencies in the M-firm is expected to have strong
impact on the ability and motivation of divisional managers to conduct effective risk taking
which positively impacts performance at the divisional level. On one hand, low-powered
incentives of divisional managers (Williamson, 1996) may discourage the latter to diligently
distinguish between “good” and “bad” risky choices which might translate into suboptimal risk
taking. In addition, the informational disadvantage of the corporate office vis-à-vis divisions
(Williamson, 1985) reduces its ability to properly monitor the effectiveness and quality of the
risky strategies undertaken by divisions (e.g. Kim et al., 2004). On the other hand, a focal
division could gain access to richer information and solutions provided by other corporate
divisions, possibly facilitating the quality of its risk taking (e.g. Hansen, 1999; 2002; Beckman
and Haunschild, 2002).
Overall, this chapter aims to develop a theory of risk taking and performance at the
divisional level. By looking at divisions I elaborate on the role of important structural and
cognitive constraints in the M-firm that impact the quality of divisional risk taking and
subsequently divisional performance. Moreover, utilizing the context of multidivisional firms
allows me to examine the impact of unexplored moderating conditions from multiple levels of
analysis, such as divisional, corporate and environmental. Thus, I am able to build a cross-level
model of the impact of divisional risk taking on divisional performance. To develop my
theoretical arguments I integrate BTOF (Cyert and March, 1963), agency theory (Jensen and
Meckling, 1976; Eisenhardt, 1989), theories of environmental dynamism, munificence and
complexity (Dess and Beard, 1984; Duncan, 1972; Palmer and Wiseman, 1999), and

104

organizational learning theories (Baum and Ingram, 1998; Beckman and Haunschild, 2002;
Chuang and Baum, 2003; Greve, 1998; Ingram and Baum, 1997a, 1997b; Madsen and Desai,
2010).
This chapter provides several contributions to existing research. First, I develop
arguments why some organizations are able to engage in beneficial risk taking and others
undertake value-destroying risky initiatives. By using the context of corporate divisions, I argue
that a key factor for successful risk taking is the ability to learn from related organizations and
utilize this knowledge in one’s own risky efforts.
Second, I bridge theories of M-firm structure (Hill and Hoskisson, 1987; Hoskisson, Hitt,
and Hill, 1991, 1993; Hoskisson and Turk, 1990) and organizational learning theories (Baum and
Ingram, 1998; Beckman and Haunschild, 2002; Chuang and Baum, 2003; Greve, 1998) to
develop a model outlining their critical role on organizational performance. More specifically, I
theorize about the impact of corporate diversification (as a central tenet of organizational
structure) on divisional vicarious learning which affects the ability of divisional managers to
undertake high-quality risky strategies.
Third, I answer calls for investigating additional contingent factors that might affect the
performance consequences of risk taking (Deephouse and Wiseman, 2000). To achieve this, I
extend prior research on the risk-return relationship by utilizing a comprehensive multilevel
approach. More specifically, I develop theoretical arguments for the moderating role of factors at
the divisional (divisional size and divisional managers’ compensation), corporate (corporate
diversification, executive compensation, and external monitoring), and environmental (industry
munificence, complexity and dynamism) levels of analysis and outline their facilitating or
inhibiting impact on the risk taking-performance relationship.

105

Finally, I employ a different methodological approach in testing the relationship between
risk and performance. On one hand, prior research following the BTOF has directly related
organizational risk to subsequent performance (e.g. Miller and Leiblein, 1996). However, by
using such an approach (see Fiegenbaum and Thomas, 1988; Miller and Leiblein, 1996;
8

Wiseman and Bromiley, 1996) , scholars have failed to pay proper attention to the managerial
risky actions (e.g. risk taking) which actually affect organizational performance. Thus, in order
to better understand the underlying mechanisms of organizational performance I follow the
recommendations of the holistic model of risk (Palmer and Wiseman, 1999) and model the
impact of risk taking on performance. On the other hand, using the firm (and diversified firms in
particular) as the unit of analysis has led to studying the risk-return relationship via an
9

aggregation approach which masks the specificities of that relationship . On the contrary, I
utilize divisions as a lower unit of analysis and thus I am able to examine more accurately the
impact of risk taking on performance.

HYPOTHESES DEVELOPMENT
Do divisions take on “good” or “bad” risks? – The moderating role of M-firm structure
Existing research investigating the impact of risk on firm performance has implicitly
assumed that managerial risky actions are relatively independent of other firms’ strategies and
initiatives. In other words, managers have the ability, authority and motivation to decide on and

8

Initially, the relationship between risk and return has been analyzed through a mean-variance
approach where risk has been operationalized as the variance of firm returns/performance.
Subsequently, scholars have begun to utilize various measures of risk, such as managerial
perceptions (Singh, 1988), earnings forecasts (Bromiley, 1991), credit risk (McNamara and
Bromiley, 1999; Wiseman and Catanach, 1997).
9
McNamara and Bromiley (1999) is one of the few studies to test the risk-return relationship
using the lending decision as a unit of analysis.
106

undertake risky choices. However, while those assumptions might be relevant for stand-alone
firms, they are relatively less applicable to divisional managers who face more structural
constraints in their risk taking strategies. More specifically, the existence of salient
interdivisional comparisons, the presence of rigid decision rules (e.g. Merton, 1968), the
hierarchical dependence, partial detachment from market pressures and low-powered incentives
(e.g. Kraatz and Zajac, 2001; Williamson, 1985) are likely to affect divisional managers’ ability
and diligence in carefully assessing the expected outcomes of risky choices. Thus, the
relationship between risk taking and performance for corporate divisions is intertwined within a
complex system of organizationally binding norms which suggests a significant deviation from
the traditional risk-return association at single-unit firms. The existence of those contingencies is
expected to affect the ability and motivation of divisional managers to identify and implement
appropriate risk taking and as a result might impact the quality of divisional risk taking which
ultimately reflects into their performance.
First, I briefly discuss why divisions might be able to undertake “good” risk taking.
Drawing on research in the organizational learning tradition (e.g. Baum and Ingram, 1998;
Beckman and Haunschild, 2002; Ingram and Baum, 1997a, 1997b; Darr, Argote, and Epple,
1995; Hansen, 1999), I argue that the existence of other divisions in the M-firm could have a
beneficial effect on a focal division’s association between risk taking and performance. The
underlying argument is that a focal division’s managers could utilize the experience of other
divisional managers, learn from them and apply this knowledge successfully in their own
division (e.g. Beckman and Haunschild, 2002). Observing other divisions, and more specifically
the outcomes of their risk taking initiatives, provides a focal division with relevant information
and clues about the causes for success or failure of other divisions’ risky choices (Baum and

107

Dahlin, 2007; Chuang and Baum, 2003). As a result, this knowledge is incorporated in the focal
division’s risky initiatives and is likely to impact positively subsequent divisional performance.
Thus, the availability of rich information inside the M-firm and the ability of divisions to learn
from the experience of other divisions allow them to make higher-quality decisions (Beckman
and Haunschild, 2002) – that is, they are more likely to undertake “good” risk taking which
translates into improved divisional performance.
Second, while M-firms provide conditions for “good” risk taking, there are also structural
contingencies that undermine the quality of divisional risk taking and might lead to “bad” risk
taking. Divisional managers are subject to low-powered incentives (Williamson, 1985, 1996) and
their motivation to serve in the corporation’s best interests is significantly reduced. As a result,
divisional managers might be less incentivized to search for optimal projects and instead might
engage in inferior risky strategies (e.g. “bad” risk taking). In a similar vein, divisional managers
might develop sense of security and complacency that the corporate office will compensate for
their mistakes (Kraatz and Zajac, 2001) and thus show lower urgency and less diligence in
searching for and implementing higher-quality risky choices. In other words, divisional
managers might become more tolerant and insensitive to inefficient risk taking. Finally, the
informational disadvantage of corporate officers vis-à-vis divisional managers (e.g. Hitt et al.,
1990) reduces the formers’ ability to control divisions (Hill and Hoskisson, 1987) and intervene
when divisional risk taking is suboptimal and detrimental for divisional performance.
The above arguments suggest that corporate structure is unlikely to have a uniform effect
on the quality of divisional risk taking. Thus, in order to better understand its role, a contingent
perspective is needed. I argue that level of diversification as an important element of corporate
structure is a critical contingency that provides an explanation for when and why divisions

108

engage in “good” and/or “bad” risk taking. More specifically, different levels of diversification
are expected to make different structural contingencies more salient to divisional managers (e.g.
Ocasio, 1997) which affects the risky choices of the latter. In particular, low to moderate
diversification (e.g. when corporate divisions are related) allows a focal division to observe
similar other divisions and thus to learn effectively from them (Baum and Ingram, 1998; Baum,
Li, and Usher, 2000; Greve, 1998). In other words, the existence of large number of similar
divisions is associated with more relevant learning and the more likely it is that a focal division
will successfully apply that knowledge in its risk taking efforts. On the contrary, when
diversification is high (e.g. divisions are unrelated and too diverse – Robins and Wiersema,
2003), the chances for successful learning diminish due to the dissimilarity and incomparability
of divisions. That is, applying unrelated divisions’ experience with risky choices to a focal
division is unlikely to improve performance because of the inapplicability of this experience to
different contexts (Baum and Ingram, 1998; Darr et al., 1995). Furthermore, as diversification
increases, it becomes harder for corporate executives to effectively monitor and control corporate
divisions (Hill and Hoskisson, 1987) and thus the impact of low-powered divisional incentives
on risk taking quality exacerbates.
Below I describe in greater details the effect of diversification on divisional risk taking
quality and subsequent performance.
Related diversification and “good” risk taking
One of the most important features of M-firms is the rich internal information they
provide to divisions and the opportunity for the latter to utilize that knowledge in their strategic
initiatives. This argument is particularly valid in moderately diversified M-firms because related
divisions could share similar information which is relevant for and fits within their strategic

109

contexts. That is, since related divisions operate in close markets, pursue similar strategies and
implement comparable investment initiatives (e.g. Miner, Kim, Holzinger, and Haunschild,
1999), their managers could observe each other, transfer operating expertise and knowledge
(Parmigiani and Holloway, 2011) and learn from the actions other divisions have undertaken
(e.g. Baum and Ingram, 1998; Ingram and Baum, 1997a, 1997b; Davis and Greve, 1997; Greve,
1998; Haunschild and Beckman, 1998).
Translating this logic to the relationship between divisional risk taking and performance,
I argue that a focal division could utilize other related divisions’ experience with risky initiatives
and enhance the quality of its own risk taking leading to increased performance. On one hand, a
focal division which observes the successful risky strategies of other related divisions is likely to
internalize this knowledge and incorporate it in its own risky initiatives (Baum and Dahlin, 2007;
Ingram and Baum, 1997a, 1997b). That is, vicarious learning from others’ success enhances the
ability of a focal division’s managers to engage in value-enhancing risky efforts. On the other
hand, a focal division could also learn from the mistakes of other related divisions and thus
utilize this information in its own strategies. Others’ failures provide salient and useful
knowledge about the potential causes of unsuccessful risky initiatives and help a focal division to
avoid repeating the same mistakes. In other words, finding out that a particular strategy does not
work in a similar environment motivates the managers of the focal division to rely on different
actions (Chuang and Baum, 2003; Madsen and Desai, 2010). The overall conclusion is that
“observing others’ successes may increase decision makers’ confidence in the accuracy of
knowledge held by their own organization and… observing the others’ failures may lead
decision makers to question their own knowledge” (Madsen and Desai, 2010: 455) and all this
experience could be successfully implemented in the subsequent strategies of the focal division.

110

Furthermore, the knowledge obtained from divisions’ risky initiatives could be
embedded in the M-firm’s routines and stored as a template for future use by other related
divisions (e.g. Madsen and Desai, 2010; Parmigiani and Holloway, 2011). As a result, when a
focal division is considering a risky initiative (e.g. undertaking a major capital investment
project), it could easily access prior codified knowledge and use it appropriately when deciding
whether and how to proceed with its strategic initiatives.
While relatedness of other divisions is a critical element of successful vicarious learning,
the number of related divisions should be equally important. In other words, the bigger the
number of relevant knowledge sources, the more likely the focal division will learn from them.
Observing multiple related divisions and internalizing their experiences provides for a larger
pool of alternatives and allows the focal divisional manager to select the most appropriate of
them (e.g. Haleblian and Finkelstein, 1999). As a result, as the number of related divisions
increases, the focal division will be more likely to implement “good” risk taking.
H1: Number of related divisions will moderate the relationship between divisional risk
taking and divisional performance, such that the slope of the relationship will be more positive
as the number of related divisions increases.

Unrelated diversification and “bad” risk taking
Under conditions of unrelated diversification divisional managers are faced with the
inability to effectively learn from other divisions and thus to successfully utilize vicarious
learning in their risk taking strategies. The main reason is the dissimilarity among unrelated
divisions and the inapplicability of others’ experiences to a focal division (e.g. Baum et al., 2000;

111

Greve, 1998). Utilizing the experience of other unrelated organizations is not only useless (Darr
et al., 1995), but could even be harmful to the focal organization (e.g. Baum and Dahlin, 2007).
Nevertheless, a focal division could still try to apply and imitate others’ knowledge
because of: a) lack of sufficient own experience with risky initiatives (Baum and Dahlin, 2007;
Baum and Ingram, 2002); b) inappropriate generalization that others’ success will translate into
success for the focal division (Kim et al., 2009); and c) internal corporate pressure to follow the
actions of other divisional counterparts (e.g. Greve, 1996). More specifically, M-firms have rigid
rules and established procedures (Dougherty and Hardy, 1996) regarding where, when and how
much to invest in risky projects and usually those rules are based on divisions’ prior successful
experiences with risk taking. Because divisional managers perceive those rules as taken-forgranted and legitimate (Scott, 1995; Shimizu and Hitt, 2004), they are likely to follow them and
thus “to restrict their search, ignore discrepant information, or be automatic in their processing”
(Sutcliffe and McNamara, 2001: 497). In other words, managers could become myopic and lax
in their risk taking (Sutcliffe and McNamara, 2001) which would negatively affect the quality of
10

their risk taking and ultimately hurt performance . Overall, unrelated diversification increases
the likelihood of misfit between rigid organizational rules for investing in risky projects and
divisional needs which reduces the usefulness and quality of those risky initiatives.
Furthermore, unrelated diversification is likely to reduce the quality of divisional risk
taking in an additional way – due to the inability of the corporate office to effectively counter the
low-powered incentives of divisional managers. While low-powered incentives exist across all
corporate divisions, the magnitude of their impact on managerial decisions is different across
related and unrelated divisions. Contrary to the former case where corporate executives are better
10

See Haleblian and Finkelstein (1999) for a discussion on the negative consequences of
inappropriate imitation.
112

equipped to control divisions (e.g. Hill and Hoskisson, 1987; Stein, 1997) and thus to alleviate
problems associated with low-powered incentives, it is extremely difficult for the corporate
office to properly monitor multiple unrelated divisions and correct their inappropriate risky
initiatives (e.g. the so-called depth-for-breath tradeoff by Williamson, 1985). Because lowpowered incentives reduce their motivation to serve in the corporation’s best interests
(Williamson, 1985), divisional managers might have lower incentives to search and implement
optimal projects (e.g. projects that entail “good” risk and promise relatively certain and high
performance outcomes). Alternatively, the “misalignment of incentives between central and
divisional managers” (Berger and Ofek, 1995: 40) could lead to implementation of inferior risky
strategies at the divisional level.
Overall, because unrelated diversification generates conditions for unsuccessful vicarious
learning and inappropriate imitation, and exacerbates the negative effects of low-powered
incentives, divisions are more likely to engage in “bad” risk taking which ultimately translates in
suboptimal performance. Formally stated:
H2: Level of diversification will moderate the relationship between divisional risk taking and
divisional performance, such that the slope of the relationship will be negative (or less positive)
for high levels of diversification.

