Edited by Gerrit De Geest
Chapter 11: Judicial organization and administration
[In: Volume 8, Chris William Sanchirico (ed) Procedural Law and Economics]
Economic analysis of substantive legal rules generally suppresses the adjudication of factual and legal disputes that a legal rule might engender. The nature of adjudication, however, will influence greatly both the content of the substantive law and the costs of dispute resolution. An understanding of the structure of adjudication is thus central to an understanding of the effects of legal rules on behavior and on the identification of socially desirable legal rules. In addition, adjudication is a complex task implemented through a set of central legal institutions that vary across time and jurisdiction.
Two distinct but related sets of questions about adjudication arise. First, what explains the structure of and variation among observed adjudicatory institutions? Court systems throughout the world and across time exhibit a number of similarities and differences. They are generally hierarchical, with appellate courts generally collegial. On the other hand, we observe variation in the selection, tenure, and dismissal practices across systems, differences in the degree of specialization, and differences in the nature of judicial output. Second, what explains the behavior of judges within a given court structure? Answers to this second set of questions, of course, will vary with the structure of the courts. After all, judges like other agents will respond to the incentives created by institutions. Moreover, different court systems require different behaviors from judges.
A vast literature in political science, economics, psychology, sociology and law have addressed these two sets of questions though far more attention has been devoted to questions of judicial behavior within institutions than to explanations of the structure of court systems. Even the literature in one of these disciplines defies succinct summary. In this review, and the review of “Appeal and Supreme Court” that addresses similar questions, I shall thus focus on rational choice approaches to courts and adjudication. Political scientists and economists introduced formal rational choice models of courts and adjudication in the 1980s. In the last decade, these models have proliferated in number and grown dramatically in power and sophistication, but there is as yet no agreement on the fundamentals of modeling.
The Encyclopedia’s division of topics into “Judicial Organization and p. 309Administration” and “Appeal and Supreme Courts” is not an analytically satisfying one. The two questions outlined above suggest a more natural division between the explanations of the structure of judicial institutions we observe and explanations of behavior within observed institutions. Unfortunately, most of the literature on specific elements of court organization asks how specific institutions of, for example, judicial selection, affect the behavior of judges rather than on the explanation of why the specific structure was chosen. The two chapters thus adopt a somewhat arbitrary division of the literature. This chapter surveys the literature on court organization other than hierarchy, a feature inherent in the idea of appeal. Chapter 2 on “Appeal and Supreme Courts” surveys the literature on hierarchy and on judicial behavior more generally.
This division is not airtight as much of the literature on court organization investigates the effects of court structures on judicial behavior. Discussion of this literature thus also requires some basic background about judicial behavior. Thus, I shall begin with a brief survey of the literature on judicial motivation, as different assumptions about motivation radically influence one’s understanding of court organization and of judicial behavior.
2. What do Judges Want?
Economic models of judicial behavior and of court organization require an assumption concerning the motivations of judges. This requirement presents the central challenge to the economic analysis of judicial behavior. Several approaches have been adopted, most of which are surveyed here. A related, supplementary discussion that emphasizes a distinction between “team” and “political” models of judges appears in Chapter 2 on “Appeal and Supreme Courts”.
2.1 Non-ideological Judicial Preferences
The literature has focused primarily on adjudication in common law jurisdictions. Suggestions have thus been attuned to the institutional structures of these courts rather than the very different organizational structure of courts in civil law jurisdictions. More seriously, the discussion of motivation has focused on the motivation of judges within the federal system of the United States. The institutional structure there is particularly spare; the judges have life tenure; their salary cannot be reduced; and they face limited prospects of promotion to a higher court. This focus has implied that lawmaking is the judicial function that has received the most motivational attention.
In this context, the literature has advanced three different types of answers to the question of judicial motivation. First, various authors, p. 310following Posner (1973, 2008), have claimed that judges, like all other economic agents, are motivated by self-interest. They value income, leisure, their reputations, power, and, perhaps in addition, they have preferences over policies and for their craft. A menu of preferential factors differs from a precise specification of a utility function. Nevertheless, this menu has spurred efforts to measure how judges respond to typical economic incentives such as promotion or retention in those jurisdictions in which judges face electoral scrutiny.
Posner (1973) sketched an approach to the problem of identification of the preferences of judges. He assumed that judges were self-interested and then briefly examined different incentive structures from which he inferred the underlying preferences of the judges. In particular, he compared the structure in many states in which judges lack life tenure and aspire to higher political office to the structure of the federal courts in the United States in which the judges have life tenure. These themes were developed in the subsequent literature.
Second, one can assume that judges meet their obligations as articulated in some jurisprudential theory of adjudication. Put differently, judges “follow the law”. In Posner’s terms, such judges place predominant importance on their craft or their reputation among peers. Team models, discussed in Section 2 of Chapter 2, attempt to ground this adherence to norms in a more economic framework.
Third, one can assume, as the predominant approach in political science does, that judges have preferences over “policies”. In political science, this approach has two variants. One, the “attitudinal model” (Segal and Spaeth 2002), adopts a behaviorist perspective, with the case serving as a stimulus and the judicial vote on the disposition of the case as the behavioral response dictated by an unobservable internal attitude. The second variant explicitly adopts an economic framework in which judges with preferences over policies act strategically. In fact, political science models generally assume that each judge has spatial preferences over a one-dimensional policy space, that is, each judge has an ideal policy X* and she prefers policy X to policy Y if and only if X is closer to X* than Y is. This formulation of preferences is discussed at greater length in the course of the review of the behavior of judges on collegial courts in Chapter 2.
A large number of studies consider the implications of these assumptions on preferences in the context of particular features of court organization. A few studies, however, address the question of motivation either theoretically or in a more direct, empirical fashion. I briefly review these few studies here.
Aranson (1990) surveys the political science literature that applies the spatial theory of voting to courts and compares it to law and economics p. 311approaches to judicial motivation. He suggests three competing views of judicial motivation. The first, that it is rule governed, parallels the first strand mentioned above and is discussed at greater length in Section 2 of Chapter 2. The second, that judicial motivation is rent redistributing, and the third, that it is wealth maximizing, describe systemic tendencies rather than the motivation of particular judges. One might understand “rent redistribution” consistently with the political model discussed in Section 3 of Chapter 2. Wealth maximization, by contrast, assumes that Posner’s claim that the common law maximizes wealth can be explained in terms of judicial success in pursuit of their common aim.
Landes and Posner (1980) argue, with the system of US federal courts in mind, that utility maximizing judges will primarily seek to maximize their power because their performance is too weakly linked to income and promotion for those concerns to have much effect on judicial behavior. They adopt the number of times a judge is cited in other cases as a measure of his power. They then argue that judges with higher salaries and more secure tenure will have greater power (and hence be cited more often) because higher salaries both attract more competent judges and reduce the judge’s incentives to distort his decisions and because longer tenure reduces turnover and the influence of politics. They provide an ingenious test of their hypotheses. They create two samples of common law appellate opinions rendered in 1950. First, they look at all 246 tort, contract and property cases decided by the federal appellate courts under their diversity jurisdiction. Second, they draw a random sample of 241 tort, property or contract cases decided by state supreme courts in 1950. The number of cases drawn from a specific state matches the proportion of federal cases decided under the law of that state. They then compare the citation rates of state and federal decisions in the two systems. They find only weak support for their hypothesis concerning higher salary and more secure tenure. Federal decisions are more likely to be cited in the state supreme courts of “other” states (i.e., states other than the one of which the federal court applied the law) than the decisions of state courts.
