Does CEO Charisma Matter?

An Empirical Analysis of the Relationships among Organizational

Performance, Environmental Uncertainty, and Top Management Team Perceptions of CEO
Charisma
Author(s): Bradley R. Agle, Nandu J. Nagarajan, Jeffrey A. Sonnenfeld and Dhinu Srinivasan
Source: The Academy of Management Journal, Vol. 49, No. 1 (Feb., 2006), pp. 161-174
Published by: Academy of Management
Stable URL: http://www.jstor.org/stable/20159752 .
Accessed: 14/06/2014 22:34

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy
of Management Journal.

http://www.jstor.org

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 ?
Academy ofManagement Journal
2006, Vol. 49, No. 1, 161-174.

DOES CEO CHARISMA MATTER? AN EMPIRICAL ANALYSIS

OF THE RELATIONSHIPS AMONG ORGANIZATIONAL
PERFORMANCE, ENVIRONMENTAL UNCERTAINTY, AND TOP
MANAGEMENT TEAM PERCEPTIONS OF CEO CHARISMA
BRADLEY R. AGLE
NANDU J.NAGARAJAN
University of Pittsburgh

JEFFREYA. SONNENFELD
Yale University

DHINU SRINIVASAN
University of Pittsburgh

This article reports the results of a study examining the relationships among strategic
charismatic leadership, organizational performance, and environmental uncertainty
with primary data from a sample of 128 CEOs ofmajor U.S. corporations. Drawing on
770 surveys from topmanagement team members in these companies, objective stock
market and accounting data, and an objective measure of environmental uncertainty,
we found that organizational performance was associated with
subsequent perceptions
of C?O charisma but that perceptions of CEO charisma were not associated with
subsequent organizational performance, even after we incorporated the potential
moderating effect of environmental uncertainty.

Are charismatic CEOs associated withbetter organ niam, 1996). Although most of this research has been
izational performance than their less charismatic performed at lower levels ofmanagement, leadership
counterparts? The academic literature has identi theories (Conger & Kanungo, 1987; Pawar & Eastman,
fied the importance of strategic leaders and in par 1997) have suggested that the same phenomenon
ticular, of CEOs, as a determinant of organizational might be seen at the strategic level of leadership as
performance (e.g., Finkelstein & Hambrick, 1996; well. However, several authors have suggested that
Hambrick, 1989). Meanwhile, leadership research charismatic leadership may have a downside for or
ers have demonstrated that charismatic leaders can
ganizational performance because of the frequent as
be more effective than their less charismatic coun sociation between charisma and dysfunctional nar
terparts (e.g., Bass, 1985; Howell & Frost, 1989; cissism (Conger & Kanungo, 1998; Maccoby, 2000;
Koene, Vogelaar, & Soeters, 2002). A meta-analysis Sankowsky, 1995).
of studies of charismatic leadership showed that it evidence on the relationship between
Empirical
is consistently related to various measures of lead CEO charisma and organizational performance is
ership effectiveness (Lowe, Kroeck, & Sivasubrama mixed. Waldman, Ramirez, House, and Puranam
(2001) found no direct relationship between CEO
charisma and subsequent organizational perfor
mance as measured
This article is an updated version (based on an ex by net profit margin, but they
panded time period and furthermeasures) of the first found that CEO charisma predicted organizational
author's doctoral dissertation, a portion of which ap
performance when managers perceived the envi
peared in the 1994 Academy ofManagement Best Paper ronment to be more uncertain. Tosi, Mi
Similarly,
Proceedings. We would like to thank the following indi and Yammarino
sangyi, Fanelli, Waldman, (2004)
viduals for their help on this project: Charles Hill, Tom
found no direct relationship between CEO cha
Jones, Patricia Kelley, Paul Collins, and Bob House. Fi
risma and organizational performance as measured
nancial resources were provided by Accenture, theDavid
Berg Center forEthics and Leadership at theUniversity of by shareholder return or return on assets but found
a positive effect of perceived environ
Pittsburgh, and the Chief Executive Leadership Institute moderating
at Yale University. We would also like to thank our mental uncertainty for the relationship between
CEO charisma and shareholder return.
Finally, un
associate editor and three anonymous reviewers for their

insightful comments. like the other two studies, Waldman, Javidan, and
161

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 162 Academy ofManagement Journal February

Varella (2004) found that CEO charisma was related In a recent editorial, Eden emphasized the impor
to subsequent performance as mea tance of replication studies, particularly those utiliz
organizational
sured by net profit margin and return on equity, but ing different samples and measures, inmoving theory
these authors found no support for a moderating forward. He wrote as follows: "Lykken (1968) stressed
effect of perceived environmental uncertainty. the importance of constructive replication, defined as
Table 1 summarizes the similarities and differ research that tests the same hypothesized relation
ences among these various studies and the study the same theoretical constructs as a
ships among
in these earlier but varies the
reported in this article. The similarities given study 'operationalization'
studies include their general research question (i.e., of those constructs. Confirmation of the same hypoth
the effect of CEO charisma on organizational per esis using different methods strengthens confidence
formance), type of informants (i.e., top manage in the validity of the hypothesized relationships"
ment team [TMT] members), and samples (i.e., pub (2002: 842; emphasis in original). Our study is a con
lic U.S. and Canadian firms). However, there are structive replication in that it tests the same hypoth
many differences, including sample size, number esized relationships earlier studies have tested,
of participants, response rates, measures, and time among the same theoretical constructs, but it utilizes
frames. For example, the current study differs from a different sample and varies the operationalization
the rest in the instrument used to measure CEO of the constructs; for instance, we have a similar, but
charisma and the measure of environmental uncer different,measure of CEO charisma, and an objective
measure of environmental uncertainty. Furthermore,
tainty. The current study uses a variety of organi
zational performance measures and, like Waldman this study overcomes many of the methodological
et al. (2001), employs a pre- and postsurvey design issues problematic in earlier studies through the use
with multiple controls. of a larger sample of firms; more respondents per
as to firm; a comprehensive ac
There are several potential explanations portfolio of perceived,
are mixed. First, as and stock-market-based measures of
why the results of prior research counting-based,
the authors of these studies noted, their studies had organizational performance; and rigorous economet
small samples and few responses per firm. These
ric methods. Such methodological improvements al
limitations prevailed because primary data on low us to have greater confidence in the results, be
CEOs is very difficult to collect. Second, theories of cause our models have greater statistical power and
this complex relationship have included a number less susceptibility to (1) a TMT member's idiosyn
of mediating and moderating variables (Canella & cratic rating of CEO charisma, (2) the peculiarities of
Monroe, 1997; Waldman & Yammarino, 1999), any particular mea
organizational performance
many of which are also difficult to disentangle, sure, and (3) underspecification.
measure, and test. Third, methodological differ
ences may contribute to the mixed findings. The
different results of prior studies could also be the THEORY AND HYPOTHESES
measures of organizational
product of the different Performance and Subsequent
Organizational
performance employed. For example, Waldman of CEO Charisma
an Perceptions
and his coauthors (2001) used just one measure,
accounting measure of performance. Also, neither Kelley's (1967) general attribution model, along
Tosi and his colleagues (2004) nor Waldman and with more specific attribution theories and re
his colleagues (2004) controlled for prior organiza search on leadership (e.g., Calder, 1977; Meindl,
tional performance, although earlier research Ehrlich, & Dukerich, 1985; Shamir, 1992) have pro
et al., 2001; Virany, Tushman, & Ro vided evidence that when organizations are per
(Waldman
manelli, 1992) emphasized the need for such a ceived as having performed well, others will at
control. Fourth, none of the earlier studies investi tribute that success to the leaders of those
them with
gated whether objective and subjective measures of organizations and will endow positive
prior organizational performance
were associated attributes. That said, Waldman, Bass, and Yam
with TMT perceptions of CEO charisma. Waldman marino (1990) noted that causation in the relation
and colleagues (2004) based their conclusion that ship between leader and organizational effective
charisma was not associated with prior organiza ness is still an open question. It is possible that
tional performance on simple correlations without charismatic leadership leads to organizational per
for firm size and CEO tenure. Their formance, but it is also possible that organizational
controlling
measure of performance is also potentially noisy, performance causes attributions of charisma to be
over a to a leader et 1985). Waldman
because they measured prior performance applied (Meindl al.,
five-year period, regardless of when a CEO had and his coauthors (1998) noted that future research
assumed that title. must be performed before any definitive statements

