The Dynamics of Collective Leadership and Strategic Change in Pluralistic Organizations

Author(s): Jean-Louis Denis, Lise Lamothe and Ann Langley

Source: The Academy of Management Journal , Aug., 2001, Vol. 44, No. 4 (Aug., 2001),
pp. 809-837
Published by: Academy of Management

Stable URL: https://www.jstor.org/stable/3069417

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 [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

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

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
? Academy of Management Journal
2001, Vol. 44, No. 4, 809-837.

THE DYNAMICS OF COLLECTIVE LEADERSHIP AND

STRATEGIC CHANGE IN PLURALISTIC ORGANIZATIONS

JEAN-LOUIS DENIS
LISE LAMOTHE
Universite de Montreal

ANN LANGLEY
HEC, Montreal

In this article, we draw on five case studies in health care organizations to develop a
process theory of strategic change in pluralistic settings characterized by diffuse power
and divergent objectives. The creation of a collective leadership group in which
members play complementary roles appears critical in achieving change. However,
collective leadership is fragile. We identify three levels of "coupling" between leaders,
organization, and environment that need to be mobilized to permit change. Since it is
difficult to maintain coupling at all levels simultaneously, change tends to proceed
sporadically, driven by the effects of leaders' actions on their political positions.

In this research, we aimed to contribute to the cost control) and multiple actors (professionals, ad-
development of process theory on leadership and ministrators, community groups, and politicians
strategic change in pluralistic organizations. Our linked together in fluid and ambiguous power re
central research question was How can leaders lationships (Bucher & Stelling, 1969; Bunderson,
achieve deliberate strategic change in organizations Lofstrom, & Van de Ven, 1998; Scott, 1982). The
where strategic leadership roles are shared, objec- sector has become more dynamic and complex in
tives are divergent, and power is diffuse? Such recent years as a variety of factors have led to the
situations are becoming increasingly common as breakdown and opening up of organizational
organizations in many industries enter into various boundaries (Meyer, Brooks, & Goes, 1990; Shortell,
forms of collaborative arrangements, as the work- Gillies, Anderson, Erickson, & Mitchell, 1996).
force becomes increasingly diversified, as internal It has been argued that the inherent tensions be-
markets, matrixes, and networks penetrate organi- tween economic values and noneconomic profes-
zational structures, and as knowledge workers play sional values in "value-rational" (Satow, 1975) or
an increasingly important economic role (Lowen- professional organizations such as health care insti-
dahl & Revang, 1998; Van de Ven, 1998). tutions can be resolved through "segmentation," a
We tackled this research question using a repli-
phenomenon in which different parts of an organiza-
cated case study method (Eisenhardt, 1989; Leo- tion function autonomously with minimal linkage
nard-Barton, 1990; Yin, 1994). Specifically, we ex-
between them (Thompson, 1967; Weick, 1976). How-
amined the dynamics of leadership and strategic
ever, while professional autonomy and loose cou-
change in two sets of case studies conducted in the
Canadian health care sector. Health care is a classic pling may encourage local incremental adaptiveness,
they do not necessarily facilitate concerted collective
pluralistic domain involving divergent objectives
action (Cohen & March, 1986). Thus, the question of
(individual patient care, population health, and
how pluralistic organizations develop enough coher-
ence among their parts to allow deliberate strategic
We are very grateful to the participating organizations change remains unanswered.
for their collaboration in this research. We also thank Our first set of three case studies dates from the
research assistants Linda Cazale, Marc Prenevost, Marc late 1980s and early 1990s, a time when many
Pineault, Melodie Faucher, Melanie Summerside, hospitals in Canada were undergoing what could
Michele Saint-Pierre, and Jean-Michel Viola for their
be described as "first-order" strategic change
help in data collection, and the Social Sciences and
Humanities Research Council of Canada, the Fonds
(Fox-Wolfgramm, Boal, & Hunt, 1998): there were
FCAR, and Universite Laval for their financial support. attempts to alter their internal practices and to
The comments of three anonymous reviewers and ofredefine their missions, but their integrity and
existence were not threatened. The second set of
special research forum coeditor Dick Woodman were
also helpful in developing this article. two studies examines the dynamics of leadershi
809

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
810 Academy of Management Journal August

and change during two mergers involving three the next section, we present the case studies of
hospitals each and taking place in the late 1990s. first-order change, summarizing what we learned
These events can be seen as representing "sec- from them, drawing partly on our own previous
ond-order" change because the nature, existence, work (e.g., Denis, Langley, & Cazale, 1996). We then
and boundaries of the organizations were more examine the merger situations, returning to the the-
fundamentally questioned. oretical elements developed earlier, examining the
We first present some theoretical ideas we devel- extent to which they can be applied and adding
oped to understand leadership and change pro- new ideas. We summarize our theoretical under-
cesses in the first set of case studies and then
standing in the final section, where we also review
the limitations of the work and set out a future
examine how these ideas can (or cannot be) gener-
alized and enriched to understand the dynamics research agenda.
observed when new dimensions of complexity and
pluralism are added. We then briefly compare our
conceptualization with other theoretical perspec-
.THEORETICAL APPROACH TO STRATEGIC
tives on change and development in pluralisticLEADERSHIP IN PLURALISTIC CONTEXTS
organizations.
Our focus here is on strategic leadership at
Our own process model identifies three levels of
top management level, a topic that has rece
"coupling," called strategic (occurring between
considerable research attention in the past 15 y
members of a leadership team), organizational (be-
(see Finkelstein and Hambrick [1996] for a revie
tween leadership team members and their internal
constituencies), and environmental (between a However, our analysis departs from traditional
proaches in four important ways related to the
team and an organization's external environment)
ralistic and dynamic context in which our st
that must be mobilized to permit change. We argue
were carried out. First, we view strategic leader
that since it is difficult to maintain coupling at all
as a collective phenomenon to which differen
levels simultaneously, change tends to proceed by
dividuals can contribute in different ways.
fits and starts, with sequential coupling and uncou-
ondly, we view strategic leadership from a proc
pling occurring over time at different levels. In fact,
sual viewpoint in which the focus is more on
strategic change in pluralistic contexts is viewed as
actions of people in leadership positions tha
a succession of episodes in which leadership group
demographics, personality, or cognition. Thir
members may promote change through their ac-
strategic leadership is viewed as a dynamic
tions, but where these actions simultaneously alter
nomenon in which participants, roles, and in
the future form and viability of the leadership
ences evolve over time. Finally, strategic lead
group because their legitimacy is constantly being
ship is seen as a supraorganizational phenome
reevaluated by powerful constituencies. Our stud-
in which leadership roles and influences on t
ies indicate that although change is feasible in such
can extend beyond focal organizational boundari
contexts, it tends to be sporadic and unpredictable.
We summarize each of these elements in turn, e
The model suggests that as the degree of complex-
plaining why this theoretical positioning is app
ity and pluralism increases and as slack resources
priate to the issues under study.
become scarcer, the conditions necessary to protect
change initiatives long enough for them to become
irreversible are more and more difficult to estab-
Strategic Leadership as a Collective Phenome
lish. In its recognition of the dynamic interplay
amongst context and action, our model reflects a Although the popular business press still te
"structurational" view of organizational processes to give inordinate attention to individual CE
(Giddens, 1984). researchers have recently stressed that strat
On the one hand, in this article we describe and leadership requires the contributions of more t
theorize about the change and development jour- a single individual. For example, Hambrick a
neys of the organizations and people studied. On Mason (1984) and Eisenhardt (1989) focused on
the other hand, since the theoretical ideas pre- management teams, and Pettigrew and Wh
sented here evolved as we moved from study to (1991) emphasized that change leadership in
study over time, this article also describes a re- large firms they studied often involved many i
search journey followed by us, as authors, over a viduals. In some cases (as when the posts of
period of eight years. We begin by introducing cer- and president are separate), collective strate
tain theoretical ideas that are central to the analy- leadership is explicitly formalized within an org
sis. We then present the methodology in the form of ization's structure, leading to a variety of pos
a chronological record of our research process. In divisions of roles (Stewart, 1991). In this researc

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 811

we deliberately focused on settings in which the ers' actions from the data, examining their observed
structural demands for collective strategic leader- relation to the evolution of change.
ship were very strong. Quebec hospitals in partic-
ular have explicit dual structures in which no one
Strategic Leadership as a Dynamic Phenomenon
individual has formal authority over all others and
leadership is shared between at least four different Most analyses of strategic leadership are cross-
actors. Moreover, the power of the top teams is sectional, as can be seen in the extensive research
further diluted by the influence of other profession- on top management team demographics, strategy,
als and by external agencies. The setting studied is and performance (e.g., Finkelstein & Hambrick,
particularly interesting in that this diffuse power 1996). However, in such arm's-length views of top
arrangement has been regulated by law since the management teams, the dynamic nature of the links
early 1970s. Although adjustments have occurred studied tends to remain mysterious. As Pettigrew
over time, the same fundamental pluralistic frame- noted, there is a need for research that directly
work remains in place, even as the organizations examines the "emergence, development, conduct,
have had to deal with many other types of change. impact and performance of management teams"
Following Hodgson, Levinson, and Zaleznik (1992a: 1974). The research described in this article
(1965), we use the term "leadership role constella- deals explicitly with the dynamic construction, de-
tion" (rather than "top management team") to refer construction, and reconstruction of leadership
to the collective leadership group. This notion im- roles over time as contextual forces evolved and as
plies not only multiple actors, but also a certain members of leadership role constellations inter-
division of roles among them. For example, one vened within their organizations. In this sense, our
member of a constellation may focus on relations research was inspired by views of structure and
with outside organizations, and another may han- action as mutually influencing one another over
dle internal operational issues. Hodgson and col- time, with individual action being constrained by
leagues (1965) posited that the effectiveness of structure and yet simultaneously modifying that
leadership role constellations depends on the de- structure (Giddens, 1984; Ranson, Hinings, &
gree of complementarity among their members. Greenwood, 1980).
Complementarity implies both adequate coverage Thus, as well as being concerned with leaders'
of all activity domains as well as the existence of actions, we were also interested in the effects of
mechanisms allowing different actors to play their those actions on organizations and on their strate-
respective roles in a concerted manner. Clearly, a gies and on the leadership role constellations them-
perfectly harmonious and complementary constel- selves. Although most of the literature does not
lation may be difficult to achieve when authority make the distinction explicit, it seems reasonable
relationships are diffuse and when individuals to suggest that leaders' actions may have at least
have different sources of expertise and influence. three different types of consequences. We have la-
However, as we shall suggest, such conditions may beled these "symbolic," "substantive," and "politi-
be an important prerequisite for the implementa- cal." Symbolic consequences concern changes in
tion of strategic change. an organization's dominant strategic frame or inter-
pretative scheme (Hinings & Greenwood, 1988;
Meyer & Rowan, 1977), which is the set of accepted
ideas about what the organization's role and strat-
Strategic Leadership as a Processual
Phenomenon
egy should be. Substantive consequences concern
the effects of leaders' actions on resource alloca-
Without denying the contributions of cognitive,tions, organizational structures, and realized strat-
egies. Political consequences concern effects i
demographic, or psychoanalytical approaches to
terms of the distribution of formal and informal
leadership, in this research we adopt what Hosking
(1988) labeled a "processual" perspective, focusing
power and the evolution of leadership roles within
a leadership constellation. These three types of
primarily on what "leaders" do to mobilize others
consequences are examined in this research.
in a system of interrelationships rather than on
what they are. We are thus interested in the tactics
leaders use to influence the course of events as well
Strategic Leadership as a Supraorganizational
as in the types of organizational actions they pro- Phenomenon
mote. Several different typologies of influence tac-
tics have been suggested in the literature (Cohen & Strategic leadership is viewed as a partly su-
March, 1986; Kipnis, Schmidt, & Wilkinson, 1982; praorganizational phenomenon in which leader
Pfeffer, 1981). We derive our own typology of lead- ship roles and influences on them extend beyond