The moderating impacts of industry, corporate and divisional characteristics on the
relationship between divisional risk taking and performance
While I argued that corporate diversification plays a focal role on the quality of divisional
risk taking, there are additional factors that are expected to influence the ability of divisions to
engage in ‘good” or bad “risk” taking. In particular, prior research offers strong evidence that

113

managerial decision making is situated in a broad social context and managerial actions are
driven by a complex set of hierarchical influences (Audia and Greve, 2006; Granovetter, 1985,
1992; McNamara and Bromiley, 1997; Sutcliffe and McNamara, 2001). Because divisions in the
M-firm are structurally embedded and their decisions are hierarchically dependent on the
corporate office, it is logical to expect that higher-order factors (e.g. corporate level) will
influence the divisional decision making processes. Furthermore, theories examining the link
between organizations and the environment provide evidence that the latter exhibits strong
impact on decision making (Goll and Rasheed, 1997; Palmer and Wiseman, 1999). Thus, I argue
that organizational decision making (and divisional risk taking in particular) is likely to vary
across “multilevel relational contexts” (Langley 1989; O’Reilly and Chatman, 1996).
In a similar vein, organizational performance has been shown to be driven by the
interaction of multilevel factors (Bowman and Helfat, 2001; Brush, Bromiley, and Hendrickx,
1999; McGahan and Porter, 1999; Misangyi et al., 2006; Rumelt, 1991). More specifically,
Misangyi and colleagues (2006) argue that “the industries and corporate parents in which
business segments are embedded serve as environments which affect business unit profitability”
(p. 581).
Based on the above arguments, it is logical to investigate how the relationship between
divisional risk taking and subsequent performance is influenced by factors at the divisional,
corporate, and environmental level of analysis. The selection of the specific factors is
theoretically driven and based on prior research outlining their impact on managerial behaviors
and decision making.
First, to analyze the industry impact I utilize Dess and Beard’s (1984) taxonomy of
environmental categories (industry munificence, dynamism, and complexity). In addition, I draw

114

on a broad stream of research investigating the role of the environment on various outcomes of
the managerial decision-making process (see Goll and Rasheed, 1997; Henderson et al., 2006;
Hough and White, 1993; Li and Simerly, 1998; Simerly and Li, 2000; Wiersema and Bantel,
1993). Second, to discuss the influence of division and corporate factors I utilize the broad
strategic management literature which outlines several organizational attributes that affect
decision makers’ actions. More specifically, I discuss the moderating role of structural and
resource endowment characteristics (Barney, 1991; Chandler, 1962; Porter, 1980) on the
relationship between risk taking and performance at the divisional level.

I. Industry factors affecting the relationship between divisional risk taking and performance
As already discussed, corporate divisions could utilize prior experience in their decision
making processes with the hope that previously successful decisions could be effectively
translated into current contexts. However, such an assumption is valid only when environmental
conditions are relatively constant. In other words, “when environments are rapidly changing,
experience-based learning may lock firms into previously successful strategies, lowering
subsequent performance” (Desai, 2008: 596; Audia et al., 2000; Baum and Ingram, 1998) and
negatively influencing the effectiveness of risky initiatives. Thus, whether divisional risk taking
could be qualified as “good” or “bad” is partially dependent on the environment in which a
division operates.
Industry munificence. Munificence is characterized by “the scarcity or abundance of
critical resources” (Castrogiovanni, 1991: 542) and the ability of the environment to support
growth (Aldrich, 1979; Dess and Beard, 1984; Starbuck, 1976). Highly munificent environments
provide managers with discretion and more opportunities to undertake various experimentations

115

and different risky projects (e.g. Hambrick and Finkelstein, 1987). Alternatively, highly
munificent environments are characterized as attractive and providing enabling conditions for
managers to execute successful strategies. Undertaking risky projects, such as production
expansion or entering new markets, in growth-supporting environments increases the chances of
those projects being successful and generating positive returns.
In contrast, lack of or low environmental munificence leads to difficult and stressful
conditions for managers (Wiersema and Bantel, 1993) and limits decision makers’ ability to
implement strategies that fit with the demands of the environment. Because scarcity leads to
“decreased information processing, … more rigid problem-solving and adherence to traditional
routines” (Wiersema and Bantel, 1993: 487) and reliance on outdated information, strategic
decisions are more likely to be inefficient. As a result, divisional managers are less likely to
diligently evaluate different risky alternatives which ultimately increases the likelihood of
implementing strategies that do not fit with the environmental requirements. Simply stated, low
munificence does not provide a buffer against environmental hostilities (Parks and Conlon,
1995) and further exacerbates suboptimal strategies. Overall, I propose that industry munificence
will affect the quality of divisional risk taking and thus moderate its relationship with subsequent
performance. Formally stated:
H3: Industry munificence will positively moderate the relationship between divisional risk taking
and divisional performance.

Industry dynamism. Dynamism is associated with high degree of uncertainty and
unpredictability in the environment which present many unclear situations to organizational
decision makers (Duncan, 1972; Li and Simerly, 1997). Highly dynamic environments are

116

characterized by instability and consistent changes that place “tremendous cognitive demands”
(Wiersema and Bantel, 1993: 488) on decision makers and limit their ability to conduct
appropriate evaluation of the external conditions. In other words, very dynamic conditions
reduce both the access to knowledge necessary for making strategic decisions (Milliken, 1987;
Simerly and Li, 2000) and the usefulness of that information and might force “managers to
perform limited search in their assessment of the environmental situation” (Li and Simerly, 1997:
171). Thus, there is an increased likelihood of misfit between managerial decisions and the
environment. All these will result in low-quality decision making which might negatively affect
performance outcomes.
Translating this logic to the quality of divisional risk taking suggests that high dynamism
will have a negative impact on the association between risk taking and performance at the
divisional level. Considering that risk taking is largely uncertain and its outcomes are unknown
to divisional managers ex ante, additional dynamism and uncertainty from the environment will
make it even harder for managers to accurately choose and implement appropriate risk taking
initiatives (e.g. McNamara et al., 2008). That is, in a highly unstable and unpredictable
environment the risk taking initiatives of divisional managers are less likely to fit with the
external conditions and more likely to lead to adverse effects on performance. Finally, as
discussed previously, divisional managers might rely heavily on existing corporate rules and
guidelines for undertaking risky investment projects; however, in highly dynamic environments
those rules are difficult to change. As a result, the latter become quickly obsolete (Henderson et
al., 2006) which further exacerbates the misfit between divisions’ risk taking strategies and
subsequent performance.

117

While highly dynamic environments are expected to have a negative impact on the
quality of divisional risk taking, very stable markets might not have a positive effect on the
relationship between risk taking and performance, either. Because such environments entail
minimum degree of change and are highly predictable (Palmer and Wiseman, 1999), it might be
difficult for managers to identify beneficial opportunities which have not been discovered in the
past. That is, since all available options have been identified and majority of managers have
access to them, there is no new information with unique value to be utilized in divisional risk
taking.
Thus, moderate degree of environmental dynamism offers the optimal conditions for
effective risk taking because: a) new opportunities and information emerge which are not
available to everyone, and b) managers could still recognize and take advantage of those
opportunities.
H4: Dynamism will moderate the relationship between divisional risk taking and divisional
performance, such that the slope of the relationship will be more positive at moderate levels of
dynamism.

Industry complexity. Complexity refers to the heterogeneity of the environment and the
multitude of factors that need to be attended to in strategic decision making (Child, 1972;
Duncan, 1972; Wiersema and Bantel, 1993). To cope with increased complexity, divisional
managers need to possess large information processing capabilities (Dess and Beard, 1984);
however, due to cognitive limitations and bounded rationality managers experience difficulties in
properly monitoring the whole range of environmental factors and events. This is echoed by
Wiersema and Bantel (1993: 489) who argue that “a wider range and greater quantity of

118

information need to be processed for effective decision making, creating strains on the
organization to achieve high quality decision making outcomes”. Hence, high environmental
complexity makes it more challenging for divisional managers to maintain necessary levels of
information processing resources which will affect the outcomes of their risk taking.
Complex environments are expected to exacerbate the ability of divisional managers to
undertake appropriate risk taking (e.g. “good” risk taking). In particular, divisional managers are
not able to properly evaluate all means-ends linkages in complex environments (e.g. Hambrick
and Finkelstein, 1987) and thus, their risk taking initiatives are less likely to fit with the
environment and more likely to result in poor performance.
However, non-complex environments might not be beneficial to divisional risk taking,
either. In particular, in markets with no complexity all available alternatives are known and
accessible by majority of managers which decreases their value and limits subsequent search
efforts in finding new beneficial opportunities. Thus, only at moderate levels of complexity
where managers are able to overcome cognitive limitations and take advantage of opportunities
not available to and understood by all firms, will divisional risk taking result in satisfactory
performance.
H5: Complexity will moderate the relationship between divisional risk taking and divisional
performance, such that the slope of the relationship will be more positive at moderate levels of
complexity.

II. Corporate governance factors affecting the relationship between risk taking and performance
The M-firm provides two distinct levels at which governance factors are expected to
affect the relationship between divisional risk taking and performance. In particular, CEO

119

incentive alignment and the presence of institutional investors and blockholders operate at the
corporate level, while divisional managers’ incentive alignment plays a key role at the divisional
level. This unique governance structure in M-firms is determined by the existence of double
agency where one set of hired agents (the corporate CEO) is responsible for overseeing another
set of agents at a lower hierarchical level (the divisional managers) (e.g. Child and Rodriguez,
2003). Thus, the M-firm and its shareholders are likely to incur two sources of agency costs –
one source is related to divisional managers undertaking “bad” risks leading to poor
performance; the other sources is associated with the corporate CEO who might protect his
personal interests rather than control the suboptimal investment decisions of his subordinate
divisional managers. However, when appropriate governance mechanisms are implemented at
both the corporate and divisional levels, shareholder interests are expected to be protected more
strongly. That is, ”good” risk taking is likely to be undertaken at the divisional level which leads
to increased performance.
Below I discuss first the separate moderating effects of corporate-level and divisionallevel governance mechanisms on the relationship between risk taking and performance.
CEO incentive alignment. A broad stream of research has investigated how managerial
incentives and ownership variables influence both firm risk taking and firm performance. For
example, evidence exists that executive option holdings are positively related to strategic risk
taking (Devers et al., 2008; Sanders and Hambrick, 2007) and strategic risky choices (Sanders,
2001). Furthermore, research shows that equity holdings by firm executives are related positively
to firm performance (Dalton et al., 2003). However, little research has been done on the
moderating effect of incentive alignment on the relationship between risk taking and
performance. In this chapter I argue that CEO contingent pay is an important factor that has a

120

positive impact on the quality of divisional risk taking. Because incentive alignment makes
CEOs better monitors of corporate value, it is more likely that they will intervene and discourage
ineffective divisional risk taking. That is, when CEOs have a portion of their income linked to
corporate performance they will be motivated to promote high-quality risky strategies (Wright,
Kroll, Lado, and van Ness, 2002).
On the contrary, low or lack of incentive alignment reduces the motivation of CEOs to
act as diligent monitors and instead encourages them to prioritize personal interests. For
example, opportunistic corporate CEOs might be more concerned for protecting own perks and
income and disregard inefficiencies across divisions (e.g. Scharfstein, 1998). Similarly, in order
to maintain political balance and friendship relationships with divisional managers, CEOs might
overlook suboptimal divisional decisions (e.g. Hill, Hitt, and Hoskisson, 1992), such as “bad”
risk taking that reduces subsequent performance.
Overall, a CEO’s contingent compensation increases his opportunity costs of not
maximizing firm wealth by allowing suboptimal divisional decisions. Alternatively, such
compensation makes CEOs more likely to intervene and correct inappropriate risky strategies by
divisional managers. This leads to the following hypothesis:
H6: CEO incentive pay will positively moderate the relationship between divisional risk taking
and divisional performance.

External monitoring. Institutional investors and large blockholders have been suggested
as an additional monitoring device to managerial actions (Hansen and Hill, 1991; Jensen and
Meckling, 1976; Schleifer and Vishny, 1986). Because of the consistent pressure they exercise
over firm managers to undertake only value-enhancing strategies, it is argued that they have a

121

positive impact on performance. This is in accordance with Pound’s (1988) efficient monitoring
hypothesis that large shareholders encourage value-enhancing decisions and oppose valuedecreasing ones. Using this logic within the context of divisional risk taking, I argue that the
presence of external monitors, such as institutional investors and blockholders, would reduce the
likelihood of “bad” risk taking by divisional managers. Since those external investors possess
sizeable portions of firm shares, they would be very sensitive to particular risky strategies by
divisions that result in suboptimal performance (e.g. Brickley et al., 1988). Institutional owners
and blockholders hold corporate CEOs directly accountable for organizational outcomes and this
constant monitoring keeps CEOs attentive to low-quality and inappropriate divisional risky
initiatives (e.g. Mikkelson and Ruback, 1991). In other words, in the presence of external
monitors CEOs are less likely to tolerate “bad” and value-decreasing divisional risk taking.
Furthermore, knowing that the CEO is under strong external monitoring and might not protect
their suboptimal risky strategies, divisional managers will exercise greater care in their selection
of risky choices. Thus, institutional ownership and blockholders will have a beneficial
moderating impact on the relationship between divisional risk taking and divisional performance.
H7a: Institutional ownership will positively moderate the relationship between divisional risk
taking and divisional performance.
H7b: Blockholders will positively moderate the relationship between divisional risk taking and
divisional performance.

Divisional managers’ contingent pay. The arguments for the role of CEO contingent pay
on the relationship between divisional risk taking and divisional performance are easily
applicable to divisional managers’ incentive alignment or what I call a “first-order” corporate

122

governance mechanism. Divisional managers are inherently risk-averse (Williamson, 1985) and
pursue personal objectives even at the detriment of the multidivisional firm. For example, they
could engage in risky initiatives which generate personal benefits at the detriment of divisional
performance (e.g. Hill et al., 1992). Hence, those divisional actions could be viewed as an
additional agency cost levied on the corporation and the wealth of its shareholders. An efficient
mechanism that could alleviate such costs, reduce the negative consequences of “bad” risk
taking, and encourage divisional managers to undertake appropriate and value-enhancing risky
strategies is contingent pay or linking their compensation to organizational outcomes (e.g. Dalton
et al., 2003; Hill et al., 1992). Seeing themselves as partial owners in the corporation, divisional
managers will have an increased motivation to carefully analyze risky choices and adopt only
those that are expected to benefit their divisions and subsequently the corporation. Alternatively,
any suboptimal decision would directly translate into reduced performance and decreased
divisional managers’ compensation. Thus, I expect contingent pay of divisional managers to
make them better monitors of corporate wealth and thus to have a positive effect on the
relationship between divisional risk taking and divisional performance.
H8: Divisional managers’ incentive pay will positively moderate the relationship between
divisional risk taking and divisional performance.

In addition to the level of incentive pay, the difference in pay (e.g. pay gap) between
divisional managers and the corporate CEO is also expected to affect the formers’ actions with
regard to the quality of risk taking. According to tournament theory wider pay gaps create
incentives for lower-level managers to be more diligent in their strategic choices and engage in
actions which protect the interests of shareholders (Henderson and Fredrickson, 2001; Lazear

123

and Rosen, 1981; Rosen, 1986). Within the context of divisional risk taking, tournament theory
arguments suggest that divisional managers will be more careful and vigilant in selecting
appropriate risky choices which have a positive effect on performance. In other words, under
conditions of high pay gaps divisional managers will have higher incentives to undertake “good”
risks which have a beneficial effect on subsequent performance.
H9: The gap between CEO pay and divisional managers’ pay will positively moderate the
relationship between divisional risk taking and divisional performance.

