Four types of individualistic contracting modes can be distinguished: (1) contract now for the specific performance of A in the future: (2) contract now for the delivery of x, contingent on event e, obtaining in the future: (3) wait until the future materializes and contract for the appropriate (specific) x at the time; and (4) contract now for the right to select a specific x from within an admissable set X, the determination of the particular x to be deferred until the future. Simon’s study of the employment relation (1957, pp. 183-195) treats contracts of the first type, which he characterizes as sales contracts, to be the main alternative to the so-called authority relation (type 4). This, however, is unfortunate because type 1 contracts, being rigid, are singularly unsuited to permit adaptation in response to changing internal and market circumstances. By contrast, contingent claims contracts (type 2) and sequential spot sales contracts (type 3) both permit adaptation. If complexity/uncertainty is held to be a central feature of the environment with which we are concerned, which it is, the deck is plainly stacked against contracts of type 1 from the outset. Accordingly, type 1 contracts will hereafter be disregarded.
1. Contingent Claims Contracts
Suppose that the efficient choice of x on each date depends on how the future unfolds. Suppose, furthermore, that the parties are instructed to negotiate a once-for-all labor contract in which the obligations of both employer and employee are fully stipulated at the outset. A complex con- tingent claims contract would then presumably result. The employer would agree to pay a particular wage now in return for which the employee agrees to deliver stipulated future services of a contingent kind, the particular services being dependent upon the circumstances which eventuate.
Contracting problems of several kinds can be anticipated. First, can the complex contract be written? Second, even if it can, is a meaningful agreement between the parties feasible? Third, can such agreements be implemented in a low cost fashion? The issues posed can all usefully be considered in the context of the framework sketched out above.
The feasibility of writing complex contingent claims contracts reduces fundamentally to a bounded rationality issue.
Recall in this connection the conclusion reached by Feldman and Kanter in their assessment of complex decision trees, to wit, “The comprehensive decision model is not feasible for most interesting decision problems (1956, p. 615). Plainly, the complex labor agreements needed for comprehensive description of the idiosyncratic tasks in question are of this kind. Not only are changing market circumstances (product demand, rivalry, factor prices, technological conditions, and the like) impossibly complex to enumerate, but the appropriate adaptations thereto cannot be established with any degree of confidence ex ante. Changing life cycle conditions with respect to the internal labor force compound the complexities.
The enumeration problems referred to are acknowledged by Meade in his discussion of contingent claims contracts. “When environmental uncertainties are so numerous that they cannot all be considered . . . or, what comes perhaps to much the same thing, when any particular environmental risks are so hard to define and to distinguish from each other that it is impossible to base a firm betting or insurance contract upon the occur- rence or non-occurrence of any of them, then for this reason alone it is impossible to have a system of contingency . . . markets” ( 1971, p. 183) . Except for bounded rationality, Meade’s concerns with excessive numbers, undefinable risks, and indistinguishable events would vanish.
But suppose, arguendo, that exhaustive complex contracts could be written at reasonable expense. Would such contracts be acceptable to the parties? I submit that a problem of incomprehensibility will frequently arise and impede reaching agreement. At least one of the parties, probably the worker, will be unable to meaningfully assess the implications of the complex agreement to which he is being asked to accede. Sequential contracting, in which experience permits the implications of various contingent commitments to be better understood, is thus apt to be favored instead.
Assume, however, that ex ante understanding poses no bar to con- tracting. Ex post enforcement issues then need to be addressed. First, there is the problem of declaring what state of the world has obtained. Meade’s remarks that contingent claims contracts are infeasible in circumstances where it is impossible, on the contract execution date, “to decide precisely enough for the purposes of a firm legal contract” what state of the world has eventuated ( 1971, p. 183) bear on this. While it is easy to agree with Meade’s contentions, I think it noteworthy to observe that, were it not for opportunism and information impactedness, the impediments to contracting which he refers to vanish. Absent these conditions, the responsibility for declaring what state of the world had obtained could simply be assigned to the best informed party. Once he has made the determination, the appropriate choice of .v is found by consulting the contract. Execution then follows directly.
