Albeit quite different in their respective roles and organizational structures, Posner (Reference Posner2010) ‘From the new institutional economics to organization economics’ enlists the judiciary of common law nations, such as the United States, and the judiciary of civil law nations, such as France, Germany, and Japan as two types of organizations that have been more successful than other public institutions in overcoming obstacles to organizational efficiency.Footnote ¹ Posner points out that the civil law judiciary is characterized by less independence of individual judges, less creativity, and a larger number of judges relative to lawyers. He explains the lack of creativity as a result of the existence of coherent codifications and as a result of the hierarchical structure of the judiciary, which reduces the freedom of individual judges. To the extent that promotion criteria are based on nonpolitical criteria, such as accuracy, timeliness, and reversal rate of decisions, independence and quality would be adequately secured by hierarchical control. However, in Posner's view, in a career judiciary the ‘desire for promotion places one at the mercy of one's superiors’. This form of hierarchical control of a career judiciary is likely to induce judges to decide in a methodologically conservative and unadventurous fashion. As Posner's puts it, a civil law judge who has bold views does not help his career.

In the following, we consider the different doctrines of precedent adopted by common law and civil law jurisdictions to explain the different impact of judicial precedents and the different incentives faced by experimentally minded judges to innovate the law. In Section 1, we develop a model to consider the effects of alternative doctrines of precedent on the social costs arising from judicial error and uncertainty in case law. In Section 2, we consider the effects of these doctrines on judicial creativity and ideological bias in judge-made law.

1. Judicial errors and the doctrine of precedent

The different role played by case law among the sources of law and the substantial historical and conceptual differences between the doctrines of precedent in common law and civil law traditions are important determinants of judicial creativity.Footnote ² As a general rule, common law systems are bound by a single precedent, whereas most civil law and mixed legal systems require a continuous line of precedents before recognizing a rule of case law which courts should follow under a doctrine of jurisprudence constante.

Common law courts started to adhere to a stare decisis principle during the sixteenth century in matters of procedure and pleading and subsequently extended the principle to apply to matters of substantive law (Berman and Reid, 1964). At the end of the nineteenth century, under Bentham's positivist influence, the doctrine of stare decisis was more formally restated. Under stare decisis, a rule established in one case serves as binding authority to resolve future cases on similar legal matters (Evans, Reference Evans and Goldstein1987; Depoorter and Parisi, Reference Depoorter, Parisi, Rowley and Schneider2003).

In civil law and mixed jurisdictions, judicial decisions have a different effect on future similar cases. Besides the different role of case law among the sources of law, judicial decisions have a delayed impact on subsequent decisions in civil law courts. A civil law judge does not consider himself bound by a single past decision on the same point of law. Rather, as Lambert and Wasserman (Reference Lambert and Wasserman1929: 14) observes with respect to French law ‘[t]he practice of the courts does not become a source of law until it is definitely fixed by the repetition of precedents which are in agreement on a single point’. Comparable doctrines of precedent are adopted in other civil law jurisdictions, establishing a sort of ‘waiting period’ before past precedents can be recognized as authority.Footnote ³ The doctrine of precedent becomes particularly articulate in some mixed jurisdictions, accustomed as they are to have practitioners with common law training in their jurisdictions. Louisiana law, for example, provides that a precedent becomes a source of law when it has become ‘settled jurisprudence’ (jurisprudence constante) through an uninterrupted line of judicial decisions. Unlike stare decisis, the doctrine of jurisprudence constante links the recognition of a judge-made rule to the emergence of a consecutive line of decisions, consistently affirming a given rule or legal principle.Footnote ⁴

These different doctrines of precedent create different incentives for both ideologically driven and experimentally minded judges.Footnote ⁵ Posner (Reference Posner2010) shows that the attitude to influence the law by creating new precedents should be more frequent for circuit judges than for district judges, given the organizational structure of the judiciary system and the way it shapes the private incentives of the judges. He clearly points out that ‘appellate judges have a greater opportunity to influence the direction of the law than trial judges’ due to a lower reversal rate: only a very small fraction of cases is in fact reviewed by the Supreme Court (actually less than 1% of cases decided by circuit judges). In this context, the principle of stare decisis can be thought of as the principal constraint for the creativity of judges. Stare decisis limits the ability of a judge to influence the direction of the law since it links this ability to the degree that the innovative judge decisions will be treated as a precedent by other judges.

