Crosssectorcollaborations.pdf

75th Anniversary
Article

John M. Bryson is McKnight

Presidential Professor in the Hubert H.

Humphrey School of Public Affairs at the

University of Minnesota. He works in the

areas of leadership, strategic management,

collaboration, and the design of engage-

ment processes. He wrote Strategic

Planning for Public and Nonprofi t

Organizations, 4th edition (2011),

and cowrote (with Barbara C. Crosby)

Leadership for the Common Good,

2nd edition (2005). He is a fellow of the

National Academy of Public Administration.

E-mail: [email protected]

Barbara C. Crosby is associate

professor in the Humphrey School of

Public Affairs, University of Minnesota, and

part of the school’s Public and Nonprofi t

Leadership Center. She focuses on

leadership and public policy and women in

leadership. She is author of Leadership

for Global Citizenship (1999) and

cowrote (with John M. Bryson) Leadership

for the Common Good, 2nd edition

(2005). She has conducted training for

senior managers of nonprofi t, business, and

government organizations in the United

States and Europe.

E-mail: [email protected]

Melissa Middleton Stone is Gross

Family Professor of Nonprofi t Management

and Distinguished University Teaching

Faculty in the Humphrey School of Public

Affairs, University of Minnesota. Her

teaching and research focus on public and

nonprofi t management and governance as

well as the design of cross-sector collabora-

tions. She has published widely in scholarly

journals and books in the fi elds of nonprofi t

studies, public management, and strategic

management.

E-mail: [email protected]

Designing and Implementing Cross-Sector Collaborations: Needed and Challenging 647

Public Administration Review,

Vol. 75, Iss. 5, pp. 647–663. © 2015 by

The American Society for Public Administration.

DOI: 10.1111/puar.12432.

John M. Bryson
Barbara C. Crosby

Melissa Middleton Stone
University of Minnesota

Editor’s Note: Cross-sector collaborations are a noteworthy addition to the tools of public administration. In
this anniversary article, John Bryson, Barbara Crosby and Melissa Middleton Stone visit themes from their
2006 article, which strongly infl uenced this line of research and practice. Th eir review of the last decade’s
research leads them to revise and add new propositions anchored in propositions they introduced in 2006.

JLP

Abstract: Th eoretical and empirical work on collaboration has proliferated in the last decade. Th e authors’ 2006
article on designing and implementing cross-sector collaborations was a part of, and helped stimulate, this growth. Th is
article reviews the authors’ and others’ important theoretical frameworks from the last decade, along with key empirical
results. Research indicates how complicated and challenging collaboration can be, even though it may be needed now
more than ever. Th e article concludes with a summary of areas in which scholarship off ers reasonably settled conclu-
sions and an extensive list of recommendations for future research. Th e authors favor research that takes a dynamic,
multilevel systems view and makes use of both quantitative and qualitative methods, especially using longitudinal
comparative case studies.

Practitioner Points
• Make sure there is a clear collaborative advantage to be gained by collaborating, meaning that collaborators

can gain something signifi cant together that they could not achieve alone. Make use of windows of opportu-
nity to advance the collaboration approach.

• View collaborations as complex, dynamic, multilevel systems.
• Collaborating parties should take a design approach to cross-sector collaboration. Th is means starting as

much as possible with the ends in mind and designing processes, structures, and their interactions in such a
way that desired outcomes will be achieved and required accountabilities met. Build ongoing learning into
the design, including learning about what goals and performance indicators should be.

• Make sure that committed sponsors, champions, and facilitators are involved throughout.
• Use inclusive processes to develop inclusive structures, which, in turn, will sustain inclusive processes.
• Adopt fl exible governance structures that can adjust to diff erent requirements across the life cycle of the

collaboration.

drawn on the design propositions presented in our
original article (e.g., Lai 2012; Simo and Bies 2007),
and that research has led us to alter some propositions
and add new ones (Bryson, Crosby, and Stone 2014;
see the appendix at the end of this article).