Cooter (1983) adopts a procedure similar to that of Landes and Posner (1979). He develops a theory of behavior of private judges and uses that as a benchmark from which to infer the preferences of public judges. He argues that both public and private judges would care about their reputations among other judges and the bar.
Higgins and Rubin (1980) address the concern for promotion somewhat more directly. They argue that judges have preferences over policy “discretion” and wealth; these preferences are conditional on their age. They argue that a judge’s ability to satisfy her preferences are constrained in two ways. First, the reversal rate depends on the judge’s policy preferences and p. 312the policy preferences of the higher court. Second, wealth depends on the reversal rate (because that affects each judge’s prospect of promotion) and age. They then derive a test for the relative effects of discretion and wealth.
They study two samples. The first sample consists of those active district court judges in the eighth federal circuit in 1974 who permitted the release of data on the total number of cases they decided in 1973 and 1974. The second sample consists of all active district court judges in the fifth federal circuit in 1966. They find that neither age nor seniority explained the reversal rates of the eighth circuit judges. They did find, however, that the estimated parameter on reversal rate had the predicted sign and was significant at the 10% level in a logit estimation of the probability of being promoted.
Greenberg and Haley (1986) argue, contrary to the conventional wisdom in general and to Posner (1985) in particular, that low judicial salaries are socially desirable because they signal a greater willingness to accept the non-pecuniary benefits of the judiciary; moreover, they argue that individuals who derive greater non-pecuniary benefits from judging make better judges.
Elder (1987) identifies two distinct mechanisms for monitoring judges: political and administrative mechanisms. These mechanisms create different incentives so that one should observe different behaviors in systems with different monitoring mechanisms. Without specifying the judicial objective function precisely, he nevertheless argues that judges deciding criminal cases will produce more trial verdicts under political monitoring than they would under administrative monitoring. He then tests this claim on 1977 data drawn from state criminal courts in 247 districts in seven states. His parameter estimates are consistent with his hypothesis.
Cohen (1992) follows up Elder’s approach. He argues that Elder implicitly assumes that each judge sought to maximize his preferences defined in terms of minimizing his workload and his reversal rate. He argues that these preferences also imply that, when the penalty range increases, a judge will increase the penalty of those defendants who request a trial more than they increase the penalty of those who plead guilty. He also argues that judges will be concerned about promotion and that this too will influence the pattern of sentencing. He then considers a sample that consists of all federal antitrust indictments from 1955 through 1980. In 1974, Congress increased the maximum penalties for antitrust violations from 50,000 to $100,000. He finds that promotion concerns are explanatory with respect to fines but not with respect to incarceration.
Cohen (1991) exploits the data generated by a “natural experiment” presented by the adoption of new sentencing guidelines by the United States federal courts. He examines 196 decisions by federal district courts p. 313that considered the constitutionality of the guidelines. He estimates a probit model of the probability of upholding the guidelines as a function of judicial ideology, caseload, promotion potential and the number of prior decisions for constitutionality. Promotion potential is measured by an index that reflects the (per district court judge) number of open seats on the appellate circuit. He finds that the parameters on workload and promotion potential have the predicted sign and are highly significant.
Katz (1988) adopts a behavioral approach. He assumes that judges decide cases on the basis of the arguments presented to it. Each party offers arguments in its favor and the court decides in favor of the plaintiff if the plaintiff’s arguments, in light of the “underlying” (or, perhaps, ex ante) merits of the case, outweigh the defendant’s arguments and some random error. He then shows that, when cases are more evenly balanced ex ante, expenditures of each party on litigation rise; and, if judicial decision is more random or variable, each party’s expenditures fall.
2.2 Policy Preferences
2.2.1 Measuring policy preferences
As already noted, most models of judicial behavior in the political science literature assume that judges have preferences over policies. Empirical tests of these models have used a variety of increasingly sophisticated, though not clearly more reliable or accurate, measures of these policy preferences. The creation and evaluation of these measures has spawned a large literature – for example Segal and Cover (1989), Epstein and Mershon (1996), Segal (2007), Cameron and Clark (2008), Martin and Quinn (2002, 2005), Epstein et al. (2007), Harvey (2006). Given their importance and prevalence in the literature, a brief discussion of them is warranted.
All these measures assume that policies lie on a one-dimensional continuum and that preferences are spatial. As a consequence, a judge’s preferences are completely described by the location of her “ideal point”, the policy that she thinks best. The measurement problem thus reduces to identifying this ideal point.
The measures fall into two classes. One set of measures uses data from during the judicial selection process and produces a relatively crude indicator of policy; the second set uses the voting behavior of the judge to construct a more refined measure. As most empirical work studies the behavior of the justices of the United States Supreme Court or, less frequently, other judges, the first set of measures looks to the process of appointment of the judge to extract a measure of the judge’s ideal point.
Federal judges in the United States are nominated by the President subject to the advice and consent of the Senate. Some measures thus p. 314attribute the “ideological” (or policy) position of the President to the appointed judge. This attribution may be more or less crude. It may simply identify the ideology in a binary way as either liberal or conservative, with Democratic presidents presumed to be “liberal” and Republican presidents presumed to be “conservative”. Or the ideology of the president may be measured by scoring the President’s announced position on roll call votes in the Congress which generates what is called a NOMINATE score that runs from –1 to 1. A refinement of these measures considers the composition of the Senate at the time of confirmation and attributes an ideological score to the judge on the basis of this composition of the Senate and the political affiliation of the President. A third measure, the Segal-Cover scores, relies on a content analysis of the editorials published in various newspapers in the United States prior to confirmation. Segal and Cover (1989) scale the ideology of justices on the unit interval, a score of –1 being most conservative and a score of 1 the most liberal. The Segal-Cover scores allow a more precise characterization of the “ideal point” of each justice and they have the additional advantage of relying on information that is exogenous to their voting behavior on the Court. A fourth measure, proposed in Cameron and Clark (2007), integrates the Segal-Cover score with a constructed NOMINATE score for each nominee.
The second set of measures constructs the ideal points of judges from their votes on cases. The most basic of these appears in Spaeth (2006), a comprehensive database of Supreme Court decisions that includes among its variables a dichotomous variable that designates each vote as “liberal” or “conservative”. Harvey (2006) criticizes this measure by considering votes assessing the constitutionality of statues and assigning the ideology of the enacting Congress to the statute.
A more refined scale is produced by considering the votes of judges on many cases. One seeks a set of ideal points that best predicts the patterns of voting, as defined by the set of coalitions that form on the Court. These ideal points are then scaled, with one end identified as conservative and the other as liberal. Bailey and Chang (2001) and Martin and Quinn (2002, 2005) proceed in this fashion. Farnsworth (2007) provides a non-technical assessment of this literature, while Bafumi et al. (2005) offer a technical introduction and assessment.
