A Model of Operational Interorganizational Network Formation
September 30, 2005
Kimberley R. Isett, Ph.D.
600 West 168th St, 6th floor
Department of Health Policy and Management
Columbia University
New York, New York 10032
(212) 342-3905
Prepared for presentation at the National Public Management Research Conference, Los Angeles, California, September 2005
(very rough) Draft- Do not Cite
For Discussion Only
A Model of Interorganizational Network Formation
Abstract
Interorganizational relationships (IORs) in the production of public services are a fact of life. IORs stem from a number of factors that are well documented in the organizational theory and public administration literatures. These factors include, but are lot limited to: environmental turbulence, political advocacy, resource scarcity, joint production functions, broad organizational missions, power and relationship asymmetry, and institutional coercion (DiMaggio & Powell 1983; Galaskiewicz 1985; Oliver 1990; Milward & Provan 2000). However, in some cases interorganizational relationships are something more than just a series of relationships between two organizations. Instead, they become a network.
Networks are created by dyadic relationships, without a doubt. But networks are a governance mechanism (formal and/or informal) among those organizations that belong to it. They have shared goals and processes that extend beyond two organizations, to a multitude of organizations that create a unified response to a given phenomenon (Chisholm 1999; Alter & Hage 1992). Interestingly, networks may be formed in several ways: from the bottom-up as a collective response by individual organizational actors; from the top- down through institutional mandate or coercion; and through some combination of the two approaches.
While we know lots about interorganizational relationships, and a good deal about interorganizational networks, one critical element of networks has thus far been overlooked by network scholars is network formation. This paper seeks to address this gap in the literature by developing a model of interorganizational network formation. This model has been developed from the extant literature on interorganizational relationships and networks, drawing heavily from the organizational theory and management literature.
While the organizational theory and management theory forms a base for the model constructed, insights from other disciplines are capitalized upon. Literature from political science is particularly emphasized, especially the Institutional Analysis and Development (IAD) framework (Ostrom 1990; Ostrom Gardner & Walker 1994). The combination of empirical and theoretical knowledge emanating from the three disciplines (public management, organization theory, and political science) creates an opportunity to develop a comprehensive model of interorganizational network formation without emphasizing either top-down or bottom-up approaches –a common shortcoming in the existing literature.
A Model of Interorganizational Network Formation
Interorganizational networks have become a prominent method of organizing social interaction in the past two decades. As such, this phenomenon has gained increasing interest among organization scholars across disciplines. The existing literature on networks mainly focuses on how networks work, the structural aspects of networks, and the organizational and social outcomes of networks. However, scholars have yet to focus on the early stages of network formation.
One potential reason for the scant attention to network formation is the close alignment of the interorganizational relationships literature to the interorganizational networks literature. Indeed, much of the research conducted on networks is a natural outgrowth of interest in and research on dyadic or organization set relationships. However, the phenomenon of interorganizational network formation is different than simple interorganizational relationships and deserves some concerted attention.
The literature on Interorganizational relationships (IORs) has been well developed. This literature focuses on why two organizations would decide to form a relationship. Seminal works on explaining why IORs form have included both management (c.f. Oliver 1990) and sociological perspectives (c.f. Galaskiewicz 1985). These works focus on the linkages between two organizations or the linkages from one organization to another. While this perspective is a valuable starting point for a network perspective, it needs to be built upon to capture the differences of networks from dyadic relationships.
The primary difference between interorganizational relationships and interorganizational networks is the need for collective action. Networks require that many organizations come together to pursue a common goal in a coordinated and concerted way. This requires more than a simple calculus of benefit for joint production or exchange. It requires a trade-off of short term and individual pay-off, for longer-term collective benefits. Just how does an organization come to a decision to participate in a network? What are the factors that contribute to network formation? This paper presents a model of interorganizational network formation to begin to address these important questions.
The model presented in this paper builds on ideas of collective action and synthesizes them with the existing IOR and networks literature to provide one conceptualization of interorganizational network formation. Through this model, I seek to elucidate the similarities and differences between interorganizational relationships and networks, as well as provide insight into the process of network formation.
