Brandeis University Web Stite

The Costs of the Costs of Accidents

Copyright (c) 2005 University of Maryland School of Law
Maryland Law Review


64 Md. L. Rev. 337

LENGTH: 8080 words


NAME: Jules Coleman*

BIO: * Wesley Newcomb Hohfeld Professor of Jurisprudence, Yale Law School; Professor of Philosophy, Yale University. I would like to thank John Gardner and members of the New York Area Tort Theory Discussion Group, especially Ben Zipursky, Mark Geistfeld, Catherine Sharkey, Anthony Sebok, Stephen Perry, and Anita Bernstein for comments on an earlier draft of this paper.

... Guido Calabresi's The Costs of Accidents is unquestionably the most important book written in tort theory during the past fifty years. ... Still, there is no denying that Calabresi was aware from the outset that any account of tort law as a technology for reducing and spreading accident costs would have to confront the extent to which efforts to do so were compatible with the requirements of fairness or, as he later called them, "other justice considerations. ... The goal of optimal deterrence requires that the same "rational" decisions be reached even though the costs of prevention (arguably) fall to the injurer and the costs of the harm fall to the victim. ... The view of tort law as a mechanism of social reform designed to solve the problem of accidents is only part of the Calabresian revolution. ... Corrective justice is the principle that in virtue of committing the first-order wrong, the injurer has made himself vulnerable, should his wrong result in harm to another, to incurring a second-order duty to make repair. ... But what we cannot allow is the Calabresian framework to blind us to the fact that this is the question we need to ask about our tort law, not the question whether it remains an attractive mechanism of social reform. ...  


Guido Calabresi's The Costs of Accidents 1 is unquestionably the most important book written in tort theory during the past fifty years. Much of what has been written since the book's publication either extends Calabresi's insights or reacts critically to them; some essays do a bit of both. Its lessons are by now so absorbed as to constitute part of the common understanding of the subject. The Costs of Accidents not only provides the intellectual framework within which the current debate occurs, but the language in which it is expressed as well. In time, the expression "cheapest cost avoider" will no doubt find its way into the Italian translation of the Oxford English Dictionary and may even appear someday as an entry in the Authorized Version. The Costs of Accidents has redefined both tort law and tort theory - often, but not always, for the betterment of each.

The primary aim of law is to regulate conduct through norms - usually rules - that create reasons, grounds, or warrants for action. Many of these reasons take the form of rights, privileges, and liberties on the one hand, and duties and other encumbrances on the other. Arguably, both the laws of tort and crime impose duties or prohibitions on agents, whereas the law of contract confers powers on individuals to create legally enforceable rights and duties.

To be sure, we can imagine certain acts currently proscribed by statute no longer being criminalized, and others not currently prohibited by law coming to be proscribed. Still, it is difficult to imagine a legal regime that does not include a criminal law, that is, a body of law designed to prohibit certain conduct and to hold those who fail to comply with its demands liable to punishment - state imposed hard treatment consisting primarily of the loss of liberty and assorted legal rights.

In contrast to the criminal law that imposes duties, contract law confers powers. In particular, it empowers individuals to create a legally  [*338]  enforceable regime of rights and duties. Duties are imposed via the process of contracting, that is, derivatively, as the result of individuals exercising the powers conferred by law. In the main, the law does not impose contractual rights and duties; it confers powers on individuals to do so and provides a regime by which the rights and duties thereby created are enforced by law. What the law does is create a system of powers that includes the power to bind oneself to others and the power to call upon the state's authority to enforce the normative structure the exercise of those powers creates.

Although the formulas for empowering individuals to change their normative relations with one another may change with time and technology, as may the mechanisms for enforcing the rights and duties thereby created, it is hard to imagine a legal regime lacking such empowering rules: hard, in other words, to imagine a legal regime forgoing an opportunity to provide individuals with this way of expressing their autonomy and of making their commitments to one another credible.

While it may be difficult to imagine a legal regime with neither criminal nor contract law, it is not nearly so difficult these days to imagine a legal regime without tort law. Indeed, for at least the last twenty-five years, legal reformers, with varying degrees of enthusiasm, have suggested one or another plan for doing away with all of tort or with significant parts of it. Reality has in fact done reformers one better, since New Zealand has already abandoned tort law as a mechanism for dealing with a broad range of ordinary accidents. 2

It is natural to ask whether there is something special about tort law that makes it more vulnerable in this sense. Why do various alternatives to tort law - for example, mandatory first-party insurance or public insurance conjoined with increased public enforcement of safety norms - seem altogether plausible, whereas efforts to abandon criminal law or contract strike us not only as implausible, but unthinkable?

