That peer groups sometimes offer advantages in relation to autonomous contracting is apparent from the discussion in Section 1. But as Section 2 makes clear, peer groups also face limitations in both opportunism and bounded rationality respects. The restricted range of control techniques to which peer groups have access, in seeking to mitigate these conditions, should be appreciated. Realizing more effective control may be possible only if the peer group structure is compromised so severely that it is altered in kind rather than merely in degree.
However, before examining the possible advantages of simple hierarchy over peer groups in bounded rationality and opportunism respects, it will be useful to first consider technological nonseparabilities as a reason for supplanting markets by internal organization. This is interesting for two reasons. For one thing, peer’group organization has little to offer as a remedy for defective market exchange when such nonseparabilities appear. Instead, simple hierarchy is apt to obtain immediately. Second, Alchian and Demsetz (1972) rely on such nonseparabilities as the leading reason to explain the origin of firms. I take exception with their view and emphasize transactional factors instead.
As indicated, Alchian and Demsetz contend that nonseparabilities occasion the shift of economic activities from market-mediated exchange to internal organization. While I concede that this is sometimes the case, I submit that: ( 1 ) there are reasons other than nonseparabilities for internal organization to appear; (2) nonseparabilities are much less widespread than is commonly believed: and (3) though nonseparabilities constitute one of the reasons for primary work groups to appear, the argument has little bearing on the joining of simple hierarchies to form complex hierarchies.
The first of these propositions follows from the discussion in Section 1.
The third proposition will be developed in the later chapters where I argue that economizing on the costs of transacting between technologically separable activities occasions the evolution of complex hierarchies.26 My concern here is accordingly restricted to proposition (2).
The manual freight loading example is used by Alchian and Demsetz to illustrate worker nonseparabilities. The task involves loading heavy cargo into trucks and in order to be done efficiently two men are required to work coordinately. Because of interaction effects between them, the marginal productivity of each w’orker cannot be determined by observing the total weight loaded during the day. Rather, where “team” production is involved, as it is here, “measuring marginal productivity and making payments in accord therewith is more expensive by an order of magnitude than for separable production functions” for which productivity is additive (1972, p. 779). Accordingly, market mediation is supplanted by internal organization in which a “boss” monitors the performance of the team and allocates rewards among members on the basis of observed input behavior. Shirking is purportedly attenuated in this way (1972, pp. 779-781).
However familiar the manual freight loading example is among econo- mists and however widespread opinion that such nonseparabilities are ubiquitous, I submit that there are many tasks for which such interaction effects are absent. The issue may be put operationally in the following w’ay. Holding individual task technologies constant, up to but not including the physical transfer of a product from one stage to the next, is it commonly possible to sever the connections between successive workers by placing an intermediate product inventory between them? Many tasks, I submit, are separable in this sense — often between individual workers and almost invariably between small groups of workers. Internal organization thus arises less on account of nonseparabilities than as a means by which to economize on buffer inventories and mitigate costly haggling between technologically separable stages — in most instances because of transactional considerations.
Consider Adam Smith’s pin-making example ( 1937. pp. 4-5). Pin manu- facture involved a series of technologically distinct operations (wire straightening, cutting, pointing, grinding, and so forth). In principle, each of these activities could be performed by an independent specialist, and work could be passed from station to station by contract. The introduction of buffer inventories at each station, moreover, would decrease the coordination requirements and thereby reduce contractual complexity. Each worker could then proceed at his own pace, subject only to the condition that he maintain his buffer inventory at some minimum level. A series of independent entrepreneurs rather than a group of employees, each subject to an authority relation, would thus perform the tasks in question.
Transaction costs militate against such an organization of tasks, however. For one thing, it may be possible to economize on buffer inventories by having the entire group act as a unit, under common direction, with respect to such matters as w7ork breaks, variable rates of production, etc. Although this could be worked out in advance and made explicit in the contract, or the authority to make such decisions could be rotated among the members of the group, the matter might usefully be assigned to a “boss,” who oversees and coordinates the entire operation and can more easily judge the fatigue and related work attitudes in the group.
The more pressing reasons for replacing autonomous contracting by hierarchy, however, turn on adaptability considerations. Suppose one of the individuals becomes ill or injured. Who nominates and chooses a replace- ment, or otherwise arranges to pick up the slack, and how is compensation determined? Reaching agreement on such matters is apt to be relatively costly in relation to having a boss reassign the work among the members of the group or make other ad hoc arrangements on the group’s behalf. Similarly, what is to be done if an individual declines to deliver the requisite quantity or quality to the next station? How are penalties determined? Litigation is apt to be costly and time consuming, and to what avail, if the individual lacks the requisite assets to compensate for the losses attributable to his deviant behavior? For the reasons developed more fully in the chapters which follow, hierarchy and an employment relationship commonly yield transactional economies in adaptability and dispute settling respects.27 By way of anticipating aspects of that argument, it may be useful to observe here that individual workers acquire relatively secure property rights claims to specific job stations under the entrepreneurial organization of tasks described above. These claims are weakened when the set of tasks in question is placed under common ownership and an employment relation is created instead.