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
812 Academy of Management Journal August

focal organizational boundaries. Institutional theo- de Ven, 1992; Yin, 1994). To date, five case studies
rists have argued that the more supraorganizational have been carried out or are underway.
influences a focal organization is exposed to, the Our approach was and is also partly deductive
more likely leaders are to respond to external (theory inspired) and partly inductive (data in-
demands for change (DiMaggio & Powell, 1983; spired). In our experience, this mixed approach can
Powell & DiMaggio, 1991). Thus, this perspective is be very fruitful because it allows one to gain cre-
particularly salient for organizations that form ative insight from the data, without necessarily de-
parts of an interrelated pluralistic network, as do nying or reinventing concepts that have been useful
health care organizations in Quebec. As noted previously (see also Fox-Wolfgramm et al., 1998;
above, although each organization has an indepen- Orton, 1997). By relying on multiple and sequential
dent (but elected) board, it receives the vast major- case studies, one can also use deductive and induc-
ity of its financing directly or indirectly from gov- tive logic iteratively to foster the development of a
ernment. Moreover, its structure and strategic richer theoretical framework over time. In this pro-
decisions are often subjected to stringent rules, reg- cess, insights from one case generate constructs or
ulations, and approval mechanisms. Thus, collec- theoretical propositions that serve as a basis for
tive leadership must mobilize support and manage probing the process of change in other cases. It is
the combination of deduction and induction
relationships not only within the organization, but
also outside it, within its network (government through time and the sequential "replication l
bodies, other health care organizations, commu- (Eisenhardt, 1989; Leonard-Barton, 1990; Yin,
nity, and so forth). Conversely, demands and op- 1994) that provides the basis for our theoretical
portunities within the network evolve over time inferences. In the following description, we at-
and will affect the patterns of collective leadership tempt to be clear about how our ideas emerged and
roles within individual organizations. The supraor- gelled over time, illustrating in the Results section
their fit with the data.
ganizational nature of leadership roles is particu-
larly important in situations in which organiza- The research journey began in 1991, when the
tional boundaries are being questioned and first author became aware that a major change was
reformed, as they were in our second series of underway in a local hospital he knew from previ-
cases. ous consulting work (labeled "Suburban Hospital"
later).
The four dimensions of strategic leadership iden- He was interested in strategic change pro-
tified above form the backdrop for the empirical cesses and suggested to the third author that they
analysis, providing a number of preliminary con- might collaborate on a case study. This hospital
was
ceptual elements and relating these to the specific negotiating with a medical school to acquire a
setting examined. We have not, however, begun teaching
to mission. At this time, Hinings and Green-
wood's (1988) work on the dynamics of strategic
theorize about how these dimensions of strategic
change in local government was very influential on
leadership affect the change process. This concep-
tualization will emerge as we present our resultsway in which we framed the study (a deductive
the
later in the article. In the following section,step).
we Their notion of organizational archetypes
describe the research journey we followed in de- change tracks appeared to be promising con-
and
cepts for understanding health care change (Hin-
veloping these ideas.
ings & Greenwood, 1988) and, indeed, these con-
cepts were mobilized in the analysis of this case
(Denis et al., 1996).
THE RESEARCH PROCESS However, other observations emerging from the
data also drew our attention (an inductive step).
General Approach
The first was the importance of collective leader-
Our research journey, which covers a long ship
periodin this context. The second was the apparent
of time and is continuing, was partly planned link and
between the types of tactics members of a
partly opportunistic. We aimed to develop leadership
process group use in promoting change and their
rather than variance theory: that is, we were inter-
credibility as leaders-indeed, their capacity to re-
ested in "describing and explaining the temporal
main in leadership positions. It was at this point that
sequence of events" involved in change (Van de to look at previous literature on collective
we began
Ven & Huber, 1990: 213) rather than in identifying
leadership (Hodgson et al., 1965; Stewart, 1991) and
relationships between variable levels of inputs
on and
power and influence tactics (Kipnis et al., 1982;
outputs. Thus, the case study method, which Pfeffer,
in- 1981, 1992) in search of theoretical ideas that
would help us conceptualize what we saw as hap-
volves tracing processes in their natural contexts,
appeared most appropriate (Pettigrew, 1992b; pening
Van in this case. Hodgson and colleagues' (1965)

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 813

notion of a leadership constellation (see above) was tively drawing on the theory developed earlier but
immediately appealing and appeared applicable to extending it inductively to this new situation.
our data (a deductive step). Drawing on this notion The five change situations studied combine two
and developing our own typology of influence tactics, critical attributes: the need to respond to new ex-
we used the data to document the nature of the lead- ternal pressures and the potential of change initia-
ership constellation, the tactics used, and their effects tives to destabilize internal patterns of influence
and interest satisfaction. The main differences
over time, classifying these effects into the three cat-
egories of symbolic, substantive, and political men- among the situations concern the intensit
tioned above (an inductive step). This process pro- change, which increases significantly when
duced our initial model of leadership and strategic switch from the analysis of a single organizatio
change under ambiguity, which was eventually pub- that of multiple organizations participating
lished (Denis et al., 1996). merger. The question examined here is how
Meanwhile, we had begun two other case stud- all) this added complexity affects the dynamics
ies. The second study was initiated in late 1992 and the change leadership process. Our final stud
was planned as a replication and test of the model sign can be described as a multiple case st
already emerging from the first. It involved the design with embedded units of analysis and
development of a new emergency care mission in a levels of replication (Yin, 1994), each contrib
smaller hospital (below, "Community Hospital"). to the generalizability of the emerging theory
The third study was started in 1993, also in a single
hospital ("University Hospital"), but for a different Data Sources and Analysis
purpose. Taking off from the notion of collective
The case studies involve multiple data sour
leadership, we were interested in determining how
including documents, interviews, and observati
a new CEO penetrated his new organization and
of meetings. Some of the data are retrospective
positioned himself within an existing leadership
some are real-time-based. Table 1 provides a c
constellation. At this point, University Hospital
nological view of data sources in relation to
could not be said to be involved in major change.
events under study for the five cases. Documen
However, as time progressed, the organization did
sources included minutes of top-level meeti
become involved in a merger (the "Metropolitan
(board or management team), internal docum
Merger"), later to be the focus of our fourth case.
and press reports. We had access to the minute
Prior to this point, we had analyzed these two
a variety of types of meetings for all cases, enablin
studies using a framework similar to that devel-
to establish chronologies based on real-time rec
oped for the first case (a deductive step). Many of and to develop hypotheses to be tested later in i
the ideas seemed applicable, but new ones views. The span of years shown in Table 1 cover
emerged. In particular, the notion of three levels of time for which minutes were available and/or direct
coupling (to be described later) was seen as useful observations of meetings were carried out.
in understanding the evolution of events (an induc- Although more subject to hindsight bias than
tive step). These ideas were presented at several documentary records, interviews allow a greater
conferences but never published. degree of understanding of why events occurred as
Then, in 1995, moves to consolidate the teaching they did and how people felt about them. We con-
hospital network in the major cities in Quebec were ducted over 100 interviews, each 1-2 hours long,
set afoot. The opportunity arose to extend our work for the full set of case studies (for details, see Table
to this new, more complex, situation. The second 1). Almost all were taped and transcribed. Inter-
author had done parallel research work at the Uni- viewees included CEOs (at least two interviews in
versity Hospital, and the three of us decided to each case), their "direct reports" (directors of pro-
embark on comparative case studies of the leader- fessional services, nursing, finance, human re-
ship and change processes involved in two mergers sources, technical services, and so forth), senior
(the Metropolitan Merger and the "Capital Merger"). medical staff members (presidents of the Medical
Data collection began in January 1997 and is still Council Executive, chiefs of medicine, surgery, and
going on. The overall objective is to develop an other departments), board members, and middle-
understanding of the processes that lead to progress level participants (managers, advisory staff mem-
in implementing mergers at both strategic and op- bers, nurses, and other professionals) involved in
erational levels. At the strategic level, issues of key decisions. (The latter were interviewed in
leadership and its relationship to change are clearly smaller numbers than were the former.) We also
important. In this article, we specifically focus on took care to ensure that people from all sites in-
this dimension of the ongoing processes, deduc- volved in the mergers were fairly represented. In

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
 TABLE 1

Data Sources and Event Coveragea

Suburban Hospital Community Hospital University Hospital Metropolitan Merger

Year Documents Interviews Documents Interviews Documents Interviews Documents Intervi

1980 X
1981 X
1982 X
1983 X X
1984 X X
1985 X X
1986 X X
1987 X X
1988 X X 1
1989 X X 1 X
1990 X X X
1991 X 3 X X
1992 X 6 X X
1993 X X 9 X 13
1994 8 1 X 10 X
1995 6 X 10
1996 X 11
1997 X 19 8
1998 X 21 16

Total interviews 17 12 29 40

Total meetings 45

a The data noted include two early interviews (with the CEO and a senior physician) for Community Hospital. These were done in the con
planning, but their content was pertinent to the subsequent study (indeed, it inspired our interest in this case).
b The Metropolitan Merger includes University Hospital. Some of the 1995 meeting data are from board meetings at University Hosp

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 815

the interviews we aimed to trace the events associ- actions of one period lead to changes in the context
ated with change, to identify the roles of members that will affect action in the next. This form of
of a hospital's leadership group, to understand the dynamic analysis was important in this research.
logic behind these actions, to gauge the reactions of For the comparative analysis of periods within
the respondent and others to these actions, and to and between cases, we used techniques for data
understand their short- and longer-term conse- reduction and presentation similar to those sug-
quences. In general, we used the three or four most gested by Miles and Huberman (1994). The three
important issues within the hospital as themes in most important categories for the analysis within
structuring interview questions. each period included (1) the characteristics of the
The third data source, direct observations of leadership constellation during that period (who
meetings, was used most extensively in the two the important members were, what roles they
merger cases because these were followed in real played, what their degree of complementarity was),
time. Indeed, for the Metropolitan Merger, we were (2) the actions of the leadership group (what was
able to follow merger negotiations as they were done, what kinds of tactics were used, and (3) the
seen by the boards of all three participating organi- effects of these actions and tactics (symbolic, sub-
zations before the formal agreement was finalized. stantive, and political). As we advanced in the
Although obviously only presenting a partial view, analysis and as theoretical ideas developed, these
observations of meetings enable researchers to see categories were refined. Typically, the analysis of
directly what positions people take and how differ- case material by individual investigators was fol-
ent individuals tend to interact during the discus- lowed by group discussion between team members,
sion of critical issues. Most meetings were not tape- all of whom had different skills and backgrounds.
recorded, but detailed notes were taken. Table 1 To further validate the data and the interpreta-
provides a summary of the timing and number of tions, preliminary analyses were circulated to key
meetings observed. The meetings included both informants. Phone calls, feedback interviews, and a
public board meetings (in the Metropolitan case) variety of other strategies were used to engage par-
and top-level internal meetings (both mergers). ticipants and ourselves in reflection on the find-
The analysis of the cases was carried out using a ings. For example, the three most important actors
temporal bracketing strategy (Langley, 1999) simi- in the leadership constellation of Suburban Hospi-
lar to what Barley (1986), Doz (1996), and some tal took part in a class discussion of strategic
others doing process research have used. This strat- change with master of health administration
egy involves decomposing the chronological data (M.H.A.) students; a preliminary case description
for each case into successive discrete time periods, was the basis for the discussion. For the University
or phases, that become comparative units of analy- Hospital case, involvement included joint presen-
sis. Phases are defined so that there is continuity in tations with the CEO and the preparation of twin
the context and actions being pursued within them, papers in a professional journal (one by the authors
and one by the CEO) on the CEO succession pro-
but discontinuities at their frontiers. In our case,
the boundaries of the chosen periods were defined cess. Finally, for the Metropolitan and Capital
either by changes in the key people involved (the cases, we conducted a joint feedback session with
leadership constellation) or by a major change in both CEOs to discuss similarities and differences
the environment. We also used informants within emerging from the preliminary analysis. In the next
the organizations to help identify and validate the section, we present the results emerging from the
chosen periods. analysis of the first three cases. In the subsequent
It is important to understand that these periods section, we extend the analysis to the mergers.
do not represent "stages" in the sense of a predict-
able sequential process, and they do not in any way
imply the type of progressive life cycle logic found COLLECTIVE LEADERSHIP AND FIRST-ORDER
in many normative process theories of change (Van STRATEGIC CHANGE
de Ven & Poole, 1995). Yet they are more than just The Formal Framework for Collective
a descriptive convenience. Specifically, they per-
Leadership in Quebec Hospitals
mit the constitution of comparative units of analy-
Hospitals have been variously described in the
sis for the exploration and replication of theoretical
ideas. This can be especially useful if feedback literature as dual hierarchies (Freidson, 1970),
mechanisms, mutual shaping, or multidirectional three-legged stools (Johnson, 1979, 1994) and, more
causality are likely to be incorporated into theori- recently, as composed of four "worlds" (Mintzberg,
zation. For example, the decomposition of data into 1997), all pulling in different directions. These im-
periods can enable the explicit examination of how ages all evoke the pluralism of their power struc-

