Trial by Market
Robert Frost once wrote, “Modern poets talk against business, poor things, but all of us write for money. Beginners are subjected to trial by market.”19 Frost, of course, was not claiming that beginning poets literally were put on trial but that market forces would provide a metaphorical trial of their work. The quotation emphasizes that though we try to avoid market forces, they are inescapable. And so it is also with civil litigation. The civil litigation process is market-driven, with litigants generally trying to gain or keep as much money as possible and lawyers seeking to make money, too. It is a cardinal principal of judicial ethics, however, that the judge cannot have any direct financial interest in a litigation over which he or she presides.20 Judicial income is largely independent of the quality of a judge’s decisions, although some lower court judges might hope that writing good decisions (whatever those are) might lead to a promotion.21 Perhaps removing financial incentives from judging frees judges to decide according to the rule of law. And yet the reverse may also occur. Because there is no financial sanction for a bad decision, judges are free to ignore the rule of law in favor of their own policy preferences. Indeed, one might wonder whether providing direct financial incentives would improve judicial performance.22 A relatively simple (albeit radical) means of accomplishing that aim would be to select randomly a subset of a judge’s decisions, assign them to another judge selected at random (perhaps from an appellate pool), and give the initial judge a bonus if the second judge agrees with the first judge’s conclusion. Judges’ identities might be kept from one another to reduce the danger of collusion. This system would tend to discourage judges from being idiosyncratic by, for example, allowing their personal ideology to influence their decision making. Some might argue that idiosyncratic decision making is beneficial (see Chapter 6), but in my view the judicial system ordinarily should prefer to adopt a view that most judges would hold to than a view that only a few would hold. Of course, the second-level judges might still decide idiosyncratically, but the first-level judges would not know in advance if a second judge would evaluate the case, and if so who would be chosen. Prediction market adjudication takes the logic a step farther. Nudging judges in the direction of deciding cases as other judges would is pushing them to make predictive decisions. Prediction markets can do this as well, and they can provide more than a nudge. Market participants may or may not care about how the case is resolved, but the provision of financial incentives will ensure that individual sympathies have little to no effect on market predictions. Even in a system that provides judges with financial incentives, a judge could decide that achieving a certain result is more valuable than the lost financial remuneration. With prediction markets, anyone trying to manipulate the market to achieve a particular result will need to overcome other market participants seeking to identify and bet against such attempts at manipulation. Attempts at manipulation not only may be expensive but also in general may not succeed. This does not necessarily mean that the prediction market’s outcomes will be heartless. The market, for better or for worse, will still be forecasting the result of traditional adjudication, involving trial judges, juries, and appellate judges. But the market should help to avoid idiosyncratic outcomes. And thus would “trial by market” become more than a metaphor. We could, of course, imagine making some modifications to trial by market. For example, we might limit participation to individuals selected through some sort of political process, such as that currently used to select judges. Further, we might allow only a small group of individuals to participate, perhaps one for trial proceedings or three for appellate proceedings. We might decide to ensure that each of these individuals has the same amount of money with which to bet on each case, prohibiting market participants from investing more in cases in which they have greater confidence. And we might require that all bets be placed at a single time, with the outcome simply the one that receives the most bets. Then we might decide not to use real money but to use play money instead, giving each of the authorized participants a play dollar to put on a particular outcome. Finally, we might call the play dollar a “vote” and the participants “judges,” and we might recognize that we are left with something akin to the existing judicial system. In each stage of this hypothetical transition from trial by market back to our current adjudicative world, a step is made that seems likely to reduce the incentives of the market participants individually and collectively to put aside their own views. Prediction markets may provide the most reliable available means of gauging how judges on average would apply the law to a particular set of facts. Arguments against trial by market thus seem likely to fall primarily into three groups: first, that prediction markets are not in fact sufficiently accurate; second, that idiosyncratic decision making is beneficial; and third, that market-based adjudication would undermine public confidence in law. All of these arguments might have merit given existing knowledge about and acceptance of prediction markets, at least for some kinds of cases. But the benefits of increasing consistency and information aggregation are sufficiently great that trial by market might prove viable in some collection of cases. There are other benefits of trial by market. One follows directly from consistency: The practice might greatly reduce frivolous lawsuits. Suppose that a plaintiff believes that there is a 10 percent chance that a lawsuit would be successful in generating a particular level of damages. Despite concerns about litigation cost, a defendant would have an incentive to settle this suit at about 10 percent of the damages. Some have an intuition that this seems like a fair result, but there is a strong argument that this is not socially desirable. The percentages reflect the fact that the vast majority of decision makers believe that the plaintiff should not recover anything. If we adopt the assumption that the majority is more likely to be correct than the minority, at