Labour contracts

Don Bellante

In the United States and in much of Western Europe, only a minority of private sector workers are covered by collectively bargained contracts. Even in the unionized sector where formal explicit contracts exist, the contracts are incomplete, in that they do not specify all of the terms of the employment relationship.

In one sense, the explanation as to why labour contracts tend to be implicit seems obvious: if they are made explicit, they are generally enforceable only on the employer, with employees free to abandon the employment contract at will. This lack of symmetry markedly distinguishes the explicit labour con­tract from the commercial contract, and employers thus would seem to have no incentive to offer explicit contracts. This explanation is not entirely satis­factory, however, as, even under existing legal conventions, employers might well have an incentive to offer explicit contracts even though such contracts would not bind employees. As long as workers value the security that a formal contract would offer, there are gains to be made by the employer in the form of a lower pay level that would be necessary in order to attract the requisite quantity and quality of labour, in accordance with the theory of compensating wage differentials (Rosen, 1986). Moreover, it is likely that workers would place a very high value on such contractual guarantees. The offering of such contracts would be consistent with profit maximization (and the Pareto-efficiency criterion) as long as the perceived additional costs of formal contracting imposed on the employer are less than the benefits.

The very concept of implicit contracts, though, originated in the literature of macroeconomics in attempts to explain the downward inflexibility of wages in the presence of a cyclically induced excess supply of labour (Gordon, 1974; Azariadis, 1975). The argument is that, in most settings, an implicit contract evolves between employer and employee with the employer promis­ing not to take advantage of temporary declines in the demand for labour by lowering wages, as might occur in a spot or auction market. By making this implicit promise, the employer gains an assurance that, during temporary shortages of labour, employees are unlikely to quit to take advantage of a higher spot price of labour, as the employer prepared to offer the spot price will not be prepared to continue it once the temporary high demand for labour is gone. Since the greater stability of wages owing to this implicit promise comes at the expense (to employees) of great instability of employ­ment, the question remains as to why this particular tradeoff necessarily enhances worker welfare. The answer partly involves the asymmetry of infor­mation between employer and employee. If the employer offered employment stability instead of wage stability, the employer would have an incentive to ‘cheat’ by acting as though the demand for labour were lower than its actual level. There is no comparable incentive to cheat if wage stability is offered: given the stable wage, the employer has no incentive to hire any less than the quantity of labour that the current level of demand renders profit maximizing.

Whether or not the rationale for an implicit promise of wage stability over employment stability is convincing, the rationale still gives no answer as to why an implicit rather than explicit contract is the dominant form. One answer lies in the problem of specification of a formal contract in the pres­ence of bounded rationality (Simon, 1951). That is, the employment relationship is long term, and it is impossible for participants in the relation­ship to foresee all or even most of the contingent circumstances that may arise over the course of the relationship. Nor is it possible to specify in advance what will be the optimal or even feasible solutions to all contingen­cies that can be foreseen. Hence it is not possible to specify completely all terms covering all possible contingencies in the contract - any labour con­tract, whether explicit or implicit, is necessarily incomplete. However, the inability to specify complete contracts puts the explicit contract at a distinct disadvantage. If, for example, lifetime employment were assured provided that certain objective conditions were met by the employee, the firm would have no legally defensible grounds to dismiss the employee if all of those objective conditions were met. The firm would be unable to protect itself against unforeseen contingencies, beyond the control of either firm or em­ployee, that made the assurance of lifetime employment no longer viable.

The greatest impediment that contract incompleteness presents to the for­mation of explicit contracts, however, concerns those circumstances that are within the control of the employee, and which present incentives for employ­ees to engage in post-contractual opportunism.

The employee’s incentive to engage in post-contractual opportunism necessitates the use of resources by the firm to ensure that employee behav­iour is in line with the pre-contractual expectations of the firm. The expenses of closely monitoring employees’ activities should thus be seen as the cost of an inability to specify the labour contract completely. There are, however, substitutes for monitoring that, in given circumstances, are more efficient. One is the payment of what are called ‘efficiency wages’ (Shapiro and Stiglitz, 1984). As with the concept of implicit contracts, the efficiency wage concept also originated in attempts to explain wage rigidity and unemployment at the macroeconomic level. The concept’s ability to explain macroeconomic phe­nomena has been questioned, for example by Bellante (1994), but the concept has found useful applications at the level of the firm. The central idea of the concept is that employers may economize on monitoring costs by providing a rent in the form of a higher than market wage, the payment of which reduces shirking to a greater extent than the expenditure of the same amount on monitoring.

