ThePromiseandProblemsofOrganizationalCulture.pdf

Group & Organization Management
2014, Vol. 39(6) 595 –625

© The Author(s) 2014
Reprints and permissions:

sagepub.com/journalsPermissions.nav
DOI: 10.1177/1059601114550713

gom.sagepub.com

Article

The Promise
and Problems of
Organizational Culture:
CEO Personality,
Culture, and Firm
Performance

Charles A. O’Reilly III1, David F. Caldwell2,
Jennifer A. Chatman3, and Bernadette Doerr3

Abstract
Studies of organizational culture are almost always based on two assumptions:
(a) Senior leaders are the prime determinant of the culture, and (b) culture
is related to consequential organizational outcomes. Although intuitively
reasonable and often accepted as fact, the empirical evidence for these
is surprisingly thin, and the results are quite mixed. Almost no research
has jointly investigated these assumptions and how they are linked. The
purpose of this article is to empirically link CEO personality to culture and
organizational culture to objective measures of firm performance. Using
data from respondents in 32 high-technology companies, we show that
CEO personality affects a firm’s culture and that culture is subsequently
related to a broad set of organizational outcomes including a firm’s financial
performance (revenue growth, Tobin’s Q), reputation, analysts’ stock
recommendations, and employee attitudes. We discuss the implications of
these findings for future research on organizational culture.

1Stanford University, CA, USA
2Santa Clara University, CA, USA
3University of California, Berkeley, USA

Corresponding Author:
Charles A. O’Reilly III, Graduate School of Business, Stanford University, Stanford, CA 94305,
USA.
Email: [email protected]

550713GOMXXX10.1177/1059601114550713Group & Organization ManagementO’Reilly et al.
research-article2014

http://crossmark.crossref.org/dialog/?doi=10.1177%2F1059601114550713&domain=pdf&date_stamp=2014-09-25

596 Group & Organization Management 39(6)

Keywords
organizational culture, CEO personality, firm performance

In the late 1970s and early 1980s, the topic of “organizational culture” cap-
tured managers and scholars’ interest. A series of poplar books (e.g., Davis,
1984; Peters & Waterman, 1982), academic conferences, and special issues
of scholarly journals (Administrative Science Quarterly, 1979, 1983; Journal
of Management, 1985; Journal of Management Studies, 1982) highlighted
the promise of organizational culture as a way to understand how organiza-
tions operate and succeed. The logic offered had two components that were
intuitive and seductively simple: (a) Cultures largely reflect the values and
actions of their senior leaders, and (b) cultures are important determinants of
firm performance.

The first premise was that organizational cultures—defined most com-
monly as “the basic assumptions and beliefs that are shared by organizational
members” (Schein, 1985, p. 9), or “a system of shared values defining what
is important, and norms, defining appropriate attitudes and behaviors”
(O’Reilly & Chatman, 1996, p. 166)—are largely created by an organiza-
tion’s senior leaders. For example, in the very beginning of his seminal book,
Schein (1985) claims that “the only thing of real importance that leaders do
is to create and manage culture” (p. 2). He concludes some 300 pages later
asserting, “The unique and essential function of leadership is the manipula-
tion of culture” (p. 317). The widespread assumption has been that cultures
reflect the values, beliefs, and actions of their senior leaders (e.g., Baron &
Hannan, 2002; Davis, 1984; Kotter & Heskett, 1992). However, in a recent
review of the culture literature, Schneider, Ehrhart, and Macey (2013) noted
that “although the theoretical literature on organizational culture is replete
with discussions of the influence of the founder and upper management have
on an organization’s culture, empirical studies of that relationship are hard to
find” (p. 372).

The second intuitively reasonable part of the argument was that organiza-
tional culture was a significant determinant of organizational performance.
Again, however, the evidence for this is mixed. Establishing a consistent
direct link between culture and objective firm performance has been ham-
pered by a number of conceptual challenges including disagreements about
defining culture and the dimensions associated with it (e.g., Schneider et al.,
2013), and a number of methodological challenges such as small samples,
measures designed for other purposes besides assessing culture, and variance
introduced by assessing multiple industries (e.g., Detert, Schroeder, &
Mauriel, 2000).

