Q: ‘Practitioners of Law and Economics claim that efficiency is the best or even the only goal that the law should pursue. Opponents of Law and Economics claim that efficiency analysis has no role to play in legal arguments. Both sides are wrong: efficiency is neither an all-pervasive value nor is it no value at all’.
-Discuss, drawing examples from any substantive area of legal analysis.
Q: ‘The practitioners of Law and Economics do not often present their most basic assumptions for scrutiny. However, those assumptions are profoundly inegalitarian. It follows that the notion of ‘efficiency’ is not a helpful guide to what the law should be’.
-Discuss, using an economic analysis of any area of law that you are familiar with. Can the notion of ‘efficiency’ be rescued to make it normatively desirable?.
The economic interpretation of law advances a bold thesis: this is that many, if not all, areas of the law can be understood to be concerned with the promotion of efficiency. The thesis has both a descriptive and a normative aspect. According to the former, analysis reveals that many legal rules in areas such as tort and contract in fact promote efficiency; while according to the latter, the theory forms a basis for the critique and evaluation of new legal rules and legislation.
The arguments turn on concepts like efficiency, superiority, optimality, allocation and distribution. The thought of Pareto has been particularly influential. The most basic notion in the economic analysis of law is efficiency or Pareto optimality. A situation is said to be Pareto optimality .However, such an explanation is quite crucial to the question whether any form of efficiency should be regarded as relevant to the legitimacy of the law.
Why transaction cost efficiency?
So what is wrong with the other notions of efficiency mentioned by Finch? Pareto efficiency provides the usual starting point for discussion of this subject. A distribution of resources (or opportunities or entitlements) is Pareto optimal (or efficient) if any further change would not make even one person better off and would make at least one person worse off, judged both times by the person’s own standards. To understand what this means, suppose there are only two people (A and B) affected by a transaction which consists of taking an asset away from A and giving it to B. The change is Pareto superior if B compensates A fully for that loss, and is still better off himself. Likewise, a change which affects numerous people is Pareto superior if all potential losers from the change have been fully compensated, and the gainers are still better off because of it. A distribution is Pareto optimal (or efficient) when all such mutually beneficial exchanges have been fully exhausted.
However appealing the notion of Pareto efficiency might seem, it is not very useful. Hardly any transaction in the world – perhaps none – satisfies the criterion of Pareto superiority. Almost every transaction affects countless parties, at least some of whom are made worse off because of it. At the very least, almost any transaction changes the pattern of demand for the resource exchanged (by satisfying someone’s demand for it). It thus affects the market price of that resource, thereby reducing the prices of identical or substitutory goods and adversely affecting the suppliers of those goods. Further, given that few Pareto superior transactions can be made, it follows that almost any state of affairs is Pareto optimal: ‘What is, is Pareto optimal’! Not much guidance can be gained about the real world from a set of criteria which outlaws virtually every transaction, and which validates almost any distribution of resources.
Kaldor-Hicks efficiency is often put forward as a more useful replacement. That it is sometimes referred to as ‘potential Pareto efficiency’ provides a clue. Consider the simple two-party transaction again, where an asset is taken from A and given to B. The transaction is Kaldor-Hicks superior if B could have compensated A fully for the latter’s loss and still been better off. The feature which makes this version of efficiency different from the previous one is precisely that no actual compensation is required here, so long as the gains to the winners are larger than are the losses to the losers.
This conception of efficiency suffers from serious problems, however. First, recall the formula just given, that a transaction is Kaldor-Hicks efficient if the gains to those who ‘win’ from it are larger than the losses to the losers. Consider how those losses are to be reckoned. The only coherent measure available is the consent of each of the losers. So a transaction would be efficient only if the party losing out could have been fully compensated to its own satisfaction, while still leaving a surplus for the party benefiting from the transaction (call this the ‘K-H surplus’). It follows, however, that if any person affected by a transaction objects very strongly to being deprived against his will of an asset or entitlement, so that a very high level of ex post compensation would have to be offered to redress his grievance at being treated thus, then the efficiency of the transaction becomes doubtful.