least more often than not, a legal system that insists on zero damages in such cases should be judged as more accurate than a legal system that allows ten percent damages. The system of trial by market could impose no damages when the legal system determines that less than 50 percent of cases would result in liability for the defendant and full damages when more than 50 percent of cases would result in liability. The result of this arrangement would be that there would be little incentive, placing litigation costs aside, for defendants to settle nuisance suits or for plaintiffs in meritorious suits to accept less than full damages. The argument is borrowed from John Coons, who sought to justify our legal system’s preponderance-of-the-evidence standard.23 This standard results in imposition of full damages when a defendant is more likely than not responsible for a plaintiff’s injuries and no damages when a defendant is less than 50 percent likely. An alternative regime would provide damages in proportion to the probability of liability. In a case in which the probability of liability is 90 percent, the choice between the preponderance regime and proportional damages is a choice about whether, given that a plaintiff will be awarded 90 percent of damages, the plaintiff will receive also the additional 10 percent of damages. There is a 90 percent chance that the plaintiff in fact should receive that 10 percent, too, and so the preponderance regime is superior to proportional damages. The argument is symmetric on the opposite side of the probability spectrum, and it can apply not just to the question of what damages should be awarded at trial but also to the question of whether the legal system should seek to discourage settlement of nuisance cases. Of course, trial by market could insist on proportional damages, and perhaps it should near the middle of the probability spectrum,24 but a significant advantage of the system is that it could facilitate reduction of proportional settlements. Trial by market itself would require a substantial subsidy in order to be successful, but there are several means by which it could reduce litigation costs. The continuous evaluation of evidence presented to a prediction market would facilitate negotiation. Ordinarily, litigants must anticipate how the decision makers (judge, jury, and appellate judges) will rule once all of the evidence has been provided. With a prediction market serving as adjudicator, litigants would be able to assess the adjudicator’s current evaluation of all available evidence. There might still be situations in which litigants will not be able to agree on the value of a legal claim because, for example, they disagree about the effect that subsequent evidence might have on the prediction. Such inherent disagreement is likely, however, to be considerably rarer. The continuous prediction, meanwhile, would make it harder for litigants to bluff about their views of the claim’s value in an effort to keep as much of the surplus from settlement as possible. The prediction market approach to adjudication thus can help eliminate two primary causes of failed settlement negotiations: asymmetric claim valuation and strategic bargaining. The continuous evaluation of evidence also might limit costs by helping reduce excessive investment in litigation preparation. For example, a rule might provide for the automatic end of the case when the probability that a particular party would win fell below some threshold, such as 10 percent. Somewhat more elaborately, a conditional prediction market might be used to assess the probability, conditional on the allowing of further evidence, that the market’s probability assessment will cross the 50 percent threshold. If there is only a low probability that any future evidence could change the market’s prediction of what a majority of decision makers would decide, then that suggests that further presentation will not do any good. Such a prediction market might lead litigants voluntarily to give up losing causes, although a rule automatically ending a case past some level of futility could still be useful, because litigants generally wish to pursue a cause longer than would be socially optimal. As a less radical example, the initial prediction market assessment might be used as a means of allocating fee-shifting. On the basis of the market assessment at some early point, the party that is expected to lose might be allowed to proceed only if the party agrees to assume the adversary’s costs. To prevent this arrangement from leading the nonpaying party to generate excessive costs, the rule could limit fee-shifting to the amount spent by the paying party on its own costs.25 This approach would alleviate a significant drawback of fee-shifting rules, namely, that they may tend to decrease settlement.26 Because trial will occur anyway only where litigants are mutually optimistic, fee-shifting makes trial seem more attractive to both sets of litigants in cases in which trial is possible. A regime of ex ante fee shifting does not produce this problem, because the trial outcome has no effect on who pays the fees.27 Ex ante fee-shifting based on prediction markets thus may be a more promising approach to eliminating both the filing of frivolous lawsuits and frivolous defenses of meritorious lawsuits. Finally, trial by market also could alleviate jurisdictional wrangling. If the prediction market is anticipating the outcome of traditional adjudication, then it can simply weigh the probabilities that this traditional adjudication might be held in various fora, such as particular state, federal, or international courts. In the existing system of adjudication, the location where the lawsuit occurs must be decided early in the litigation process, so that a judge can be selected to supervise the remainder of the adjudication. Even in cases that are likely eventually to settle, the adjudication of the jurisdictional questions may occur while negotiations are in progress and can be expensive. With trial by market, the probability that a case might end up in different jurisdictions would simply be priced into the market’s ultimate assessment of the probability of victory. The same holds for disputes about which substantive law should apply. Of course, there might still be considerable disputes about whether individual cases should be tried by market or by courts in the first place.
Leave a Reply