1. The Internal Labour Market
A description of the major ‘contracting modes’ (the spot contract, the state-contingent contract and the authority relation) has already been presented at the end of Chapter 2, along with a brief outline of Williamson’s criticism of these modes. Bounded rationality and opportunistic behaviour deriving from task idiosyncrasy, he argues, imply that alternative contractual arrangements have advantages over each of these forms of contract. As has been recognised at many points in this book, contracting in circumstances of asymmetrically distributed information is facilitated by ‘reputation’ and ‘trust’, but it is fundamental to our approach to the firm that instead of assuming the existence of this ‘reputation’ and ‘trust’, institutional mechanisms are seen as the means by which confidence in the integrity of others can be encouraged.
In Williamson’s view, the internal labour market represents an institutional response to the problem of opportunism, and elicits more cooperative behaviour from employees than would be possible under spot or state-contingent contracting. As seen in Chapter 2, Williamson (1985) sees ‘unified governance’ as becoming established under conditions of uncertainty, transaction-specificity and high frequency of repeat dealing. More recently, Williamson (2000) succinctly distinguishes the transactions cost and principal-agent approaches to contract. ‘Moving beyond the agency theory tradition of ex ante incentive alignment, transaction cost economics turns its attention – additionally and predominantly – to the ex post stage of contract’ (p. 599). In other words, the theory which was reviewed in sections 2 to 6 above concerned the problem of selecting the right transactional terms ex ante (the best payment schedule or the best distribution of prizes in a tournament) – before the outcomes are known. Even when the passage of time was explicitly recognised – as in the sections on seniority payments and bond-posting – implicit contracts were simply seen as being put into effect supported by the forces of reputation. Transactions cost theory asks what happens at the post-contractual stage. As information is revealed how will the transactors behave? Will they try to renegotiate the contract? How disruptive and costly will this process be? What mechanisms will be used to reduce the costs of ex post adjustment? Notice here that the Grossman-Hart-Moore property rights theory of the firm introduced in Chapter 4, while emphasising ex post opportunism as a problem, is not a transactions cost approach in the spirit of Coase or Williamson. This is because ex post hold-up is accurately predicted by the transactors ex ante while ex post bargaining costs are assumed to be zero. Transactions cost economics, in contrast, is largely about the costs of ex post bargaining.
2. Specific Human Capital
Consider yet again the problem faced by the firm. It requires a person to perform some task. This task involves skills which are not general and cannot be used to aid performance in other lines of activity. They are specific skills (to use Becker’s (1964) terminology). They are useful in the highly particular context of the firm, and some may be acquired only by on-the-job training, as in the case of ‘knowledge of time and place’ mentioned in section 7. Human capital theory concludes that the costs of general training, which increases the productivity of people in lines of activity both inside and outside the firm, will be borne by the individual concerned, whereas the costs of specific training will be borne by the firm. This argument depends on the idea that the firm will not pay for the general training of employees because they may then be enticed away to higher-paid jobs elsewhere and the firm will lose its investment. The firm cannot trust the employee. On the other hand, the employee will not pay the cost of specific training because this implies accepting a lower reward than could be obtained elsewhere during the training period. Higher rewards would be forthcoming after the training period, but this implies that the employee must trust the firm, and the traditional argument of human capital theory assumes this will not happen. Instead the firm pays for the specific training.
As we saw in section 7, however, a firm which pays for specific training might still itself be the victim of opportunism. An employee, by threatening to leave, is capable of inflicting large costs on the firm which would face the prospect of having to invest more resources in the training of another outsider. In the context of full employment the threat would be costless and hence credible (see section 5). Thus, the conventional distinction between types of training, general and specific, and the conclusion that these are financed by the individual and the firm respectively, depends upon the assumption that employees do not trust firms but that in the realm of specific training firms have to trust employees.29 Note that the opportunism of the employee is here a form of rent seeking and that the implicit challenge is to the employer’s property rights in the ‘specific human capital’ he or she has financed. Just as with property rights in information, property rights in human capital30 are difficult to establish and police for obvious reasons, and are therefore challengeable.