In civil law and mixed jurisdictions that adopt a jurisprudence constante principle, the incentives to innovate are further reduced, since no single decision is capable of becoming a ‘leading case’. As pointed out by Posner (Reference Posner2010), given the incentives provided by a civil law system to the judiciary, the output of a judge will be characterized by a greater level of uniformity in quality, with a lower variance and a lower level of judicial creativity. Linking a new legal rule to a single precedent fosters judicial creativity, but also increases the variance and the risk of error in case law. Posner identifies agency costs as a consequence of judicial independence. These may become a relevant source of social costs associated with the judicial system of common law jurisdictions, taking either the form of shirking or the form that Posner labeled as ‘sabotage’ – described as the tendency of the judge to redefine his judicial agenda in conformity to his personal ideology, disregarding the lawmakers’ objectives. In this light, judicial creativity may be interrelated to phenomena of ‘sabotage’, made easier by the higher degree of independence of common law judges. Among the costs borne by society in a common law system, Posner does not explicitly include those associated with the possibility that judicial creativity and independence may result in judicial errors, with a resulting welfare loss from bad legal precedents.

In the following, we develop a model to consider the effects of alternative doctrines of precedent on the social costs arising from judicial error and uncertainty in case law.

In a jurisprudence constante system, precedents cannot become binding until a consecutive line of decisions, say Nconsecutive decisions, has emerged. We can think of Nas a ‘waiting period’ before judicial decisions become binding case law. A system where a single precedent is capable of establishing a binding judge-made rule, can be viewed analytically as N =1. We define Nas the institutionally chosen ‘waiting period’. More specifically, Nrepresents the number of consecutive decisions from different judges above which the line of past decisions becomes binding case law. For the purpose of our analysis, we consider N as an institutional choice variable, and examine the factors that would influence the optimal choice of N*.

There are two costs associated with the choice of the critical threshold N. First, there is a cost associated with the possibility of erroneous decisions becoming ‘bad law’. Assume that α is the likelihood of ideological or erroneous decisions for an individual judge, and 1 − α is the likelihood that the judge makes the correct decision. We assume that courts are more likely than not to decide correctly, i.e. α < ½. The probability that bad law may emerge through a process of jurisprudence constante reflects the possibility that different judges make the same legal error in N consecutive decisions. This probability equals α^N if errors are independent, measuring the likelihood that N consecutive cases are judged erroneously.Footnote ⁶ Analytically, in a stare decisis system where a single decision is capable of setting a binding precedent, the probability of error in a judicial precedent is initially equal to the probability of error of an individual decision, α.

Let H be the present value of the harm occasioned when an erroneous precedent becomes law. Then α^NH represents the expected harm of a decision issued on a new legal issue or problem in the absence of a past legal precedent. As N increases, the emergence of a bad line of legal precedents becomes less likely, reducing the expected social cost of judicial errors.

The second cost associated with the choice of the threshold N is due to the legal uncertainty that characterizes the legal system when there is insufficient consistency in judicial precedents to establish case law. When a threshold N has not been met, a binding legal precedent does not yet exist. Hence, citizens and other legal actors face legal uncertainty. Legal uncertainty imposes social costs in terms of inability to plan and reduced ability to predict the outcome of legal cases. Legal uncertainty is the result of split case law and occurs with probability[1−α^N−(1−α)^N], i.e. when no binding precedent, good or bad as it may be, has yet emerged due to inconsistent judicial decisions.

Let D be the present value of the loss derived from legal uncertainty for any given legal issue. The expected social cost of legal uncertainty hence becomes [1−α^N−(1−α)^N]D. As the waiting period for binding case law, N, increases, the legal environment incurs in longer periods of legal uncertainty and the social costs associated with this uncertainty increases. For the sake of analytical comparison, note that in the limiting case of a stare decisis with N = 1 the cost of uncertainty after a first decision is null, since even a single decision constitutes authority as a legally binding precedent.

The social problem is to choose an optimal doctrine of precedent, N* such as to minimize the sum of the expected costs of legal error and legal uncertainty

The first-order condition for this minimization problemFootnote ⁷ would implicitly determine the optimal critical threshold N*

Comparative static results show that the optimal waiting period for legal precedents becoming binding case law N* is higher when the probability of judicial error α and/or the cost associated with the emergence of a bad case law H increase, and is instead lower when the cost of legal uncertainty, D, increases: i.e. dN*/dα > 0, dN*/dH > 0, and dN*/dD < 0.

Sabotage phenomena – as those discussed in Posner (Reference Posner2010) may increase the incidence of harm caused by wrongful decisions of judges inspired by the individual ideology of the judge in contrast with the lawmaker's goals. This will in turn affect more heavily common law systems, where judicial creativity may affect and ‘shape entire fields of the law’. It is therefore important that the fraction of active judges inclined to influence substantially the law remains small in a common law system.

