The Common Law and the Labour Market

Richard A. Epstein

I. No to the Status Quo

In examining the relationship between the common law and the labour market, we cannot accord any presumption of legitimacy to the current status quo. That status quo has been created by a set of elaborate compromises, which have been fuelled by political deals, some savoury and others not. The present law may embody the correct system, but, if it does, that outcome has not been legitimated by a long and harmonious tradition of voluntary interaction. It is not the result of a customary or a spontaneous order.

By the same token, I think it would be improper for me to make my life easy and comfortable as I would sometimes like it to be, and to assume that we ought to dismiss the current structure out of hand because of its rather dubious pedigree. Some victories are so easy that they become hollow. All I ask therefore is that we look at the question of labour relations from a neutral position, without any presumptions one way or the other. Without using any presumption we can then see how a system of labour relations emerges by applying a more general legal theory. How do we make a prescriptive analysis that follows in consistent fashion from a defensible set of first principles?

II. Common Law Building Blocks

Generally speaking there are three branches of substantive law that have a lot to do with the organisation of the labour market. I shall talk about one of these in great detail, and will necessarily touch upon the other two. The first branch of law is the law of property; the second is the law of contract, which I think everyone will understand is very much a part of the labour market; and the third is the law of tort, that fancy-sounding word that deals with the situations in which the law does, or should, give redress when one person inflicts a harm, either physical or economic, upon another. Tort is that branch of the law that protects against all forms of aggression, from the overt use of force to the more subtle forms of intimidation or threat.

Within this system, the common law rules of property form the baseline for all subsequent analysis of the law of contract. The theory of contract allows us to trade only those things which we have already acquired, either by original acquisition or voluntary transaction. Where one deals with entitlements acquired through voluntary transaction, there is the problem of infinite regress. One transfer may rest on the one that precedes it. But there has to be a first title that is not dependent upon prior contract or grant. The key question is how is that initial entitlement obtained, and why is it respected?

The basic common law contains one key, but unreflective assumption, that is more astute than may first appear. It is the assumption of individual self-ownership all labour, talents and abilities for each person. This rule has a powerful degree of formal equality built into it: everybody has certain talents and endowments, and these will differ from person to person. The common law rule of natural ownership of labour thus creates of necessity a decentralized market, because there is no single person, no group of individuals, who can decide who is going to own what, or to decide who is entitled to deal with whom on the transfer and use of labour or talents.

With the original position of self-ownership thus established, the office of the law of contract is to facilitate trades of labour, or material things acquired through first possession. The common law rules on transfer do not treat labour relations as something distinctive or special. When I studied law initially at Oxford, there was no separate branch of the common law governing labour relationships. Instead you first applied the property rules of self-ownership and first possession; next you applied the general rules of contract with respect to the endowments so acquired, and afterwards you applied the tort laws to see that no impermissible actions of aggression took place. The body of labour law simply fell out of these general considerations, not as a special edict or legislative command. The common law achieved a very high level of coherence and universality with its very few generative principles of rights and duties.

III. Abstractions and Persons

In analysing contractual validity, then, the legal system did not speak about one set of rules for employers and another for employees. It did not create one set of rules for people who were rich and powerful and another for those who were frail or meek. instead we spoke about those two hardy standbys in all contractual arrangements: A and B. These people were colourless, odourless and tasteless. They were self-conscious abstractions that were known to be false as descriptions of the world, and useful precisely because of the extent of their detachment. A and B (and their friends C to Z) moved on high and only occasionally made contact with the hurly-burly world of commerce. The common law theory was that so long as you knew there was an A and a B, you had all the information necessary to figure out their self-definition by contract. It did not matter to the law (although it surely did to the parties) that one side to the agreement was the employer and the other was the employee. These relationships, these categories were not imposed upon private persons through the operation of the law; these roles were freely chosen by individuals acting within the system. A would decide to work for B, and B to work for A, and each to assume other roles in relation to other parties. There was, and is, no inconsistency in A working for B in one transaction and hiring C in a second.