Clark and Lauderdale (2010) adopt a different approach. They consider the opinions in a given area of law – in their case, religious expression – and count the favorable and unfavorable citations by the justices of those citations. This yields an ideology score for each justice that is in fact correlated with the Martin Quinn score.
Brace et al. (2000) have produced a measure of p. 315ideology for justices of supreme courts of several states. Their measure uses a measure of citizen ideology for justices of supreme courts in which such judges are elected and a measure of the ideology of the political “elite” in states in which supreme court justices are appointed.
2.2.2 Are judges ideological?
As noted earlier, a significant part of the American political science literature on adjudication seeks to determine the effects of the “ideological” views of the judges on judicial decisions. This literature in general implicitly accepts the political model of adjudication and it generally finds at least some influence of “ideology” on outcome, where the judge’s ideology is measured either by her political affiliation or by the affiliation of the appointing President. In this section, I consider some recent studies within the legal-economic framework.
The first study, Eisenberg and Johnson (1991), examine 118 federal district court opinions and 66 federal circuit court opinions in racial discrimination cases decided under the fourteenth amendment to the US Constitution. These opinions constituted all opinions on this issue published between 7 June 1976 and 6 February 1988. Eisenberg and Johnson found no effect of ideology, measured either by party of judge or of appointing President, on outcomes at either the district or appellate level. They attempted to evaluate the effects of the selection of cases for trial on their results by comparing trial success rates and success rates on appeal in the class of cases that they studied to these success rates in other classes of cases.
Ashenfelter, Eisenberg and Schwab (1995) (“AES”) also found no effects of ideology. In a clever research design, they studied 2258 federal civil rights and prisoner cases filed in three federal districts courts in fiscal 1981. Unlike most prior studies, AES examine the effects of ideology not only on cases with published opinions but on all case dispositions. AES determined whether a case settled and, if it did not settle, which party prevailed. They also collected data, including party and party of appointing President, on each of the 47 different judges who sat on some case in the sample. AES then analyzed this data by district. This district analysis permitted them to exploit the federal practice of random assignment of cases to judges. Any observed differences in behavior across judges could be attributed to differences in judges rather than differences in cases. They found that judges had relatively little effect on case disposition. Moreover, “ideological” variables did not explain the small effects observed.
Revesz (1997) studied the industry and environmental challenges to EPA rulemaking in the 1970s and the late 1980s through early 1990s. This sample of cases has two features that make it a nearly ideal sample for analysis. First, all such challenges were heard in the Court of the Appeals p. 316for the District of Columbia. Second, virtually all challenges to EPA rulemaking are appealed. Given the legal context, there is no opportunity for the agency to settle with dissatisfied litigants. Following AES, Revesz focused on time periods in which a large number of judges had continuous tenure on the Circuit; he restricted his analysis to these judges. He measured the judge’s policy preferences by the party of the appointing President, assuming that Republicans would favor industry challenges and disfavor environmental challenges relative to Democrats. For the late 1980s through early 1990s, he observed that ideology had a clear effect on industry challenges and a more ambiguous effect on environmental challenges. (A different statistical test, however, would likely reveal a stronger ideological effect in environmental challenges.)
Pinello (1999) conducts a meta-analysis of 84 studies that use party as an indicator of judicial ideology. He concludes that ideology explains 38% of the variation in decisions, though the explanatory effect of ideology varies across institutional settings. Thus, he argues that it is less successful an explanatory variable in state courts than in federal courts and that, within the federal system, it is most important in explanations of Supreme Court decisions.
Sisk and Heise (2005), however, argue that Pinello overstated the significance of ideology on Supreme Court decisions. In their analysis, ideology explains only 7% of the variation of federal judicial votes. Songer, Sheehan and Haire (2000) had earlier provided a similar estimate of the size of the effect for the federal intermediate courts of appeal.
Lindquist and Klein (2006) attempt to determine the role of both attitudinal or ideological factors and of jurisprudential factors in Supreme Court decision making. They model the role of “law” in terms of the Supreme Court’s responsiveness to lower court rulings. To do so, they study cases that involve conflicts among the circuits. They find that the Court is more likely to adopt the majority position on a conflict; less likely to adopt positions that have attracted dissents or confusing concurrences below, and more likely to endorse the view of a prestigious lower court judge.
3. The Functions of Courts
3.1 What Do Courts Do?
We may answer this question at several different levels. We might ask it within a “legal” framework or more “globally”. The first phrasing asks what differentiates courts from other political institutions such as legislatures and the executive. The second asks how courts facilitate societal success or welfare.
3.1.1 p. 317The legal functions of courts
Adjudication resolves disputes. Dispute resolution itself consists of at least four different tasks. First, courts must determine the facts of the dispute. Second, they apply the law to those facts. Third, the court may, in some instances, enforce its judgment against one of the parties. Finally, in some instances, the court may have to compel one or more of the parties to submit to the jurisdiction of the court.
In some legal systems, the judicial function is limited to these four tasks; in others, courts also make law. In civil law theory, for example, courts simply apply the law announced by legislatures to resolve disputes. In practice, civil law courts do make law, but the theoretical fiction has consequences for the development of the law. Legal theorists in common law countries, by contrast, contend that courts resolve disputes through the application of law, but also promulgate new legal rules; courts thus play a lawmaking function as well.
This difference between civil and common law perceptions of adjudication, as well as various structural differences between the two systems, presents a challenge to economic analysts which has yet to be answered. The sequel in this review and in Chapter 2 tries to draw out these differences in specific areas.
Analyses of lawmaking and error-correction are generally addressed in a specific context and are addressed in various subsections below or in Chapter 2.
3.1.2 The social functions of courts
In the last ten years, a vigorous debate has arisen about the role of courts in economic development. La Porta et al. (1998) triggered this debate with its claim that societies with common law judicial systems led to superior economic performance of financial markets than societies with civil law systems. This claim was then generalized to economic growth in Mahoney (2001). Moreover, legal origin has been used as an instrument for institutional differences across countries in studies such as Acemoglu et al. (2001). A more comprehensive survey of the literature appears in Levine (2005).
Three questions arise. First, is the claim true? Do common law systems in fact cause societies to grow more quickly than civil law systems? Second, if the claim is true, through what causal mechanism does the difference in legal regimes cause differential growth? Of course, to answer this question – indeed to answer the first – we need some characterization of the differences between the two systems. Third, what explains the emergence of different legal systems in England and France? I briefly address these questions in turn.
For several reasons, it is difficult to characterize the differences between p. 318civil law legal systems and common law systems. Conventionally, commentators point to four sets of institutional differences. First, in theory at least, courts in common law legal systems are a source of law, while courts have no such power in civil law legal systems. Second, common law processes are adversarial, while civil law systems are inquisitorial. Crudely, this distinction generally means that common law judges play a less active role in the production of the facts needed to resolve the dispute and that common law advocates play a larger role in developing legal arguments. Moreover, common law trials are primarily oral, while civil law trials are primarily written. Often, and almost always at their origins, juries found the facts in common law adjudication, while judges did so in civil law adjudication. (While juries continue to play an important role in the United States, they no longer find facts in many other common law jurisdictions.) Third, the judiciary in common law legal systems is drawn from the practicing bar, with lawyers ascending to the bench in the middle of or late in their careers, while the judiciary in civil law legal systems are government bureaucracies, with entry generally at the beginning of the jurist’s career. Finally, Djankov et al. (2003) claim that civil law legal systems are more formalistic than common law ones.