Hence, one would expect shorter or null waiting periods, such as stare decisis doctrines, to be appropriate in legal systems that have a relatively higher valuation for legal certainty over judicial errors (N* decreases as the ratio increases, i.e. dN*/d(D/H) < 0). This is consistent with the observation that legal systems that are in their early phase of development (e.g., early Roman law, early English law, etc.), adopted and heavily relied on stare decisis for the consolidation of their young legal systems (higher value of given the importance to establish binding case law and to foster certainty in the law). As legal systems grow older or can rely on other sources of law, such as codified law, to provide legal certainty, the benefit of quickly establishing binding legal precedents becomes lower and legal systems can benefit for longer waiting periods. A lower value of indeed makes stare decisis less desirable, due to the relatively higher harm caused by judicial errors, once certainty in the legal system is sufficiently established.

2. Judicial creativity and ideology: the effects of anonymity

According to Posner (Reference Posner2010), judicial creativity can be regarded as a result of the way the organizational structure of common law system shapes the private incentives of the judges. The point was already made in an earlier work, where Posner (Reference Posner1994) argued that the incentives for judicial creativity and good performance are influenced by the judge's desire for recognition by peers and prestige in the legal community. A salient issue is therefore the method of evaluation of judges’ performance and the development of a ranking to boost reputational incentives among judges within the judicial organization. However, as pointed out by Posner (Reference Posner2010), evaluating a judge's performance through a quantitative measure is a difficult task, since any multidimensional ranking is subject to a high degree of subjectivity and arbitrariness. Given the varying weights that could be assigned to each dimension, judicial creativity may be fostered or penalized.

Posner does not take into account that another reason for which the incentives for judicial creativity may be undermined is constituted by the practice of anonymous decision making. Common law judges are often remembered for their decisions (many common law opinions are often associated with the name of the judge that wrote them). Although the number of per curiam decisions in common law courts has increased in recent years, signed opinions and concurring and dissenting opinions continue to characterize common law adjudication for appellate and supreme court cases.

These decision drafting techniques are not generally seen in civil law jurisdictions. Unlike common law judges, anonymity is often associated with the use of per curiam decisions in civil law courts. In civil law jurisdictions, panel decisions are generally taken on a simple majority vote, and judges on a panel have no opportunity to associate themselves with a majority or a minority view. Dissenting and concurring opinions are not admitted in most civil law countries. A judge who holds a minority view has no opportunity to dissent or to disassociate herself from the majority decision. Indeed, it is not infrequent that a judge may be drafting and signing an opinion on behalf of the court even though she had voted against that decision as a member of the court.Footnote ⁸ This practice severs the connection between a judicial decision and any individual judge that participated in its deliberation. This has the dual effect of shielding individual judges from the blame of a wrong decision, but also from the reputational benefits of being known for drafting correct or creative decisions. Furthermore, in the traditional decision of a civil law court, outcomes are often presented as the unavoidable result of existing legal rules. Comparative legal scholars frequently observed that civil law courts rarely discuss the doubts that they encounter in reaching a decision, apparently thinking it necessary to fulfill the expectation of legal certainty. Decisions are formulated as if they were dictated by law, applied in accordance with established canons of statutory interpretation and logic. The mechanistic nature of civil law adjudication hides much of the drama of adjudication and takes away much of the intellectual gratification associated with creative legal thinking.

An important effect of anonymity is that the views of individual judges are not transparent to the general public or to the community of lawyers. This has an effect on the possibility of adverse selection in the selection of disputes brought for litigation and on the resulting evolution of judicial remedies and judge-made law. The transparency of judges’ ideological views increases the opportunity for adverse selection, exacerbating the tendency of legal remedies to expand in number and scope of application. As pointed out by Fon and Parisi (Reference Fon and Parisi2003, Reference Fon and Parisi2006), the entire process of creation and change of legal precedent under stare decisis, is driven by a pervasive adverse selection mechanism. When judges differ from one another for their ideological views, plaintiffs decide whether to file suit based on the likelihood of success in the specific court. As shown in their model, the resulting choices of the plaintiffs create a strong bias toward filing in pro-plaintiff jurisdictions. New legal precedents will be created on the basis of the resulting subset of litigated cases, with a greater opportunity for judges with a pro-plaintiff ideology to innovate the law and create new legal precedents. Anonymous decision making, while undermining judicial creativity for the reasons pointed out by Posner (Reference Posner1994), will thwart the adverse case selection and weaken the bias in the evolution of the common law identified in the recent literature.