We will not get very far in answering this question if we accept without reflection the inference that doing away with criminal law and contract, for example, is in fact unthinkable. While it may be hard today to imagine a legal regime without either, it was not always so. It was not all that long ago that Grant Gilmore proclaimed the death of  [*339]  contract: a death owing in part to the fact that contract law was being swallowed up by - of all things - tort law. 3

And there have always been reform movements that would have us abandon the criminal law in favor of a system of "treatment." How familiar at one time was the thought that so-called criminal wrongdoing does not so much evidence morally culpable conduct worthy of punishment as it reveals the existence of a psychosocial disorder that calls for health care, not hard time!

Many areas of the law have been the object of the reformer's eye - not just tort law. The vulnerability of areas of the law to the reformer's knife can depend on a variety of factors, often having little to do with the intrinsic nature of the law itself. In the case of the criminal law, for example, reform movements have largely sprung from a view about the nature of human motivation, or, in some cases, skepticism about the possibility of free will.

What explains the particular vulnerability of tort law at this point in time? I want to suggest that three factors have played a prominent role in the current conceptualization of tort law, and that this way of conceptualizing tort law has had the effect of rendering it vulnerable to the reformer's knife. Two of these factors receive their most important expression in Calabresi's The Costs of Accidents. I fear I have to take at least partial responsibility for the third. Some good, but more bad, has come out of the reconceptualization of tort law. In this Essay, then, I want to demonstrate the contributions that both Calabresi and I have made to the reconceptualization of tort law.

Calabresi's key contribution in this regard is not his emphasis on the goal of efficiency as a standard for assessing tort law. Rather, it is his claim that the domain of tort law is "accidents taken as a whole," as well as his emphasis on "costs" as the normatively most important conceptual category for thinking about accidents. My contribution has been to emphasize that philosophical appropriateness of separating two sets of questions from one another. The first is the distinction between the grounds of the claim to repair and the grounds of the duty to compensate. The second is the distinction between the grounds of each and the modes of each, that is, between the sources of rights to repair and duties to provide it on the one hand, and the institutional mechanisms for doing so on the other. Unlike my teacher, in retrospect, I think this reconceptualization has proven more harmful than insightful.

 [*340]  Part of what I hope to do then is rectify the wrongs I have committed to tort theory.

I. Accidents and Costs
In The Costs of Accidents Calabresi asks a straightforward question: How should the costs of accidents be distributed? To answer that question, he introduces two general criteria of assessment: cost avoidance and fairness. 4 Cost avoidance is further divided into three categories: primary, secondary, and tertiary. 5 Roughly, these categories cover the costs of avoiding accidents - that is, deterrence - the costs of bearing the costs of accidents - that is, cost spreading - and the costs of administering systems of deterrence and spreading. In general, we want to avoid accidents that are not worth their costs and to spread the costs of accidents that are not worth the costs of preventing them. We do not want to deter or spread without regard to the costs of doing so. We want optimally to reduce and spread risk. Or to put it in the language that has come to be represent the very idea of the efficient allocation of accident costs: We want to minimize the sum of the costs of accidents and the costs of avoiding them.

And more. We want not only to optimize accident costs, but we want to do so in a way that is fair or that meets the demands of justice. The relationship between fairness and justice in Calabresi's work has always been somewhat unclear. Nor is it clear whether in Calabresi's view justice (and fairness) are side constraints on the pursuit of cost avoidance or simply other goals to be balanced against the goal of optimally reducing accident costs. 6 These issues can be and have been much debated, but they are of no concern to us here.

As I see it, tort law was changed not by the answers Calabresi gave to his question or by the criteria he introduced as appropriate to evaluating various answers to it. Tort law was changed by the nature of the question Calabresi took to be the appropriate one to ask. For what Calabresi did was take the subject matter of tort law, namely, particular harms or accidents between private persons, and transform it into a different subject: accidents. Instead of asking whether A should compensate B for the harm he caused, Calabresi asked the very different question: What should we, as a society, do about accidents? Accidents are the problem, not this or that accident. To be sure, we could deal with the social problem of accidents by focusing on particular  [*341]  accidents as they occur, and by asking whether there is a particular person responsible for an accident who should therefore be responsible for its costs. But we don't begin with that question. We begin with the very different question of what we ought to do about accidents. Tort law is one of many possible institutional arrangements - technologies - available for answering that question - the societal-level question. 7

Second, precisely what is the societal problem of accidents? That is the question to which the Calabresian answer is, "their costs." It's not just that accidents happen. They do of course. Rather, it is that accidents impose costs - internal and external costs - on those involved in them and on others as well. More than that, preventing accidents is costly as well. So we don't want to prevent accidents if the costs of doing so are too high. We want a policy according to which we prevent those accidents we can prevent in a cost-justified manner. We have a social goal, and that is optimally to reduce accident costs. And again, one way of optimally reducing accident costs might be by imposing their costs on a case-by-case basis on either a particular injurer or victim. In other words, tort law is an available technology of cost avoidance: a potential tool of social policy.

Of course, if tort law is a potential tool of social reform, it may not in fact be a particularly effective one. The case for tort law is simply whether it is the best available technology for optimally dealing with the social problem of "accidents and their costs." Since the publication of Calabresi's book, several reformers have had their doubts about whether in fact it is.

