The concern here is less with progressive distortions that occur as the firm grows in size than it is with the types of distortions that are prospectively associated with any transaction which a firm of medium to large size and complexity chooses to discharge. The social psychology and organization theory literatures are the sources for most of the argument. Rather than express the argument in the language of markets and hierarchies from the outset, I first present the issues in the language which the social psychologists and organization theorists have employed. A brief interpretation in terms of the framework is then offered. Although some of the richness of description is lost in the translation, so that reverting to the original language may often be more instructive, I trust that the applicability of the organizational failures framework to the study of internal organizational frictions is made apparent.
1. Sources of Transactional Distortion
Among the more severe goal distortions of internal organization are the biases which it experiences that are favorable to the maintenance or extention of internal operations. Biases of three types are discussed: internal procurement, internal expansion, and program persistence. Communication distortion supports all three. Systems rationality versus subgroup rationality conflicts are posed throughout. The distinction is this: systems rationality entails global optimization whereas rational subgroup behavior is that which enhances the individual and collective objectives of those parties who effectively control the transaction, or related set of transactions, in question.
1.1. INTERNAL PROCUREMENT
The internal procurement bias is supported by a number of factors. For one thing, the internal supplier that produces mainly for internal uses may be judged to be at a relative disadvantage in the marketplace. The internal supplier may lack both the large and experienced marketing organization and the established customer connections that nonintegrated external suppliers have access to. In consideration of these conditions, and since fixed costs are sunk, a “preference” for internal procurement might seem appropriate at least so long as the external price exceeds the variable cost of internal supply.
This may be a specious argument, however, since the fixed costs in question may easily be overstated (there may be a second-hand market for the machinery in question) and individual equipment renewal decisions ought eventually to be made with reference to the long-run viability of the internal facility. Managers are notably reluctant, however, to abolish their own jobs – even in the face of employment guarantees. The problems with such guarantees are that while continued employment may be secure, assurances that status will be maintained, when a position is eliminated, and that promotion prospects will not be upset, upon being romoved from a known promotion ladder, are unenforceable. A preference for internal supply is thus to be expected and may manifest itself by urging that each equipment replacement decision be made serially, in semi-independent fashion. A fundamentally nonviable internal capability may be uncritically preserved in this way. Internal cross-subsidization will then exist.
More generally, the argument is that the existence of an internal source of supply tends to distort procurement decisions. Subgoals of a group or bureaucratic sort are easily given greater weight in relation to objective profitability considerations. This is supported by common membership and the structure of social relations, including a system of shared beliefs and orientations, that emerge.
A norm of reciprocity easily develops. Equivalence of return between the parties, moreover, is inessential to this result. The opportunities for 68 reciprocity are simply more extensive internally (I buy from your division, you support my project proposal or job promotion, and so forth) than in the market, where reciprocity is mainly limited to commodity trades that are a matter of record. Distinguishing logrolling distortions from “constructive cooperation” between internal parties is thus made difficult by the various and subtle forms that internal reciprocity can take. Exceptions from a systems rationality procurement standard are relatively easy both to self-justify and implement in these circumstances.
1.2. INTERNAL EXPANSION
The expansionary biases of internal organization are partly attributable to its dispute settling characteristics. Fiat is efficient for reconciling instrumental differences, but it is poorly suited for mediating disputes that have internal power consequences. A common method of dealing with internal system strain is to adopt a compromise solution by which concessions are made to subsystems rather than require them to give up essential functions or resources (Katz and Kahn, 1966, p. 101). This size preserving tendency is favored by the positive association of both pecuniary and nonpecuniary rewards with size, at least among the functional parts of the enterprise (Williamson, 1970, pp. 51-52). An expansionary bias commonly obtains (Marshall, 1932, pp. 321-322).
Persistent conflict or generalized malfunctioning results frequently in role proliferation: . . . the creation of specialized roles is a generalized solution for all organizational problems… It is the modal solution for insuring the role requirements of organization are met; a new role (or many) is added to the organization, the requirements of which are solely to see to it that other role occupants are performing in the required manner, at the approved pace, and in the prescribed relation, one to another” (Katz and Kahn, 1966, p. 203). Thus, in addition to the generalized expansionary bias described in Section 2, below, internal organization specifically favors the extension of the compliance machinery. Policing costs easily become a disproportionate share of the total if these tendencies are not checked. Where markets may be said to work well, therefore, the firm might consciously resist the internalizing of incremental transactions for this reason as well.
1.3. PERSISTENCE
Consider the matter of persistence behavior. Partly, this is simply a sunk cost phenomenon (March and Simon, 1958, p. 173; Downs, 1967, p. 195): existing activities embody sunk costs of both organizational and tangible types while new projects require initial investments of both kinds. The sunk costs in programs and facilities of ongoing projects thus insulate existing projects from displacement by alternatives which, were the current program not already in place, might otherwise be preferred. Moreover, unlike plant and equipment, specialized investments in organizational infrastructure may experience little depreciation. Persistence of a wholly objective variety thus can continue over a long interval.
