We have already noted that an organizational form gives unitary character to a population of organizations. But what does this mean concretely? A number of answers to this question have been given in the population- ecology literature.
1. Organizational Genetics
One set of answers looks for an analogue to the genetic structure that reproduces biotic forms. When viewed abstractly, such genetic mechanisms can be thought of as blueprints in the spirit of Monod (1971). the structure of DNA molecules contains a set of instructions for building biotic structures. We have suggested (Hannan and Freeman 1977) that organizational forms be defined analogously, as instructions for building organizations and for conducting collective action.
The blueprint imagery refers to qualitative patterns—a blueprint codes specific, discrete instructions. It has the advantage of allowing variations in structures produced from the same blueprint, because the relation between blueprint and final structure depends on the cultural conceptions and training of those who implement it, the quality of tools and resources at hand, and the nature of environmental shocks during the period of assembly. So this conception of form recognizes the possibility that populations with different forms may overlap in terms of observable characteristics. That is, modest overlap of quantitative characteristics is consistent with qualitative differences between underlying blueprints.
This approach has the strong disadvantage that blueprints for organizations are not observable, making this conception of form far from concrete. Earlier (Hannan and Freeman 1977) we proposed that blueprints could be inferred from formal structure, patterns of activity, and forms of authority. But these observable features of populations often overlap considerably, as in the case of the core and peripheral forms discussed earlier. In the absence of strong theory it is hard to locate the underlying invariance; and no strong theory exists for this problem.
2. Taxonomy Based on Organizational Genetics
A related proposal for identifying forms makes a stronger appeal to the analogue to population genetics and makes a strong claim for the priority of taxonomy over explanation (McKelvey 1982; McKelvey and Aldrich 1983). It treats the problem of classifying organizational forms as analogous to classifying biotic species and tries to find the analogue to the individual gene. For example, McKelvey (1982) argues that bits of productive knowledge, which he calls “comps,” serve as a useful analogue. Tracing flows of comps between organizations specifies family trees and allows classification of forms based on considerations of organizational genetics.
We have serious doubts about the utility of an organizational genetics approach to the problem of classification. First, there is the question of whether comps exist as individual properties disembodied from organizational context. In particular, can they be transferred unchanged from one organization to another? Even the simpler biotic case involves strong non- linearities; structure depends on combinations of genes in a highly nonadditive way. The best example of this kind of process concerns the so-called regulator gene that controls the timing of developmental events. Small differences in timing have enormous consequences for structure.
The routines that coordinate elementary production in organizations are analogous to regulator genes. Even when the competencies of individual members are standardized across organizations, as in the case of certain crafts and professions, the coordinating routines tend to be highly specific to organizational forms and to particular organizations.12 It would be very difficult to discern the differences between, say, public and private universities by simply recording a list of the competencies held by members of the two kinds of organizations. The difference in form would seem to lie elsewhere.
Second, the process of transmission of structure is not unitary for orga- nizations as it is for biotic species. In the biotic case, virtually all of the information governing structure is passed from parent to offspring in the single event of reproduction. The genetic material possessed by an individual is essentially constant over the lifetime except for mutations in cell reproduction. In the case of organizations, there are two important differences. Transmission is not unitary; information comes from diverse sources, which means that there is no clear-cut parent. In addition, the flow of information about buildiûg structure never ceases for an organization because members come and go continually over its lifetime. These
differences greatly complicate the prospect of building an evolutionary taxonomy of organizations.
We also wonder whether definition of forms based on organizational genetics is necessary for organizational ecology. The usual presumption is that a Darwinian theory of organizations cannot be built until Linnaean classifications have been constructed. But it is far from obvious that this was true even in the case of Darwin and Linnaeus. Darwin relied less on a complete set of species classifications than on his own naturalistic observations, both the well-known studies in the Galapagos and his lifelong studies of barnacles. In fact, if history had worked out differently and Darwin had known and used Mendel’s theory of particulate inheritance, the conventional wisdom would no doubt say that Darwinian theory could not have been built without an understanding of these genetic mechanisms.
Darwin built a successful theory of evolution using the wrong theory of genetics—a theory of blending inheritance—and an inaccurate set of species classifications. The connection between creating useful theories of change and correct classification is a loose one. In fact, what matters is that classifications accurately reflect discontinuities in nature. All that mattered for Darwin is that he understood the main distinctions among species of finches and among species of barnacles. This is not to say that organizational taxonomy, based on organizational genetics or on other principles, is irrelevant to developing theories and empirical research on organizational ecology and evolution; rather, the success of the former is not a precondition of the success of the latter.