Nevertheless, it is this formulation of the problem, not so much the answer Calabresi gives to it, that has been of primary importance to tort theory ever since. Its central elements are: (1) conceptualizing the problem facing tort law as the need to provide an effective instrument for solving a societal issue: the problem of accidents; (2) the normatively important feature of accidents is their "costs."

 [*342]  It might be helpful to note that there are many obvious other features of accidents that could have been offered as normatively significant. Among the obvious alternatives are "harms," "wrongs," "wrongdoings," and "rights." One could easily have said, for example, that the problem of accidents is that they often involve harms or wrongs. And if the problem is a social one and if tort law is to be responsive to it, perhaps its aim should be to distinguish accidents that involve harms from those that do not - though both involve costs. Or, one could just as easily have said that the problem with accidents is that sometimes they are the result of wrongs, sometimes not; the aim of tort law is to deal in an appropriate way with the difference between them.

Or, one could just as easily have said that the problem with accidents is that they sometimes are the result of wrongdoing, and that the purpose of tort law is to distinguish among the cases accordingly and respond appropriately to them based on the difference between them.

Or, one could have suggested that some accidents involve the invasion or infringement of rights and that it is the aim of tort law to distinguish those that do from those that do not. Both may involve the imposition of costs, but that feature of them is not what is most interesting from the moral point of view. The interest lies in the invasion of a right, and tort law should be designed to address the difference and respond appropriately to it.

In the cases of harms, wrongs, wrongdoings, and rights invasions there are costs, but the concern of tort law is not with their costs - or so one might have argued. Indeed, one might have argued this way even granting Calabresi's first premise: that the overarching problem is that of accidents - not this or that accident.

And just to be complete, one need not have accepted the idea that tort law is about accidents at all - whether on an individual or social basis. For one could just as plausibly suggest that tort law is about harms or wrongs or wrongdoings and so on. Some of these involve accidents, but many do not.

These points are significant because we might otherwise fail to notice just how radical a reconceptualization of tort law the Calabresian revolution wrought - so familiar has the phrase "accidents and their costs" become.

II. The Liberal Reform
Importantly, The Costs of Accidents was a product of its time. It was written in the late 1960s and published in 1970. Fueled by a liberal  [*343]  ideology, this was a period flush with confidence in the liberal state's ability to cure social ills of all variety through law. This is the period in which reflective political theory was dominated by the likes of Rawls's A Theory of Justice, 8 and Dworkin's Taking Rights Seriously. 9 This is the period of belief in social engineering for the production of desirable social ends: whether the reduction in the costs of accidents or the just distribution of benefits, opportunities, or burdens throughout society.

The Costs of Accidents is a product of its time in another sense as well. By the 1960s, the basic tort of A hits B has been replaced by group torts with many wrongdoers and just as many, if not more, victims. The immediacy of harm has been replaced by wrongdoing at one time and resulting harm occurring much later. The technology of injury is undergoing a major shift, as is the proximity of the causal connection between action and harm. The nature of the one-to-one relationship of injurer and victim is changing as well. The simple model of A hitting B, celebrated as one of the four paradigms of causal relations in Richard Epstein's A Theory of Strict Liability 10 (ironically published at just about the time The Costs of Accidents was), was for many no longer helpful in our ability to grasp the essence of tort law.

Calabresi's book was embedded in a political and legal culture that was particularly receptive to it. His great innovation in taking advantage of this framework was to present the problem of tort law not as one between particular actors, but as a social problem: the problem of accidents. This framing of the problem was welcomed by both the liberal social engineers and the legal realists. The conventional categories had lost their usefulness. The law, and its various institutions, including courts, were viewed as engines of social change and not merely as passive agents responsive to the claims brought to them by litigants and argued before them from the narrow point of view of the interests of the litigants.

Though very much a product of the liberal social reform movement of the late 1960s and 1970s, Calabresi's book was a significant departure from the other works of the period by its emphasis on the law as a mechanism for promoting efficiency. Though different in some ways, the pursuit of efficiency is closely tied with a broadly speaking utilitarian moral and political theory. In contrast, the leading philosophical works of this period were decidedly anti-utilitarian. This is true of Rawls's A Theory of Justice, Dworkin's Taking Rights Seriously,  [*344]  and, equally importantly, Nozick's Anarchy, State, and Utopia. 11 Whereas Rawls and Dworkin are egalitarian liberals, Nozick is a libertarian and decidedly non-egalitarian. But even the libertarian Nozick picked out the utilitarian as one of his two main targets (the other being the liberal egalitarian).

Calabresi shared with the liberals an unwavering confidence in the state as the engine of social change. This separated them all from Nozick, who argued for a minimal, night-watchman state partially reflecting a skepticism about the possibility (and the desirability) of state-sponsored engineering for social justice.

