Why judges are imperfect rule-makers

The rule model of common law decision-making treats rules announced in precedent cases as authoritative in later decision­making. When rules are misconceived or obsolete, judges may overrule them.

The rule model is defensible only insofar as precedent rules are good rules. A good rule need not prescribe the morally best outcome in every case to which it applies; it is enough that, if regularly followed, the rule will prevent more error than it causes. When a rule fails to meet this test, judges should overrule it and return to unconstrained moral and empirical reasoning.

7.1.1 The salience of a particular case

A significant difficulty of the rule model of decision-making, as applied to the common law, is that judges are not focused single-mindedly on the task of designing good rules. Instead, they must combine rule­making with determining the outcome of a particular case. These two responsibilities are not always compatible.

According to the traditional understanding of judicial decision­making, the common law developed through the courts, but the courts did not make this law. Instead, the common law was understood as a body of norms located in custom and sound reason rather than judi­cial opinions. Judges were understood to be learned in custom and reason, and consequently their statements and decisions served as evidence of common law. But they had no personal authority to make law by announcing rules: instead their role, and their special skill, was to discover and apply the law that was immanent in social practice (see Tubbs 2000; Hale 1713; Simpson 1973; Blackstone 1765; Postema 2003). This understanding of judicial decision-making did not deter early courts from developing a comprehensive body of law, but it pre­vented them from acknowledging lawmaking as part of their role.

Modern judges recognize that the opinions they write in explanation of their decisions affect both future decisions and the conduct of pri­vate citizens. They are quicker to admit that they can and do create law, and some are quite explicit in announcing rules (see Posner 2006; Tiersma 2005). Nevertheless, rule-making is a secondary concern for most courts; their primary task is to consider and resolve a particular dispute. As a result, they may pay less attention than legislatures pay to the full range of consequences their rules might bring about.

For example, a judge deciding a nuisance case involving a pet bear might state that ‘wild animals are nuisances per se when kept in resi­dential neighborhoods'. The image in the judge's mind is a large animal such as a bear, and with this picture in mind, she may fail to realize that the rule as stated covers smaller, less-threatening animals, such as pet rats or ferrets or maybe an uncontrolled population of squirrels. The result is an overly broad nuisance rule. It might be possible to counter­act any damage resulting from the judge's statement by assuming, as a matter of interpretation, that the judge did not intend an authoritative rule, but only an explanation of her decision. If so, no harm is done. But if the judge clearly characterized her statement as a rule intended to govern future cases, the result is an authoritative but suboptimal rule, skewed by the judge's focus on one particular set of facts.

This is not to say that legislators are ideal rule-makers. For a variety of reasons, including their allegiances to particular constituencies and their need to compromise, legislatures are also capable of enacting bad rules. Enacting rules, however, is their primary task, so they are more likely than judges and better situated than judges to take into account the future consequences of their rules.

7.1.2 Cognitive bias

A second, closely related problem with judges as rule-makers is that, even if the judge is fully aware that she is issuing a rule that will govern a range of future cases, irrelevant features of the immediate case may affect her judgment about the likely consequences of the rule. Cognitive psychologists have shown that human reasoners commonly rely on a variety of assumptions, mental habits, and biases that affect their moral and empirical assessments (see Gilovich, Griffin and Kahneman 2002;

Tversky and Kahneman 1982; Symposium 2002). Often, these biases are useful shortcuts that allow reasoners to move forward under condi­tions of uncertainty. Yet, because they replace fully reflective reasoning with simple, indirect strategies for reaching decisions, they also can lead reasoners into error.

Judges are not immune from these effects. Their education and professional obligations may lead them to exercise care but will not completely insulate them from cognitive errors. Researchers inter­ested specifically in judicial behavior have identified biases, decisional strategies, and cognitive shortcuts that lead judges to err in the course of fact-finding and other aspects of the trial and appellate process (see, for example, Rachlinski and Wistrich 2017; Wistrich, Rachlinski, and Guthrie 2015; Sharfinan 2005; Rachlinski 2006; Guthrie, Wistrich, and Rachlinski 2001). Cognitive biases of this type can affect not only the outcomes of litigation but also the design of judicial rules. When the facts of a particular case are prominent in a rule-maker's mind, they may lead the rule-maker to miscalculate the overall future effects of the rule (see Schauer 2006; Alexander and Sherwin 2001). Concrete facts provide an example of how the rule will operate in action, but they may also distract the rule-maker from consideration of other, less desirable consequences of the language used in the rule.