Unfortunately, both systems are quite complex. An historical perspective aggravates this complexity. Much of the literature in this area identifies legal origins with the imperial power that governed a particular region four or five centuries ago. Common law institutions today differ dramatically from those of the sixteenth century. For instance, common law adjudication in the United States was much more formalistic in the nineteenth century that it is today. Civilian legal systems did not yet exist, as they are a creation of the French Revolution and the Napoleonic wars. The structure of legal systems within Europe prior to the spread of the Napoleonic codes may have varied greatly. While it may be true that the countries coded for common law legal systems generally had better economic performance than those coded for civil law legal systems (or Scandanavian legal systems), attributing the observed difference to the structure of the court system requires a giant inferential leap. Berkowitz et al. (2002), for example, argue that it is not the origin of the legal system that matters but the manner in which it was received; countries in which the shift was imposed fared worse than those that freely adopted the system.
Through what mechanism does the structure of the legal system operate? Most accounts of development contend that economic development requires secure property rights and effective enforcement of contract law. Why should common law systems provide more protection of property rights and better contract enforcement than civil law systems? Glaeser p. 319and Shleifer (2002) argue that common law systems are more decentralized than civil law systems largely because of the common law institution of the jury. They trace this divergence back to the origins of each legal system in the twelfth and thirteenth centuries. Klerman and Mahoney (2007) argue persuasively that, in fact, at their origins, the common law courts emerged in Britain as an institutional means of centralized law enforcement. The creation of royal courts served to displace adjudication from local, manorial courts. They note that the common law courts were centralized in the sense that both the judiciary and the bar were quite small; in France, by contrast, they note that the size of the judiciary presented severe monitoring problems. Moreover, they argue that juries did not serve a lawmaking function. Rather juries served as a collective form of testimony. Finally, they contend that, as jurors were chosen by a royal official, they were not in fact outside of royal control.
Lee (2003) offers a more plausible story. He characterizes the structure of a legal system with five variables: the status and power of judges (measured by the tenure of the judges of the highest court and by whether judicial systems serve as a source of law) and the structure of adjudication of “administrative claims” (or claims against the government) (measured by the tenure of the judges of the highest adminstrative court and by whether the administrative claims are adjudicated by a separate court system). He finds that unitary adjudication of administrative claims and judicial decisions as a source of law have the greatest effect on growth.
What explains the different evolution of the English and French legal structures? Glaeser and Shleifer (2002) argued that conflict between the barons and the King in Britain led to the institution of the jury trial in the thirteenth century in Britain. On their account, by contrast, the balance of power in France was different, with the King able to impose a centralized legal system. Klerman and Mahoney (2007) again contest this explanation. They argue rather that the English Civil War (and the subsequent Glorious Revolution) and the French Revolution were the central events that triggered the divergence in the structure of the common law and civil law legal systems.
Roe (2007) offers a different rebuttal of Glaeser and Shleifer. He attacks not the claim about the origins of the jury and the inquisitorial system; rather, he argues that juries are not central to the protection of investors’ rights. Roe makes two distinct arguments. First, he notes that Delaware, a primary source of investor protection in the United States, decides corporate cases in a chancery court that operates without a jury. Second, he argues that Britain did not generally transplant the jury system to its colonies as that might have interfered with imperial policy in those locations.
3.2 p. 320Public versus Private Adjudication
Adjudication is commonly considered a quintessential function of government. In recent years, however, there has been a substantial privatization of at least some of these functions through a growing use of arbitration and other “private judges”. This phenomenon has prompted scholars to study the extent to which government must supply judicial functions.
Landes and Posner (1979) provide one of the earliest inquiries into the choice between private and public provision of adjudicatory services. Working within a common law framework, they identify the judicial functions as both dispute resolution and rule generation. They argue that only the dispute resolution function is suitable for (partial) privatization. Some public provision of dispute resolution might be necessary because (1) enforcement of judgments might require public authority and (2) in some instances, public authority may be necessary to compel one or more parties to submit the dispute to adjudication in the first place. A private market in adjudicatory services, however, would meet all other requirements of a dispute resolution system. They argue that a competitive market would produce competent and impartial judges because both qualities would be necessary to induce all parties to a dispute to consent to adjudication of their dispute by a given judge. Cooter (1983) offers a similar argument.
Landes and Posner, however, argue that private provision of rules will not, in general, be desirable. Rule generation is a public good and a private judge who announces a rule will not capture all the benefits that the announcement of a rule creates. Therefore, in a system of private adjudication, rules will be undersupplied. In addition, competing judges may generate competing sets of rules. They do not, however, explain why rules cannot be adequately supplied by legislatures.
Shavell (1995) considers the choice of private dispute resolution in the context of an established public system of courts. He addresses two questions: why would parties resort to private dispute resolution and when is it socially desirable? He distinguishes ex ante invocation of private dispute resolution, in which the parties agree prior to the emergence of a dispute to resort to private resolution of their dispute, from ex post invocation of it, in which the resort to these private methods occurs after the dispute arises. He argues that ex ante invocation is socially desirable for three reasons: (1) it may lower the costs and risks of dispute resolution; (2) it may create better incentives to perform because of the greater accuracy of private resolution; and (3) it may reduce the number of litigated disputes. He argues that ex post invocation of private dispute resolution is not desirable.
4. p. 321The General Organization of Courts
Courts are complex organizations that form only one branch of government. This section addresses global questions about court organization. The first concerns the relation of courts to other branches; specifically I review the literature on judicial independence. This literature largely considers independence at a macro or systemic level. In Chapter 2, I consider the strategic implications for the behavior of judges deciding specific cases. The second question concerns the appointment and tenure of judges.
In this section, I discuss some of the issues raised by this variety of court organization. I begin, however, with a discussion of the question of judicial independence.
4.1 Independence of the Judiciary
Judicial independence is much praised and almost universally urged upon developing countries as a necessary pre-condition for economic and political development, though the experience of China suggests that this claim is false. The various definitions offered are not wholly consistent. Moreover, judicial independence is instrumentally, rather than intrinsically, valuable; we value judicial independence because of the benefits it provides. Three interrelated benefits are sometimes attributed to judicial independence. (1) An independent judiciary provides impartial adjudication of private disputes; (2) an independent judiciary prevents expropriation of wealth (or private property) by the government; and (3) an independent judiciary assures the civil and human rights of its citizens. Notice that these functions might be exercised in a jurisdiction without judicial review of legislation on a constitutional basis. British courts, for example, had no power of judicial review over parliamentary statutes, prior to the establishment of the European Union (EU).
Economic analysts have devoted much attention to explanations for, and implications of, “independent” judiciaries. An “independent” judiciary is one that is free from external, primarily, political influence. This conception of independence is somewhat at odds with the ascription of a political motivation to judges themselves because independence is generally seen as guaranteeing a more “objective” resolution of disputes. Though the literature has not attended much to this specific problem, there are suggestions that a more general model of constitutional structure might justify an independent branch of politically motivated judges.