Our abiding aim is to help public managers and
integrative leaders from any sector design and sustain
eff ective cross-sector collaborations when such eff orts
are likely to produce public value. We are therefore
pleased to have been asked to assess the theoretical
and empirical work on cross-sector collaboration
over the last decade and the infl uence that our article
might have had. Much has happened, but some

Designing and Implementing Cross-Sector Collaborations:
Needed and Challenging

Our 2006 Public Administration Review article “Th e Design and Implementation of Cross-Sector Collaborations” appeared when
collaboration theory, research, and practice were accel-
erating. Cross-sector collaboration has now become a
staple of public management research as governments
are called on to partner with organizations across sec-
tors and civil society in order to address public prob-
lems that they cannot successfully address alone (Kettl
2015). In the last decade, researchers have continued
to highlight cases of successful cross-sector collabora-
tion, but many report cases of failure and uneven
results (e.g., Andrews and Entwistle 2010; Hodge and
Greve 2007). Some research, including our own, has

648 Public Administration Review • September | October 2015

or more sectors to achieve jointly an outcome that could not be achieved by
organizations in one sector separately (Bryson, Crosby, and Stone 2006,
44). We assert that collaboration occurs in the midrange of a con-
tinuum of how organizations work on public problems (Crosby and
Bryson 2005). At one end are organizations that hardly relate to each
other, and at the other are organizations merged into a new entity. In
the middle, toward the formal end of interorganizational confi gura-
tions, are collaborative entities that fi t the foregoing defi nition.1

Collaborations are given a variety of labels (e.g., consortium,
alliance). We use the term “collaborations” interchangeably with
“partnerships” because of the prevalence of the latter term in the
literature. However, we make an important distinction between col-
laborations and public–private partnerships. We see public–private
partnerships as a particular type of cross-sector collaboration based
on formal, contractual relationships between two or more entities
(Minnesota Department of Transportation 2011). We cite studies of
public–private partnerships in this review when fi ndings are relevant
to understanding cross-sector partnerships more generally (rather
than how to structure and administer contracts).

Frameworks from Earlier in the Decade
Building on prior work (e.g., Gray 1989; Huxham and Vangen
2005; Ostrom 1990; Ring and Van de Ven 1994), several holis-
tic frameworks were published about the same time as our own.
Th ese include Th omson and Perry (2006), Ansell and Gash (2008),
Agranoff (2007, added to in 2012), and Provan and Kenis (2008).

Th ese frameworks have clear similarities. All attend to important
general external antecedent conditions, more proximate initial
conditions, internal processes, structural elements, and outcomes.
Antecedent conditions include the availability of varied resources,

things have not changed. Recent research only reinforces our earlier
conclusion that cross-sector collaboration is hardly an easy answer
to complex public problems. Indeed, it is typically frustrating for
participants (although sometimes exhilarating as well) and full of
opportunities for what Huxham and Vangen (2005) call “collabora-
tive inertia.” Yet those who seek to combat problems such as poverty
and urban traffi c congestion often have no choice but to work across
sector lines to develop shared understandings of the problem and
commitments to shared solutions.

In this article, we review the most signifi cant theoretical frameworks
developed in the last decade for exploring cross-sector collaboration.
We then review some of the most important empirical studies since
2006. Finally, we off er conclusions and suggest next steps for this
important research and practice fi eld.

Signifi cant Frameworks for Understanding Cross-Sector
Collaboration
Several important holistic frameworks for understanding cross-sec-
tor collaboration have been published in the last decade. Th is review
focuses on those developed by Bryson, Crosby, and Stone (2006);
Th omson and Perry (2006); Ansell and Gash (2008); Agranoff
(2007, 2012); Provan and Kenis (2008); Emerson, Nabatchi, and
Balogh (2011); and Koschmann, Kuhn, and Pfarrer (2012). Table 1
provides a summary of each theoretical framework.