It is hazardous, however, to permit the best informed party unilaterally to make state of the world declarations where opportunism can be anticipated. If the worker is not indifferent between supplying services of type xi rather than xk, and if the declaration of the state of the world were to be left to him, he will be inclined, when circumstances permit, to represent the state of the world in terms most favorable to him. Similar problems are to be expected for those events for which the employer is thought to be the best informed party and unilaterally declares, from among a plausible set, which e, has eventuated. Moreover, mediation by a third party is no answer since, by assumption, an information impacted- ness condition prevails with respect to the observations in question.
Finally, even were it that state of the world issues could be settled con- clusively at low cost, there is still the problem of execution. Did the worker really supply xi, in response to condition ei, as he should, or did he (oppor- tunistically) supply xi instead? If the latter, how does the employer show this in a way that entitles him to a remedy? These are likewise information impactedness issues. Problems akin to moral hazard are posed.
Ordinarily, bounded rationality renders the discription of once-for-all contingent claims employment contracts strictly infeasible. Occasions to examine the negotiability and enforcement properties of such contracts thus rarely develop. It is sufficient for our purposes here, however, merely to establish that problems of any of these kinds impair contingent claims contracting. In consideration of these difficulties, alternatives to the once- for- all supply relations ought presumably to be examined.
2. Sequential Spot Contracts
Alchian and Demsetz take the position that it is a delusion to characterize the relation between employer and employee by reference to fiat, authority, or the like. Rather, it is their contention that the relation between an employer and his employee is identical to that which exists between a shopper and his grocer in fiat and authority respects (1972, p. 777):
The single consumer can assign his grocer to the task of obtaining whatever the customer can induce the grocer to provide at a price acceptable to both parties. That is precisely all that an employer can do to an employee. To speak of managing, directing, or assigning workers to various tasks is a deceptive way of noting that the employer continually is involved in renegotiation of contracts on terms that must be acceptable to both parties . . . . Long term contracts between employer and employee are not the essence of the organization we call a firm.
Implicit in their argument, I take it, is an assumption that the transition costs associated with employee turnover are negligible. Employers, therefore, are able easily to adapt to changing market circumstances by filling jobs on a spot market basis. Although job incumbents may continue to hold jobs for a considerable period of time and may claim to be subject to an authority relationship, all that they are essentially doing is continuously meeting bids for their jobs in the spot market. This is option number three, among the contracting alternatives described at the beginning of this section, done repeatedly.
That adaptive, sequential decision-making can be effectively imple- mented in sequential spot labor markets which satisfy the low transition cost assumption (as some apparently do, for example, migrant farm labor)35 without posing issues that differ in kind from the usual grocer-customer relationship seems uncontestable. I submit, however, that many jobs do not satisfy this assumption. In particular, the tasks of interest here are not of this primitive variety. Where tasks are idiosyncratic, in nontrivial degree, the worker-employer relationship is no longer contractually equivalent to the usual grocer-customer relationship and the feasibility of sequential spot market contracting breaks down.
Whereas the problems of contingent claims contracts were attributed to bounded rationality and opportunism conditions, sequential spot contracts are principally impaired only by the latter. (Bounded rationality poses a less severe problem because no effort is made to describe the complex decision tree from the outset. Instead, adaptations to uncertainty are devised as events unfold.) Wherein does opportunism arise and how is sequential spot contracting impaired?
Recall from the discussion of opportunism in Chapter 2 that opportunism posed a contractual problem only to the extent that it appears in a small- numbers bargaining context. Otherwise, large-numbers bidding effectively checks opportunistic inclinations and competitive outcomes result. The problem with the tasks in question is that while large-numbers bidding conditions obtain at the outset, before jobs are first assigned and the work begun, the idiosyncratic nature of the work experience effectively destroys parity at the contract renewal interval. Incumbents who enjoy nontrivial advantages over similarly qualified but inexperienced bidders are well- situated to demand some fraction of the cost savings which their idiosyncratic experience has generated.