O’Reilly et al. 597

More than 40 years later, these two fundamental assumptions, with some
minor modifications, remain intact: Organizational culture is largely shaped
by an organization’s leaders and is presumed to be important because it can
have consequential effects on firm performance. Yet, the empirical evidence
for these claims remains fragmented and inconclusive (Hartnell, Ou, &
Kinicki, 2011). In a recent review, Sackmann (2011) concluded that even as
research on culture is becoming more methodologically sophisticated,
researchers’ use of diverse measures of culture and performance is stalling
paradigm development. Almost no studies have attempted to simultaneously
test these two fundamental assumptions by providing an empirical test of the
effects of senior leadership personality on organizational culture and the sub-
sequent effects of culture on objective indicators of organizational perfor-
mance. In doing this, we provide a clearer picture of the origins of organizational
cultures and clarify how culture can affect organizational performance.1

We begin by reviewing previous research on the effects of CEO personal-
ity and leadership on culture and firm performance. We then use data from
more than 1,000 respondents to revalidate a measure of organizational cul-
ture originally developed by O’Reilly, Chatman, and Caldwell (1991) and
investigate the associations between CEO personality, culture, and firm per-
formance for 32 high-technology firms over a 3-year period.

CEO Personality and Organizational Culture

How do senior leaders affect organizational culture? When culture is con-
ceived of as a consensus about norms (e.g., Cooke & Rousseau, 1988; Schein,
1985), then the recurring patterns of behavior of senior leaders becomes a
critical source of information about the normative for those in the orga-
nization (Bandura, 1986). Based on this social learning perspective, several
authors have identified the mechanisms through which managers might
develop and change cultures. O’Reilly and Chatman (1996) argue that the
mechanisms for developing and changing culture can be seen in the social-
psychological processes of normative and informational influence. Schein
(1985) and others have suggested similar mechanisms that act to signal the
desired normative , including systems, structures, and processes
designed to reinforce ways of thinking and behaving. While useful, these do
not answer the question of where the desired behavioral regularities come
from. Several scholars have suggested that the true origins of culture can be
found in the fundamental dispositions (values and personalities) of the orga-
nizations’ leaders (Schein, 1985). In this sense, leaders’ values and personali-
ties may be the primary building blocks of organizational culture (Baron &
Hannan, 2002; Detert et al., 2000; Fu, Tsui, Liu, & Li, 2010).2

598 Group & Organization Management 39(6)

Schneider and Smith (2004) define personality broadly to refer to those
individual attributes that “give form, structure, and consistency to people’s
behavior over time and situations” (p. 347). Personality traits are patterns of
thought, emotion, and behavior that are relatively consistent over time and
across situations. Similar to personality, values are enduring subjective judg-
ments or perspectives on what is seen as important that reflect basic disposi-
tions. Values represent one translation of dispositions into situational
preferences (Parks & Guay, 2009). As such, personality and values are impor-
tant precursors of patterns of behavior. With regard to organizational culture,
the patterns of behavior of the CEO may then become a salient source of
information about the normative .

During the past several decades, researchers have accumulated an impres-
sive body of findings providing convincing evidence that (a) personality and
values can be assessed with great accuracy (e.g., Funder, 2012; John,
Naumann, & Soto, 2008), (b) values and personality are related to a range of
important individual and life outcomes (e.g., Ozer & Benet-Martinez, 2006;
Roberts, Kuncel, Shiner, Caspi, & Goldberg, 2007), and (c) the myriad of
potential personality and value constructs can be reliably captured by five
essential personality constructs, the so-called Big Five or the Five-Factor
Model (FFM), that integrates decades of earlier research (e.g., John et al.,
2008). The five underlying dimensions include (a) Extraversion, (b)
Agreeableness, (c) Conscientiousness, (d) Neuroticism, and (e) Openness to
Experience.