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
816
Academy of Management Journal August

tures, whose roots lie in the nature of professional services, another position mandated by law in Que-
work itself and in the historical evolution of the bec hospitals. This is a medically trained manage
health professions and the institutions they inhabit who reports to the CEO and is administratively
(Abbott, 1988; Friedson, 1970; Mintzberg, 1979; responsible for medical services offered in the hos
Reed, 1996; Satow, 1975). It is not our purpose to pital. The importance of this role is highly variable
explore these root causes here. However, in order to from situation to situation. As we shall see, he o
understand the dynamics of leadership and change she can become an important player in the leader-
in the organizations we studied, it is first necessary ship constellation, serving as a critical bridge be
to explain more clearly how this pluralistic power tween the administration and the medical staff,
structure is formalized within the particular set- occupying a "structural hole" (Burt, 1992). How-
tings we examined.1 Specifically, who are the prin- ever, this role tends to be more secondary if deal-
cipal actors in their leadership groups? How are ings between the CEO and doctors are more direct.
they related to one another? What are their sources Other potential top-level leaders include clinica
of power and legitimacy? What formal levers do department heads (of surgery and medicine for in-
they control? stance), who report officially to the director of pro
The first key actor is obviously the CEO. He or fessional services, and other senior administrators.
she is appointed by a board of directors and is However, despite Mintzberg's (1997) positioning o
formally responsible for all aspects of hospital nursing as one of a hospital's four worlds, and
management, including professional services, nurs- despite the obvious importance of nursing to oper-
ing, support services, financial management, and ations, in our cases, nursing directors, like othe
human resources. The second key pole of leader- administrators (finance, human resources) were
ship is the board. This is a partly elected body usually seen as subordinate to the CEO in leader-
including representatives of the medical staff (doc- ship terms.
tors), nurses, other professionals, other employees, In summary, as shown in Figure 1, the variou
hospital users, and members of the community. candidates for membership in a collective leader
The board chairperson cannot be a hospital em- ship group or leadership role constellation in a
ployee and is a part-time volunteer. However, as hospital are linked by hierarchical relationships,
the hospital's final arbiter, the board and, particu- and each has certain official roles and responsibil
larly, its chair can play an important leadership ities. However, the formal prescribed structure
role. The third pole of leadership is the Medical does not give dominant power to any of the key
Council Executive (MCE), a committee democrati- actors. Thus, they have to find some kind of modu
cally elected from among the medical staff for a vivendi or, to put it more, theoretically, a "negoti-
one- or two-year mandate. (Physicians are not em- ated order" (Bucher & Stelling, 1969; Strauss
ployees but are paid on a fee-for-service basis by a Schatzman, Ehrlich, Bucher, & Sashin, 1963), in
government agency other than the one that regu- which, through ongoing interactions based on con-
lates hospitals.) The Medical Council Executive is trol of key uncertainties, affinity, and expedience,
legally responsible for the quality of care within the different individuals negotiate their space and
hospital, but it reports formally to the board, not to come to take on different roles that may be more o
the CEO. Partly because of this, and also because of less harmoniously aligned (that is, complementary).
the dominant role of physicians in operations, the In fact, although the prescribed structure is the
president of the Medical Council Executive often same within any hospital in Quebec, each organi
takes on a key role within the leadership constel- zation can develop over time a certain emergen
lation. Indeed, this committee often becomes in- idiosyncratic equilibrium between the poles o
volved in management issues beyond its legislated leadership that extends beyond the particular indi-
mandate.
viduals that occupy the roles (Hinings & Green-
These three poles basically represent the three wood, 1988). For example, the long tenure of a
legs (administration, board, and physicians) of the leader with a particular style is likely to create
traditional "three-legged stool" (Johnson, 1979). tradition in which other leadership team members
However, individuals occupying other positions will tend to play symbiotic roles that are adjuste
can also take on important leadership positions. to this style. A new leadership participant who
Critical among these is the director of professional enters this situation will find it difficult to single
handedly alter the patterns of role sharing previ
ously established, as the constellation at the top
1 Note that as all hospitals in this research are regu- will have become embedded in a network of rela-
lated by the same agency, they all have the same basictionships and expectations that extends across the
structure. entire organization. This constitutes a major chal-

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 817

FIGURE 1
Basic Structure of Public Hospitals in Quebec

University

;fluidb; 'oun I i I rl -
fluid boundary " r *^ :^ >_ > :=_ -i - M 1 I

Medical staff

a Strictly speaking, this organizational chart is valid only prior to 1992, when the regional councils became region
hierarchically placed between the hospitals and the Ministry of Health and Social Services.
b Bold type indicates the actors most likely to be members of a hospital's leadership constellation.

lenge for change initiatives. To achieve significant

form of a vignette accompanied, in the right-hand
change, a tightly knit leadership group that can column,
act by extracts from interviews that illustrate
in concert may be needed. But change in itself often
the dynamics described.2 Each vignette begins with
implies a need to reformulate power relationships,
a description of the initial situation and leadership
thus requiring the destruction of past patterns of
constellation and then presents the case history
embedded equilibrium. divided into phases. For each phase, we succes-
In closing this section, it is also important for sively
us describe the actions of the leadership team
to note that in this public sector context, the lead-
during each phase and then their substantive, sym-
ership potential of all internal actors is dilutedbolic,
by and political consequences. A comparative
the influence of outside agencies, which include a
analysis of the three cases is given in Figure 2 and
regional coordinating body and the province's Min- in Table 5, and brief synopses are provided below.
istry of Health, the main source of hospitals' reve- The three stories are different in many ways but
nue. The pressures and constraints associated with similar in others. The first story shows how, after
these dependencies contribute to the challengeanof abortive change attempt, a new leader emerged
change. We will now examine empirically how
changes occurred or did not occur in our first series
of cases.
2 For more detailed descriptive information on the
cases, see Denis et al., (1996, case 1): Denis, Langley,
Case Descriptions Prenevost, and Cazale (1994); and Prenevost (1994, case
2); Denis, Langley, and Pineault (2000, case 3); and De-
It is not easy to present complex process stories nis, Lamothe, and Langley (1999, merger cases). The
succinctly. Our approach is to provide, in Tables 2, conference paper from this list (Denis et al., 1994) is
3, and 4, a descriptive summary of each case in the available on request from the authors.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
818 Academy of Management Journal August

TABLE 2

Case 1, Suburban Hospital: Acquiring a Teaching Mission

Vignettesa Interview Extracts

Context. A new CEO enters a leadership constellation dominated by medical staff

Pre 1985. This hospital was the largest nonteaching hospital in the country andThe had[previous] CEO was the executor of
attracted physicians interested in developing high-volume practices. decisions that were taken by everyone
Traditionally, the medical staff played a dominant role, with administrators else.
handling the day-to-day mechanics of running the hospital but not intervening in
the clinical domain. The Medical Council Executive had direct influence with the It was management by the gang of doctors
board and used its many political connections to protect its interests and as a who set off to see the government in
means of ensuring that the hospital had adequate resources. In 1985, a new CEO Quebec. If you have a problem, they'll
settle it with the minister.
was hired with a mandate to modernize management methods. A leadership
constellation dominated by physicians had to integrate a new CEO who wanted
to play a more strategic role.

Phase 1. Confrontation: Alignment to external constraints -> Weakened

constellation

1985-87. The CEO's initial efforts at management control included the creation of The
a new CEO arrived and wanted to
strategic planning process with a consulting firm and attempts to satisfy "straighten out" the hospital. It didn't
government demands for budget restraint. These efforts were rebuffed by the work.
medical staff, who withdrew their collaboration, claiming that strategic planning
He was unable to get the doctors and the
was their exclusive domain and that the organization could not cut budgets
directors involved in the process.
because it was already underfunded. Under pressure from the physicians, the
board withdrew its support for the CEO's strategic planning process.
Substantively, the actions of this period produced no change, although
symbolically, they may have planted an idea that could become more acceptable
later. Politically, collaboration between physician leaders, the administration, and
the board was at a low point, and the constellation was weakened.

Phase 2. Mobilization: Symbolic management -> Constellation building

1988. After a waiting period, the CEO resuscitated the planning process by It was a plan that united the
delegating its coordination to the head of the hospital's community health administration, physicians, nurses,
department. This physician was able to build consensus around certain everyone.
mobilizing developments (including university affiliation), and offered his
services to help produce justifications for an increased budget allotment from the There was a chemistry between the DPS
government. Symbolically, these actions produced a positive vision of the future, and the Medical Council president, and
but so far, no substantive changes. Politically, these popular developmental they had good relations with XYZ on
orientations increased the physician's visibility and credibility, and led to the board, which moved things along.
successive promotions, first to DPS and then to director of operations. They also
contributed to the development of a strong alliance with a new president of the
Medical Council and with key board members, creating the nucleus of a tightly
knit constellation committed to change.

Phase 3. Implementation: Forcing change - Delegitimizing the constellation

1989-90. The united top management group organized to implement a number of The university hospital idea was linked to
major changes. The director of operations acted as catalyst and visionary with the obtaining a balanced budget.
support of the board and a new DPS, while the Medical Council president played
a key political role in gaining access to government decision makers. University The people in place were functioning in a
affiliation was achieved by obtaining external support in exchange for respecting kind of delirium and with a process
budgetary constraints and making a complementarity agreement with another that was not transparent.
hospital. However, many physicians began to be unhappy with the changes when
You can't keep on breaking china. The
it became clear they might have a negative effect on their incomes. It was also felt
orientation was not fully appropriated.
that they had been forced through using somewhat underhanded tactics. Growing
distrust resulted in the replacement of the Medical Council president in an
election by someone who opposed the teaching affiliation. The director of
operations left. Substantively, the actions of the period produced considerable
change, though symbolically, the vision lost its glow. Politically, although the
board, the CEO, and the DPS were still united, the constellation was weakened
overall by the departures of two key members. From 1990 onwards, the pace of
change slowed down.

a DPS stands for director of professional services.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 819

TABLE 3

Case 2, Community Hospital: Saving a Declining Organization

Vignettesa Interview Extracts

Context. A new CEO is imposed on a hospital with a laissez-faire leadership

tradition.

Pre 1986. This hospital was founded in the 1960s as a center for obstetrical The CEO managed the hospital according to his
care and general surgery, run mainly by general practitioners. However, mood and the weather.
declining birth rates, a failure to adopt new styles of practice, and an aging
medical staff left the hospital stagnating. The new CEO was imposed by It was a hospital in free fall. There were major
government in the midst of a crisis over declining admissions, complaints quality problems.
about quality, difficulties in budget management, and conflict between
physicians, nursing, and the previous CEO (resulting in his departure). The
new CEO faced a conservative medical leadership, weakened by negative
quality assessments from professional bodies.

Phase 1. Transition: Finding a new raison d'8tre -> Renewing the

constellation

When the new CEO arrived, you felt more

1986-87. The new CEO rapidly built a strong alliance with a new board chair,
replaced most of the administrative team, and instituted a strategic enthusiasm on the part of employees, the
planning process. To maintain the support of medical staff and others, heboard, and the doctors; everyone got on
initially included development of obstetrics in the strategic plan and board.

challenged the doctors to recruit new obstetricians. This effort failed,

admissions continued to decline, and criticisms of quality persisted. We had to find something else to do as
Pointing to these problems, the CEO and a new DPS persuaded the Medical [obstetrics] admissions were declining.
Council and the board to accept closure by negotiating with the government
to retain the resources associated with obstetrics to develop a new
emergency care mission. Substantively, the obstetrics department was
closed. Symbolically, although there was distress at the loss, the new
mission provided hope for survival. Politically, the CEO, the board, and the
DPS had developed a strong alliance around the new orientation.

Phase 2. Turnaround: Implementing the new mission -> Broadening the

constellation

1988-90. The DPS embarked energetically on a program of reorganization. [The emergency mission] was a way of saving
With the support of the board and the CEO, he took personal control of the the acute care status of the hospital.
weaker clinical departments. He attracted a group of younger doctors that
With the new Executive, we succeeded in
got themselves elected to the Medical Council Executive and who
enthusiastically promoted the new mission. A united leadership group getting an acceptable level of functioning.
appeared to be in place, and the emergency mission began to be
implemented. Substantively, productivity improved, although older
physicians had difficulty adjusting to the new rhythm. Symbolically, there
was greater pride in the institution. Politically, the Medical Council
Executive was now associated with the constellation. However, the
increased influence of the DPS and pressure for resources from new
physicians was beginning to create tensions with the CEO.

Phase 3. Confrontation: Pursuing divergent objectives -> Splitting the

constellation

1991-92. The DPS and the new Medical Council Executive began to push for We wanted to develop intensive care, a trauma
faster change than the CEO and the board chair felt was prudent and center but the CEO wasn't buying it.
feasible. The physicians wanted to develop intensive care to ensure higher
volumes and adequate revenues for new recruits. The CEO was concerned I think the DPS was after the CEO's job.
with maintaining a balanced budget. Distrust between the DPS and the CEO
There was a withdrawal of the main medical
over this issue escalated into a deadly power struggle. When the dust
leaders and a series of departures by
settled, the DPS had left the organization, taking several new physicians
specialists.
with him, and the CEO's own position was severely weakened. Power
within the Medical Council reverted to physicians less favorable to the newWe were on the way up and then we crashed.
orientation. Substantively, this resulted in continued decline. Symbolically,
the battle was disastrous for internal motivation and external reputation.
Politically, the leadership constellation was weakened and divided. The
organization had moved full circle.

a DPS stands for director of professional services.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
820
Academy of Management Journal August

TABLE 4

Case 3, University Hospital: Developing an Academic Medical Center

Vignettes Interview Extracts

Context. A new CEO enters a prestigious institution with strong medical

leadership
Pre 1992. This large and prestigious teaching hospital had a tradition of strong The board, the [previous] CEO, and the
medical staff involvement in management issues but with the three poles of Executive always agreed on the main
influence (board, Medical Council Executive, and CEO) working closely issues.
together in a tightly knit constellation. The previous CEO had launched a
number of major development and construction projects, and the hospital We're proud. We're the best.
was perceived as performing well both in clinical and financial domains. What we wanted was a new captain, not
There was no perceived need for major change. There was, however, a someone to change the ship.
concern about increasing resources to finance the new initiatives and about
acquiring formal designation as an academic medical center under new
legislation. This meant upgrading certain areas and improving the hospital's
external relations.

