Review of Administrative Agency Legal Decisions
The hard-look doctrine represents only one form of judicial review that courts perform in assessing the decisions of administrative agencies. In addition to considering whether the agency’s decisions are supportable as a policy matter, a court reviews the agency’s factual determinations,22 as well as whether the agency’s decision is legally permissible. This second inquiry is necessary because agencies are constrained by statutes passed by Congress, although statutes often contain a great degree of ambiguity. According to the Chevron doctrine,23 the courts defer to any reasonable construction of an ambiguous statute by the agency that administers it. Courts, however, do not accept an agency’s claim that a statute is ambiguous at face value. They use a wide range of legal tools to try to determine whether the statute is ambiguous. Consider, for example, the Supreme Court case Babbitt v. Sweet Home Chapter of Communities for a Great It is perhaps not surprising that the split was along ideological lines, with the Justices regarded as generally more conservative in the dissent. The task of determining whether a statute is ambiguous is not inherently more objective than the task of determining what the correct meaning of a statute is. Presumably, the resolution of any particular interpretive issue falls on a continuum, with unambiguous resolution one way occupying one end, ambiguity in the middle, and unambiguous resolution the other way occupying the other end. Chevron requires the courts to determine whether an issue falls in the amorphous middle of this continuum. De novo review, the approach that the courts used with issues identified as purely legal before Chevron,26 required the courts simply to determine on which side of the continuum a particular issue fell. It is not obvious that the Chevron approach reduces the number of cases in which interpreters will disagree. Even if it does, there will remain some cases in which interpreters will generally agree that X is the better interpretation but defer to the agency’s choice of Y based on a conclusion that there is sufficient ambiguity in the statute. Because statutory interpretation is in part subjective, a normative prediction market may be a useful way of gauging how decision makers, on average, would approach a particular question of statutory interpretation. In such a market, the ex post decision maker might be asked to determine whether the statute is somewhat ambiguous. A superior approach, however, might be simply to ask the decision maker which interpretation is the better one.27 Though it is the job of courts to determine whether a statute is ambiguous, this approach should provide a sound guide, identifying cases in which there is significant disagreement. In existing Chevron cases, judges are effectively predicting whether different interpreters of the statute would agree sufficiently often to justify a conclusion that the statute is clear. A normative prediction market can make this type of prediction in a more rigorous way. Of course, prediction markets might help not only when judges are assessing agency statutory interpretations but also when agencies are initially engaging in interpretation. Ideally, agencies routinely would create normative prediction markets when faced with legal ambiguity. The primary incentive of agencies to do this would be to persuade the courts that they are not engaging in excessively creative statutory interpretation. There is at least one body that generally could benefit from statutory interpretation markets: the legislature. Individual legislators sometimes might prefer creative to plausible interpretations of particular statutory issues, but in general the legislature will prefer that agencies and courts not usurp its power in cases in which it did not intend to delegate that power. If normative prediction markets develop a reputation for accuracy, the legislature someday might require them. The closer the ex post decision makers are to the original drafters in outlook, the better normative prediction markets will reflect the interpretations that the original drafters would offer. If, for example, the normative prediction market will forecast the future assessment of the administrator of the Environmental Protection Agency, and it is predictable that the administrator will on average be more liberal than the legislature, then the market predictions will not generally reflect what the legislature intended. One possibility is to structure the market so that it predicts an assessment by a randomly selected member of the legislative body. Despite change over time, a legislature might wish prediction markets to forecast its own decisions rather than those of administrative officials or judges. Such a proposal might seem to duplicate one that we have already seen: the variation I suggested to Einer Elhauge’s proposal that courts seek to resolve statutory ambiguities according to current legislative preferences (see Chapter 5). There are, however, two key differences. First, instead of predicting what the entire legislature will do, conditional on the legislature’s making a decision, this proposal is to predict what a single randomly selected member of the legislature will do. At least in a unicameral, majoritarian legislature, this should amount to the same thing, given the assumption that the legislature’s output will reflect the views of its median member, but it may demand less legislative time. Second, the market would not simply be predicting the current legislature’s general political preferences. Rather, it would be predicting how legislators would engage in interpretation of the earlier legislature’s views. This approach therefore might be more palatable to someone who believes that statutory ambiguities should be resolved with reference to the enacting and not the current legislature. The possibility of using prediction markets to discipline statutory interpretation emphasizes that normative prediction markets need not assess open-ended policy reasoning. Statutory interpretation involves normative concerns that differ from those of the hard-look doctrine, and separate prediction markets could be used to perform the two tasks in the same case. Central to statutory interpretation is the premise that adherence to legislative intent ultimately advances rule-of-law values, but there is always a danger in a particular case that it will be more appealing for an administrative agency or a judge to advance a salient policy objective at some very small cost to the rule of law. This is less of a danger with normative prediction markets, because the ultimate decision being predicted has no direct effect on policy; it affects only the payouts of market participants. Making an ideological decision would be an empty gesture, but attempting genuine interpretation would improve overall confidence in the system. Even if ideological decision making still occurs, there will generally be little reason ex ante to think that it will be biased in one direction rather than in another.
Leave a Reply