Of course, substitution of efficiency wages for monitoring cannot be com­plete, as the fear of loss of the rent is not effective if there is a zero probability that shirking will be detected. Besides reducing the incentive to shirk, the payment of an efficiency wage should reduce labour turnover and its associ­ated costs, a benefit that does not result from monitoring expenditures. More precisely the benefits and costs of both monitoring and efficiency wage payments are respectively subject to decreasing marginal returns and increas­ing marginal costs. Hence there is an optimal division of expenditure on monitoring and efficiency wages, though in some settings a corner solution involving no efficiency wage is optimal. It can be argued, however, that not all payments of efficiency wages are the result of a deliberate, optimizing decision of the firm. The ability of unions to obtain higher rents has been well documented, and is extensively surveyed in Lewis (1986). Despite the inabil­ity of firms to resist incorporation of these rents into explicit collectively bargained contracts, the firm may, ex post, reap the benefits of efficiency wages, even if the benefits do not cover the costs. These results may explain the counterintuitive finding of Freeman and Medoff (1984) that unions raise productivity levels of workers.

Another alternative to monitoring is incentive-based pay. There are severe limits to the use of incentive-based pay, so that its use tends to be fairly narrow. While pay for performance evolved as a partial mitigation of the consequences of the inability to specify labour contracts completely, incen­tive-based pay would reward effort rather than performance, but effort cannot be directly measured. Performance is subject to influences that are beyond the control of the employee; thus employees are subjected to risk in any real- world incentive pay scheme.

The choice of the use of monitoring, efficiency wages or incentive pay (or the optimal mix of the three) obviously depends on conditions peculiar to a given employment relationship, but the conditions described above lead to some recognizable patterns. For example, sales commissions are used far more extensively with outside than with inside sales personnel in light of the relatively lower cost of monitoring the latter. Another pattern is that piece­rate compensation, though rare, tends to be used most widely in agriculture, where discerning individual contributions is not difficult and where, within wide limits, product quality is not significantly affected by the worker re­sponse to the quantity incentive. It is further apparent that, in manufacturing, particularly in small, owner-managed firms, monitoring costs will be com­paratively low. Hence in such firms there is a preponderant use of monitoring and the payment of fixed hourly wages that, at least for unskilled labour, contain little or nothing in the way of efficiency wage premia. However, in large multi-plant manufacturing firms, monitoring is less cost-effective be­cause of the more intense principal-agency problem (for example, the monitors must be monitored). Hence the intensity of the use of efficiency wages can be expected to rise with firm size, and this may explain the often observed (beginning with Lester, 1968) positive relationship between wage levels and firm size. In any event, it is clear that the costs of all three devices for coping with potential shirking come under the general category of agency costs, in that they arise out of the incongruence between the objectives of firms and their employees (Okuno-Fujiwara, 1987).

The necessarily implicit nature of most employment contracts, combined with the lack of enforceability on employees, raises other problems of oppor­tunistic behaviour. In most labour markets involving complex jobs, the firm makes a number of investments in its workers. Chief among these is training, which may be either formal or on-the-job in nature. The more important distinction is whether training is general or specific. If the training is suffi­ciently general in nature, firms will face little incentive to provide it. Hence training in general skills such as accountancy, nursing and so on will rarely be provided by firms. When training would provide firm-specific skills, that is, skills that do not have much value to firms other than the one providing them, a potential ‘hold-up’ problem exists on both sides of the employment relationship. If the employee were to undertake all of the costs of specific training on the basis of a promise of future pay increases that would fully compensate for the cost of the training, the very specificity of the training would leave the employee with no market alternative to ensure that the promise would be kept. In principle, an explicit contract to pay future com­pensation would protect workers from being held up. But, for the reasons mentioned previously, the inability to specify completely all the terms of employment renders explicit contracting generally infeasible. Alternatively, the firm that provides and completely pays for specific training is unable to obtain a commitment from employees to stay long enough to ensure that the investment will be returned. The firm facing an inability to obtain commit­ment could still invest in some provision of specific training, but only up to the point where the return over the expected duration of employment is sufficient to return the investment. As compared to an ideal situation where commitment could be obtained from workers, this situation will produce less than the optimal (that is, value-maximizing) level of investment in specific human capital.