One possible adaptation is for employers to avoid idiosyncratic techno- logies and techniques in favor of more well-standardized operations. Although least-cost production technologies are sacrificed in the process, pecuniary gains may nevertheless result since incumbents realize little strategic advantage over otherwise qualified but inexperienced outsiders. Structuring the initial bidding in such a way as to permit the least-cost tech- nology and techniques to be employed without risking untoward contract renewal outcomes is, however, plainly to be preferred. Two possibilities warrant consideration: (1) extract a promise from each willing bidder at the outset that he will not use his idiosyncratic knowledge and experience in a monopolistic way at the contract renewal interval: or (2) require incumbents to capitalize the prospective monopoly gains that each will accrue and extract corresponding lump sum payments from winning bidders at the outset.
The first of these can be dismissed as utopian. It assumes that promises not to behave opportunistically are either self-enforcing or can be enforced in the courts. Self-enforcement is tantamount to denying that human agents are prone to be opportunists, and fails for want of reality testing. Enforcement of such promises by the courts is likewise unrealistic. Neither case by case litigation nor simple rule-making disposition of the issues is feasible. Litigation on the merits of each case is prohibitively costly, while rules to the effect that “all workers shall receive only competitive wages” fail because courts cannot, for information impactedness reasons, determine whether workers put their energies and inventiveness into the job in a way which permits task-specific cost savings to be fully realized— in which case disaffected workers can counter such rules by withholding effort.
The distinction between consummate and perfunctory cooperation is important in this connection. Consummate cooperation is an affirmative job attitude — to include the use of judgment, filling gaps, and taking initiative in an instrumental way.B Perfunctory cooperation, by contrast, involves job performance of a minimally acceptable sort — where minimally acceptable means that incumbents, who through experience have acquired task-specific skills, need merely to maintain a slight-margin over the best available inexperienced candidate (whose job attitude, of necessity, is an unknown quantity). The upshot is that workers, by shifting to a perfunctory perfor- mance mode, are in a position to “destroy” idiosyncratic efficiency gains. Reliance on pre-employment promises as a means by which to deny workers from participating in such gains is accordingly self-defeating.
Consider, therefore, the second alternative in which, though worker participation in realized cost savings is assumed to be normal, workers are required to submit lump sum bids for jobs at the outset. Assuming that large numbers of applicants are qualified to bid for these jobs at the outset, will such a scheme permit employers to fully appropriate the expected, discounted value of future cost savings by awarding the job to whichever worker offers to make the highest lump sum payment?
Such a contracting scheme amounts to long-term contracting in which many of the details of the agreement are left unspecified. As might be anti- cipated, numerous problems are posed. For one thing, it assumes that workers are capable of assessing complex future circumstances in a sophisticated way and making a determination of what the prospective gains are.
Plainly, a serious bounded rationality issue is raised. Second, even if workers had the competence to complete such an exercise, it is seriously to be doubted that they could raise the funds, if their personal assets were deficient, to make the implied full valuation bids. As Malmgren has observed, in a somewhat different but nevertheless related context: ” . . . some [individuals] will see opportunities, but be unable to communicate their own information and expectations favorably to bankers, and thus be unable to acquire finance, or need to pay a higher charge for the capital borrowed” (1961, p. 416). The communication difficulties referred to are due to language limitations (attributable to bounded rationality) that the parties experience. That bankers are unwilling to accept the representations of loan-seekers at face value is because of the risks of opportunism.
Third, and crucially, the magnitude of the estimated future gains to be realized by workers often depends not merely on exogenous events and/or activities that each worker fully controls but also on the posture of coworkers and the posture of the employer. Problems with coworkers arise if, despite steady state task separability, the consent or active cooperation of workers who interface with the task in question must be secured each time an adap- tation is proposed. This effectively means that related sets of workers must enter bids as teams, which complicates the bidding scheme greatly and offers opportunities for free riding. Problems also arise if gains cannot be realized independently of the decisions taken by management with respect, for example, to the organization of production, complementary new asset acquisitions, equipment repair policy, and so forth. Lump sum bidding is plainly hazardous where workers are entering bids on life cycle earnings streams that are repeatedly exposed to rebargaining.36
Finally, but surely of negligible importance in relation to the issues already raised, there is the question of efficient risk-bearing: which party is best situated to bear the risks of future uncertainties —the individual workers or the firm? That individual workers may be poorly suited to bear such risks and, as a group, can pool risks only with difficulty seems evident and further argues against the bidding scheme proposed.