III. Divisional factors affecting the relationship between risk taking and performance
Divisional size. Divisional size is associated with two main characteristics: structural
constraints and bargaining power. Because of their size, bigger divisions are faced with more
bureaucracy, inertia, and reliance on established routines (Haleblian et al., 2012; Hannan and
Freeman, 1984; Mintzberg, 1979). As a result, they are more likely to utilize prior investment
strategies and less likely to carefully evaluate among alternative risky initiatives (e.g. Shimizu
and Hitt, 2004). In other words, a big division faces a lower probability of selecting “good” risk
taking that fits within the strategic context of the division.
Size also gives divisions a bargaining advantage over the corporate office – as the size of
a division increases, it becomes more important for the corporation and divisional managers gain
more influence and power vis-à-vis corporate executives (e.g. Kim et al., 2004). Thus, divisional
managers could leverage this power to bargain for and receive more resources. While a larger
resource base allows divisions to invest in projects and undertake various initiatives, it does not
automatically translate into successful projects and higher returns. Quite the opposite, because
large divisions could develop the perception that they are entitled to larger capital funds, they

124

might exhibit less care in how they invest those funds. Simply stated, large divisions could
engage in suboptimal risky initiatives knowing that this will not affect significantly the future
capital allocations they receive from the corporate office. Furthermore, large divisions have the
power to oppose the monitoring attempts of the corporate office regarding their suboptimal risky
choices. For example, Kim and colleagues show that within keiretsu organizations (which are
similar to M-firms) large member firms leverage their power to block attempts by corporate
authorities to emphasize efficiency and profitability (Kim et al., 2004).
On the contrary, smaller divisions are expected to exhibit a stronger relationship between
risk taking and performance. Since those divisions have less power and are relatively less
important within the M-firm, they receive smaller amount of corporate funds and try to utilize
them in the best way possible. Moreover, if small divisions make suboptimal investment
decisions which reduce their performance, the corporate office will monitor them closely and is
highly likely to divest them (Kim et al., 2004). Thus, small divisions will be very careful and
diligent in their risky choices and are more likely to “place greater emphasis on profitability”
(Kim et al., 2004: 619) by undertaking “good” risks. Formally stated:
H10: Divisional size will moderate the relationship between divisional risk taking and divisional
performance, such that larger divisions are more likely to engage in “bad” risk taking.

METHODS
For Chapter Two I utilize two separate samples which allows me to test the proposed
relationships in a more robust manner and reach conclusions with higher external validity. More
specifically, Sample 1 includes a wider range of firms and focuses on a broader construct of risk
taking – e.g. divisional capital expenditures. Also, I am able to obtain all variables necessary for

125

testing the proposed hypotheses. Sample 2 focuses on the banking industry and offers a more
fine-grained examination of banking divisions’ risk taking. In particular, I analyze banking
managers’ decisions to alter the structure of their risk portfolios by changing the relative weights
of various risk-laden categories comprising the portfolio. Due to data limitations, the banking
sample allows me to test only some of the proposed relationships.
Sample 1
Data was drawn from several sources: corporate and segment financial data comes from
Compustat, executive compensation and governance measures are accessed through Execucomp,
and divisional managers’ compensation data is obtained through manual search of corporate
annual reports. The sample includes all US-based firms from 1998 to 2009. The sample is
subject to the following screening criteria. First, units with less than $10 million in sales or assets
are excluded (McGahan and Porter, 2003; McNamara, Vaaler, and Aime, 2005). Second, I drop
units which are classified as “corporate” or “other” because they are not active units. Third, if the
unit’s four-digit SIC category is defined as “not elsewhere classified” or “nonclassifiable
establishment”, this unit is not included in the sample.
Hypothesis 8 is tested on a subsample of Compustat firms. During the period of the study
(1998-2009) I identified all diversified S&P 500 corporations and matched them to the initial
sample. For all corporations that had no missing values on all variables in the model I executed a
manual search of their DEF 14A statements which provide information on the compensation of
the top five highest paid executives. After identifying the compensation of a divisional executive,
I matched it back to the initial sample. One caveat should be recognized – because not all
divisional executives are among the top five highest paid executives during a year, there is a
large underrepresentation in this subsample and it is likely that it is biased and not generalizable

126

to the entire population of multidivisional firms. Thus, the results should be interpreted with
caution (see Appendix 2).
Dependent variable
The dependent variable is divisional performance, operationalized as operating profits
(OPS) at period t.
Independent variables
Risk taking is measured as capital expenditures. Capital expenditures are an appropriate
measure of managerial risk taking because: a) they represent difficult-to-reverse, long-term
investments (Audia and Greve, 2006); and b) “their consequences are uncertain and depend
partly on difficult to predict environmental factors” (Desai, 2008: 598). Thus, capital
expenditures represent an ex-ante decision of allocating financial resources to projects with
uncertain outcomes. In addition, research by Miller and Bromiley (1990) identifies capital
expenditures as one of the key dimensions of strategic risk. Finally, capital expenditures have
received solid support as a good operationalization of risk taking (Audia and Greve, 2006; Desai,
2008; Sanders and Hambrick, 2007).
Diversification in this chapter is argued to impact the ability of divisions to learn from
each other and the quality of their risk taking. Thus, both degree of (dis)similarity and number of
similar divisions are expected to moderate the relationship between divisional risk taking and
performance. To test Hypothesis 1, I measure diversification as the count of same-corporation
divisions (excluding the focal division) that have the same 2-digit SIC code with the focal
division. This variable is expected to positively moderate the relationship between risk taking
and performance. To test Hypothesis 2, I measure diversification as the unrelated component

127

(DU) of the entropy measure (Jacquemin and Berry, 1979). This variable is expected to have a
negative moderating effect on the relationship between divisional risk taking and performance.
Institutional ownership is calculated as the total number of shares owned by institutional
investors, such as pension funds, mutual funds, and hedge funds, divided by the total number of
outstanding firm shares (Brickley et., 1988).
Blockholders is the number of stock owners with at least 5% stake in the company
(Hoskisson et al., 1994).
CEO contingent pay is measured in two ways – CEO option pay and CEO stock
ownership. CEO option pay is measured as the value of options granted during a year using
Black and Scholes method. CEO stock ownership is measured as the value of shares owned by
the CEO (Sanders, 2001; Wright et al., 2002).
Industry munificence is calculated by following prescriptions of prior research (Dess and
Beard, 1984; McNamara, Haleblian, and Dykes, 2008). More specifically, for each 4-digit SIC
industry I calculate the total value of sales and regress it on a year-counter variable. I use 5-year
time windows. The regression coefficients for each industry are divided by mean industry sales.
Industry dynamism is calculated by dividing the standard errors of the same regression
coefficients by mean industry sales. After standardizing the variable, I calculated the absolute
value and multiplied by -1. Thus, higher values indicate moderate levels of dynamism and low
values indicate the two extremes (e.g. low and high dynamism).
Industry complexity is usually represented as a concentration ratio (Boyd, 1995; Palmer
and Wiseman, 1999). Thus, I divide the sales of the four largest firms for each 4-digit SIC
industry by the total sales in that industry. I use the inverse of this ratio, so that lower
concentration translates into higher complexity. After standardizing the variable, I calculated the

128

absolute value and multiplied by -1. Thus, higher values indicate moderate levels of complexity
and low values indicate the two extremes (e.g. low and high complexity).
Divisional relative size is calculated as the ratio of divisional to corporate assets.
Divisional managers’ ownership is measured as the value of corporate shares listed in the
annual corporate report.
Divisional managers’ stock options reflect the annual award of options.
Pay gap is calculated as the difference in total compensation between the corporate CEO
and the next highest paid executive (Henderson and Fredrickson, 2001).
Control variables
Divisional size is measured as the log of divisional assets. Corporate slack is
operationalized as potential slack (inverse ratio of debt to equity) and free cash flow (Bourgeois,
1981; Haleblian et al., 2012; Iyer and Miller, 2008). Prior divisional performance is based on
divisional OPS at period t-2. Divisional sales growth is measured as the annual change in sales
(Haleblian et al., 2012).
All independent and control variables are measured in period t-1.
Sample 2
In this sample I utilize data on the banking industry and more specifically, I collect
variables for parent and divisional banking establishments. The data is obtained from Federal
Reserve Bank Regulatory (Call Reports) databases. In these databases parent banks are
represented as bank holding companies that are comprised of several commercial banks which
operate like divisional units. For example, Citicorp Holdings is the corporate parent of several
nationally represented commercial banks, including Citibank NV NA, Citibank NY NA,
Citibank South Dakota NA, Citibank Delaware.

129

The cut-off criterion for inclusion of banks in the sample was $80 million in line with
recommendations from prior research (McNamara et al., 2002).
Hypothesis 8 (moderating role of divisional managers’ incentive pay) will not be tested in
the Banking sample due to data unavailability. Hypotheses 6, 7a, 7b, and 9 will be tested on a
subsample of banks. After estimating the models with the initial available data from the Call
Reports, I identified all bank holding corporations with no missing data on all variables. Then, I
manually searched the ticker symbols of those corporations and used those tickers to obtain data
for CEO compensation and external governance factors from Compustat. After adding those
supplementary variables to the rest of the sample, I reestimated my models. It should be noted
that this subsample is much smaller (about 11% of the initial sample), it is likely to be biased and
not representative of the entire banking population. Thus, the results from this subsample,
presented in Appendix 2, should be interpreted with caution.

Dependent variable
Bank divisional performance is measured as return on assets (ROA) – it is the most
commonly used measure of performance in the banking industry research (Haveman, 1993;
McNamara et al., 2002).
Independent variables
Bank divisional risk taking. Following recommendation by Miller and Bromiley (1990)
and Wiseman and Catanach (1997) regarding the multidimensional character of risk taking, I use
three types of risk taking – bank, loan, and bank liquidity risk – which have been empirically
validated in prior research (Shrieves and Dahl, 1992; Wiseman and Catanach, 1997). Bank risk is
an indicator of capital adequacy and solvency (Marcus, 1983) and reflects the risk position of the

130

banking institution. It is measured as the ratio of risk weighted assets to total assets (Shrieves and
Dahl, 1992). Loan risk represents an operational risk stemming from accepting loans with high
probability of borrower default. It is calculated as sum of non-accrual loans and one half of loans
past due 90 days to total loans (Meeker and Gray, 1987; Shrieves and Dahl, 1992). Liquidity risk
indicates the ability of a bank to stay solvent and pay its obligations (c.f. Hambrick and D’Aveni,
1988; Wiseman and Catanach, 1997). It is measured as the inverse ratio of liquid assets to total
assets (Hambrick and D’Aveni, 1988).
Level of diversification is measured in two ways. For Hypothesis 1, I use the count of
bank divisions in a bank holding corporation (excluding the focal bank division) that operate in
the same Metropolitan Statistical Area (MSA). For Hypothesis 2, I use the count of banking
divisions within the same bank holding corporation that operate in a different MSA. To account
for the impact of geographic distance between banking divisions, I devise a supplementary
measure which consists of the weighted sum of the above two measures. I expect higher values
of this measure to have a negative moderating effect.
Institutional ownership is calculated as the total number of shares owned by institutional
investors, such as pension funds, mutual funds, and hedge funds, divided by the total number of
outstanding firm shares (Brickley et., 1988).
Blockholders is the number of shareholders with at least 5% stake in the company
(Hoskisson et al., 1994).
CEO contingent pay is measured in two ways – CEO option pay and CEO stock
ownership. CEO option pay is measured as the value of options granted during a year using
Black and Scholes method. CEO stock ownership is measured as the value of shares owned by
the CEO (Sanders, 2001; Wright et al., 2002).

131

Pay gap is calculated as the difference in total compensation between the corporate CEO
and the next highest paid executive (Henderson and Fredrickson, 2001).
Bank divisional relative size is calculated as the ratio of divisional assets to bank holding
corporation assets.
Industry munificence is calculated by following prescriptions of prior research (Dess and
Beard, 1984; McNamara, Haleblian, and Dykes, 2008). More specifically, for each MSA I
calculate the total value of deposits and regress it on a year-counter variable. I use 5-year time
windows. The regression coefficients for each MSA are divided by mean MSA deposits.
Industry dynamism is calculated by dividing the standard errors of the same regression
coefficients by mean MSA deposits. After standardizing the variable, I calculated the absolute
value and multiplied by -1. Thus, higher values indicate moderate levels of dynamism and low
values indicate the two extremes (e.g. low and high dynamism).
Industry complexity is usually represented as a concentration ratio (Boyd, 1995; Palmer
and Wiseman, 1999). Thus, I divide the deposits of the four largest banks for each MSA by the
total deposits in that MSA. I use the inverse of this ratio, so that lower concentration translates
into higher complexity. After standardizing the variable, I calculated the absolute value and
multiplied by -1. Thus, higher values indicate moderate levels of complexity and low values
indicate the two extremes (e.g. low and high complexity).
Control variables
I use several control variables. Bank divisional size is measured as log of assets of the
banking division. Bank divisional slack is operationalized as potential slack and free cash flow

132

11

(Bourgeois, 1981; Haleblian et al., 2012) . Bank divisional age is measured as the time in years
since incorporation. Bank divisional growth is represented as the annual change in deposits.
Prior divisional performance is based on divisional ROA and is measured in period t-2.
All independent and control variables are measured in period t-1.
Analysis
The data for my dissertation has a panel structure where repeated observations over time
(e.g. divisional risk taking) are nested within divisions, which are nested within corporations and
industries. That is, I have variables (divisional size, divisional managers’ contingent pay, and
sales growth) which exist at the divisional level, variables (level of diversification, corporate
slack, institutional ownership, blockholder ownership, CEO contingent pay, and corporate size)
which exist at the corporate level, and variables (munificence, dynamism, and complexity) which
should be modeled at the industry level of analysis. The nesting of lower levels within higher
levels leads to lack of independence between observations and might result in incorrect estimates
of the regression coefficients (Raudenbush and Bryk, 2002) if standard OLS regression is used.
Thus, I use multilevel linear modeling (MLM) to account for the interdependence between lower
levels. MLM allows for the appropriate modeling across time of the relationship between the
variables in my model while simultaneously considering the nesting structure of the data (Bryk
and Raudenbush, 1989). Overall, by using MLM I could obtain “unbiased and efficient estimates
of the regression coefficients and their standard errors despite the dependence among
observations” (Fong et al., 2010: 637; Bryk and Raudenbush, 1989). An additional benefit of

11

In the Banking sample I measure slack at the divisional level rather than at the corporate level.
First, the availability of data allows me to do this. Second, calculating slack for the bank holding
corporation reduces largely my sample due to underrepresentation of bank holding corporations’
identification variables (RSSD 9348). Third, I do not see any theoretical rationale for a different
impact of divisional versus corporate slack on risk taking.
133

MLM is that by partitioning the variance across levels it “effectively controls for industry
effects” (Fong et al., 2010: 638; Bloom and Milkovich, 1998).
I rely on the statistical software STATA XT MIXED to conduct the panel data estimation
where level-1 equation represents the within-division variance in risk taking and higher-level
equations model the cross-sectional (e.g. between divisions) variance in divisional risk taking.
Below are the respective level equations.
(1) Divisional performancetij = π0ij + π1ij (Risk Taking)t-1ij + π2ij (Divisional relative size)t-1ij +
π3ij (Divisional contingent pay)t-1ij + π4ij (Risk taking* Divisional relative size)t-1ij +
π5ij (Risk taking* Divisional contingent pay)t-1ij + π6ij (Divisional sales growth)t-1ij +
π7ij (Prior divisional performance)t-2ij + π8ij (Divisional size)t-1ij + etij
(2a) π0ij = β00j + β01j (Level of diversification)ij + β02j (Institutional ownership)ij +
β03j (Blockholder ownership)ij + β04j (CEO contingent pay)ij + β05j (Corporate slack)ij +
u0ij
(2b) π1ij = β10j + β11j (Level of diversification)ij + β12j (Institutional ownership)ij +
β13j (Blockholder ownership)ij + β14j (CEO contingent pay)ij + u1ij
(3a) β00j = γ000 + γ001 (Industry munificence)j + γ002 (Industry dynamism)j +
γ003 (Industry complexity)j + µj
(3b) β10j = γ100 + γ101 (Industry munificence)j + γ102 (Industry dynamism)j +
γ103 (Industry complexity)j + µj

134

(4a) π2ij = β20j; (4b) π3ij = β30j; (4c) π4ij = β40j; (4d) π5ij = β50j; (4e) π6ij = β60j; (4f) π7ij = β70j;
(4g) π8ij = β80j
(5a) β01j = γ010; (5b) β02j = γ020; (5c) β03j = γ030; (5d) β04j = γ040; (5e) β05j = γ050;
(6a) β20j = γ200; (6b) β30j = γ300; (6c) β40j = γ400; (6d) β50j = γ500; (6e) β60j = γ600;
(6f) β70j = γ700; (6g) β80j = γ800
(7a) β11j = γ110; (7b) β12j = γ120; (7c) β13j = γ130; (7d) β14j = γ140
Equation (1) represents the time level where performance in year t by division i in
industry j is regressed on risk taking (year t-1), divisional relative size (year t-1), divisional size
(year t-1), divisional contingent pay (year t-1), sales growth (year t-1), the interaction term of
risk taking and divisional relative size, and the interaction term of risk taking and divisional
contingent pay. Thus, π0ij is the mean divisional performance over time for division i in industry
j, accounting for the impact of the independent variables.
In Equation (2a) I use the intercept of Equation (1), π0ij, as the dependent variable and
regress it on level of diversification, corporate slack, institutional and blockholder ownership,
and CEO contingent pay. Thus, I am able to account for the direct effect of these explanatory
variables on between-division variance in performance.
Equation (2b) models the slope coefficient, π1ij, as a dependent variable regressed on the
variables (level of diversification, institutional and blockholder ownership, and CEO contingent
pay) predicted to moderate the relationship between divisional attainment discrepancy and
divisional risk taking. Thus, Equation (2b) allows for testing of cross-level interactions.