But persistence is also favored by other factors. “If the . . . administrative system has committed itself in advance to the correctness and efficacy of its reforms, it cannot tolerate learning of failure” (Campbell, 1969, p.10). March and Simon, in their discussion of innovation, suggest that such commitment is common. They distinguish program elaboration, which corresponds roughly to the program proposal and development stages, from program execution, which involves continuing operations, and note that decisions made at the program elaboration stage are rarely re-examined at the execution stage (1958. p. 187).
As a comparative institutional matter, the question is whether firms and markets can be expected to display differential attitudes toward or experience differential opportunities for persistence. One basis for expecting such differences is that the internal procurement bias referred to above would presumably support project persistence within firms that is unavailable between firms across markets. Information impactedness may also contribute to this result. Thus it may be impossible, at reasonable cost, to easily distinguish between faulty and meritorious internal performance. Responsible parties who are unable to reveal the objective causes of failure and be absolved of fault are thereby induced to press for program extensions beyond objectively rational cut- off limits in the hope that the environment will change and save their reputations. The lack of a cross-subsidization base prevents managers of specialized firms, who may be similarly inclined but are more fully subject to the discipline of the market, from effecting such persistence preferences.
Sequential decision-making procedures, designed to permit project review on the merits, if they exist at all, are often overwhelmed by partisan appeals due to the tie-in of advocacy and administration (Campbell, 1969, p. 10). Although this tie-in may have the effect of mobilizing energies that would not otherwise be available, in that individuals will expend super-normal efforts to rectify error, the error admission properties of internal organization would appear to be defective. Fixed costs are not sunk but need to be justified (Wolf, 1973, p. 667); the decision to proceed is transformed into a commitment to “succeed” — whatever the costs. By contrast, interfirm contracts, which separate advocacy from administration, are apt explicity to provide for periodic review and favor tougher-minded, more calculative assessments if “failure” is objectively indicated.
Drucker contends in this connection that while “[n]o institution likes to abandon anything,” budget based institutions are more prone to persist with unproductive or obsolete projects than are revenue based institutions — since the necessary support for the latter will be removed by the market (1973, p. 52). Shifting the incremental transaction from the market to the firm generally results in greater budget based support, whence vertical integration gives rise to persistence tendencies.
1.4. COMMUNICATION DISTORTION
One of the serious problems to which market exchange is subject is that the information exchanged between the parties in small-numbers bargaining situations may be manipulated for strategic advantage. Integrating these transactions serves to attenuate these effects. It would be incorrect to conclude, however, that internal communication is not subject to distortion whatsoever. Although incentives are intendedly harmonized by internalization, members of the organization may seek to promote personal goals by diverting the communication system to their own uses (Simon, 1961, p. 171). This is implicit in the discussion of internal reciprocity, expansion, and persistence tendencies discussed above and is explicit in the discussion of idiosyncratic exchange (in employment relation and inside contracting respects) in earlier chapters.
Communication distortions can take either assertive or defensive forms. Defensively, subordinates may tell their supervisor what he wants to hear; assertively, they will report those things they want him to know (Katz and Kahn, 1966, p. 246). Recognition of these assertive tendencies is reflected in the common law of testimony in which demonstrated self-interest on the part of a witness can lead to discredtting of testimony (Campbell, 1958, P.350). Distortion to please the receiver is especially likely where the recipient has access to extensive rewards and sanctions in his relations with the transmitter, as in up-the-line communication in an administrative hterarchy (Campbell, 1958, P. 351). The cumulative effects across “ stve hierarchical levels of these and related adjustments to the data easilv result in gross image distortions (Boulding, 1966. p. 8) and contribute to l limitation of firm size (Williamson, 1970, Chap. 2).
Related to this matter of communication distortion for assertive per- sonal or group purposes is the possibility that disaffected members of the organ|zat;on may, rather than quit the organization, choose to subvert ” Organizational loyalties are imputed to and strategic privileges are accorded o members that are denied to outsiders. (In consideration of the privileges to which members have access, subversion in time of war is commollv punishable by death.) The disaffected employee whose estrangement is unknown may deliberately plant misinformation or disclose sensitive in- TorThe’en,t0,0tUhSldrrS *n WayS ‘hat impair the Perf°™ance of the firm, the extent that large, complex organizations are more easily subject such subversion — which presumably they are-smaller, less complex (hence, nonmtegrated) organizations are preferred.
Campbell notes that human links in communication systems are prone o a whole senes of biases. Some of these biases are little affected whether transactions are organized through the market or internally. There are circumstances, however, where specialized (hence, nonintegrated) organi-zations may have a special advantage.