This massive oversimplification of the social universe treats all persons as though they are as similar as the letters of the alphabet, and thus ignores or rejects every effort for the common law to take into account the difference between an individual worker of limited means and a huge industrial corporation. The major attack on the common law system was its common law refusal to take into account differences in role and status that were, and are, apparent to all who care to look at the operation of the legal system. The reformers want to dismiss the common law approach as a pitiful caricature and replace the slippery generalities of contract, property and tort with a far more reticulated set of rules that takes into account known and particular differences in roles and context. The attack on the common law seeks to make its rules more sensitive to the dominant features of social relationships.

Now why should any analyst try to blink at the evident social truths and speak only of the abstract As and Bs when specific information about context and role is indispensable for understanding the way in which individuals behave and markets operate? To answer this question, it is necessary to step back from matters of detail in order to focus again on points of general principle. It is true that all individuals differ in natural endowments, in social role and institutional expectations. But there is one thing that each of them want, and want just as As and as Bs. And that is more. We may not be able from a disinterested perch to answer the question "more of what?" or explain why more of this is preferred to more of that. Indeed there are many occasions where people themselves are unable to articulate why they want more of one commodity or goods than they do of another. But regardless of these gaps in knowledge we do know that people who have more can always end up with less by giving away some of what they have. But people with less can only end up with more by one of two routes: force or trade.

This characteristic of wanting more is something of a universal. It applies with equal force to greedy and rapacious firms as it does to self-interested individuals. Indeed, wanting more is not a characteristic for which we should want to condemn people. The desire for more is one of the few features that is indispensable for human progress and advance. The right question to ask is not why we desire more? It is how are we prepared to go about getting the more that we all want.

There are only two kinds of games (two patterns of behaviour) that people can play in a world in which everyone desires more and fears less. The first game is to get more for yourself by taking it from someone else. The second is by entering into a voluntary exchange with another person. All ostensible third cases (e.g. misrepresentation) are essential variations on either coercion or consent.

Faced with these general truths, the argument for thinking about legal relationships in terms of the lifeless As and Bs runs as follows: Each A and B has certain initial endowments: these are then augmented by things which are acquired through first possession or voluntary exchange. When entering therefore into voluntary arrangements, this constant pressure to end up with more instead of less implies that two inequalities are simultaneously satisfied. The first inequality says that when A surrenders any thing (be it labour or material goods) he will value it less than the thing that he receives in exchange. The second inequality says that when B surrenders any thing she will value it less than she receives in exchange.

IV. Positive and Negative Sum Games

Taken together, these inequalities have enormous social significance. They mean that voluntary exchanges are positive sum games for the participants. That proposition holds regardless of how A or B received their initial endowments. And it holds even if we as analysts have no strong theory of value which explains why A wants to order his preferences in one direction and B wants to order hers in another. So long as we know that all As and all Bs desire more, there is sufficient warrant for allowing voluntary exchange to take place. Both sides are better off than they were before. Theft for its part is at best a constant sum game, for what one party gains the other necessarily loses. In truth it is rarely that, given the enormous resources that have to be spent to work the theft, given the resistance and opposition to it that is sure to follow.

This general proposition---all voluntary exchange yields positive sum game for the parties---is an enormously fertile idea. Once it works, it should be used over and over again. If it works to improve allocations when A and B each have only their initial talents and endowments, it will work when each has acquired some endowments by labour or by trade. So long as trade works well, it should he encouraged over and over and over again. One good idea is all that we need. A can take the endowments that he has received from B and trade all or part of them with C. He can then take some portions of those endowments, mix them with those he has, and trade them with D. There is no natural end point or final period to this game, for the outputs of one voluntary transaction can easily become the inputs of the next. As the game continues to run its course the transactions may become very complex, but the legal rules that generate these transactions remain as simple as they were at the outset. There is no need to speak of common law rules that ensure the primacy of the market. There is no reason for the law of contracts to talk of markets at all. The series of voluntary exchanges that emerge spontaneously (as Hayek would say) yield a social marketplace rich and variegated by the repeated operation of a single legal rule.