Unlike his liberal egalitarian allies, Calabresi saw the private law as an appropriate instrument for pursuing efficiency. If Nozick was taken aback by the liberal egalitarian's view of law as apt for social reform through redistribution, he would have been rendered apoplectic by the suggestion that the private law is an appropriate institution for pursuing utility. Private rights of action enforce the substantive rights litigants have against one another. The responsibility of the court is to identify those claims and to enforce them. The court has no authority to fix rights and duties ex post so as to promote overall utility. Litigants do not come to court to provide judges with the opportunity to display their social reformist tendencies.

The Costs of Accidents is weakest in its discussions of the roles considerations of justice and fairness should play in accident law. Calabresi is not clear whether justice or fairness are goals of accident law or constraints on the pursuit of its goals. Nor is he clear what the content of justice or fairness is. Still, there is no denying that Calabresi was aware from the outset that any account of tort law as a technology for reducing and spreading accident costs would have to confront the extent to which efforts to do so were compatible with the requirements of fairness or, as he later called them, "other justice considerations." 12

Considerations of fairness play two central roles in The Costs of Accidents. The first concerns the justification for risk spreading. The primary goal of accident law is optimal deterrence. Optimal deterrence requires first that we determine which accidents ought to be prevented. Only when the costs of preventing an accident are less than the (expected) costs of the harm ought an accident to be prevented. If the costs of prevention exceed the expected costs of the  [*345]  harm, then preventing harm is not rational, and it would be a mistake to induce a potential injurer to take irrational levels of precaution. So we want to induce potential injurers to take precautions whenever it is rational for them to do so, but not otherwise. Fair enough.

Another way to put the same point is this: One ought to be liable only for those harms that one ought to have avoided. From an economic perspective, one ought to avoid all and only those accidents that could be prevented by incurring cost-justified precautions. Accidents that can be prevented only by taking precautions that are themselves more costly than the accidents themselves ought not to be prevented. That means that the costs of those accidents will fall on victims. No one claims that the victims are to bear these costs because they are responsible for having brought them about. Not at all. They are being required to bear the costs because there is no good economic-efficiency-related reason why anyone else (in particular, the injurer) should. And so there is a straightforward sense in which victims are being asked to bear costs for the pursuit of some social goal - namely, optimal accident cost avoidance. In that case, arguably, the costs of pursuing optimal deterrence are falling unfairly - or at least, disproportionately - on victims. Why ought victims to be conscripted in this way?

Calabresi's emphasis on risk spreading as a goal of tort law is partially responsive to this concern. The efficiency or accident avoidance goal of tort law requires indifference as to the distribution of costs. This is easy to see. If one person engages in an activity that benefits him to the tune of $ 100, then he will continue to do so even if the harm or costs to him of doing so are quite high - provided the costs fall below $ 100. Once the costs exceed $ 100, he stops the activity. In other words, when the costs of preventing the harm (avoidance of the activity) are less than the harm, a rational agent takes the precautions. When the cost of prevention (avoiding the activity) exceeds the costs of the harm, he suffers the harm as a cost of securing the benefit. In this case, both the costs and benefits are his to bear.

The goal of optimal deterrence requires that the same "rational" decisions be reached even though the costs of prevention (arguably) fall to the injurer and the costs of the harm fall to the victim. Collectively rational action, which is just another way of referring to optimal deterrence, requires that our rules encourage individuals to take precautions as if they were both the injurer and the victim. In fact, of course, they are not, and so the costs of accidents that ought not to be avoided - those for which it would be collectively irrational to avoid -  [*346]  fall to victims. But why in the name of collective rationality ought they disproportionately to bear the costs?

Calabresi's view is that they ought not. Those costs ought to be spread. Again, risk or cost spreading should be accomplished through mechanisms that are consistent with both justice (and fairness) and efficiency. There is no unique way to spread risk, but several ways are compatible with justice: private insurance contracting and public insurance, for example. Any mechanism that spreads risk maximally over persons and time and in doing so minimizes its potential impact on individuals is permissible provided it is compatible with more fundamental concerns of fairness. In this way, Calabresi saw that there would always be two questions about accidents and their costs understood as a social problem: their minimization and their distribution. The former is the concern of efficiency, broadly understood; the latter is the concern of justice or fairness.

III. A World Without Tort
It is easy to see how the Calabresian approach articulated in The Costs of Accidents - what I think of as the social-engineering view of tort law (tort law as a technology of social reform) - makes a world without tort law altogether imaginable, and perhaps even desirable. For if we accept the Calabresian approach to tort as a technology for solving the social problem of reducing accident costs fairly, the question is whether tort law - roughly characterized above - is the right kind of instrument. The answer, of course, is that it depends.

In order to reduce accidents at the lowest cost, incentives have to be imposed on those individuals who are in the best position to reduce accidents at the lowest cost. But in any particular lawsuit it is always an open question whether the parties to the suit include the individual or individuals in the best position to reduce accidents at the lowest cost. That is, taking seriously the goal of optimal deterrence at least invites us to consider whether limiting the allocation-of-loss decision to particular injurers and victims is the best alternative. Clearly, it is not the only alternative.