Unlike the conclusions of human capital theory, the upshot of Williamson’s approach is a greater emphasis on the employee trusting the firm. The primary characteristic of the ‘internal labour market’ is that employees do not bargain individually over terms and conditions with the employer. Instead, a person joining a firm will be assigned to a certain grade in the hierarchy and will receive whatever remuneration attaches to that grade: ‘The internal labour market achieves a fundamental transformation by shifting to a system where wage rates attach mainly to jobs rather than to workers’(Williamson et al., 1975, p. 270). By ‘job’, here, the authors are not referring to a single idiosyncratic task but to a broad category of tasks attached to which are standard terms of employment. Because individual bargaining is ruled out, ‘the incentives to behave opportunistically …are correspondingly attenuated’ (p. 271). The employee voluntarily accepts these constraints on his or her freedom of action in the knowledge that others will be similarly constrained. Of course, while this procedure mitigates the problem of rent seeking through attempts to mislead the employer in an unending bargaining process, it does not in itself prevent shirking. Williamson argues, however, that the promotion ladder is designed to counter this form of rent seeking and to encourage ‘consummate’ instead of ‘perfunctory’ cooperation. ‘Consummate cooperation is an affirmative job attitude – to include the use of judgement, filling gaps, and taking initiative in an instrumental way. Perfunctory cooperation, by contrast, involves job performance of a minimally acceptable sort’ (Williamson et al. ,1975, p. 266). Thus Williamson’s approach is very closely related to the ideas of Lazear and others (sections 5 and 6) on the incentive effects of a rising profile of earnings over time. The additional element is an explanation of the hierarchical characteristic that each individual is not treated as a special case but, at any one time, is the holder of a ‘standard’ contract applying to a particular grade. From this perspective, it no longer follows therefore that firms will finance specific training. If the worker accepts a lower wage at first, which reflects his or her lower productivity as s/he gradually becomes acquainted with the idiosyncrasies of the task, s/he effectively ‘posts a bond’, as discussed at length earlier. In so doing s/he is forced to trust the firm. Why, though, should we expect the employee to trust the firm rather than the other way around?
For any exchange transaction to take place, even of the most basic ‘bread for beer’ variety, someone has to trust someone else. At the end of section 5 it was shown that the firm’s ‘reputation’ would suffer in the long run if it cheated on its implicit agreements. A strand in the argument was missing at that stage, however. In principle, the same ‘reputation-protecting’ argu- ment could be applied to employees to show that they might lose in the long run if the firm trusted them and they cheated. Clearly ‘reputations’ in the labour market may be important, and we shall see later that some writers have used this possibility as a limiting factor to managerial cheating of shareholders (Fama, 1980). However, it is usual to argue that, for most types of employee, ‘reputations’ are difficult; that is, costly to create. As Klein (1984) puts it, A firm generally has lower costs of creating brand name capital and hence contract fulfilment credibility because of its increased repeat purchase frequency’ (p. 333). In other words, a single firm’s integrity is tested each period in its dealings with each of its employees. Because it is a ‘central contractual agent’ it is party to many contracts, whereas each employee is party to only one. Further, the working lifetime of a joint-stock firm is not limited in the same way as the working lifetime of employees (although bankruptcy can ‘terminate’ a firm’s life). Thus, ‘cheating firms are likely to become known more quickly than cheating workers, reducing the short-run cheating potential of firms relative to workers’ (p. 333). A single employee may only serve a few customers (employers) in a lifetime, and although a good ‘track record’ of reliability is not worthless, the difficulties of communicating such a record to potential employers (the source of the adverse selection problem) imply high costs of creating this type of ‘brand name capital’ as Klein calls it. Contrary to conventional human capital theory, therefore, ‘the worker can be expected to make much of the specific investment and the firm guarantee that it will not hold up the worker by reducing his wage below the value of his marginal product’ (Klein, 1984, p. 333).