In most respects, this problem of specific human capital is no different from the general asset-specificity problem faced by firms subject to the hold­up problem in dealing with other firms. For example, a railroad and a phosphate mining firm would both be reluctant to pay for the laying of track that would be of no use other than for hauling phosphate from the location of a specific firm’s mine unless some mutual protection against hold-up could be pro­vided. Given bounded rationality, this is a difficult problem, and thus not all potential net value-creating investments will be made. However, the strongest solution to the asset-specificity problem in commercial relationships between firms is not available in employment relationships. The potential hold-up problem between firms arising out of asset specificity can be solved by merger or vertical integration (Joskow, 1985), but, since this solution is not available in employment relationships, other solutions have to be employed. Basically, there are two solutions: reliance on reputation as a contract en­forcement mechanism, and the construction of (usually) implicit contracts that are said to be ‘self-enforcing’.

It can be argued that the effect of loss of reputation can, under proper conditions, provide a sufficiently strong incentive not to renege on the terms of an implicit contract (Klein and Leffler, 1981). It can also be argued that potential loss of reputation provides much more protection against reneging for the employee than it does for the employer. The value of reputation depends on repeated transaction. An employee will transact with only a few employers over a finite working life; a corporation deals with many employ­ees over its potentially perpetual life. Efficiency considerations, supported by game-theoretic analysis, suggest that implicit contracts tilt the risk towards the party with the least to lose through loss of reputation (Kreps et al., 1982; Kreps and Wilson, 1982). Hence, to the extent that reputational concerns affect the division of specific training costs between employer and employee, they will tend towards requiring the employee to bear the cost, most likely in the form of a below-market wage during the period of training. It should be noted, however, that reputation considerations affect, not just payment for specific training, but the entire structure of implicit labour contracts.

If potential loss of reputation offers some protection for employees, it does so imperfectly; moreover, it offers little or no protection for the employer. Primary protection against opportunistic behaviour comes from structuring implicit relational contracts in such a way that they are ‘self-enforcing’. That is, both the employer and employee must face a strong incentive to honour the terms of the implicit contract. Consequently, firms and employees typi­cally share the costs of training, but the firm’s share is in large part deferred. The necessity of self-enforcement is one of the explanations offered by economists for the fact that earnings tend to rise with tenure to a greater extent than can be explained by rising productivity over time. In this explana­tion, employees accept lower pay during the initial phases of employment (during which a high proportion of the employee’s specific training is re­ceived) in exchange for increasingly higher wages in future periods. Pensions that are fully funded but for which ‘vesting’ is deferred are a very common means of deferred compensation that allows firms strongly to induce commit­ment by workers and to ensure a return on the firm’s investment. The United States now in general requires full vesting to take place within five years. This change is seen as nullifying existing implicit contracts and is expected to reduce the willingness of firms to offer training (Bellante and Porter, 1990).

For implicit contracts involving deferred compensation to be self-enforc­ing, several conditions should be met. First, the deferred payments must be sufficiently large in discounted present value terms to match the present value of alternative employment opportunities that do not offer such training, and at the same time sufficiently small to yield at least a competitive return on the firm’s investment in training. The employer thus receives some assurance that employees have a motive not to leave the firm. Moreover, it induces job applicants to reveal their private knowledge of their own ex ante expectations about commitment to staying with the firm (Salop and Salop, 1976). The fact that the training provided will cause productivity to rise over time (though not as rapidly as wages) gives the employer a somewhat reduced incentive to renege by firing employees in later stages of employment, in addition to whatever effect potential reputation loss provides. Moreover, replacing older workers with new employees would necessitate additional investment in train­ing and other start-up costs of employment - costs that are sunk, so far as current employees are concerned. Further protection against opportunistic behaviour on the part of employers is provided by explicit seniority policies. And, by making employees more willing to undertake training and accept the associated deferment of wages, such policies also serve the purpose of firms by reducing the lifetime wage premium that must be paid to such workers.