Obviously, the greater the harm to the losing party, the higher the compensation required to remedy it, and so the greater the possibility that there is no K-H surplus. All it takes to make the universe of Kaldor-Hicks-efficient transactions an empty set is one person who sincerely cannot be bought – that is, a person who values autonomy, either his own or that of others, so highly that no amount of after-the-fact compensation could possibly leave him as well off as he would have been[,] had the loss never been inflicted (without consent) in the first place. Again, transactions in the real world affect countless people. This is even truer of a rule or policy, potentially governing multiple transactions, that we might seek to test using the criterion of Kaldor-Hicks efficiency. This raises the probability that among those affected by that rule or policy would be someone ‘who sincerely cannot be bought’ in the sense specified above. Even if that is not the case, it is likely there would be little or no K-H surplus in many, perhaps most, instances, given the large number of adversely affected persons. So Kaldor-Hicks efficiency is scarcely more useful than the Pareto version.
Secondly, arguments based on Kaldor-Hicks efficiency violate the fundamental egalitarian principle that all persons are to be regarded as equals. One embodiment of this moral equality is the maxim that they must be treated as ends, not as means. However, reliance on Kaldor-Hicks efficiency to defend a rule or policy would imply that the mere fact of a gain for a large group of people justifies losses to a smaller one. It would follow that those losing out because of the rule or policy were being used merely as means towards the ends of others.
Even great harm to them would be acceptable, as long as a sufficiently large number of others derived small individual benefits which, taken together, were even larger. To illustrate this objection, we can imagine a transaction that inflicts loss on A, such that d100 of compensation would be required to make him whole. The same transaction also, however, brings benefit to two hundred other people, each of whom would be willing to pay d1 for that benefit. The K-H surplus is then d100 (the gain of d200 minus the loss of d100), and so the transaction is efficient. The problem, however, is that, once the existence of the K-H surplus is established, no compensation need actually be provided to A for the requirements of Kaldor-Hicks efficiency to be satisfied. As noted, this is unacceptable to anyone committed to the basic egalitarian principle.
Finally, consider a problem common both to the Kaldor-Hicks and the Pareto versions of efficiency. On either conception, whether the transfer of an asset from X to Y is efficient depends on whether Y could compensate X for the loss and still be better off. However, Y’s ability to compensate X depends (among other things) on the prior distribution of resources. Very simply put, a rich Y would be able to offer a larger payment to X for the transfer. What is more, note the dependence of both types of efficiency on the preferences of individuals for a resource or an entitlement. Efficiency is achieved when such a resource or entitlement is vested in the person who places the greatest value on it, thus satisfying the recipient’s relevant preference. But a person’s preferences themselves are a function (inter alia) of his initial wealth and income. This is easier to understand if we remember that as defined in economics, wealth includes a person’s native endowments, his skill at a sport or his possession of two healthy kidneys, etc. So a physically weak person would have a greater preference for a legal entitlement to bodily integrity than someone stronger. And a person might have a greater preference for leisure at one income level than at another. It follows that ‘If income and wealth were distributed differently, the pattern of demands might also be different and efficiency would require a different deployment of our economic resources.’ This shows also that simply because a transaction is efficient says nothing about its normative merits. Efficiency depends on the prior distribution of income and wealth amongst the parties, and that distribution may itself be unjust. If the initial distribution of resources is unjust, attaining an efficient allocation of those resources would be simply meaningless, since it would make that allocation no more desirable from the ethical point of view than if it were inefficient. So efficiency of either sort says nothing about the ultimate justifiability of any transaction, and thus, a fortiori, of any rule or policy.
The relationship between fairness and efficiency
For all these reasons, we must conclude that efficiency in itself does not provide a goal that any area of the law should aim at. It creates no sufficient reason for the law to be one way rather than another. This shows why efficiency concerns cannot be on par with fairness concerns, which do create good reasons for the law to be one way rather than another. But note the emphasised qualification, that efficiency does not ‘in itself’ provide any goals for the law. Here is how this is to be understood.
Let us draw a distinction between what might be called procedural and substantive goals of the law. A substantive goal of a certain branch of the law (or indeed of the law as a whole) would be an end that it seeks to pursue. Taking corporate insolvency law as an example, the substantive goals of this law might be stated, at different levels of abstraction, as ‘To be just to all the relevant parties’, ‘To treat parties as equals’, ‘To provide a fair scheme of co-operation under the circumstances peculiar to insolvency’, ‘To show equal concern and respect for the interests of all those facing such circumstances’, and so on. Substantive goals, then, are those which justify the existence of this part of the law by showing it in its best light, by demonstrating why it is worthwhile having it.