135

Equations (3a) and (3b) represent the cross-level impact of environmental factors –
munificence, dynamism and complexity – on the relationship between divisional risk taking and
performance.
Finally, Equations 4a-4g, 5a-5e, 6a-6g, and 7a-7d represent the coefficients from
equations (1), (2a), and (2b) as dependent variables at a higher level of analysis.
All coefficients are modeled as fixed effects.
Equations (1), (2), and (3) have separate error terms – etij is the across-time residuals, and
uij is the between-unit residual.
RESULTS
Tables C1 and D1 present descriptive statistics and correlations for the variables in the
study12. I calculated VIFs and all of them are below 3 which is an indication that
multicollinearity is not a concern.
The results of the MLM regressions are presented in Tables C2 and D2. To control for the effect
of unobserved factors I have included year dummies in the analyses13. To conserve space, those
variables are not included in the tables. All variables are standardized with a mean of 0 and
standard deviation of 1.
Hypothesis 1 argues that number of related divisions would have a beneficial moderating role on
the relationship between risk taking and performance. However, in the Compustat sample I find
contrary results which are graphed in Figure A9 (p<0.05). Results are somewhat similar for

12

I have winsorized the variables in the study at 99% to eliminate the possibility that the results
are driven by extreme outliers.
13
For the Banking sample, in a supplementary analysis I created a dummy variable to distinguish
between bank holding corporations consisting of only bank-type divisions and bank holding
corporations consisting of multi-purpose divisions (e.g. credit card, services, etc.). The results
remained consistent across these two estimations.
136

liquidity risk in the Banking sample – Figure B6 (p<0.01). Results are insignificant for bank and
loan risks. Thus, there is no support for Hypothesis 1.
In Hypothesis 2 I stipulated that as diversification increases, it will negatively moderate the
relationship between risk taking and performance. The graphs in Figures A10 and B7 show that
when risk taking is measured as capital expenditures (p<0.001) and bank risk (p<0.01) the results
are opposite to H2. However, when loan risk is used to represent risk taking I find support for H2
(p<0.001) – see Figure B8.
Hypothesis 3 proposes a positive moderation of munificence on the risk-performance
relationship. However, munificence negatively moderates the main relationships between capital
expenditures and performance (p<0.001), loan risk and performance (p<0.001), and liquidity risk
and performance (p<0.05). Results in the model with bank risk are insignificant. Thus, there is
no support for Hypothesis 3 (see Figures A11 and B9).
In Hypothesis 4 I argued that dynamism will moderate the risk-performance relationship but only
moderate levels of dynamism will have a positive moderating impact. In the Compustat sample
results are significant and in the expected direction (p<0.05). In the Banking sample I found no
significant results for any type of risk. Overall, I find very weak support for H4 (see Figure A12).
Hypothesis 5 tested the moderating effect of complexity on the relationship between risk taking
and performance. In the Compustat sample, I found opposite results where moderate levels of
complexity negatively moderate the main relationship between risk taking and performance
(p<0.05). In the Banking sample, I found support for H5 in the model with bank risk (p<0.05)
and the model with loan risk (p<0.001). Results were insignificant when liquidity risk was used.
Overall, support for H5 is mixed (see Figures A13, B10-11).

137

Hypotheses 6, 7, and 9 tested the moderating role of internal and external corporate governance
mechanisms. Hypothesis 6 argued that CEO contingent pay will strengthen the risk-performance
relationship. When I used CEO shares as a moderator, I found strong support (p<0.001) for this
argument in the Compustat sample. Results were insignificant in the Banking sample. However,
when using CEO options as a moderator, I found the opposite results when capital expenditures
(p<0.001) and bank risk (p<0.01) are used as measures of risk taking. Results in models with
loan risk and liquidity risk were insignificant. Thus, Hypothesis 6 receives only partial support.
Hypothesis 9 argued that pay gap between the CEO and the divisional managers will positively
moderate the relationship between risk taking and performance. Results for capital expenditures
(p<0.001), bank risk (p<0.001), and loan risk (p<0.001) are strongly significant. The result for
liquidity risk is insignificant. Overall, there is strong support for H9 (see Table AP8 and Figures
A14-16).
Hypotheses 7a and 7b stipulated that the relationship between risk taking and performance will
be strengthened when external monitoring, such as institutional ownership and blockholders, is
higher. However, for H7a I find statistically insignificant results. For H7b results are opposite to
predictions when using capital expenditures (p<0.001) and liquidity risk (p<0.01). Results are
insignificant for bank and loan risks – see Figure A17.
In Hypothesis 8 I argued that divisional managers’ contingent pay will positively moderate the
relationship between risk taking and performance. In a supplementary analysis on a subsample of
S&P 500 corporations I tested this hypothesis. Results in Table AP7 show that divisional shares
positively moderate the main relationship (p<0.01), while divisional options are insignificant.
Thus, I find some support for H8. However, results from this subsample should be interpreted

138

with caution due to the smaller sample size (around 6% of the original sample) and the likelihood
of this sample being biased towards larger corporations (see Figure A18).
Finally, in Hypothesis 10 I argued that size would negatively moderate the relationship between
risk taking and performance. Results from the Compustat sample are insignificant. In the
Banking sample, Figure B13 shows that as the relative size of banking divisions increases, they
take more “bad” loan risks which hurt performance (p<0.01). Thus, across the two samples I find
at best marginal support for H10.

DISCUSSION
The results of this chapter extend prior research on the risk-return relationship by
developing a multi-level model where factors from multiple levels of analysis in the organization
play a critical moderating role on the impact of risk taking on performance. The findings support
prior calls for developing a contingency perspective on organizational risk taking (Deephouse
and Wiseman, 2000) and serve as evidence that this relationship is more nuanced than previously
argued. Summary of hypotheses and results are presented in Table S1. In particular, there are
several key conclusions from this chapter.
First, across the Compustat and Banking samples I found that various types of risk
exhibit different associations with performance. Strategic risk represented by capital
expenditures was positively related to performance. Considering that capital expenditures are
major strategic investments and involve large commitment of resources, it is expected that
divisional managers will be extremely careful in ex-ante evaluation of those projects and will be
willing to undertake them only when the expected returns are satisfactorily high. On the contrary,
bank risk, loan risk, and liquidity risk exhibited a consistently negative relationship with

139

performance. Those findings are consistent with prior research on the risk-return relationship in
the banking industry and suggest that banking managers “may systematically misestimate the
risk and return of the decisions they face” (McNamara and Bromiley, 1999: 332) and require a
risk premium that is insufficient for the risky choices they made (Wiseman and Catanach, 1997).
Bank risk corresponds to the exposure of the bank to a portfolio of assets with various degree of
risk, including loans and letters of credit (e.g. Shrieves and Dahl, 1992). Thus, bank risk reflects
the investment decisions of banking managers. Loan risk represents non-accrual loans and loan
overdue 90 days and also reflects prior lending decisions. Considering that banking managers
often underestimate the riskiness of borrowers (McNamara and Bromiley, 1997) and fail to
require appropriate risk compensation (McNamara and Bromiley, 1999), it could be expected
that bank and loan risks represent failing prior decisions that materialize in performance losses.
Liquidity risk exists when the bank does not have enough liquid assets to cover short-term
obligations and in order to satisfy those obligations its managers have to convert less liquid
assets into cash at a cost (Wiseman and Catanach, 1997).
Nevertheless, those inconsistent findings across the two samples raise important
implications for future research. First, it is advisable for scholars to avoid treating risk taking as a
uniform construct and instead focus on its various dimensions. Using a single operationalization
of risk could lead to misleading results and draw inappropriate conclusions regarding the true
relationship between risk and return. Second, considering that some risks are detrimental to
performance, the question that stands out is why managers do not limit those risks and emphasize
those that have a beneficial effect on performance? It might be that managers unintentionally
follow established organizational rules and routines and myopically take risks with inadequate
returns (e.g. McNamara and Bromiley, 1999). Understanding why such a suboptimal selection of

140

risks exists in organizations is critical not only to theory but also to practitioners for devising
recommendations about how to prevent such managerial inefficiencies.
Second, diversification plays an important role on the ability of divisions to implement
successful and high quality risky strategies. Contrary to hypothesized, I found that higher
number of related divisions negatively moderates the relationship between risk taking and
performance. There are several viable explanations for those results. First, the opportunities for
learning and knowledge transfer across divisions might be more limited than hypothesized. For
example, Barnett and colleagues (1994) argue that transferring lessons and experience across
divisions might be a challenge in multiunit organizations. In addition, as the number of related
divisions increases, the corporate office would face additional constraints and cognitive
limitations (March and Simon, 1958) in providing appropriate advising and counseling to
banking divisions. This idea is in line with Hill and Hoskisson (1987) who argue that increasing
number of divisions put a strain on information processing capabilities and might prevent the
corporate office from executing optimally its advising role. Second, as the number of related
divisions increases, the competition for internal resources might also rise (Liebeskind, 2000).
Thus, two or more related divisions in the same market or industry might see each other as
competitors and ultimately decide not to share information and successful past experiences
regarding risky choices. Finally, the corporate office of a corporation might intentionally saturate
a given market with several related divisions in order for them to share physical assets and thus
to obtain economies of scope. As a result, the corporate office would allocate resources
collectively (rather than individually to each division) to those related divisions and would
encourage shared decision making in order to achieve synergies (Palich et al., 2000). However,
such an approach necessitates compromises among related divisions and often divisions make

141

individual sacrifices for the common objective. In other words, divisions might undertake
strategic investments which while improving the collective performance do not improve their
individual divisional performance.
The findings for Hypothesis 2 are to a large degree in line with the findings for
Hypothesis 1. While I argued that level of diversification would have a negative moderating role
on the relationship between risk taking and performance, I find opposite results. Higher levels of
diversification usually translate into more different divisions operating in separate markets. As a
result, those divisions have the ability to tailor their risky strategies to the requirements of the
market (Barnett et al., 1994) and execute them successfully without major intervention from the
corporate office.
The one distinction is the opposite findings for loan risk – this result should be
interpreted with caution considering the nature of loan risk and the fact that it represents alreadymade inappropriate loan decisions.
Third, I find that market conditions have a strong influence on the relationship between
risk taking and performance. I proposed that munificence would create beneficial conditions for
banking divisions to undertake effective risky choices. In other words, such an environment
would be more receptive to various risky strategies (e.g. Hambrick and Finkelstein, 1987).
However, high munificence could also reduce the diligence of divisional managers to identify
optimum risky strategies. Knowing that the market would allow a larger margin for error,
managers might be less careful in which risky choices they select (Goll and Rasheed, 1997).
In terms of the moderating role of dynamism, I found in the Compustat sample that
medium levels of dynamism had a beneficial moderating impact on the risk-return relationship.
Alternatively, moderate degree of environmental dynamism offers the optimal conditions for

142

effective risk taking because: a) new opportunities and information emerge which are not
available to everyone, and b) managers could still recognize and take advantage of those
opportunities (Milliken, 1987; Simerly and Li, 2000).
Regarding the role of complexity, I found that it has a positive moderating role in the
Banking industry and a negative moderating role in the Compustat sample. These findings
indicate that the moderating impact of complexity across different empirical settings is not
universal and ponders further work in the future.
Fourth, I argued and found evidence that corporate governance factors provide important
contingencies for the risk-return relationship. In particular, I looked at how internal and external
corporate mechanisms enhance or reduce the ability of divisional managers to engage in highquality risk taking. The findings indicate that contrary to classical arguments of agency theory
(Jensen and Meckling, 1976), not all governance mechanisms uniformly contribute to managers’
successful risky initiatives. In terms of the internal factors, I examined the impact of CEO shares,
CEO options and pay gap. As prescribed by agency theorists, I found that CEO shares
successfully align the interests of managers and shareholders in that when CEOs possess more
shares they are more likely to oversee the strategic initiatives of divisional managers and
encourage them to undertake performance-beneficial risks. However, I found that CEO options,
due to their asymmetric risk properties, have an opposite moderating role of the relationship
between risk taking and performance. Because options provide an upside potential for gains but
limit the downside losses of corporate executives (Sanders, 2001), the latter are likely to
encourage sub-optimal risk taking at the divisional level – corporate managers have nothing to
lose if those divisional strategies are unsuccessful; however, if somehow the investment
initiatives turn beneficial, corporate executives gain positive outcomes. In a similar vein,

143

corporate executives might be less diligent in overseeing divisions’ risky choices due to the
protection from downside losses if those risky choices do not pan out in the future. Overall, CEO
options might encourage some “good” risk taking at the divisional level but on average those
risks prove to be unsuccessful and hurting divisional and overall corporate performance (e.g.
Sanders and Hambrick, 2007). As predicted, pay gap motivates divisional managers to be more
diligent (Henderson and Fredrickson, 2001; Lazear and Rosen, 1981) and undertake risks that
have a beneficial influence on performance.
From the external governance mechanisms, blockholders do not appear to have a
beneficial role on the risk-return relationship. More specifically, I found that they exhibit a
negative moderating role on the risk-return relationship. Prior research provides findings that
large equity holders are represented by families that have a sizable stake and control of
corporations (Faccio, Lang, and Young, 2001; LaPorta, Lopez-de-Silanes, and Shleifer, 1999).
Evidence also exists that those family owners may not necessarily pursue financial performance
but instead could prioritize other objectives. For example, Gomez-Mejia and colleagues (2007)
develop arguments and provide support that family owners focus on strategic decisions that trade
off lower profitability for non-financial benefits. More specifically, the strong identification with
and the emotional attachment to the corporation could encourage family stakeholders to overlook
financial benefits in order to preserve the family dynasty, maintain control (Gomez-Mejia et al.,
2007) and/or enhance the reputation of the corporation. With regard to H7b in this chapter, it
could be argued that blockholders may be willing to undertake low-performing capital
investments if those initiatives increase the size and visibility of the corporation or enhance their
legacy as owners.

