Cascading Organizational Change

This article develops a formal theory of the structural aspects of organizational change. It concentrates on changes in an organization’s architecture, depicted as a code system. It models the common process whereby an initial architectural change prompts other changes in the organization, generating a cascade of changes that represents the full reorganization. The main argument ties centrality of the organizational unit initiating a change to the total time that the organization spends reorganizing and to the associated opportunity costs. The central theorem holds that the expected deleterious effect of a change in architecture on the mortality hazard increases with viscosity and the intricacy of the organizational design. (Organizational Change; Cascades; Organizational Mortality) Introduction The theory of structural inertia (Hannan and Freeman 1977, 1984) has motivated a considerable body of empirical research on the effects of change in core organizational features on the hazard of mortality.1 This research generally supports the basic insights of the theory. However, the theory has not developed in parallel with empirical research on these issues. We think that further progress in understanding organizational inertia and change requires sustained attention to theoretical foundations. This article has two broad objectives. First, we propose a formal language that is suitable for expressing the insights of the original theory and also supporting efforts to deepen and broaden it. Second, we use this language to embed the original arguments in a richer model of organizational structure and to show that some key assumptions of the original theory can be derived as theorems in the new one. This effort involves elaborating what might be called microfoundations for organizational ecology.2 At least four types of processes delay—and often prevent—organizational change: (1) structural processes, including the consequences of intricacy and viscosity (sluggishness of response); (2) institutional processes, involving identities and the “moral” character of structural arrangements; (3) political processes, involving interests and interest-group politics; and (4) learning processes, involving feedback over time. While any specific major organizational change likely activates several, or even all, of these processes, we think it clarifies the analytical picture to deal with them separately. This article addresses only some of the basic structural processes; it makes no claim to comprehensiveness. The original theory of structural inertia pertains to changes in the “core” features of organizations, not peripheral ones. In developing this theory, Hannan and Freeman (1984) provided some specifics—They claimed that four features constitute a generalized core: (1) the organizational mission, (2) the form of authority and the nature of the exchange between the organization and its members, (3) the basic technology used to transform inputs into outputs, and (4) the organization’s general marketing strategy. Limiting the scope of the theory in this way—the restriction to a set of purported canonical features—now seems unfounded. Moreover, this limitation on scope does not constrain empirical research: A review of studies testing the main implications of the theory shows that researchers claim a very wide variety of organizational features as among the canonical core features (Barnett and Carroll 1995, Carroll and Hannan 2000). We advance here a structural alternative to specifying coreness. We examine changes in an organization’s architecture (defined as a code system) and analyze their impacts in terms of the cascades of subsequent changes throughout the organization. We follow the original theory in positing that such architectural changes initiate periods of reorganization. We define such reorganization periods precisely in two ways: (1) the total time spent reorganizing in all parts of the organization even if some occur simultaneously and (2) the temporal span required to bring all the organization’s feature values in line with the new architecture. We assume that the total time spent reorganizing increases with the intricacy of an organization’s design, where intricacy is defined as a strong and 1047-7039/03/1405/0463 1526-5455 electronic ISSN Organization Science © 2003 INFORMS Vol. 14, No. 5, September–October 2003, pp. 463–482 MICHAEL T. HANNAN, LÁSZLÓ PÓLOS, AND GLENN R. CARROLL Cascading Organizational Change complex pattern of interconnections among the organization’s component units and with viscosity, the typical time it takes for a unit to respond to changes and bring local architecture into conformity. The second part of the argument ties these considerations to organizational mortality using the standard assumption that an organization’s hazard of mortality is proportional to its stock of resources. We assume that reorganization diminishes an organization’s ability to take advantage of opportunities. It then follows that organizations undergoing reorganization miss more opportunities relative to those that are not reorganizing, thereby lowering resources and increasing the hazard of mortality. The key theorems hold that the expected effect of organizational change on the mortality hazard increases with the intricacy of design and viscosity. Formalizing Organizational Ecology Organizational ecology and demography have spawned a variety of thriving theory fragments and bodies of related empirical research. These include theories of: • structural inertia and change (Hannan and Freeman 1984, Haveman 1992, Barnett and Carroll 1995, Hannan et al. 2003), • age dependence (Barron et al. 1994, Sørensen and Stuart 2000), • niche width (Hannan and Freeman 1977, Dobrev et al. 2001), • resource partitioning (Carroll 1985, Carroll and Swaminathan 2000, Park and Podolny 2000), • density dependence (Hannan 1989, Carroll and Hannan 1989, Hannan and Carroll 1992, Barron 1999, Ruef 2002), • localized competition (Hannan and Freeman 1977, McPherson 1983, Baum and Singh 1994, Podolny et al. 1996), • organizational identities and forms (Zuckerman 1999, Ruef 2000, McKendrick and Carroll 2001, Baron 2002), • social movement forms (Minkoff 1999, Ingram and Simons 2000, Olzak and Uhrig 2001, Swaminathan and Wade 2001, Sandell 2001), • red-queen evolution (Barnett and Hansen 1996, Barnett and Sorenson 2002), and • recruitment-based competition (Sørensen 2000). These strands can sensibly be regarded as fragments in a larger research program because they (1) build on a common conception of the organizational world as shaped by processes of selection and (2) share methodological presumptions and practices (Carroll and Hannan 2000). A considerable amount of formalizing activity has already taken place in this arena. Various fragments have been subject to rational reconstruction and logical analysis designed to test the soundness of the arguments. Fragments analyzed in this manner include: structural inertia and change (Péli et al. 1994, 2000), niche width (Péli 1997), and age dependence (Hannan 1998; Pólos and Hannan 2002, 2003). These efforts took seriously the “frozen” published texts; that is, they translated the natural-language renderings of the arguments into a formal language and checked the proofs of the claims in that language. Although this work has been valuable in establishing the soundness of the main arguments and in filling gaps in arguments, it has taken a largely passive role with respect to moving the theories forward. In particular, because these efforts consider each fragment in isolation from the others, they did not clarify the relationships among the fragments. The relationships among organizational ecology’s theory fragments do require clarification. The preponderance of effort over the last 25 years has focused on empirical testing, with relatively little attention paid to issues of theoretical integration. A lack of progress on integration makes it hard to envision exactly which empirical projects would move the larger program forward substantially. In a pair of previous articles, we began a series of projects aimed at integrating the theory fragments by developing a formal language that allows a precise definition of the key units of organizational ecology: form and population (Pólos et al. 2002) and the niche (Hannan et al. 2003). This article continues the effort by turning to the Hannan-Freeman (1984) theory of structural inertia. We attempt to embed the theory in a richer model of organizational structure such that some of what the original theory assumed can be derived. Moreover, we want to construct this model broadly so that it can also serve as a foundation for other related fragments. We use logic as the modeling tool because it allows us to keep the structure of the argument close to the original formulations and the key intuitions while also establishing the soundness of the arguments. We use a particular nonmonotonic logic that regards the causal claims of the arguments to be rules with exceptions. The Appendix sketches the logic; full details can be found in Pólos and Hannan (2003). We build models on two levels. We specify processes holding for units in an organization and we derive implications at the organization level. We concentrate on the possibility that actions in one unit can set off cascades of actions in other units. We argue that long cascades complicate organizational action and heighten the risk of failure. Our formal representation of these arguments 464 Organization Science/Vol. 14, No. 5, September–October 2003 MICHAEL T. HANNAN, LÁSZLÓ PÓLOS, AND GLENN R. CARROLL Cascading Organizational Change Table 1 Notation: Logical Constants, Predicates, Functions,