Phase 1. Assimilation: Defending the hospital -> Reinforcing the constellation

1992-94. The new CEO came from a smaller hospital where he had developed The CEO succeeded in taking the local
a reputation as an innovator and communicator. He quickly demonstrated culture, adopting it and adapting it, and
energy and effectiveness in defending the hospital externally, facilitating his taking the lead.
integration into the tripartite alliance that previously existed (CEO, board,
Medical Council Executive). His administrative initiatives (such as total There's even greater collaboration with the
quality management) were less successful. However, support from medical staff.
physicians and the board enabled him to overcome these difficulties and to
instigate a move toward ambulatory care. Substantively, change was modest,
though admission volumes increased. Symbolically, the medical staff
became convinced of the advantages of ambulatory care. Politically, the
tripartite leadership constellation was reinforced, and the CEO had high
internal and external credibility.

Phase 2. Identity crisis: Negotiating a merger -> Dividing the constellation

1995-96. In 1995, the hospital was forced by the government into merger talks Under budget restraints, the only way to
with two other teaching hospitals. The CEO, the board chair, and the develop a real academic medical center
president of the Medical Council Executive took part in these negotiations. was by broadening the playing field.
They continued to defend the hospital's position but were gradually
convinced by promises of new resources that merger was the only way to The doctors were worried that the merger
ensure that the hospital could develop and acquire formal designation as an would take place on their backs and they
academic medical center. Their shift towards acceptance of a merger on wanted more transparency from the
Medical Council Executive.
three sites was unpopular with some members of the medical staff, who
succeeded in overturning the Medical Council Executive in an election.
However, the new team was unable to redirect the negotiations.
Substantively, the merger agreement was finalized. Symbolically, the merger
was a traumatic decision but one that still held out hope for development.
Politically, the constellation was divided. However, the CEO still
maintained considerable credibility, and he was widely supported internally
as a potential candidate for CEO in the merged hospital (see the
Metropolitan Merger case).

and mobilized enthusiasm internally around theport for a new mission that would enable the organ-
development of a teaching hospital mission. He ization to survive. However, the stronger group of
built a strong alliance that then moved aggressively physicians put in place to implement the new mis-
to implement this mission. However, the unitedsion pushed to develop it more extensively than
leadership team moved too fast for many organiza- the CEO and the board felt was possible. Conflict
tion members, and an election caused its dissolu- led to departure of the physicians and a return to
tion, producing a slowdown in the change process. stagnation.
The second story shows how a new leader was The third story shows how a new CEO became as-
imposed on a declining organization, how he re- similated into the leadership constellation of a strong
placed the management team, built credibility with organization, fitting into the promotional role
the board, and achieved internal and external sup- expected of him and strengthening the leadership

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 821

FIGURE 2
Progression in Leadership Constellations: The Three Cases

Time Suburban Hospital Community Hospital University Hospital

Initial

constellation | o OA0 o 0| |0 O 0
Board Board Board

CE CEMCE MCE CEO MCE

Medical staff domination Weak coupling Strong coupling

Phase 1 ? 100
Integrationof v | I?
o ool
o 0 o0o 0
ol
n g oard Io \ o Board Y Board
CEO MC l CE C MCE MC

Conflict in constellation, paralysis Recoupling with e

Phase2 I o 0 oo | o 0 ol
Board I S Board / Bo ardo l

CEJ MCE CEO CE | CEO MCE

Building momentum by stimulating Implementing change, Negotiat

and selling organizational aspirations broadening constellation detach

Phase3 O O O o

Zmplementing change but detaching Conflict in constellation,

from organizational base paralysis

End of study
period 00O ,OloO
O o
o O
\ O I

I r,Board ard Board

CEO MCE CEO CCE MCE

Slower change Weakened organization y Forced merger, hope in CEO

O Individual member of leadership constellation

* Key member of leadership constellation in this time period
- Coupling within constellation and organization (thickness of line indicates degree of coupling)
Conflict within leadership constellation
Coupling of constellation with environment (thickness of line indicates degree of coupling)
-> Direction and intensity of influence between constellation and environment
O Internal groups

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
 TABLE 5

Dynamics of Collective Leadership and First-Order Cha

Characteristics Suburban Hospital Community Hospital Universit

Initial constellation Moderately strong Weak Strong

Politically connected medical staff Weakened medical staff Influential
with links to board Laissez-faire administration Strong li
Passive administration Board threatened with trusteeship CEO

Environment-organization Moderate Low (environmental uncoupling) High

coupling

Phase 1
Focal actors New CEO New CEO, board, DPS New CEO

Dominant tactics Confronting organization Salvaging organization Aligning w

Aligning with environment Negotiating with environment Promotion vs
Strategic planning before ensuring Alliance of CEO with board Adopting
support from other people in Hiring new directors Active p
constellation Last-ditch try to save obstetrics Quality i
Attempt to respect external Closure of obstetrics Selling a
constraints Defining new mission

Consequences Substantive change Substantive change Substantiv

None Obstetrics closed Minimal
Symbolic change Symbolic change Symbolic c
Unfreezing New mission Ambulato
Political change Political change Political ch
Weaker constellation (strategic and Stronger constellation Stronger
organizational uncoupling) Heightened CEO credibility Heightene
Weakened CEO credibility

Phase 2
Focal actors Community health physician DPS, CEO, board CEO, board,

Dominant tactics Aligning with organization Changing organization Leading org

Promotion vs. environment Meeting environment demands Negotiating
Strategic planning with Recruiting new physicians Protecting
developmental focus Control of weak clinical departments Negotiatin
Budget claim with government Balancing budget

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
 Consequences Substantive change Substantive change Substantive ch
None Implementation of new mission Merger agre
Symbolic change Symbolic change Symbolic chan
Belief in teaching mission New mission established Trauma but
Political change Political change Political chang
Building strong constellation, Strong constellation extended to include Weakened
including director of operations, CEO, board, DPS, Medical Council (organi
Medical Council president, DPS, Executive election
board Heightened credibility and authority of Credibili
Heightened credibility and authority DPS-but seen as threat to CEO and old be CEO
(physician becomes DPS and guard
director of Operations)

Phase 3
Focal actors Director of operations, DPS, Medical DPS, Medical Council Executive
Council president, board

Dominant tactics Changing organization Aligning with (new) organization

Negotiating with environment Confronting CEO and board
"Forcing" university mission Promoting development
Balancing budget Contesting legitimacy of CEO
Negotiating complementarity

Consequences Substantive change Substantive change

Affiliation, budget cuts, etc. Loss of dynamism in new mission
Symbolic change Symbolic change
Disenchantment Disenchantment
Political change Political change
Weakened constellation Weakened constellation
(organizational uncoupling) (strategic uncoupling)
Loss of key actors (election, etc.) Loss of DPS, new physicians
Need to rebuild credibility Weakened CEO credibility

a DPS stands for director of professional services.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
824 Academy of Management Journal August

group. Once assimilated, he was able to convince can become disconnected in at least three different
organization members of certain changes, and he led ways:
other members of the constellation in merger negoti- (1) Constellations become disconnected when
ations imposed from outside. However, these negoti- conflicts develop between their members. This is
ations were controversial within the organization, illustrated in Figure 2 by the diagrams for Suburban
and an election led to turnover in the team (without Hospital in phase 1 and Community Hospital in
changing the result of the negotiations). phase 3. Note that although conflict may have a
personal component, in these cases, it developed at
least partly around diverging conceptions of the
Analysis and Theorization relative importance of internal organizational aspi-
rations and external environmental constraints.
Analyses of these stories using the categories we
(2) Constellations become disconnected when
mentioned earlier are presented in Figure 2 and
they become detached from their organizational
Table 5. Figure 2 provides a synthetic portrait of the
bases, leading to possible political turnover in a
evolution over time of the organizations and their
subsequent phase. This is illustrated at Suburban
leadership constellations. To help represent these
Hospital in phase 3 and at University Hospital in
dynamics as clearly as possible, we have simplified
phase 2. The position of the president of the Med-
the diagram of the basic structure of public hospi-
ical Council is most vulnerable to this type of un-
tals in Quebec (Figure 1) to show the pattern of
coupling, which is caused by a perception among
alliances within a leadership group and between it
organization members that the leadership team is
and other stakeholders. In each miniature stylized
not adequately representing its interests.
organizational chart, the shaded large central circle
(3) Finally, constellations become disconnected
represents the key members of the leadership con-
when they become detached from their environ-
stellation. The lower half of each diagram repre-
ments to such an extent that performance declines
sents the organizational base, which includes sev-
and the constellations become vulnerable to polit-
eral internal groups, depicted as midsize white
ical turnover either imposed from outside or insti-
circles. The circle on the left can be viewed as
gated by others (for instance, a concerned board).
representing administrative and support areas,
This had clearly happened to Community Hospital
while that on the right represents the dominant
factions of the medical staff. The environment prior to the arrival of the CEO in phase 1.
These three sources of fragility can be rede-
(ministry, university, other organizations) is repre-
fined in terms of three types of coupling that
sented by the rectangle at the top. Links between
appear to be important in ensuring constellation
the various groups at different phases are shown by
viability over the long term. Strategic coupling
lines and arrows, as explained in the legend. Table
refers to the internal harmony between members
5 adds a tabular comparative analysis of leadership
of a leadership constellation. Organizational cou-
actions and their consequences for each phase.
pling refers to the support the leadership constel-
lation receives from its organizational base. This
is related to the perceived conformity between
Key Observations
the objectives it is pursuing and the interests of
Inspection of Figure 2 and Table 5 leads to fourkey organization members. Environmental cou-
observations that we will develop in our theoriza- pling reflects the degree of coherence between a
tion of the dynamics of collective leadershipleadership
and constellation's vision and aspirations
change. and the demands and constraints of its environ-
First observation: Change and united leader- ment. We use the generic word "coupling" to
ship are associated. The first and simplest obser- indicate support for the goals and initiatives pur-
vation is that periods of major substantive change sued by individuals within the constellation and
tend to be associated with complementary and between the constellation as a whole and groups
united leadership constellations (shown by the inside and outside the organization.3
connections between all the actors within the cen-
tral circle). This state is observed at Suburban Hos-
pital in phase 3, at Community Hospital in phase 2,
3 This notion of "coupling" is different from the notion
and at University Hospital in phase 2. Collective
of "fit" used in the strategic management literature in
leadership seems to be important for change. that it concerns shorter-term political linkages between
Second observation: Constellations are fragile.
specific groups rather than long-term compatibility be-
Despite the apparent importance of "collective-tween structures, processes, and environments (e.g.,
ness," leadership constellations are fragile. They Miles & Snow, 1978; Miller, 1992). It is also narrower