144

Another plausible explanation for the results presented here is that large shareholders are
diligent monitors of managerial investment strategies (e.g. Mikkelson and Ruback, 1991) but
those investment strategies, such as capital expenditures, need more time to materialize into
performance improvements. If that is the case, the one-year lagged structure of my data would
not allow me to capture those effects. In a post-hoc analysis I tested the regression models with
performance measured two, three and four years after the capital expenditures were made. The
results for the interaction term between capital expenditures and blockholders did not change
significantly which is further evidence that blockholders may prioritize non-financial over
financial objectives. However, at year t+4 the interaction term between capital expenditures and
institutional ownership became positive and statistically significant (the term was marginally
significant at t+3). This finding could partially explain why in my main model I did not find
support for H7a and serves as tentative evidence that the temporal element in the risk-return
relationship should not be overlooked in future research. Alternatively, some of the inconsistent
findings in prior research could be attributed to the use of risk measures with different time
horizons.
Another implication of this post-hoc analysis is the likely distinction between the roles of
institutional shareholders and blockholders. Contrary to prior theorizing about their similarly
beneficial roles, my findings indicate that institutional owners and blockholders may pursue
different objectives and thus impact the strategic risk taking of corporations in a differentiated
way. A further theoretical and empirical analysis of those two groups of external monitors is
warranted.
Finally, divisional size exhibited a negative moderating role on the main relationship.
While those findings should be interpreted with caution (the only significant interaction was in

145

the model with loan risk), they indicate that smaller divisions might have more urgency and
exercise greater diligence in their risky strategies (e.g. Kim at al., 2004).
This chapter has its own limitations. The reliance on secondary data precludes from
understanding the real motives and perceptions of managers regarding their risky choices.
Although secondary data provides me access to a larger number of companies and banks, a tradeoff is made between external validity and deeper understanding of managerial decision
processes. The choice of the banking sample and its risk measures could be questioned due to the
timing of the study. Some could argue that during this period the banking industry was
expanding and bank managers might have perceived their investments and risky choice as less
uncertain. However, this should create a common bias among all managers and thus should not
influence the variability in amount of risk taken and the quality of risk across multiple banks. In
addition, if such a bias exists, it would make it more difficult to find statistically significant
results; the fact that I find significant results indicates that this is a conservative estimation of the
proposed relationships. The selection of capital expenses could also be put under question.
Although it could be argued that capital expenditures reflect risky choices because they are major
capital outlays with uncertain outcomes (Desai, 2008), capital expenditures represent only one
dimension of risk taking (e.g. Miller and Bromiley, 1990). In addition, the banking measures of
risk could be viewed as realized risks rather than ex ante evaluation of the riskiness of different
investments. Utilizing ex ante versus ex post measures of risk could generate different empirical
results and thus lead to different conclusions.

CONCLUSIONS AND FUTURE DIRECTIONS

146

This chapter contributes to research on the risk-return relationship in several ways. It
develops a contingency perspective and shows that the sign and strength of this relationship is
dependent on various factors residing at different levels of analysis. This is one of the first
studies to introduce a multi-level perspective on organizational risk taking and build theoretical
arguments that the embeddedness of risk taking in a larger organizational context impacts its
performance consequences. Ignoring the existence of multi-level influence on risk taking could
lead to incomplete theorizing and mask the complex multidimensional nature or organizational
risk taking. Additionally, by utilizing a multi-level framework towards risk taking I could
identify the relative impact of those moderating factors. In other words, are some of those factors
more important than others to either enhance or reduce the intensity and quality of risk taking?
Results indicate that the interaction terms for pay gap and CEO options exhibit the largest
associations with risk taking and performance. This serves as tentative evidence that governance
mechanisms might have the largest impact on risk taking in organizations; from a managerial and
practitioner standpoint this finding might have significant implications for organizational design
and subsequently for organizational efficiencies. Considering that compensation of top and
middle executives could be changed relatively easily, then appropriate adjustments to executive
pay might lead to critical improvements in the quality of managerial decisions and firm
performance. An under-explored area is the potential interaction of those cross-level factors and
the possibility of them being complements and/or substitutes. Future research should focus on
those interactions and try to identify whether there is an optimal configuration of those factors
that provides the most beneficial influence to managerial decisions.
This chhapter also shows that depending on what measures of risk are used, the
relationship could be either positive or negative. This could help in the long-standing and

147

inconclusive debate whether more risk is good or bad for corporate returns. The evidence
presented here turns this debate not into a “good/bad risk” dichotomy but suggests a contingency
view depending on what type of risk is undertaken. In addition, this dissertation confirms prior
theorizing that risk is a multidimensional construct and it should be treated as such (Miller and
Bromiley, 1990).
Overall, this chapter tries to extend our understanding of the risk-return relationship by
emphasizing the role of various multi-level factors and focusing the attention on the
multidimensional nature of risk taking. As a result, it provides ample opportunities for future
research in the area of strategic decision making and risk taking.

148

Table S1: Summary of hypotheses and results
Hypothesis
H1: Number of related divisions will moderate the relationship
between divisional risk taking and divisional performance, such
that the slope of the relationship will be more positive as the
number of related divisions increases.
H2: Level of diversification will moderate the relationship between
divisional risk taking and divisional performance, such that the
slope of the relationship will be negative (or less positive) for high
levels of diversification.
H3: Industry munificence will positively moderate the relationship
between divisional risk taking and divisional performance.

Compustat findings
No support – opposite results

Banking findings
Opposite results (liquidity risk)
NS – bank and loan risks

No support – opposite results

Opposite results (bank risk)
Support (loan risk)
NS – liquidity

No support – opposite results

H4: Dynamism will moderate the relationship between divisional
risk taking and divisional performance, such that the slope of the
relationship will be more positive at moderate levels of dynamism.
H5: Complexity will moderate the relationship between divisional
risk taking and divisional performance, such that the slope of the
relationship will be more positive at moderate levels of complexity.
H6: CEO incentive pay will positively moderate the relationship
between divisional risk taking and divisional performance.

Support

Opposite results (loan and
liquidity risk)
NS – bank risk
NS

H8: Divisional managers’ incentive pay will positively moderate
the relationship between divisional risk taking and divisional
performance.

Support (bank and loan risks)
NS - liquidity

Support for CEO shares
Opposite results for CEO
options
NS

* NS for CEO shares
Opposite results for CEO
options (only with bank risk)
* NS

Opposite results

H7a: Institutional ownership will positively moderate the
relationship between divisional risk taking and divisional
performance.
H7b: Blockholders will positively moderate the relationship
between divisional risk taking and divisional performance.

No support – opposite results

* Opposite results for liquidity
risk
NS for bank and loan risk
NA (not tested)

** Support for divisional
shares
NS for divisional options

149

TABLE S1 (cont’d)
H9: The gap between CEO pay and divisional managers’ pay will
positively moderate the relationship between divisional risk taking
and divisional performance.
H10: Divisional size will moderate the relationship between
divisional risk taking and divisional performance, such that larger
divisions are more likely to engage in “bad” risk taking.

Supported

Support for bank and loan risk
NS for liquidity risk

NS

Support (loan risk)
NS – bank and liquidity risk

* tested on a smaller sample. ** Results are based on a sample of manually collected data for S&P 500 corporations.

150

FIGURE 1: Main and moderated relationships between attainment discrepancy and divisional risk taking

Corporate
diversification

Corporate
governance
factors

H2-H3 (-)

Divisional
attainment
discrepancy

H5-H8 (+)

Divisional
risk taking

H1 (+)

H9 (+)

H4 (+)

Corporate
slack
resources

Divisional size

151

FIGURE 2: Moderated relationships between divisional risk taking and divisional performance

Environmental
factors
(munificence,
dynamism,
complexity)

H3-H5 (+, , )

Number of
related
divisions

H1 (+)

Corporate
diversification

H2 (-)

Divisional
risk taking

Divisional
performance

H6-H9 (+)

H10 (-)

Corporate
governance
factors

Divisional size

152

TABLE C1: Descriptive statistics and correlation coefficients for the study risk taking – performance (Compustat sample)

1. Performance
2. Sales growth
3. Div. size
4. Potential slack
5. Free cash flow
6. Risk
7. Relative size
8. Related divisions
9. DU
10. CEO shares
11. CEO options
12. Pay gap
13. Inst. ownership
14. Blockholders
15. Munificence
16. Dynamism
17. Complexity

M
203.8
0.07
6.1
1.1
0.05
143.7
0.35
1.2
0.33
1838.1
2197.4
7.6
0.55
1.7
5.84
0.22
0.39

SD
569.1
0.25
1.85
1.35
0.4
355.4
0.26
1.14
0.34
6256.6
4120.4
1.3
0.31
1.58
16.1
0.23
0.25

1
1.00
0.07
0.58
-0.09
0.06
0.62
0.18
-0.08
0.14
0.13
0.30
0.36
-0.02
-0.18
0.03
0.00
-0.02

2

3

4

5

6

7

8

1.00
0.07
0.02
0.09
0.05
0.06
0.02
-0.04
0.02
0.01
0.04
0.09
-0.02
0.03
0.02
0.01

1.00
-0.31
0.03
0.54
0.34
-0.01
0.13
0.16
0.32
0.47
0.01
-0.20
0.05
0.05
-0.05

1.00
0.03
-0.09
0.08
0.02
-0.14
0.03
-0.04
-0.20
-0.01
0.05
0.01
-0.06
0.04

1.00
0.01
-0.01
-0.01
0.03
-0.01
0.04
0.10
0.08
-0.08
-0.02
-0.01
0.05

1.00
0.23
-0.01
0.03
0.19
0.20
0.24
-0.05
-0.12
0.10
-0.03
-0.10

1.00
-0.14
-0.19
0.00
-0.06
-0.10
0.03
0.04
-0.04
-0.02
-0.03

1.00
-0.52
0.02
-0.05
-0.01
-0.01
0.03
0.03
0.01
0.02

153

TABLE C1 (cont’d)

9
1. Performance
2. Sales growth
3. Div. size
4. Potential slack
5. Free cash flow
6. Risk
7. Relative size
8. Related divisions
9. DU
10. CEO shares
11. CEO options
12. Pay gap
13. Inst. ownership
14. Blockholders
15. Munificence
16. Dynamism
17. Complexity

N=5329

10

11

12

13

14

15

16

1.00
0.05
0.15
0.18
-0.04
-0.11
-0.06
0.04
0.00

1.00
0.14
0.10
-0.09
-0.08
-0.01
-0.02
0.03

1.00
0.56
0.02
-0.16
0.04
-0.02
0.04

1.00
0.19
-0.09
0.00
0.03
0.03

1.00
0.58
-0.03
0.01
0.10

1.00
-0.07
-0.02
0.05

1.00
0.01
-0.09

1.00
0.06

Correlations larger than 0.02 and smaller than -0.03 are significant at p<0.05

154

TABLE D1: Descriptive statistics and correlations for the study risk taking – performance (Banking sample)

M
1. ROA
0.7
2. Sales growth
0.08
3. Divisional size 12.97
4. Divisional age 63.8
5. Slack
-0.01
6. Bank risk
0.7
7. Loan risk
0.01
8. Liquidity risk 0.92
9. Relative size
0.29
10. Related
1.4
diversification
11. Unrelated
5.5
diversification
12. Munificence 1.3
13. Dynamism
0.17
14. Complexity
0.3
N= 7742

SD
0.67
0.15
1.6
44.6
0.09
0.14
0.01
0.10
0.27
3.9

1

2

3

4

5

6

7

-0.01
0.01
0.08
0.11
-0.10
-0.40
-0.08
-0.04
0.02

1.00
0.14
-0.11
0.15
0.05
-0.06
0.02
0.10
-0.02

1.00
0.22
-0.04
0.15
0.05
-0.05
0.36
-0.08

1.00
-0.10
-0.04
-0.02
-0.02
0.24
-0.07

1.00
0.01
-0.24
0.01
0.04
0.03

1.00
0.10
0.35
0.10
-0.12

1.00
0.01 1.00
0.05 -0.01 1.00
-0.07 -0.07 -0.25 1.00

8.2

0.02

-0.02 0.07

-0.02

-0.08 0.08

9.7
0.29
0.21

0.01 0.02
-0.02 0.03
-0.02 0.00

-0.01 0.02
0.10 -0.11
-0.12 -0.09

8

-0.02 0.02

9

10

-0.46 0.00

0.01 0.05 -0.03 0.03 -0.03 0.19
-0.01 0.02 0.03 -0.02 -0.03 0.00
0.03 -0.08 0.01 -0.11 -0.07 0.38

Correlations larger than 0.02 and smaller than -0.02 are significant at p<0.05.

155

11

12

13

1.00
-0.02 1.00
0.03 -0.54 1.00
-0.12 -0.04 0.22

TABLE C2: Relationship between risk taking and performance
Variables
Constant
Sales growth
Divisional size
Potential slack
Free cash flow
Performancet-2
Risk taking
Relative divisional size
Related divisions (DR)
Unrelated diversification (DU)
CEO shares
CEO options
Pay gap
Institutional ownership
Blockholders
Munificence
Dynamism
Complexity

Model 1
Coeff
204.6***
33.6***
92.8***
7.3
6.8
315.98***
92.8***
5.9
-11.2
20.4**
8.4
-3.5
27.8***
-12.4
-10.1†
-1.97
-6.7
6.7

Risk taking x Relative size
Risk taking x DR
Risk taking x DU
Risk taking x CEO shares
Risk taking x CEO options
Risk taking x Pay gap
Risk taking x Inst. ownership
Risk taking x Blockholders
Risk taking x Munificence
Risk taking x Dynamism
Risk taking x Complexity
N

Model 2
Coeff
168.8***
30.8***
98.2***
1.2
8
288.1***
22.6
15.9*
-11
28.9***
0.7
8.9
38.7***
-15.7†
-19.6**
-9.5
-2.5
-0.7

SE
(22)
(3.9)
(9.8)
(7.7)
(6)
(8.5)
(20.6)
(6.6)
(7.7)
(7.1)
(7.5)
(5.8)
(5.96)
(9.2)
(5.98)
(7.3)
(6.4)
(11.3)

7.9
-23.1*
43.9***
15.6***
-100.4***
103.58***
-9.91
-44***
-51.5***
16.2*
-28.64*

SE
(22.8)
(4.1)
(9.7)
(7.8)
(6.4)
(8.5)
(9.4)
(6.4)
(7.6)
(7)
(7.5)
(5.9)
(6.1)
(8.9)
(6.1)
(7.5)
(6.6)
(11.5)

(8.5)
(11.2)
(9.4)
(4.3)
(5.1)
(7.7)
(15.7)
(8.7)
(9.9)
(6.95)
(14.1)

5132
4864
2

Prob > χ
Variance within divisions
Variance between divisions
Variance between firms
Variance between industries

0.000

0.59
0.07
0.3
0.03

Wald χ

5132
5252

0.000

2

0.54
0.11
0.32
0.03

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001
156

TABLE D2: Relationship between risk taking and performance

Variables
Constant
Sales growth
Divisional size
Divisional age
Slack
Performancet-2
Risk taking
Relative
divisional size
# Related
divisions
Unrelated
diversification
Munificence
Dynamism
Complexity
Risk taking *
relative size
Risk taking *
related divisions
Risk taking *
unrelated
diversification
Risk taking *
Munificence
Risk taking *
Dynamism

Model 1 – using
Bank risk
Coeff
SE
0.83*** (0.03)
-0.004 (0.01)
0.01 (0.01)
0.04*** (0.01)
0.06*** (0.01)
0.13*** (0.01)
-0.03** (0.01)

Model 2 – using
Model 3 – using
Model 4 – using Model 5 – using
Bank risk
Loan risk
Loan risk
Liquidity risk
Coeff
SE
Coeff
SE
Coeff
SE
Coeff
SE
0.83*** (0.03) 0.80***
(0.03) 0.79*** (0.03) 0.84*** (0.03)
-0.004 (0.01)
-0.01 (0.01)
-0.01 (0.01)
-0.003 (0.01)
0.02 (0.01)
0.01 (0.01)
0.02 (0.01)
0.01 (0.01)
0.04*** (0.01) 0.04*** (0.01) 0.04*** (0.01) 0.04*** (0.01)
0.05*** (0.01)
0.02* (0.01)
0.02* (0.01) 0.05*** (0.01)
0.12*** (0.01) 0.10*** (0.01) 0.10*** (0.01) 0.12*** (0.01)
0.005 (0.02) -0.18*** (0.01) -0.14*** (0.01)
-0.005 (0.01)