More generally, if the goals are optimally to deter unjustifiably risky behavior and also to compensate those who suffer misfortune at the hands of others, then in principle the best approach might be to separate these two goals from one another. Liability is a cost, and we should use it as an instrument to promote deterrence. If the goal is to reduce risk, then we should want to discourage unjustifiably risky activity whether or not the risk materializes as harm to others. In contrast, tort law only imposes costs on those whose unjustifiably risky  [*347]  activity is harmful to others. This not only runs the risk of leaving too much wrongdoing inadequately deterred, it also leaves too much to chance. At least in theory a better approach might be to identify all those who unjustifiably risk harm to others and to impose a fine or sanction on them adequate to deter their mischievous conduct.

On the other hand, if we want to compensate victims fully, we should not hold their claims hostage to the uncertainties of litigation. We should create a compensation fund from which the victims of conduct secure compensation for the losses they have suffered. Claims to repair are claims against the fund, not against particular wrongdoers or those who are otherwise at fault in harming them.

But to separate the questions of liability and compensation, as pursuing the general goals suggest we would have reason to do, invites the dismantling of tort law as we know it. For tort law as we know it may simply not stack up as an attractive instrument for distributing the costs of accidents - for solving the societal problem of accidents and their costs.

In many ways, the more significant point is that once we reconceptualize tort law along Calabresian lines, the case for having a tort law as we know it rests on the claim that it provides an optimal compromise, not an ideal. Tort law is not, to use the economist's jargon, a first-best solution to the social problem it is designed to handle. In an ideal world with low search and information costs, separating the concerns of deterrence from those of compensation would be the right thing to do. And it would mean imposing costs or liabilities on individuals whose conduct is harm-threatening even as it is not harmful. And that is not how tort law proceeds.

Even if in the ideal world it might be better to treat those who put others at risk on a par with one another by subjecting equally risky conduct to a comparable "sanction" or "penalty," it is extremely costly to do so. Take a simple example. If we wanted to impose equal fines on all negligent drivers, then we would have to increase our police forces dramatically and put most of them on traffic patrol. As it is, the vast majority of negligent driving goes undetected and the vast majority of negligent drivers are not fined or otherwise subject to liability. To hold all negligent drivers equally subject to an expected liability penalty, we might need a police officer at every intersection with the sole purpose of being on the lookout for negligent driving.

Instead of relying on public enforcement of the norm against careless or negligent driving, the costs of which would be outrageous and wasteful given that the police will be distracted from guarding against other serious forms of mischief, we can rely on tort law to provide  [*348]  a source of private enforcement. It can only work its charms as a mechanism of private enforcement, however, by bringing injurers and victims together in a way that prevents us from pursuing deterrence and compensation as separate goals. It is a compromise on the ideal: defensible in the real world, but a compromise nonetheless.

Tort law is a technology of social reform for the enhancement of overall human well-being. It is to be assessed and appreciated alongside other possible instrumentalities. The key is to identify the set of social ills to which tort law might be part of the remedy. It is a tool not merely of deterrence and compensation, but of private enforcement as well. Calabresi, for one, is prepared as well to see it as a tool for wealth redistribution.

The view of tort law as a mechanism of social reform designed to solve the problem of accidents is only part of the Calabresian revolution. After all, this picture does not identify which features of accidents are normatively important. The other part of the Calabresian framework is the emphasis on costs. Here there are many factors at work, which taken together pretty much make the emphasis on costs appear so natural as to have it pass relatively unnoticed; certainly, the emphasis on costs has for too long gone completely unchallenged.

The most important factor contributing to the emphasis on the moral significance of costs is the formulation of the aims of tort law as deterrence and compensation, and the larger framework of seeing the law in terms of the role it might play in improving human welfare. Accidents are bad because they are costly. After a point the costs they impose are wasteful. Collectively we are all worse off on balance when accidents impose costs that could be reasonably avoided. Our collective well-being is reduced.

On the other hand, the problem with accidents from the point of view of victims is that accidents represent a misfortune in their lives that imposes a cost upon them. It sets them back because they have fewer resources as they pursue their projects, plans, and goals. Ultimately, their welfare suffers.

A rational policy for dealing with accidents will deter accidents that ought to be deterred and thereby optimally reduce costs and increase welfare; at the same time it will attend to compensating the victims of misfortune insofar as doing so can be accomplished without incurring additional costs. Indeed, if it weren't for their costs, why would we be interested in accidents in the first place?

Oddly, this approach was made more palatable by the fixation in the early 1970s in philosophy on distributive justice, which was equated with principles and institutions for distributing benefits and  [*349]  burdens (or costs) throughout society. There was little if any emphasis on notions of corrective justice, almost no discussion of the centrality of the concept of personal responsibility to the demands of justice, and so on. The language of justice and the language of the law were merged through the discourse of costs and their distribution.