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 825

As illustrated in Figure 2, it was difficult to main- that forms the heart of our process theory of lead-
tain the three levels of coupling simultaneously. A ership and change in pluralistic organizations.
fundamental problem underlying this fragility Third observation: Change is cyclical. Change
seems to be the tension between organizational as- occurs in a cyclical fashion in which opposing
pirations and environmental demands. In a context forces tend to be reconciled sequentially rather
of limited resources, close alignment with organi- than simultaneously. We have just described how
zational aspirations tends to result in a neglect of this happened at University Hospital. The same
environmental constraints, and vice versa. Plural- dynamics are evident at Suburban Hospital. The
istic organizations are by definition settings in period of mobilization and promotion of organiza-
which a multiplicity of actors and groups pursue tional interests in phase 2 was necessary to con-
varying goals. A leadership constellation must try struct a strategically coupled leadership constella-
to accommodate these variations and take actions tion committed to change. Only then could the
that are meaningful to different sets of actors. group
Re- move ahead aggressively to connect its aspi-
rations to the environment. But to achieve this, the
source scarcity provides less potential for a com-
leadership group had to make concessions and ac-
promise that will be satisfactory to all groups. Fail-
cept certain environmental constraints. This pro-
ure to accommodate divergent perspectives means
that organizational coupling will remain fragile.
cess eventually led to organizational uncoupling of
Success in reconciling objectives may come at the
theconstellation. Yet during the short period when
cost of weak coupling with the environment. the constellation remained intact, the change
The fact that members of the constellation derive project was made irreversible and arguably en-
their personal legitimacy from somewhat different hanced the hospital's chances of long-term survival
sources (with managers more concerned about and development. Although in the short term, the
demonstrating fiscal responsibility and physiciansprocess of change slowed down, the organization
more concerned about satisfying their peers) canwould never be the same.
also lead to internalization of the fundamental ten- At Community Hospital, the dynamics were also
sions within the constellation (that is, strategic un- cyclical, but they started from a different position.
coupling). University Hospital, at the beginning of Here, environmental coupling took precedence
the case, was clearly the organization that was clos- over organizational coupling as the new CEO was
est to achieving simultaneous coupling at all three imposed on an organization that had neglected its
levels, perhaps because its size and prestige gave it environment for years. But for the changes to stick,
more power to determine its own environment they had to become rooted in organizational aspi-
(Pfeffer & Salancik, 1978). However, tensions be- rations and capabilities. As soon as a connection
tween environmental and organizational forces was achieved, however, organizational forces be-
were present even here. The new CEO arrived with came more powerful and demanding, an increase
a strong reputation for innovation and initially that eventually led to a break in the constellation
pushed to introduce new ideas from outside, such and collapse of the change process. Yet, for the
as quality management and ambulatory care. He short period when key leaders were working in
quickly realized, however, that his survival in the harmony, an important, irreversible decision was
organization required him to establish his alle- made. In other words, change is possible, but it
giance to medical staff interests first and foremost. proceeds in fits and starts and seems to involve
Only when his organizational legitimacy was estab- see-sawing attention to environmental and organi-
lished was he able to convince them to consider zational preoccupations.
certain changes and to lead them in merger negoti-Fourth observation: Leadership actions affectpo-
litical positions. Finally, the cases suggest that the
ations. This brings us to the third key observation
driving force behind these cycles is the impact of
constellation members' individual or collective ac-
tions on their political positions. Table 5 is an attempt
than the concept of coupling used by Weick (1976) and to summarize these actions and their effects at each
Orton and Weick (1990) in their discussion of loosely
phase for the three cases, focusing particularly on the
coupled systems. These concepts refer to links between
individuals or groups whose actions were most crit-
structural elements of an organization that impose more
ical at each phase. In a context of fluid power rela-
or less consistency in modes of functioning. The organi-
zations we studied were quite loosely coupled in the tionships, the judgments of others concerning the
Weickian sense, but the definition of coupling we use appropriateness of one's behavior are crucial for long-
here is restricted to compatibilities in orientations term
be- survival in a leadership position. Actions that
tween people in a leadership group, other organizationtend to enhance survival prospects can be called
members, and the environment. credibility enhancing, and those that tend to dimin-

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
826 Academy of Management Journal August

ish it are credibility draining (see also Denis et al., potentially opposing forces: organizational objec-
1996). Changes in credibility directly or indirectly tives associated with professional development
affect the capacity of an individual or leadership and environmental pressures and opportunities in-
group to act in the future. Increased credibility wid- cluding financial constraints and political forces.
ens the scope for action. Reduced credibility dimin- In a pluralistic organization characterized by frag-
ishes it and may lead to leader turnover. mented power and multiple objectives, where rec-
Actions can have both process (tactical) and con- onciliation by fiat is not an option, these opposing
tent (strategic) dimensions. In process terms, tactics forces are in constant dynamic tension. Before ex-
can be more or less aggressive, more or less trans- amining to what extent our conceptualization can
parent, and more or less participatory. If a constel- be extended to explaining more complex change
lation is tightly coupled and covers all power bases, (the merger situations), it is worth more closely
aggressiveness, secretiveness, and authoritarianism examining the conditions that make coupling pos-
can be effective in getting things done in the short sible, despite the underlying tensions. Pushing the
term; examples can be seen in the vignettes in the metaphor, the see-saw is not totally unstable in the
tables for Suburban Hospital, phase 3, and Commu- cases we observed. If it were, nothing could ever get
nity Hospital, phase 2). However, in the long term, done. It is as if its base had a certain width of
these tactics are credibility-draining. Indeed, a lack resistance that allowed pressure to build up on on
of transparency was cited several times as a reason side for a while before tipping over. What are th
for leadership turnover. In contrast, greater sensitivity forces that allow temporary stability within inst
to constituents was associated with the creation of bility? What are the forces that allow substant
strategic and organizational coupling. This phenom- organizational change to become irreversible befo
enon is one source of cyclicality. political change makes it impossible?
However, another source of cyclicality is the stra- One potential force for stability is slack resourc
tegic dimension of a constellations' actions. This (Cohen, March, & Olsen, 1972; Cyert & March
dimension has been emphasized in Table 5 and 1963). When our case studies began, the hospit
concerns leadership actors' chosen alignment be- were coming out of a period of relative envir
tween organizational objectives and environmental mental munificence. These circumstances began t
demands. Whether any particular alignment en- change in the late 1980s and early 1990s, the per
hances or diminishes credibility depends on the covered by our cases. Nevertheless, for at leas
relative weights of those who make these judg- while some capacity to promote the organizati
ments. Thus, although alignment with the medical with respect to the external environment remain
staffs' aspirations was generally credibility-enhanc- This capacity gave leaders the time to build u
ing at both Suburban Hospital and University Hos- their credibility internally and create strategical
pital, at Community Hospital, the medical staff's coupled leadership constellations empowered
extreme weakness at the beginning of the case re- move more aggressively. However, although sl
versed this situation temporarily (see Table 5). To resources encourage stability, they are not alw
build a constellation requires emphasizing a type of productive of adaptive change and can lead to sta
alignment that will enhance credibility, enabling nation, pasting over inefficiencies and contrad
members to acquire formal and informal power and tions until they overflow and can no longer
to develop a successful way of working together de- ignored (as in the Community Hospital case).
spite their having different skills and different A second source of stability is internal soci
sources of power. However, the resolution of environ- embeddedness. This refers to the degree to wh
mental and organizational tensions eventually de- leaders become involved in richly interconnec
mands an adjustment in orientation. Because these social networks within their organizations and
demands are often imperfectly compatible, the con- quire tacit knowledge about how things are do
stellation will either fail to deliver on the promises For example, many of the physician leaders e
that allowed it to acquire power, or it will be obliged countered had been with their organizations a lo
to negotiate a compromise. Either way, it risks losing time, knew them well, had developed effective i
credibility: its members may ultimately break ranks ternal networks, and could easily gauge the expe
or lose their positions. Before this happens, however, tations of different groups. Leaders who beha
it may be possible to implement change. according to expectations could be given a cert
latitude. Newcomers would be expected to fit wit
the patterns the old hands had established
Conditions for Coupling
norms. Those who took the time to learn approp
Clearly, this see-saw theory of collective leader- ate behavior patterns and gain the trust of power
ship and strategic change implies a need to connect groups might eventually be given a chance to ma

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 827

an impact (the CEO at University Hospital is an changed. In addition, other organization members
example). Those who did not would be rebuffed may not notice for a while what the leaders are up
(the CEO at Suburban). In these circumstances, a to, or may not believe that their efforts are feasible.
would-be participant in a leadership role constel- For example, interviewees made statements like
lation must first take actions that satisfy other pow- "Some thought it was 'Alice-in-Wonderland' and It
erful people if he or she wishes to enhance his or was a dream that would never get off the ground."
her role. Yet, paradoxically, the very process of Formal position also offers some protection from
acquiring the power to act seems likely to dilute political exclusion, at least for a time. For example,
any actions that may be taken. Social embedded- Medical Council Executive elections occur every
ness is thus a source of stability in constellations one or two years, and CEOs are not instantaneously
that can be used to protect change initiatives. How- fired. These circumstances surrounding leaders
ever, as a force for stability, it is also a barrier to seem a fragile base on which to build a change
change in itself.4 process. And yet, they were apparently useful in
The third source of stability in the see-saw is allowing individual leaders to achieve as much as
more dynamic. It lies in the creative opportunism they did at both Suburban and Community Hospitals.
of constellation members. If they can find a plausi- These four sources of stability in the cyclical
ble space in which a variety of organizational aspi- process of reconciling tensions between environ-
rations and capabilities seem reconcilable with en- mental and organizational forces are not unrelated
vironmental pressures and opportunities, there is to Poole and Van de Ven's (1989) suggested modes
some chance that change can gather sufficient mo- of paradox resolution. The see-saw effect itself il-
mentum to survive. At Suburban Hospital, the com- lustrates the use of time to resolve paradox, and
munity health physician's pursuit of the teaching creative opportunism and the use of slack can be
hospital mission happened to fit with the need of a seen as different means of synthesis that provide
small university's faculty of medicine to broaden space in which connections can be made. We now
its base in order to ensure survival. At Community move on to the examination of the two merger
Hospital, the proposed new emergency mission situations in an attempt to determine to what extent
happened to be of interest to other local hospitals the themes emerging from our comparative analysis
with overflowing emergency rooms. At University of the first series of cases are useful in the context of
Hospital, the move to ambulatory care could be second-order change.
sold as a way to increase patient volumes (and
physician revenues) while meeting budget con-
straints. These orientations all had costs and un-
COLLECTIVE LEADERSHIP AND SECOND-
popular consequences for certain groups, and the
reconciliation between environmental demands ORDER CHANGE

and organizational aspirations was imperfect,Context

as for Change
subsequent events showed. However, they gave the
In February 1995, the Quebec Minister of He
see-saw more potential stability than (for example)
a simple proposal to cut budgets without any and Social Services announced a proposal to m
obvious corresponding gain (Suburban Hospital, three large teaching hospitals in each of two m
phase 1). cities. There was general agreement that there
The fourth source of stability is a useful but tem- too many teaching hospitals, that consolida
porary support for the third. It is related to time, was required to generate the critical mass nece
to support highly specialized services, and
inattention, and the protection of formal position. It
takes a certain amount of time for leaders to be there was also a need to move health care de
toward an emphasis on ambulatory care. How
which hospitals would become part of the
"academic medical centers" and the precise
4 Note that our notion of social embeddedness implies
a high density of ties between leaders and other organi-
sion of roles was naturally a very contentious
zation members. It is different from the notion of organi-
Committees were established in both cities under
zational embeddedness used by Granovetter (1985),
the leadership of senior public servants to negoti
which refers primarily to the embeddedness of focal or-
the terms of the two mergers. These negotiati
ganizations in larger institutional and moral structures.
were
Of course, our organizations were also deeply embedded
extremely difficult, more so for the Metrop
itan
in Granovetter's sense, because of the importance of the Merger than for the Capital Merger. Howeve
agreements
health care network and the strong institutional forces to merge were obtained under pressu
from
(including regulations and professional norms) permeat- the government, and new integrated boa
ing it. were formed, in October 1996 for the Metropolit

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
828 Academy of Management Journal August

Merger and in December 1995 for the Capital tion in which the Medical Staff Executive was dom-
Merger. inant (the Superstar), and one had a constellation
In principle, the formal structural framework for in which the administration was strongest and
collective leadership at each of the merged institu- where the Medical Council Executive played a pas-
tions was exactly the same as that we described sive role (the Street Worker). Finally, the third
earlier (see Figure 1). The difference, of course, was showed more balance between poles of influence
that now this same structure capped three different (the Aristocrat). These labels, which were inspired
operating institutions on three different sites, each by interviewees, reflect the history and strengths of
with its own traditions. In addition, both merged the three organizations. The Superstar site, the hos-
institutions initially added a certain number of in- pital with the most entrepreneurial medical staff
house rules promoting equitable representation of and the strongest academic credentials, saw itself
the three sites on the board and the Medical Coun- as the leader in the merger. The Aristocrat was the
cil Executive, and (less explicitly) on the adminis- oldest institution and was perceived as serving an
trative team. To the extent that people from the elite clientele. The Street Worker served a poorer
three sites were equally represented within each downtown population and was the weakest aca-
individual pole of leadership, the leadership team demically.
can be described as "holographic," including
within itself the preexisting tensions between the
sites.5 An extra dimension of "collectiveness" has
Case Descriptions
potentially been added. Using the same approach as for the cases of first-
The challenge of the mergers was not simply one order change, we provide phase-by-phase process
of governance, however. Each merger involved the descriptions for the two merger stories in the form
rationalization of activities among the three sites, of vignettes with accompanying interview citations
thus requiring "micromergers" between myriad in Tables 7 and 8. Below are brief synopses of these
clinical services currently operating separately and stories.
demanding the fundamental transformation of the The Metropolitan Merger occurred in a highly
mission of some or all of the sites; for instance, conflictual context. A board with equal representa-
transformation to an ambulatory care center had to tion from each hospital was set up to implement a
occur in one case. Thus, besides maintaining three rather unrealistic merger agreement (the Protocol).
operating institutions and learning to work collabo- The board hired a civil servant CEO who built up
ratively with former rivals, the leaders had to im- the administrative team. However, she was ham-
plement fundamental, second-order change. pered in change efforts by budget cuts, dissension
Before describing how the leadership constella- among board members and physicians, and con-
tions in each of the two mergers evolved and how frontation with the Medical Council Executive and
they attempted to move toward integrating and de- the unions. After a change in Medical Council pres-
veloping their organizations, it is important to re- ident, she created a proposal for a new ambulatory
call our earlier argument, that despite hospitals' care center and joined with physicians in demand-
having common formal structures, each develops ing government resources for development. This
its own negotiated order between poles of leader- strategy was partly successful, and some resources
ship that become embedded in organizational tra- were promised. With this, she began to build ten-
ditions, illustrating Stinchcombe's (1965) notion of uous support. However, the overall project was still
"imprinting." Clearly, one important background contested by influential internal and external
for understanding the evolution of leadership roles factions.
in these mergers is the nature of these preexisting The Capital Merger began more propitiously as
modi vivendi. Table 6 provides brief profiles of the the participating hospitals were initially more col-
six hospitals involved in the two mergers. As can laborative. They deliberately tried to avoid perpet-
be seen, there are remarkable similarities between uating intersite rivalry by naming a neutral board
the two cases, but sharp differences between the and an outside CEO. He developed links with re-
sites within each merger. One site had a constella- gional bodies and promoted a development project
that would make the merged hospital the central
component of a new academic medical center in-
5 The word "holographic" is used metaphorically, corporating elements from other city hospitals.
as suggested by Morgan (1986), to represent an entity This move was popular internally and helped pro-
vide time to build an administrative base and links
in which the whole is reproduced in each of its parts.
(Holographic images produced by laser technology have with medical staff. However, it was unpopular with
this property.) the other city hospitals, which lobbied against the

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 829

TABLE 6

Profiles of Merging Hospitals

Hospital" Metropolitan Merger Capital Merger

The Superstar
History and clientele Founded in 1880 as a lay hospital to give home- Founded in 1968 in Upper Town to
grown francophone medical education. concentrate all medical teaching on premises.