Of particular relevance for organizational research, a number of studies
have linked Big Five dimensions to both leadership and job performance
(e.g., Barrick & Mount, 1991; Hoffman & Jones, 2005; Lim & Ployhart,
2004). But the vast majority of these studies have focused not on senior lead-
ers but on leader emergence and laboratory studies using student subjects.
When the focus was on senior leaders, dispositions were less useful as predic-
tors (Hoffman, Woehr, Maldagen-Youngjohn, & Lyons, 2011). Only a few
studies have attempted to link senior level leaders’ personality to culture. For
example, in an archival study of 17 CEOs, Peterson and his colleagues inves-
tigated how the personality of the CEO affected the dynamics and norms of
the senior team (Peterson, Smith, Matorana, & Owens, 2003). They found
that CEOs higher on Agreeableness had teams rated as higher in cohesion and
decentralization. Giberson et al. (2009) found some associations between Big
Five measures and organizational culture but did not link culture to organiza-
tional performance. In an attempt to more directly link CEO values, culture,
and firm performance, Berson, Oreg, and Dvir (2008) collected data from 26
CEOs and 256 of their subordinates. Their results showed that different val-
ues (self-direction, security, and benevolence) were associated with different

O’Reilly et al. 599

cultures (innovation oriented, bureaucratic, and supportive). Interestingly,
the differing cultures were differentially related to firm outcomes. More inno-
vative cultures had higher sales growth, more bureaucratic cultures were
more efficient, and more supportive cultures had higher levels of employee
satisfaction but lower sales growth.

Overall, the evidence suggests that personality as manifested in values and
behavior is associated with leadership at the CEO level (Peterson et al., 2003;
Tsui, Zhang, Wang, Xin, & Wu, 2006) and that these may affect the culture of
the organization, although the specific form of these relationships is not clear.
One implication of this argument is that an organization’s senior leaders,
because of their salience, responsibility, authority, and presumed status, have
a disproportionate impact on culture and may be a significant source of cul-
tural influence.

Organizational Culture and Firm Performance

Given the widespread interest in the potential effects of culture on firm per-
formance, it is noteworthy how little clarity there is about this connection. In
an early study, Siehl and Martin (1990) concluded that a link between culture
and firm performance “has not been—and may well never be—empirically
demonstrated” (p. 242). Almost 20 years later, Gregory, Harris, Armenakis,
and Shook (2009) observed that “few empirical studies have provided
detailed insight into the relationship” (p. 673). In a recent review of the asso-
ciations between culture and organizational effectiveness broadly defined,
Hartnell et al. (2011) found significant correlations between culture and
employee job satisfaction, obtained mixed results for culture and subjective
ratings of organizational processes and performance, but found too few stud-
ies of studies of objective performance indicators and culture to come to any
conclusions.

Beginning with the 84 studies identified by Hartnell and including others
not in their sample (e.g., Balthazard, Cooke, & Potter, 2006; Bezrukova,
Thatcher, Jehn, & Spell, 2012), we identified 31 studies that appeared to
explicitly investigate both culture and performance. A review of these showed
that only 9 studies reported associations between culture and objective, firm-
level financial performance outcomes. Of these, several used very small
samples (Calori & Sarnin, 1991; Gordon & DiTomaso, 1992; Siew & Yu,
2004). Only 6 studies had a reasonable sample size and objective perfor-
mance measures (Berson et al., 2008; Christensen & Gordon, 1999; Gordon,
1985; Kotter & Heskett, 1992; Peterson et al., 2003; Sørensen, 2002). Results
from these ranged from no associations between culture and objective firm
performance (Gordon, 1985; Sørensen, 2002) to mixed results (Christensen

600 Group & Organization Management 39(6)

& Gordon, 1999) to positive findings under specific conditions (Berson et al.,
2008; Kotter & Heskett, 1992; Peterson et al., 2003). Thus, while there is
evidence that organizational culture seems to be positively associated with
employee attitudes and subjective assessments of performance (e.g.,
Bezrukova et al., 2012; Denison & Mishra, 1995), there is little evidence
definitively linking organizational culture to objective firm-level outcomes.