It is tempting always to write a revisionist history in which oneself and one's heroes come out smelling like a rose, but there is no such story to be told here. The language of the day was the language of costs, not the language of harms, wrongs, wrongdoings, breaches of duty, responsibility, or causation. Indeed, if anything, concerns about causation were recast in terms of luck. If two persons both engage in unreasonably risky activity and one is fortunate enough to escape injuring another, whereas the other is not, why should fortuity be all there is that justifies imposing liability on the latter but not on the former?

IV. Making It Worse
Beginning in the early 1970s, I raised questions about Calabresi's account. What I didn't realize until much later was that in every relevant respect I had completely bought into the overall framework. As a result, my work on the moral foundations of tort law, even where it was critical of Calabresi, did more to entrench the larger project than to redirect attention from it.

I felt that Calabresi and those who followed him had done an excellent job of exploring the efficiency of the tort system and alternatives to it. They had done less well at determining what justice required in the way of an appropriate policy towards accidents.

Mind you, it is not that Calabresi had failed to address the issue. Indeed, as I mentioned above, the argument for secondary cost avoidance - or risk spreading - is in part an argument from fairness. In the case of accidents that could have been and should have been avoided, the loss should fall to the person who failed to take appropriate prevention measures, presumably as a way of deterring him and others similarly situated.

But in the case of accidents whose prevention is not worth the costs, no liability is imposed on the injurer. The reason is a societal one; doing so in effect shrinks the pie. The question is why injurers should bear the costs of a policy that benefits everyone. Shouldn't the costs be spread among the beneficiaries as a whole? That is an argument from fairness.

 [*350]  Again, Calabresi develops the concept of an "at-fault" pool as an alternative to the current tort system. 13 The idea here is that the current system puts too much emphasis on luck. Equally faulty risk takers should be liable equally for their conduct. Their liability ought not reflect what happens beyond their control. Each should be equally liable for equal wrongdoing, whatever the differences may be with respect to the causal upshots of their mischief. This too is an argument from justice, analogous in most ways to the argument from justice for punishing criminal attempts on a par with successes.

So it is not as if Calabresi was insensitive to the claims of justice with respect to accidents and their costs. But he did not explicitly explore what constraints, if any, principles of retributive and especially corrective justice might impose on pursuing accident cost avoidance. That is where I came into the picture.

In On the Moral Argument for the Fault System, 14 I asked whether considerations of retributive justice would restrict imposing liability only to those who are at fault for the harms they cause others: those who are morally to blame for the harm their conduct occasions. I argued for three conclusions. First, I argued that the concept of fault that is presupposed by the principle of retributive justice is much stronger than the notion of fault that is required for liability judgments in accident law. Secondly, I argued that even if tort law imposed a standard of fault comparable to that required to establish moral culpability and thus within the province of retributive justice, retributive justice would not require the fault system for the allocation of accident costs. Finally, I showed that retributive considerations would more likely support something akin to Calabresi's at-fault pool for distributing accident costs.

Then, in a series of papers beginning with Justice and the Argument for No-Fault 15 and culminating with Corrective Justice and Wrongful Gain, 16 I asked the same questions about accidents and corrective justice (which for quite a while I treated as interchangeable with "compensatory justice"). The question I was most concerned to ask was this: What constraints, if any, did the principle of corrective justice impose on the way we deal with accidents and their costs?

The first order of business of course was to offer a conception of corrective justice. For quite some time I defended the view that corrective  [*351]  justice is the principle that wrongful gains and losses must be annulled. I characterized gains and losses as a kind of distortion in the distributive scheme, and the burden of justice to correct for them - correct for them by seeing to it that they were annulled. There were many niceties to the view, but the key to fleshing it out was a series of distinctions.

As I mentioned above, the first of these was the distinction between the grounds of a claim to repair on the one hand and a duty to compensate on the other. The second was the difference between the grounds of the claims of either sort mentioned above and the mechanisms that were appropriate to discharging all warranted claims.

In the main, I argued that corrective justice would allow separating questions of deterrence and compensation. It might even allow that we compensate all individuals who have a claim to repair in justice, because they have been injured by the wrongful conduct of others, without requiring that those responsible have a unique duty to compensate. For it may be that the most efficient and otherwise desirable way to compensate those who have a right to repair in corrective justice is to allocate the costs of making repair among the population as a whole. I also showed how an at-fault pool would be compatible with corrective justice, so conceived, but by no means uniquely required by it.

The long and short of it is that I had given a philosophical credibility to the "do away with tort law" movement. For tort law does not separate these questions: If a victim has a right to recover, she holds an enforceable claim against the defendant, who has a correlative duty to repair. If anything, tort law was as crude a mechanism for securing the aims of retributive and especially corrective justice as it was a mechanism for promoting the efficient allocation of accident costs.