3. Duality of Niche and Form
An alternative approach tries to identify forms in terms of the niche structure of populations.13 The niche of a population consists of combinations of resource abundances and constraints in which members can arise and persist. There is a fundamental duality here: niches define forms and forms define niches. Niche structure can be summarized by a fitness function, which is a rule relating levels of environmental conditions to growth rates of the population. So a reasonable idea is to infer differences between forms from empirical fitness functions. That is, we might use empirical estimates of fitness functions to decide whether a claimed distinction between a pair of similar populations reflects some fundamental difference in form. Unfortunately, fitness functions are not directly observable and are costly to estimate; thus a great deal of research is required just to define forms in this way.
This strategy of using the duality of niches and forms to define the latter can be approximated by defining populations in terms of a set of core properties. For example, we suggested earlier that four properties—stated goals, forms of authority, core technology, and marketing strategy—provide a useful basis for classifying organizations into forms because an organization’s initial configuration on these dimensions commits it to a set of environmental dependencies and to a long-term strategy (Hannan and Freeman 1984).
As we discuss at length in Chapters 5 and 6, a fitness function implies a carrying capacity. Thus propositions about fitness functions can be restated as propositions about carrying capacities. In particular, the structure of the niche can be defined in terms of effects of environmental conditions on carrying capacities. In this sense, two populations can be said to be distinct if their carrying capacities are affected differently by the same environmental conditions or by different sets of environmental conditions.
McPherson (1983) used the dependence of carrying capacities on envi- ronmental conditions to analyze the niche structure of a set of voluntary associations. Using data on memberships in a list of conventionally defined kinds of associations such as labor unions, sports associations, and youth- serving associations, he measured niches in terms of concentration of membership in a space defined in terms of attributes of potential members, including age, occupation, sex, and education. Competition between pairs of kinds of associations for members is measured as the degree of overlap of these niches. Assuming aggregate equilibrium, McPherson used these estimates to solve the Lotka-Volterra competition equations14 for the fun- damental carrying capacity (the carrying capacity in the absence of com- petitors) of each form of organization. Finally, he showed that inserting estimates of intrinsic carrying capacities into the model gives a reasonably good fit to observed population sizes which, under the assumption of equi- librium, equal the realized carrying capacities. McPherson notes that this result confirms the validity of the typology of associations used in his analysis.
4. Structural Equivalence
DiMaggio (1986) points out that our definition of forms is “based firmly in the logic of structural equivalence.” The notion of structural equivalence, developed mainly by Harrison White and his collaborators (White 1963; Lorrain and White 1971), has become most familiar in the formal structure and operational procedures of blockmodeling (White, Breiger, and Boorman 1976). A blockmodel is a hypothesis about a set of ties among some population of actors. Suppose that each tie is represented in the population as a square matrix whose entries tell whether the tie exists between the indicated pair of actors. Arrange each matrix so as to create blocks (that is, sets of actors) for whom the tie does not exist; these are the so-called zero- blocks. Consider the set of all the relevant matrices. A blockmodel is a hypothesis that some single partition of the population will reproduce the zero-blocks for all of the tie-matrices. The set of actors identified as members of the same block are said to stand in a structurally equivalent position with respect to the ties examined and the boundaries specified for the population.
A number of algorithms exist for developing blockmodel images of data on ties. DiMaggio (1986, p. 360) argues that these procedures provide a natural way to implement our conception of forms: “The power of the population ecology perspective may, in certain circumstances, be enhanced by returning to an operational definition of niche and form as mutually defined by observable patterns of relations among sets of actors.” The idea is to obtain data on flows of resources among organizations and to use blockmodeling procedures to identify structurally equivalent sets of organizations. Such sets are then considered to be populations having a common organizational form and occupying the same niche.
This proposal has the advantage of directing attention to patterns of dependence on environments including other organizational populations. We wonder, however, whether this approach can provide temporally stable classifications of forms. It seems likely that small changes in ties in a network, resulting perhaps from the demise of one of the earlier organizations, will cause numerical clustering or blockmodeling procedures to yield quite different block structures.15 That is, it is likely that such a procedure is sensitive to relatively small perturbations in the observable data. If so, this approach will have difficulty identifying enduring bases of unit character of populations. However, it remains to be seen whether this approach offers advantages that offset this likely disadvantage.
The various approaches reviewed in this section diverge in considering internal attributes of organizations versus external dependencies as bases for classification. However, they converge in focusing on the content of the organization structures or the character of external relations. The implicit underlying idea is that content and relations define the boundaries around populations. That is, boundaries should be drawn in organizational space to produce uniformity within populations in terms of observable attributes and relations.
Source: Hannan Michael T., Freeman John (1993), Organizational Ecology, Harvard University Press; Reprint edition.