Thus far I have emphasized how simple and impersonal legal rules are sufficient to define and organize markets. There is, however, a sense in which it does matter who A and B are after all, even in a world of voluntary exchange. If it turns out that there is only one unique A sitting there with a monopoly type position, while there are lots of Bs who are pretty close substitutes for each other, then A is going to be in an exceedingly strong position with respect to any particular B, because that B can't go anywhere else to find a substitute A, while A can pick and choose among a large array of more or less identical Bs.

So given this initial distribution of roles, it is instructive to try to figure out how the pattern of exchanges will evolve. In order to do that it is necessary to know the threat position of each party, what that party can do to better his lot on his own when there is no bargain to be made with some other person. That threat position will be exceedingly strong for A, who can choose from an array of Bs. But it will be weak for each B, who has to deal with A if there is to be any agreement at all. We can be pretty confident that in a world of one A and many Bs, A will end up with a very large proportion of the gain that comes out of any A-B agreement. The strength of a common law system, and the weakness of centralized systems of regulations, relates directly to the ratio between the As and the Bs in any market. By having its decentralized system of original ownership rights, the common law tends to guarantee, both by natural selection and migration, that there will be lots of As and lots of Bs in any organized market. The chief virtue of the common law is that it allows no one to say "no one can enter this side of a market, except for me". Free entry in all possible market rules is a necessary corollary of the basic common law rules of self-ownership and voluntary exchange.

V. The Many and the One

The institutional implications are both clear and striking. No government, or group of individuals with special privileges or franchises, will be able to preclude entry on any side of the market. With this condition in place the uniqueness of any A or B vanishes. There can be lots of persons who can undertake any particular social role. Accordingly all persons have a reasonably good threat position vis-à-vis the rest of the world. If you do not wish to do business with me, there are lots of other people in the world to whom you (and I) can turn.

It follows that the A and B model is extraordinarily robust if two assumptions are satisfied:

A. Multiple players, and

B. Mutually advantageous exchanges.

By starting from this very modest material, the law can facilitate (not compel, but facilitate) sizeable productive interactions which will continue to expand over time and transactions until it embraces all individuals who have the minimum capacity to engage in contracting at all. The system goes forward in a benevolent fashion because we, as lawyers or perhaps more as economists, know that the exchanges are mutually beneficial: otherwise the parties would not choose to enter into them. We don't have to know why any given exchange will be beneficial. It is quite enough to know that, given the manifestation of consent, this will be the case. The legal system can tolerate exchanges that it does not fully understand. indeed it must do so if it is to function effectively at all.

Nonetheless by stepping out of our sceptical mode, it is possible to say something about the forms of voluntary exchange that should emerge over time, namely that there will be sorting by role, specialization in the market. In order to defend that conclusion we need one other additional assumption which is as critical as it is accurate. We must assume that there is some level of natural variation across people in a wide range of different dimensions. The obvious ones will be such things as height and weight, but others, which are less obvious will be critical as well, of which attitude and taste are perhaps the most important. We may not know exactly how these are distributed, but we can say with perfect confidence that all people are not all concentrated at a single point: people will not have identical heights and weights, and so too with their other natural endowments, characteristics and abilities. The biology and the sociology of the situation make us confident that some people will be better at some things, and others at other things. The dimensions over which these differences exist can be quite subtle and important. It seems clear (given the behaviour of professional gamblers and traders) that some persons are better than others at managing risk, while others may be better at handling details. The first could become traders and the second brokers, both in the same financial marketplace.

Similarly within a contractual arrangement, persons who are better at bearing risk may become owners of the firm and employers, while those who are less able to bear it may be quite content to take the fixed salary and greater security associated with being an employee. Nor is there any reason why employment relationships must be so stark and divided. It is possible to incorporate commissions and bonuses into hiring arrangements that reflect some complex system of risk sharing between the parties. The parties can mix and match to their hearts' content. The patterns of combination and the terms of combination can yield all the variation and subtlety anyone desires even if the law of contract continues to treat A and B as placeholders instead of persons.