Strengths and weaknesses Full coverage of all specialties; strong in Widest coverage in city; strong in teaching and
specialty teaching, moderate in research. research.

Leadership constellation Doctor's workshop: Domination by physicians Doctor's workshop: Domination by physicians
with complicity of management and board. with complicity of management and board.

The Aristocrat
History and clientele Founded in 1642 by nuns. First medical Founded in 1639 by nuns. Historical links with
education in the city. Perceived elite university. Upper Town society clientele.
clientele.

Strengths and weaknesses Had lost several key specialties, stagnating. Strengths in some specialties; strong in
Moderate in teaching, strongest in research. teaching, moderate in research.

Leadership constellation Fragile partnership: Management based on Mature partnership: Management based on
tradition, rigor, and hierarchy with tradition, rigor, and hierarchy with
participation of medical staff (but weakened participation of medical staff; politically
by recent divisions over hospital move). influential board.

The Street Worker

History and clientele Founded in 1908 to care for indigent children Founded in 1914 in Lower Town, poorer
and poor. Poorer downtown clientele. clientele.

Strengths and weaknesses Specialties related to poverty, drug addiction, Maternity care; moderate to weak academic
and prostitution; moderate academic role. role.

Leadership constellation Managerial fiat: Centralized management, Managerial fiat: Centralized management,
interventionist board, passive Medical informal style, passive Medical Council
Council Executive. Executive.

a "Superstar," "Aristocrat," and "Street Worker" designat

Metropolitan Merger, the Superstar was the University H

project and succeededas in blocking

the result it.
of an irresistable drive from This
a powerful d
government (a form
pointment led to disaffection by of certain
coercive isomorphism,
physi in
DiMaggio
groups and ultimately to and Powell's [1983] terms) and
a government a trans-
proposa
actional bargaining process
undo the merger and reconfigure that produced
city a mo-
hospital
totally different way. mentarily acceptable compromise. In fact, transac-
tional negotiation and momentary compromise-
rather than the creation of a unified leadership
Analysis and Theorization
constellation committed to change-has been the
These stories share certain obvious characteris- pattern in this more complex situation. The num-
ber of interests and actors that need to be mobilized
tics with the stories of first-order change but are
different in other ways. We will now return to the
to cover all poles of authority, expertise, and legit-
four themes identified in the first-order cases to imacy has suddenly multiplied to such an extent
examine to what extent they are reproduced in
that even temporary integration into one comple-
these more complex situations. mentary and unified team has become a virtual
Theme 1: Substantive change and collective impossibility. Although links between the two
leadership constellations. The only significant CEOs and their board chairs have helped stabilize
substantive changes that we observed were the for- their positions, it has been difficult for these hos-
malization of the mergers themselves in the cre- pitals to build more complete constellations be-
ation of unique (but holographic) governance struc- cause as soon as an individual leader has an alli-
tures. These events cannot be explained as the ance with one potential coleader, he or she
result of a united leadership constellation but more becomes detached from another. Links are there-

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
830 Academy of Management Journal August

TABLE 7

Case 4, Metropolitan Merger: Building a Star Hospital from Bitter Rivalsa

Vignettes Interview Extracts

Phase 1. Forcing a merger agreement (The "protocol") -> Power to a holographic board
1995-96. A merger "agreement" was imposed after a year of acrimonious negotiations The merger wasn't a marriage of love, it
led by a well-known public administrator. A broad consulting process with wasn't even a marriage of reason. It
physicians was led by the dean of an affiliated medical school. One factor ofwas a forced marriage.
convergence was competition from another university hospital network. The
The merger of [the other network] was
agreement was signed by all parties under government pressure. It was precise in
mentioned a lot: "Look at them: they're
some ways but left key issues unresolved to enable all parties to sign. It designated
ahead of us."
the Aristocrat an ambulatory care center, the Superstar a site for specialized services,
and the Street Worker the site for basic medical education. It also included a There was so much pressure to sign the
government commitment of resources for development. Substantively, the merger was protocol that major problems weren't
finalized. Symbolically, the protocol explicitly recognized the continued existence of settled.
three sites and provided a strategic orientation that had been formally accepted (but
not internalized by all). Politically, power passed to a transitional board with equal
representation from the boards and the medical staff of the three sites.

Phase 2. Drifting and divergence -- Recruitment of outside (civil servant) CEO

1996-97. After the merger agreement, the transitional board (headed by an Aristocrat The board took up more and more space.
leader) did not name an interim CEO but gave "horizontal" tasks to the three ex- They're the same people that
CEOs. A joint Medical Council was also formed with equal representation of the negotiated the protocol. It'll take a
hospitals but with a Superstar president. A planning process was initiated, but this strong CEO to rebalance things.
holographic form of leadership perpetuated divisions among the three institutions.
Everything drifted as before. Each site
Because of these divisions, none of the ex-CEOs could be appointed CEO, and
tried to get away with things here and
recruiting was difficult. After nine months, the board hired a senior civil servant
there. It was a period of limbo.
from the Ministry of Health, despite opposition from medical staff members of the
board. Substantively, this period produced no changes. Symbolically, the confusion
heightened uncertainty about the future. Politically, the board established itself as the
major strategic actor in the new constellation, showing its willingness to confront the
medical staff.

Phase 3. Conflict, paralysis, return to the protocol -> Contested leadership

1997-98. The new CEO chose an administrative team and confirmed departure fromThe thenew CEO was highly contested by
Superstar's physician-dominated approach by refusing to submit to pressure from thecertain destructive forces, ... a residue
Medical Council president. A consultative committee with newly appointed clinicalof the old "doctor power."
department heads was formed in an attempt to secure an alternative form of
People were very disappointed. We
collaboration with physicians. Budget cutting was a priority, and conflicts with
looked for the best solution. Everyone
unions were frequent. Various planning efforts were initiated to implement the
worked very hard. But the board chair
merger but were stopped at the board level by reaffirmation of the protocol.
announced, "We're sticking to the
Substantively, two departments were moved in crisis. Symbolically, the protocol was
protocol."
seen as both the only source of stability and as a block to progress. Politically,
although the CEO's position was still contested, she maintained the support of the
board chair. The Medical Council president's destructive approach lost him
credibility, and he was replaced in elections.

Phase 4. Promoting development -- Groping toward leadership stability

1998-99. With the help of a consulting firm, the CEO developed a major project forI thoughtan the CEO had jumped the fence
ambulatory care center. This initiative gave value to the role of the Aristocrat in the when I read her letter to the regional
merger while enabling its people to make concessions departing from the position board.
they had adopted in the protocol. Again, a process of consultation with physicians
Unfortunately, I haven't seen more
was implemented. With this project, for the first time the CEO also took on the role
gestures like that. But we did a lot of
of promoter of physicians' demands for new financial resources from the government.
work since August and I think we came
Actions had a positive symbolic impact (expectations of investment, innovative
out of it stronger.
ambulatory care center) and some negative impacts (fear of marginalization in the
other two hospitals). Substantively, little changed; indeed, the board was unable to
agree to a transition plan. Politically, there were the beginnings of a fragile alliance
between the CEO, the board chair, and some physicians, but the board and Medical
Council Executive were still holographic.

a "Superstar," "Aristocrat," and "Street Worker" designate a hospital's ethos, not its identity. See footnote a, Table 6.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 831

TABLE 8
Case 5, Capital Merger: Reconfiguring Academic Medicine in the Citya

Vignettes Interview Extracts

Phase 1. Constrained collaboration -* preparing the ground for new leadership

1995-96. The three hospitals involved in this merger were relieved to be among those
There was a good base in the ground before
chosen for academic medical center status and worked together to prepare thethe new CEO arrived. We proceeded
ground. The merger was supported by the government in spite of resistance from
harmoniously. A sort of framework was
other city hospitals and ambiguous support from the university involved. Thedefined beforehand, the mother-child
participants in negotiations favored equity by creating a board with no members
center.

from the old boards and by choosing a Medical Council president on a rotation
We have alliance problems even with the
principle. The retiring CEO of the Street Worker was named interim CEO and
worked with the new board to create a vision of an academic medical center with a university.
consolidated mother-child center at the Street Worker site. Internal resistance was
low: there was a wait-and-see attitude. At the end of the period, a new CEO with
merger experience was hired from outside. Substantively, the merger was made
official. Symbolically, inclusion in the academic medical center and the new
mother-child center provided a stimulating vision. Politically, the emphasis on
equity created a promising context for the new leadership.

Phase 2. Building regional vision for the hospital -> Creating a context for
stabilization

1996-97. The new CEO had good contacts with regional government. He negotiated The new emerging culture is oriented by
budget concessions and participated in regionwide planning that would make the the CEO and by the objectives: a great
merged institution into the heart of a unified academic medical center including North American teaching hospital.
parts of other city hospitals. With board support, he built his internal legitimacy by
promoting this concept, selecting an administrative team according to the equity The Medical Council president is of
principle (including all three ex-DPS),b and creating links with the medical staff.tremendous help. He is involved in many
Symbolically, the actions produced a positive vision of the future, although issues. There were many meetings with
substantive changes were to come. Politically, the board and CEO were united, and physicians.
administrative and medical bases were being reconstructed. However, other
hospitals were not happy with the emerging regional plan that threatened their
autonomy and/or survival.

Phase 3. Setback from outside pressures -* Struggling to rebuild confidence

1997-98. Efficient resistance of other hospitals in the city created a break with theSome of our doctors-those not happy with
ministry and regional board. The academic medical center and the location of thethe location of the mother-child center-
mother-child center were questioned. This generated dissatisfaction among some are lobbying with government and
physician groups, who began to lobby government directly and negotiate with other negotiating with hospital X.
hospitals. In the face of uncertainty, the board and the CEO attempted to save the
situation by tightening deadlines and establishing division of roles between sites.When things started to slow down, the
Medical Council President decided to
The Medical Council president took on an important role in attempting to maintain
take matters into his own hands ... there
commitment among medical staff, and he developed direct collaborative links with
the CEO (weakening the role of the DPS). Symbolically, the vision appeared more is a power struggle between the DPS and
Medical Council Executive.
distant, although the salvage actions helped maintain some confidence.
Substantively, there were still no changes. Politically, the Medical Council
Executive began to play a more important role in a somewhat weakened leadership
team.

