A possible view of the role of an adjudicator is that it is to obtain the information that is needed to apply a well-articulated legal rule. That is, the task of an adjudicator in a case is to gather and verify the information that is called for to employ a legal rule, but that once the required information is in hand, the adjudicator's work is done, for the rule then fully determines the outcome. This view is not, of course, descriptive of reality. Adjudicators (as the word "adjudicate" itself implies) typically enjoy discretion, both in their ultimate decisions and in their conduct of litigation: the legal rules that adjudicators apply usually do not dictate their decisions; rather, adjudicators tend to possess some freedom of authority in making decisions. And this is true not only of adjudicators in our formal legal system, it holds also for individuals playing adjudicative roles outside of that domain. For instance, when supervisors make promotion decisions about employees or when college admissions officers make acceptance decisions about applicants, they typically possess some discretionary power. It thus seems to be a fundamental aspect of our legal and other rule-based systems that adjudicators frequently enjoy a measure of discretionary sway. The object of this article is to explain why permitting discretion may be valuable to society (or to an organization) and how the problems associated with allowing discretion can be ameliorated. The kernel of the rationale for granting discretion to adjudicators is familiar: discretion allows adjudicators to make decisions on the basis of information that is not included in the rule (such as the degree of remorse shown by a criminal defendant). A basic question that arises in relation to this explanation is why the information that it is hoped adjudicators with discretion will use properly in their decisions is not included in the legal rule in the first place (thereby eliminating the need for discretion). Two distinct answers are offered. One is that the information (such as a defendant's remorsefulness) would not be easy for an appeals court (more generally, a supervening tribunal) to verify, making its inclusion in the legal rule unworkable. The other is that the information might pertain to conditions that are too particular to have been well described and incorporated into the rule by lawmakers, given the costs and the inherent difficulty that they would face in fashioning highly detailed rules (how would a lawmaker describe degree of remorsefulness?). Hence, it seems inevitable that there will be information that, on one hand, is not included in legal rules, but that, on the other hand, society would in principle wish to be taken into account in decisions. This is why there is a potential value in permitting discretion. The article first examines the tradeoff between the advantage of allowing discretion and the disadvantage of so doing, namely, that adjudicators may not use the information as society wants but rather to further their own objectives (perhaps adjudicators favor more lenient punishment than does society) or as they incorrectly believe society wants. After this tradeoff and the conditions under which discretion is desirable to allow, ways of controlling discretionary deviation are investigated. A direct method that is noted is to limit the extent of discretion (for instance, stipulate that a sentence for a crime be in the range between 1 and 5 years). Then three major, indirect methods of controlling discretionary deviation are examined. The first is changing the incentives of adjudicators, by giving them "decision-based" incentives: a reward or penalty based on their decisions (such as enhanced promotion possibilities for imposing higher sentences). This can counter discretionary deviation, but not perfectly. A second means of controlling discretion is through monitoring of adjudication itself (such as by having spotchecks of proceedings); a threat of punishment for discretionary deviation can induce better decisionmaking. A third method of control is the appeals process, under which disappointed litigants, who see that deviations from desirable use of discretion have occurred, bring appeals. Like monitoring, the prospect of appeal can induce better decisionmaking by adjudicators. The foregoing points are developed informally in Section 2 and formally in Section 3; Section 4 concludes. Much has previously been written on the general subject of discretion by legal commentators. Many legal writers discuss the basic point that discretion allows additional information to be employed, yet has the drawback that it may not be used desirably. However, the question of why the law does not explicitly include the variables that adjudicators are implicitly relied upon to incorporate into their decisions (that is, if remorse is important to take into account, why does not the law do that in the first place? why rely on adjudicators?) does not seem to have been emphasized before. Also, the development of an explicit model of discretion seems to me to be clarifying. In any case, I am not aware of systematic discussion of decision-based incentives or of monitoring as remedies for discretionary deviation. The general point that the appeals process can curb discretionary deviation is well-appreciated; but how it does so, and the limits of the appeals process in this regard (having to do with, as will be discussed, the information that the appeals tribunal can obtain) may be of interest to the reader. Another body of literature of relevance to the subject here is the principal-agent literature, for the adjudicator may be considered an agent of society. I note the relationship of this literature to the model of discretion in Section 3F.
