What a wonderful honor to have our

What a wonderful honor to have our 1984 article
(Hambrick & Mason, 1984) acknowledged in
this way by the editorial board of AMR. AMR has
published a number of highly influential works
over the years, so it is surprising—and humbling—
to learn that our contribution is seen as
somehow worthy within such a lustrous league.
In this update on upper echelons theory, I follow
the suggestion of AMR’s editor, tracing the
evolution and refinements of the upper echelons
approach and articulating some of the challenges
that still need to be explored.1 I hasten to
add that the editor asked me “to educate the
field a bit concerning . . . [my] ongoing research
agenda in this area”; otherwise, this piece
would run the risk of seeming shabbily selfpromotional.
Since I’m currently engaged in research
on narcissism in CEOs, I am keenly attuned
to the obnoxiousness of first-person
boosterism. In this case, though, I’m just following
directions.
RECAP OF THE INITIAL ARTICLE
The central idea in our original paper, and the
core of upper echelons theory, has two interconnected
parts: (1) executives act on the basis of
their personalized interpretations of the strategic
situations they face, and (2) these personalized
construals are a function of the executives’
experiences, values, and personalities. As such,
the theory is built on the premise of bounded
rationality (Cyert & March, 1963; March & Simon,
1958)—the idea that informationally complex,
uncertain situations are not objectively
“knowable” but, rather, are merely interpretable
(Mischel, 1977). If we want to understand
why organizations do the things they do, or why
they perform the way they do, we must consider
the biases and dispositions of their most powerful
actors—their top executives.
The paper also introduced two subordinate
ideas, each of which seems to have stimulated
major streams of research. The first of these was
that a focus on the characteristics of the top
management team (TMT) will yield stronger explanations
of organizational outcomes than will
the customary focus on the individual top executive
(e.g., CEO) alone. Leadership of a complex
organization is a shared activity, and the collective
cognitions, capabilities, and interactions of
the entire TMT enter into strategic behaviors. In
this vein, many subsequent studies have verified
that organizational outcomes depend, at
least in part, on TMT composition (e.g., Bantel &
Jackson, 1989; Carpenter & Fredrickson, 2001)
and processes (Eisenhardt & Bourgeois, 1988; Simons,
Pelled, & Smith, 1999). The upper echelons
perspective does not require a focus on TMTs
(and a number of significant contributions have
examined CEOs or other individual leaders), but
attention to executive groups, rather than to individuals,
often yields better explanations of organizational
outcomes.
This article is dedicated to Phyllis Mason, who died in
March 2006.
1 I do not attempt to provide a comprehensive review of
upper echelons theory and research. For that, see Finkelstein
and Hambrick (1996), Carpenter, Geletkanycz, and
Sanders (2004), Hambrick (2005), and Finkelstein, Hambrick,
and Cannella (in press).

Academy of Management Review
2007, Vol. 32, No. 2, 334–343.
Copyright of the Academy of Management, all rights reserved. Contents may not
be copied, emailed, posted to a listserv, or otherwise transmitted without the
copyright holder’s express written permission. Users may print, download, or email
articles for individual use only.
334
The second subordinate idea is that the demographic
characteristics of executives can be
used as valid, albeit incomplete and imprecise,
proxies of executives’ cognitive frames. Given
the great difficulty obtaining conventional psychometric
data on top executives (especially
those who head major firms), researchers can
reliably use information on executives’ functional
backgrounds, industry and firm tenures,
educational credentials, and affiliations to develop
predictions of strategic actions. Granted,
the use of demographic indicators leaves us at a
loss as to the real psychological and social processes
that are driving executive behavior,
which is the well-known “black box problem”
(Lawrence, 1997). Nonetheless, researchers have
generated substantial evidence that demographic
profiles of executives (both individual
executives and TMTs) are highly related to strategy
and performance outcomes (Boeker, 1997;
D’Aveni, 1990; Eisenhardt & Schoonhoven, 1990).
And some researchers have ventured to peer
inside the black box, examining the psychological
and social processes that mediate between
executives’ demography on the one hand and
their behaviors on the other (e.g., Smith, Smith,
Olian, Sims, O’Bannon, & Scully, 1994; Simons et
al., 1999).