As the process runs its course, there is a slow transformation in the way in which the world operates. While we start with isolated and undifferentiated As and Bs, some of the As start coalescing and some of the Bs start coalescing. Certain fundamental asymmetries start-to emerge, and with time one can see the variegated world that we observe today, filled with complex organisations and individual entrepreneurs, much as in evolution we see organisms of different sizes filling the various niches that are available in nature. And throughout it is a process of self-definition, not of external compulsion.

But is there anything to fear in this rapid process of differentiation and (re)combination? So long as each person has the power to refuse one offer, and to go somewhere else, then no person will enter into a bargain that leaves him or her worse off than before. If that statement is correct, then the model of 'exploitation' by contract which has such broad appeal in common discourse is wholly inapposite. Similarly, the kindred notion of 'economic' duress that has long held such great appeal to lawyers cannot be treated simply as an intelligent extension of the idea of physical duress. Physical duress arises when a person forces another to enter into a contract with the threat that he will inflict harm on him if he does not. Any exchange so made will leave that person better off than he was after the threat was made, but that is not the vantage point from which the exchange should be evaluated. Rather the critical point in time for making the comparison is before the threat has been made, and relative to that reference point the party subject to duress has been left worse off than before. Economic duress, if it only refers to hard bargaining under conditions of scarcity, does not lead to an outcome where one party is left worse off after the interaction than before. The ever present existence of scarcity, far from being a reason why we should limit the scope of contractual freedom, is the reason why we need to resort to contractual freedom in the first place. The class of transactions that ought to be set aside on the grounds of economic duress is basically an empty set.

Important implications follow once the distinction between these two forms of duress is observed. In the first place, once the idea of economic duress is placed to one side, then the entire case for mandatory terms in labour contracts becomes wholly misguided. It thus becomes unwise for the state to say that a certain minimum wage should be imposed as a matter of law, or that certain contractual forms, e.g. the contract at will, should be ruled out of bounds, or that certain institutional and organisational arrangements, e.g. mandatory collective bargaining, should be imposed as a matter of course. Arguments of this form all depend upon a showing of exploitation that cannot be made if our two assumptions set out above (multiple actors and mutually advantageous trades) are satisfied.

The problem of physical duress, however, remains more acute, and is representative of the class of grounds, long incorporated into the common law, on which contracts could be properly set aside. Fraud, misrepresentation, coercion, undue influence, infancy and insanity are all issues that fall into this general class. These are contract defences that are consistent with, and indeed necessary to, the operation of free markets as we understand them. Only the opponents of laissez-faire may be bold enough to claim that the principle of freedom of contract is so absolute that these are all excluded. Its proponents recognize that the power of contract lives and dies with the supposition of mutual gain for self-interested parties on which it rests.

VI. From Within the Market

Thus far I have celebrated the power of a few simple ideas to generate powerful social results, and have consistently adopted the view of the social observer to the exclusion of the market participant. But there is another task related to the first which requires that I depart the lofty perch as lawyer cum economist, and assume the role of market participant. In that capacity my question is what contract should I choose to enter into given the wonderful freedom that the law has allowed me. At this point the inquiry turns from contractual freedom to contracting strategy---how is it that individuals who know that they are self-interested, and who know that their trading partners are self-interested as well, formulate their contracts? More concretely, what kind of devices do, and should, contracting parties use to insure that an agreement that is welcome today is going to be an agreement that remains viable over time? And there is reason to fear defection from contractual promises. If Mr A performs first and Ms B ends up with all the dough, only to turn around and flee, then we do not have a contract that creates reciprocal obligations to mutual advantage, but a breach by B that turns out to be nothing more than a fancy species of theft. Any individual, knowing of this risk in advance, will seek to find devices that ensure that contracts designed for mutual gain will not turn into instruments of one-sided advantage and systematic breach. The problem of contract performance remains even when freedom in contract formation is assured.