Model 6 – using
Liquidity risk
Coeff
SE
0.84*** (0.03)
-0.003 (0.01)
0.01 (0.01)
0.04*** (0.01)
0.05*** (0.01)
0.12*** (0.01)
0.01 (0.01)

-0.03* (0.01)

-0.03* (0.01)

-0.02†

(0.01)

-0.02† (0.01)

-0.03**

(0.01)

-0.03** (0.01)

0.01 (0.02)

0.01 (0.02)

0.01

(0.02)

0.00 (0.02)

0.01

(0.02)

0.02 (0.02)

-0.01
0.01
0.01
-0.01

(0.01)
(0.01)
(0.01)
(0.01)

-0.02
0.001
0.02
-0.01

(0.01)
(0.01)
(0.01)
(0.01)

-0.01
0.001
0.02
-0.01

(0.01)
(0.01)
(0.01)
(0.02)

-0.03*
0.002
0.02
-0.01

(0.01)
(0.01)
(0.01)
(0.02)

-0.01
0.01
0.01
-0.01

(0.01)
(0.01)
(0.01)
(0.02)

-0.02†
-0.001
0.02
-0.01

(0.01)
(0.01)
(0.01)
(0.01)

0.01 (0.01)

-0.02** (0.01)

-0.01 (0.01)

0.002 (0.01)

0.01 (0.01)

-0.04** (0.01)

0.03** (0.01)

-0.06*** (0.01)

0.02† (0.01)

-0.01 (0.01)

-0.02*** (0.01)

-0.02* (0.01)

0.00 (0.01)

0.01 (0.01)

0.02 (0.01)

157

TABLE D2 (cont’d)
Risk taking *
Complexity
N

0.04* (0.02)

0.05*** (0.01)

0.01 (0.02)

7766
1643

2

Wald χ

2

Prob > χ
Within division
variance
Among division
variance
Among firm
variance
Among industry
variance

7766
1657

7736
2326

7736
2443

7766
1626

7766
1648

0.000

0.000

0.000

0.000

0.000

0.000

0.8

0.8

0.8

0.81

0.79

0.79

0.04

0.04

0.06

0.05

0.04

0.04

0.14

0.14

0.12

0.12

0.14

0.15

0.02

0.02

0.01

0.01

0.02

0.02

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

158

FIGURE A9: Interaction between risk taking and # related divisions (Compustat sample)

250
200

150

Performance

100
50

0
_

low

Risk taking
Performance
low related divisions
high related divisions

0
high

high/low +1/-1 SD

FIGURE B6: Interaction between risk taking and # of related divisions (Banking sample using
liquidity risk)

0.9
0.88
0.86
0.84

Performance

0.82

0.8
0.78
0.76
_

low

Risk taking
Performance
low related divisions
high related divisions

159

0
high

high/low +1/-1 SD

FIGURE A10: Interaction between risk taking and unrelated diversification (Compustat sample)

300
250
200
Performance

150
100

50
0
_

low

Risk taking
Performance

low unrelated diversification
high unrelated diversification

0
high

high/low +1/-1 SD

FIGURE B7: Interaction between risk taking and unrelated diversification (Banking sample
using bank risk)

0.88
0.86
0.84
0.82
Performance

0.8
0.78
0.76

0.74
0.72
0.7
_

low

Risk taking
Performance

low unrelated diversification
high unrelated diversification

160

0
high

high/low +1/-1 SD

FIGURE B8: Interaction between risk taking and unrelated diversification (Banking sample
using loan risk)

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

Performance

_

low

Risk taking
Performance

low unrelated diversification
high unrelated diversification

0
high

high/low +1/-1 SD

FIGURE A11: Interaction between risk taking and munificence (Compustat sample)

300
250
200
Performance

150
100

50
0
_

low

Risk taking
Performance
low munificence
high munificence

161

0
high

high/low +1/-1 SD

FIGURE B9: Interaction between risk taking and munificence (Banking sample)

0.87
0.86
0.85
0.84
Performance

0.83
0.82
0.81

0.8
0.79
0.78
_

low

Risk taking
Performance
low munificence
high munificence

0
high

high/low +1/-1 SD

FIGURE A12: Interaction between risk taking and dynamism (Compustat sample)

195
190
185
180
175
170
165
160
155
150
145
140

Performance

_

low

Risk taking
Performance

low and high dynamism
moderate dynamism

162

0
high

high/low +1/-1 SD

FIGURE A13: Interaction between risk taking and complexity (Compustat sample)

250
200

150

Performance

100
50

0
_

low

Risk taking
Performance

low and high complexity
moderate complexity

0
high

high/low +1/-1 SD

FIGURE B10: Interaction between risk taking and complexity (Banking sample using bank risk)

0.92
0.9
0.88
0.86
Performance

0.84
0.82
0.8

0.78
0.76
0.74
_

low

Risk taking
Performance

low and high complexity
moderate complexity

163

0
high

high/low +1/-1 SD

FIGURE B11: Interaction between risk taking and complexity (Banking sample using loan risk)

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

Performance

_

low

Risk taking
Performance

low and high complexity
moderate complexity

0
high

high/low +1/-1 SD

FIGURE A14: Interaction between risk taking and CEO shares (Compustat sample)

250
200

150

Performance

100
50

0
_

low

Risk taking
Performance
low CEO shares
high CEO shares

164

0
high

high/low +1/-1 SD

FIGURE A15: Interaction between risk taking and CEO options (Compustat sample)

300
250
200
Performance

150
100

50
0
_

low

Risk taking
Performance
low CEO options
high CEO options

0
high

high/low +1/-1 SD

FIGURE A16: Interaction between risk taking and Pay gap (Compustat sample)

400

350
300

250
Performance

200
150

100
50

0
_

low

low pay gap

Risk taking
Performance

high pay gap

165

0
high

high/low +1/-1 SD

FIGURE A17: Interaction between risk taking and Blockholders (Compustat sample)

300
250
200
Performance

150
100

50
0
_

low

Risk taking
Performance
low blockholders
high blockholders

0
high

high/low +1/-1 SD

FIGURE B12: Interaction between risk taking and divisional relative size (Banking sample)

Performance

_

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
low

Risk taking
Performance
low relative size
high relative size

166

0
high

high/low +1/-1 SD

APPENDICES

167

APPENDIX 2
In this appendix I present supplementary analyses using hand-collected divisional managers’ compensation data (for the Compustat
sample) and internal and external governance data (for the Banking sample).

TABLE AP6: Correlation table for the reduced sample testing H8
M
1. Performance
203.8
2. Sales growth
0.07
3. Div. size
6.1
1.1
4. Potential slack
0.05
5. Free cash flow
6. Risk taking
143.7
7. Relative size
0.35
623.6
8. Divisional shares
9. Divisional options 1180.2
10. Related divisions
1.2
11. DU
0.33
12. CEO shares
1838.1
13. CEO options
2197.4
14. Pay gap
7.6
15. Inst. ownership
0.55
16. Blockholders
1.7
17. Munificence
5.84
18. Dynamism
0.22
19. Complexity
0.39

SD
569.1
0.25
1.85
1.35
0.4
355.4
0.26
1131.3
1683.6
1.14
0.34
6256.6
4120.4
1.3
0.31
1.58
16.1
0.23
0.25

1
1.00
0.10
0.69
0.11
0.12
0.65
0.47
0.51
0.17
-0.27
0.11
0.23
0.15
0.27
-0.15
-0.27
0.08
-0.03
-0.05

2

3

4

5

6

7

1.00
0.10
-0.01
0.11
0.11
0.11
0.08
0.06
-0.02
-0.04
0.08
-0.03
-0.06
0.14
-0.01
0.00
0.10
-0.06

1.00
-0.05
0.04
0.66
0.55
0.42
0.20
-0.19
-0.04
0.26
0.14
0.33
-0.17
-0.26
0.10
0.05
-0.12

1.00
-0.01
0.17
0.26
0.00
0.00
-0.17
-0.05
0.29
0.29
0.13
0.02
-0.07
0.07
-0.29
0.13

1.00
0.00
0.11
0.01
-0.02
-0.10
0.07
-0.02
-0.09
-0.11
0.19
0.07
-0.03
0.06
0.05

1.00
0.51
0.34
0.06
-0.19
-0.08
0.50
0.03
0.15
-0.17
-0.13
0.02
-0.03
-0.22

1.00
0.10
0.09
-0.32
-0.09
0.37
0.06
0.05
0.05
-0.12
-0.12
-0.12
-0.13

168

8

9

10

1.00
0.32 1.00
-0.13 -0.10 1.00
0.06 0.13 -0.53
0.17 0.12 -0.20
0.19 0.48 -0.18
0.26 0.38 -0.12
-0.22 0.00 0.10
-0.19 -0.09 0.31
-0.01 -0.12 -0.09
0.09 0.04 0.05
-0.09 0.00 0.17

TABLE AP6 (cont’d)
11
1. Performance
2. Sales growth
3. Div. size
4. Potential slack
5. Free cash flow
6. Risk taking
7. Relative size
8. Divisional shares
9. Divisional options
10. Related divisions
11. DU
12. CEO shares
13. CEO options
14. Pay gap
15. Inst. ownership
16. Blockholders
17. Munificence
18. Dynamism
19. Complexity

N=302

12

1.00
-0.04
0.14
0.13
-0.16
-0.21
-0.03
0.01
-0.02

1.00
0.05
-0.05
-0.29
-0.05
-0.09
-0.03
-0.31

13

14

15

16

17

18

1.00
0.63 1.00
-0.10 0.04 1.00
-0.25 -0.13 0.45 1.00
0.15 0.14 -0.17 -0.14 1.00
-0.12 -0.07 0.01 0.04 -0.33 1.00
-0.01 -0.05 0.04 0.08 -0.10 -0.10

Correlation smaller than -0.12 and larger than 0.12 are significant at p<0.05.

169

TABLE AP7: Relationship between risk taking and performance testing H8
Model 1
Variables
Constant
Sales growth
Divisional size
Potential slack
Free cash flow
Performancet-2
Risk taking
Relative divisional size
Divisional managers’ shares
Divisional managers’ options
Related divisions (DR)
Unrelated diversification (DU)
CEO shares
CEO options
Pay gap
Institutional ownership
Blockholders
Munificence
Dynamism
Complexity

Coeff
-138.3
62*
141.9
38.2
57.4
159.9***
204.9***
82.1
154.8***
-3.1
-6.9
116.8†
-8.9
-42.2
25.4
214.0†
-161.0*
-25.0
-63.0
14.5

Risk taking x Relative size
Risk taking x Divisional shares
Risk taking x Divisional options
Risk taking x DR
Risk taking x DU
Risk taking x CEO shares
Risk taking x CEO options
Risk taking x Pay gap
Risk taking x Inst. ownership
Risk taking x Blockholders
Risk taking x Munificence
Risk taking x Dynamism
Risk taking x Complexity

Model 2
SE
Coeff
SE
(235.1)
-135.1
(219.1)
(31.3)
60.6*
(28.8)
(95.6)
183.4†
(98.4)
(125.5)
-85.4
(116.1)
(44.7)
87.6*
(41.2)
(35.6)
196.6***
(34.3)
(42.2)
124.0
(126.9)
(59.0)
105.8†
(60.6)
(32.5)
58.1
(40.5)
(56.3)
-7.6
(61.9)
(53.4)
-15.8
(49.1)
(61.1)
79.9
(57.3)
(53.3)
11.5
(62.9)
(63.2)
25.3
(61.5)
(70.3)
14.2
(67.7)
(127.6)
39.0
(127.1)
(62.1)
-73.9
(59.3)
(72.6)
1.6
(68.8)
(76.0)
15.1
(71.3)
(97.3)
-76.8
(92.9)
51.8
46.9**
44.1
-140.2†
48.5
7.3
-106.3*
56.9
266.1**
-21.1
9.3
466.4**
31.8

170

(51.4)
(13.9)
(33.9)
(72.4)
(45.7)
(21.7)
(44.1)
(53.4)
(85.7)
(46.3)
(71.3)
(136.7)
(97.2)

TABLE AP7 (cont’d)
N
2
Wald χ

294
369
0.000

0.000

0.56
0.13
0.25
0.06

2

Prob > χ
Variance within divisions
Variance between divisions
Variance between firms
Variance between industries

294
522
0.53
0.24
0.08
0.13

Standardized coefficients are reported. Standard errors are in parentheses.
† p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

171

FIGURE A18: Interaction between risk taking and divisional shares

150

100
50

0
Performance

-50
-100

-150
-200

-250
_

low

Risk taking
Performance
low divisional shares
high divisional shares

172

0
high

high/low +1/-1 SD

TABLE AP8: Relationship between risk taking and performance (Banking sample)

Variables
Constant
Sales growth
Divisional size
Divisional age
Slack
Performancet-2
Risk taking
Relative
divisional size
# Related
divisions
Unrelated
diversification
CEO shares
CEO options
Pay gap
Institutional
ownership
Blockholders
Munificence
Dynamism
Complexity
Risk taking *
relative size
Risk taking *
related divisions

Model 1 – using Model 2 – using Model 3 – using Model 4 – using Model 5 – using Model 6 – using
Bank risk
Bank risk
Loan risk
Loan risk
Liquidity risk
Liquidity risk
Coeff
SE
Coeff
SE
Coeff
SE
Coeff
SE
Coeff
SE
Coeff
SE
1.12*** (0.07) 1.05*** (0.1) 0.97*** (0.1) 0.99***
(0.1) 0.99*** (0.11) 1***
(0.11)
-0.05* (0.02)
-0.05* (0.02)
-0.02 (0.02)
-0.03 (0.02)
-0.05* (0.02)
-0.06* (0.02)
-0.04 (0.04)
-0.01 (0.04)
-0.02 (0.04)
-0.01 (0.04)
-0.02 (0.04)
0.01 (0.05)
0.01 (0.03)
0.02 (0.03)
-0.01 (0.03)
-0.03 (0.03)
0.00 (0.03)
0.01 (0.03)
0.02 (0.02)
0.00 (0.02)
0.00 (0.02)
-0.01 (0.02)
(0.03 (0.02)
0.01 (0.02)
0.11*** (0.02)
0.08** (0.02) 0.12*** (0.02) 0.11*** (0.02) 0.12*** (0.02) 0.09*** (0.02)
0.10*** (0.03)
-0.07 (0.07) -0.10*** (0.02)
-0.10 (0.07)
0.04† (0.02)
-0.08 (0.10)
-0.06 (0.04)

-0.05 (0.04)

-0.04 (0.04)

-0.18*** (0.05)

-0.18** (0.06)

-0.17** (0.06)

-0.11**
0.13***
-0.12***
0.17***
0.06
-0.04
0.01
0.02
-0.14†

(0.03) -0.13*** (0.04)
-0.08* (0.03)
(0.03) 0.13*** (0.03) 0.12*** (0.03)
(0.03) -0.11*** 0.03) -0.11*** (0.03)
(0.03) 0.17*** (0.03) 0.15*** (0.03)
(0.04)
(0.03)
(0.04)
(0.04)
(0.07)

0.08*
-0.05
0.00
0.03
-0.10

(0.04)
(0.03)
(0.04)
(0.04)
(0.07)

0.07*
-0.04
0.02
-0.01
-0.13†

(0.04)
(0.03)
(0.03)
(0.04)
(0.07)

-0.03 (0.04)

-0.05 (0.04)

-0.08†

(0.04)

-0.14* (0.06) -0.18*** (0.05)

-0.13*

(0.07)

(0.03) -0.10** (0.03) -0.10**
(0.04) 0.12*** (0.03) 0.12***
(0.03) -0.12*** (0.03) -0.11***
(0.03) 0.16*** (0.03) 0.18***

(0.04)
(0.03)
(0.03)
0.03)

(0.04)
(0.03)
(0.03)
(0.04)
(0.07)