Phase 4. Attack from environment -> Greatly weakened leadership

1998-99. Influenced by dissatisfaction among other local hospitals and by doctors
The regional board is going back on
opposed to the projected site for the mother-child center, the regional board previously made decisions. .... They are
announced a plan to break up the Capital Merger and reconfigure local hospitals
now working on a consortium without
into an entirely new combination. This bombshell destabilized both administrative
the participation of the hospitals.
leadership (many departures) and medical commitment (departures and direct
negotiations with other hospitals). The board and CEO worked to defend the
position of the Capital Merger, supported by frontline staff. Substantively and
symbolically, the merger lost much of its momentum. Politically, the CEO's
legitimacy was compromised.

a "Superstar," "Aristocrat," and "Street Worker" designate a hospital's ethos, not its identity. See footnote a, Table 6.
b DPS stands for director of professional services.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
832 Academy of Management Journal August

fore tenuous and reflect the interests of the mo-

environmental forces are no longer easily distin-
ment. Bargaining and voting may allow decisions
guished. Rather, leaders are struggling with the rec-
to pass on a given day. But the lack of profound
onciliation of multiple tensions structured only
commitment from a powerful concerted group partly by organizational boundaries. The ambigu-
makes irreversibility elusive and consistency of vi- ous role of physicians on the regional scene in the
sion unlikely. Capital Merger case is illustrative.
Theme 2: Fragility and the three modes of cou- Themes 3 and 4: The cyclicality of change and
pling. As a first approximation, the idea that three the political effects of leadership actions. The
different but opposed forms of coupling are needed cases and the previous discussion confirm the cy-
to enable change is reinforced by these new cases. clical nature of change processes and also the po-
For example, we see from the Capital Merger how litical effects of leadership actions as drivers of
the CEO and board's initial expansionist vision of these cycles. Leaders' credibility rises and falls de-
the hospital produced the beginnings of strategic pending on process tactics and strategic alignment
and organizational coupling and how this then col- with various interests, and this variation in turn
lapsed as soon as it became clear that the environ- affects their capacity to act in the future. However,
ment would not cooperate. For the Metropolitan there seems to be much greater fragility and insta-
Merger, we see how the new CEO was only able to bility in the cycles. Vast amounts of energy are
begin to build confidence internally once she had consumed in processes that produce tiny effects
aligned herself explicitly with organizational inter- that are then overturned or overwhelmed. The
ests against the environment. many opposing tensions to be reconciled ag
However, the cases also place in dramatic relief seem to lie partly behind this pattern. Stretchi
the added challenges of second-order change. In metaphor, the seesaw with one degree of freedo
fact, the first effect of the mergers was to destroy or replaced by a plane balanced on its center p
weaken both strategic and organizational coupling that can be destabilized from any number of d
while simultaneously making such coupling all the tions. The stabilization of change has become m
more necessary. We have already mentioned the more difficult in this hyperfluid, hyperplurali
enormous challenge of achieving strategic coupling context. This argument leads us to reexamine
among three or more poles of leadership in a situ- four conditions for coupling that we identified
ation where each of the three poles reflects within the analysis of the first-order cases.
itself the divisions between three old institutions.
However, the mergers also destroyed organiza-
Conditions for Coupling
tional coupling. On day one of the merger, all in-
dividuals in transorganizational leadership posi- The four factors that we identified as enab
tions, whatever their origins and whatever their the sequential reconciliation of opposing forc
level, were uncoupled from at least two-thirds ofthe first-order cases were slack, social embedded-
their new organizations. In a merger, the internalness, creative opportunism, and time, inattention,
legitimacy of all individual leaders and any poten-and formal position. A more pluralistic context cre-
tial team needs to be completely reconstructed, and ates an even greater need for stabilization. How-
the interests and aspirations of more heterogeneous ever, for the second-order cases we observed, these
groups must be reconciled. Our emerging theoriza- sources of stability were either missing or beyond
tion would suggest that only dramatic support orthe capabilities of the organizations at the time we
tolerance from the environment would provide the observed them.
time and resources necessary to build a legitimate For example, it is well established that mergers
and harmonious leadership constellation capable need investment and slack resources (Haspelagh &
of achieving substantive change. In the cases weJemison, 1993). Yet the first item on the agenda for
described, such support was not there. Indeed, both mergers was budget cuts. Indeed, the mergers
since part of the goal of both mergers was rational-existed partly because of a belief that they would
ization, environmental pressures tended to hinderproduce rationalization.
rather than help. We have already mentioned the dislocation of
These second-order-change situations also lead leaders from their organizational base (that is, the
us to view modes of coupling in a more complexabsence of embeddedness) as a source of instability
manner. As the organizations get larger, the distinc- that hampered progress. Leaders who try to play a
tion between organization and environment be- horizontal role simply do not know how most of
comes increasingly artificial. Leaders are playing in their new organization functions. Moreover, if they
a supraorganizational field in which the bound- do have roots in one of the original organizations,
aries of their action are unclear. Organizational and they are deeply distrusted by the others. Time and

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 833

resources are needed to build new organizational Hodgson and colleagues (1965), is often fo
roots. in a literature that still tends to glorify ind
heroes (Meindl, Erlich, & Dukerich, 1985).
Creative opportunism is still an available option
in this type of change, but it is much more difficultSecond, unified collective leadership is
to achieve than in the earlier cases, given the sary but is always fragile in a context of
broader range of internal and external forces that power and multiple objectives, where lead
must be simultaneously or sequentially reconciled. at least partly by the consent of the led. In
This approach was nevertheless used by the CEO atular, a constellation may be shattered by
the Capital Merger when he participated in the rivalry (strategic uncoupling), by dislocation
development of a plan for a regional academic med- its organizational base (organizational unco
ical center. However, the project that he helped or by inadaptiveness to the needs of the e
create was so big that it threatened the autonomy ment (environmental uncoupling). We argu
and survival of other local organizations, generat- particularly in times of change, it is diff
ing a vicious backlash that produced negativemaintain the three forms of coupling sim
rather than positive change. neously because of tensions between envir
Finally, in this fluid situation, time pressure istal pressures, organizational aspirations,
intense, leaders are constantly under the micro- preferences of incumbent leaders. Sometim
scope both internally and externally, and, in manysions may appear between the desire to m
cases, their positions grow more fragile. the integrity and harmony of a constellat
In summary, these mergers combine multiplica- the need to introduce people with new sk
tion of the dimensions of pluralism with removal experience
of to respond to new environmen
some of the key forces for stability that keep plu- lenges. At other times, the leadership conste
ralistic organizations from collapse. We do not may attempt to respond to external pressu
know how the mergers will turn out in the end, but lose its alignment with its organizational
this summary conveys how we see them now. Thisvice versa). Finally, partly because of their di
is as far as we have come in our research journey. skills and sources of legitimacy, individua
Hopefully, as we keep on following events, we willbers of a leadership constellation may develo
learn more. Indeed, if Weick (1999) was right about ferent conceptions of the appropriate course
the value of real-time research, the record we have tion, leading to conflict and an eventual cha
set down here and now should one day enable us tomembership.
increase our understanding, when we go back toThird, because of the difficulty of main
compare it with what actually happened next. alignment at all three levels, we argue that
in pluralistic organizations tends to occur
DISCUSSION AND CONCLUSIONS clical manner in which opposing pressures ar
onciled sequentially rather than simultan
We began this article with a simple research Cyclical or sporadic patterns of change ha
question: How can leaders achieve deliberate stra-
empirically observed in other pluralistic s
tegic change in organizations where strategic lead-
(Hinings & Greenwood, 1988), and the idea
ership roles are shared, objectives are divergent, ing sequential attention to contradictory for
and power is diffuse? Our response has six meansmainto enable change has its roots in Cy
components that we will now attempt to March's summa-(1963) behavioral theory of the firm
rize in a more general form. The six components also been popularized as a practical manag
represent an emerging process theory of leadership approach by Hampden-Turner (1990) and o
and strategic change in pluralistic settings. Fourth, First, we argue that the effect of lead
we argue that major substantive change in plural-
tions on their political positions drives c
istic organizations is more likely to be establishedchange. To build up the organizational and
under unified collective leadership in which gic each
coupling necessary to allow substantive
member of a "leadership constellation" playsina adis- pluralistic setting may require orienting
tinct role and all members work together harmoni- actions around promotion of an organiza
ously. The logic behind this idea is that if no single
least initially. But concrete implementation
individual can impose his or her vision and pref-
ganizational aspirations demands negotiatin
erences on an organization and bring about sub-
environmental forces. If leaders are perce
stantive change, a team assembling a variety of too many concessions or, alternativ
offering
skills, expertise, and sources of influence and legit-
their projects are seen as different in reality
imacy may be able to achieve such change. whatThiswas promised in theory, these perc
simple and attractive idea, suggested originally will inby turn decrease leaders' credibility and

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
834 Academy of Management Journal August

slowdown of change. The variety of goals pursued difficult to achieve irreversibility in such a fragile
by individual organization members increases the context. Similarly, we noted how sequential atten-
likelihood that certain groups will ultimately see tion to different goals is also rendered more com-
leaders' actions as unacceptable. This relationship plex under extreme pluralism by the enormity o
between leaders' actions, their political effects, and the proposals that must be put forward to reconcil
cyclical patterns of change was very evident in the a range of heterogeneous interests. We argue tha
cases we studied. It recalls what Blau (1964) called increased pluralism thus increases the need for
the "leadership dilemma" (see also Denis et al., counterbalancing sources of stability, such as slack
1996). Leadership is most effective when it is legit- social embeddedness, creative opportunism, and
imated by the approval of the led (see also Pfeffer, time, inattention, and formal position. (Unfortu
1992). Yet forceful leadership action is likely to be nately, our empirical observations of the merger
somewhat incompatible with social approval. Blau indicate that conditions are exactly the opposite.
(1964) suggested that sequential rather than simul- The organizations we studied were professiona
taneous insistence on the two aspects of leadership bureaucracies (Mintzberg, 1979) or "value-rational
is one solution to the dilemma. This is the essence organizations (Satow, 1975). These organizations
of our argument here. are inherently pluralistic, permeated by a variety of
Note that exogenous factors may also create dis-external professional influences and norms that ar
continuities. For example, an external event may
not always reconcilable with efficiency objective
transform the context for action, as occurred when (DiMaggio & Powell, 1983). As organizational
the government intervened in the initiation of the boundaries become increasingly fluid and as
two mergers. Or individual leaders may withdraw knowledge workers become more important in
for reasons that have nothing to do with their prior many areas of the economy (Lowendahl & Revang
actions (reasons like illness or retirement). Our 1998), the conditions of pluralism we observed are
point is simply that leaders' actions do often have likely to become more and more common. This
political as well as substantive and symbolic con- study provides some indication of how change may
sequences and that this politicality contributes to proceed in contexts where no one has full control,
generating cyclical change processes. where divergent objectives are pursued by different
Fifth, we argue that, despite the presence of op- groups, and where the legitimacy of change agents
posing forces, four factors can contribute in differ- and change initiatives cannot be taken for granted.
ent ways to the stabilization of change in a plural- To what extent does the process theory we have
istic setting: slack; social embeddedness; creative offered here add to, contradict, or confirm other
opportunism; and time, inattention, and formal po- conceptualizations of change processes in pluralis-
sition. Slack makes the reconciliation of contradic- tic settings? Clearly, certain aspects of our discus-
tory forces easier; social embeddedness ensuression recall Cohen and colleagues' (1972) garbage
both realism in change proposals and a certain can theory of organizational choice. Indeed, we
loyalty among colleagues; and creative opportun-found the term "organized anarchy" to have great
ism allows the invention of win-win proposals.resonance among several of the people we inter-
Time, inattention, and formal position sometimesviewed. The randomness and transience of organi-
allow leaders to survive controversial leadership
zational decision making described by Cohen et al.
actions long enough to help change occur. How- (1972) are evident in the merger situations we stud-
ever, almost by definition, in a pluralistic organi- ied. Moreover, the conditions of slack and of time,
zation, this factor offers weak protection. In con- inattention, and formal position also form part of
trast, formal position is the base from which many both Cohen and colleagues' and our conceptualiza-
unitary hierarchical organizations plan and imple- tion. However, our analysis suggests much greater
ment strategic change. dynamic continuity in issues and in actors, and it
Finally, through the merger case studies, we ex-emphasizes the role of creative individuals and
amined the effect on the capacity to achieve change committed unified groups in proactively moving to
of multiplying the number of pluralistic dimen- make change possible. From this point of view, our
sions. We argue that extreme pluralism adds to theperspective is closer to that put forward by King-
difficulty of forming unified leadership constella-don (1984). He argued that in public policy arenas,
tions because no group can unite all sources of streams of problems, policies (solution ideas), and
politics (trends in public opinion) are "coupled"
power, expertise, and legitimacy and still remain
grounded in its own organizational base. Leader-
together by "policy entrepreneurs" knowledgeable
ship instead tends to be only partially coupled, andenough to profit from "windows of opportunity" to
change staggers through a series of shifting alli-
get the ideas onto policy agendas. However, King-
ances and momentary compromises. However, it is don did not explore the cyclicality of the change