There is a second obstacle that successful contracting must also confront. The number of possible permutations in any contract between A and B, especially in the employment context, is very large. There are all sorts of sudden and unforeseen turns in the road, and it is often impossible or too costly to specify their contractual solution in advance. It is very expensive to negotiate every contract term by term. What large numbers of people have done therefore is to develop a set of default provisions that are designed to imitate what a typical or model person in a particular role would want under certain sorts of circumstances. In fashioning the right set of default provisions, it is no longer possible to take refuge in the useful abstractions of A and B. Now it is necessary to have (or acquire) a powerful experiential base within a trade or business, or a shrewd sense of what sound economic theory would say that rational contracting parties who have assumed certain characteristic roles (buyer and seller, etc.) would in general desire. Sometimes there are implicit conflicts between what customary practice requires and what economic theory predicts, and it is very difficult to choose between them. But rather than stress the difficult cases of tension, it is better to give a bit of the flavour of how some of these problems are handled.

One major problem in contracting concerns the sequence of performance, the problem to which I referred before. Thus the ordinary labour contract raises the question of whether the work has to be performed before the money is paid, or whether the money has to be paid before the labour is performed. This whole question thus gives rise to that exotic body of law in contract, called the law of "conditions". The promises that people make need not be instantly and immediately enforceable. It may well be that the one promise need be honoured only after the other has been honoured. How then is that order determined? Essentially the right way to analyze this issue is to figure out what sequence of performance will minimize the likelihood of breach by either party, because that solution will, on balance, tend to yield the greatest level of gain from the contract in question. In normal situations I think that a fairly powerful argument can be made in favour of a regime that has labour perform first, with the employer obligated to pay in full only on the completion of the work. (The issue gets very messy if there is part performance that has value to the employer, who nonetheless must cover from some other source.) And it should be clear why we want to take this approach, even though it exposes labour to the serious risk of non performance on the other side.

One possible explanation depends on the theory of exploitation that I attacked earlier. The worker always gets the short end of the deal. But I think that this explanation is mistaken, as there are two more powerful explanations that are consistent with the general theory of contracting for mutual benefit outlined above. Thus one question to ask is, which way is the risk of insolvency the greater? If the employer has sufficient wealth, then that wealth works to the advantage of the worker by providing a fund that makes any judgement obtained in court collectible. On the other hand, if you have an employer who pays a worker who then does not perform, there is a more serious risk that this wealth will be dissipated before any suit could be brought for breach. In addition, the remedy of the worker is easy to determine, it is the wages owing. But the loss to the employer is far more difficult to measure as the value of the work that should have been performed is not a liquidated sum. The remedial system that works far better with employee suits than it does with employer suits.

The second element that must be factored into the equation is reputation. Often an employer has to be in the market all the time. The requirements of the business are such that it is not just a question of hiring A or B or C. It is a question of hiring them all at the same time. If the employer gets the reputation of not keeping promises, the risk that he faces is not legal suit from disgruntled employees as much as it is the inability to find people who are willing to work without being paid in advance. The reputation of being an untrustworthy employer translates quickly into dollar and cents losses. The employer that faces great reputational losses therefore is under constraints that do not bind as powerfully for individual workers. Leaving the worker with the residual legal risk therefore will usually make sense, and there can be some alteration in the underlying wage rate to compensate the worker for the loss that has to be borne. And if the reverse sequence of performance, pay now work later, is desirable it can be adopted explicitly in particular cases, or perhaps by industry custom and practice. And for the still more complicated cases, such as difficult construction contracts, a system of progress payments can be used to split the risks between the parties. It is therefore possible to develop some mechanisms that allow contracting to go forward even though some individuals are scurrilous and markets are imperfect. The world will never be ideal, and there will surely be some shipwrecks no matter what default rules are in place. But the shipwrecks do not show that the rules are wrongly conceived, for so long as the costs of enforcing contracts are positive, we have to expect some shipwrecks even under the ideal set of rules. Our task is to minimize the risk of contract breach. But that risk can never be driven to zero no matter how ingenious our contractual strategies.