“Evolutionary considerations lead to the expectation that no constant errors will be found m an environment ecologically typical… Where constant errors have [an ecologically atypical] origin, they will be found to be part and parcel of psychological processes of general adaptive usefulness (Campbell, 1958, p. 3^0) .9 To the extent that organizations are required to deal with ecologically atypical conditions, an isolated organization hat trams its personnel to be sensitive to the special characteristics of the atypical environment and faithfully to reproduce observations – without introducing adjustments that, in general, have “adaptive usefulness” but which in the particular case, are dysfunctional — is indicated. Although full structural disassociation between the parts of an integrated enterprise that deal with ecologically atypical conditions and the rest of the firm is possible, integration can scarcely be expected to yield advantages in relations to the market in these circumstances.
2. Transactional Disabilities Interpreted
Recall that the joining of bounded rationality with uncertainty makes contractual completeness expensive (if not infeasible) to attain, while incomplete contracts expose market-mediated exchange to the hazards of profit haggling if small-numbers bargaining conditions obtain. Because it is able to suppress or avoid opportunistic profit haggling, internal organization is able to tolerate contractual incompleteness — thereby to economize on bounded rationality, by employing an adaptive, sequential decision-process, without incurring contractual hazards. But internal organization is nevertheless vulnerable to opportunism in other respects.
Again, however, it is not opportunistic inclinations by themselves which explain the problem. Rather, it is opportunism in conjunction with both a small numbers and an information impactedness condition that accounts for the transactional disabilities that internal organization experiences.
Shifting a transaction from the market to the firm is significant not because a small-numbers exchange relation is eliminated but rather because the incentives of the parties are transformed. Indeed, the typical internal transaction is really a small-numbers exchange relation writ large. Investment by a firm in fixed plant and organizational infrastructure serves to insulate internal transactions from competition in the product market — with the result that, in the short run at least, there may be no credible alternative source of supply whatsoever. 76 Functional managers, naturally, are not unaware of this condition. Coupled with the information impactedness advantage that they enjoy, a nontrivial degree of managerial discretion obtains.
Internal opportunism takes the form of subgoal pursuit – where by subgoal pursuit is meant an effort to manipulate the system to promote the individual and collective interests of the affected managers.^ Such efforts generally involve distorting communications in a strategic manner. Sometimes this may be done by individual managers without the support of others. More often, however, internal distortions are due to the cooperative efforts of two or more managers. ( The internal reciprocity phenomenon is a specific manifestation of such collective distortion.)
Although internal auditing and experience-rating serves to check egregious distortions, the general management in the firm is nevertheless severely limited in information impactedness respects. It is simply too expensive (or, for bounded rationality reasons, perhaps infeasible) for the general management to be apprised of everything that is going on at the operating level and adjust compensation accordingly. Moreover, what is of special importance here, these problems become more severe as the firm becomes progressively more complex.
The problem with successively adding related transactions, holding size constant, 14 is that even small degrees of overlap give rise to joint responsibility. Imputations cannot be made with confidence when the nature of the interaction is costly to sort out and the parties enter conflicting claims.
Furthermore, fiat cannot easily be invoked where equity issues are at stake. Norms of internal justice, which support quasimoral involvement, check attempts at vigorously implementing the compliance machinery so long as the “defendant” can establish a reasonable doubt by asserting joint responsibility. The resulting confounding of accountability impairs incentives. Thus, although internal organization is (potentially) better able than the market to distinguish between random events and meritorious performance, and in this respect is superior to the market as a mechanism for assigning rewards to deeds, the inference difficulties that internal auditing experiences as the firm grows in complexity eventually limit its power in this respect.
The upshot of this is that distortion-free internal exchange is a fiction and is not to be regarded as the relevant organizational alternative in cir- cumstances where market exchange predictably experiences nontrivial frictions. Implicitly, whenever a transaction is shifted from the market to the firm, a tradeoff is struck in which it is understood that some irreducible degree of opportunism will continue.
It is significant, moreover, to realize that many of the costs of opportunism will not show up immediately; the firm that is considering shifting from a market to an internal transaction needs to be especially sensitive to future distortions. Thus, the decision to continue a project, once it is begun, the decision to renew or expand the internal facility, once it is in place, and the decision to procure internally, once that capability exists, may not be decisions for which the ordinary profitability calculus will govern. Proposals to internalize a transaction might therefore reasonably be made to “promise” an immediate, nontrivial cost advantage. Required rates of return on new programs should accordingly be set at uncommonly high levels in order to offset predictable, later stage distortion tendencies. (The widely observed practice of stipulating higher marginal rates of return on new programs than the actual rate that the firm realizes on existing programs is consistent with the argument. The same practice should presumably be followed in evaluating government investments, though I suspect that it is not.)
Source: Williamson Oliver E. (1975), Markets and hierarchies: Analysis and antitrust implications, A Study in the Economics of Internal Organization, The Free Press.