Most industrialized countries assert the independence of their judiciaries, but the structures that guarantee “independence” differ greatly. In the federal system in the United States, for example, judges have life tenure and their (nominal) salaries cannot be decreased during their lifetime. In many states of the United States, however, judges serve for a term of p. 322years; moreover, in many states, they may be elected in either partisan or non-partisan elections. In both the state and federal systems, as in common law jurisdictions generally, judges are, however, drawn from the general bar. In many civil law countries by contrast, law graduates choose to enter practice or the judiciary at the outset of their career; a judicial bureaucracy then creates incentives for that judge. The insulation of that bureaucracy from “normal” politics then determines the extent of judicial independence. This institutional variation across jurisdictions permits a comparative study of judicial independence. As a consequence, studies of independence, unlike studies in other areas of judicial organization, have been primarily comparative in nature.
One might investigate judicial independence in at least two ways. One might consider it functionally and ask what features of polities lead to court systems that exhibit judicial independence. Alternatively, one might examine particular institutional structures – a particular selection, promotion and tenure scheme, for example – and whether it functions acceptably. Thus much of the discussion below concerning selection, promotion and tenure and of judicial review might be understood as discussions of judicial independence.
Landes and Posner (1975) offered the earliest explanation for the emergence of judicial independence. Their argument rests on the claim that an independent judiciary will, in statutory interpretation, enforce the original legislative understanding. Individual legislatures accede to this practice because they want to increase the time a statute prevails. Landes and Posner then claim that the degree of independence should increase with the size of the jurisdiction because larger jurisdictions provide broader scope for rent-seeking. They then attempt to test their theory on data concerning 97 statutes declared unconstitutional by the United States Supreme Court between 1789 and 1972.
Ramseyer (1994) extends the analysis of Landes and Posner (1975) by drawing on a comparison of Japan and the United States. Ramseyer defines judicial independence as the extent to which politicians do not manipulate careers of sitting judges. He asks why some politicians provide an independent judiciary and others do not. He claims that a political structure will provide for an independent judiciary if (i) politicians believe that elections will continue indefinitely and (ii) politicians believe that their prospects of continued victory are low. He then does three case studies: the United States, which satisfies both antecedent conditions and has an independent judiciary; contemporary Japan, which satisfies the first condition but not the second and does not have an independent judiciary; and imperial Japan, which satisfied condition (ii) but not condition (i) and did not have an independent judiciary.
p. 323Stephenson (2003) places Ramseyer’s argument within a formal model. In his model, parties compete for control of the legislature. The winning party announces a policy. Then the courts rule on the legality of the policy. The legislature may then either ignore the ruling or amend the policy. Judicial independence exists if an equilibrium exists in which the legislature acquiesces in judicial rulings of illegality. Stephenson proves that an equilibrium exhibiting judicial independence exists if the partisan politics are sufficiently competitive, the judiciary uses a “moderate” rule in determining the illegality of legislative actions, and the competing parties are sufficiently risk averse and care enough about the future.
Hanssen (2004b) provides a formal model for the Ramseyer model. The model extends Maskin and Tirole’s (2004) analysis of electoral accountability. There are three periods. In period 0, the ruling party chooses between two court structures – a dependent judiciary in which the ruling party can choose judges each period and an independent judiciary in which the initial party chooses judges for life. “Politics” occurs in periods 1 and 2. In each period, a “judge” must choose one of two policies a or b. There are two parties A and B with A preferring a to b and B preferring b to a. In period 1, the constitutional designer has the power to appoint a judge who will then implement a policy. There are three parameters to the model: the probability p that the party in power in either period can determine the policy preferences of his appointees, the probability x that the party in power in period 1 will be re-elected in period 2, and the distance d between the two policies a and b.
Hanssen proves two results. First, as the probability x of re-election increases, the higher the quality p of screening must be to make judicial independence desirable. That is, as elections grow less competitive, judicial independence becomes less desirable. Second, as the distance d between policies increases, the lower the quality p of screening must be to make judicial independence desirable. Hanssen then tests his model using data from the states of the United States.
Ramseyer and Rasmusen (2003) provide an extensive study of the case of contemporary Japan. They study the careers of roughly 790 judges who entered the Japanese judiciary between 1959 and 1968. As a measure of political ideology, they use membership in a leftist organization, the Young Jurist League. As those who become judges in Japan have foregone lucrative careers as private attorneys, Ramseyer and Rasmusen assume that the judges have preferences for successful careers, understood as rapid promotion through the bureaucratic ranks and postings to desirable cities. After controlling for quality of judge (with various measures of academic quality), they find that ideologically left judges have somewhat less successful careers than non-members of the Young Jurist League. Ramseyer p. 324and Rasmusen also identify classes of cases in which the ruling party in Japan, the LDP, would have a particular interest. They then compare the career paths of judges who rule against the LDP interest with the career paths of those who rule according to the LDP interest. They conclude that, in some instances, the judges who rule against the LDP interest have less successful careers. Ramseyer and Rasmusen (2006) extend their analysis to the era after the collapse of the political dominance of the LDP.
Haley (2007) argues that Ramseyer and Rasmusen’s argument has a fatal gap. The Japanese judiciary, as Ramseyer and Rasmusen acknowledge, is governed by a Secretariat that makes all promotion and posting decisions. Haley contends that Ramseyer and Rasmusen have not excluded the hypothesis that the Secretariat acts disinterestedly to further the careers of the “best” judges.
Cooter and Ginsburg (1996) also deploy comparative data to study the question of judicial independence. They work with a different conception of independence than Ramseyer. Ramseyer’s definition referred to the structure of appointment, pay, promotion and tenure in the judiciary. Cooter and Ginsburg have a more substantive view of independence; they look to the courts’ ability to make law that diverges from the views of the legislature. They argue that the degree of judicial independence will depend on both political and constitutional features of a society. In particular, they argue that societies in which a cohesive party dominates politics will be less likely to have judicial independence. On the other hand, the more “legislative vetoes” that the constitution builds into its political process, the more independence the courts have. They asked experts in comparative law to rank the daringness of the judiciary of various countries. They then regressed daringness on the number of vetoes and the duration of the governing coalition. Despite the small sample size, the parameters on both independent variables had the predicted sign and were statistically significant; moreover, they explained a substantial part of the variance.
As noted at the beginning of this section, the meaning of judicial independence is clear. Feld and Voigt (2003, 2004) define judicial independence in terms of the absence of short-term political pressure on judges. They distinguish two “types” of judicial independence: de jure and de facto. They measure de jure judicial independence, however, in terms of structural features of the court system. They focus on the highest court within a state and then consider twelve structural features, such as constitutional standing for the highest court, the nature of the appointments process, the tenure of judges, and the extent of the Court’s control over its docket. They measure de facto judicial independence in terms of eight “performance” variables, such as the actual average term length of justies, the frequency of removal, the constancy of judical salary, the frequency with p. 325which legal rules governing the court are changed. These two measures led to dramatically different rankings of court systems. The United States, for example, ranked 30th out of 75 on the de jure scale, while Switzerland ranked 67th. Moreover, the two measures led to very different rankings; no country ranked in the top ten under the de jure measures ranks in the top ten under the de facto measure. In the earlier study, they found no significant correlation of growth in GDP with the de jure measure of judicial independence, while the de facto measure of judicial independence was significantly correlated with GDP growth. In a larger data set used in their later article, they found GDP was significantly correlated with constitutionally specified selection procedures and powers.