Conceptual Background
Networks are self-governing institutional arrangements. These arrangements may be explicitly designed, as in many governmental service delivery systems (see Provan and Milward 1995 and Isett and Provan 2005), or may be more organically developed, as in many environmental management focused networks (see Ostrom, Gardner, and Walker 1998). Regardless of the way in which they were developed, all networks have a common archetype.
All networks are flat governance mechanisms in which actors come together to devise rules of interaction focused on joint goal attainment. The goal attainment may be productive, social, or conservation oriented, but requires collective action to achieve those goals. Thus, the actors in the network come together to share decision making efforts and authority in regard to the network goal (Alter and Hage 1993; Powell 1990; Scott 1981).
One perhaps obvious statement about networks that is worthwhile to state is that network members come together to both give and take organizational resources to/from the network. Incentive to participate in networks comes mainly from an organization having the opportunity to achieve a goal that it otherwise would not be able to achieve on its own, similar to why organizations-proper form. However, an organization must give the other network members a reason for including them in the network as well as a reason for sharing their organizational resources through relationships. Thus, organizations must have something to give as well as to take away from networks. While this exchange may not be symmetrical, there is still some benefit to the exchange for the parties involved.
Coordination of activities is an expensive and time consuming (Dill and Rochefort 1989). Thus, it is not worthwhile for organizations to participate in network activities if there is little to gain, or if the organization can achieve its purposes without the network. Although organizations will subsume network goals and agendas to their individual goals and agendas, the network still aids the organization to accomplish something on its agenda. Thus, the benefit of network activities must come into balance with individual organizational activities and provide some real benefit (tangible or intangible) for the organization. Further, to reinforce the point above, if an organization offers no reason or benefits to other organization to interact with them in a network, then it is difficult to discern why an alter organization would spend the resources needed to coordinate and establish a relationship with the ego organization when their resource would be better used developing other relationships.
Further, these exchanges in networks are not caused by serial interdependencies (Thompson 1967). Serial interdependencies can easily be solved through traditional market transactions. Instead, networks are more of a function of complex integrated interdependencies. Integrated interdependencies means that actors must put their resources into the production of an outcome at the same time and have those contributions interact with other organizations’ contributions; without this interaction and synchronous contributions the outcome or product of the network will not be achievable.
While networks serve network members by guiding action and activities (Isett and Provan 2005), they also serve members in other ways. Network members capture benefits of remaining formally independent and small, such as speed of innovation, while reaping economies of scope and scale as well through the structure of other organizations (Powell, Koput, and Smith-Doerr 1996). Thus networks offer its members flexibility in terms of the resources of time, transition, and cost (Sabel 1989).
Despite this knowledge of how networks work and the benefits members receive by belonging to them, the knowledge on how and why networks form is still scant. The early literature on interorganizational relationships suggests that organizations form relationships in order to compensate for resource dependencies and uncertainty. More recent literature on interorganizational relationships has developed reasons such as legitimacy, information needs, and stability.
James Thompson (1967) suggested that organizations will buffer their core technologies from environmental turbulence. This means building structures at the governance level of the organizations that focuses on boundary spanning and acquiring needed organizational resources from that environment. Buffering is particularly important in turbulent environments.
Turbulent environments present uncertainty for organizations in that conditions (political, environmental, market, and/or social) are changing and organizations have to face contingencies about how and where they will capture organizational resources, a type of state uncertainty (Milliken1987). Turbulence in the environment leads organizations to seek interorganizational relationships in order to smooth out environmental perturbations and to have more certainty with regard to resource acquisition (Oliver 1990; Galaskiewicz 1985). More turbulence leads to greater need for interorganizational relationships, and perhaps more of them.
Not only do interorganizational relationships help organizations even out their resource acquisition, but it also helps them in their planning and strategic management. Organizations seek out stability from their environment and their operations. While this is a similar point to the idea of buffering core technologies and dealing with uncertainties, it merits further explication.