Th e defi nitions of collaboration used by the frameworks’ authors
vary, but they are generally similar to our own in that they stress a
continuum of progressively more intense interorganizational relation-
ships. We defi ne cross-sector collaboration as the linking or sharing of
information, resources, activities, and capabilities by organizations in two

Table 1 Major Cross-Sector Collaboration–Related Theoretical Frameworks

Publication Bryson, Crosby, and Stone (2006) Thomson and Perry (2006) Ansell and Gash (2008) Agranoff (2007, 2012)

Theory base Diverse, including organization
theory, public administration
theory, leadership theory, strategic
management theory

Diverse, including organization theory,
public administration theory, strategic
management theory

Diverse, including organization
theory, public administra-
tion theory, policy studies,
planning and environmental
management studies

Diverse, including
organization theory,
public administra-
tion theory, strategic
management theory

Major components Initial conditions
Formal and informal processes
• Agreements
• Leadership
• Legitimacy
• Trust
• Confl ict management
• Planning

Formal and informal structures
• Membership
• Structural confi gurations
• Governance structures

Contingencies and constraints
• Type of collaboration
• Power imbalances
• Competing institutional logics
Outcomes and accountabilities

Antecedents
Processes
• Governance
• Administration
• Organizational autonomy
• Mutuality
• Norms of trust and reciprocity
Outcomes

Starting conditions
Collaborative process
• Face-to-face dialogue
• Trust building
• Commitment to process
• Shared understanding
• Intermediate outcomes
• Facilitative leadership
Outcomes

Plus core contingencies: time,
trust, and interdependence

Decision networks
versus nondecision
networks

Processes
• Activation
• Framing
• Mobilizing
• Synthesizing

Particular emphases
vis-à-vis the others

Cross-sector collaboration, institu-
tional logics, planning, contingen-
cies, power and the importance
of remedying power imbalances,
the need for alignment across
components

Learning, organizational autonomy, leader-
ship, administration

Face-to-face dialogue, incen-
tives and disincentives, the
importance of remedying
power imbalances

Leadership through
a whole range of
roles, processes, and
structures, public
value, capacity build-
ing, and learning

Designing and Implementing Cross-Sector Collaborations: Needed and Challenging 649

Table 1 Major Cross-Sector Collaboration–Related Theoretical Frameworks, Continued

Provan and Kenis (2008) Emerson, Nabatchi, and Balogh (2011) Koschmann, Kuhn, and Pfarrer (2012)

Theory base Network theory Diverse, including organization theory, public adminis-
tration theory, confl ict management theory, planning
and environmental management studies

Communication theory

Major components Ideal types of network governance:
• Participant governed
• Lead organization
• Network administration organization

Critical contingencies:
• Degree of trust, number of members,

goal consensus, need for network-
level competencies (nature of the
task and kinds of external demands)

Persistent tensions:
• Effi ciency versus inclusion
• Internal versus external legitimacy
• Flexibility versus stability

Evolution of these governance systems
over time

System context
• Drivers
Collaborative governance regime
Collaboration dynamics
• Principled engagement
• Capacity for joint action
• Shared motivation
Actions
Impacts
Adaptation

Communication practices
• Increasing meaningful communication
• Managing centripetal and centrifugal

forces
• Creating a distinct and stable identify
Development of authoritative texts

Trajectory of authoritative texts

Communication practices to assess over-
all cross-sector partnership value

• External intertextual infl uence
• Accounts of capital transformation

Particular emphases
vis-à-vis the others

Governance structures Collaborative regimes, what makes collaborations work,
capacity building

Pulling out collaborative actions from overall impact/
outcomes

Authoritative texts and their effects on
activities and partners

characteristics of the institutional environment, and the need to
address complex public issues. Initial conditions encompass the sig-
nifi cance of preexisting histories and relationships (whether positive
or negative), some agreement on collaborative aims and perceived
interdependence among members, and the availability of leadership.
Processes emphasize fostering trust, organizing inclusive participa-
tion, developing a shared understanding of the problem, building
commitment to collective goals and actions, and formal advance
planning or emergent planning. Structures highlighted by all the
frameworks include norms and rules that emerge to promote these
processes as well as to reach agreement on collaborative goals and
actions. Governance at the intersection of processes and structures is
a feature of each framework. Th ese early works also emphasize several
diff erent types of endemic tensions or confl icts facing collabora-
tions. For example, Th omson and Perry (2006) emphasize ongoing
tensions between the need for partner autonomy versus the inter-
dependence of collective interests, while Kenis and Provan (2008)
focus on tensions between stability versus fl exibility, inclusivity versus
effi ciency, and internal versus external legitimacy. Several frameworks
highlight power imbalances (Agranoff 2007; Bryson, Crosby, and
Stone 2006), and ours raises the issue of confl ict as a result of multi-
ple institutional logics.