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
2006 Agle, Nagarajan, Sonnenfeld, and Srinivasan 163

TABLE 1
of Empirical Studies of CEO Charismatic Leadership and Organizational Performancea
Comparison

Waldman, Ramirez,
House, and Puranam

Study Current Study (2001) Waldman (2004) Tosi (2004)

128 public U.S. firms 48 public U.S. firms in 69 public U.S. and Canadian 59 public U.S. firms in
Sample
in various various industries firms in various various industries.
industries. (only 32 with CEO industries.
in place for entire

period).

Informants team Top management team Top management team

Top management Top management
team members. members. members. members.

Number of informants 770 survey number of Fewer than 193 survey 112 survey
Unspecified
respondents; survey respondents respondents; average of respondents; average
average of 6 per (estimated as 125); fewer than 3 per CEO. of fewer than 2 per
CEO (at least 3 per average of fewer CEO.

CEO). than 3 per CEO.

rate 26 percent of 500 20 percent of 28 percent of executives 37 percent of

Response
executives
CEOs agreed to executives surveyed. surveyed. surveyed.
participate,
constituting 80

percent of
executives

surveyed.

Measure of Adaptation of scale Multi-Factor Multi-Factor Leadership Multi-Factor

charismatic by Podsakoff et al. Leadership Questionnaire (Bass & Leadership
leadership (1990). Questionnaire (Bass Avolio, 1990). Questionnaire (Bass
& Avolio, 1990). & Avolio, 1990).

Measures of organiza Accounting-based: Accounting-based: Accounting-based: Industry Accounting-based:

tional performance Industry-adjusted Industry-adjusted adjusted ROA, NPM, and Industry-adjusted
ROA, ROS (NPM), NPM. sales growth. ROA.
ROE, and sales

growth.
Stock-market-based: Stock-market-based:

Industry-adjusted Industry-adjusted
stock returns. stock return.

Subjective: TMT
evaluations of firm

performance.

Measure of Objective: Volatility Subjective: Perceived Subjective: Perceived Subjective: Perceived

uncertainty of stock returns environmental environmental uncertainty environmental

(Core, Holthausen, uncertainty (Khandwalla, 1976). uncertainty
& Larcker, 1999). (Khandwalla, 1976). (Khandwalla, 1976).

Control variables CEO tenure, organiza CEO tenure, organiza CEO tenure (before and Organization size, time
tion size, and prior tion size, prior firm after), organization size. trend.
firm performance. performance.

Repeated-measures Pre- and postsurvey Pre- and postsurvey Only postsurvey period Only postsurvey period
measures. No control for measures. No control
design period accounting period accounting
and stock-market based performance prior firm performance. for prior firm
based measure. Control for performance.
performance
measures. Control priorfirm
for prior firm performance.
performance.

a
Waldman et al. (2001) in the Academy ofManagement Journal; Waldman et al. (2004) and Tosi et al. (2004) appeared in the
appeared
Leadership Quarterly.

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 164 Academy ofManagement Journal February

