**1. Introduction**

There is a growing literature on the dynamics of cooperation among organizations (see Cropper and Palmer 2008), though Bell, den Ouden, and Ziggers (Bell et al. 2006) argue that the field is "*fragmented, lacks coherence, and has produced non-comparable research*". In their review of 22 longitudinal cases of inter-organizational collaborations, Majchrzak et al. (2015) start to draw together insights into dynamics, conceived as changes in the characteristics of inter-organizational collaborations and as patterns of relationships between sources and characteristics of inter-organizational collaborations. Most studies of dynamics focus on alliances between two firms, often taking the perspective of one side of the alliance (Bell et al. 2006). In this paper, by comparison, we focus on associations composed of multiple organizations, drawing on, and appraising, aspects of the theory of meta-organization (Ahrne and Brunsson 2005, 2008). Majchrzak et al. (2015) include only one study on the dynamics of meta-organizations, the study of Sematech by Browning et al. (1995), and no specific attention is given to the difference in dynamics. The (relatively limited) literature on organizational federations and associations does contain some studies of importance (e.g., Lowndes and Skelcher 1998; Selsky 1998; Traxler 2002). However, insights into the specific character of meta-organizations remain limited and fragmentary (Berkowitz and Bor 2018; Berkowitz and Dumez 2016).

This paper presents changes in the membership composition of a meta-organization over a 17-year period and discusses the significance of these changes for the character of the meta-organization. In the case study, we draw on the annual reports of Partners in Paediatrics (PiP), from its initiation in 1997 until 2014, to capture what we term the 'compositional dynamics' of this formalized association of organizations (Cropper 2001), which meta-organization theory suggests will have distinctive characteristics. We understand 'character' in institutional terms. As Selznick (1957) observed, "*This patterning [character] is historical, in that it reflects the specific experiences of the particular organization; it is functional in that it aids the organization to adapt itself to its internal and external social environment; and it is dynamic, in that it generates new and active forces, especially internal interest-groups* ... *.the emergence of organizational character reflects the irreversible element in experience and choice*". In this paper, we ask whether closer attention to the institutional environment, to change over time, and to changes in membership composition would both indicate and explain changes to meta-organizational character and strengthen the core claims of the theory of meta-organization.

The paper is organized as follows. We start by discussing the characteristics of meta-organizations and review the literature on dynamics relevant to this associational form of organization. We describe our methodology and the data included in the study. This is followed by an account and discussion of the dynamics of the case against the terms of meta-organization theory.

### **2. Characteristics of Meta-Organizations and Key Sites of Dynamics**

In a recent series of publications, Ahrne and Brunsson (2005, 2008, 2011) have proposed that meta-organizations are a distinct type of inter-organizational entity (Cropper et al. 2008). They stipulate some characteristics of the form: "*meta-organizations are all associations; membership is voluntary and members can withdraw at will. The purpose of a meta-organization is to work in the interests of all its members, with all members being equally valuable and membership being based on some form of similarity*" (Ahrne and Brunsson 2008, p. 11). Although, in these senses, meta-organizations may be likened to other associations, Ahrne and Brunsson (2008) emphasize the significance of the composition of meta-organizations: Members are organizations not individuals.

Ahrne and Brunsson (2008) also stipulate a formality of character. A meta-organization "*is not the same as a network, class, or society. For people to believe that something is an organization, it must have members, a hierarchy, autonomy, and a constitution*." (Ahrne and Brunsson 2008, p. 45). Such formalization, seen as a process and an outcome (Vlaar et al. 2006), creates the conditions for a 'decided order' (Ahrne and Brunsson 2011; Ahrne et al. 2016). Decisions concerning members specify who are in, and who are out and so set the boundary. The hierarchy exists as a center of authority: Though this may be no more than a mechanism for making decisions, it may equally hold the right to issue commands and rules prescribing members' actions. Such rights and rules are laid down in constitutions, which may also describe the goals or tasks of the organization. Ahrne and Brunsson (2008) observe that a meta-organization can gain a degree of autonomy, and become recognized as an organization in their own right. Formalization can have positive effects in terms of a meta-organization's presence and agency; however, it may also provide false impressions of comprehensibility and controllability (Vlaar et al. 2006).