Iaryczower et al. (2002) define judicial independence as “the extent to which judges can reflect their preferences in their decisions without facing retaliation measures by Congress or the President.” In their model, the court evaluates the constitutionality of a piece of legislation and then, if the court strikes it down, the President, when he has Congressional support, may sanction the court. In equilibrium, the court upholds the legislation when its preferences correspond to the President’s or when the President has control over the Congress. They test this model on data from the Argentine Supreme Court and find that the court is more independent than conventionally believed.
Helmke (2005) provides a model that challenges this definition. In her model, justices are forward-looking and they act in an uncertain world. In rendering decisions, they must determine whether the current government will remain in power; if it is likely to fall, then the justices have an interest in deciding according to the interests of the successor government.
Klerman and Mahoney (2005) suggest that judicial independence means that judges can make decisions based “on the law without fear of reprisal from the executive or the legislature.” They then consider several structural features of court systems as indicia of judicial independence, among them tenure, a non-manipulable appointments procedure, and a relatively fixed jurisdiction. In an empirical study of eighteenth century England, they find that increased judicial independence, particularly as measured by “tenure during good behavior”, was strongly correlated with growth of GDP.
4.2 Appointment and Tenure of Judges
The appointment, promotion and tenure prospects of judges vary greatly across court systems. The literature often takes the federal system of the United States as paradigmatic, but it is, in fact, rather unusual. In the federal system, judges have life tenure and limited prospects of promotion from the trial bench to the courts of appeal and even less prospect p. 326of promotion to the Supreme Court. Nevertheless, as the prior discussion on judicial motivation suggested, these modest prospects of promotion apparently affect judicial behavior.
The 50 states of the United States offer a wide variety of distinct appointments and tenure policies. Some states have institutions that parallel those of the federal government. Many elect judges; these elections are themselves conducted in a variety of ways. Some are contested, partisan elections. Others are contested non-partisan elections. Some have periodic, unconstested retention elections. Still other states have nomination procedures that appoint judges for some term.
In most of these states, as in most common law countries, the line between the judiciary and the bar is relatively fluid, with practicing lawyers often entering the judiciary for greater or lesser stays sometime in the middle of their career.
Civil law systems have a very different structure. In these systems, the judiciary is part of the state bureaucracy. Individuals thus choose at the outset of their careers whether to practice law or to serve in the judiciary. These judges face bureaucratic incentives quite distinct from the variety of incentives found in common law systems. The discussion of judicial independence in Japan in the prior subsection illustrated at least one consequence of this different incentive structure. Guarnieri (2004) provides a description of the court systems in these countries.
In this section, I briefly survey some of the literature on appointments, promotion and tenure. Again, this literature has been heavily influenced by US institutions, so much of it focuses on electoral mechanisms. The studies here might equally have been discussed in the section studying judicial motivation.
Substantial evidence suggests that judges respond to electoral pressures, at least in prominent cases. Shepherd (2007) studied the effects of retention elections on state supreme court justices. She found that judges who faced retention decisions by Republicans were more likely to vote for business litigants over individual litigants, for employers in labor cases, for defendants in medical malpractice and other tort cases and for the state in criminal cases. Judges facing retention decisions by Democrats had the reverse tendencies. The tendencies were greatest when judges faced partisan elections.
Helland and Tabarrok (2002) study the effect of partisan elections on the level of tort awards. Their study contrasts both awards in states with partisan elections to those with non-partisan elections and to awards by elected state judges versus appointed federal judges. They find that judges in states with partisan elections made higher awards than both federal judges and judges elected in non-partisan elections.
p. 327Choi, Gulati, and Posner (2007) measure the comparative performance of appointed versus elected judges. They measure the quantity of output by the number of opinions and quality by the number of citations. They find that, though appointed judges write higher quality opinions, elected judges write more opinions. Total influence is thus roughly the same. Moreover, they find that elected judges exhibit as much judicial independence as appointed ones.
Hanssen (2002) attempts to explain why some states appoint judges and others elect them. He argues, on the basis of empirical data, that merit appointment procedures are correlated with increased litigation. As lawyers prefer more litigation to less, he claims that this increase explains their support for merit appointment.
5. The Organization of Adjudication
Adjudication involves a number of distinct tasks. Courts must find facts. They must apply the law to the facts; if no law exists, they must make the law. Once judgment is rendered, courts may be asked to enforce that judgment. Each of these tasks might be further subdivided. Finding facts requires a trial procedure with rules governing what evidence may be presented, in what order, and with what burden of production of evidence and burden of proof. These questions of trial procedure are considered elsewhere in this volume.
These tasks might be organized in radically different ways that exhibit greater or lesser degrees of division of labor and specialization of functions. All functions might be entrusted to a single court or they might be divided among them. Such a division of labor might occur in a number of different ways. In common law countries, since the merger of law and equity, courts usually have general jurisdiction. A single court may hear cases almost regardless of subject matter. Even in this system, however, some specialization exists. Within state courts, for example, generally family matters (issues concerning divorce, custody of children, delinquency) and probate have distinct courts of first instance, though appeals are taken to the appellate courts of general jurisdiction. In the federal system, many administrative issues, such as immigration status, claims under the social security act, and labor disputes, are heard first by administrative judges, with appeal sometimes permitted from these administrative courts to the appellate courts of general jurisdiction. A distinct appellate court exists to hear intellectual property cases.
Civilian legal systems, by contrast, are much more specialized. In France, for example, distinct court systems exist for hearing cases arising from government administration and for ordinary civil and criminal matters. In addition, a separate body hears constitutional questions. p. 328Courts hearing ordinary civil matters generally do not have general jurisdiction. Commercial cases and labor cases, for example, are heard in separate fora.
5.1.1 Case and controversy requirement
Courts generally resolve disputes, but the dispute must generally be, in some way, “live” and “real”. In the United States, this requirement for the existence of a real dispute is embodied in the “case or controversy” requirement of Article III of the Constitution. The requirement, however, is not a peculiarity of United States federal courts. With the exception of some constitutional courts, such as those of France and Germany, that have jurisdiction to issue opinions on the constitutionality of legislation prior to its enactment, most judicial systems require some similar trigger. Why should this be so? And who should be allowed to press a claim concerning a “real” dispute?
Jensen, Meckling, and Holderness (1986) address this latter question of who should be allowed to press a claim in a live dispute. They argue for limited rules of standing because, they claim, a liberal standing rule increases the costs of engaging in any rule-governed transaction.
Landes and Posner (1994) address the first question concerning the justification of a requirement for a concrete case to trigger the adjudicatory power. They extend to the question of the case or controversy requirement a model of legal advice developed by Shavell (1988, 1992) and Kaplow and Shavell (1992) to investigate questions concerning the use of attorneys. Landes and Posner focus on the distinction between type 1 and type 2 errors – the wrongful attribution of liability to defendants versus the wrongful excusal from liability of defendants. They argue that this distinction can be deployed to explain much of the observed pattern of exceptions to the case or controversy requirement. Specifically, they note that anticipatory adjudication, in violation of the concreteness requirement, increases both errors. They then examine in detail the institutions of declaratory judgments, res judicata, advisory opinions, and preliminary injunctions.