Th ese frameworks also diff er, especially
regarding attention to specifi c aspects of
collaboration. For example, some emphasize
the importance of particular leadership roles
(Bryson, Crosby, and Stone 2006; Th omson
and Perry 2006), while others focus on leader-
ship activities (Ansell and Gash 2008) or the
structure of a “leadership core” (Agranoff
2007, 2012). Another diff erence pertains
to the extent to which collaborative struc-
ture is considered as a separate component (Bryson, Crosby and
Stone 2006; Kenis and Provan 2008) or is subsumed under process

dynamics (Agranoff 2007; Emerson, Nabatchi, and Balogh 2012;
Th omson and Perry 2006). Frameworks also diff er with regard to
whether they focus on collaborations involving multiple sectors
or primarily those led by public agencies. Our own framework
highlights the role of sector failure as a prompt to cross-sector col-
laboration as a way of making up for the shortcomings of particular
sectors.

Noticeably, most of these early frameworks do not pay special atten-
tion to outcomes and accountabilities, but they do off er important
insights. For example, most argue that evaluative assessments should
consider substantive as well as process outcomes for individual par-
ticipants, member organizations, the collaboration as a whole, and
the community. Agranoff adds that collaborations should be judged
on whether they produce public value from the standpoint of vari-
ous stakeholders, while Th omson and Perry and we argue that the
issue of accountabilities is especially problematic in collaborations.

More Recent Frameworks
Emerson, Nabatchi, and Balogh’s (2011) “collaborative governance
regime” framework extends the view of collaboration as a system
embedded in, and interacting with, a larger environment. Th ey are

focused principally on public organizations
as collaborators but note the likely involve-
ment of actors from other sectors as well.
Determinants of the collaborative governance
regime are rooted in the external context,
including resource conditions, policy and
legal frameworks, and politics and power
conditions. A set of drivers separate from
system context include leadership, conse-
quential incentives, recognized interdepend-
ence, and uncertainty. Th e authors highlight

three important internal collaboration dynamics: principled engage-
ment, leading toward shared motivation, and ending with capacity for

Determinants of the collabo-
rative governance regime are

rooted in the external context,
including resource condi-

tions, policy and legal frame-
works, and politics and power

conditions.

650 Public Administration Review • September | October 2015

complicated dynamic systems. Within these systems, a better
understanding is needed of, for example, the interactions between
managerial actions, processes and structures over time; the eff ects
of various internal and external contingencies; how collaborations
are aff ected by being embedded in, or directly aff ected by, existing
hierarchies; and what helps create eff ective performance. Finally,
these articles advocate for a variety of research designs and the use of
both quantitative and qualitative methods. Most call for longitudi-
nal research in order to capture the dynamic contexts and responses
of collaborations over time.

In a review of the collaboration fi eld, O’Leary and Vij argue that
“the study and practice of collaborative public management is
generally fragmented with a low level of consensus; from a research
perspective, it is a low-paradigm fi eld” (2012, 518). Th ey make
several suggestions for improvement, including the need for “more
precise theoretical models of behavior, and agreement on the meas-
urement of relevant variables” (507). We concur but note that such
a state of aff airs is not surprising at this stage of the fi eld’s develop-
ment, and beyond that, in the last 10 years, considerable theoretical
and empirical advances have produced greater understanding of
collaboration. Collectively, the frameworks reviewed here provide a
rich picture of the importance of seeing collaboration as embedded
in larger systems; what is involved in collaboration and cross-sector
collaboration; its inherently interdisciplinary nature; its systemic,
multilevel, multiactor nature; and the array of signifi cant constitut-
ing elements. Th ese frameworks have provided important guidance
for empirical research, to which we now turn.