Suppose, for example, that a judge is presiding over a nuisance case involving a pit bull that attacked a small child without provocation.

Another cognitive bias that may influence judges in their dual capac­ity as rule-makers and adjudicators is ‘affect', meaning the emotional impact of particular facts or images on decision-making (see Rachlinski and Wistrich 2017). As a cognitive bias, affect manifests itself in a number of ways. The most important of these in the context of judicial rule-making is that facts or images that provoke an emotional response tend to carry more weight in decision-making than emotionally neu­tral facts. Emotional responses to facts play a useful role in human life: triggers such as fear or sympathy help people respond quickly to threats to themselves or others. But a judge's emotional reaction to the facts of a particular case may interfere with neutral, reasoned rule-making. Evidence that the defendant's pit bull has mauled a child may lead the presiding judge to announce a stricter nuisance rule than statistics about the probability of dog-inflicted harm would otherwise support.

Another cognitive mechanism relevant to judicial rule-making is ‘anchoring' (Chapman and Johnson 2002). Anchoring occurs when a decision-maker must give numerical answers to questions about matters such as value or probability. Cognitive research indicates that recent exposure to a number - even a completely irrelevant number - can have significant effects on the decision-maker's conclusions, and that judges are as susceptible as others to this phenomenon (Rachlinski, Wistrich, and Guthrie 2015). Thus, if the defendant in a nuisance case owns three pit bulls, and two of them have attacked children, then if overall statistics on pit bull attacks are at all mixed, the judge is likely to be drawn to a high rather than a low estimate of the likelihood of pit bull violence in residential neighborhoods.

There are other cognitive biases that may affect judges as rule-makers. Reasoners whose attention is focused on a single observation may con­fuse statistical questions about the likelihood of similar events with questions about the likelihood of the particular event. Thus, a judge who has heard evidence describing an attack by a particular pit bull and then is called on to assess the likelihood of pit bull attacks may think in terms of how likely this dog was to attack rather than how likely similar dogs are to attack. Familiarity with a single event may also lead reasoners to underestimate the likelihood of other, similar but non-identical, events. A judge who has heard evidence of a pit bull

attack may underestimate the likelihood of an Airedale attack. Judges familiar with particular evidence may also make attribution errors, attributing responsibility to personal traits of those involved rather than background conditions affecting their behavior.

The process of adjudicating particular cases can also have positive effects on judicial rule-making. A ready example may sometimes clarify judicial decision-making about a more general problem. The fact that a judge's attention is focused on one set of facts rather than an issue posed in abstract terms may also lead the judge to announce a nar­rower rule that will do less damage if it turns out to be misconceived. Overall, however, it seems likely that the special salience of a particular dispute pending before the judge is likely to interfere with, rather than sharpen, the type of reasoning needed to draft a sound general rule for a category of future cases.

7.1.3 When to overrule a precedent rule

Under the rule model of the common law, judges have power not only to make authoritative rules but also to overrule rules established in prior cases. The rule model does not distinguish between overruling a precedent rule and modifying or distinguishing a precedent rule. In either case, the effect of the current action is to eliminate the prior rule and replace it with a new rule.

Ideally, judges should overrule precedent rules only when they are unjustified, or could be improved upon, as rules. The fact that a rule will produce a bad result in the case now before the court is not a reason to overrule if, over time, the rule will prevent more error - by correcting reasoning mistakes, settling moral controversy, or support­ing coordination - than it causes. Overruling (or modifying, which entails overruling) is appropriate in just two circumstances. If, due to poor design or obsolescence, the rule is likely to cause more error than it prevents over the range of cases to which it applies, the rule is unjustified and should be eliminated. Alternatively, if a different rule would prevent more error than the existing rule, and the benefits of the new rule exceed the loss of settlement or coordination value that would follow from a change, then the existing rule should be replaced. Otherwise, the current court should treat the rule as authoritative and follow it according to its terms.

