Received wisdom advances two broad reasons why government is entitled to impose its will on its subjects, and why the subjects owe it obedience, provided its will is exercised according to certain (constitutional) rules. One reason is rooted in production, the other in distribution—the two aspects of social cooperation. Ordinary market mechanisms produce and distribute the national income, but this distribution is disliked by the majority of the subjects (notably because it is ‘too unequal’) and it is for government to redistribute it (making it more equal or bend it in other ways, a function that its partisans prefer to call ‘doing social justice’). However, the market is said to be deficient even at the task of producing the national income in the first place. Government is needed to overcome market failure. A society of rational individuals would grasp this and readily mandate the government to do what was needful (e.g. by taxation, regulation and policing) to put this right.

In this two-part essay, I claim that at least some, if not the whole, of the market failure argument fails to prove its case. In Part I, I look at the problem of contract enforcement and in Part II at the provision of public goods. There have been other writings using related arguments to the same effect, but one more such will not be too many.

1. One-Off Contract Execution

The division of labour implies exchange and exchange is the execution of a tacit or overt contract. In standard theory, if one party to a contract executes his part by delivering as agreed, the other party’s optimal course of action is to take the delivery and walk away without delivering his part. The first party knows this and correctly concludes that his best course of action is not to deliver. The second party knows that this is the case. Therefore the parties will not contract and the mutually advantageous exchange of deliveries will not take place. (The well-rehearsed model of this interaction is, of course, the notorious prisoners’ dilemma which has been a cornerstone of arguments for political authority from the 1950s to the 1980s, though it has since been somewhat eroded by the widening understanding of game theory.)

If circumstances permit the two parties to execute simultaneously, the problem disappears, since each delivery is contingent on the other, so that both parties are best off if each delivers. Plainly, however, it is not always convenient or efficient to insist on cash-on-the-barrelhead dealings. A modern economy is inconceivable without the bulk of exchanges being non-simultaneous. Do contracts involving credit or other non-simultaneous execution require a third party, such as the state, to see to it that both parties fulfil their commitments?

It used to be thought that in a small-scale, ‘face-to-face’ society, say the village cattle market, no third-party enforcer is necessary, because no party to an exchange could risk to default and face loss of reputation and even retaliation in some unpleasant form. In large groups of ‘faceless’ contracting parties, on the other hand, each could default with impunity. Hayek, for one, strongly argues that in the ‘great society’ where anonymous dealings prevail, a firm legal framework was needed to underpin the free market which could not function at all without it. His ‘spontaneous order’ emerged inside this (non-spontaneous) framework.

This type of ‘market failure’ argument, that comes strangely from a Hayek who is widely venerated as a champion of classical liberalism, fails mainly by getting the facts wrong. The most obvious one is the unworldly idea of contracts between anonymous parties who can walk away from the contract without delivering their side without anyone knowing who they were. There are no anonymous contracts. Where thousands of faceless customers stream through the checkout counters of a supermarket, they have a contract with the bank who issued their credit card, and the card company has a contract with the supermarket, each party to each contract being duly named and identified. In wholesale trading dealers in the same trade know a good deal about their counterparties half a world away and if they do not, their bankers and brokers do. Default risk is shifted, often to specialised intermediaries, to whomever will assume it at the least cost because best able to minimise it. For relevant purposes, the wide world is a face-to-face society, or at any rate functions much like one.

The other fact of life that standard market failure theory does not get right is that while many market exchanges are done in the form of one-off contracts that are fully executed once each party has made one delivery, many more—probably the greater part of aggregate market exchanges—is not. It is run on continuing contracts providing for repeated executions, often an indefinite number of times.

2. Repeated Executions

The example that first springs to mind is the labour contract, where the employee agrees to render some service week by week, month by month, and the employer agrees to pay him at regular intervals, for a period or until either party terminates the contract by giving due notice. Similar contracts with repeated delivery often govern the supply of parts and materials to manufacturers and the supply of finished goods to commerce. They typically run for an indefinite yet uncertain duration.

Unlike the one-off kind, such contracts do not obey the logic of the prisoners’ dilemma where ‘take the money and run’, i.e. deliberate default, is the best strategy. Defaulting on any given delivery at any link of the chain of deliveries breaks the chain and normally wrecks the contract. Therefore it pays only if the gain made by defaulting on a single delivery outweighs the present value of all future gains that would accrue if the contract went on to its indefinite term.

The balance in favour of continuing to deliver as agreed (or pay as agreed) will be vastly strengthened if the potential defaulter loses, not only the anticipated gains from the contract he would break, but also the potential gains from other contracts that third parties would decline to conclude with him after they learned that he was a defaulter. The forgone gains from potential contracts, added to the forgone gains from the contract the defaulter has actually broken, create a strong conjecture that carrying out commitments under the system of repeated contracts is a self-enforcing convention.

This conclusion parallels the deduction, made by numerous theorists and therefore known as the Folk Theorem, that mutual cooperation through a series of indefinitely repeated games, each of which has the structure of a prisoners’ dilemma, is a possible equilibrium.

3. Free Riding

Little is left, then, of the market failure argument which holds that the market cannot spontaneously generate the contract enforcement required for its own functioning. If this argument were valid, a really free market would be a logical impossibility. ‘Real existing’ markets would all depend for their very existence on the scaffolding of an enforcing apparatus.

It so happens that most ‘real existing’ markets do make some use of the enforcement service provided by the legislator, the courts and the police. Why is this the case if the market failure argument is invalid and there are adequate incentives for rational economic agents to adhere to a self-enforcing convention of contract fulfilment?

The short answer is that punishing and hence deterring default is rarely costless. Even passively boycotting the defaulter involves some cost in inconvenience, even though incurring the cost may be the means of preventing a greater loss. If much the same result can be got without incurring any cost, that method will be preferred.

Once legislatures, courts and police—in one word, the government—is in place, maintained by the taxes it has the power to exact, firms and individuals will rationally prefer to entrust the task of enforcement to it and enjoy the illusion of getting something for nothing, instead of making the effort themselves. They perceive this as a chance to free-ride on the taxes paid by everybody else, and do not perceive that ultimately their own taxes must increase to cover the cost of all the free riding others will also prefer to do. The tendency fits nicely into an important objective of every government, namely the goal of discouraging private enforcement and vesting in the state the monopoly of all rule enforcement.

4. The Enforcement Agency

It is a long way from putative market failure to the risks of overwhelming political power, but that long way must nonetheless be travelled. Textbook theory rather blithely teaches that since contracts are inherently default-prone, their binding force must be assured by the services of a specialised enforcement agency (such as the state). However, if the agency is to be bound by tacit or overt contract (such as a constitution) to a best-effort service in the interest of all bona fide economic agents, that contract itself needs enforcement, for why else should the agency not go slack or biased or otherwise abusive? Plainly, however, the enforcement agency cannot be entrusted with enforcing such a contract against itself. The supposed remedy could well be much worse than the disease. Perhaps herein lies the ultimate failure of the market failure thesis.


*Anthony de Jasay is an Anglo-Hungarian economist living in France. He is the author, a.o., of The State (Oxford, 1985), Social Contract, Free Ride (Oxford 1989) and Against Politics (London,1997). His latest book, Justice and Its Surroundings, was published by Liberty Fund in the summer of 2002.

The State is also available online on this website.

For more articles by Anthony de Jasay, see the Archive.