Signifi cant Empirical Work in the Last Decade
Th is section highlights areas in which empirical work is fi lling in the
foregoing theoretical frameworks. Occasionally, we will cite earlier
work when it is important for the argument.

We developed the list of possible works using the time parameter
of 2007 to early 2015 and certain keywords, such as “cross-sector
collaboration” and “partnership,” and keyword combinations,
including “collaboration” and “accountability,” “outcomes,” “power,”
“antecedents,” and so forth. In all, we reviewed 196 articles and
three books. We removed the articles that were either nonempiri-
cal or added little to existing theory or empirical work. (We use
“empirical” broadly to mean qualitative or quantitative studies,
meta-analyses, or literature reviews of empirical work.) Th e team
further refi ned the list using fi ve criteria: the study (1) was situ-
ated within a larger systems view of collaboration; (2) was focused
on cross-sector collaboration; (3) illuminated a key aspect of the
frameworks discussed earlier; (4) had an empirical grounding; and
(5) specifi ed infl uence relationships, or contextual constraints, such
as rules, resources, or settings that alter the range of possible eff ects.
Th e team divided up responsibility for reviewing articles and books
and then discussed collectively specifi c fi ndings and overarching
themes. In spite of the limitations of our search procedures, we
are reasonably confi dent that the included references provide an
important overview of empirical fi ndings over the last decade and
additional needed work. We describe the studies according to the
major categories encompassed in the frameworks just described:
general antecedent conditions; initial conditions, drivers, and link-
ing mechanisms; processes, structures, and links between them;

joint action, which includes important elements of structure. Th ese
collaboration dynamics create action, which may further infl u-
ence collaboration dynamics as well as aff ect the external system or
context. Th e possible broad impacts are similar to those in our own
framework. Th e framework is quite comprehensive, although it may
excessively privilege process over structure. Th e framework improves
on previous frameworks in its move toward articulating causal con-
nections and calling for greater attention to exploring which causal
relationships matter in which contexts.

Th e fi nal framework we review is that of Koschmann, Kuhn, and
Pfarrer (2012), which theorizes communication as constitutive of
organizations and collaborations. Th ey emphasize the importance
of formulating “authoritative texts” that can include implicit norms
of cooperation or bylaws, mission statements, and memoranda of
understanding. Th ese texts indicate the collaboration’s “general
direction and what it is ‘on track’ to accomplish” (337). Ideally,
these texts will marshal the willing consent of partners, attract other
resources, and help collaborators exercise collective agency. Th e
authors argue that communication practices increase the potential
impact of cross-sector collaborations by shaping public perception
as well as member perspectives on issues and providing accounts
of intellectual, social, and organizational capital transformation.
Th is framework off ers two noteworthy contributions: the addition
of communication theory, especially the idea of communication as
constitutive of collaborations, and attention to authoritative texts.

Summary
All frameworks imply some causality among particular components
but eschew simple causal connections, instead focusing on important
contextual contingencies. For example, a common theme across these
frameworks is their attention to the infl uence of antecedent condi-
tions on collaboration eff ectiveness. Ansell and Gash are especially
explicit about particular aspects of starting conditions, including the
need for a collaboration to rectify signifi cant power or resource asym-
metries, acknowledge interdependence among members, and mitigate
negative prehistories. Likewise, Emerson, Nabatchi, and Balogh begin
their causal logic model with specifi c elements of the external system
in which the collaboration is embedded. Provan and Kenis also pro-
pose more explicit causal connections where appropriate governance
structures are determined by the number of collaboration members,
degrees of goal consensus and trust, and need for network-level
competencies. We emphasize contingencies related to the infl uence
of collaboration type, power imbalances, and competing institutional
logics, while Agranoff adds the eff ects of collaborative capacity.

Th e frameworks leave out some important considerations. None of
the frameworks delves very deeply into the eff ects of the broader
technical and institutional environments on collaboration, including
the eff ects of adjacent and often competing social fi elds (Fligstein
and McAdam 2012; Scott and Davis 2006); how collaboration
might diff er depending on the nature of the issue or task to be dealt
with; the eff ects of technology broadly construed; the need for struc-
tural and processual ambidexterity (O’Reilly and Tushman 2013)
across the life cycle of a collaboration; and the array of attitudes,
competencies, and capacities needed for eff ective collaboration.