VII. The Contract at Will

The basic logic of sequential performance in contracts is, I think, pretty widely accepted today. A far more controversial application of the same general approach arises in connection with the so-called contract at will. Now what a contract at will basically says is that neither side needs to offer any public justification for the termination of a contractual relationship. The employer can fire a worker without offering any justification. It is only necessary to say 'you're fired'. By way of a set-off, and it is a critical set-off, if a worker doesn't want to stay on the job, she is not required to give a long song and dance routine either. All that she has to say is 'I quit', and the relationship is forthwith terminated.

The central legal question is whether a contract of this sort ought to be permitted. There is, to be sure, a school of thought that generally says that private choice is to be severely constrained, and that the only serious social options are whether certain practices should be prohibited or required. There will be nothing that people are just allowed to do. I think that it is necessary to escape that double bind and to recognize that while a contract at will may not be appropriate in many circumstances, there will still be lots of cases, the lion's share of cases, where it turns out to be the value maximizing solution for both the employer and the employee. As before one should not make that judgement in the naive belief that there will never be a shipwreck, for cases will occur (and these are the ones most likely to be brought into court) in which the at-will result will appear to be unfortunate and unjust. But there are enormous risks in trying to tailor the entire law of contract to fit the occasional case of disaster. It is not possible in advance to simply prohibit the contract at will in the cases where it does not work. We do not have the luxury of hindsight at the time of contract formation. Instead it will be necessary to alter the entire fabric of contract law, and to create a legal environment that makes it more difficult for good employers and good employees to structure their relationships. An unjust dismissal law is always intrusive, for it exposes every decision in labour markets to second guessing by either a court or an administrative agency.

So what are the advantages of the contract at will? Paradoxically, the first is that its apparently precarious nature provides both sides with a secured obligation. The point sounds strange given that one side can quit, and the other side can fire, without any explanation. For most people the idea of security connotes a mortgage or a lien on some form of property. But it is appropriate to expand our horizons on this point. The employer that decides to fire a worker has to pay a price: he will no longer be able to reap the benefits of the worker's labour. Conversely if the worker decides to quit, she will no longer be able to command the wage. Each obligation is hostage to the other. Before quitting or firing, one has to make a hard decision of whether the benefit foregone is worth the labour or the wages, as the case may be, that can now be retained. But once that decision to sever the arrangement is made, the security on the other side is instantly realized, without the formalities and delay of foreclosure proceedings. The worker instantly recovers her labour, and the employer his cash. Knowing the efficiency of the security arrangement people will move with caution, given that it is always costly to exercise the right to quit or to fire.

The unquestioned right to quit or to fire has powerful, and desirable, incentive effects. In particular it serves as an effective check against the advantage taking that either side can resort to in a continuous relationship. Thus if the employer decides to chisel the worker, the threat to quit becomes more credible because the worker who leaves has less to lose given the greedy behaviour on the other side; and again the argument holds as well in reverse. The lazy worker is easier to fire because he is of that much less value to the employer. The party that tries to take too much of the co-operative surplus, the joint gain from trade, runs the risk of losing it all.

These moderating influences therefore lend a certain durability to the contract at will that contrasts with its fragile legal nature. Employers that spend enormous amounts of money in training workers for particular jobs are not eager to see them depart. And workers who have specific capital tied up to their jobs are not eager to leave. The contract at will is thus a versatile legal arrangement that is compatible with a wide range of business arrangements.