Transactional difficulties thus beset both contingent claims and sequential spot market contracting for the idiosyncratic tasks of interest in this chapter. Consider, therefore, the so-called authority relation as the solution to the contracting problems in question.
3. The Authority Relation
Simon has made one of the few attempts to formally assess the employ-ment relation. Letting B designate the employer (or boss), W be the employee (or worker), and x be an element in the set of possible behavior patterns of W. he defines an authority relation as follows (1957, p. 184):
We will say that B exercises authority over W if W permits B to select x. That is. W accepts authority when his behavior is determined by B‘s decision. In general. W will accept authority only if x0 the x chosen by B. is restricted to some subset (W’s “area of acceptance”) of all the possible values.
An employment contract is then said to exist whenever W agrees to accept the authority of B in return for which B agrees to pay W a stated wage (1957 p. 184).
Simon then asks when will such an employment relationship be pre- ferred to a sales contract, and offers the following two conjectures (1957. p. 185):
- W will be willing to enter into an employment contract with B only if it does not matter to him “very much” which x (within the agreed upon area of acceptance) B will choose, or if W is compensated in some way for the possibility that B will choose .v that is not desired by W (i.e.. that B will ask to perform an unpleasant task).
- It will be advantageous to B to offer W added compensation for entering into an employ ment contract if B is unable to predict with certainty, at the time the contract is made, which x will be the optimum one. from his standpoint. That is. B will pay for the privilege of postponing, until some time after the contract is made, the selection of x.
He then goes on to develop a formal model in which he demonstrates that the employment contract commonly has attractive properties, under conditions of uncertainty, provided that the alternatives are (1) the promise of a particular x in exchange for a given wage w (what he considers to be the sales contract option), or (2) a set of X from which a particular x will subsequently be chosen in exchange for a given wage w’ (the employment contract option).
Put differently, the deterministic sales contract is shown to be inferior to an incompletely specified employment relation in which W and B do not agree on all terms ex ante, but “agree to agree later” – or better, “agree to tell and be told. But plainly the terms are rigged from the outset. As noted previously, the particular type of sales contract to which Simon refers in attempting to establish the rationale for an authority relation is the only one of the three types of sales contracts described at the beginning of this section that lacks adaptability in response to changing market circumstances. Since employment contracts of both the contingent claims and sequential spot marketing kinds are not similarly flawed, a better test of the authority relation would be to compare it with either of these instead.
Simon’s modeling apparatus, unfortunately, does not lend itself to such purposes. It is simply silent with respect to the efficiency properties of alter- native contracts in which adaptability is featured. Not only is it unable to discriminate between the authority relation, contingent claims contract, and spot market contracting in adaptability respects, but Simon’s model fails to raise transaction cost issues of the types described here.
This is not, however, to suggest that the authority relation has nothing to commend it. To the contrary, such a relation does not require that the complex decision tree be generated in advance, and thus does not pose the severe bounded rationality problems that the contingent claims contracting model is subject to. The authority relation also, presumably, reduces the frequency with which contracts must be negotiated in comparison with the sequential spot contracting mode. Adaptations in the small can be costlessly accomplished under an authority relation because such changes, to the worker, do not matter very much.
Assuming, however, that the parties are prospectively joined in a long-term association and the jobs in question are of the idiosyncratic kind, most of the problems of sequential spot contracting still need to be faced. Thus, how are wage and related terms of employment to be adjusted through time in response to either small, but cumulative, or large, discrete changes in the data? What happens when hitherto unforeseen and unforeseeable con tingencies eventuate? How are differences between parties regarding state of the world determinations, the definition of the task, and job performance to be reconciled? Substantially all of the problems that are posed by idiosyn- cratic tasks in the sequential spot contracting mode appear, I submit, under the authority relation as well.
Source: Williamson Oliver E. (1975), Markets and hierarchies: Analysis and antitrust implications, A Study in the Economics of Internal Organization, The Free Press.