Common Tragedies

Thoughts on Environmental Economics

Puny-tive damages

Posted by Rich Sweeney on July 9, 2008

In one of it’s less talked recent decisions, the Supreme Court dramatically reduced the ammout of punitive damages Exxon was forced to pay for allowing one of its tankers to spill 11 million gallons of crude oil into the Prince William Sound in Alaska in 1989. Writing for the majority, Justice Souter appeared to substantially limit the use of punitive damages, apparently capping them at the level of awarded compensatory damages. As an economist and someone who’s worked in litigation consulting, I’m inherently weary of our ability to effectively set numerical limits on future damages. While Souter didn’t actually set a dollar cap, capping a ratio seems equally as arbitrary to me. Moreover, in this instance, I’m afraid that the shortcomings of applying such a ratio will be particularly burdensome on the environment.

One of the main functions of punitive damages is to deter entities from taking similar actions in the future.* Compensatory damages, on the other hand, are awarded to place a clamaint in the position he would have been in had the wrongful act not taken place. In this decision, SCOTUS effectively says that the latter can never legally exceed the former. (At least in maritime cases. It’s not really clear to me how broadly this rule will apply). Yet I find Souter’s justification of this ratio deficient for two reasons. First, he never discusses the theoretical relationship between the two types of damages. Second, he sets his cap based on historical ratios, ignoring the endogeniety of risky behaviour with respect to the limit of potential punishment.

In a world where transaction costs were zero, all transactions were legal, and Ricardian intergenerational concerns prevailed, punitive damages would be unnecessary. An intelligent judge could fully and perfectly compensate all affected parties, and society would be indifferent as to whether or not a similar act occurred again. We wouldn’t worry about deterrence, and market forces would prevail.

The real world is, of course, much messier: Transaction costs abound, contracts are imperfect, society outlaws certain acts (like the sale of organs or children), and intergenerational compensation is complicated. On top of all that, our judicial system is a blunt instrument. Court costs and rules of merit/ standing serve to exclude most small or tangential claims. For all these reasons, it seems reasonable that society would seek some additional penalty, beyond just compensation of affected parties, in order to deter acts where the harm extends beyond the parties involved. We recognize that our legal system can’t perfectly compensate all effective parties, but still try to protect fringe claimants by deterring actions that will adversely affect them.

Pegging punitive damages to compensatory damages will dramatically limit our ability to protect potential fringe claimants and bias risky behaviour towards areas that we already do a poor job of valuing. Up until now, assuming the probability of being successfully prosecuted is 1, a firm considering a risky act (in this case, hiring an alcoholic to captain its oil tanker) had to respect the fact that a judge and jury could subsequently tack on punitive damages to the point where the total costs outweigh the benefits. Now however, they would only be deterred up to the point where the value of the risky act is less than 2X prospective compensatory damages. As I just described above, actual awarded compensatory damages become a smaller and smaller component of total damages as the harm done becomes more opaque or spread out over a larger group of individuals. We can surely think of lots of heinous acts where the harm is far away and difficult to directly quantify (although Souter apparently can’t), but right off the bat this seems to obviously describe environmental disasters such as Valdez.

Ok, quickly, on endogeniety and legal rules of thumb. Souter cites a host of statistics on past damages ratios in developing his proposed 1:1 limit. He writes that the majority of these cases were probably decided justly and therefore suggests we just impose this mean amount on the population as a whole in order to curb unfair outliers. However, this assumes that the distribution of acted upon risky endeavors will remain unchanged after the cap is enacted. What he fails to consider is the effect that the threat of unlimited punitive damages had a on preempting particularly harmful acts during the time period he’s studying. Previously, firms were concerned solely with the total damages amount, and the threat of massive punitive damages probably deterred a lot of risky environmental behavior. Now, however, disipated harm looks comparatively cheap, and therefore firms are going to purchase more of it, shifting the very ratio Souter is treating as fixed.

This is something I argue about with Johnny Walker, my Borkish** budding corporate lawyer roommate, all the time. He wants to cap medical malpractice damages, but refuses to acknowledge that this would encourage the most risky behaviour at the tails of the damages distribution. I have a feeling he’ll have something to say about this post. Most likely that I don’t understand the law at all.

* There are other possible legal justifications for awarding punitive damages aside from determent. Souter has a nice overview of the history of punitive damages beginning on p. 17 of the decision.