We should not necessarily presume that the worker will implicitly finance all the specific human investment in the firm. Some sharing of the initial burden might occur, especially if the investment is so substantial that the temptation to the firm to renege is a serious one. Note that although both the ‘bond-posting’ and ‘specific-human-capital’ approaches to wages suggest that they will be observed to rise with length of tenure, the two theories differ in important respects. In the bond-posting or deferred compensation scheme, wages rise faster than the productivity of the worker – starting out lower and finishing higher than actual productivity. Where worker and firm share in the finance of specific human capital, wages will rise less fast than productivity – starting out somewhat higher (though below the ‘market wage’) and finishing lower (as the firm takes a return on its share in the human capital formation). Further, because the wage finishes below productivity according to the human capital view, it is not so necessary to insist on retirement at a particular date. Thus, whether a person is seen as receiving ‘enforcement rents’ or ‘rents on specific human capital’ makes a difference to the time path of wages.31
3. Behavioural Norms and Perceptions of Fairness
Williamson (1975, pp. 37-8) argues that ‘atmosphere’ and ‘supplying a satisfying exchange relation’ are part of the economic problem. The governance structure of the firm therefore aims not only to develop an environment of trust but also to generate a desire to act cooperatively in the interests of the team effort. There are close parallels here with the ‘human relations school’ of organisation theory mentioned briefly in section 1. The British system of relying on voluntary blood donors is a much-discussed example of the importance of establishing emotionally satisfying arrangements. A market in blood could be established, but would affect the individual’s perception of the act of supplying blood. Titmus (1970) argues that the gift relationship is not only a worthy ideal in itself, but may also produce in certain circumstances a more effective system. The argument is of far wider significance than the blood example alone suggests. Akerlof (1982, 1984) applies the idea of ‘gift exchange’ to the whole area of labour contracts. Employers pay workers more than market-clearing wages (a gift) and in return they hope for a gift of greater loyalty and effort than would otherwise be forthcoming. He cites evidence from social psychology (Adams, 1965) indicating that people regarding themselves as ‘overpaid’ are more productive than those merely receiving ‘the rate for the job’. The approach is thus a variety of the ‘efficiency-wage’ theory but relies on the existence of behavioural norms which require reciprocity.
Many analysts of labour markets have argued that observed behaviour is difficult to explain without reference to some notions of ‘fairness’ and implicit mechanisms of the ‘gift exchange’ kind. Schlicht (1992) uses the idea of ‘wage generosity’ and gift exchange to explain the fact that collectively set wages at the industry level in Germany are usually well below those actually paid by firms. Yet, bargaining over the collective norms is often tough and changes do feed through to actual wages paid. No firm likes to be seen as falling short of established standards and most prefer to be generous. Solow (1990) emphasises the development of norms of fairness in explaining the functioning of labour markets. As in the exchange game set out in Chapter 1, and the Hawk-Dove game of Chapter 4, norms of behaviour can evolve in a repeated game setting. These norms may prevent unemployed workers from trying to undercut the employed, or firms from trying to grab an unacceptably large share of the rents available. Solow argues (p. 23) that, ‘norms of behaviour can be modelled as constraints on decisions’, but, as Alchian notes in the quote at the head of Chapter 4, constraints on decisions can be modelled as property rights. Governance structures, by contributing to the establishment of prevailing norms of behaviour, are implicitly supporting the property claims of workers to income flows that are otherwise vulnerable to opportunism.
4. Company Unions and Disputes Procedures
Behavioural norms cannot cover every set of circumstances. The internal labour market reduces the return to opportunism and encourages cooperation by replacing individual bargaining with a set of standard contracts and the promise of promotion. These standard contracts, concerning as they do the relationship of employer and employee over a substantial period of time, cannot be comprehensive in their provisions because of the problem of bounded rationality. As time advances, therefore, disputes will arise concerning the precise interpretation of the vague terminology of the standard contract. In my own case, for example, a contract requires that as a professor of economics I ‘perform the duties which are customary for such posts in universities in Great Britain and shall include teaching, research, examining, and if so directed administration’. Whether a daily chore of making the tea for the vice-chancellor could be considered as ‘administration’ or ‘research’ would have to be settled by arbitration procedures. The establishment of these arbitration and grievance procedures is thus an integral part of the internal labour market, and is closely associated with the role of trade unions or other associations of employees. Trade unions can be seen as monitoring the firm’s commitment to its implicit obligations (Malcolmson, 1982), making ‘shirking’ by the firm less likely and increasing the confidence of employees in the working of the internal labour market.
Source: Ricketts Martin (2002), The Economics of Business Enterprise: An Introduction to Economic Organisation and the Theory of the Firm, Edward Elgar Pub; 3rd edition.