The contract at will is also strengthened by the same reputational forces that are at work in cases of sequential performance. The employer with a large work force is constrained in the way in which he deals with any particular employee. Firing the first worker for reasons that other workers perceive as unfair will have powerful ripple effects throughout the firm. Other workers will become uneasy and in consequence will re-evaluate their own prospects: the job that once looked good will not seem as good as it did before. The best workers may be the first to leave. The decision therefore to fire one worker, if done unwisely, can come back to haunt the employer with the 8, 20 or 1000 workers that remain in his employ. In practice, it is a lot harder to fire a worker under a contract at will than it is for a worker to quit. And it is a great mistake to stress only the legal modes of enforcement when the social ones are at least as powerful and important in practice.

The third point about the contract at will that ought never to be underestimated is the ease of its enforcement. The legal position is this: I quit, or you fire me, judgement for the defendant. The entire system takes about two words to explicate in the standard case. Anything goes within the legal system precisely because anything will not go in the market setting. And both sides can share in the administrative savings in the form of higher profits and higher wages. Ex ante only the lawyers lose when the contract at will is fully enforceable.

No one can claim that the contract at will never results in any miscarriages of justice. There is too much temptation on both sides for such a happy result to obtain. But the benchmark of a legal practice is not perfection. It is the next best alternative, and to see what that looks like it is necessary to ponder for a moment the law of unjust dismissal, which is its major rival. The one point that the "just cause" rule has in common with the at will rule is that both are two words long. But while at will is a phrase that gets the courts out of the employment business, just cause is one that immerses them in an absolute infinitude of litigation. There are any number of possible ways in which an employer may be said to misbehave and no industry or standard practices that allow these to be evaluated. If the employer fires a worker some five months after he has asked her out for a date, is her refusal to go the reason why she was fired? And if it was is the reason a legitimate one? Does it make a difference whether they associated with each other outside of work in the interim? Or before? The variations on this one theme resist any easy classification.

The same problem arises when other types of cases come up. Is there just cause to dismiss a worker whose assignment is inconsistent with her religious beliefs? Is there just cause to dismiss a worker because her division has been closed down, because the firm's cost structure is too extensive, or because a product line has been discontinued? Each of these questions has been the source of extensive litigation, and the most that can be said of any of them is that the decision made must be "reasonable under the circumstances". The upshot of this structural indecision is that virtually any dismissal for whatever reason, from redundancy to manifest incompetence may well be subject to legal challenge. And the message will not be lost on employers who will be slow to hire and slow to fire in order to minimize their exposure to liability under the unjust dismissal laws.

Nor does the problem stop with dismissal, for the insidious effects of the "regulatory pyramid" are at work in this area as in so many others. Once it is clear that dismissals are subject to judicial scrutiny, an employer may be tempted to reduce the worker's pay (say to a penny a day) or to assign her boring and unsuitable tasks. Yet the law cannot stand by and tolerate these "constructive dismissals" if it is to make sure that no unjust dismissal takes place. So these actions have to be regulated and supervised as well. But once they are, then all transfers, pay reductions, demotions must be reviewed as well, because there is nothing in the unjust dismissal law that limits its application to cases of manifest abuse. So the system grows of its own weight, so that judicial intervention can become not the exception but the rule.

There is a moral here with which it is useful to close. The contract at will is widely used for positions of great power and positions of small consequence. It is used for chief executives and for day labourers. (One common exception is tenure for university professors, itself a complex subject that cannot be discussed here.) Its ubiquity should not be treated as a sign of widespread corruption but of widespread utility. In order to go against common practice, one needs to have enormous confidence in his own judgements about right and wrong. And typically those who know most about the subject are aware of the subtle variations between individual cases and are least willing to intervene in the affairs of others. But for those who have not faced the challenges of running a business, it is easy to disparage practices that are not understood. The law and economics of labour contracts and labour markets is a complex business, whose outlines have only been well explicated in the last generations. Imperfections are the order of the day in all markets, and anyone who thinks that the legal system can be operated without substantial error and cost is unduly optimistic about the power of law in general and of regulation in particular. But as we learn more about labour markets this universal law should apply: those who know the most seek to govern least.