C. NESSON, THE EVIDENCE OR THE EVENT? ON JUDICIAL PROOF AND THE ACCEPTABILITY OF VERDICTS
98 Harv. L. Rev. 1357 (1985)
Many decision-theory modes suggest that factfinders should base their decisions on the laws of probability in order to minimize the costs of erroneous judicial decisions. These models ignore the judicial function of generating acceptable verdicts which reflect and project substantive legal rules.... A court must generate an acceptable account of what actually happened as a predicate to imposing a sanction for violation of a substantive legal rule.... Many procedural and structural mechanisms of the legal system serve to enhance the acceptability of judicial verdicts.... The goal of generating acceptable verdicts is not met simply by choosing the verdict that is most probably accurate. Acceptable verdicts and probable verdicts might appear to coincide, given that one obvious way to gain public acceptance is to search for truth. But the correlation between probability and acceptability is not exact: a probable verdict may not be acceptable, and an acceptable verdict may not be probable.
Cases of naked statistical proof present the most provocative example of probable verdicts that are unacceptable. In these cases, the evidence suggests a sufficiently high numerical probability of liability, but the absence of deference-inducing mechanisms in the judicial process is such that the public is unable to view a verdict against the defendant as a statement about what actually happened. The statistical nature of the evidence precludes both acceptance of the verdict against the defendant and internalization of the underlying norms.
Decision theorists have tried, with great difficulty, to accommodate the blue bus case in their models. The logic of the standard decision-theory model holds that the plaintiff is entitled to win because he has shown, much more probably than not, that a bus owned by the Blue Bus Company ran him off the road. But most decision theorists have suggested that courts should grant a directed verdict to the defendant. The problem for the decision theorists, then, has been to explain their answer to the blue bus hypothetical without having to abandon their theory.
Professor Tribe has attempted to solve the decision theorists' problem. He has asserted that verdicts are based on subjective probability assessments. Although the plaintiff's objective proof indicates an 80% likelihood that the defendant's bus caused him to be injured, a juror is not bound to believe this probability. The very tenuousness of the plaintiff's evidence, Tribe argues, may cause the juror to feel some skepticism about the plaintiff's case. This skepticism may be enough to reduce the juror's subjective probability assessment of the defendant's liability to less than 50%. The juror presumably believes that if the defendant's bus really had forced the plaintiff off the road, the plaintiff would have had better proof; all told, the odds are less than fifty-fifty that it was the defendant's bus.
Tribe's argument explains how a juror might find against the plaintiff, but the actual cases involving the blue bus hypothetical do not pose that problem. Plaintiffs in such cases would almost certainly lose by directed verdict; the evidence would never reach the jury. Tribe's argument explains why a court should refuse to grant a directed verdict to the plaintiff, but his analysis does not explain why the judge should throw the plaintiff out of court. The jurors could arrive at a subjective probability higher than 50%; no objective evidence compels a juror to drop his subjective probability so drastically that the plaintiff must lose. The logic of Tribe's argument leads to the conclusion that the case should reach the jury, and the jury's verdict should be upheld, no matter which way it comes out.
Other commentators have rationalized granting a directed verdict against the plaintiff on the ground that any other result would impose too large a burden on the defendant. If a court held the Blue Bus Company liable in this case, courts would have to hold the company liable in all similar cases, even though it was responsible for only 80% of them. Some decision theorists consider this result unfair to the defendant and to all similarly situated plaintiffswould dislocate the market: it would disproportionately burden defendants like the Blue Bus Company and subsidize their smaller competitors. Because these competitors would then have little incentive to drive carefully, accident rates would increase. Simple application of a more-probable-than-not rule would not maximize utility. Posner would thus apply a special rule of proof in cases like the blue bus case, a requirement that the plaintiff offer some evidence of liability in addition to the statistical information. In effect, Posner's call for additional evidence reflects a need for a judicial mechanism that will induce deference to the jury's decision and thus promote an acceptable verdict.
Proportionate Damages.--One obvious alternative to a directed verdict for the defendant in the blue bus case would be to make the Blue Bus Company pay 80% of the plaintiff's damages. This solution responds to both the economic and fairness arguments advanced by decision theorists. It requires the company to pay in total (in this and all similar cases) for no more than the damage it probably caused over the long run. Forcing the Blue Bus Company to pay 80% of the damages in all such cases might still cause economic dislocation if the company's competitors had to pay nothing. This dislocation could be avoided, however, by requiring competitors to pay their proportionate shares as well. If there were a 20% probability that a defendant caused the plaintiff's damages, then the defendant would pay 20%. The transaction costs of litigating claims might suggest some minimal percentage (or dollar) threshold for liability, but it would surely not be as high as 50%. Requiring the Blue Bus Company's competitors to pay their fair share would give the plaintiff full recovery and properly balance the relative economic impact of the damage awards on the bus industry.