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 835

dynamics associated with the role of policy entre- tive elites linked to organizational bases). There is
preneurs and the political effects of their actions. need to broaden the collective leadership persp
Complexity theory offers another theoretical per- tive vertically as well as horizontally to people an
spective from which organizational researchers processes at other levels who are crucial to stim
have begun to analyze change in pluralistic con- lating and implementing substantive change.
texts (Browning, Beyer, & Shetler, 1995; Stacey, fact, change leadership in pluralistic organizati
1995). Underlying this theory are the notions of could be viewed as a cascading process that
opposing forces, nonlinear relationships, and feed- volves chains of interlocking role constellation
back loops that produce unpredictable and cyclical different levels. There is a need for much more
patterns of change. Complexity theorists argue that attention to the flow of leadership and change
a great error many managers (and researchers) make through organizations. As the two mergers de-
is their attempt to impose structure and simplicity scribed in this article are pursued and as attemp
on complex situations, because pressures at one are made to integrate hospitals at the clinical oper-
point will generate backlashes at another and per- ating unit level, our own research journey will
haps amplify the cycles that one was attempting to likely proceed in this direction while continuing to
control (Stacey, 1995). Our perspective is that in develop our understanding of the dynamics of stra
itself, complexity theory is a metaphor that could tegic leadership and change under conditions o
be applied to the situations we describe but that it extreme pluralism and complexity.
needs to be fleshed out to improve understanding
of organizational processes. Without deriving our
inspiration from this source, we have described in REFERENCES
concrete terms here how opposing forces, nonlin-
ear relationships, and feedback loops manifested Abbott, A. 1988. The system of professions. Chi
themselves in a real organizational setting. We have University of Chicago Press.
also tried to show that there are ways in which Barley, S. R. 1986. Technology as an occasion for str
change can be stimulated, but not entirely con- turing: Evidence from observations of CT scann
trolled, provided certain forces for stability are and the social order of radiology departmen
present. Administrative Science Quarterly, 31: 8-108.
Although providing some new perspectives on Blau, P. 1964. Exchange and power in social life
the dynamics of collective leadership and strategic York: Wiley.
change, this article leaves several issues for further Browning, L. D., Beyer, J. M., & Shetler, J. C.
research. For example, we have barely touched on Building cooperation in a competitive indust
the personal and interpersonal dynamics within SEMATECH and the semiconductor industry.
collective leadership constellations. In doing this Academy of Management Journal, 38: 113-151.
research, we encountered a set of extremely color- Bucher, R., & Stelling, J. 1969. Characteristics of profes-
ful and interesting people whose views of the sional organizations. Journal of Health and Social
world and whose personalities clearly had a major Behavior, 10: 3-15.
effect on evolving events. We deliberately avoided Bunderson, S., Lofstrom, S., & Van de Ven, A. 1998.
dealing with such microlevel variables by focusing Administrative and occupational models of organ-
on what leaders did, not on what they were. How- izing. Working paper, Carlson School of Manage-
ever, there is a clear opportunity for further re- ment, University of Minnesota, Minneapolis, St.
search into these causally prior concepts. Such in- Paul.
vestigation should extend beyond correlational Burt, R. S. 1992. Structural holes. Cambridge, MA: Har-
studies of team consensus, diversity, and perfor- vard University Press.
mance (e.g., Priem, 1990; West & Schwenk, 1996) to
Cohen, M. D., & March, J. G. 1986. Leadership and
the processual and biographical aspects of team ambiguity. Boston: Harvard Business School Press.
evolution and interaction over time. In addition,
although our analysis implicitly suggests a relation- Cohen, M. D., March, J. G., & Olsen, J. P. 1972. A garbage
can model of organizational choice. Administrative
ship between change leadership and career reversal
Science Quarterly, 17: 1-25.
in pluralistic organizations, we examined the ef-
fects of these events on the organizations, not on Cyert, R. M., & March, J. G. 1963. A behavioral theory of
the firm, Englewood Cliffs, NJ: Prentice-Hall.
the people who lived them. There is room for more
research on how change leadership affects change Denis, J.-L., Lamothe, L., & Langley, A. 1999. The struggle
leaders as individuals. to implement teaching hospital mergers. Canadian
Finally, our research focused mainly on the do-Public Administration, 42: 285-311.
ings of fairly narrow leadership elites (albeit collec-
Denis, J.-L., Langley, A., & Cazale, L. 1996. Leadership

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
836 Academy of Management Journal August

and strategic change under ambiguity. Organization Kipnis, D., Schimdt, S., & Wilkinson, I. 1982. Intraorga-
Studies, 17: 673-699. nizational influence tactics: Explorations in getting
one's way. Journal of Applied Psychology, 65: 440-
Denis, J.-L., Langley, A., & Pineault, M. 2000. Becoming a
452.
leader in a complex organization. Journal of Man-
agement Studies, 37: 1063-1103. Langley, A. 1999. Strategies for theorizing from proces
Denis, J.-L., Langley, A., & Prenevost, M., & Cazale, L.,
data. Academy of Management Review, 24: 691-
710.
1994. Leading strategic change in organized anar-
chies. Paper presented at the Strategic Management Leonard-Barton, D. 1990. A dual methodology for case
Society Conference, Paris. studies: Synergistic use of a longitudinal single site
DiMaggio, P., & Powell, W. W. 1983. The iron cage revis-
with replicated multiple sites. Organization Sci-
ence, 1: 248-266.
ited: Institutional isomorphism and collective ratio-
nality in organizational fields. American Sociologi- L0wendahl, B., & Revang, 0. 1998. Challenges to existin
cal Review, 28: 147-160. strategy theory in a post-industrial society. Strateg
Doz, Y. L. 1996. The evolution of cooperation in strategic
Management Journal, 19: 755-774.
alliances: Initial conditions or learning processes? Meindl, J. R., Erlich, S. B., & Dukerich, J. M. 1985. Th
Strategic Management Journal, 17(special issue): romance of leadership. Administrative Science
55-83. Quarterly, 30: 78-102.

Meyer, A. in
Eisenhardt, K. M. 1989. Making fast strategic decisions D., Brooks, G. R., & Goes, J. B. 1990. Environ-
mental jolts and industry revolutions: Organiza-
high-velocity environments. Academy of Manage-
ment Journal, 31: 543-576. tional responses to discontinuous change. Strategic
Management Journal, 11(special issue): 93-110.
Finkelstein, S., & Hambrick, D. 1996. Strategic leader-
ship. St. Paul: West. Meyer, J. W., & Rowan, B. 1977. Institutionalized organi-
Fox-Wolfgramm, S. J., Boal, K. B., & Hunt, J. G. 1998.
zations: Formal structure as myth and ceremony.
American Journal of Sociology, 83: 340-363.
Organizational adaptation to institutional change: A
comparative study of first-order change in Miles, M. B., & Huberman, A. M. 1994. Qualitative data
prospec-
tor and defender banks. Administrative Science analysis. Newbury Park, CA: Sage.
Quarterly, 43: 87-126.
Miles, R. E., & Snow, C. C. 1978. Organizational struc-
Freidson, E. 1970. Professional dominance. New York: ture and process. New York: McGraw-Hill.
Atherton.
Miller, D. 1992. Environmental fit versus internal fit.
Giddens, A. 1984. The constitution of society. Berkeley: Organization Science, 3: 159-178.
University of California Press. Mintzberg, H. 1979. The structuring of organizations.
Hambrick, D. C., & Mason, P. A. Upper echelons: The Englewood Cliffs, NJ: Prentice-Hall.
organization as a reflection of its top managers. Mintzberg, H. 1997. Towards healthier hospitals. Health
Academy of Management Review, 9: 193-206. Care Management Review, 22(4): 9-18.
Hampden-Turner, C. 1990. Charting the corporate Morgan, G. 1986. Images of organization. Thousand
mind. New York: Free Press.
Oaks, CA: Sage.
Haspelagh, P. C., & Jemison, D. B. 1991. Managing Orton, J. D. 1997. From inductive to iterative grounded
acquisitions: Creating value through corporate theory: Zipping the gap between process theory and
renewal. New York: Free Press.
process data. Scandinavian Journal of Manage-
Hinings, C. R., & Greenwood, R. 1988. The dynamics of ment, 13: 419-438.
strategic change. Oxford, England: Blackwell. Orton, J. D., & Weick, K. E. 1990. Loosely coupled sy
Hodgson, R. C., Levinson, D. J., & Zaleznik, A. 1965. The tems: A reconceptualization. Academy of Manage
executive role constellation. Boston: Harvard Busi- ment Review, 15: 203-223.
ness School Press.
Pettigrew, A. 1992a. On studying managerial elites. Stra-
Hosking, E. 1988. Organizing, leadership and skillful tegic Management Journal, 13(special issue): 163-
process. Journal of Management Studies, 25: 147- 182.
166.
Pettigrew, A. M. 1992b. The character and significance o
Johnson, R. L. 1979. Revisiting the "wobbly three legged strategy process research. Strategic Management
stool." Health Care Management Review, 4(3): 15. Journal, 13(special issue): 5-16.
Johnson, R. L. 1994. HCMR perspective: The purpose of Pettigrew, A. M., & Whipp, R. 1991. Managing chang
hospital governance. Health Care Management for competitive success. Oxford, England. Basil
Review, 19(2): 81-88. Blackwell.

Kingdon, J. 1984. Agendas, alternatives and public pol- Pfeffer, J. 1981. Power in organizations. Marshfield, MA
icies. Boston: Little, Brown. Pitman.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms
2001 Denis, Lamothe, and Langley 837

Pfeffer, J. 1992. Managing with power. Boston: Harvard Van de Ven, A. H. 1992. Suggestions for studying strategy
Business School Press. process: A research note. Strategic Management
Pfeffer, J., & Salancik, G. 1978. The external control of Journal, 13(special issue): 169-188.
organizations. New York: Harper & Row. Van de Ven, A. H., & Huber, G. P. 1990. Longitudinal
Poole, M. S., & Van de Ven, A. H. 1989. Using paradox to field research methods for studying processes of or-
build management and organization theories. Acad- ganizational change. Organization Science, 1: 213-
219.
emy of Management Review, 14: 562-580.
Powell, W. W., & DiMaggio, P. J. 1991. The new institu-Van de Ven, A. H., & Poole, M. S. 1995. Explaining
tionalism in organizational analysis. Chicago: Uni- development and change in organizations. Academy
versity of Chicago Press. of Management Review, 20: 510-540.
Weick, K. E. 1976. Educational organizations as loosely
Prenevost, M. 1994. L'urgence au secours de l'hopital
[Emergency: Save the hospital]. Unpublished M.Sc. coupled systems. Administrative Science Quar-
thesis, Universite de Montreal. terly, 21: 1-19.

Priem, R. L. 1990. Top management group factors, con-Weick, K. E. 1999. That's moving: Theories that matter.
sensus, and firm performance. Strategic Manage- Journal of Management Inquiry, 8: 134-142.
ment Journal, 11: 469-478. West, C. T., & Schwenk, C. R. 1996. Top management
Ranson, S., Hinings, C. R., & Greenwood, R. 1980. The team strategic consensus, demographic homogeneity
structuring of organizational structures. Administra- and firm performance: A report of resounding non-
tive Science Quarterly, 25: 1-17. findings. Strategic Management Journal, 17: 571-
576.
Reed, M. I. 1996. Expert power and control in late mo-
dernity: An empirical review and theoretical synthe-Yin, R. K. 1994. Case study research. Newbury Park, CA:
sis. Organization Studies, 17: 573-598. Sage.

Satow, R. L. 1975. Value-rational authority and profes-

sional organizations: Weber's missing type. Admin- Jean-Louis Denis is a professor of health administration
at Universite de Montreal. He holds a chair in the trans-
istrative Science Quarterly, 20: 526-531.
formation and governance of health care organizations
Scott, W. R. 1982. Managing professional work: Threefrom the Canadian Health Services Research Foundation
models of control for health organizations. Health
and the Canadian Institutes for Health Research and is a
Services Research, 17: 213-240.
member of the interdisciplinary research group on health
Shortell, S. M., Gillies, R. R., Anderson, D. A., Erickson, at Universite de Montreal. He received his Ph.D. in com-
K. M. & Mitchell, J. B. 1996. Remaking health care munity health from Universite de Montreal. His current
in America. San Francisco: Jossey-Bass. research deals with leadership and change in health care
Stacey, R. D. 1995. The science of complexity: An alter- organizations and with the role of evidence in the adop-
native perspective for strategic change processes. tion of clinical and organizational innovations in the
health care sector.
Strategic Management Journal, 16: 477-495.
Stewart, R. 1991. Chairmen and chief executives: An Lise Lamothe is an associate professor of management at
exploration of their relationship. Journal of Man- Universite de Montreal and was previously assistant pro-
agement Studies, 28: 511-527. fessor of management at Universite Laval in Quebec City.
She received her Ph.D. in management from McGill Uni-
Stinchcombe, A. 1965. Social structure and organiza-
versity in Montreal. Her research focuses on the transfor-
tions. In J. G. March (Ed.), Handbook of organiza-
mation of health care organizations, the formation of
tions: 142-193. Chicago: Rand McNally.
integrated delivery systems, and the dynamics of profes-
Strauss, A., Schatzman, L., Ehrlich, D., Bucher, R., & sional organizations.
Sashin, M. 1963. The hospital and its negotiated
order. In E. Friedson (Ed.), The hospital in modern Ann Langley is a professor of strategy at lcole des Hautes
society: 147-169. London: Collier-MacMillan. ltudes Commerciales (HEC) in Montreal. She was on the
faculty at Universite du Quebec a Montreal until Novem-
Thompson, J. D. 1967. Organizations in action. New ber 2000. She received her Ph.D. from icole des Hautes
York: McGraw-Hill.
Etudes Commerciales. Her research interests include
Van de Ven, A. 1999. Theme: change and development strategic change processes, strategic leadership decisio
journeys into a pluralistic world. Call for papers, making, innovation, and interorganizational relati
annual meeting of the Academy of Management, ships, with particular emphasis on health care organiz
Chicago. tions.

This content downloaded from

183.219.253.71 on Wed, 10 Feb 2021 07:14:16 UTC
All use subject to https://about.jstor.org/terms