Stearns (1995, 1996) argues that the case or controversy requirement serves to restrict the occurrence of cycling that plagues institutions of social choice. For a fuller discussion, see the section on collegiality in Chapter 2 on “Appeal and Supreme Courts”.
5.1.2 Specialized courts versus courts of general jurisdiction
Over what disputes should a court have jurisdiction to decide? Two methods for defining the subject matter jurisdiction of courts predominate. First, and most p. 329commonly, a court might have jurisdiction over any dispute that arises within a specified geographic area. Virtually all court systems consist primarily of these courts of general jurisdiction. Second, a court might have jurisdiction over disputes with a specified subject matter. Notice that these organizational patterns could be applied to courts of first instance (that find facts as well as apply law) only, to appellate courts only, or to both courts of first instance and to appellate courts. Moreover, in a system with specialized courts of first instance but appeal to courts of general jurisdiction, one could limit appeal to a single general court of jurisdiction or one could allocate non-exclusive jurisdiction to a number of appellate courts. In the United States, one may understand the administrative law structure as establishing administrative agencies as specialized courts to determine the facts and make initial legal rulings that are then appealed to courts of general jurisdictions. Thus most labor disputes, welfare disputes, and immigration disputes are first heard in administrative agencies and then, if necessary, appealed to the federal courts of appeals. In these instances, each of the federal courts of appeal has jurisdiction to hear appeals from these specialized tribunals. Many environmental disputes, by contrast, may only be appealed to the US Court of Appeals for the DC Circuit. Finally, there are some specialized courts of appeal such as the tax court and the federal circuit for Court of Appeals for the Federal Circuit which has jurisdiction over patent and other intellectual property disputes. What is the appropriate way to allocate jurisdiction among courts?
Posner (1985) argues that specialized appellate courts are likely to be more ideological than courts of general jurisdiction because judges on specialized courts are likely to be more focused on the subject matter of their jurisdiction and hence more likely to be sensitive, and responsive, to controversy.
Revesz (1990) analyzes the desirability of vesting appellate authority over administrative agencies in specialized courts. His analysis emphasizes the effect of the nature of the appellate court’s jurisdiction on legislative ability to control agencies. He argues that review by specialized courts reduces the effectiveness of congressional delegation to administrative agencies. He develops a simple principal-agent model in which Congress is the principal and the administrative agency the agent. They have different policy preferences because commissioners in agencies have terms that do not correspond to the terms of the commissioners and because there may be a divergence in preferences between Congress and the President who appoints the head of many agencies. Congress has three mechanisms for the control of agencies: it may overrule particular decisions, it may exercise oversight through one or more committees, and it may alter the agency’s budget. Congress, however, might also use the courts to monitor p. 330the agency. Revesz argues that a court of general jurisdiction is a better monitor.
5.1.3 Constitutional courts
In some political systems, courts have the power to declare legislative enactments and actions of the government executive unconstitutional. This power of judicial review may, as in the United States, be exercised by the courts of general jurisdiction. Or the power may, as in Germany or France, be exercised by a distinct body governed by its own jurisdictional and standing rules. In some instances, adjudication must occur prior to the legislation coming into force and only specific public actors may challenge the constitutionality of a rule. In other instances, standing may be broad and challenges may be brought at any time.
5.2 Court Congestion
Comparative law scholars have often noted the variation in “litigiousness” across countries. There is little economic literature that seeks to explain this cross-cultural variation, but there is a substantial literature analyzing the causes of “congestion” in the courts of one notably litigious society, the United States.
Virtually every proposal to reduce court congestion in the United States federal courts recommends the abolition of diversity jurisdiction, which grants federal courts the authority to resolve disputes between citizens of different states, even when they involve only questions of state law. One justification for diversity is that it prevents discrimination against out-of-state residents. Goldman and Marks (1980) tested this claim by looking at two samples of attorneys drawn from the US District Court for the Northern District of Illinois in 1976. They randomly sampled 200 attorneys tied to specific questions and asked them their reasons for litigating in federal court. They had a 62% response rate. In addition, they randomly sampled 205 attorneys from the law division of Cook County District Court in 1976; this court had a $15,000 amount a controversy minimum so that the cases within its jurisdiction were reasonably comparable to those within the jurisdiction of the federal court. This survey had only a 37% response rate. Only 40% of the attorneys in federal court listed local bias as a reason for choosing federal court. Attorneys drawn from state court cases were asked to consider a hypothetical case identical to the one litigated but in which their client was an out-of-state resident. Roughly 25% said they would file in federal court. Local bias thus had very little influence on the choice of forum.
Noam (1981) attempts to assess the social cost of court congestion by calculating the effects of congestion on criminal sentences and then on the crime rate. He argues that plea bargain reached between prosecutor p. 331and defendant will depend on the caseload of the court. The higher the caseload per judge, the lower the average sentence. Moreover, he argues that the per capita crime rate is a function of the average sentence. Using simple specifications of these functional relations, he derives an equation that represents the marginal effect of additional resources devoted to the criminal courts. He then estimates his equation using FBI data on crime rates for four types of crimes against property in the District of Columbia. This estimation yields very high marginal returns to increased investment in the court system.
To resolve disputes, courts must determine the facts. A substantial portion of procedural rules govern the fact-finding process. Each of these rules influences the structure of the judicial system because each influences the cost of litigation relative to the cost of settlement and to self-help remedies. (The literature on evidence and the choice between settlement and litigation are summarized elsewhere in this volume.) Some procedural rules play a more central role in the organization of court systems. In this section, I examine the structure of trials and the use of juries, a characteristic element of many but not all court systems.
6.1 Sequential versus Unitary Trials
Most disputes present more than one factual issue for resolution. Should these be resolved simultaneously or sequentially? In the United States, factual issues are generally resolved simultaneously, but often the question of liability and the question of remedy are decided sequentially, sometimes by different fact-finders.
Landes (1993) presents the primary investigation of this issue. He modifies his own, early model of the choice between settlement and litigation (see Landes 1971) to analyze several issues concerning the effect of trial structure on the settlement rate. The analysis rests on the insight that a sequential structure to the litigation reduces the expected cost of litigation. Hence, plaintiff’s incentive to sue increases, which implies that the number of lawsuits will increase. Landes argues further that sequential trials reduce the probability of settlement because they narrow the range of acceptable settlements.
Many legal systems divide the law-finding (or law-applying) and fact-finding functions of trial courts. These systems delegate fact-finding to a jury, generally a group of lay individuals chosen more or less randomly to decide one case. This procedure raises several questions.
6.2.1 p. 332Jury selection
Bowles (1980) compares the cost of jury trials in Britain to the cost of a system of fact-finding by a three-judge court. The cost of a jury trial consists primarily of the production foregone by persons serving on the jury. Bowles notes that high wage earners will attempt to avoid jury service more vigorously than low wage earners. He concludes that a jury trial will be less expensive if the cost of a judge is more than three times the cost of a juror. He does not, however, correct for the speed of the trial.