Procedural goals, on the other hand, are about how the law goes about attaining its substantive goals. For example, procedural goals would be concerned with the methods the law adopts to implement a fair scheme of co-operation in the circumstances of corporate insolvency. This perspective enables us to clarify the statement made above: efficiency can never be a substantive goal of the law. It can never by itself confer justification on any part of it. However, efficiency – understood properly – is quite indispensable as a procedural goal. Once a set of substantive goals has been exogenously specified (e.g. using a theory of justice), efficiency can be used to judge between various proposed schemes for implementing it.
Suppose we specify that a branch of the law should attain a set of substantive goals. Suppose also that there are two proposals about how to bring about these goals. Now it is obvious that no proposal can be operationalised and implemented costlessly. Some resources will inevitably be consumed simply in putting in place any such proposal, and in maintaining it in operation. Remember also that most resources are scarce. So the more resources that are consumed in implementing a particular scheme, the fewer that will be available for other worthwhile objectives. It follows that if there are two methods of bringing about a certain goal in these circumstances, we must choose the method which is less costly to implement, other things being equal. Any other decision would amount to wasting resources, since the same objective could have been attained and in addition, a surplus would be available for application towards other valued goals. This waste is morally objectionable, then, to the extent that the attainment of those others goals is morally desirable.
This provides an understanding of transaction cost efficiency. A method of implementing a set of substantive goals is efficient in this way when the resources it consumes in the process of implementation are lower than would be consumed by adopting any other feasible method of implementation. Put differently, a method is efficient, given a particular amount of resources dedicated towards implementation, when it can operational the set of substantive goals to a greater degree than would be possible for any other feasible method. It is obvious that to attain transaction cost efficiency should be a (procedural) goal of every part of a morally defensible legal system. This, it is submitted, is how efficiency analysis should be employed in arguments about the law: to determine whether it attains its substantive goals in the cheapest feasible way.
This also helps us understand why we must reject Finch’s repeated assertion that fairness may sometimes have to be traded off against efficiency. These two ‘rationales’ cannot ‘pull in opposite directions’ because substantive goals and procedural ones, ends and means, do not compete. This is perhaps the most deeply engrained confusion in Finch’s work, and it creates glaring inconsistencies in it. For example, she emphasises at one point that efficiency is about producing, ‘at minimal cost and without waste[,] whatever social and distributional goals are set by society’. This is an accurate statement of the relationship between substantive and procedural goals: the latter can be attained only by reference to the former, which means that substantive goals must be settled upon before procedural ones are chosen. Put another way, we cannot even understand what count as efficiency ‘costs’ and ‘waste’ until we decide what ends we wish to pursue.
This might sound strange, so consider this simple example. Suppose a company’s directors have signed contracts with the company which create incentives for them to run it so as to further the interests, predominantly, of the shareholders as a group. Now if the end is to maximise shareholder value, then the directors’ contractual allegiance to shareholders is desirable and should be encouraged. On the other hand, if the end is to ensure that directors operate impartially as amongst shareholders, creditors and employees, then their exclusive allegiance to shareholders is a ‘cost’ and should be mitigated. Now consider the following sort of assertion: ‘In assessing whether the law’s treatment of employees leads to efficiency in insolvency processes, it is necessary to keep the efficiency question separate from the issue of fairness or distributional justice’.
Returning to efficiency, there are two broad (and overlapping) categories of transaction costs (i.e. the costs inherent in implementing a set of substantive goals) relevant here. Co-ordination costs are the costs arising from the fact that there are limits on what people in the real world can foresee, and on their cognitive capacity for selecting the appropriate response to a set of circumstances presented to them. These costs also arise because there are informational asymmetries, for example, where information relevant to the common plans of both A and B is available only to the former, or is available to him to a greater degree than to the latter. An example from the corporate insolvency context is that of the coordination problems faced by the creditors of a company on the verge of insolvency. One implication, given the aim of implementing any set of substantive goals requiring co-operation, is that some resources would have to be expended either ensuring that information is available to a fuller and more uniform degree to all the relevant actors, or that the adverse effects of this not being the case are remedied.