In calls for future research, a number of themes are apparent,
including, most obviously, the need to view collaborations as

Designing and Implementing Cross-Sector Collaborations: Needed and Challenging 651

General Antecedent Conditions
Recent research confi rms that the institutional environment is espe-
cially important for partnerships focused on public policy or public
problem solving because it includes broad systems of relationships
across jurisdictional areas (Sandfort and Moulton 2015; Scott and

endemic tensions or points of confl ict; and outcomes and account-
abilities. Figure 1 presents a reasonably parsimonious and pragmatic
synthesis of the theoretical frameworks and empirical studies we
reviewed. (Th e fi gure represents a signifi cant adaptation of a similar
fi gure in our 2006 article.)

Figure 1 Summary of Major Theoretical Frameworks and Findings from Empirical Studies, 2006–15. Bolded elements are from
both the theoretical frameworks and recent empirical studies; elements in italics are new elements from empirical studies

GENERAL ANTECEDENT CONDITIONS
Resources
Institutional environment

Mandates
Window of collaborative opportunity
Vulnerable to policy/political change

Need to address public issue
Sector failure
Resources from nongovernmental
partners to solve issue

INITIAL CONDITIONS, DRIVERS, AND
LINKING MECHANISMS

Agreement on initial aims
Authoritative texts

Recognized interdependence
Preexisting relationships

Initial leadership
Specific leader characteristics

Consequential incentives

Nature of task

COLLABORATIVE PROCESSES
Trust and commitment
Shared understanding of
problem
Communication mechanisms
Legitimacy

Internal legitimacy
Formal and emergent planning

COLLABORATION
STRUCTURES

Development of norms and
rules or practices of
engagement
Dynamic and
particularistic structures

Structural ambidexterity

ENDEMIC CONFLICTS AND TENSIONS
Power imbalances
Vulnerable to exogenous and endogenous shocks
Multiple institutional logics
Tensions

Flexibility versus stability
Inclusivity versus efficiency
Unity versus diversity
Autonomy versus interdependence (or self-interest
versus collective interest)

ACCOUNTABILITIES AND OUTCOMES
Complex accountabilities

Formal and informal
Tangible and intangible outcomes

Immediate, intermediate, and long
term
Multiple levels of outcomes
Resilience and reassessment learning
Public value creation

LEADERSHIP

GOVERNANCE
Emergent, dynamic

Contingent

TECHNOLOGY

CAPACITY AND
COMPETENCIES

652 Public Administration Review • September | October 2015

Other important drivers or initial conditions have also been
identifi ed in recent studies. First, in line with several theoretical
frameworks, there is evidence about the importance of initial, albeit
general, agreement on the problem defi nition that also indicates
the interdependence of stakeholder organizations when it comes
to addressing the problem (Simo and Bies 2007). While informal
agreements about the collaboration’s composition, mission, and
process can work, formal agreements have the advantage of sup-
porting accountability. Th e absence of formal agreements can
make collaboration more diffi cult, especially when accompanied
by a lack of administrative capacity (Babiak and Th ibault 2009).
Furthermore, Koschmann, Kuhn, and Pfarrer (2012) emphasize
that naming the collaboration, weaving a compelling story of the
collaboration’s work, and developing an “authoritative text” can help
marshal the willing consent of partners and foster collective agency.
Th is text may include implicit norms of operation or by-laws, mis-
sion statements, and memoranda of understandings; it indicates
the collaboration’s “general direction and what it is ‘on track’ to
accomplish” (Koschmann, Kuhn, and Pfarrer 2012, 337). (“Text”
in this case clearly overlaps with structure and governance, discussed
later.) Additionally, the need for diff erent types of initial agreements
and the reworking of agreements are likely to increase as collabora-
tions grow to include more geographically dispersed partners and
diverse actors within a problem domain (Clarke and Fuller 2010;
Sandström, Bodin, and Crona 2015; Vangen and Huxham 2012).

Second, the role of prior relationships or …