** Ok he’s not really Borkish at all, just pro tort reform. Speaking of Mr. Bork, for anyone else who thinks they’re for capping torts, I refer you to the single best editorial the New York Times has ever written. Irony isn’t even the word.

12 Responses to “Puny-tive damages”

  1. AC said

    I’m a Texas lawyer with a products-liability practice. I think you’re well off the mark here.

    The goal of tort law is efficient deterrence. If the costs of avoiding a harm exceed the cost of the harm (discounted by its probability of occurring), we do not want the defendant to avoid the harm. Over-deterring is just as bad as under-deterring.

    You assume that juries make accurate determinations, and I think that is your mistake. Jury verdicts are largely random, and the randomness biases punitive damages awards towards over-deterrence. The reason is that large outliers cannot be offset by small outliers (there are no “negative” punitive damages).

    Suppose you submit the identical punitive-damages case to 100 different juries. Let’s assume the verdicts follow a roughly normal distribution, with a median of .6 and a standard deviation of .6. Almost all juries, in other words, conclude that the efficient multiplier is less than 1.8.

    But 5 of the juries award more; 1 of the 100 juries awards 100 times the compensatory damages. The average (mean) ignoring the outliers may be 1.0, say, but the outliers may raise the average to 3.0.

    Which multiplier is efficient?

    You have to have an argument why it is more efficient to include the outliers in the mean. I don’t think there is one because juries are subject to systematic biases. Peremptory strikes introduce bias; availability for a lengthy trial introduces bias (panels are weighted towards people with a lot of time on their hands); and juries are plainly subject to emotional appeals. Even if you assume biases are equally likely in either direction, the few large verdicts cannot be offset by a few small awards; there are no “negative” multipliers. Jury bias systematically skews puntive multipliers too high.

    Bias cannot be observed; it can only be inferred. And a court doesn’t get a 100-verdict sample. It has to decide the case based on a sample of one. The Supreme Court in Exxon was essentially trying to figure out, “How do you determine whether you have an outlier based on a sample of one?”

    It concluded that the historical distribution of punitive damage multipliers is the appropriate population. Absent any data showing that, say, environmental contamination cases historically have produced larger multipliers, that’s the best information we have. It is impossible (indeed, meaningless) to ask whether a jury’s verdict is an outlier just by looking at the specific facts of the case. That is something the Supreme Court understands from institutional experience.

  2. Johnny Walker said

    Rich, you don’t understand the law at all.

    Of course, neither do I. However, I may be able to correct a few flawed assumptions that you have made.

    You base your justification for high punitive damages awards on the need to account for future actual harm to individuals who are not parties to the lawsuit for harm resulting from the wrong giving rise to the litigation (particularly, environmental disasters):

    “As I just described above, actual awarded compensatory damages become a smaller and smaller component of total damages as the harm done becomes more opaque or spread out over a larger group of individuals. We can surely think of lots of heinous acts where the harm is far away and difficult to directly quantify (although Souter apparently can’t), but right off the bat this seems to obviously describe environmental disasters such as Valdez.”

    The Supreme Court has held that it is an unconstitutional deprivation of property without due process of law for a judge to enter a punitive damages award in which the jury took into account harm done to individuals who were not parties to the lawsuit. This makes sense, right? In law, you can’t punish people for harm that they may or may not have caused. “Opaque and spread out” harm as you call it. You have to prove it with evidence. This also blurs the line between compensatory damages, the pecuniary elements of which must be proven, and punitive damages. Besides the fact that compensatory damages require proof of their amount, the principle factor in determining whether to award punitive damages is the outrageousness of the conduct itself, not the degree of the harm caused. So these awards are never going to serve the purpose you want for them.

    To the contrary, primary purpose of a punitive award is much the same as a criminal sanction; this is why ratios are necessary to ensure that civil defendants are not unconstitutionally deprived of their property. As you know, the criminal justice system provides several procedural protections for defendants: facts must be proven beyond a reasonable doubt, the rules on vicarious liability of an entity for the actions of its agents are stricter, and the kind of evidence admissible to show criminal fault verses that admissible to shoe eligibility for punitive damages is much more limited. The due-process ratios serve the interests of justice, not pragmatism. They ensure some degree of protection for civil defendants who face something pretty indistinguishable from criminal penalties (except that they can often far exceed them), in a system that does not offer criminal defendant protections.