With this in mind, note that co-ordination costs are a function, inter alia, of expertise. This is for several reasons. First, on the assumption that there are inherent differences amongst individuals as to the potential to develop certain skills, some of the social resources earmarked towards the imparting of those skills would have to be spent finding the right candidates to impart them to. Then there are search costs involved for a potential buyer of expertise in locating it. Information that there is an expert who can do the job within the constraints set by her financial resources, say, has to be discovered by the buyer, information that there is a set of tasks his expertise would allow him to perform rewardingly has to be discovered by the seller, and there will seldom be a perfect overlap in what the buyer requires and the supplier provides. Third, the greater the complexity of the tasks involved and the more knowledge available on how to perform them, the greater will be the division of labour amongst the experts in the area, and so the greater the co-ordination costs of their working together. Finally, there is the risk, given the expert’s cognitive limitations, of his choosing the non-optimum strategy for dealing with the task at hand, one requiring ten hours of work when it could have been performed in eight, say. Take these together and it is clear that Finch’s ‘expertise’ benchmark is very much an aspect of efficiency, since it is mostly about such costs, and efficiency is about attempting to reduce these costs.
The second category of transaction costs is motivation costs. These arise because, once again along with asymmetric information, different actors have different incentives about how to behave in any situation, since they perceive their interests not to lie in the same direction. Given the aim of implementing a set of substantive goals requiring co-operation, some resources would have to be expended aligning their interests in such a way that the actors would be encouraged to pursue those goals. In English corporate insolvency law, a good example is found in the wrongful trading provisions in section 214 of the Insolvency Act 1986, which seek to align the interests of managers of firms on the verge of insolvency with the interests of the firm’s creditors. Once again, it is suggested that, to the extent that ‘accountability’ is concerned with issues distinct from those considered under the ‘expertise’ benchmark, it is a function of motivation costs: the greater the latter, the greater the need for the former, and vice versa.
The burdens of ‘fairness’
Most importantly, a distinction was drawn between the ultimate ends of the law, and the methods it adopts in attempting to achieve those ends. It was pointed out that the ‘three’ benchmarks so far considered are about means not ends, and therefore, that their invocation is entirely question begging unless it can be connected to a clear statement of the ultimate aims the law is attempting to attain. Only in that case would we be able to evaluate the fruitfulness of Finch’s analysis in helping us understand the law’s efficiency in attaining those ends. Her ‘fairness’ benchmark provides a suitable candidate – as noted, fairness can undoubtedly serve as a substantive goal of the law. In view of this, her discussion of fairness bears a heavy analytic burden. So finally in this review, that is what we must consider.
As noted above, we get a hint that ‘fairness considers issues of substantive justice and distribution’. Yet even this minimal understanding of what she might mean by ‘fairness’ is not secure. This is because, in criticising contractarian justificatory models, Finch asserts that in them, the ‘distinctiony between principles of fairness and justice and principles governing the allocation of other goods such as wealth isy problematic’. Casting an eye back on Finch’s own suggestion about what fairness is, the reader would see why this muddies the waters considerably. How can there be a distinction of this sort, when Finch herself recognises that fairness is (in some appropriate way) about the justice of distributions? Surely this includes distributions of ‘goods such as wealth’? Nor is it plausible to suppose that she regards this ‘insight’ too occult to have been accessible to contractarians. Amongst those contractarians who are explicitly interested in fairness and justice (so that the early Thomas Jackson, for example, is excluded), it is well known that many regard the justice of a decision to be dependent upon the fairness of the decision making process. And, to take some examples mentioned by Finch herself, Rawls understands justice to be about the distribution of certain ‘primary goods’, including wealth; Korobkin takes justice in insolvency situations to be a matter of the distribution of influence over the strategies and actions of the distressed corporation; and for this reviewer, it is about the distribution of rights in the insolvent company’s assets. Since Finch acknowledges that contractarian theories are concerned with ‘principles of fairness and justice’, then, and since she also accepts that fairness and justice are (at least partly) about principles governing distributive issues, it is difficult to understand what ‘distinction’ between these and the so-called ‘principles governing the allocation of other goods such as wealth’ she would require contractarians to render more ‘unproblematic’. What is more, given that this makes it seem that there might for Finch be such a ‘distinction’, it becomes even more difficult to understand what she means throughout the book when she talks about ‘fairness’.