LATER REFINEMENTS OF THE THEORY
Although the original articulation of upper
echelons theory, in the 1984 AMR article, engendered
a sizable and enduring stream of empirical
investigations, there have also been several
enhancements of the theory itself. Among the
most notable refinements have been the introduction
of two important moderators—managerial
discretion and executive job demands—
which affect the theory’s predictive strength.
Two Moderators of Upper Echelons Predictions
Sydney Finkelstein and I introduced the concept
of managerial discretion (Hambrick &
Finkelstein, 1987) as a way to reconcile two thenopposing
views about the effects of top executives
on organizational outcomes. One view,
coming out of the prevailing tradition of strategic
management, was that top executives
greatly influence what happens to their organizations.
The competing view, coming out of population
ecology (e.g., Hannan & Freeman, 1977)
and new institutional theory (e.g., DiMaggio &
Powell, 1983), was that executives have little
effect because organizations are exceedingly inertial,
swept along by external forces, and constrained
by a host of conventions and norms.
Finkelstein and I argued that both of these
views are conditionally valid, depending on
how much managerial discretion—or latitude of
action—exists. Discretion exists when there is
an absence of constraint and when there is a
great deal of means-ends ambiguity—that is,
when there are multiple plausible alternatives.
Discretion, we proposed, emanates from environmental
conditions (e.g., industry growth),
from organizational factors (e.g., a weak board),
and from the executive himself or herself (e.g.,
tolerance for ambiguity).
The implications of managerial discretion for
upper echelons theory are straightforward—
and profound: upper echelons theory offers good
predictions of organizational outcomes in direct
proportion to how much managerial discretion
exists. If a great deal of discretion is present,
then managerial characteristics will become reflected
in strategy and performance. If, however,
discretion is lacking, executive characteristics
do not much matter. Several studies have shown
that managerial discretion is a pivotal moderator
of upper echelons predictions (e.g., Crossland
& Hambrick, in press; Finkelstein & Hambrick,
1990).
More recently, Finkelstein, Mooney, and I
(Hambrick, Finkelstein, & Mooney, 2005) introduced
another moderator of upper echelons predictions:
executive job demands. Although the
prevailing image is of CEOs carrying very
heavy loads and operating under great pressure,
the reality is that executives’ (including
CEOs’) jobs differ widely in how difficult they
are. For example, some CEOs operate in munificent
environments, with well-fortified strategic
positions, and have very capable subordinates,
whereas others have none of these cushions. We
argue that executive job demands stem from
three sets of factors: task challenges (e.g., difficult
strategic conditions), performance challenges
(e.g., demanding owners or board), and
executive aspirations (e.g., strong personal desire
to deliver maximum performance).
As with managerial discretion, we envision
executive job demands as a potentially important
moderator of the basic predictive strength
2007 Hambrick 335
of upper echelons theory.2 Executives who are
under heavy job demands will be forced to take
mental shortcuts and fall back on what they
have tried or seen work in the past; thus, their
choices will reflect their backgrounds and dispositions.
Conversely, executives who face minimal
job demands can afford to be more comprehensive
in their analyses and decision making;
thus, their choices will more greatly match the
objective conditions they confront.
Empirical work on executive job demands has
not yet commenced, and we anticipate that measurement
will be difficult. Still, it may be possible
to identify conditions of objectively light demands
(e.g., monopolies or cartels) and heavy
demands, or to examine abrupt increases in job
demands (e.g., deregulation), or even to use lab
designs to manipulate job loads. These methods
could be used to explore a host of phenomena,
including this central proposition: the greater
the executive job demands, the stronger the relationship
between executive characteristics
and strategic choices.
Additional Refinements
Beyond identifying managerial discretion and
executive job demands as basic moderators of
the upper echelons logic, refinements of the theory
have extended in other directions as well.
Among these, two elaborations on the basic concept
of the TMT stand out: intra-TMT power distributions
and TMT behavioral integration.
Finkelstein (1992) took the lead in demonstrating
that TMT characteristics yield stronger predictions
of strategic behavior when the differing
amounts of power of TMT members are accounted
for. For instance, he found that the
greater the proportion of TMT members with finance
backgrounds, the more acquisitions companies
would make. But when he added weights
for how much power the finance-oriented executives
had within their TMTs, the results became
even stronger. Finkelstein’s methodology
for measuring executive power was comprehensive
and well-validate