A second condition that must be met is that employees will not ‘over collect’ from the firm by staying with it too long. Mandatory retirement policies can be seen as meeting this condition (Lazear, 1979). The desires of national governments in recent years to prohibit mandatory retirement have curtailed the use of this device. Alternatively, pension plans can be structured to induce retirement by providing a schedule of benefits whose expected present value begins to decline at the age for which the firm desires retire­ment to take place. Lazear (1983) has provided evidence that a large fraction of firms paying pensions do in fact structure them in this manner.

Despite all that can be done to make implicit contracts self-enforcing, market conditions may change so as to make their abrogation seem neces­sary, perhaps even a prerequisite for the firm’s survival. However, the costs to the firm of abrogation are obvious and likely to be severe. A question worthy of further investigation is whether and to what extent mergers and spin-offs are in some instances devices for minimizing the costs of abrogation through unilateral ‘rewriting’ of the implicit contract.

References

Azariadis, C. (1975), ‘Implicit contracts and underemployment equilibria’, Journal of Political Economy, 83, 1183-202.

Bellante, D. (1994), ‘Sticky wages, efficiency wages and market processes’, Review of Austrian Economics, 8, 21-33.

Bellante, D. and P. Porter (1990), ‘A subjectivist economic analysis of government-mandated employee benefits’, Harvard Journal of Law and Public Policy, 13, 657-87.

Frank, J. (1986), The New Keynesian Economics, New York: St. Martin’s Press. Freeman, R. and J. Medoff (1984), What Do Unions Do?, New York: Basic Books.

Gordon, D. (1974), ‘A neo-classical theory of Keynesian unemployment’, Economic Inquiry, 12, 431-59.

Joskow, P. (1985), ‘Vertical integration and long term contracts: the case of co al-burning electric generating plants’, Journal of Law, Economics and Organization, 1, 33-79.

Klein, B. and K. Leffler (1981), ‘The role of market forces in assuring contractual perform­ance’, Journal of Political Economy, 89, 615-41.

Kreps, D., P. Milgrom, J. Roberts and R. Wilson (1982), ‘Rational cooperation in the finitely repeated prisoner’s dilemma’, Journal of Economic Theory, 27, 245-52.

Kreps, D. and R. Wilson (1982), ‘Reputation and imperfect information’, Journal of Economic Theory, 27, 253-79.

Lazear, E. (1979), ‘Why is there mandatory retirement?’, Journal of Political Economy, 87, 1261-84.

Lazear, E. (1983), ‘Pensions as severance pay’, in Z. Bodie and J. Shoven (eds), Financial Aspects of the United States Pension System, Chicago: University of Chicago Press, pp. 57­85.

Lester, R.A. (1968), ‘Pay differentials by size of establishment’, Industrial Relations, 7, 57-67. Lewis, H.G. (1986), Union Relative Wage Effects: A Survey, Chicago: University of Chicago Press.

Milgrom, P. and J. Roberts (1992), Economics, Organization and Management, Englewood Cliffs, NJ: Prentice-Hall.

Okuno-Fujiwara, M. (1987), ‘Monitoring cost, agency relationships and equilibrium modes of labor contracts’, Journal of Japanese and International Economies, 1, 147-67.

Rosen, S. (1986), ‘The theory of equalizing differences’, in O. Ashenfelter and R. Layard (eds), Handbook of Labour Economics, New York: North-Holland, pp. 641-92.

Salop, J. and S. Salop (1976), ‘Self-selection and turnover in the labour market’, Quarterly Journal of Economics, 90, 619-28.

Shapiro, C. and J.E. Stiglitz (1984), ‘Equilibrium unemployment as a worker discipline device’, American Economic Review, 74, 433-44.

Simon, H.A. (1951), ‘A formal theory of the employment relationship’, Econometrica, 19, 293­305.

19