0.03
0.00
0.02
0.00
-0.14†

(0.04)
(0.04)
(0.04)
(0.04)
(0.08)

-0.08*
0.16***
-0.11**
0.12***
0.04
-0.04
0.04
-0.01
-0.08

0.05 (0.04)
-0.03 (0.03)
0.01 (0.04)
0.01 (0.04
-0.13† 0.07)

-0.03 (0.04)

-0.06† (0.03)

-0.01

(0.03)

-0.05 (0.04)

-0.07 (0.06)

-0.07

(0.11)

173

TABLE AP8 (cont’d)
Risk taking * unrel.
diversification
Risk taking *
CEO shares
Risk taking *
CEO options
Risk taking *
Pay gap
Risk taking *
Inst. ownership
Risk taking *
Blockholders
Risk taking *
Munificence
Risk taking *
Dynamism
Risk taking *
Complexity
N
2
Wald χ

0.13*** (0.03)

-0.12** (0.05)

0.15**

(0.05)

0.02 (0.02)

-0.07† (0.04)

0.01

(0.05)

-0.05** (0.02)

-0.03 (0.04)

-0.06

(0.04)

0.06*** (0.02)

0.07*** (0.02)

-0.04

(0.03)

0.01 (0.04)

0.02 (0.05)

0.08†

(0.05)

-0.03 (0.03)

-0.02 (0.02)

-0.17**

(0.06)

0.02 (0.03)

-0.02 (0.02)

0.04

(0.03)

(0.02

0.03

(0.03)

-0.03 (0.05)

-0.14†

(0.08)

-0.03 (0.03)

0.01

-0.18** (0.06)
869
392

2

Prob > χ
Within division
variance
Among division
variance
Among firm variance
Among industry
variance

869
455

860
395

860
477

869
372

869
413

0.000

0.000

0.000

0.000

0.000

0.000

0.88

0.85

0.83

0.82

0.85

0.83

0.00

0.03

0.00

0.00

0.00

0.00

0.00
0.12

0.01
0.1

0.02
0.16

0.04
0.14

0.00
0.14

0.03
0.14

Standardized coefficients are reported. Standard errors are in parentheses. † p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001
174

REFERENCES

175

REFERENCES
Aldrich, H. E. 1979. Organizations and Environments. Prentice-Hall, Englewood Cliffs, NJ.
Aldrich, H., & Auster, E. 1986. Even dwarfs started small: liabilities of age and size and their
strategic implications. In Research in Organizational Behavior, Staw BM, Cummings,
LL (eds). JAI Press: Greenwich, CT: 165–198.
Amburgey, T. L., & Miner, A. S. 1992. Strategic momentum: The effects of repetitive,
positional, and contextual momentum on merger activity. Strategic Management
Journal, 13: 335-348.
Audia, P. G., & Brion, S. 2005. Reluctant to change: Self-enhancing responses to diverging
performance measures. Organizational Behavior and Human Decision Processes,
102(2): 255-269.
Audia, P. G., & Greve, H. R. 2006. Less likely to fail? Low performance, firm size, and factory
expansion in the shipbuilding industry. Management Science, 52: 83–94.
Audia, P. G., Locke, E. A., & Smith, K. G. 2000. The paradox of success: An archival and a
laboratory study of strategic persistence following a radical environmental change.
Academy of Management Journal, 43: 837–853.
Audia, P. G., Sorenson, O. & Hage, J. 2001. Tradeoffs in the organization of production: Multiunit firms, geographic dispersion and organizational learning. Advances in Strategic
Management, 18.
Barkema, H. G., & Vermeulen, F.. 1998. International expansion through start-up or acquisition:
A learning perspective. Academy of Management Journal, 31: 7–26.
Barney, J. B. 1991. Firm resources and sustained competitive advantage. Journal of
Management, 17: 99-120.
Baum, J. A. C. 1990. Adaptation and selection in organizational populations. Unpublished
manuscript, Stern School of Business, New York University.
Baum, J. A., & Dahlin, K. B. 2007. Aspiration performance and railroads’ patterns of learning
from train wrecks and crashes. Organization Science, 18: 368–385.
Baum, J. A., & Ingram, P. 1998. Survival-enhancing learning in the Manhattan hotel industry,
1898–1980. Management Science, 44: 996–1016.
Baum, J. A., Li, S. X., & Usher, J. M. 2000. Making the next move: How experiential and
vicarious learning shape the locations of chains’ acquisitions. Administrative Science
Quarterly, 45: 766–801.

176

Baum, J. A. C., Rowley, T. J., Shipilov, A. V., & Chuang, Y. T. 2005. Dancing with strangers:
Aspiration performance and the search for underwriting syndicate partners.
Administrative Science Quarterly, 50: 536–575.
Baysinger, B., & Hoskisson, R. E. 1989. Diversification strategy and R&D intensity in large
multiproduct firms. Academy of Management Journal, 32 310–332.
Beckman, C., & Haunschild, P. 2002. Network learning: The effects of partners heterogeneity of
experience on corporate acquisitions. Administrative Science Quarterly, 47: 92–124.
Berger, P., & Ofek, E. 1995. Diversification’s effect on firm value. Journal of Financial
Economics, 37: 39–65.
Bloom, M., & Milkovich, G. T. 1998. Relationships among risk, incentive pay, and
organizational performance. Academy of Management Journal, 41(3): 283–297.
Bourgeois, L. J. 1981. On the measurement of organizational slack. Academy of Management
Review, 6: 29–39.
Bowman, E. H. 1980. A risk/return paradox for strategic management. Sloan Management
Review, 21(3): 17–31.
Bowman, E. H. 1982. Risk seeking by troubled firms. Sloan Management Review, 23(4): 33–42.
Bowman, E. H. 1984. Content analysis of annual reports for corporate strategy and risk.
Interfaces, 14(1): 61–71.
Bowman, E. H., & Helfat, C. E. 2001. Does corporate strategy matter? Strategic Management
Journal, 22(1): 1–23.
Boyd, B. K. 1995. CEO duality and performance: A contingency model. Strategic Management
Journal, 16(4): 301-312.
Bradley, S. W., Aldrich, H., Shepherd, D. A., & Wiklund, J. 2011. Resources, environmental
change, and survival: asymmetric paths of young independent and subsidiary
organizations. Strategic Management Journal, 32(5): 486-509.
Brickley, J. A., Lease, R. C., & Smith, C. W. 1988. Ownership structure and voting on
antitakeover amendments, Journal of Financial Economics, 20: 267-291.
Bromiley, P. 1991. Testing a causal model of corporate risk taking and performance. Academy of
Management Journal, 34: 37–59.
Brüderl, J., & Schüssler, R. 1990. Organizational mortality: The liabilities of newness and
adolescence. Administrative Science Quarterly, 35: 530–547.

177

Brush, T. H., Bromiley, P., & Hendrickx, M. 1999. The relative influence of industry and
corporation on business segment performance: an alternative estimate. Strategic
Management Journal, 20(6): 519–547.
Bryk, A. S., & Raudenbush, S. W. 1989. Methodology for cross-level organizational research.
Research in the Sociology of Organizations, 7: 233–273.
Castrogiovanni, G. J. 1991. Environmental munificence: A theoretical assessment. Academy of
Management Review, 16: 542–563.
Chandler, A.D. 1962. Strategy and structure: Chapters in the history of the American industrial
enterprise. Cambridge, MA: The MIT Press.
Child, J., & Rodriguez, S. B. 2003. The international crisis of confidence in corporations.
Journal of Management and Governance, 7: 233–240.
Chuang, Y. T., & Baum, J. A. C. 2003. It’s all in the name: Failure-induced learning by multiunit
chains. Administrative Science Quarterly, 48: 33–59.
Connelly, B. L., Tihanyi, L., Certo, S. T., & Hitt, M. A. 2010. Marching to the beat of different
drummers: The influence of institutional owners on competitive actions. Academy of
Management Journal, 53(4): 723-742.
Cyert, R. M., & March, J. G. 1963. A behavioral theory of the firm. Englewood Cliffs, NJ:
Prentice-Hall.
Dalton, D. R., Daily, C. M., Certo, S. T., & Roengpitya, R. 2003. Meta-analyses of financial
performance and equity: Fusion or confusion?. Academy of Management Journal, 46:
13-26.
D'Aveni, R. 1989. Dependability and organizational bankruptcy: An application of agency and
prospect theory. Management Science, 35(9): 1120-1138.
Darr, E., Argote, L., & Epple, D. 1995. The acquisition, transfer, and depreciation of knowledge
in service organizations: Productivity in franchises. Management Science, 31: 1750–
1762.
Deephouse, D. L. 1996. Does isomorphism legitimate?. Academy of Management Journal, 39:
1023–1039.
Deephouse, D. L. 1999. To be different, or to be the same? It’s a question (and theory) of
strategic balance. Strategic Management Journal, 20(2): 147–166.
Deephouse, D. L., & Wiseman, R. M. (2000). Comparing alternative explanations for corporate
risk-return relations, Journal of Economic Behavior and Organization, 42: 463-482.

178

Desai, V. 2008. Constrained growth: How experience, legitimacy, and age influence risk taking
in organizations. Organization Science, 19(4): 594-608.
DeSarbo, W. S., & Grewal, R. (2008). Hybrid strategic groups. Strategic Management Journal,
29(3): 293–317.
Dess, G. G., & Beard, D. W. 1984. Dimensions of organizational task environments.
Administrative Science Quarterly, 29: 52–73.
Devers, C.E., McNamara, G., Wiseman, R.M., & Arrfelt, M. 2008. Moving closer to the action:
Examining compensation design effects on firm risk. Organization Science, 19(4): 548566.
Dougherty, D., & Hardy, C. 1996. Sustained product innovation in large, mature organizations:
Overcoming innovation-to-organization problems. Academy of Management Journal,
39: 1120-1153.
Duncan, R. B. 1972. Characteristics of organizational environments and perceived environmental
uncertainty. Administrative Science Quarterly, 17: 313–327.
Eisenhardt, K., 1989. Agency theory: An assessment and review. Academy of Management
Review, 14: 57-74.
Fama, E. F. 1980. Agency problems and the theory of the firm. Journal of Political Economy,
88: 288–307.
Fiegenbaum, A., & Thomas, H. 1985. An examination of the structural stability of Bowman's
risk-return paradox. Academy of Management Proceedings: 7-11.
Fiegenbaum, A., & Thomas, H. 1986. Dynamic and risk measurement perspectives on Bowman's
risk-return paradox for strategic management: An empirical study. Strategic
Management Journal, 7: 395-407.
Fiegenbaum, A., & Thomas, H. 1988. Attitudes towards risk and the risk return paradox:
Prospect theory explanations. Academy of Management Journal, 31: 395–407.
Fong, E. A., Misangyi, V., & Tosi, H. 2010. The effect of CEO pay deviations on CEO
withdrawal, firm size, and firm profits. Strategic Management Journal, 31: 629-651.
Galbraith, J. 1977. Organization Design. Addison-Wesley: Reading, MA.
Garud, R., & Van de Ven, A. H. 1992. An empirical evaluation of the internal corporate
venturing process. Strategic Management Journal, 13: 93–110.
Gavetti, G. 2005. Cognition and hierarchy: Rethinking the microfoundations of capability
development. Organization Science, 16: 599–617.
179

Goll, I., & Rasheed, A.M. A., 1997. Rational decision-making and firm performance: the
moderating role of environment. Strategic Management Journal, 18(7): 583–591.
Gomez-Mejia, L. R., Takacs Haynes, K., Nunez-Nickel, M., Jaconson, K. L., & MoyanoFuentes, J. 2007. Socioemotional wealth and business risks in family-controlled firms:
Evidence from Spanish olive mills. Administrative Science Quarterly, 57: 106-137.
Gomez-Mejia, L. R., & Wiseman, R. M. (1997). Reframing executive compensation: An
assessment and outlook. Journal of Management, 23: 291-375.
Granovetter, M. 1985. Economic action and social structure: the problem of embeddedness.
American Journal of Sociology 91: 481–510.
Granovetter, M. 1992. Problems of explanation in economic sociology. In Nohria and R. Eccles
eds. Networks and Organizations: Structure, Form and Action. Harvard Business
School Press, Boston, MA, 25–56.
Greene, W. H. 2003. Econometric analysis (5th ed.). New York: Pearson Education.
Greve, H. R. 1998. Performance, aspirations, and risky organizational change. Administrative
Science Quarterly, 44: 58–86.
Greve, H. R. 2003a. A behavioral theory of R&D expenditures and innovation: Evidence from
shipbuilding. Academy of Management Journal, 46: 685–702.
Greve, H. R. 2003b. Organizational learning from performance feedback: A behavioral
perspective on innovation and change. Cambridge, U.K.: Cambridge University Press.
Greve, H. R. 2011. Positional rigidity: low performance and resource acquisition in large and
small firms. Strategic management Journal, 32(1): 103-114.
Gupta, A. K., 1987. SBU strategies, corporate-SBU relations, and SBU effectiveness in strategy
implementation. Academy of Management Journal, 30, 477-500.
Haleblian, J., & Finkelstein, S. 1999. The influence of organizational acquisition experience on
acquisition performance: A behavioral learning perspective. Administrative Science
Quarterly, 44: 29–56.
Haleblian, J., McNamara, G., Kolev, K., & Dykes, B. 2012. Exploring firm characteristics that
differentiate leaders from followers in industry merger waves: A competitive dynamics
approach. Strategic Management Journal.
Hambrick, D. C., & D’Aveni, R. A. 1988. Large corporate failures as downward spirals.
Administrative Science Quarterly, 33: 1–23.

180

Hambrick, D. C., & Finkelstein, S. 1987. Managerial discretion: a bridge between polar views of
organizational outcomes. In Research in Organizational Behavior (Vol. 9), Staw B. M.,
Cummings L. L. (eds). JAI Press: Greenwich, CT; 369–406.
Hannan, M. T., & Freeman, J. 1984. Structural inertia and organizational change. American
Sociological Review, 49: 149–164.
Hansen, M. 1999. The search-tarnsfer problem: The role of weak ties in sharing knowledge
across organization subunits. Administrative Science Quarterly, 44(1): 82-111.
Hansen, M. 2002. Knowledge networks Explaining effective knowledge sharing in multiunit
companies. Organization Science, 13(3): 232-248.
Hansen, G., Hill, C. W. L., 1991. Are institutional investors myopic? A time-series study of four
echnology-driven industries. Strategic Management Journal, 12 (1): 1–16.
Haveman, H. A. 1993. Organizational size and change: Diversification in the savings and loan
industry after deregulation. Administrative Science Quarterly, 38: 20–50.
Hayes, R. H., & Abernathy, W. J. 1980. Managing our way to economic decline, Harvard
Business Review, 62: 95-101.
Hayward, M. L. A., & Shimizu, K. 2006. De-commitment to losing strategic action: Evidence
from the divestiture of poorly performing acquisitions. Strategic Management Journal,
27(6), 541-557.
Henderson, A. D., & Fredrickson, J. W. 2001. Top management team coordination needs and the
CEO pay gap: A competitive test of economic and behavioral views. Academy of
Management Journal, 44(1): 96-117.
Henderson, A. D., Miller, D., & Hambrick, D. C. 2006. How quickly do CEOs become obsolete?
Industry dynamism, CEO tenure and company performance. Strategic Management
Journal, 27: 447–460.
Hill, C. W. L., Hitt, M. A., & Hoskisson, R. E. 1988. Declining U.S. competitiveness:
Reflections on a crisis. Academy of Management Executive, 2: 51-60.
Hill, C. W. L., Hitt, M. A., & Hoskisson, R. E. 1992. Cooperative versus competitive structures
in related and unrelated diversified firms. Organization Sciences, 3: 501-521.
Hill, C. W. L. & Hoskisson, R. E. 1987. Strategy and structure in the multiproduct firm.
Academy of Management Review, 12: 331-341.
Hitt, M. A., Hoskisson, R. E., & Ireland, D. R. 1990. Mergers and acquisitions and commitment
to innovation in M-form firms. Strategic Management Journal, 11: 29–47.

