Thinking about punitive damages
Punitive damages is something of a misnomer: they are punitive, not damages.
Damages typically refer to an injury and the compensation necessary to make a victim whole. Punitive damages, however, are not designed to make the victim whole. They are intended to ensure that the defendant never does the wrongful act again, and to help deter others from committing similar acts. They are, thus, closely related to criminal law punishment in their goals.
Punitive damages, even thought of as a fine, differ from criminal punishment in several significant respects related to their civil classification:
- no statutory violation need be proven to find liability;
- proof is only by a preponderance of the evidence, not beyond a reasonable doubt;
- the jury is civil (which may be smaller than a criminal jury and need not decide by unanimous vote);
- private citizens may commence the action without consent of a public prosecutor answerable to the political process;
- only monetary fines, not imprisonment or hard labor, are imposed;
- typically, juries rather than judges set the fine in the first instance; and
- the fine goes to the private plaintiff instead of the state .
Moreover, to date we have seen no move in tort comparable to the criminal sentencing guidelines seeking to mandate that similar offenses receive similar sentences; punitive damages are set (by juries in the first instance) with little guidance regarding the comparable punishment imposed in comparable cases.
On the other hand, punitive damages do not seem to bear the same reputational weight as a criminal conviction: they are not seen as quite as strong a social condemnation.
I. The Background: Compensatory damages in ordinary tort law.
Ordinary tort law struggles with the problem of determining how many accidents are appropriate. Negligence law accepts that people will injure other people and their property. Not only does it not condemn this, it doesn’t even start from an assumption that those who injure others should apologize or fix the damage they have done. Rather, the basic negligence rule (in contrast to strict products liability) is that those who injure are liable to repair the damage they’ve done only if they failed to act with “reasonable care” and hurt someone to whom they owed a duty in a fairly direct fashion (i.e., proximately).
Even when ordinary tort law requires tort feasors to compensate their victims, the usual rule is distinctly undercompensatory.
First, tort does not seek to impose the bargain that parties would have reached had they bargained freely: it rarely asks juries to imagine how much the victim would have asked for (ex ante) to agree to suffer the injury. Rather, tort juries are directed to imagine the market value of the injury: the difference between the price a buyer might be willing to pay for the property (or person!) in its uninjured and injured state. Since we know that sell prices are consistently higher than buy prices, asking juries to imagine buy prices will consistently generate a price lower than what the injured party would have required to agree to the injury.
Because it uses buy prices rather than sell prices, tort, therefore, allows private parties to engage in a kind of eminent domain. They can force a party to sell (their property or personal integrity) so long as they compensate at the market price. But one thing we know about markets is that those who have not sold wouldn’t be willing to sell at the market price: otherwise they would have sold already.
Second, the American rule provides that each side pays its own costs in a litigation. Thus, a tort feasor can be assured of one thing: it will never pay the full costs to the victim, because it will not be asked to pay the cost to the victim of making the tort feasor pay. Victims, not tort feasors, pay their litigation costs.
Paradoxically, even though tort law theoretically aims at undercompensation, in practice its results may appear overcompensatory.
Torts often involve bodily injury, and placing a price on bodily injury is notoriously difficult. No market price exists, perhaps because ordinarily it would be illegal to sell someone the right to hurt you, and perhaps because very few people (and only ones in desperate or strange circumstances) would be willing to sell–or buy—those rights if it were legal. So jury verdicts valuing physical injuries are necessarily somewhat arbitrary. Moreover, since we give juries few uniform guidelines about how to value these injuries, they are also inconsistent.
Tort also allows compensation for pain and suffering. If putting a price on a lost limb or life is difficult, pricing pain is even more difficult. At least when a jury has to value a shattered bone or a missing eye, it can be reasonably certain that it understands what it is trying to price, even if has little guidance in how to price it. Pain and suffering are far harder to prove: how is a jury supposed to compare one person’s suffering to another’s, even before it reaches the problem of pricing it?
Because of the difficulties of pricing life, limb and pain, jury verdicts are highly contestable, and even critics who accept my claim that tort in theory aims to undercompensate may conclude that in practice it still overcompensates, if they think that juries systematically overvalue the injuries they are faced with. In fact, we know from a consistent body of social science that people in general consistently are willing to spend far more money to rescue or cure identifiable individuals than to prevent accidents ex ante. Accordingly, we should also expect that juries will consistently find higher damages (even using buy prices) than the same juries would be willing to spend to prevent the same accidents in the abstract. So it should not be surprising that there will be critics who, imagining the problem as one of abstract, ex ante, accidents, believe that juries are overly generous.