Martin (1972) also estimates the social cost of the jury system in the United States. He first estimates the occupational distribution of jurors on the assumption that juror days of service are distributed identically to jurors. He then multiplied by the median daily wage rate for each occupation and estimated the social cost at $233 million in 1958 dollars. He then compares the cost of two systems of jury selection: random selection and a “keyman” procedure in which lists are constructed through consultation with community leaders. He finds that random selection is significantly less expensive because community leaders are more likely to draw jurors from high wage occupations than random selection. Finally Martin argues that voluntary service would reduce costs even further.
6.2.2 Jury size and jury voting rules
Models of jury size and the effects of jury voting rules must specify how juries deliberate and vote. Many models assume no deliberation. Moreover, until recently, the literature assumed that each juror voted conscientiously; her vote expressed her view of the guilt or innocence of the defendant. When one assumes in addition that each juror is more likely than not to decide correctly and that jurors’ judgments are independent of each other, the Condorcet jury theorem applies and one can easily show first, that a unanimity rule minimizes the probability of wrongful convictions; second, that majority rule maximizes the probability of a correct decision; and third, that, as the number of jurors increases, the probability of wrongful conviction under a unanimity rule decreases towards zero and the probability of a correct decision under majority rule increases towards one. Recently, however, analysts have introduced models of strategic voting by jurors and these models yield dramatically different results.
Austen-Smith and Banks (1996) and Feddersen and Pesendorfer (1996) show how radically the assumption of strategic behavior by jurors undermines the Condorcet jury theorem. When a juror acts strategically, the information aggregation feature of the Condorcet jury theorem disappears because each juror now decides how to cast her vote in light only of those instances in which her vote will be pivotal. As a consequence, as Austen-Smith and Banks show clearly, a juror’s vote may not reveal her p. 333actual view concerning the case. Feddersen and Pesendorfer show, more dramatically, that the probability of wrongful conviction may be higher under a unanimity rule than under a different voting rule. In their model, moreover, non-strategic voting is not an equilibrium. Guarnaschelli, McKelvey and Palfrey (2000) offer some experimental support for the conclusion that jurors vote strategically, but they find unanimity leads to more correct decisions.
These results, however, are sensitive to the specification. Coughlan (2000), for example, analyzes two variants of the Feddersen and Pesendorfer model; in one, there is a possibility of a hung jury and, in the second, the jurors may deliberate. In these environments, unanimity does minimize the probability of wrongful conviction and non-strategic voting is an equilibrium.
These models raise important questions about the appropriate way in which to model juries. These questions parallel those presented in the study of collegial courts, discussed in Chapter 2, where team models in which judges act in a fashion somewhat analogous to the assumption underlying the naive Condorcet jury theorem models contend with political models in which judges act strategically.
In the United States, until the 1970s, juries typically consisted of twelve individuals and required unanimity to render a verdict. Then, the United States Supreme Court ruled that neither the number twelve nor unanimity were constitutionally required in state criminal proceedings. These decisions spurred research into the importance of these requirements.
Klevorick and Rothschild (1979) provide a model of jury deliberation in order to determine whether non-unanimous juries or unanimous juries of fewer than twelve jurors would yield different verdicts than the “standard” twelve-person unanimous jury. Their analysis uses a stationary Markov model to provide a dynamic model of the majority persuasion hypothesis, derived from Kalven and Zeisal (1966), that the final verdict is the same as the majority position on the first ballot. They offer both a discrete and continuous model of jury deliberation. The cores of the models are identical. The jurors enter the jury room with a view on the merits and cast an initial ballot. The vote then changes incrementally as time passes, with one person changing her vote at each instant until unanimity is reached. The transition probabilities are given by the current vote: the probability of one more vote for the plaintiff equals the percentage of jurors who favor plaintiff’s case. In this model, one can calculate both the expected number of jury ballots until unanimity is reached (conditional on the initial vote) and the probability that a non-unanimous jury will agree with a unanimous one. Klevorick and Rothschild show that a move to a requirement of ten votes for judgment rather twelve always alters the probability of conviction by less than p. 3340.0055. Their result, of course, depends critically on the manner in which majority pressure operates in the jury room. When they adopt a different assumption concerning transition probabilities, they calculate that the move to ten for judgment may lead to as many one in six of the cases being decided differently. Klevorick and Rothschild also show that this shift substantially reduces the expected number of ballots prior to judgment.
p. 335Schwartz and Schwartz (1992, 1995, 1996a) offer a somewhat different approach to jury decision making. The earliest paper addresses three questions concerning the decision rules of juries. First, why do juries fail to reach a verdict? Second, what verdict will a jury reach when it may convict for lesser included offenses or when multiple offenses are litigated simultaneously? Third, how, if at all, will the jury decision rule alter the charges filed by prosecutors? To this end, they assume that a defendant might be charged with any number of counts within some interval. They consider two regimes. In one, the prosecutor chooses one count in the interval, and the jury must choose among three outcomes, conviction on that count, acquittal, or no decision. In the second regime, the prosecutor chooses two counts, one entailing a lesser punishment than the other, and the jury chooses among four outcomes, acquittal on both counts, conviction on the lesser count, conviction on the more severe count, or no decision. In the event of no decision, a retrial before a different jury occurs. Each jury consists of four people, chosen randomly from a population. Each potential juror is an expected utility maximizer, characterized by spatial “preferences” for punishment of the defendant; that is, each juror has an ideal outcome that represents the crime for which she believes the defendant should be convicted. The simple model implies that a defendant always prefers a non-unanimous, super-majority rule to unanimity because his expected verdict is lower under the non-unanimous rule. This result occurs because, though the probability of conviction is greater under non-unanimity, the probability of acquittal rises more rapidly. When jury may convict on a lesser included offense, the analysis is more complex and less clear-cut.
Schwartz and Schwartz (1995) extend the analysis of their prior article in the context of the liability and punishment phases of capital offenses in the United States. They argue that the decision process for fact-finding has three elements: the voting rule, the jury selection process, and the characterization of the set of outcomes among which the jury may choose. They argue that, in a multi-stage fact-finding process, the voting rule must be the same in each stage in order to avoid the jury at one stage nullifying the law governing the jury at a different stage. In addition, only reciprocal voting rules will avoid hung juries (where a rule is reciprocal if “conviction” requires c of n votes, then “acquittal” requires n – c of n votes).
Schwartz and Schwartz (1996a) carry the argument of the prior two papers further. They argue that any voting rule should satisfy at least two properties: (1) it should be decisive; and (2) it should satisfy the “one person/one vote” criterion. This second property does not eliminate any anonymous voting rule, whether majority, supermajority, or submajority. (Super and submajority rules would violate neutrality.) The first criterion identifies majority rule as the unique decisive voting rule when there are only two possible outcomes. To satisfy the desire to minimize wrongful convictions that the unanimity rule is said to promote, Schwartz and Schwartz argue that the size of the jury should be increased perhaps to fifteen.
7. Concluding Remarks
Economic analysis of judicial organization and administration, though it has greatly increased understanding, is only in its infancy. No question has received an exhaustive treatment and many have not been examined at all.
This area suggests a multitude of comparative questions, most of which have not received any attention at all. They are ripe for analysis.