In their account of dynamics, Ahrne and Brunsson (2008) focus substantially on the composition and order of the meta-organization, arguing that "*meta-organizations can best be understood as being in a transitional phase between a weak organization with strong members and a strong organization with weak members*" (p. 132). Like other accounts that focus on inherent instabilities (Das and Teng 2000), tensions (Huxham and Vangen 2000), or dialectics (De Rond and Bouchikhi 2004) in inter-organizational entities, meta-organization dynamics are held to arise from the interplay between members, the meta-organization, and their collective activities. Although Ahrne and Brunsson (2008) explain the formation of a meta-organization as a move on the environment, they say less about the way the

meta-organization might continue to respond to field influences as a source of change. The boundary between organizations and their environment is moved, but a new boundary and set of relationships both within the meta-organization and between the meta-organization and its environment are produced. Other accounts of organizational change, including the co-evolutionary perspective, recently reexamined by Rodrigues and Child (2003, 2008), frame dynamics essentially in terms of the mutual and reciprocal influence between an organization as a collective effort or strategy and its environment (see also Barnett et al. 2000; Selsky 1998). We move on to discuss institutional environment and membership composition as sources of dynamics using a third lens—that of patterned transitions through time, i.e., history.

### *2.1. Institutional Environment*

The environment of organizations is unpredictable and meta-organizations are an attempt at organizing salient parts of the environment (Ahrne and Brunsson 2008). As Ahrne and Brunsson (2008, p. 56) note, "*Creating meta-organizations entails the reduction of environment and an increase in organization—transforming part of what was once the members' environment into organization. Instead of constituting each other's environment, the organization's members become members in the same organization*." A new and additional organization is formed while no organization is 'lost' (unlike in a merger or acquisition). Members, thus, "*retain their organizational boundaries, but a new boundary is placed around them all*." (p. 64). While the meta-organization organizes part of the environment, the environment does not disappear. Ahrne and Brunsson note, "*To a great extent, organizations must accept the environment as it is and adapt to it rather than attempt to control it*." (Ahrne and Brunsson 2008, p. 56). However, the meta-organization still needs to deal with its environment and so do the members. Frequently, meta-organizations seek deliberately to influence and to change their environments (and that of the members) for the benefit of the members. Meta-organization theory says little about this ongoing work. In his analysis of the developmental dynamics in non-profit federations, Selsky (1998) observes that the federation's dynamics include both forms of adaptation and efforts at construction. Adaptation consists in the federated referent organization's (FRO; Selsky's term is conceptually close to meta-organizations) responses to "*influences and pressures from constituent member organizations, domain elites, and the wider context of resource and policy environments in which they operate*." (p. 298). Construction is where the FRO seeks to shape the context through its own presence and actions. This can be understood in terms of co-evolution (Rodrigues and Child 2003), whereby the environment influences the meta-organization, but, at the same time, the meta-organization influences the environment. Rodrigues and Child (2003) recognize three system levels, that of the macro (the general environment), the meso (the immediate environment), and the micro (the organization internal), which interact. Preempting Rodrigues and Child (2008) work, Selsky concludes that the developmental dynamics of the FRO are characterized by a continuous interplay between the FRO and its context, in which it seeks to maintain effective alignment between its strategy and the significant elements of the field within which it is set. However, becoming an actor that can influence its environment requires a high degree of coordination and a strengthened common identity, according to Ahrne and Brunsson, and may increase the need for similarity among members. In addition, one of the issues with meta-organizations is that the member organizations as individual organizations also remain embedded within their environment; they can act and attempt to influence their environment as well. We expect much more of an active and complex relationship to the remaining environment than simple acceptance.