Getting to the right level of accidents.
Normally, preventing accidents has a cost. Thus, usually there will be a tradeoff between greater safety and other values. Tort, therefore, normally does not seek to ban accidents. Instead, it attempts to aim at some reasonable level of accident.
One way of aiming at the right level of accident is to think of compensatory damages as a kind of a bargain (parallel to the theory of the efficient breach in contract). If (and as we’ve seen above, this is a big if) tort gets the level of compensation “right,” we could take a utilitarian view of accidents. That is, if a tortfeasor is willing to fully compensate its victims and still commit the tort, then it must follow that the tortfeasor values the tort more than the victim values freedom from it – in the aggregate, society is better off if the tort occurs.
Notice that this explanation of tort law should be highly controversial.
First, many of us, much of the time, reject the basic utilitarian view that it is permissible to use A for the benefit of B without A’s consent – even if A ends up “no worse off”. Rawls refers to this as a “Kantian constraint” and in one form or another it underlies most theories of human rights, civil rights and private property. If you take the Kantian position, tort law is far too weak: individuals are being forced to suffer injuries without their consent, and the goal of law ought to be to prevent that, not simply to compensate them.
Second, even if you accept the notion that it is permissible for B to force A to enter into a transaction so long as B’s fully compensates, the difficulties we’ve seen in setting the “right” level of compensation should make you fairly skeptical about the likelihood of tort fully compensating victims. But if victims are not fully compensated, then the basic claim that A is no worse off but B is better off is incorrect. Instead, tort is simply presiding over B’s forcible expropriation of A and allowing B to get away with it, or at least with part of it.
Putting the argument into economic terms doesn’t change it dramatically. Clearly, economic activity – innovation, production, factories, and so on – is going to hurt people from time to time. Equally clearly, in the aggregate we’d rather have the factories and the injuries than no factories and no injuries.
But in specific instances, the way capitalism should work is that consumers will buy products that they think are worth the price – and the price should reflect the cost to society of producing the product. If pricing works this way, when consumers buy a particular product, they are making a judgment that it is more socially worthwhile than the alternatives.
Tort law is an essential part of this process. If a product causes accidents (or producing the product causes accidents), the accidents are a social cost of the product. For capitalist allocation methods to work properly, those costs should be included in the price of the product. But absent tort law, the cost of the accidents will be borne by victims – they won’t be part of the price. Consumers will buy too much of the product because they will pay too low a price.
Ordinary tort law can be understood, then, as forcing producers to include the cost of their accidents (measured by tort damages, see above) into the price of the product, so that consumers buy the “right” amount. If consumers are unwilling to pay that price, that is a sign that the product isn’t socially worthwhile and shouldn’t be produced. Ordinarily, there is no reason for society to subsidize products (by imposing a tax on tort victims) that consumers don’t want to pay full price for. And even if a particular product should be subsidized, it seems grossly unfair to make accident victims the source of the subsidy.
This justification for tort also suggests that current tort law is distinctly undercompensatory. First, as we’ve seen above, tort uses the “buy” price for injuries. But for the “price” of injuries to be comparable to the price of other inputs into production, it must be the “sell” price, not the buy price. Producers are not permitted to expropriate fuel, labor or raw materials (even if they compensate at market prices). They must convince, not force, owners to sell. The same should be true of injuries.
Second, the “reasonableness” rule means that producers need not compensate for injuries they have unquestionably caused, so long as the producer acted “reasonably”. But when a producer gets large enough, risk turns into certainty. A producer that takes a 1 in a million chance is certain to injure roughly 1000 people if it repeats the risk a billion times. Those 1000 injuries are costs of production just as much as the fuel or wages it pays to make the product, and they should be included in the company’s costs and therefore the product’s price. The reasonableness test, which considers risks from a non-statistical perspective, simply doesn’t generate the right answer.
II. Back to Punitive Damages.
The argument for punitive damages is fairly clear. Sometimes all the complicated arguments above are just inappropriate. The issue is not getting to the right level of accidents or deciding how to get there. Instead, the right level of accidents is zero. The issue is how to convince tortfeasors to stop doing whatever is causing the accident.