What Calabresi had done for the economic analysis of tort, I unwittingly did for the justice side of the ledger. If economic considerations suggest the reasonableness of disentangling what tort law treats as unified, considerations of justice would certainly provide no bar to doing so. Indeed, I gave ample reasons for thinking that corrective justice might be better served by severing the injurer from the injured. I had done my share to out-instrumentalize the instrumentalists. Doing away with tort law was not only imaginable, it was downright desirable.


V. Rectifying the Wrongs I Have Done
I have spent the last decade trying to undo the wrong I have done to tort theory. 17 Of course, some, like Calabresi, will see me as having recently lost my way. He may be right, but I don't think so.

The truth is that I made two fundamental mistakes in my early work. To be honest, these mistakes led to much more interesting, creative, and controversial claims than I might otherwise have made, and I surely don't regret going down what I now see as a mistaken path. My biggest mistake was to see the problem of accidents in largely Calabresian terms: as about their costs. What principles of justice ought to figure in determining how the costs are to be allocated? My second mistake was to misunderstand the nature of corrective justice.

Spurred by the very thoughtful criticisms of Stephen Perry 18 and Ernest Weinrib 19 (and others), I reconsidered first my understanding of corrective justice, and ultimately with that my mistaken emphasis on costs as the normatively important feature of accidents.

Corrective justice is the normative principle that connects our first-order duties of care to others to a second-order duty of repair in the event we fail to discharge the first-order duties we have. At bottom is the idea that our relations with one another are structured in part by rights and duties we owe one another. We do not see each other merely as potential holding bins of some share of the world's welfare. Among the duties we owe others is the duty to take appropriately into account the impact of our conduct on their safety and other of their important interests. We have a duty to take proper care with respect to those interests. Determining both to whom we owe such a duty and the content of the duties we owe is part of the point of tort law, and efforts to do away with it have simply ignored this essential feature of it.

A failure to take appropriate care may or may not lead to loss or harm. It is nevertheless a wrong to another. If it leads to a setback to the protected interests of others, it is a harm: a harm that is the result of a wrong. The harm will likely be accompanied by tangible loss for which we can associate a dollar figure of some sort.

The point to emphasize first is that in failing to discharge a duty of care, it is not that injurers may have done something bad or regrettable,  [*353]  or something worthy of blame. Indeed, they may have done nothing blameworthy at all; what they have done is a wrong to a person. Corrective justice is the principle that in virtue of committing the first-order wrong, the injurer has made himself vulnerable, should his wrong result in harm to another, to incurring a second-order duty to make repair. Corrective justice is not primarily about costs and their allocation. It is about some of the first-and second-order duties we have to one another and the relationship between them: why, in particular, a failure to discharge a certain first-order duty of care is grounds of a second-order duty of repair. 20

It is an interesting but very different question to ask what is the content of that duty of repair. I have argued at length that the content of the duty of repair depends in part on the practices we have. In tort law, that duty is one to make good the losses that have resulted. What is plain is that the emphasis on loss and its distribution is of secondary, not primary concern, to the principle of corrective justice. Instead, the normatively significant concepts include those of responsibility, duty, wrong, and harm.

In saying this, the corrective justice theorist is claiming that something important has been lost within the Calabresian framework of accidents and their costs. In turning the problem of an accident - in which the question is whether what happened to you is something that is my problem to address because I have wronged you or inadequately taken into account your interests and the impact of my choices on those interests - into the problem of accidents - in which the question is how should the costs of accidents as a whole be minimized and distributed - the Calabresian framework diminishes, if it does not eliminate altogether, the rich moral character of tort law. Tort law is not a mere technology; it is a way of expressing the nature of responsible agency to one another in our legal practices.

 [*354]  None of this is to suggest that we could not at the end of the day abandon tort law. There is nothing sacred about it. After all, we may develop other institutional and non-institutional arrangements for expressing the principles and values of responsible agency to one another, and in doing so we may no longer need tort law as the fundamental institution for doing so. But what we cannot allow is the Calabresian framework to blind us to the fact that this is the question we need to ask about our tort law, not the question whether it remains an attractive mechanism of social reform.

Calabresi's most important contribution in The Costs of Accidents is that it shifts the paradigm of tort law, fittingly, to accidents and their costs. This is a paradigm from which we have not recovered, no thanks to my own contributions. Beyond that, no interpretation of the book should fail to emphasize its historical context. It was penned in the period of Rawls's A Theory of Justice and the beginning of Dworkin's work on equality. This was a period of great confidence in the power of government to solve social ills. It may be worth noting that this great optimism about law as a vehicle of social reform - what I think of as an "aspirational" view of law - figured prominently in the claims both Calabresi, and especially Dworkin, made about their respective areas of law. In Calabresi's case, he has continued to press a view of the common law as a set of categories that come to us and which it is the burden of judges both to respect and to rework in the light of contemporary social problems the law addresses. The law's categories "fault," "causation," and "foreseeability," for example, are to be given their content in the light of the law's role in reducing and distributing accident costs - their content derived from the social purposes of the law. In Dworkin's case, the very idea of law as well as interpretation of the law of any particular political community requires seeing it in its best light - as pursuing or exhibiting a commitment to a range of values and ideals: thus, the "constitution of principle." 21 Thus, the continued attraction of both Calabresi and Dworkin to the contemporary liberal legal academy. The Costs of Accidents is not merely a great work; it is an important one.