2. Bounded Rationality
Supplanting the all-channel network in the peer group by a wheel network, in which each member is connected only with the center, can yield savings in both information transmittal and decision-making respects. Such a radial network will be especially efficacious if, once apprised of the relevant data by the various members of the group, the central coordinator is able, with little assistance or consultation, to reach the correct decision and disseminate it among the members of the group.
It is elementary that the advantages of centralization vary with the degree of interdependence among the members — being negligible in an unrelated task group, possibly great in a common task group (especially if indivisible physical asset utilization problems are posed), and almost certainly great in an integrated task group. If occasions for adaptation between integrated tasks are unpredictable yet common, coordinated responses may be difficult to secure unless the parties have reference to a common signal.
Arrow summarizes the case for the centralization of information flows and decision-making by setting out a series of four propositions ( 1974, p. 68) :
- Since the activities of individuals interact with each other, being sometimes substitutes, sometimes complements, and frequently compete for limited resources, joint decision on the choice of indi- viduals’ activities will be superior to separate decisions.
- The optimum joint decision depends on information which is dis- persed among the individuals in society.
- Since transmission of information is costly, in the sense of using resources, especially the time of individuals, it is cheaper and more efficient to transmit all the pieces of information once to a central place than to disseminate each of them to everyone.
- For the same reasons of efficiency, it may be cheaper for a central individual or office to make the collective decision and transmit it rather than retransmit all the information on which the decision is based.
What is especially relevant to the choice of peer group or simple hier- archy is that, to the extent that the requisite information-processing and decision-making talents are not widely distributed, efficiency will be served by reserving the central information collection and decision-making position to the one or few individuals who have superior information processing capacities and exceptional oratorical and decision-making skills.11 Something of an elite thereby results, as the select subset bears an asymmetrical relation to everyone else. Not only does the peak coordinator enjoy the power which authority and expertise accord him (Katz and Kahn, 1966. Chap. 11). but having more complete information gives him a strategic advantage over everyone else. The peak coordinator has inordinate influence over both the value and factual premises of other members of the group. It is really a fiction, when such an elite develops, to maintain that a peer group any longer exists —even if. in principle, the group can always challenge and even reverse individual decisions. Simple hierarchy effectively obtains.
Upset. Trow, and Coleman’s study of democracy in the International Typographical Union is noteworthy in this regard. They conclude ( 1956, p. 404):
The structure of large-scale organization inherently requires the development of bureaucratic patterns of behavior. The conditions making for the institutionalization of bureaucracy and those making for democratic turnover in oifice are largely incompatible…[Bureaucracy] gives an incumbent administration great power and advantage over the rank and file or even an organized opposition. This advantage takes the form of control oxer iinaneial resources and i nternal communications, a large, permanently organized political machine, a claim to legitimacy, and a near monopoly oxer political skills.
While their remarks apply to the life cycle of bureaucracy in large scale voluntary organizations, similar considerations apply in smaller scale enter- prises as well — albeit in attenuated degree.
As described in Chapter 1, the provision of ex post auditing and exper- ience-rating capabilities served to relieve the moral hazard problems besetting the insurance markets. Assuming that the insurance market model generalizes in this respect, the free-rider problems of peer groups should likewise be mitigated in these ways. The peer group, however, is inherently- limited as an auditing and experience-rating instrument. Designating a member of the peer group to perform the necessary productivity audits and make corresponding adjustments in the compensation of group members violates both the letter and the spirit of peer groups. Making wages contingent on productivity is really a concession that the peer group form of organization is not viable; whatever it may be called, it is functionally equivalent to introducing hierarchy.
The advantages of hierarchy in which one individual, the supervisor, is expressly assigned the tasks of auditing and experience-rating have been discussed by Alchian and Demsetz (1972). As they point out, auditing serves to overcome information impactedness, while paying compensation in accordance with observed productivity discourages malingering. Admission to the work group, moreover, is made conditional on acceptance by the worker of the supervisor’s constitutional authority to perform auditing and experience-rating functions. These constitutional differences between markets and internal organization are central to an understanding of the employment relationship.
Not only does auditing and experience-rating influence the post- admission behavior of members of the work group but it affects the compo- sition of, and standards for, admission to the group as well. Recall in the insurance example that good risks and/or those who were less given to ex post opportunism could be induced to join at the mean premium by the prospect that they (and others) would subsequently be experience-rated. The same holds true in the employment context. That productivity potential cannot be accurately discerned ex ante and that individuals vary in their sense of responsibility to a job are less critical if the organization is able to discern actual productivity ex post and pay the appropriate discriminating wage. High productivity types and/or those who would be prepared to self- enforce covenants not to malinger can thus be induced to affiliate at a low wage by the assurance that this condition will be rectified as information accumulates and more discriminating wage assignments can be made. Moreover, given that those with low productivity and/or high proclivities to malinger will be unable to long exploit the system, easier admission standards can be allowed.28
Source: Williamson Oliver E. (1975), Markets and hierarchies: Analysis and antitrust implications, A Study in the Economics of Internal Organization, The Free Press.