181

Hoskisson, R.E., & Hitt, M.A. 1988. Strategic control systems and relative R&D investment in
large multiproduct firms. Strategic Management Journal, 9: 605-621.
Hoskisson, R. E., Hitt, M. A. & Hill, C. W. L. 1991. Managerial Risk Taking in Diversified
Firms: An Evolutionary Perspective. Organization Science, 2(3): 296-314.
Hoskisson, R. E., Hitt, M. A., & Hill, C. W. L. 1993. Managerial incentives and investment in
R&D in large multiproduct firms. Organization Science, 4: 325-341.
Hoskisson, R. E., Johnson, R. A., & Moesel, D. D. 1994. Corporate divestiture intensity in
estructuring firms: Effects of governance, strategy, and performance. Academy of
Management Journal, 37: 1207–1251.
Hoskisson, R. E., & Turk, T. 1990. Corporate restructuring: Governance and control limits of the
internal market. Academy of Management Review, 15: 459-477.
Hough, J. R., & White, M. A. 2003. Environmental dynamism and strategic decisionmaking rationality: An examination at the decision level. Strategic Management
Journal, 24 (5): 481-489.
Ingram, P., & Baum, J. A. 1997a. Chain affiliation and the failure of Manhattan hotels, 1898–
1980. Administrative Science Quarterly, 42: 68–102.
Ingram, P., & Baum, J. A. C. 1997b. Opportunity and constraint: Organizations’ learning from
the operating and competitive experience of industries. Strategic Management Journal,
18: 75–98.
Iyer, D. N., & Miller, K. D. 2008. Performance feedback, slack, and the timing of acquisitions.
Academy Management Journal, 51(4).
Jacquemin, A. P., & Berry, C. H. 1979. Entropy measure of diversification and corporate growth.
Journal of Industrial Economics, 27: 359-369.
Jelinek, M., & Schoonhoven, C. B. 1990. The Innovation Marathon: Lessons from High
Technology Firms. Blackwell, Oxford, UK.
Jensen, M. C., & Meckling, W. H. 1976. Theory of the Firm: Managerial Behavior, Agency
Costs and Ownership Structure. Journal of Financial Economics, 3: 305-360.
Jensen, M. C., & Murphy, K. J. 1990. CEO incentives: it’s not how much you pay, but how.
Harvard Business Review, 68(3): 138–149.
Johnson, R., 1996. Antecedents and outcomes of corporate refocusing. Journal of Management,
22: 439–483.
Kim, J., & Finkelstein, S. 2009. The effects of strategic and market complimentarity on
182

acquisition performance: Evidence from the US commercial banking industry, 19892001. Strategic Management Journal, 30: 617-646.
Kim, H., Hoskisson, R. E. & Wan, W. P. 2004. Power dependence, diversification strategy and
performance in keiretsu member firms. Strategic Management Journal, 25: 613-636.
Kim, J. Y., Kim, J., & Miner, A. S. 2009. Organizational learning from extreme performance
experience: The impact of success and recovery experience. Organization Science,
20(6): 958-978.
Kraatz, M. S., & Zajac, E. J. 2001. How organizational resources affect strategic change and
performance in turbulent environments: Theory and evidence. Organization Science,
12(5): 632–657.
Lamont, O. 1997. Cash flow and investment: Evidence from internal capital markets, Journal of
Finance, 52: 83-109.
Langley, A. 1989. In search of rationality: The purposes behind the use of formal analysis in
organizations. Administrative Science Quarterly, 34(4): 598–631.
Lant, T. K. 1992. Aspiration level adaptation: An empirical exploration. Management Science,
38:623–644.
Lant, T., & Hurley, A. 1999. A contingency model of response to performance feedback. Group
and Organization Management, 24: 421–437.
Lant, T. K., Milliken, F. J., & Batra, B. 1992. The role of managerial learning and interpretation
in strategic persistence and reorientation: An empirical exploration. Strategic
Management Journal, 13: 585–608.
Larcker, D. F., 1983. The association between performance plan adoption and corporate capital
investment. Journal of Accounting and Economics, 5: 3-30.
Laughhunn, D. J., Payne, J. W. & Crum, R. L. 1980. Managerial risk preferences for below
target returns. Management Science, 26: 1238-1249.
Levinthal, D. A. 1991. Random walks and organizational mortality. Administrative Science
Quarterly, 36: 397–420.
Li, M., & Simerly, R. L. 1998. The moderating effect of environmental dynamism on the
ownership and performance relationship. Strategic Management Journal, 19(2): 169–
179.
Liebeskind, J.P. 2000. Internal capital markets: Benefits, costs, and organizational arrangements.
Organization Science, 11: 79–92.

183

Madsen, P. M., & Desai, V. 2010. Failing to learn? The effects of failure and success on
organizational learning in the global orbital launch vehicle industry. Academy of
Management Journal, 53(3): 451-476.
March, J. G. 1981. Footnotes to organizational change. Administrative Science Quarterly, 26:
563–577.
March, J. G., & Simon, H. 1958. Organizations. New York: Wiley.
Marcus, A. J. 1984. Deregulation and Bank Financial Policy. Journal of Banking and Finance,
8: 557-65.
McGahan, A. M, & Porter, M. E. 1999. The persistence of shocks to profitability. Review of
Economics and Statistics, 81: 143–153.
McGahan, A. M., & Porter, M. E. 2002. Source: What do we know about variance in accounting
profitability? Management Science, 48(7): 834–851.
McNamara, G., F. Aime, P. Vaaler. 2005. Is performance driven by industry- or firm-specific
factors? A response to Hawawini, Subramanian, and Verdin. Strategic Management
Journal, 26(11): 1075–1081.
McNamara, G., & Bromiley, P. 1997. Decision making in an organizational setting: Cognitive
and organizational influences on risk assessment in commercial lending. Academy of
Management Journal, 40: 1063– 1088.
McNamara, G., & Bromiley, P. 1999. Risk and return in organizational decision making.
Academy of Management Journal, 42: 330-339.
McNamara, G., Haleblian, J., & Dykes, B. 2008. The performance implications of participating
in an acquisition wave: early mover advantages, bandwagon effects, and the moderating
influence of industry characteristics and acquirer tactics. Academy of Management
Journal, 51(1): 113-130.
McNamara, G. M., Luce, R. A., & Tompson, G. H. 2002. Examining the effect of complexity in
strategic group knowledge structures on firm performance. Strategic Management
Journal, 23: 153–170.
McNamara, G., Moon, H., & Bromiley, P. 2002. Banking on commitment: Intended and
unintended consequences of an organization’s attempt to attenuate irrational
commitment. Academy of Management Journal, 45: 443–452.
Merton, R. K. 1957. Bureaucratic structure and personality. In Classic Readings in
Organizational Behavior, Ott JS (eds.)Vol. 30: Wadsworth Pub. Co., Belmont: 377–344.

184

Merton, R. K. 1968. Bureaucratic structure and personality. In Social Theory and Social
Structure, Free Press: New York: 249–260.
Mikkelson, W. H., & Ruback, R. S. 1991. Targeted Repurchases and Common Stock Returns,
RAND Journal of Economics, 22: 544–61.
Milgrom, P., & Roberts, J. 1990. Bargaining costs, influence costs, and the organization of
economic activity, In James Alt and Kenneth Shepsle, (eds.): Perspectives on Positive
Political Economy. Cambridge University Press, Cambridge.
Miller, K. D., & Bromiley, P. 1990. Strategic risk and corporate performance: An analysis of
alternate risk measures. Academy of Management Journal, 33: 756-779.
Miller, K. D., & Leiblein, M. J. 1996. Corporate risk-return relations: Returns variability versus
downside risk. Academy of Management Journal, 39: 91–122.
Milliken, F. J. 1987. Three types of perceived uncertainty about the environment: state, effect,
and response uncertainty. Academy of Management Review, 12: 133– 43.
Miner, A. S., Kim, J.-Y., Holzinger, I., & Haunschild, P. R. 1999. Fruits of failure:
Organizational failure and population level learning. In Miner A.S. & Anderson P.
(Eds.), Advances in strategic management: Population level learning and industry
change, vol. 16: 187–220. Greenwich, CT: JAI Press.
Mintzberg, H. 1979. An emerging strategy of "direct" research. Administrative Science
Quarterly, 24, 580-589.
Misangyi, V. F., Elms, H., Greckhamer, T., & LePine, J. 2006. A new perspective on a
fundamental debate: A multi-level approach to industry, corporate, and business-unit
effects. Strategic Management Journal, 27: 571-590.
Ocasio, W. 1997. Towards an attention-based view of the firm. Strategic Management Journal,
18: 187–206.
O’Reilly, C., & Chatman, J. 1996. Culture as social control: Corporations, cults, and
commitment. Research in Organizational Behavior, 17: 157-200.
Palich, L. E., Cardinal, L. B., & Miller, C. C. 2000. Curvilinearity in the diversificationperformance linkage: An examination of over three decades of research. Strategic
Management Journal, 21: 155-174.
Palmer, T. B., & Wiseman, R. M. 1999. Decoupling risk taking from income stream uncertainty.
A holistic model of risk. Strategic Management Journal, 20: 1037–1062.
Parks, J. M., & Conlon, E. J. 1995. Compensation contracts: Do agency theory assumptions
predict negotiated agreements? Academy of Management Journal, 38(3): 821–838.
185

Parmigiani, A., & Holloway, S. S. 2011. Actions speak louder than modes: Antecedents and
implications of parent implementation capabilities on business unit performance.
Strategic Management Journal, 32(5): 457-485.
Porter, M. 1980. Competitive Strategy. Free Press: New York.
Pound, J. 1988. Proxy contests and the efficiency of shareholder oversight, Journal of Financial
Economics, 20: 237-266.
Rajan, R. G., Servaes, H. & Zingales, L. 2000. The cost of diversity: Diversification discount and
inefficient investment. Journal of Finance, 55: 35-80.
Rappaport, A . 1978. Executive incentives vs. corporate growth. Harvard Business Review,
56(3): 28-46.
Raudenbush, S. W., & Bryk, A. S. 2002. Hierarchical Linear Models: Applications and Data
Analysis Methods (2nd edn). Sage: Thousand Oaks, CA.
Reger, R. K., Duhaime, I. M., & Stimpert, J. L. 1992. Deregulation, strategic choice, risk, and
financial performance. Strategic Management Journal, 13: 189-204.
Robins, J., & Wiersema, M. F. 2003. The measurement of corporate portfolio strategy: analysis
of the content validity of related diversification indexes. Strategic Management Journal,
24(1): 39–59.
Ross, J., & Staw, B. M. 1993. Organizational escalation and exit: Lessons from the Shoreham
nuclear power plant. Academy of Management Journal, 36: 701–732.
Rumelt, R. P. 1991. How much does industry matter? Strategic Management Journal, 12(3):
167–185.
Sanders, W. J. 2001. Behavioral responses of CEOs to stock ownership and stock option pay.
Academy of Management Journal, 44: 477–492.
Sanders, W. G., & Boivie, S. 2004. Sorting things out: Valuation of new firms in uncertain
markets. Strategic Management Journal, 25: 167–186.
Sanders, W.G., & Hambrick, D.C. 2007. Swinging for the fences: The effects of CEO stock
options on company risk-taking and performance. Academy of Management Journal,
50(5): 1055-1078.
Scharfstein, D. S. 1998. The dark side of internal capital markets II. Working paper, MIT.
Scharfstein, D. S., & Stein, J. C. 2000. The dark side of internal capital markets: Divisional rentseeking and inefficient investment. Journal of Finance, 55: 2537-64.
186

Scott, W., R. 2008. Institutions and organizations, 3rd [1995] ed. Thousand Oaks, CA: Sage.
Shimizu, K. 2007. Prospect theory, behavioral theory, and threat-rigidity thesis: Combinative
effects on organizational divestiture decisions of a formerly acquired unit. Academy of
Management Journal, 50(6): 1495-1514.
Shimizu, K., & Hitt, M. A. 2005. What constrains or facilitates divestitures of formerly acquired
firms? The effects of organizational inertia. Journal of Management, 31: 50–72.
Shimizu, K., & Hitt, M. A. 2004. Strategic Flexibility: Organizational Preparedness to Reverse
Ineffective Strategic Decisions. Academy of Management Executive, 18(4): 44-59.
Shin, H., & Stulz, R. 1998. Are internal capital markets efficient?. Quarterly Journal of
Economics, 113: 531-553.
Shleifer, A., & Vishny, R. W. 1986. Large shareholders and corporate control. Journal of
Political Economy, 94: 461-488.
Shrieves, R. E., & Dahl, D. 1992. The Relationship between risk and capital in commercial
banks. Journal of Banking and Finance, 16(2): 439-457.
Simerly, R., & Li, M. 2000. Environmental dynamism, capital structure, and performance: A
theoretical integration and an empirical test. Strategic Management Journal, 21: 31–49.
Singh, J. V. 1986. Performance, slack, and risk taking in organizational decision making.
Academy of Management Journal, 29: 562–585.
Sitkin, S. B., & Pablo, A. L. 1992. Reconceptualizing the determinants of risk behavior.
Academy of Management Review, 17: 9-38.
Starbuck, W. H. 1976. Organizations and their environments. In M. Dunnette (Ed.), Handbook
of industrial and organizational psychology (pp. 1069-1124). Chicago: Rand McNally.
Staw, B. M. 1981. The escalation of commitment to a course of action. Academy of
Management Review, 6: 577-587.
Staw, B. M. 1997. The escalation of commitment. In Z. Shapira (Ed.), Organizational decision
making: 191–215. New York: Cambridge University Press.
Staw, B. M., Sandelands, L. E., & Dutton, J. E. 1981. Threat-rigidity effects in organizational
behavior: A multi-level analysis. Administrative Science Quarterly, 26: 501–524.
Sutcliffe, K., & McNamara, G. 2001. Controlling decision making practice in organizations.
Organization Science, 12(4): 484-501.

187

Thomas III, L. G., & Waring, G. 1999. Competing capitalisms: Capital investment in American,
German, and Japanese firms. Strategic Management Journal, 20(8): 729-748.
Tushman, M. L., & Romanelli, E. 1985. Organizational evolution: A metamorphosis model of
convergence and reorientation. In L. L. Cummings & B. M. Staw (Eds.), Research in
organizational behavior, vol. 7: 171-222. Greenwich, CT: JAI Press.
Tversky, A., & Kahneman, D. 1986. Rational choice and the framing of decisions. Journal of
Business, 59: 251-278.
Vissa, B., Greve, H. R., & Chen, W. R. 2010. Business group affiliation and firm search
behavior in India: Responsiveness and focus of attention. Organization Science, 21(3):
696-712.
Wiersema, M. F., & Bantel, K. A. 1993. Top management team turnover as an adaptation
mechanism: The role of the environment. Strategic Management Journal, 14(7): 485504.
Williamson, O. E. 1975. Markets and Hierarchies: Analysis and Antitrust Implications. Free
Press: New York.
Williamson, O.E. 1985. The Economic Institutions of Capitalism. Free Press, New York.
Williamson, O. 1996. The Mechanisms of Governance. Oxford University Press: New York.
Wright, P., Kroll, M., Lado, A., & Van Ness, B. 2002. The structure of ownership and corporate
acquisition strategies. Strategic Management Journal, 22(1): 41–53.
Wiseman, R. M. 1992. An empirical examination of strategic change in the savings and loan
industry. Unpublished dissertation.
Wiseman, R. M., & Bromiley, P. 1991. Risk-return associations: Paradox or artifact? An
empirically tested explanation. Strategic Management Journal, 12: 231-242.
Wiseman, R. M., & Bromiley, P. 1996. Toward a model of risk-taking by declining
organizations. Organization Science, 7(5): 524-543.
Wiseman, R., & Catanach, A. 1997. A longitudinal disaggregation of operational risk under
changing regulations: Evidence from the savings and loan industry. Academy of
Management Journal, 40: 799-830.

188