There are several understandable reasons for this lack of clarity. First,
designing studies and obtaining data that allow for the assessment of culture
across organizations, especially with the CEO’s participation, has been a
daunting task, often resulting in studies with very small samples and low
power (e.g., Calori & Sarnin, 1991; Gordon & DiTomaso, 1992). For exam-
ple, Denison and Mishra (1995) used archival data on five firms to develop a
theory of culture and then used survey data in an attempt to refine their theory.
While useful, they acknowledge that, “Neither the survey instrument nor the
traits operationalized were ideal for culture research” (p. 207). Similarly, other
researchers have made use of pre-existing surveys that were not designed for
culture research but, post hoc, relabeled the constructs as “culture” (e.g.,
Marcoulides & Heck, 1993). Further compounding the issue is that the rela-
tionship between culture and firm performance has been shown to vary across
industries (e.g., Christensen & Gordon, 1999) such that a significant result
obtained in one setting may not apply in another. This is not to criticize these
efforts but to simply note the difficulty that culture research poses.

Second, there have been disagreements about the definition and measure-
ment of both culture and performance that has resulted in the use of different
frameworks and metrics that make aggregation of results difficult (e.g.,
Schneider et al., 2013). Hartnell et al. (2011) concluded that one reason for the
failure to find culture-performance relationships may be that simple measures
of culture may be too broad. In one of the first published articles on organiza-
tional culture, Andrew Pettigrew (1979) echoed this concern against the use of
simple categorizations: “While providing a general sense of orientation, culture
treated as a unitary concept in this way lacks analytical bite” (p. 574).

Responding to the concern that simple measures of culture might fail to
capture the complexity of culture across different types of organizations,
O’Reilly et al. (1991) developed a more variegated and comprehensive
approach to developing a framework for categorizing organizational culture.
Just as the Big Five personality attributes represent a mid-range theory of
personality, the Organizational Culture Profile (OCP) was designed to empir-
ically identify a set of dimensions that reflect a more comprehensive set of
organizational norms to accurately reflect the complexity, uniqueness, vari-
ety, and range of an organization’s culture. This approach has been refined
and validated by several researchers (e.g., Barber & Wesson, 1998; Judge &

O’Reilly et al. 601

Cable, 1997; Sarros, Gray, Dentsen, & Cooper, 2005; Siew & Yu, 2004). Just
as the Big Five provides a framework for summarizing the effects of person-
ality, the OCP methodology offers a comprehensive way to characterize orga-
nizational cultures on a variety of dimensions.

Finally, as researchers have explored the possible associations between
organizational culture and firm performance, there has been an evolution in
understanding the form that this relationship might take, ranging from a sim-
ple direct association to contingent relationships dependent on firm strategy
and environmental conditions (e.g., Christensen & Gordon, 1999; Khazanchi,
Lewis, & Boyer, 2007; Sørensen, 2002). However, in spite of the strong intu-
ition that organizational culture should be directly linked to firm effective-
ness, the empirical results remain equivocal.

Hypotheses Linking CEO Personality and
Organizational Culture

The argument proposed thus far is that a leader’s personality is manifested in
regularities in his or her attitudes and behaviors and these, in turn, shape cul-
tural norms and expectations. Although there is no expectation that a CEO’s
personality should directly affect firm performance, their patterns of behavior
(expressed in what questions they ask, what they pay attention to and reward,
the types of people they hire, etc.) are likely to shape their firm’s culture (e.g.,
norms regarding what people pay attention to, what behaviors are seen as
important) through a process of social learning. Thus, we expect that certain
CEO personality attributes, expressed in terms of the Big Five, may be asso-
ciated with certain types of organizational culture. Culture, in turn, may be
associated with subsequent firm performance.

Previous research has shown that, under certain conditions, each of the
Big Five dimensions may be associated with leader emergence, job perfor-
mance, culture, and possibly even the organization’s strategy (Berson et al.,
2008; Giberson et al., 2009; Judge, Bono, Iles, & Gerhardt, 2002; Nadkarni
& Herrmann, 2010). Although one could easily hypothesize how combina-
tions of the Big Five dimensions might affect organizational culture, for sim-
plicity, we focus here solely on the potential direct effects on organizational
culture.