Frederick Schauer has pointed out that judges are likely to have dif­ficulty in overruling precedent rules when and only when they should (Schauer 2006). One problem is a conflict built into the overruling standard itself. As adjudicators, judges must follow authoritative prec­edent rules even when they produce morally unsatisfactory results in a particular case. At the same time, as potential makers and overrulers of rules, judges must evaluate the overall justification of existing rules before deciding whether to retain them or eliminate them. Reconciling these tasks is not easy. When judges fail to distinguish between bad outcomes and unjustified rules, they may jeopardize the values of set­tlement and coordination in order to accommodate the ‘equities' of particular cases; or, alternatively, they may forgo the opportunity to eliminate a defective rule.

The problem of overruling precedent rules too quickly in response to compelling but unrepresentative facts is a problem of ordinary human sympathy and desire for fairness, coupled with the type of cognitive shortcuts we have already discussed at some length. Suppose a judge has before her a nuisance case against the owner of a well-trained, calm, and friendly pit bull. The owner is an older person who relies on the dog to help her get around. A previous nuisance case in the same jurisdiction established the rule that ‘pit bulls in residential neighbor­hoods are nuisances per se'. Assume this is, overall, a good rule, one that will prevent more error than it causes over time. Yet, when the current judge evaluates the rule, the salient image of the defendant's good-tempered dog, and the judge's sympathy for the defendant, may distract her from full and fair consideration of background statistics about pit bull attacks and other justifications for the precedent rule. As a result, she may overrule the rule or replace it with a modified rule that makes an exception for good-tempered pit bulls. If the existing rule is, overall, the best rule for most pit bull disputes and serves the ends of settlement and coordination, then a response of this type is mistaken even if the immediate result is morally and practically satisfying.

The opposite problem, of retaining rules when they are no longer good rules, or when another rule would be better, is exacerbated in judicial settings by selection effects (Schauer 2006). Judges confront precedent rules only when they are asked to apply them to particular cases. When the facts of a case fall squarely within the terms of an existing rule, parties are likely to settle their disputes rather than take them to court. As a result, judges will rarely confront cases in which the rule applies unambiguously to the facts. They are more likely to hear cases in which the application of the rule to the facts is in doubt, but in these cases the judge can escape the rule through interpreting it narrowly and thus avoid the need to overrule it. For that reason, obsolete or misconceived precedent rules may never receive serious judicial attention, and so may linger on the books. When a set of facts arises in which the out­come of a rule is particularly harsh, an affected party may be motivated to challenge the rule. This solution is not ideal, however, because, as we have noted, a harsh outcome may not indicate that the rule itself is faulty.

7.1.4 Summary: weaknesses of judicial rule-making

The rule model of the common law casts judges as both adjudicators and rule-makers, and the circumstances of judicial decision-making place emphasis on the adjudicatory role. The dual role of judges in the legal system is likely to affect the quality of judicial rules in a number of ways. The demands of adjudication and common law traditions that relegate rule-making to a secondary position may lead judges to pay less attention than they ideally should to the potential consequences of the precedent rules they announce. Cognitive effects triggered by judges' attention to particular facts can aggravate the problem by caus­ing judges to miscalculate, or fail to attend to, the full consequences of judicial rules. Adjudication may also have adverse effects on judicial oversight of judicial rules over time.

Despite these difficulties, we do not assume that legislation is always or obviously superior to the common law as a source of settlement and coordination within a society. Moral and empirical deliberation by elected representatives is notoriously subject to interest group politics and collective action problems in addition to cognitive biases and ordinary reasoning problems. We do not undertake to compare judicial and legislative rule-making; nor do we comment on collective decision-making by judicial panels. We believe, however, that Schauer is correct when he observes that the need to resolve a particular case tends to hinder rather than assist judges in their role as rule-makers. Not only hard cases, but more generally the need to settle disputes, can result in bad law (Schauer 2006).