The usual test for punitive damages is that the tortfeasor has gone well beyond negligence to something considerably worse than mere unreasonableness: a “willful and wanton” disregard for human life and safety, “malice,” or “wanton and reckless disregard of the rights of others.” Someone acting this way is anti-social, taking themselves out of society and placing themselves in opposition to it.
The usual mode of tort law assumes that injury was inadvertent – the result of unreasonable risk taking in negligence law, or ordinary capitalist behavior in strict products liability law. Even intentional torts do not necessarily include malice: one may intend to scare someone (assault) or to touch them in a manner the ordinary person would find offensive (battery) without hatred or stepping outside of normal social constraints. Here, however, we are faced with a tortfeasor who meant to hurt someone, who has deliberately flouted the usual rules of human relations. Pricing the injury done to the direct victim may seem utterly inadequate.
At this point, tort law becomes much closer to criminal law. It is an expression of social disgust and disapproval (unlike ordinary tort law, particularly in the pricing view, which does not condemn those who commit injuries – it just demands that they pay for them, just as the law demands that anyone who purchases something pay for it.) It seeks to deter the tortfeasor – to convince it never to do this again – and others – convincing them not to try either. It may even seek revenge.
As in criminal law, the primary goal of punitive damages should be to prevent the defendant or others similarly situated from committing the offense again. The primary limitation should be the level of proportionality – that the punishment fit the crime – that distinguishes punishment from state oppression.
The damage done to the particular individual (or class) who brought the action – compensatory damages -- has at best a loose connection to these goals. Indeed, just as in criminal law, the injury may be as much to society as a whole as to the specific victim. A murder kills not only the victim but also trust and solidarity in the broader society. Fraud harms not only its victims but all those who thereafter fail to trust or be trusted. Producers that sell unsafe products harm all consumers and all producers, causing everyone else to have to spend extra money testing or verifying producer claims and generally decreasing economic growth and transactions throughout the economy. When these crimes are prosecuted as torts, the values being vindicated are (at least in part) public and not merely private, and the private costs to the victim are only the beginning of the injuries that must be considered.
Punitive damages, then, are public and criminal-law like. This should make them quite controversial: as we have seen, they lack many of the normal safeguards meant to keep criminal law from overreaching. Perhaps the fact that they are enforced by private citizens reduces some of the need for those safeguards: they aren’t as available for selective use as tools of a overpowerful executive seeking to repress dissent, or a popular majority seeking to keep an uppity or unpopular minority in its place, as criminal law could be. On the other hand, perhaps the opposite is true: private citizens can press for punitive damages unrestrained by democracy. On balance, the appropriate size of punitive damages should raise nearly every issue raised by criminal punishment as well as the distinctive ones related to tort’s status as a private, civil, action.
One final point: punitive damages go to the private citizen or class who brought the lawsuit, not to the state. This may seem unfair: why should a single citizen obtain a sum that, in essence is forfeit by the loser for an offense against the public. But I don’t find that complaint very convincing. Most wealth in a capitalist society is unfair, arbitrary and lottery like in exactly the same way. Capitalists get rich as side effects of systems that are created for the public benefit; they have no more (or less) claim to the wealth that flows their way than the tort plaintiff has to punitive damages. If envy is a problem here, it should be generally, and the obvious solution would be to reinstate a functioning income and wealth tax system.
A better argument would be that this is an enormously expensive way to enforce public norms. If we increased the budgets and pay of the public prosecutors – or even permitted them to sue for punitive damages under current rules, but with the damages to go to the state, or the prosecutors office, rather than private parties – we’d get a good deal more norm enforcement for our buck. That, it seems to me, is unquestionably correct. But current political mores seem unlikely to allow prosecutors to be that aggressive or to endorse directly spending the public money on the public good. Punitive damages are a backhanded way of avoiding taxes, more expensive in the long run like so many other tax avoidance schemes, but probably preferable to no regulation at all.
Compensatory damages, with their clear tendencies towards absolution – what I’ve referred to as the “make a mess if you wish, but clean it up” view of tort – are fundamentally inadequate when the message the system wishes to send is, instead, “just don’t do that.” Punitive awards make that latter, quite different, point.
-- Dec '04