    Even as a practical matter, though, ratios and by extension reasonable damages caps are good policy. As Cass Sunstein, law’s emissary into the dangerous and scary world of economics, has detailed in a book he co-authored, Punitive Damages: How Juries Decide, juries are notoriously bad at calculating awards. If only we could have a panel of twelve citizens with your dedication to calculating a proper figure for the deterrence of future conduct, the system would operate much more smoothly. Instead what you get is twelve folks randomly selected from the voter registration roles in some random small town who have spent the past two weeks staring at one table packed with family members crying because a chainsaw manufactured by Extremely Wealthy, Inc. cut off their daddy’s legs, and another table packed with someone from the general counsel’s office of EWI and its four or five corporate attorneys. Faced with such unpredictable actors and the potential for a “fine” in the amount of whatever number they pull out of their heads, a company would have to be insane to even step foot in the courthouse.

    And this is MY problem with unrestrained punitive damages awards (along with unrestrained nonpecuniary compensatory damages): they lead to results based on a cost-benefit rubric in which the merits of the case itself play no role. Rather than submit the fate of the company to twelve jurors, the smart company will settle regardless of merit (knowing that juries often act regardless of merit) and save itself the anxiety.

  3. Rich Sweeney said

    Lawyers! Thanks for the great comments. I’m clearly no expert on the law, and, despite the tone of my post, don’t usually pretend to be. My point was not to say that the Exxon case was decided incorrectly, or to imply that Souter did not have some good reasons for instilling the cap. In the decision, he talks at length about the need for fairness and predictability in the law. AC, you add the good point that efficiency is important, and that juries have demonstrated the potential for horribly irrational and inconsistent damages. (Although Souter does write that “A survey of the literature reveals that discretion to award punitive damages has not mass-produced runaway awards”). Johnny I basically agree with just about everything you say here, although I think (know) we disagree about what these words mean in practice. In particular, I appreciate your points about juries only concerning themselves with proven facts and present defendants. Yet when you say “reasonable” caps are good policy, you assume that we can determine ex ante what a reasonable cap is.

    Which brings me to the only point of my post. Souter’s rationale for the 1:1 cap was flawed for two reasons: (1) He assumed that the observable ratio of punitive to compensatory damages revealed the global ratio of all possible risky activity, and (2) he ignored the fact that altering the threat of punishment in courts will alter the behaviour of individuals and corporations in the real world. I also described what the theoretical relationship might be under the current, unlimited framework, and concluded that this will lead to riskier behaviour in situations where compensatory damages would have a lower expected value relative to total actual damage.

    Now none of this is to say that its not worth it to limit damages anyways if you think that the efficiency or applicability/ fairness benefits of caps outweigh these costs. I’m just saying that the probability of another drunken oil tanker captain spewing crude into the ocean again in the future just went up.

  4. Rich Sweeney said

    Ok I guess I had one other point in my original post. Souter acknowledges that one of the main objectives of punitive damages is to deter. Seems to me that if that’s a stated objective, we should think about how to actually achieve it. Now for all the reasons Johnny mentioned this may be impossible, and maybe we should leave deterrence of the sort to the legislative branch. But if this is the case Souter should have decreed that from now on punitive damages will no longer seek to deter, and will simply be known as “additional damages”.

  5. AC said

    “(2) [Souter] ignored the fact that altering the threat of punishment in courts will alter the behaviour of individuals and corporations in the real world.”

    Yes, altering the threat of punishment will alter the behavior of potential defendants. That is a good thing if they are too cautious today because of the danger of outliers.

    Basically, you are worried about what I’d call “first-order efficiency” — using punis to get incentives just right. You are absolutely correct as a matter of theory. But we live in a world where juries get things wrong. The SC was attempting to resolve “second-order” efficiency problems — how do you eliminate the risk of outliers, which over-deter?

  6. AC said

    Johnny Walker:

    The SC has held that, as a matter of due process, states may allow juries to consider the harm to others within the same state. (State Farm Mut. Ins. Co. v. Campbell).

  7. Johnny Walker said


    I would be interested to hear what language you derive that from. Specifically, the 2007 case Phillip Morris v. Williams states as follows:

    “In our view, the Constitution’s Due Process Clause forbids a State to use a punitive damages award to punish a defendant for injury that it inflicts upon nonparties or those whom they directly represent, i.e., injury that it inflicts upon those who are, essentially, strangers to the litigation. For one thing, the Due Process Clause prohibits a State from punishing an individual without first providing that individual with ‘an opportunity to present every available defense.'”