CEO Openness to Experience

Openness to Experience is the tendency to be imaginative, unconventional,
and independent. Previous research has shown mixed associations between
Openness and leadership (Hoffman & Jones, 2005); however, we expect that

602 Group & Organization Management 39(6)

those CEOs who are high on Openness are more likely to create cultures that
value innovation and change. For instance, Nadkarni and Herrmann (2010)
reported that CEOs who were higher on Openness were also more likely to
adapt their strategies in the face of change. Thus, we predict that CEOs who
are higher on Openness will be more likely to have cultures that value inno-
vation, speed, experimentation, and risk-taking:

Hypothesis 1: CEOs who are higher on Openness to Experience will be
more likely to be associated with cultures that emphasize adaptability
(innovation, speed, and risk-taking).

CEO Conscientiousness

Conscientiousness refers to the tendency to control impulses and tenaciously
pursue goals. At very high levels, those high on Conscientiousness can also
be careful, compulsive, preoccupied with rules, and concerned with avoiding
mistakes. Therefore, at the CEO level, high levels of Conscientiousness may
produce cultures that are more rule oriented, centralized, and careful (e.g.,
Peterson et al., 2003). Thus, we expect that CEOs who are high on
Conscientiousness will be more likely to be associated with cultures that are
more detail oriented and emphasize analysis, precision, and attention to
detail.

Hypothesis 2: CEOs who are higher on Conscientiousness will be more
likely to be associated with cultures that are detail oriented.

CEO Agreeableness

Individuals high on Agreeableness are typically seen as modest, helpful, and
willing to compromise (e.g., Peterson et al., 2003). People who are low on
Agreeableness are more competitive than cooperative and can be seen as
skeptical, unconcerned about others’ feelings and antagonistic. There is some
evidence that low Agreeableness can lead to higher performance (e.g., Lepine
& Van Dyne, 2001). At the CEO level, we predict CEOs who are lower on
Agreeableness will have cultures that are more competitive and achievement
oriented with higher expectations for performance.

Hypothesis 3: CEOs who are lower on Agreeableness will be more likely
to be associated with cultures that are more results oriented (e.g., high
expectations for performance, achievement oriented).

O’Reilly et al. 603

CEO Neuroticism

People who score high on Neuroticism tend to be anxious, emotionally unsta-
ble, defensive, and upset by minor threats or frustrations. Those who are low
on Neuroticism are seen as emotionally stable, relaxed, and secure. In a meta-
analysis, Judge et al. (2002) found that Neuroticism was negatively associ-
ated with leader emergence. Because of this, leaders who score high on this
dimension are seen as more likely to be associated with cultures that are less
collaborative. Thus, we predict the following:

Hypothesis 4: CEOs who are higher on Neuroticism will be more likely
to be associated with cultures that are less collaborative.

CEO Extraversion

The most obvious aspect of Extraversion is the propensity to prefer extensive
interactions with others. However, extraverts are also characterized by opti-
mism, energy, and a preference for excitement (e.g., Judge et al., 2002).
Extraverts have been shown to be socially engaging and able to involve oth-
ers. For example, Giberson et al. (2009) found that CEOs who were higher on
Extraversion were associated with more market-oriented cultures. Thus, we
expect that CEOs who are more optimistic and sociable to be more likely to
create cultures that emphasize a customer orientation than those who are
more introverted.

Hypothesis 5: CEOs who are higher on Extraversion will be more likely
to be associated with cultures that are more customer oriented.

Organizational Culture and Firm Effectiveness

As reviewed earlier, the evidence for a link between organizational culture
and objective measures of firm performance has been mixed, with no consis-
tent evidence showing that culture contributes to financial performance. We
believe that there are two reasons for this lack of clarity. First, organizational
culture is multidimensional and has been measured in a myriad of ways. It
may be that a given facet of culture may be relevant in some circumstances
and irrelevant in others. For instance, in a stable industry with a low-cost
strategy, a hierarchical culture (internally focused-stable) might be positively
associated with success, while under more dynamic conditions this culture
could either be irrelevant or negatively associated with performance. This is
especially important for studies where the sample of firms spans multiple

604 Group & Organization Management 39(6)

industries where different cultural dimensions may be more or less valuable
(e.g., Kotter & Heskett, 1992). Research on organizational culture that uses
samples where there are likely to be wide variations in industries and firm
strategies may miss the subtle differences in cultures that drive performance
unless the design permits these to be controlled or accounted for.

Second, the association of a firm’s culture and performance may also
depend importantly on the outcome variables studied. For instance, previous
studies have documented associations between more people-friendly cultures
and employee attitudes (e.g., Berson et al., 2008). However, those aspects
that promote positive employee attitudes may be unrelated—or even nega-
tively related—to a firm’s financial performance. In contrast, those aspects of
a firm’s culture that promote financial performance (e.g., a strong emphasis
on delivering results) may be unrelated—or negatively related—to employee
attitudes. Market-based measures of a firm’s value (e.g., Tobin’s Q) in which
analysts and investors estimate the future value of a company through its
stock price may value cultural attributes like a firm’s ability to innovate even
if that aspect of the culture comes at the expense of short-term profit (e.g.,
Amazon prioritizes long-term growth over short-term profit). The fact that
some cultural dimensions may be positively related to some outcomes and
negatively related to others may account for some of the mixed results in
studies of culture and firm performance.

In spite of these complexities, research does suggest that certain cultural
dimensions may be important for firm performance and broadly related to
short-term financial performance regardless of the specific strategy adopted.
First, as previous research has shown, adaptability appears to be a critical
cultural element in promoting firm performance (e.g., Chatman, Caldwell,
O’Reilly, & Doerr, 2014; Kotter & Heskett, 1992). Second, and related, a
culture that emphasizes a results-orientation appears generally useful regard-
less of the strategy a firm pursues (e.g., Detert et al., 2000). Finally, in terms
of strategic execution, firms that are more detail oriented are more likely to
perform well when compared with those that are not, especially in competi-
tive markets (e.g., Khazanchi et al., 2007). This suggests the following three
hypotheses:

Hypothesis 6: Organizations whose cultures emphasize adaptability more
will perform better than those that emphasize adaptability less.
Hypothesis 7: Organizations whose cultures emphasize results more will
perform better than those that emphasize results less.
Hypothesis 8: Organizations whose cultures emphasize detail orientation
more will perform better than those that are less detail oriented.

O’Reilly et al. 605

Method

Research Design and Sample

There were two steps in our research design. First, to assess the culture in our
sample organizations, we used a slightly revised version of the OCP (O’Reilly
et al., 1991) to collect culture data in 2009 from a set of large, publicly traded,
high-technology firms headquartered in the United States (n = 56 firms, n =
880 respondents) and a separate set of privately held firms headquartered in
Ireland (n = 44 firms, n = 378 respondents). This full panel of data was used
for our factor analyses, that is, to identify specific dimensions of organiza-
tional culture that characterize an organization. Second, once we had identi-
fied these dimensions, we used a subset of the U.S. sample to collect data on
CEO personality as rated by company employees and firm performance for
2011. This subset of U.S. firms was qualified based on the reliability of
responses within the firm. We used only this U.S. subsample of data to test
hypotheses about CEO personality, culture, and firm effectiveness.

U.S. firm sample. We identified 60 firms to invite to participate in this study
using the following criteria: The firms were publicly traded, U.S.-headquar-
tered, had their primary operations in the high-technology sector (hardware,
software, Internet services—SIC 35xx, 36xx, 38xx, 73xx; GIC Sector 45;
S&P Economic Sector 940), and concurrently employed a minimum of 20
alumni from three focal West Coast business schools. Alumni from these
schools are highly prone to joining high-technology firms post-graduation,
and thus expedited the identification of current employees within our target
industry.

Alumni of these business schools provided culture assessments of their
employing organizations using the revised OCP. In fall 2009, we sent pro-
spective informants an email inviting them to participate in an online survey
assessing their organization’s current culture. We specified that informants’
culture assessment responses were confidential and would not be identified
to their employers, and that the study results would not identify their organi-
zations by name. We received a total of 880 culture assessments from infor-
mants in 56 of the 60 firms. We included all 880 responses from U.S.-based
employees in the factor analysis described below. Eighty-nine percent of the
56 firms were included in the list of the Fortune 1000, representing the larg-
est …