Stability of relationships and within the environment helps organizations to better adjust to market conditions and to other contingencies it encounters. We see this concept most clearly from the contracting literature in both management and public administration. When organizations contract with other organizations, they must adjust their operations to accommodate the style and requests of their partners. This is especially poignant if the partner is a major source of the focal organization’s business or service recipients. Thus, if a contractor changes, the organization must go through a new adjustment period which may disrupt operations and takes time to learn and anticipate partner needs (Uzzi, 1997; Larson 1992).
Therefore, contracting can become a stabilizing or disruptive part of an organization’s operations, based on frequency of principal changes. So a principal in a contracting partnership must balance the need for accountability and flexibility through frequent contracting, with the partner/agent’s need for adjustment and operational process norming through less frequent contracting (Milward and Provan 1993).
Stability concerns were found to be an important variable in several empirical network studies in the past decade. Provan and Milward (1995) found that mental health networks had improved effectiveness if there were stability related to the Network Administrator and their connections with the state. Likewise, Isett and Provan (2005) illustrated that networks may also have stable network structures, which helps to smooth out operations by having a stable roster of organizations with which an organization interacts, consistent with the buffering core technologies and pre-empting environmental uncertainty concepts discussed above.
The vulnerabilities created by turbulence in the environment are enacted. Just as organizations enact their environments, choosing which elements to pay heed to and to address, they enact their vulnerabilities as well. Organizations will choose to perceive vulnerabilities that they feel are important. More importantly, however, is that the vulnerabilities themselves are perceived (Van de Ven and Walker 1984). One such vulnerability that has been increasingly cited in the new institutional literature is legitimacy.
Vulnerabilities, or perceptions of vulnerabilities, are an important reason for organizations to come together. While most vulnerabilities and uncertainties have some tangible quality to them, the conferring of organizational legitimacy is one organizational resource in which the focal organization has little control of acquisition.
Organizational legitimacy is the idea that other actors in the organizations environment accept the organizations actions, processes and outcomes as appropriate (Suchman 1995; Zucker 1991). Legitimacy is socially conferred from outside the organization and can not be manufactured from within (Zucker 1991). Thus, the organization is dependent upon others in their environment to bestow legitimacy upon them.
One way that some organizations approach legitimacy acquisition is through developing relationships with other organizations. This strategy is consistent with upward appeal activities of organizations where being associated with a prestigious alter organization produces a halo effect for the ego organization (Stinchcombe, 1965). Not only does this organizational behavior produce some legitimacy effects, it also helps the organization to overcome some of its inherent liabilities, such as newness and size (Stinchcombe 1965; Powell, Koput and Smith-Doerr 1996).
Legitimacy is a powerful attribution and can impact organizational outcomes. Tolbert and Zucker (1983) illustrate the pressures to adopt civil service reform among public agencies in order to achieve or retain legitimacy. Early adopters became organizational and field leaders, while late adopters experienced pressures to mimic the successes of those governments with the reform already implemented. However, while Tolbert and Zucker illustrated the power of adopting innovative reform, it is less obvious that the reform itself must attain a level of legitimacy itself before adoption is viewed as desirable.
Human and Provan (2000) demonstrate the power of legitimacy for innovations. In their study of two small manufacturing networks, they show how legitimacy works on three levels in network: network as interaction, network and entity, and network as form. If any of these three levels of legitimacy is not satisfied, then the network will not be effective, and will likely disband. Thus, organizational legitimacy, in this case network legitimacy, is directly related to organizational survival.
One way that legitimacy can be developed by an organization is through the building of reputation. Reputation is the idea that alters can predict reliably future or current behavior based on past behavior (Ostrom 1990). Alters can predict behavior based on a pattern that and ego has established in past or repeated interactions. Reputation attributions are stronger the longer established and the more consistent a pattern of behavior is. Thus, reliability is, more or less, the reliability that other have in predicting an ego’s behavior.