This proportionate-award approach could be applied to all cases in which the jury is uncertain of the facts on which a defendant's liability is predicated, including cases not based on statistical evidence. If the jury, after hearing a mass of conflicting evidence in a negligence case, is 40% certain that the defendant acted negligently, then the proportionate award approach would require the defendant to pay an award of 40% of the total actual damages, instead of paying nothing.
The proportionate-award approach addresses the concerns of the decision theorists so well that a question arises as to why our legal system is so firmly committed to the all-or-nothing rule. The answer is that a proportionate award projects a substantive legal rule that differs from the rule projected by an award of full damages. The former might project a behavioral message that differs from the message the court would convey by the standard application of the rule. Courts have resisted the proportionate-award approach because in most cases they have not desired to project the new legal rule and the new behavioral message that would accompany such an award.
The blue bus case illustrates this proposition. A court faced with such a case might choose to apportion damages according to the probability that the bus company injured the plaintiff. But a proportionate award against the company would not exemplify the basic negligence rule and its behavioral norm. The award would instead project a very different and seemingly perverse legal rule: courts will hold the defendant liable, notwithstanding the possibility that he committed no wrong, if the nature of his activity places him within a class of suspect persons. The behavioral norm embodied in this rule is not one of care and safety. In response to the rule, the Blue Bus Company might attempt to minimize exposure to liability by running fewer buses, rather than by trying to drive more carefully; the verdict sends a message about the volume of business, rather than one about safety standards. Indeed, proportionate awards of this kind make safety precautions more difficult to justify in economic terms. Thus, judicial hesitation to award proportionate damages in cases like the blue bus case may well arise, in part, from a reluctance to project the legal rule and behavioral message that would accompany such an award.
Courts have awarded proportionate damages, however, when such awards would convey desirable behavioral norms. In Summers v. Tice, for example, the plaintiff sued two hunters who had negligently fired in his direction. The court held that in the absence of any evidence as to which hunter had actually shot the plaintiff, the hunters were jointly liable for the plaintiff's injury. Similarly, in Sindell v. Abbott Laboratories, plaintiffs brought a class action against manufacturers of the drug DES. The court held that in the absence of evidence as to which manufacturers had made the product that caused the plaintiffs' injuries, each manufacturer was liable for the proportion of the judgment represented by its share of the DES market.
These cases are distinguishable from the blue bus hypothetical. In the hypothetical, only one bus company was negligent, and there is no basis for determining which one. A verdict that the defendant is liable would project the substantive rule that one need not act negligently to be liable. In Summers and Sindell, all of the defendants acted negligently, even though one cannot know which defendant actually caused the injury. The legal rule that these cases project is that when we do not know the identity of the person who caused an injury, we will award damages against all negligent parties in proportion to their probable responsibility for the specific harm. The rule and its behavioral message seem sensible enough. In Summers and Sindell, the courts imposed liability because the defendants acted wrongly, the courts thereby projected a message about the result of wrongful activity....
The outcomes of the cases turn on substantive issues that relate to the effects of generating new legal rules. In Summers and Sindell, the courts generated a new definition of what was relevant to a finding of liability and thus generated a new rule of substantive law. Whether the courts generated a good rule depends on one's assessment of the rule and of the process of judicial lawmaking. Casting these issues as problems of proof serves only to obscure them. The cases concern changing the elements of the substantive legal rule; the problem of proof is simply that of generating acceptable conclusions about those elements.
Although the traditional logic of proof rules can inhibit judicial efforts to find liability when the evidence is merely statistical, we should recognize that courts can nevertheless find liability and generate new substantive law by redefining the elements of the legal rule and the sanction so they reflect the statistical nature of the evidence. The current reluctance of the judicial system to impose liability in such situations cannot be overcome by changing the grammar of proof, because this grammar is essential to achieving the projection and affirmation of the law's behavioral norms. Instead, reform must come, and should be welcomed, by bringing about changes in the factual elements that must be proved. An instinctive reaction against probabilistic proof should not constrain efforts to restructure substantive law by changing the rules that govern what must be proved.