can be made about Thus, we ex

this relationship. dence that changes can be positive. Shamir and
pect perceptions of CEO charismato be higher Howell (1999) noted that the vision of charismatic
when an organization has performed well. leaders helps followers to see opportunities in
change and provides them with the hope and con
Hypothesis 1. Organizational
performance and fidence in the future that allow them to mobilize
subsequent top management team perceptions their energy in pursuit thereof. Finally, obligation
of CEO charisma have a positive relationship. inertia (commitments to constituencies) can be
overcome through leaders' ability to change current
contractual (both legal and social) relationships
Charismatic Leadership and Subsequent
with various stakeholders. Thus, when top manage
Organizational Performance
ment team members perceive their CEO as charis
Although several theories of charismatic, trans matic, we expect higher subsequent organizational
formational, or visionary leadership and organiza performance. Therefore:
tional performance have been offered (e.g., Bass,
& Kanungo, Hypothesis 2. Top management team per
1985; Burns, 1978;
Conger 1987;
CEO charisma and or
House, 1977; Shamir, House, & Arthur, 1993), in an ceptions of subsequent
have a
integration of these theories House and Shamir ganizational performance positive
(1993) proposed that charisma is the central con relationship.
cept in each. Recent work has
suggested that the
positive relationship between charismatic leader
CEO Charisma, Organizational Performance, and
ship and performance found in earlier studies will
Environmental Uncertainty
also hold true at the strategic (CEO) level (Waldman
et al., 2004). The essence of extant theories is that As Waldman and his coauthors (2001), Shamir
charismatic leaders may influence organ
strategic and Howell (1999), and others have noted, environ
izational performance because of their ability to mental uncertainty can affect relationships be
overcome the three major inertial forces (cognitive, tween various predictors of corporate leadership
motivation, and obligation [Gersick, 1991]) that and organizational performance. Because the per
keep organizations from successfully adapting to ceived risk of organizational failure is greater when
new environments (Agle, 1993); their ability to in the environment is more unstable or uncertain,
spire and motivate employees and other corporate CEOs' discretion may be enhancedduring periods
stakeholders (Agle, 1993; Shamir et al., 1993); their of environmental uncertainty (Hambrick & Finkel
ability to create cohesion among the members of stein, 1987). With increased discretion, the influ
organizations with their vision (Waldman & Yam ence of their leadership is magnified.
marino, 1999) and power (Finkelstein & Hambrick, Similarly, leaders in unstable, risky, or crisis sit
1996); the influence their values have on their stra uations may also take on greater symbolic impor
tegic choices (Finkelstein & Hambrick, 1996); the tance. Factors such as greater follower effort and
cascading effect their leadership has on followers greater follower cooperation that mediate the rela
throughout their organizations; and
their willing tionship between charisma and organizational per
ness to implement strategic change (Waldman & formance depend on the followers' willingness to
Yammarino, 1999). Indeed, Waldman and Yam accept the influence of their leaders. Under condi
marino (1999) provided a full theoretical model tions of uncertainty and crisis, followers feel the
demonstrating how CEO charismatic leadership need for greater direction and guidance (Bass,
leads to organizational performance through both 1990), and their inclination to accept influence
close and distant relationships. may be greater (Shamir & Howell, 1999). Thus,
Adapting to a changing environment by over willingness to follow a leader may be more pro
coming organizational inertia is viewed as a partic nounced in unstable environments. This emphasis
important conditionfor improving organiza on the importance of uncertainty corresponds with
ularly
tional (Tichy & Devanna, 1990). sociological perspectives on the importance of cri
performance
Charismatic leaders overcome cognitive inertia (in ses in the demonstration of charisma and its effects
ability to think outside one's own schema) because (Beyer & Browning, 1999).
their strong values shape choices concerning strat Finally, Pawar and Eastman (1997) suggested
egy. For instance, these leaders create exciting vi that firms in turbulent environments tend to be
sions of the future and promote unconventional dominated by their boundary-spanning units, as
approaches. Motivational inertia opposed to their dominant technical cores. Because
problem-solving
(desire to avoid change) can be overcome through a the tasks in these types of units are not well defined
leader's ability to provide followers with confi and thus cannot be standardized, such organiza

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 2006 and Srinivasan 165
Agle, Nagarajan, Sonnenfeld,

tions yield themselves to a greater influence from to the factor scores, a global variable labeled "total
their charismatic CEOs. Therefore, CEO charis charisma" was computed; itwas the sum of the raw
matic leadership will be more related to scores of all the questionnaire items. Because of
highly
organizational performance when an organization's strengths and weaknesses of various interrater
environment is uncertain and volatile. agreement measures, we computed both the intra
class correlations statistics ICC(l) and ICC(2), and
Hypothesis 3. The positive relationship be
the interrater agreement statistic &
tween CEO charisma and subsequent (r*wg^ ) (Lindell
organi
Brandt, 1999; Lindell, Brandt, & Whitney, 1999;
zational performance is stronger under condi
Winer, 1962). The Appendix shows the high values
tions of high uncertainty.
of these measures for the five factors. In addition,
the alpha coefficient, ICC(l), ICC(2), and for
r*wga)
METHODS the global charisma scale used in this study were
.94, .34, .90, and .80, respectively. Overall, the very
Sample Characteristics and Questionnaire
Administration high interrater agreement and scale reliability jus
tified the combination of individual manager's re
Our questionnaire data were gathered in 1992. The sponses into a single measure of charisma for each
sample for this study was drawn from the Monitor CEO. Because statistical analyses utilizing the five
Publishing Company's Financial 1000 and Corporate charismatic leadership factor scores and the global
1000 Yellow Books. We randomly chose 500 CEOs charismatic leadership scale were very similar for
from the sample of firms appearing in these volumes each of the hypotheses tested, this article reports
to participate in this research. Of these, 128 CEOs the results of the analyses using the global charis
agreed to participate, a number representing a 26 matic leadership scale.
percent response rate. The members of the top man Operationalizing organizational performance is al
agement team of each firm were sent the question ways a challenge because the strategy, accounting,
naire with an accompanying cover letter indicating and finance literatures suggest that both accounting
their CEO's support for the project and assuring re and market-based measures sufferfrom measurement
spondent anonymity and confidentiality of the infor and controllability issues, and that these measures
mation. Out of 960 questionnaires sent, 770 were may not converge to represent the same construct of
returned (an 80 percent response rate), resulting in an organizational performance (Fryxell & Barton, 1990).
average of six respondents per firm. We compared Hence, we used both primary and secondary sources
our sample (128 firms) with the nonresponding (372) of data, and objective and perceived modes of assess
firms on a number of dimensions, such as age, size, ment, as suggested by Venkatraman and Ramanujam
abnormal stock returns, and capital intensity, and (1987). We utilized themeasure of perceived organi
CEO tenure. The i-tests demonstrated no significant zational performance developed by Ramanujam,
differences. The CEOs in the sample had an average Venkatraman, and Camillus (1986), which they based
tenure of 6.6 years at the time the questionnaire was on four questions assessing a firm's increase in sales,
administered and stayed for an average of 4.5 years earnings, market share, and return on investment
after the questionnaire administration. The sample (ROI) in relation to its competitors since the time of
firms averaged 55 years of age, $6.5 billion in assets, its CEO's ascension (measured on a seven-point Lik
and 16,000 employees, and they covered the spec ert scale ranging from "much better" to "much
trum of industries. worse"). Factor analysis utilizing principal compo
nent analysis also placed these items into a single
factor with respective loadings of .88, .89, .87, and
Measures
.88. Our measure of perceived organizational perfor
We measured the various theorized characteris mance, constructed as the average of the sum of the
tics of charismatic leaders using a refinement of a raw scores for these four items, had an appropriate

larger instrument developed by Podsakoff, Mac level of scale reliability and interrater agreement [a =
= = =
Kenzie, Moorman, and Fetter (1990; see Agle [1993] .90; ICC(l) .69; ICC(2) .89; r*wg .86).1
on the refinement). To encourage a high
for details We further operationalized organizational perfor
response rate from these top executives, we limited
our survey to one page; this design criterion pre
cluded the use of the multifactor questionnaire 1
We did not throw out any outliers in the construction
(MLQ; Bass & Avolio, 1990), which at the time of of the TMT subjective performance and charisma mea
our survey was available only in its entirety. sures. However, we also constructed these measures after
The Appendix displays five major factors that deleting firms with three or fewer respondents. The re
emerged from this refined instrument. In addition sults did not change.

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 166 Academy ofManagement Journal February

manee by constructing a stock return measure and Waldman et al., 2001, 2004), we utilized an objec
various accounting measures. In general, CEOs may tive measure of environmental uncertainty?firm
have greater control over (internal) accounting risk?which is commonly used in accounting, fi
measures than over (external) stock market mea nance, and strategy research (e.g., Core, Holt
sures because of their ability to control levels of hausen, & Larcker, 1999; Gray & Canella, 1997;
investment, adjust discretionary accruals, and shift Miller, Wiseman, & Gomez-Mejia, 2002) and that
earnings across periods as assessment
(Murphy, 1999), and be may be interpreted the collective of
cause stock market measures are subject to greater uncertainty by stock market participants as op
levels of extraneous noise than are accounting mea posed to the assessment ofmanagers. More impor
sures. However, itmust be noted that there are no tantly, high stock volatility, perceived as
high risk
theoretical reasons to suppose that a CEO may have by the market, may also raise the cost of capital,
greater leverage over one accounting measure than generating further financial uncertainty. We used
another. For example, accounting measures that the volatility of stock returns during the postques
have been popular with researchers, such as return tionnaire2 (performance) period
as our proxy for
on assets, equity, or sales, differ only in their de uncertainty and estimated it as the annualized stan
nominators. CEOs potentially can influence all dard deviation of daily stock returns of a firm ad
these measures to a similar extent. Because all in justed for average industry volatility. Following
dicators of organizational performance are likely to prior studies (Virany et al., 1992; Waldman et al.,
be imperfect, we considered several to verify that 2001) that have stated that organizational size, CEO
the hypothesized relationships were robust across tenure, and prior performance should be controlled
variations in measurement, thus increasing the for in research seeking to relate CEO characteristics
value of constructive replication (Eden, 2002). to organizational performance, we included these
To construct the stock return measure, we first control variables: organizational size, measured as
measured the cumulative stock return of each firm the value of total assets in 1991 (the year immedi
over the prequestionnaire period (from the an ately prior to the questionnaire date), CEO tenure3
nouncement of a new CEO until the questionnaires (measured as the number of years the CEO held that
were completed in 1992) and the postquestionnaire position prior to the questionnaire date), and prior
(from the time the questionnaires were com (pre-92) industry-adjusted organizational perfor
period
pleted in 1992 until the same CEO's departure or mance. To reduce problems associated with multi
until the end of 1999, whichever was first).We then collinearity, as did Waldman et al. (2001), we used
subtracted the corresponding average cumulative Cronbach's transformation by centering the cha
stock return of a firm's industry (defined by four risma and uncertainty variables around theirmeans
digit SIC code) to obtain the industry-adjusted before forming the interaction term.
stock return and then annualized it.
We next constructed three accounting-based RESULTS
measures following prior studies in accounting re
Table 2 presents the correlations among the vari
search: (1) return on assets (ROA), computed as
ables of interest. Charisma is significantly corre
income before extraordinary items divided by the
lated with the perceived performance measure <
total assets of a firm in a year, (2) return on equity (p
<
as income before extraordinary .01), prequestionnaire period stock return (p .05),
(ROE), computed <
ROA (p .05), and sales growth (p < .01). How
items divided by the total stockholder equity of a
firm in a year, and (3) return on sales, or profit
ever, for the postquestionnaire period, it is signifi
cantly correlated (p < .05) only with sales growth.
margin (ROS), computed as income before extraor
CEO tenure is associated with some measures of
dinary items divided by the net sales in a year (this
organizational performance. However, all the cor
measure corresponds to the net profit margin
relations must be interpreted with caution because
[NPM] used byWaldman et al. [2001, 2004]). We associations and do not
also constructed a growth measure for each firm as they represent pairwise
control for other factors that may impact organiza
the percentage change in sales dollars (or revenues) tional performance.
from one year to the next. Data from the COMPU
STAT database were obtained for each measure for
each firm for the years of our sample period. We 2
Our results were similar when the prequestiormaire
corrected all these measures by subtracting indus was used to estimate uncertainty.
period
3
try averages and averaged each industry-adjusted The minimum CEO tenurewe allowed was one year.
annual measure over the pre- and postquestion The results remained qualitatively similar when CEOs
naire periods. with this tenure requirement were either included or
Unlike earlier researchers (Tosi et al., 2004; excluded.

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 2006 Agle, Nagarajan, Sonnenfeld, and Srinivasan 167

TABLE 2
Descriptive Statistics and Correlations'1

Variable Mean s.d. 1 2 3 46 7 8 5 9 10 11 12 13 14

1. CEO charisma 115.30 14.90

2. Perceived 17.06 4.84 .49**

performance,
before 1992
3. Industry-adjusted 3.89 10.41 .14* .19**
ROA, before 1992
4. Industry-adjusted 3.76 16.51 -.11 -.01 .24*
ROE, before 1992
5. Industry-adjusted 7.16 13.89 -.03 .03 .53** .12
ROS, before 1992
6. Industry-adjusted 4.78 15.97 .24** .38** .01 -.02 -.02
sales growth,
before 1992
7. Stock returns, 3.77 19.53 .17* .24** .08 .07 .02 .17**
before 1992
8. Industry-adjusted 5.94 12.00 .11 .07 .56**-.02 .44** .06 .02
ROA, after 1992
9. Industry-adjusted 7.54 19.48 -.08 -.11 .01 .16 .02 -.06 -.01 .15
ROE, after 1992
10. Industry-adjusted 10.61 17.05 -.12 -.20 .33** .03 .63**-.12 -.05 .62** .06
ROS, after 1992
11. Industry-adjusted 4.44 15.44 .22** .25** .04 -.01 -.10 .06 .08 .16* -.02 -.03
sales growth,
after 1992
12. Industry-adjusted 8.79 57.39 .05 .08 .22** .04 .10 .02 .02 .33** .14 .06 .26**
stock return, after
1992
13. Uncertainty 0.13 0.04 .04 .08 -.08 -.10 -.02 -.05 .01 -.05 -.05 -.06 .04 -.02
14. Total assets, in 6.50 12.63 .07 -.05 -.05 .07 .01 .13 -.06 -.08 -.02 -.02 -.05 -.03 -.21**
billions
15. CEO tenure, in 6.64 6.05 .14 .29** .13 .18* .05 .08 .22**-.04 .08 -.07 -.05 .15* .01 -.09

years

an = 128.
* <
p .05
< .01
**p

To test Hypothesis 1,we regressed CEO charisma Prior organizational performance is the only signif
separately on the perceived measure of prior organ icant variable that explains much of subsequent
izational performance and also on four objective organizational performance, except for the model
(financial) measures of prior organizational perfor with stock return, forwhich prior-period organiza
mance, after controlling for organizational size and tional performance is not significant. These results
CEO tenure. Table 3 presents the relation between suggest that subsequent organizational perfor
CEO charisma and various measures of prior per mance is not associated with charisma or the inter
formance. The results suggest that the perceived action between charisma and uncertainty, thus fail
measure of organizational performance is associ ing to support either Hypothesis 2 or 3. The results
ated with subsequent perceptions of CEO charisma. of hierarchical regression analyses (not shown) also
Further, some objective measures of organizational suggest that charisma and environmental uncer
performance?that is, stock return, ROA, and sales tainty do not predict any significant variance in
growth?are significantly associated with subse subsequent organizational performance and are not
quent perceptions of CEO charisma, supporting Hy associated with it, even when they are included
pothesis 1. Table
4 presents the results of regres one at a time in the regression models. We also
sion analyses (which are similar to those of measured organizational performance and environ
Waldman et al. [2001]) for tests of Hypotheses 2 mental uncertainty over short periods; specifically,
and 3. The explanatory power of the models (mea a one-year period before and after the questionnaire
sured as adjusted R2) with various measures of date. Our results did not change in the presence of
organizational performance ranges from .03 to .38. these alternative measures. Our reported results are

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 168 Academy ofManagement Journal February

TABLE 3
The Regression of CEO Charisma on Perceived, Stock Market, and Accounting Measures of
Prior Organizational Performance41

Predictor Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Total assets .07 .05 .04 .04 .04 .01

CEO tenure .02 .13 .12 .17 .15 .09
Perceived performance, .55**
before 1992

Industry-adjusted Stock .27*

Return, before 1992

Industry-adjusted ROA, .17*

before 1992

Industry-adjusted ROS, .12

before 1992

Industry-adjusted ROE, -.08

before 1992

Industry-adjusted Sales .37*

growth, before 1992

R2 .31 .06 .04 .03 .02 .09

Adjusted R2 .29 .04 .03 .01 .01 .08

F 18.60** 3.42** 1.67 1.32 0.83 3.66**
127 125 126 124 125 125

a
Standardized regression coefficients are shown.
b
All the results are based on a sample that was obtained after deletion of observations that had r-Student values of more than 2 (since
our size was 128, as can be seen from the table, up to four outliers were deleted for different models).
sample
*
p < .05
< .01

based on the sample after deleting outliers defined DISCUSSION

as those observations with r-Student values ofmore
Theoretical and Practical Implications
than two (Belsley, Kuh, & Welsch, 1980). Our re
sults were similar when we did not delete the out In this research, we found evidence of (1) a rela
liers. Overall, our findings are robust to various tionship between perceptions of CEO charisma and
alternative specifications. perceived measures of prior organizational perfor

TABLE 4
Regression of Subsequent Organizational Performance on Prior CEO Charismaa

Industry-Adjusted Industry-Adjusted Industry-Adjusted Industry-Adjusted Industry-Adjusted

Predictor Stock Return ROA ROS ROE Sales Growth

Total assets .01 -.11 .02 -.04 -.24

CEO tenure .04 -.12 .01 .08 -.08
Prior performance, .02 .49** .75* .11 .43*
before 1992

Uncertainty -.02 -.08 -.04 .02 -.03

Charisma .03 .11 -.07 .02 .06
Charisma X -.03 .02 .01 .02 .08

uncertainty

R2 .04 .27 .41 .06 .24

Industry-adjusted R2 .03 .23 .38 .04 .21

F .88 7.19* 12.77* 1.63 5.69*
127 125 125 124 125

a
Standardized coefficients are shown.
regression
b
All the results are basedon a sample that was obtained after deleting observations that had r-Student values of more than 2 (since our

sample size was 128, as can be seen from the table, up to four outliers were deleted for different models).
< .01
**p

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 2006 Agle, Nagarajan, Sonnenfeld, and Srinivasan 169

manee, (2) a relationship between perceptions of different effect from that found in studies that used
CEO charisma and some objective measures of perceived measures. A top management team's
prior organizational performance, and no relation own perception of the uncertainty of its environ
ship with other objective measures, and (3) no re ment or level of firm crisis might lead TMT mem
lationship between charisma and subsequent objec bers to be more accepting of change (Tichy &
tive measures of organizational performance, even Devanna, 1990) than an objective (stock market)
after the moderating effect of environmental uncer assessment of the firm's uncertainty.
tainty was considered. Some of these findings are Some theoretical reasons also might explain why
similar and the rest are contrary to the findings of the various studies have failed in general to find a
other recent studies examining these relationships. direct relationship between CEO charisma and or
As illustrated in Table 1 and as discussed in the ganizational performance. First, empirical testing
introduction, some of the reasons for themixed find of this relationship is in its early stages. Charis
ings may be the methodological differences in the matic leadership has been theorized on a number of
studies. Another factor, not yet introduced, is the dimensions, not all of which have been tested in
potential for same-source and same-method biases. In the research. For example, notions of behavior nov
our study, when testing Hypothesis 1 using perceived elty (unconventionality) and challenges to the sta
measures of charisma and organizational perfor tus quo (Conger & Kanungo, 1998) have not been
mance, we found that same-source bias existed and tested. Shamir and Howell (1999) suggested that
minimized it by using the split-sample technique the effectiveness of charismatic leaders is greater in
(Rousseau, 1985). However, same-method bias might the early and later stages of organizational life cy
still have existed. The presence of a stronger relation cles, a potential moderator not yet included in the
ship between CEO charisma and perceived prior or research. Waldman and Yammarino's
(1999) model
ganizational performance than between CEO cha of the relationship
between charisma and or
CEO
risma and objective measures of prior organizational ganizational performance included a number of
performance obtained from sources other than the mediating and moderating variables. To better un
questionnaire exposes this bias. It is possible that past derstand the causality of this relationship, further
findings may also have had this limitation. Waldman research needs to include such mediating and mod
and his colleagues (2001) noted that none of the stud erating variables to examine the black box contain
ies meta-analyzed by Lowe and his coauthors (1996) ing CEO charisma and organizational performance.
used hard measures of actual financial performance. An example of such research is a longitudinal ran
In addition, Waldman et al. (2001) and Tosi et al. domized field experiment by Dvir, Eden, Avolio,
(2004) found a significant interactive effect for per and Shamir (2002), which showed that transforma
ceived environmental uncertainty, but their per tional leadership had positive effects on follower
ceived measures of charisma and uncertainty came development and performance.
from the same source and through the same method. Also, leaders in small organizations generally
Neither of these studies showed a direct relationship have
greater latitude of action (Hambrick & Finkel
between a perceived measure of charisma and an stein, 1987) than the leaders sampled in existing
objective measure of performance. Only after adding studies. Thus, in a sample with smaller firms, one
the interaction between perceived charisma and per may find a greater effect of CEO charisma on organ
ceived uncertainty did they find any relationship. izational performance. In fact, Koene and his col
We also gained insight from the different mea leagues (2002) found that managerial charisma was
sures used in these studies. Through use of the correlated with the performance of small supermar
MLQ, Waldman and his coauthors (2004) substi kets but was not correlated with that of large super
tuted intellectual stimulation for charisma and markets. Also, the function of CEOs in large firms is
found a significant moderating role for environ often different from that in smaller firms. These
mental uncertainty in the relationship between CEOs are often more involved with public relations
CEO charisma and organizational performance. activities and with capital allocation and corporate
This resultsuggests that researchers should look development (e.g., mergers, acquisitions, divesti
more closely at the intellectual stimulation portion tures) than with the actual "running of the busi
of charismatic leadership as a potential mediator of ness" (Garten, 2001). In fact, the role of the CEO in
CEO charisma and organizational performance. In many multidivision firms is to serve as a substitute
tellectualstimulation might be a particularly im for the capital markets, evaluating the performance
portant aspect of leadership when business condi of the leaders of the corporate divisions and allo
tions are changing and followers need to be able to cating capital and changing leadership accordingly;
think outside the box. Similarly, our use of an this is a more evaluative than inspirational role. In
objective measure of uncertainty might have had a such firms, the line operating executives' level of

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 170 Academy of Management Journal February

charisma might actually be more important than prior research that gave solid evidence of relation
the CEOs'. In addition, in such large firms, boards ships between leader charisma and leadership ef
of directors and the financial community also re fectiveness can be generalized to all situations.
strain the CEOs' power. Future studies might in
clude as variables the degree to which a CEO is
Limitations and Conclusion
involved in the actual operations of a firm, as well
as the degree towhich this research used a much
the firm's board of directors Although larger sam
and the financial community constrain his or her ple and more measures of the phenomena under
power. However, given the absolute complexity investigation than similar studies, it is not without
and diversity of the variables that contribute to its limitations. Primary among these is the self
organizational performance in large firms, it
may be selected nature of the sample. We were unable to
very difficult to find an effect of CEO charisma on determine if the sample was skewed for CEO cha
organizational performance. risma and, thus, it is possible that our sample was
Alternatively, itmay be that charismatic leader overly populated with CEOs who agreed to partic
ship at the CEO level is not more effective leader ipate in the research because considered them
they
ship. Yukl (1999) suggested a number of reasons selves to be charismatic. Also, as noted earlier, this
that this might be the case, such as inappropriate, sample consisted solely of individuals from large
radical changes resulting from charismatic leader firms; this restriction of range might have con
ship, misuse of power by charismatic leaders, and stricted the relationships revealed.
polarization of organization members. Conger and In addition, in spite ofmeasuring organization
Kanungo (1998) andMaccoby (2000) proposed that al performance at two different times?both be
the narcissistic tendencies of charismatic leaders fore and after questionnaire administration?it
might neutralize the benefits of their leadership. was still not possible for us to completely disen
Such negative consequences of charisma have not tangle the causality in the relationship between
been tested in extant research. CEO charisma andorganizational performance.
To answer the question posed in the title of this Earlier works using the same data (Agle, 1993;
article, "Does CEO charisma matter?," our evidence Agle & Sonnenfeld, 1994) reported a relationship
suggests that CEOs who are perceived to be more between CEO charisma and organizational per
charismatic appear to be perceived as more effec formance, but their authors acknowledged that
tive. In this subjective sense, CEOs matter. How given the sequence of measures in their research
ever, despite evidence of possible initial effects, the (measurement of organizational performance pre
lack of long-term corroborating evidence from ob ceding measurement of CEO charisma), they
jectively assessed CEO performance suggests that could not demonstrate this causal direction em
the search for charismatic CEOs may be based more pirically. Yukl's (1999) argument that charisma is
on implicit theory or halo effects than on solid transitory, and Weber's (1947) classic argument
evidence that charisma really does make CEOs that charisma cannot remain stable, but becomes
more effective. The present study is not able to either traditionalized or routinized over time,
provide stock analysts, investors, and boards of both suggest that CEO charisma's impact on or
directors with evidence that CEO charisma neces ganizational performance is greatest earlier in a
sarily benefits future financial performance, even CEO's tenure. It could also be evidence of the
under conditions of uncertainty, and suggests that "low-hanging fruit" effect, whereby more acces
they need to be cautious when considering the sible interventions are addressed early on. Thus,
potential benefits of charismatic leaders. In the because our measurement of charisma took place
same vein, the title of Khurana's (2002) book exam on average after six years of CEO tenure, one may
executive succession, a still argue that the significant, positive associa
ining Searching for Corpo
rate Savior: The Irrational Quest for Charismatic tion we found (see Table 2) between charisma
CEOs, refers skeptically to the phenomenon of and prequestionnaire organizational performance
firms' looking for charismatic CEOs. He suggests measured over the earlier part of CEO's
that directors have been under the illusion that tenure is evidence that charisma matters for or
charisma is the key element of leadership perfor ganizational performance. Those who would sup
mance and writes that they should be looking for port this interpretation could refer to research
specific firm-related leadership skills in potential suggesting that TMT members, who are very close
CEOs instead of charisma. The efficacy of a leader's to their firms' CEO, are less likely to attribute
charisma for directly shaping organizational perfor their perceptions of CEO charisma to prior organ
mance may therefore vary by the leader's hierarchi izational performance (Shamir, 1995).
cal level; this idea raises questions about whether To go beyond this study to investigate the impact of

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 2006 and Srinivasan 171
Agle, Nagarajan, Sonnenfeld,

on organizational Burns, J.M. 1978. Leadership. New York: Harper & Row.
CEO charisma performance, one
would need to create a design similar to the one Calder, B. J. 1977. An attribution theory of leadership. In
House, Spangler, and Woycke (1991) used to test for B. M. Staw & G. R. Salancik (Eds.),New directions in
the effects of presidential leadership. In such a cha organizational behavior: 179-204 Chicago: St. Clair.
risma-performance study, CEO charisma would be
Canella, A. A., & Monroe, M. J. 1997. Contrasting per
measured before each CEO ascended to that position. on leaders: Toward a more realis
spectives strategic
A researcher could meet this difficult challenge by, tic view of top managers. Journal ofManagement,
for instance, measuring the charisma of a large group 23: 213-237.

of top executives likely to rise to the CEO level within & Kanungo, R. N. 1987. Toward a behav
Conger, J. A.,
a short time period and then testing for organizational ioral theory of charismatic leadership in organiza
performance effects after they become CEOs. Alterna tional settings. Academy ofManagement Review,
tively, one could measure charisma immediately after 12: 637-647.

the appointment of a CEO or via archival materials & Kanungo, R. N. 1998. Charismatic lead
Conger, J.A.,
(articles, speeches, or videotapes) generated before an ership in organizations. Thousand Oaks, CA: Sage.
individual assumed the CEO title. All these designs
Core, J. E., Holthausen, R. W., & Larcker, D. F. 1999.
would help researchers explore the short-term versus
governance, chief executive officer com
Corporate
long-term impact of charisma on organizational per pensation and organizational performance. Journal
formance while the confounding effect of 51: 371-406.
eliminating of Financial Economics,
prior organizational performance on perceptions of
& Shamir, B. 2002.
Dvir, T., Eden, D., Avolio, B., Impact
charisma. The present that the signifi
results show
of transformational leadership on follower develop
cant relationship between prior performance and rat ment and performance: A field experiment. Acad
ings of CEO charisma is not complemented by evi emy ofManagement Journal, 45: 735-744.
dence of any relationship between ratings of CEO
Eden, D. 2002. scientific
Replication, meta-analysis,
charisma and subsequent performance, yet the con
progress, and AMJ's publication policy. Academy of
clusion that CEO charisma does not affect perfor 45: 841-846.
Management Journal,
mance cannot be ruled out. Many unresolved issues
that could cast a different light on this relationship Finkelstein, S., & Hambrick, D. C. 1996. Strategic lead
remain. These be fruitful areas for future ership: Top executives and their effects. Minneap
may olis: West.
research.
G. E., & Barton, S. L. 1990. and con
Fryxell, Temporal
textual change in the measurement structure of fi
REFERENCES nancial performance: Implications for strategy re
search. Journal ofManagement, 16: 553-569.
Agle, B. R. 1993. Charismatic chief executive officers:Are
theymore effective?An empirical test of charismatic Garten, J. E. 2001. The mind of the CEO. New York:
leadership theory.Unpublished doctoral dissertation
Basic Books.

(publication no. AAT 9416983), University ofWash Gersick, C. J.G. 1991. Revolutionary change theories: A
ington, Seattle. Available fromUMI Proquest Digital multilevel exploration of the punctuated equilib
Dissertations, www.lib.umi.com/dissertations. rium paradigm. Academy of Management Review,

Agle, B. R., & Sonnenfeld, 1994. Charismatic chief exec

16: 10-36.

utive officers:Are theymore effective?An empirical

Gray, S. R, & Canella, A. A. 1997. The role of risk in
test of charismatic leadership theory. Academy of executive compensation. Journal of Management,
Management Best Papers Proceedings: 2-6. 23: 517-540.

Bass, B. M. 1985. Leadership and performance beyond D. C. 1989. Guest editor's introduction: Put
Hambrick,
New York: Free Press.
expectations.
ting topmanagers back in the strategy picture. Stra
Bass, B. M. 1990. Bass and StodgilTs handbook of lead tegic Management Journal, 10: 5-15.

ership: Theory, research and managerial applica Hambrick, D. C, & Finkelstein, S. 1987. Managerial dis
tions. New York: Free Press. cretion: A bridge between polar views of organiza
Bass, B. M., & Avolio, B. J. 1990. The Multifactor Lead tional outcomes. In L. L. Cummings & B. M. Staw

ership Questionnaire. Palo Alto, CA: Consulting (Eds.), Research in organizational behavior, vol. 9:
Psychologists Press. 369-406. Greenwich, CT: JAIPress.

Belsley, D. A., Kuh, E., & Welsch, R. E. 1980. Regression House, R. J. 1977. A 1976 theory of charismatic leader
diagnostics. New York: Wiley. ship. In J.G. Hunt & L. L. Larson (Eds.), Leadership:
& 1999. an
The cutting edge: 189-207. Carbondale: Southern
Beyer, J.M., Browning, L. D. Transforming
in crisis: Charisma,
Illinois University Press.
industry routinization, and sup
portive cultural leadership. Leadership Quarterly, House, R. J.,& Shamir, B. 1993. Toward the integration of
10: 523-530. transformational, charismatic and theo
visionary

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 172 Academy ofManagement Journal February

ries. InM. Chemers & R. Ayman (Eds.), Leadership Ramanajum, V., Venkatraman, N., & Camillus, J. C. 1986.

theory and research: Perspectives and directions: Multiobjective assessment of effectiveness of strate
81-107. San Diego: Academic Press. gic planning: A discriminant analysis approach.
House, R. J., Spangler, W. D., & Woycke, J. 1991. Person Academy ofManagement Journal, 29: 347-372.
ality and charisma in theU.S. presidency: A psycho Rousseau, D. M. 1985. Issues of levels in organizational

logical theory of leader effectiveness. Administra research: Multi levels and cross level perspectives.
tive Science Quarterly, 36: 364-396. In L. L. Cummings & B. M. Staw (Eds.), Research in
vol. 12: 43-72. Green
Howell, J.M., & Frost, P. J. 1989. A laboratory study of organizational behavior,

charismatic leadership. Organizational Behavior wich, CT: JAIPress.

and Human Decision Processes, 43: 243-269. D. 1995. The charismatic leader as narcis
Sankowsky,
sist: Understanding the abuse of power. Organiza
Kelley, H. H. 1967. Attribution theory in social psychology.
Nebraska Symposium onMotivation, 15: 192-240. tional Dynamics, 23(4): 57-71.

Khandwalla, P. N. 1976. The design of organizations. Shamir, B. 1992. Attribution of influence and charisma to
New York: Harcourt Brace Jovanovich. the leader: The romance of leadership revisited. Jour
nal ofApplied Social Psychology, 22: 386-407.
Khurana, R. 2002. Searching for a corporate savior:
The irrational quest for charismatic CEOs. Prince Shamir, B. 1995. Social distance and charisma: Theoret
ton,NJ: Princeton University Press. ical notes and an exploratory study. Leadership

Koene, B. A. S? Vogelaar, A. L. W., & Soeters, J. L. 2002. Quarterly, 6: 19-48.

Leadership effectson organizational climate and finan Shamir, B., House, R. J., & Arthur, M. B. 1993. The motiva
cial performance: Local leadership effect in chain or tional effectsof charismatic leadership: A self-concept
ganizations. Leadership Quarterly, 13:193-215. based theory.Organization Science, 4: 577-594.
Lindell, M. K, & Brandt, C. J. 1999. Assessing interrater Shamir, B., & Howell, J.M. 1999. Organizational and
agreement on the job relevance of a test:A compar contextual influences on the emergence and effec
ison of the CVI, T, rWG(7)and r*WGC^indexes. Journal tiveness of charismatic leadership. Leadership
of Applied Psychology, 84: 640-647. Quarterly, 10: 257-284.
Lindell, M. K., Brandt, C. J.,& Whitney, D. J. 1999. A
Tichy, N., & Devanna, M. 1990. The transformational
revised index of interrater agreement formulti-item leader. New York: Wiley.
ratings of a single target. Applied Psychological
Tosi, H. L., Misangyi, V. F., Fanelli, A., Waldman, D. A.,
Measurement, 23: 127-135.
& Yammarino, F. J. 2004. CEO charisma, compensa
Lowe, K. B., Kroeck, K. G, Sivasubramaniam, N, 1996.
tion, and firm performance. Leadership Quarterly,
Effectiveness correlates of transformational and trans
15: 405-421.
actional leadership: A meta-analytic review of the
N., & Ramanujam, J. 1987. Measurement of
MLQ literature.Leadership Quarterly, 7: 385-425. Venkatraman,
business economic performance: An examination of
M. 2000. Narcissistic leaders. Harvard Busi
Maccoby, method Re
convergence. Journal of Management
ness Review, 78(1): 68-78.
view, 11: 801-814.
Meindl, J.R, Ehrlich, S. B., & Dukerich, J.M. 1985. The M. 1992.
Virany, B., Tushman, L., & Romanelli, E. Exec
romance of leadership. Administrative Science
utive succession and outcomes in tur
30: 78-102. organizational
Quarterly,
bulent environments: An organization learning ap
Miller, J.,Wiseman, R. M., & Gomez-Mejia, L. R., 2002. 3: 72-91.
proach. Organization Science,
The fit between CEO compensation design and firm
Waldman, D. A., Bass, B. M., & Yammarino, F. J. 1990.
risk.Academy ofManagement Journal, 2002: 745
756. Adding to contingent-reward behavior: The aug
menting effect of charismatic leadership. Group &
K. 1999. Executive In O. Ash
Murphy, compensation.
Organization Studies, 15: 381-394.
enfelter & D. Card (Eds.), Handbook of labor eco
D. A., M., & Varella, P. 2004. Charis
nomics, vol. 3B: 2485-2566. Amsterdam: North Waldman, Javidan,

Holland. matic leadership at the strategic level: A new appli

cation of upper echelons theory. Leadership Quar
Pawar, B. S., & Eastman, K.K. 1997. The nature and
15: 355-381.
terly,
implications of contextual influences on transforma
tional leadership: A conceptual examination. Acad Waldman, D. A., Ramirez, G A., House, R. J., & Puranam,

Review, 22: 80-109. P. 2001. Does leadership matter? CEO leadership

emy ofManagement
attributes and profitability under conditions of per
Podsakoff, P. M., MacKenzie, S. B., Moorman, R. H., &
ceived environmental uncertainty. Academy of
Fetter, R. 1990. Transformational leader behaviors
Journal, 44: 134-143.
and their effects on follower's trust in leader, satis Management

faction, and organizational citizenship behaviors. Waldman, D. A., & Yammarino, F. J. 1999. CEO char

Leadership Quarterly, 1: 107-142. ismatic leadership: Levels-of-management and

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 2006 and Srinivasan 173
Agle, Nagarajan, Sonnenfeld,

levels-of-analysis effects. Academy of Manage Winer, B. J. 1962. Statistical principles in experimental

ment Review, 24: 266-285. design. New York: McGraw Hill.
Weber, M. 1947. The theory of social and economic Yukl, G 1999. An evaluation of conceptual weaknesses
organization [A.M. Henderson & T. Parsons, eds. & in transformational and charismatic leadership the
trans.] New York: Oxford. ories. Leadership Quarterly, 10: 285-405.

APPENDIX

Factor Cronbach's
Varimax-Rotated Factors Loadings Alpha ICC(l) ICC(2) '"B?F)

Dynamic leadership
Our chief executive officer. . . .92 .85 .91 .82
is dynamic. .83
has the ability to excite a group of people, .82
is charismatic, .82
when communicating, drives to motivate .76
with
every word, story, and inflection,
communicates an exciting vision of the .71
future of the organization,

paints an exciting picture of the future of .67

the organization.

Exemplary leadership .90 .78 .89 .82

Our chief executive officer.. .
is trusted by members of the organization, .79
is respected by members of his/her top .77

management team,
sets a good example, .75
a good model forme to follow, .70
provides
will not sacrifice his/her moral standards, .65
has a clear understanding of where we are .61

going,
gives us the feeling that, succeed or fail, .52
we are all in this together.

Personal leadership .91 .55 .91 .76

Our chief executive officer. . .

gives me special recognition when my .87

work is especially good,
informs others in the organization when I .83
do outstanding work,

gives me positive feedback when I .83

perform well,
encourages me to feel positive about .80

myself if I do an assignment especially

well,
demonstrates total confidence in me. .59
looks out formy personal welfare. .57

Leader expectations .66 .65 .81 .75

Our chief executive officer.. .

expects me to give 100% all of the time, .79

expects less than other supervisors with .75

whom I have worked,
insists on only the best performance. .64

Leader risk .85 .63 .83 .83

Our chief executive officer. . .
risks substantial personal loss in order to .85
achieve his/her vision,
has voluntarily risked a great deal on the .84
success or failure of our mission.

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions
 174 Academy ofManagement Journal February

-/&

Bradley R. Agle (agle@katz.pitt.edu) is an associate profes the Lester Crown Professor ofManagement Practice and
sor in theKatz Graduate School of Business at theUniver associate dean of theYale School ofManagement, where
sityof Pittsburgh,where he is also the director of theDavid he is also the president of theChief Executive Leadership
Berg Center for Ethics and Leadership. He received his Institute.He received his doctorate from Harvard Univer
Ph.D. from theUniversity ofWashington. His research in sity,where he served on the faculty for a decade. His
terests include CEO leadership, stakeholder management, research focuses on CEO succession, careers,
leadership,
ethical leadership, measurement of ethics in organizations, and board governance.
and religious influences on business ethics.
Dhinu Srinivasan (dhinus@katz.pitt.edu) is an associ
Nandu J.Nagarajan (nagaraja@katz.pitt.edu) is a profes ate professor in the Katz Graduate School of Business
sor at the Katz Graduate School of Business at the Uni at the University of Pittsburgh. He received his Ph.D.
versity of Pittsburgh. He received his Ph.D. from North from the University ofMinnesota. His research inter
western University. His research interests focus on
ests include governance, mea
corporate performance
governance, measurement and
corporate performance surement and incentives in organizations, nonfinan
managerial incentives, and corporate restructuring. cial performance measures, and design of cost systems

JeffreyA. Sonnenfeld (jeffrey.sonnenfeld@yale.edu) is and cost management.

-fa

This content downloaded from 194.29.185.109 on Sat, 14 Jun 2014 22:34:18 PM

All use subject to JSTOR Terms and Conditions