Legal Topics:

For related research and practice materials, see the following legal topics:
GovernmentsLegislationTypes of Statutes


Click here to return to the footnote reference.n1. Guido Calabresi, The Costs of Accidents: A Legal and Economic Analysis (1970) [hereinafter The Costs of Accidents].

Click here to return to the footnote reference.n2. See generally Donald A. Rennie, The New Zealand Accident Compensation Scheme, in Compensation for Personal Injury in Sweden and Other Countries 143 (Carl Oldertz & Eva Tidefelt eds., 1988).

Click here to return to the footnote reference.n3. Grant Gilmore, The Death of Contract (1974).

Click here to return to the footnote reference.n4. The Costs of Accidents, supra note 1, at 24.

Click here to return to the footnote reference.n5. Id. at 26-29.

Click here to return to the footnote reference.n6. See id. at 25.

Click here to return to the footnote reference.n7. My argument is that Calabresi's most important contribution is his reconceptualization of tort law as accident law, his redefining the problem as the social problem of accidents, and his identifying the most important normative category as that of "costs." In saying that, I do not mean to suggest that Calabresi was the first person to view part of the law of tort as replaceable or modifiable. After all, earlier in the past century some of the law of accidents had already been replaced by workmen's compensation plans that were similar to the legal realist's so-called "enterprise liability" approach. Calabresi's contribution in this regard was synthesizing these considerations within an overall theoretical framework. He didn't merely advance a policy change; he theorized the reconceptualization - as did Coase in the very much related area of causation.

Click here to return to the footnote reference.n8. John Rawls, A Theory of Justice (1971).

Click here to return to the footnote reference.n9. Ronald Dworkin, Taking Rights Seriously (1977).

Click here to return to the footnote reference.n10. Richard A. Epstein, A Theory of Strict Liability, 2 J. Legal Stud. 151, 167-68 (1973).

Click here to return to the footnote reference.n11. Robert Nozick, Anarchy, State, and Utopia (1974).

Click here to return to the footnote reference.n12. See generally The Costs of Accidents, supra note 1, at 24-33; Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules, and Inalienability: One View of the Cathedral, 85 Harv. L. Rev. 1089, 1093, 1102-05 (1972).

Click here to return to the footnote reference.n13. See The Costs of Accidents, supra note 1, at 304-06.

Click here to return to the footnote reference.n14. Jules L. Coleman, On the Moral Argument for the Fault System, 71 J. Phil. 473 (1974).

Click here to return to the footnote reference.n15. Jules Coleman, Justice and the Argument for No-Fault, 3 Soc. Theory & Prac. 161 (1975).

Click here to return to the footnote reference.n16. Jules Coleman, Corrective Justice and Wrongful Gain, 11 J. Legal Stud. 421 (1982).

Click here to return to the footnote reference.n17. See generally Jules L. Coleman, The Practice of Principle: In Defense of a Pragmatist Approach to Legal Theory (2001).

Click here to return to the footnote reference.n18. Stephen R. Perry, Comment on Coleman: Corrective Justice, 67 Ind. L.J. 381 (1992).

Click here to return to the footnote reference.n19. Ernest J. Weinrib, The Idea of Private Law (1995).

Click here to return to the footnote reference.n20. In saying this, I am suggesting in effect that fault and strict liability are best conceived as ways of giving content to the duty of care we owe one another. Sometimes our duty is to take reasonable precautions not to harm or injure, and that is best thought of as expressing the requirement of "fault." Other times our duty is to exercise sufficient precautions so as not to harm others, and that is best thought of as expressing a standard of strict liability. This way of thinking invites the idea that strict liability always involves a wrong, the failure to discharge an appropriate duty of care, but the inference may be unwarranted. As I have argued in Risks and Wrongs, sometimes strict liability plays a slightly different role in tort law - as setting out a condition of justifiably engaging in an activity. See Jules L. Coleman, Risks and Wrongs 342-44 (1992). In other words, blasting, for example, is a permissible activity for one to engage in only if one is prepared to compensate those injured as a result of doing so. Whether it is possible to render the idea of strict liability in tort entirely coherent remains; on the other hand, what is required to do so remains as well.

Click here to return to the footnote reference.n21. See Ronald Dworkin, Life's Dominion: An Argument About Abortion, Euthanasia, and Individual Freedom 119 (1993).

[PHIL 1A] [Syllabus] [Handouts] [Home] [Bio] [CV] [PHIL DEPT.] [E-MAIL]

Send comments to: Andreas Teuber
Last Modified: 01/16/05
Instructor's Toolkit
Copyright © The President and Fellows of Harvard College