  8. Johnny Walker said


    Let me try to address your points as you restate them more directly. Of Justice Souter’s analysis you say:

    “(1) He assumed that the observable ratio of punitive to compensatory damages revealed the global ratio of all possible risky activity, and (2) he ignored the fact that altering the threat of punishment in courts will alter the behaviour of individuals and corporations in the real world.”

    I don’t think he really did either of those things. As to point (1), Justice Souter used the observable ratio, not to reveal the global ratio of all possible risky activity, but to determine a just (though, in your view, less effective) amount. And past jury awards are a classic way of evaluating jury awards for being unjustly excessive (e.g., this method is also used by the California Supreme Court in evaluating California jury verdicts). And I think it is a good method. Verdicts are supposed to represent the opinion of one’s peers, so why not establish a ceiling by taking the largest possible sample of these peers as is available?

    As to point (2), again, Justice Souter’s agenda is just punishment where yours is the most effective result, but I think you are wrong here either way. The breadth of corporate activity that can be subject to a punitive damages award is great. These lawsuits can often be founded upon theories of the slightest negligence and most plaintiff’s attorney’s will argue for punitives whether or not there is a basis for them in the facts. It would be difficult if not impossible for a company to conform its behavior such that it would be immune from the potential for a large puntitive damages award under our tort system. But AC has already aptly made this point about over-deterrence.

    Which leads me to the most important point I really want to make (and this really just gets deeper into what AC seems to have already acknowledged). Those responsible for judicial policies do a great disservice to justice when they get too cocky about the merits of our system. It is a good system, but in this age of judicial realism, we have to acknowledge that it is not the role of our, nor any other, judicial system to establish some “absolute truth.” Its role is to produce an outcome to a dispute by applying admissible and available evidence to an appropriate evidentiary standard. Under such a system, sometimes the result conforms with the truth (such as it exists) and sometimes it does not. This is demonstrated by the example that the correct result in a criminal murder case may be the incorrect result in a wrongful death case where the evidentiary standards are different. “Untrue” results (maybe it is more accurate to call them unpredictable) are more frequent in the tort system as the evidentiary standards are far lower. An enlightened and humble judicial system must acknowledge this reality by limiting the punishments it is willing to subject people to: whether that be the death penalty or unlimited punitive damages. This may not be why Justice Souter finds unpredictable and unlimited damages to be unacceptable, but I think it is a reason worth considering.

  9. M said

    Without having had the chance to read the opinion because of bar study, I don’t have a serious problem with the 1:1 limit ONLY in the oil spill context because the punitives can still come out really high, and there’s still a strong incentive not to spill oil because you lose the money you would have made if you sold it. I do have a problem with the 1:1 ratio being declared a constitutional standard (what clause anyway? due process? whatever…)

    My other gripe is people thinking that they need to understand “the law” in order to understand these cases. It really comes down to balancing business costs versus compensation and deterrence, and that’s not very hard to understand. Figuring out the right balance and implementing it, on the other hand, is really hard. That’s why we let the legislatures continually fine tune their instruments: damages caps, regulatory measures, taxes, pleading standards, etc. Trial and error should produce results that are as close to efficient as possible; the court has no business declaring that the constitution somehow requires a 1:1 ratio, which will be really hard to change if it turns out that it doesn’t adequately deter.

  10. Rich Sweeney said

    M, you say “I don’t have a serious problem with the 1:1 limit ONLY in the oil spill context because the punitives can still come out really high, and there’s still a strong incentive not to spill oil because you lose the money you would have made if you sold it.”

    But “high” is a relative term when it comes to damages as the incentive provided by punitive damages depends on your ability to pay. As the LA Times reported, Exxon made over $40 billion in profits last year, which means they’ll pay of the punitive damages in about 5 days.

  11. M said

    They’ll actually pay off the damages in 12 hrs. “The new figure represents about 12 hours of revenue for the giant oil company, which reported record profits of $40.6 billion in February 2008.”

    But if we want to take their money, there are other ways to do it. Our government makes a conscious effort to keep business costs down for big oil to reduce our reliance on middle east oil. Why should 12 people be able to defeat our government’s policy decisions? Besides, a $5 billion punitive damages award (the original amount) would probably just get passed on to consumers, rather than cutting into anyone’s profits.

  12. Rich Sweeney said

    Actually I’m pretty sure a large firm specific cost like this would come out of profits, not get passed on to consumers. Exxon’s big, but it’s still a price taker and that price is set on a large multi-firm global market. Since this cost only applies to them, I think they’d just have to eat it. Unless they’re supplying the marginal unit I guess….

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: